Letter to the editor

Surg Endosc (1998) 12: 1091 Springer-Verlag New York Inc. 1998

Telesurgical laparoscopic cholecystectomy

On March 3 1997, a telesurgical laparoscopic cholecystectomy was performed for the first time in history at the St. Blasius hospital in Dendermonde, Belgium. The device used was the Mona from Surgical Intuitive, Mountain View, California, USA. After clearance from the local ethical committee and after informed consent had been obtained, the patient, a 72-year-old woman with a body mass index of 42 kg/m2 was put under general anesthesia with endotracheal intubation. Four 10-mm and one 5-mm trocar cannulas were inserted according to Dubois technique [1]. One 10-mm cannula harbored a straight-looking optical system connected to a three-dimensional camera system. Another 10-mm cannula was placed to the left of the umbilicus and used for clip placement only. The 5-mm trocar contained a probe for liver retraction. The two remaining active ports were connected to two fully mobile mechanical arms attached to the siderails of the operating table. They harbored two articulated tools (end effectors): a grasper and an electrocautery hook commanded by the surgeon sitting at a working console approximately 15 feet away from the patient. As the surgeon watched the three-dimensional image of the operative field, he manipulated two handles that transmitted impulses to and from the end effectors via a computer interface. Sensory input and downscaling of the surgeons motions 4 to 1 were secured. The procedure was successfully performed in 82 min, and the patients recovery was uneventful. The use of a computer interface between surgeon and patient has many more advantages than the ability to perform operations on a site remote from the patient. The translation of virtual manipulations into commands to robot arms allows the surgeon to deal with the three most significant shortcomings of laparoscopic surgery: (a) the reduction in degrees of freedom, (b) the lack of tactile feedback, and (c) the impaired dexterity attributable to the use of long and rigid instruments introduced through a fixed point in the abdominal wall and manipulated in an often awkward position [2]. With the present master-slave system, seven degrees of freedom are acquired because the end effectors have an additional articulation (wrist) inside the abdominal cavity, perfectly mimicking all motions as they are performed by the surgeon. Moreover, the system is capable of reducing (downscaling) the virtual manipulations of the operator, hereby eliminating minor flaws such as physiologic tremor. This, combined with the ergonomically optimal position, the visual immersion, and the sensory feedback on the surgeons part, allow for a more intuitive, nearly perfect surgical approach, hence more precision, and thus improved safety for the patient. More complex endoscopic operations such as microanastomosis in the cardiovascular field will now become feasible with the same accuracy as in the open chest procedure. References
1. Dubois F, Berthelot G, Levard H (1991) Laparoscopic cholecystectomy: historic perspective and personal experience. Surg Laparosc Endosc 1: 5257 2. Wapler M (1995) Medical manipulators: a realistic concept? Minim Invasive Ther 4: 261266

J. Himpens1,2 G. Leman1 G. B. Cadiere2

1

Correspondence to: J. Himpens

Department of Surgery Saint Blasius Hospital 50, Kroonveldlaan Dendermonde 9200, Belgium 2 Department of Surgery Saint Pierre University Hospital Brussels, Belgium

Surg Endosc (1998) 12: 10611063

Springer-Verlag New York Inc. 1998

Transthoracic induction of a hiatal hernia in domestic swine

F. J. Brody, J. Hunt, J. Sackier
Department of Surgery, Suite 3B, The George Washington University Medical Center, 2150 Pennsylvania Avenue NW, Washington, DC 20037, USA Received: 22 August 1996/Accepted: 29 January 1997

Abstract Background: With the common performance of laparoscopic Nissen fundoplication for gastroesophageal reflux disease, there is renewed interest in the pathophysiology and potential histologic consequences of hiatal hernias. However, in vivo model exists that both reliably reproduces the hiatal hernia and is amenable to subsequent laparoscopic repair. Methods: A transthoracic approach was used to induce a hiatal hernia surgically in female James pigs (50160 kg; n 5). Results: Hiatal hernias were successfully induced in all pigs and verified with barium swallow, endoscopy, and/or laparoscopy. Laparoscopic reduction and Nissen fundoplication were subsequently completed on each animal on postoperative day 30. One postoperative death occurred on postoperative day 4 after thoracotomy. Conclusions: We describe the induction of a hiatal hernia via a transthoracic approach in domestic swine. The hiatal hernia is amenable to subsequent laparoscopic repair, enabling surgeons to acquire the technical skills required to correct this defect in the laboratory. To our knowledge, this is the first report of a reproducible model of a transthoracically induced hiatal hernia that allows subsequent laparoscopic repair. We suggest that in addition to refinement of surgical skills, our model may provide new information to researchers regarding the potential indications for antireflux procedures, as well as the natural history and appropriate management of hiatal hernias. Key words: Gastroesophageal reflux Pig Laparoscopic Nissen fundoplication Hiatal hernia Animal model

medically without surgical intervention. The remaining 20%, however, require surgical intervention to repair the hiatal hernia and prevent pathophysiologic reflux [1]. Surgical interventions such as Nissen fundoplication are successfully performed laparoscopically. To our knowledge, there is no in vivo model of a transthoracically induced hiatal hernia that can subsequently be repaired laparoscopically. We sought to test the hypothesis that a hiatal hernia could be transthoracically induced for subsequent laparoscopic repair in an animal model.

Materials and methods

Five female James (50160-kg) pigs were used. The animals were anesthetized with intramuscular injections of 1.0 mg/kg of xylazine, 6.0 mg/kg of telazol, and 0.01 mg/kg of atropine. The pigs were intubated and mechanically ventilated with anesthesia maintenance using 2.0% isoflourane. Selective intubation of the right mainstem bronchus was performed with a flexible bronchoscope. The animals were positioned on their right side, and a left posterolateral thoracotomy was performed through the sixth intercostal space. A rib spreader was applied for retraction. The inferior pulmonary ligament was divided, and the left lower lobe of the lung was gently retracted superiorly to expose the distal esophagus and aorta. The pleura overlying the esophagus was incised exposing the vagus nerves. A Penrose drain was secured around the esophagus, and vagus nerves and cephalad traction was applied. The phrenoesophageal membrane was visualized at the gastroesophageal junction, and circumferential dissection with Metzenbaum scissors and electrocautery was performed to incise this structure thoroughly. The dissection completely freed the esophagogastric junction from its diaphragmatic attachments and allowed retraction of the lower esophagus and gastric fundus into the posterior mediastinum. To allow elevation of the gastric fundus into the chest, the most proximal short gastric vessel was ligated. The gastric body was secured to the posterior aspect of the left crus of the diaphragm with one suture. An 18 French red rubber catheter was inserted in the left chest through a stab wound above the eighth rib for evacuation of the pneumothorax. The chest wall musculature was reapproximated in three layers and the skin closed with a continuous subcuticular suture. A Heimlich valve was applied to the distal end of the red rubber catheter, and the pig was weaned from ventilatory support after discontinuation of anesthesia. The Heimlich valve was removed and a 60-cc syringe was attached to the red rubber catheter to ensure complete resolution of the pneumothorax. The red rubber catheter was removed and the pursestring suture secured. The animal was extubated and given postoperative analgesia consisting of buprenorphine (0.1 mg/kg intramuscularly.) The pig was restarted on a liquid diet 6 hours postoperatively, and a regular diet was resumed 12 hours postoperatively. On postoperative day

Approximately 33% of individuals with hiatal hernias suffer from gastroesophageal reflux disease [6, 7]. The vast majority of these individuals (80%) are successfully treated
Correspondence to: J. Sackier

1062

Fig. 1. Under flouroscopy, the posteriorly displaced hernia was visualized (black arrow). The pig was positioned in the left lateral decubitus position. Fig. 2. With endoscopy, the hiatal hernia was clearly seen under direct vision (white arrow).

25 an upper gastrointestinal (UGI) series was completed with barium, and on day 30 the animal underwent esophagoscopy and gastroscopy followed by laparoscopic reduction and repair of the hiatal hernia. The pig was subsequently killed.

Results The UGI series was completed with the animal in the left lateral decubitis position and documented the posterior location of the hiatal hernia (Fig. 1). Hiatal hernias were documented in all animals with either an UGI series or endoscopy (Fig. 2). According to the radiographic and endoscopic findings, these hernias were all type II paraesophageal hernias. A standard Nissen fundoplication with crural closure was completed on all animals without complications.

Discussion Previous reports regarding hiatal hernias have focused on the pathophysiologic and mechanical relationship of hiatal hernias to gastroesophageal reflux. After performing diaphragmatic crural myotomies in cats through an abdominal incision, Mittal and colleagues [4] performed postoperative esophageal manometry and pH monitoring. They showed a higher frequency of gastroesophageal refux after crural myotomy than with control animals. Laparoscopic repair was not attempted in this model. Patterson and Kolyn [8] demonstrated esophageal shortening in an opossum model after the induction of esophagitis secondary to intraluminal perfusion of hydrochloric acid. Esophageal shortening was associated with a significant decrease in lower esophageal sphincter pressures. On the basis of their study, the authors suggest that esophagitis with subsequent esophageal shortening may contribute to the development of hiatal hernias. No attempt was made at operative intervention and repair of this abnormality. We could not find any reports of an animal

model with a hiatal hernia or gastroesophageal reflux amenable to laparoscopic repair. We believe that the one postoperative death of a 160-kg pig was attributable to respiratory compromise secondary to a fractured rib that occurred intraoperatively. The fractured rib probably impaired normal ventilation throughout the animals enormous thoracic cavity, resulting in respiratory failure. On the basis of this complication, we use 40 to 50-kg pigs with careful application of the rib spreader intraoperatively. Since making this change, we have had no subsequent morbidity or mortality. The barium study seen in Fig. 1 details the posterior location of the hiatal hernia. We did not observe gross esophageal reflux (GER) on barium examination. Although GER was not grossly documented on the UGI series, we plan to investigate potential reflux through pH monitoring and esophageal manometry in subsequent trials. With the addition of physiologic data from future trials, we hope to correlate the anatomic role of the crura, phrenoesophageal membrane, and the fundoplication to GER in our model. To our knowledge, this is the first animal model that allows transabdominal surgical repair of a hiatal hernia via a laparoscopic approach. In addition to the anatomic accuracy of the hiatal hernia, our model may provide a physiologic paradigm of gastroesophageal reflux disease including esophagitis and dysplasia. We suggest that the described animal model may be used to address the following subjects: (1) the selective use of antireflux procedures with paraesophageal hernias [5], (2) the natural history of lowgrade esophageal dysplasia after medical therapy or an antireflux procedure [2], and (3) the anatomic relationship of the esophagogastric junction in a hiatal hernia and its pathophysiologic association with gastroesophageal reflux disease [3]. Regardless of the outcome of future protocols with pH monitoring and esophageal manometry, this model will continue to provide excellent anatomic representation for training purposes.

1063 Acknowledgments. The authors thank Drs. Michael Salem and Paul Lin for their editorial help with this manuscript and Ethicon Endo-Surgery staff for their generous support. mechanism of gastroesophageal reflux in cats. Gastroenterol 5: 740 747 Myers GA, Harms BA, Starling JR (1995) Management of paraesophageal hernia with a selective approach to antireflux surgery. Am J Surg 170: 375380 Ott DJ, Glauser SJ, Ledbetter MS, Chen MY, Koufman JA, Gelfand DW (1995) Association of hiatal hernia and gastroesophageal reflux: correlation between presence and size of hiatal hernia and 24-hour pH monitoring of the esophagus. AJR Am J Roentgenol 165: 557559 Ott DJ, Ledbetter MS, Chen MY, Koufman JA, Gelfand DW (1996) Correlation of lower esophageal mucosal ring and 24-hour pH monitoring of the esophagus. Am J Gastroenterol 91: 6164 Patterson WG, Kolyn DM (1994) Esophageal shortening induced by short-term intraluminal acid perfusion in opossum: a cause for hiatal hernia? Gastroenterol 107: 17361740

5.

References
1. Barlow AP, Hinder RA, Demeester TR (1989) Principles of 24 hour pH monitoring and its clinical applications. Gastroenterology 98: A27A29 2. Edwards MJ, et al. (1996) The rationale for esophagectomy as the optimal therapy for Barretts esophagus with high-grade dysplasia. Ann Surg 223: 585591 3. Mittal RK (1993) Hiatal hernia and gastroesophageal reflux: another attempt to resolve the controversy. Gastroenterol (1995) 105: 941943 4. Mittal RK, et al. (1993) Effect of crural myotomy on the incidence and

6.

7.

8.

Surg Endosc (1998) 12: 10821084

Springer-Verlag New York Inc. 1998

Laparoscopic cholecystectomy under epidural anesthesia in patients with chronic respiratory disease
K. G. Pursnani,1 Y. Bazza,1 M. Calleja,2 M. M. Mughal1
1 2

Department of Surgery, Chorley and South Ribble District General Hospital, Preston Road, Chorley, Lancashire PR7 1PP, England Department of Anaesthesia, Chorley and South Ribble District General Hospital, Preston Road, Chorley, Lancashire PR7 1PP, England

Received: 11 July 1997/Accepted: 28 October 1997

Abstract Background: Laparoscopic cholecystectomy (LC) has become firmly established as a procedure of choice for gallstone disease. The procedure usually necessitates general anaesthesia and endotracheal intubation to prevent aspiration and respiratory embarrassment secondary to the induction of pneumoperitoneum. There is a paucity of data in the literature on the procedure being performed under regional (epidural) anaesthesia, especially in patients with coexisting pulmonary disease and pregnancy, who are deemed high risk for general anaesthesia. We report our preliminary experience with LC using epidural anaesthesia in patients with chronic obstructive pulmonary disease (COPD). Methods: We performed LC in six patients (one man and five women), with a median age of 56 years (range, 3874), under epidural anaesthesia over an 8-month period. All patients were ASA grade III/IV and the mean FEV1/FVC was 0.52 (range, 0.40.68), due to chronic asthma (two cases) and COPD (four cases). They were admitted a day prior to surgery for pulmonary function tests, nebulisers, and chest physiotherapy. An epidural catheter was introduced at T10/ 11 intervertebral space, and a bolus of 0.5% Bupivacaine was administered. Depending on the patients pain threshold and the segmental level of analgesia achieved, incremental doses of 2 ml of 0.5% Bupivacaine along with boluses of intravenous 100 mcg Alfentanil was given to each patient. The patients were breathing spontaneously. No nasogastric tube was inserted, and a low-pressure (10 mmHg) pneumoperitoneum was created. LC was performed according to the standard technique. Results: All the patients tolerated the procedure well and made an uneventful postoperative recovery. Median operating time was 50 min; average length of hospital stay was 2.5 days (range, 24). The epidural catheter was removed the morning after the operation. Only one patient required postoperative opioid analgesia. Two patients complained of persistent shoulder tip pain during surgery and required inCorrespondence to: M. M. Mughal

traoperative analgesia (Alfentanil). There was no change in the patients cardiorespiratory status, including pO2 and pCO2, and no complications occurred either intra- or postoperatively. Conclusions: LC can be performed safely under epidural anaesthesia in patients with severe COPD. Intraoperative shoulder tip or abdominal pain does not seem to be a major deterrent and can be effectively controlled with small doses of opioid analgesia. Key words: Laparoscopic cholecystectomy Epidural anaesthesia Chronic respiratory disease

For decades, the management of symptomatic cholelithiasis in high surgical risk patients has remained contentious. Since its advent in 1988, laparoscopic cholecystectomy has become firmly established as a procedure of choice in the management of symptomatic cholelithiasis [7, 14, 15]. The procedure usually necessitates general anaesthesia and endotracheal intubation to prevent aspiration and respiratory embarrassment secondary to the induction of pneumoperitoneum. There have been several case reports of successful laparoscopic cholecystectomy performed under epidural anaesthesia in pregnant patients during the 3rd trimester [4, 5, 17] and patients with cystic fibrosis [6, 11], who are deemed high risk for general anaesthesia. However, little has been reported about the possibility of performing the procedure under regional (epidural) anaesthesia in patients with significant pulmonary disease. Cholecystitis and cholelithiasis in patients with chronic obstructive pulmonary disease (COPD) pose several management problems for the surgeon. Because of the high risk associated with the induction of anaesthesia in patients whose pulmonary status is compromised, surgery is sometimes delayed or avoided. It is generally agreed that the condition is best managed conservatively and that surgical intervention should be reserved for patients who fail to re-

1083

spond or develop complications. Laparoscopic cholecystectomy (LC) usually necessitates general anaesthesia and endotracheal intubation to prevent aspiration and respiratory embarrassment secondary to the induction of pneumoperitoneum. Furthermore, in patients with COPD, CO2 pneumoperitoneum could have detrimental effects secondary to splinting of the diaphragm and systemic CO2 absorption. With the advent of LC and anaesthetic techniques such as epidural blockage, we have another option that may be safe for many of these patients. We report our preliminary experience with laparoscopic cholecystectomy using epidural anaesthesia in patients with COPD. Patients and methods
Six patients (one male, five female), with a median age of 56 years (range, 3874), underwent laparoscopic cholecystectomy under epidural anaesthesia over an 8-month period in our institution. All patients were American Society of Anaesthesiologist (ASA) grade III/IV due to chronic asthma (two cases) and COPD (four cases). Spirometric studies (performed with a respiradyne pulmonary function monitor) showed a mean tidal volume of 300 ml, forced vital capacity (FVC) of 1.94 L (50% of predicted value for age, weight, and height) and peak expiratory flow rate (PEFR) of 146 ml/s. Forced expiratory volume in 1 s (FEV1) was 1.02 L (30% of predicted value), and FEV1/FVC 100 was 52% (range, 4068%). All patients were admitted a day prior to surgery for pulmonary function tests, nebulisers, and chest physiotherapy.

Table 1. Cardiorespiratory function after LC performed under epidural anaesthesia in six patientsa Mean FEV1 (L) FVC (L) PEFR (ml/s) EtCO2 (%) SpO2 (%) Heart rate (beats/min) Blood pressure (mmHg)
a

Preop 1.02 (0.921.28) 1.94 (1.882.12) 146 (122180) 4.5 (3.85.4) 86 (8292) 92 (84102) 122/80

Intraop 4.8 (4.15.5) 90 (8894) 100 (94110) 110/76

Postop 1.12 (0.911.36) 1.98 (1.872.10) 142 (128188) 4.6 (4.15.2) 87 (8594) 94 (90102) 118/84

All values are mean (range), except blood pressure, where only median values are given.

Epidural anaesthesia
The patients were premedicated with 10 mg diazepam 1 h before the procedure. A 20-gauge epidural catheter via a Tuohy-Huber needle was introduced at T10/11 intervertebral space; the tip of the catheter was advanced 3 cm cephalad beyond the tip of the needle. After a test dose of 3 ml, an 8 ml bolus of 0.5% Bupivacaine was injected. Depending on the segmental level of anaesthesia achieved, incremental doses of 2 ml of 0.5% Bupivacaine were administered to reach the desired segmental block. In most cases a block up to T4/5 was achieved, as determined by temperature sensation using ethyl chloride spray. Depending on the patients pain threshold and the amount of shoulder tip pain they experienced, boluses of intravenous 100 mcg Alfentanil were given to each patient during laparoscopic cholecystectomy. Along with measuring heart rate and arterial pressure, the monitoring also included ECG, pulse oximetry, and expired capnography.

time was 40 min (range, 3060), and the average length of hospital stay was 2.5 days (range, 24). The epidural catheter was removed the morning after the operation. Only one patient required postoperative opioid analgesia. Two patients complained of persistent shoulder tip pain during surgery and thus required intraoperative analgesia (Alfentanil). Two patients elected to view the procedure, thereby producing an enhanced sham surgical response. Both these patients were communicative throughout the procedure. They did not express any distress or discomfort. They made an extremely rapid postoperative recovery and were discharged home the following day. They gave a positive response on direct questioning whether they would have a similar procedure done again under the same conditions. There was no change in the patients cardiorespiratory status including SpO2 (oxygen saturation) and EtCO2 (end-tidal CO2), and no complications occurred either intra- or postoperatively (Table 1).

Discussion In addition to changes to mucociliary transport associated with anaesthetic agents, abdominal surgeryparticularly upper abdominal surgeryis associated in normal individuals with adverse effects on respiratory mechanics such as functional residual capacity (FRC), vital capacity (VC), tidal volume (TV), and closing volume [8, 13]. Because mucociliary clearance is an important pulmonary defense mechanism against infection, general anaesthesia using inhalational or intravenous agents may be deleterious to the patient with COPD undergoing surgical procedure. Furthermore, it has been shown that patients with COPD are at risk of developing pulmonary complications after upper abdominal surgery [1, 9]; therefore, these patients may benefit from laparoscopic surgery. The goal of anaesthesia management in these patients should include avoidance of anaesthetics that depress mucociliary transport, provision of postoperative pain relief adequate to prevent deterioration of respiratory mechanics, and ambulation as early as possible. Epidural anaesthesia fulfills all of the above criteria and aids in the quick and uneventful postoperative recovery of these patients. Langenberg et al. [12] evaluated the feasibility of performing LC under locoregional anaesthesia in 25 patients without any evidence of respiratory disease. The procedure

Laparoscopic cholecystectomy
Laparoscopic cholecystectomy was performed according to the standard technique [7, 14, 15]. Sequential compression device stockings were used in all patients, and 1.2 g Augmentin (coamoxiclav/clavulinic acid) was administered intravenously intraoperatively. The patients were breathing spontaneously, no nasogastric tube was inserted, and a low-pressure (10 mmHg) pneumoperitoneum using CO2 was created. A 10-mm trocar was inserted via the umbilical port to accommodate the laparoscope and visualise the peritoneal surface of the abdominal cavity. Trocars were placed in the anterior axillary line (5 mm), midclavicular line (5 mm), and midepigastrium (10 mm) just beneath the costal margin. The gallbladder was grasped through the 5-mm ports. Dissection, clip application, and electrocauterisation were performed through the 10-mm epigastric port. The gallbladder was dissected free from the liver bed and the neck delivered through the epigastric port. In some cases, the gallbladder was decompressed using a small suction instrument and removed intact from the abdomen.

Results All the patients tolerated the procedure well and made an uneventful postoperative recovery. The median operating

1084

was successful in 20 patients (80%) using epidural anaesthesia, allowing satisfactory surgical conditions and rapid postoperative recovery. In another study [10], 43 of 45 patients (95%) (without COPD) underwent successful LC under epidural anaesthesia and intravenous propofol sedation. They also reported excellent operating conditions and exceptionally pleasant postoperative recovery. The present study is the only report of LC being performed successfully under epidural anaesthesia in patients with COPD. The pneumoperitoneum should progress slowly, and a low-pressure (<10 mmHg) insufflation is preferred. Most patients complained of mild to moderate brachioscapular pain, possibly due to irritation of the subdiaphragmatic peritoneum or traction of the mesentery. The symptoms appeared to correspond to the level of innervation of diaphragm (C4) by the phrenic nerve. The pain responded well to administration of intravenous Alfentanil in doses of 11.5 mcg/kg. A pertinent intraoperative manouvre that helps to minimize the shoulder tip pain is gentle liver retraction and minimal irrigation during the procedure. Another important operative principle is to ask the patient to withhold breathing voluntarily for a few seconds during clipping and cutting the cystic artery/duct in order to prevent inadvertent injury to other vital structures during respiration. In our group of patients, continuous low thoracic epidural anaesthesia provided excellent anaesthesia and muscle relaxation without the use of muscle relaxants for surgery, as well as excellent pain relief in the postoperative period. Hence, it was possible to restore respiratory function to the preoperative level more rapidly without the use of narcotics [3, 16]. In high surgical risk patients such as pregnant patients (3rd trimester) and patients with severe COPD, the laparoscopic procedures can be optimised using the laparolift or abdominal wall lift or gasless laparoscopy [2, 5]. The technique would be an excellent alternative to pneumoperitoneum in patients with severe COPD who are CO2 retainers, where not only the increased intraabdominal pressure but also systemic CO2 absorption prove detrimental to their already compromised respiratory status. The abdominal wall lift with zero pressure pneumoperitoneum is similar in many respects to the traditional open technique of cholecystectomy [2, 5]. However, to use the technique under epidural anaesthesia might prove difficult, because the wires inserted to lift the abdominal wall may encroach on the lower cervical dermatome, sometimes even requiring paralysis of the diaphragm. Additional studies and more experience are required to determine the feasibility, risks, and complications of using a combination of these two techniques. Conclusions Symptomatic cholelithiasis in patients with COPD can be difficult to manage. In experienced hands, laparoscopic cholecystectomy can be performed safely under regional anaes-

thesia in this group of patients. The procedure may be contraindicated in patients with severe COPD who are CO2 retainers and in whom systemic CO2 absorption secondary to the pneumoperitoneum could prove detrimental. In these patients, the laparolift technique under epidural anaesthesia may be a viable option. Also contraindicated for this combined technique are those patients in whom epidural anaesthesia per se is contraindicated. These include patients with spinal deformities and blood dyscrasias. Although the number of such cases is so small that we cannot suggest that it be regarded as the preferred method of surgery, we believe that COPD is no longer a contraindication to laparoscopic cholecystectomy, especially under epidural anaesthesia.

References
1. Becqeumin JP, Piquet J, Becquemin MH, Melliere D, Harf A (1985) Pulmonary function after transverse or midline incision in patients with obstructive pulmonary disease. Intensive Care Med 11: 247251 2. Banting S, Shimi S, Vander-Velpen G, Cuschieri A (1993) Abdominal wall lift: low pressure pneumoperitoneum laparoscopic surgery. Surg Endosc 7: 5759 3. Craig DB (1981) Postoperative recovery of pulmonary function. Anesth Analg 35: 319322 4. Costantino GN, Vincent GJ, Mukalian GM, Kleifoth Jr WL (1994) Laparoscopic cholecystectomy in pregnancy. J Laparo Endoscopic Surg 4: 161164 5. Edelman DS (1994) Alternative laparoscopic technique for cholecystectomy during pregnancy. Surg Endosc 8: 794796 6. Edelman DS (1991) Laparoscopic cholecystectomy under continuous epidural anaesthesia in patients with cystic fibrosis [Letter] Am J Dis Child 145: 723724 7. Gadacz TR, Talamini MA, Lillemoe KD, Teo CJ (1990) Laparoscopic cholecystectomy. Surg Clin North Am 70: 12491262 8. Hansen G, Diablos PA, Steinert R (1977) Pulmonary complications, ventilation and blood gases after upper abdominal surgery. Acta Anaesthesiol Scand 21: 211215 9. Hedenstierna G (1989) Mechanisms of postoperative pulmonary dysfunction. Acta Chir Scand [Suppl] 550: 152158 10. Jack NTM, Lo GL, Rouwet EFAM, Sprakel JAH (1992) Cholecystectomy under epidural anesthesia and propofol sedation. Regional Anaesth 17(suppl): 35 11. Kang SB (1982) Continuous thoracic epidural anesthesia for biliary tract surgery and for postoperative pain relief in a patient with cystic fibrosis. Anesth Analg 61: 793795 12. Langenberg CJM, Huygan F, Go PMNYH, Gouma DJ (1992) Thoracic epidural anaesthesia for laparoscopic cholecystectomy. Regional Anaesth 17(suppl): 35 13. Meyers JR, Lembeck L, OKane H, Bane AE (1975) Changes in functional residual capacity of lung after operation. Arch Surg 110: 576582 14. Reddick EJ, Olsen DO (1989) Laparoscopic laser cholecystectomy: a comparison with minilap cholecystectomy. Surg Endosc 3: 131133 15. Soper NJ, Barteau JA, Clayman RV, Ashley AW, Dunnegan DL (1992) Comparison of early postoperative results for laparoscopic versus open standard cholecystectomy. Surg Gynecol Obstet 174: 114 118 16. Spence AA, Smith G (1971) Postoperative analgesia and lung function: a comparison of morphine with extradural block. Br J Anaesth 43: 144148 17. Weber AM, Bloom GP, Allan TR, Curry SL (1991) Laparoscopic cholecystectomy during pregnancy. Obstet Gynecol 78: 958959

Case report
Surg Endosc (1998) 12: 10851087 Springer-Verlag New York Inc. 1998

Laparoscopic treatment of duodenal carcinoid tumor

Wedge resection of the duodenal bulb under endoscopic control
T. Toyonaga,1 K. Nakamura,1 Y. Araki,2 H. Shimura,1 M. Tanaka1
1 2

First Department of Surgery, Kyushu University Faculty of Medicine, 3-1-1 Maidashi, Higashi-ku, Fukuoka, 812-8582 Japan Third Department of Internal Medicine, Kyushu University Faculty of Medicine, 3-1-1 Maidashi, Higashi-ku, Fukuoka, 812-8582 Japan

Received: 27 January 1997/Accepted: 4 December 1997

Abstract. A 46-year-old man with epigastralgia and slight elevation of urinary 5-hydroxyindole acetic acid (5HIAA) was found to have a well-demarcated carcinoid tumor in the duodenal bulb. The tumor measured 8 mm in size, and showed submucosal involvement but no metastasis to the liver and regional lymph nodes. After laparoscopic exposure and lifting of the duodenal wall around the tumor, wedge resection of the duodenal bulb including the tumor was performed successfully with a laparoscopic endostapler under direct endoscopic control. The postoperative course of the patient was uneventful. Laparoscopic wedge resection of the duodenum would be an appropriate minimally invasive treatment for selected duodenal neoplasms with special preoperative assessments and intraoperative considerations. Key words: Laparoscopic surgery Intraoperative endoscopy Duodenal carcinoid Endoscopic ultrasonography

A gastroduodenal carcinoid tumor is characterized by slow growth and low metastatic potential to the liver or lymph nodes until the late stage of the disease [4]. A conventional method of treatment for the duodenal carcinoid has been surgical excision or endoscopic mucosal resection [3]. The advent of laparoscopic surgery has opened a new pathway for treating benign and malignant diseases of the gastrointestinal tract [1]. We report a case of a duodenal carcinoid tumor treated by laparoscopic wedge resection of the duodenal bulb with the aid of intraoperative endoscopy. Case report
A 46-year-old Japanese man suffering from epigastralgia after meals for 2 months was admitted to the Kyushu University Hospital on June 6, 1996.

Fig. 1. Upper gastrointestinal series showing a well-demarcated tumor located in the duodenal bulb (arrow).

Correspondence to: T. Toyonaga

Physical examination showed no remarkable findings except for mild epigastric tenderness. Laboratory studies showed an elevated level of urinary 5HIAA of 8.3 mg/day (1.06.0), but normal serum concentrations of serotonin (0.16 g/ml, normal < 0.35) and 5HIAA (4.2 ng/ml, normal < 6.1). Duodenal endoscopy and radiology demonstrated a submucosal tumor,

1086

Fig. 2. Endoscopic ultrasonogram demonstrating the tumor involving the mucosa and the submucosa (arrow). Fig. 3. Intraoperative view. The duodenal wall including the tumor is exposed, lifted up, and grasped to simulate and confirm complete excision. Fig. 4. Resected specimen. A well-demarcated tumor with sufficient surgical margins. Fig. 5. Microphotograph of the carcinoid tumor. The tumor involves the mucosal and submucosal layers of the duodenum (hematoxylin-eosin stain, original magnification 2.5). Tumor cells with round or oval nuclei proliferate in a trabecullar and microglandular pattern (inset, original magnification 100).

8 mm in diameter, located in the lesser curvature of the duodenal bulb (Fig. 1). Endoscopic biopsy revealed a carcinoid by histology. Endoscopic ultrasonography demonstrated a hypoechoic tumor in the submucosal layer of the duodenum and no metastasis to periduodenal lymph nodes (Fig. 2). Computed tomography and ultrasonography of the abdomen revealed no metastasis to the liver or para-aortic lymph nodes. Endoscopic mucosal resection was abandoned because the tumor was located too close to the pylorus. Under general anesthesia the patient was put in the supine position. After creation of pneumoperitoneum via a Hasson trocar placed by open laparotomy at a subumbilical region, a 10-mm trocar was inserted on the pararectal line at the right lower quadrant and a 5-mm trocar at the right upper quadrant just below the costal margin. At laparoscopy, the stomach and duodenum showed no deformity or adhesion. Intraoperative duodenoscopy confirmed the site of the tumor on the posterior wall of the duodenal bulb just distal to the pylorus. The duodenal bulb was mobilized by devascularizing the lesser curvature from the pylorus to the duodenal descending portion. While precisely locating the tumor by duodenoscopy, the laparoscopic surgeon placed two stitches through all layers of the duodenum longitudinally 5 mm apart from the tumor edges, and lifted up the duodenal wall including the tumor (Fig. 3). Then wedge resection of the duodenum including the tumor with proper margins was performed by the use of a 60-mm EndoGIA. Before firing the stapler with its arms closed, we checked its correct placement by endoscopy. The duodenum was checked

for leaks, and the specimen was retrieved through the right lower 10-mm trocar. Although the oral side of the staple line was close to the pylorus, no deformity of the pyloric ring was observed by duodenoscopy. The staple line was reinforced with seromuscular sutures and covered with the omentum. The patient tolerated the whole procedure, which took 4 h and 20 min. The resected specimen showed a smooth-surfaced whitish submucosal tumor, which was hard in consistency and 9 8 mm in size (Fig. 4). A histologic examination revealed a well-demarcated carcinoid tumor located in the submucosal layer. The tumor was composed of small cells with uniform, round, or oval nuclei arranged in a trabecullar and microglandular pattern (Fig. 5). Postoperative recovery of the patient was uneventful. He was able to walk the next day, and his postoperative urinary 5HIAA was within the normal range.

Discussion Therapeutic options for a duodenal carcinoid tumor include surgical resection and endoscopic excision [3]. Although most pathologists consider all extra-appendiceal carcinoids as potentially malignant, solitary tumors smaller than 1 cm in diameter and confined to the submucosal layer rarely

1087

have metastasis [2]. Therefore, as shown in this case, endoscopic ultrasonography is of great value in determining the depth of the tumor involvement and in demonstrating the presence or absence of metastasis to the adjacent lymph nodes [3]. Endoscopic mucosal resection should be attempted if the lesion is localized within the submucosa and unaccompanied by lymph node metastasis. Endoscopic mucosal resection, widely performed in Japan for excising small gastric tumors, has some limitations when applied to duodenal tumors, even when they are confined to the mucosa or submucosa [3]. The thin wall and narrow lumen of the duodenum may lead possibly to higher risks of perforation and luminal stenosis as well as apparent difficulties of endoscopic control of the resection procedure. Considering the location in the duodenal bulb and the second portion as seen in most cases [3], duodenal carcinoids may be treated laparoscopically if the tumor is solitary, smaller than 1 cm in size, and free of metastasis. Preoperative assessments by endoscopic ultrasonography and intraoperative endoscopic luminal visualization are essential as shown in the present case. The latter facilitated the exact

location of the tumor and also ensured the complete resection of the tumor with proper margins. The patient should be informed about the possibility of conversion to a standard open procedure in the event of technical difficulties or unexpected findings such as lymph node metastasis.

References
1. Burke AP, Sobin LH, Federspiel BH, Shekitka KM, Helwig EB (1990) Carcinoid tumors of the duodenum: a clinicopathologic study of 99 cases. Arch Pathol Lab Med 114: 700704 2. Kaplan EL, Udekwu A (1990) The carcinoid syndromes. In: Friesen SR, Thompson NW (eds) Surgical endocrinology: clinical syndromes, 2nd ed, Lippincott, Philadelphia, pp 181209 3. Ohgami M, Kumai K, Otani Y, Wakabayashi G, Kubota T, Kitajima M (1994) Laparoscopic wedge resection of the stomach for early gastric cancer using a lesion-lifting method. Dig Surg 11: 6467 4. Yoshikane H, Tsukamoto Y, Niwa Y, Goto H, Hase S, Mizutani K, Nakamura T (1993) Carcinoid tumors of the gastrointestinal tract: evaluation with endoscopic ultrasonography. Gastrointest Endosc 39:375383

Review article
Surg Endosc (1998) 12: 10091012 Springer-Verlag New York Inc. 1998

Active electrode monitoring

How to prevent unintentional thermal injury associated with monopolar electrosurgery at laparoscopy
T. G. Vancaillie
Department of Endogynecology, Royal Hospital for Women, Barker Street, Randwick NSW 2031, Sydney, Australia Received: 26 June 1997/Accepted: 10 December 1997

Abstract Background: In recent years, the use of minimally invasive surgery (MIS) has expanded to a wide variety of surgical specialties. The increased popularity of the procedure, however, has been accompanied by its share of complications, including trocar lacerations and inadvertent thermal injuries to nontargeted tissues during monopolar electrosurgery. Methods: A survey on electrosurgical thermal injuries and three case studies are presented. The new technology of active electrode monitoring (AEM) is described. Results: AEM eliminates stray currents generated by insulation failure and capacitive coupling. Conclusions: To reduce the incidence of injury by monopolar electrosurgery at laparoscopy, there is a need for advanced technology, such as AEM. In addition, laparoscopic surgeons should be encouraged to study the basic concepts of the biophysics of electrosurgery. Key words: Laparoscopy Electrosurgery Insulation failure Capacitive coupling Active electrode monitoring

In the late 1980s, the development of videolaparoscopy led to an explosion in the use of minimally invasive surgery (MIS). Beginning with a small number of gynecological procedures, MIS has been applied to a wide array of surgical specialties, including gastrointestinal, oncologic, and general surgery. Survey results indicate that by the year 2000, 50% of general surgery procedures and 70% of gynecology procedures will be performed via MIS [9]. Laparoscopic surgery is favored by both surgeons and patients over conventional surgery. Patients usually heal faster and suffer less postoperative pain than with traditional
Correspondence to: T. G. Vancaillie

open surgery, expediting discharge from the hospital and requiring shorter convalescence. As with open surgery, monopolar electrosurgery is the preferred technique for tissue cutting and hemostasis in laparoscopy. Monopolar electrosurgery is employed by >85% of surgeons who perform laparoscopic procedures [6]. Monopolar instruments enable the delivery of a significant level of energy to targeted tissue, accounting for the versatility of the procedure. Surgeons can perform smooth cuts by using a continuous low-voltage current, fulgurate tissue with a damped current, or combine the two functions simply by varying the current or voltage level delivered to the tip of the active electrode. These adaptive features of monopolar electrosurgery have made it an invaluable tool in both open and laparoscopic surgery. Though it is popular, cost-effective, and versatile, the combination of monopolar electrosurgery and laparoscopy can be dangerous [3, 5, 8]. The reduced field of view inherent to laparoscopic surgery prevents the surgeon from directly observing any tissue located away from the tip of the active electrode. Because of this restricted view, the surgeon is less likely to detect thermal damage caused by stray energy [5]. Electrosurgical burns result from insulation failure or capacitive coupling. Small defects in the layer of electrical insulation surrounding the shaft of the active electrode allow energy to leak from the instrument during surgery. These instrument defects can be hard to detect, even with careful visual inspection. Repeated handling and passes of the electrode shaft through trocars can compromise the insulation, as can disinfection and sterilization. Moreover, the high voltage associated with certain current modes can stress the insulation over time, making it vulnerable to cracks [4]. Capacitive coupling is electrical current that is established in tissue or in metal instruments running parallel tobut not directly in contact withthe activated electrode. The electromagnetic field around the active electrode

1010

created by the alternating current induces electrical energy in any nearby parallel conductor. Capacitive coupling is, at best, containedit cannot be eliminated [5, 8]. Various other laparoscopic surgical techniques, such as bipolar, laser, and the harmonic scalpel, have been evaluated by surgeons in an attempt to circumvent the problem of electrosurgical burns during laparoscopic monopolar electrosurgery. The clinical efficacy of these techniques, however, is limited, and they have not been widely adopted. A survey conducted at a 1993 conference of the American College of Surgeons (ACS), in fact, found that only 12% of surgeons who perform laparoscopic surgery use bipolar techniques and only 2% use laser energy.

Surgeon perspectives and case histories The results of the 1993 ACS conference survey indicated that there is a high level of awareness among surgeons of the danger of electrosurgical thermal injury to patients. When questioned, 86% of the 506 responding surgeons acknowledged the potential for burns to tissues outside the surgical field during laparoscopic monopolar electrosurgery. A number of the surgeons reported firsthand experience with complications resulting from insulation failure or capacitive coupling, and over half (54%) stated that they knew of colleagues whose patients had suffered complications [7]. Despite this understanding of the potential for electrosurgical burns from laparoscopic surgery, complications such as direct trocar or needle puncture wounds and instrument lacerations are more widely acknowledged in daily practice by the surgical communityin part, because they are easier to diagnose and treat when they occur. Inadvertent burn injuries to nontargeted tissues outside the surgeons view during laparoscopic monopolar electrosurgery, on the other hand, are difficult to diagnose and thus are less well understood. Symptoms of electrosurgical thermal injury are often delayed, making it difficult to determine the etiology of the problem. The injured area may also be compromised by secondary infection, making histologic diagnosis complex. In addition, an area of coagulative necrosis may be missed on microscopic examination, or a pathologist may not be aware of the unique histological characteristics of thermal injury and mistakenly attribute the injury to some other cause, such as trocar puncture or mechanical laceration. For these reasons, the prevalence of electrosurgical burns is likely underreported and underestimated by the surgical community. Among the consequences of thermal injuries are bowel perforation and peritonitis, which are associated with significant morbidity and even death. Fecal peritonitis, for example, has a mortality rate as high as 25% [1]. Recently, we encountered a case involving thermal injury to a patient undergoing routine laparoscopic surgery. The patient was a 37-year-old woman who presented with pelvic pain and metrorrhagia. A laparoscopy was performed for resection of endometriosis. Monopolar electrosurgery was used to resect the affected area of the rectovaginal septum. Hemostasis was accomplished by monopolar electrodesiccation and fulguration. At the end of the procedure,

a bandlike strip of unintentional thermal damage was noted along the left side wall that included the inner aspect of the left ureter, necessitating the placement of a 28-cm 6-Fr double J stent. Fortunately, the patient did not suffer any long-term adverse effects from the burn since the injury was immediately detected and treated. Not all patients are so fortunate. Reliable incidence figures on burn injuries during laparoscopic monopolar electrosurgery are difficult to obtain given the aforementioned diagnostic challenges. In recent years, however, an increasing number of case histories have appeared in the medical literature. The following case history illustrates that complications arising from undetected burns during monopolar electrosurgery can have serious and long-term morbidity. A 38-year-old nurse was seen by a gynecologist for left lower quadrant pain. The patients surgical history included wedge resection of the left ovary for endometriosis. The gynecologist diagnosed pelvic adhesions of the ovary and performed laparoscopic surgery. Monopolar electrosurgery was used to cauterize adhesions from the ovary to the pelvic side wall. The power setting of the electrosurgical generator was 30 W, and the electrode was activated for 5 sec. The patient was discharged from the hospital on the same day of surgery, but she was admitted to the emergency room with a low-grade fever and leukocytosis on the 7th postoperative day. A CT scan found free air in the abdomen. Exploratory laparotomy revealed multiple necrotic areas in the distal ileum that resembled burns. Several areas of the colon appeared compromised, with one area showing perforation. Peritonitis was localized to the right lower quadrant. Microscopic examination of the small bowel showed focal full-thickness necrosis. Examination of the large intestine revealed areas of mucosal ulceration and full-thickness wall necrosis. During the follow-up surgery, 40 cm of the ileum were removed, and a temporary colostomy was performed. After the laparotomy, the patient developed a wound infection requiring further treatment. She was not sufficiently well to return to normal activities until 6 months after the initial laparoscopic surgery [11]. As this case study suggests, electrosurgical thermal injury should be suspected in patients who have undergone monopolar laparoscopic electrosurgery and who demonstrate symptoms associated with organ perforation or peritonitis. The restricted visual environment in laparoscopy increases the risk of unseen electrosurgical burns. Electrical interference on the electrosurgical units video monitor or reduced power at the tip of the electrode are unreliable indicators of electrosurgical burn injury potential, but they should arouse the suspicion of the presence of stray currents. In addition to the potential clinical risks to patients, surgeons performing laparoscopic monopolar electrosurgery may encounter medicolegal liability. At the 1995 meeting of the Society of Laparoendoscopic Surgeons, 13% of members surveyed reported involvement with one or more active malpractice cases associated with a laparoscopic electrosurgical procedure [7]. A malpractice case in 1994 illustrates the liability risks surgeons face as a result of undetected thermal injury. In this case, an Oregon woman underwent laparoscopy in 1994

1011

for gallbladder removal. One week after the surgery, a laparotomy revealed a high-grade stricture of the common hepatic ducta complication that required additional surgeries for repair and dilation. Following her recovery, the patient sued the surgeon for negligence and received a substantial damage award. The operating room record indicated that significant electrical interference on the video monitor had hindered completion of the surgery. A witness for the surgeon testified that these periods of interference most likely signaled the existence of stray electric currents that produced the burns to the hepatic duct [2]. Physician insurance companies have responded to the medicolegal risk posed by laparoscopic monopolar electrosurgical procedures through rate adjustments and training incentives. Some providers of malpractice insurance offer their members free accredited postgraduate training courses in electrosurgery and risk management [5].

Active electrode monitoring: a technological advance in the prevention of electrosurgical burns due to stray currents The continued reports of clinical and medicolegal problems directly associated with thermal injury during laparoscopic monopolar electrosurgery reinforces the view that the most common protective measures (e.g., inspection of electrodes of insulation cracks, specialized training for surgical personnel, etc.) have not eliminated the risk of burns to nontargeted tissues during minimally invasive monopolar electrosurgery. Even the most able and experienced surgeons who consistently use strict safety protocols cannot transcend the immutable conditions of the electrical environment encountered during laparoscopic monopolar electrosurgery. An alternative technological solution is necessary to ensure patient safety during laparoscopic monopolar electrosurgery. The selection of a particular technology should be evaluated relative to its capability of eliminating inadvertent tissue injury due to stray electrical currents, require a minimal amount of training or modification in surgical methods, and offer a cost-effective solution. Moreover, it should overcome the current deficiencies in the maintenance and testing of laparoscopic instruments by ensuring that when such instruments do fail, they fail safely. Active electrode monitoring (AEM) (ElectroScope, Boulder, CO, USA) meets these criteria. AEM offers the ultimate safety in monopolar electrosurgery by combining added electrical insulation, conductive shielding, and an electronic current monitoring system. Stray currents that may be released through faulty insulation are absorbed by the additional electrical insulation and conductive shielding (Fig. 1). The conductive shielding within the insulation itself becomes capacitively coupled to the active electrode, instead of any metal surgical instruments or the patients tissue, eliminating the incidence of tissue burns from capacitive coupling. The conductive sheath is electrically connected to the return electrode of the electrosurgical unit, allowing for harmless dissipation of capacitively coupled currents. If stray energy levels become sufficiently high to damage nontargeted tissues, the AEM circuit interrupts the flow of energy from the electrosurgical unit and sounds an alarm.

Fig. 1. Schematic representation of active electrode monitoring. The active electrode is surrounded by three successive layers: a first layer of insulation, a conductive sheath, and a second layer of insulation. The conductive sheath captures the current generated by capacitive coupling, which cannot be avoided, and other stray currents. These stray currents are analyzed by the active electrode monitor. If the amount or character of the stray currents exceeds or differs from preset norms, the AEM will shut down the electrosurgical generator (ESU).

The Emergency Case Research Institute, a nonprofit organization that evaluates medical devices, tested this monitoring system and concluded that it offers maximum protection against patient injury due to either insulation failure or capacitive coupling [10]. The contribution of AEM to the safe application of monopolar laparoscopic electrosurgery has also been recognized by the American Association of Gynecological Laparoscopists, which has urged surgeons to consider the use of AEM when performing laparoscopic monopolar electrosurgery. Conclusions Laparoscopic monopolar electrosurgery is a highly versatile and effective tool that is used in a wide variety of surgical specialties. Monopolar electrosurgery is by far the superior and preferred technique for tissue cutting and hemostasis, eclipsing alternative measures. It can, however, place patients at risk for unintended burns to nontargeted tissues outside the surgeons view, resulting from stray electrical currents associated with insulation failure or capacitive coupling. The clinical and medicolegal risks, combined with the anticipated growth in the number and type of laparoscopic surgery applications, necessitate a shift in electrosurgical practice. This shift should ideally encompass the study of the basic concepts in biophysics of electrosurgery and the introduction of more sophisticated technology. AEM minimizes the risk of nontargeted tissue burns associated with minimally invasive electrosurgery while allowing both surgeons and patients to reap the many benefits of laparoscopic monopolar electrosurgery. References
1. Berry SM, Ose KJ, Bell RH, Fink AS (1994) Thermal injury of the posterior duodenum during laparoscopic cholecystectomy. Surg Endosc 8: 197200 2. Golden TR (1993) Laparoscopic cholecystectomy verdict. Trial News 20

1012 3. Grosskinsky CM, Hulka JF (1995) Unipolar electrosurgery in operative laparoscopy. J Reprod Med 40: 549552 4. Luciano AA, Soderstrom RM, Martin DC (1994) Essential principles of electrosurgery in operative laparoscopy. J Am Assoc Gynecol Laparosc 1: 189195 5. Odell RC (1993) Electrosurgery in laparoscopy. Infert Reprod Med Clin North Am 4: 289304 6. Southern Surgeons Club (1991) A prospective analysis of 1518 laparoscopic cholecystectomies. New Engl J Med 324: 10731078 7. Tucker RD (1995) Laparoscopic electrosurgical injuries: survey results and their implications. Surg Laparosc Endosc 5: 311317 8. Vancaillie TG (1994) Electrosurgery at laparoscopy: guidelines to avoid complications. Gynaecol Endosc 3: 143150 9. Wetter PA (1994) Trends study. Presented at the Annual Meeting of the Society of Laparoendoscopic Surgeons, Westin Hotel, Seattle 10. Anonymous (1995) Focus on laparoscopy. Health devices 24: 338 11. Trudy Karl v. Rufus S. Armstrong, M.D. (1993) Fla Jury Verdict Rep 14: 4748

Surg Endosc (1998) 12: 10551060

Springer-Verlag New York Inc. 1998

Laparoscopic Collis gastroplasty and Nissen fundoplication

A new technique for the management of esophageal foreshortening
A. B. Johnson, M. Oddsdottir, J. G. Hunter
Department of Surgery, Emory University Hospital, Room H124C, 1364 Clifton Road, N.E., Atlanta, GA 30322, USA Received: 8 September 1997/Accepted: 17 December 1997

Abstract Background: The short esophagus increases the difficulty and limits the effectiveness of laparoscopic Nissen fundoplication. In our experience, 2025% of esophagi judged by preoperative criteria to be foreshortened will, after dissection, be insufficiently long to allow 2 cm of esophagus to reside below the diaphragm without inferior distraction (i.e., tension free). Collis gastroplasty combined with Nissen fundoplication has become the standard approach for the creation of an intraabdominal neoesophagus and fundic wrap. Methods: After developing methods of performing totally laparoscopic stapled gastroplasty in the cadaver lab in 1994, we started applying the technique clinically in 1996. We performed 220 laparoscopic antireflux procedures between January 1996 and July 1997. Of these 220 patients, 26% were suspected to have esophageal foreshortening based on preoperative barium studies and/or endoscopy. Results: After hiatal dissection, nine patients, or 16% of those suspected to have esophageal foreshortening and 4% of the entire population, required the laparoscopic CollisNissen procedure. There was symptomatic improvement in all patients as assessed by patient-initiated symptom scores. Conclusions: The management of patients with esophageal foreshortening is a complex problem. We believe that our technique of laparoscopic Collis-Nissen provides an effective means of achieving intraabdominal placement of the fundic wrap while maintaining the benefits of a minimally invasive approach. Key words: Hiatal hernia Paraesophageal hernia Gastroesophageal junction Esophageal stricture Collis gastroplasty Laparoscopic Nissen fundoplication

The shortened esophagus not only increases the difficulty but also limits the effectiveness of laparoscopic Nissen funCorrespondence to: J. G. Hunter

doplication. It has long been known to complicate the work of anti-reflux surgery and paraesophageal hernia repair. Esophageal foreshortening is found more frequently in association with a gastroesophageal (GE) junction that is >5 cm above the hiatus on barium swallow, esophageal stricture, type III (mixed) paraesophageal hernia, and Barretts esophagus (with or without stricture) [8]. Even when these preoperative findings are noted, it is often difficult to predict which patients will have a truly foreshortened esophagus, because the esophagus, when adequately mobilized and transposed to the anterior hiatus, is still long enough to allow the GE junction to reside below the hiatus without tension. Previous investigators have demonstrated that only 20% of esophagi believed preoperatively to be foreshortened will prove to be foreshortened intraoperatively [11]. Collis gastroplasty in combination with complete or partial fundoplication has become the standard approach to create an antireflux valve in patients with esophageal foreshortening. Collis originally described the performance of gastroplasty through a thoracoabdominal incision; however, it is now usually performed with a transthoracic approach and followed by a partial (Belsey) or complete (Nissen) fundoplication [2]. Traditional teaching has emphasized the need for extensive mediastinal dissection in order to adequately mobilize the esophagus for a tension-free transthoracic repair [1]. Though it provides adequate exposure, thoracotomy subjects patients to significant pain and morbidity. The desire to avoid inconvenience to the patient led to the development of abdominal gastroplasty techniques by Steichen and Henderson and Marryatt to manage the shortened esophagus [4, 10]. Two descriptions have been published of thoracoscopic Collis gastroplasty combined with laparoscopic fundoplication [3, 11]. Because of the additional requirements of thoracoscopy (e.g., double-lumen anesthesia, additional videoendoscopic equipment, chest preparation), we have developed a laparoscopic approach to esophageal lengthening. Our technique and results are described in this report.

1056 Fig. 1. Trocars are placed in the baseball diamond configuration with the 5-mm right epigastric and 12-mm left epigastric operating ports in the third and first base positions, respectively. There is also a 10-mm camera port to the left of the umbilicus, a 5-mm right subcostal port, and a 5-mm left subcostal port. Fig. 2. A 48-Fr dilator is placed to calibrate the width of the gastric tube. The burn mark is placed 3 cm inferior to the angle of His and 1 cm away from the dilator. Fig. 3. The Keith needle and attached anvil is passed through the posterior wall of the stomach 1 cm away from the dilator.

Technique The laparoscopic Collis gastroplasty was developed in the fresh-tissue laboratory by one of our group (M.O.) in 1994 [7]. The technique utilizes the same five trocars and trocar positions as laparoscopic Nissen fundoplication (Fig. 1) [5]. The initial dissection technique is also identical to that used for uncomplicated fundoplication. The hiatal hernia is reduced, and the hernia sac is dissected from the crura. The

gastric fundus is mobilized from the spleen and the left hemidiaphragm by dividing the short gastric vessels, the posterior gastric artery, and all posterior gastrophrenic attachments. The retroesophageal window is developed, and a Penrose drain encircles the esophagus. This maneuver provides optimal inferior traction for esophageal dissection in the mediastinum. Dissection proceeds into the meddiastinum for 46 cm. It is occasionally possible to mobilize the entire distal third

1057

Fig. 4. The 21-mm circular stapler is introduced through an incision to the left of the xiphoid and then docked with the anvil to create a sealed gastric window. Fig. 5. A 30-mm endoscopic stapler is inserted and fired adjacent to the dilator, thus lengthening the esophagus by 4 cm. Fig. 6. The completed wrap with the fundic staple line behind the esophagus. The apex of the staple line abuts the esophagus on the right. Fundoplication sutures are placed on either side of the staple line.

of the esophagus through the hiatus laparoscopically. It is difficult to mobilize higher than this in most patients, since access to the mediastinal esophagus is limited by anterior deflection of the esophagus around the heart and close apposition of the pleura to the middle third of the esophagus. Once the esophagus is completely mobilized, it is verified that the GE junction cannot be adequately reduced to allow

2 cm of esophagus to reside in the abdomen without tension. If the Penrose drain creeps above the hiatus after release, a Collis gastroplasty should be performed. Next, it is extremely important to remove the hernia sac and epiphrenic fat from the angle of His anteriorly and posteriorly. We find that the Ultracision LCS (Smithfield, RI, USA) accomplishes this task bloodlessly. When it be-

1058

comes clear that esophageal lengthening is necessary, the 10-mm laparoscope port (above and to the left of the umbilicus) is replaced with a 12-mm trocar to accommodate the linear cutting stapler. A 48-Fr dilator is placed to calibrate the width of the gastric tube (Fig. 2). A burn mark is then placed 3 cm inferior to the angle of His and 1 cm from the dilator on the anterior wall of the stomach to indicate the exit point of the anvil of a 21-mm circular stapler. This procedure will create a 4-cm neoesophagus. A 2-cm vertical mini-laparotomy is made just to the left and slightly inferior to the xiphoid and divided down through the peritoneum with electrocautery. The minilaparotomy is dilated with a large hemostat. A 2-0 Prolene suture on a Keith needle is attached to the open hole on the plastic skewer of a 21-mm circular cutting anvil (CLH-21; Ethicon Endosurgery, Cincinnati, OH, USA) and popped into the peritoneal cavity through this mini-laparotomy. The skin is closed with towel clips in order to maintain the pneumoperitoneum. The anvil is placed in the lesser sac by elevating the well mobilized gastric fundus. The greater curvature of the stomach is held anteriorly with two graspers. The Keith needle and attached anvil are then passed through the posterior wall of the stomach, exiting the anterior wall of the stomach at the burn mark 1 cm away from the dilator (Fig. 3). The body of the 21-mm circular stapler is introduced through the xiphoid incision, docked with the anvil, and fired to create a sealed window through both gastric walls (Fig. 4). The laparoscope is shifted to the left subcostal position, and a 30-mm linear cutting stapler (Ethicon Endosurgery) is inserted through the primary trocar near the umbilicus and fired adjacent to the dilator (Fig. 5). The staple line is oversewn with two running vertical mattress sutures of 2-0 braided nylon, one starting at the GE junction and one starting at the distal margin of the staple line on the fundus. This suture reinforcement of the staple line may not be essential, but it provides additional security against leak or bleeding. The sutures are tied to each other when they meet in the region of the circular staple line. After an appropriate crural closure, a 2-cm floppy Nissen fundoplication is sutured. The fundic staple line lies behind the esophagus, with its apex becoming the middle point of the fundic suture line to the right of the esophagus (Fig. 6).

Table 1. Patient demographics Patient no. 1 2 3 4 5 6 7 8 9 Age/ sex 70/F 35/M 65/M 83/M 48/M 67/M 69/F 68/M 57/F O.R. time (min) 285 351 289 394 332 210 232 287 269 Length of stay (days) 6 3 2 4 2 3 2 2 3

Cause Large HH/stricture HH/stricture Para HH/stricture Large Para HH Large HH/esophagitis Large HH/stricture Para HH/stricture Para HH HH/stricture

Complications atelectasis none none A-fib none none none none none

Para, paraesophageal; HH, hiatal hernia

examination suggests that the fundoplication may have come apart. Between January 1996, when we performed the first Collis gastroplasty, and July 1997, we did a total of 220 laparoscopic antireflux procedures. Of this population, 58 patients (26%) were suspected to have esophageal foreshortening prior to surgery. Of these 58, nine patients (16% of those suspected, 4% of entire population) required esophageal lengthening with a laparoscopic Collis gastroplasty. As compared to a group of patients that underwent laparoscopic fundoplication, operative time was longer and length of stay was longer, but there was no additional morbidity. Using a five-point patient-initiated symptom score pre- and postoperatively (0 no symptoms, 1 rare symptoms, 2 moderate symptoms, 3 severe symptoms, 4 inccapacitating symptoms), we found no differences in clinical response between these patients and our control group of 300 patients undergoing Nissen fundoplication (Table 2) [6].

Discussion Laparoscopic Nissen fundoplication has become a routine procedure; many centers have reported >200 procedures. In most centers, Collis gastroplasty is not considered necessary. Our development of the Collis gastroplasty technique started in the cadaver lab shortly after we had reached 100 procedures and was not applied until we had reached 400 procedures. We believed that this additional procedure was occasionally necessary because of the high rate of paraesophageal herniation following laparoscopic Nissen fundoplication reported by others as well as ourselves (range, 38%) [6, 13]. One of the avoidable causes of postoperative paraesophageal hernia is the need to ensure adequate esophageal length to allow the GE junction to reside in the abdomen without tension. When we specifically looked for a cause of paraesophageal herniation among the 3% of our patients that developed this problem, esophageal foreshortening contributed in less than a third of patients (1%). While it may be enticing to attribute the high frequency of postoperative dysphagia reported during the learning curve of laparoscopic fundoplication to a short esophagus and paraesophageal herniation, this does not appear to be the case. In most cases, persistent postoperative dysphagia occurs be-

Clinical experience Our technique of laparoscopic Collis-Nissen was performed in nine patients between January 1996 and July 1997 (Table 1). Large hiatal hernias were present in all patients. Two patients had intrathoracic stomachs, and both of these patients had esophageal strictures requiring dilation. There has been symptomatic improvement in all patients in whom we have utilized this technique; however, one patient suffered a recurrence of moderate dysphagia from a distal esophageal stricture present before the operation. This patient has responded to dilation and standard doses of omeprazole, a dosage that did not control his heartburn preoperatively. We believe that the additional length of this patients esophagus may be contributing to symptom control, since endoscopic

1059 Table 2. Typical symptoms pre- and postoperatively of Collis-Nissen patients (n 9) and population undergoing laparoscopic fundoplication (n 300) 1 year following operation Preop Collis-Nissen n 9) SSSa Heartburn Regurgitation Dysphagia
a

Postop Nissen (n 253) Collis-Nissen (n 9) 2 22% 13% 13% 34 70% 31% 31% 01 89% 100% 89% 2 11% 0 11% 34 0 0 0 Nissen (n 253) 01 92% 95% 88% 2 4% 3% 7% 34 4% 2% 5%

01 56% 49% 78%

2 0 22% 11%

34 44% 33% 11%

01 21% 56% 62%

Symptom Severity Score (SSS): 0 no symptoms, 1 rare symptoms, 2 moderate symptoms, 3 severe symptoms, 4 incapacitating symptoms

cause the fundoplication has been misformed. In our experience, most of these patients have had the Rosetti modification of the Nissen fundoplication [12]. The key to managing patients with esophageal foreshortening is making an accurate intraoperative determination that the esophagus is truly foreshortened. After dissecting 46 cm up into the mediastinum with the GE junction retracted inferiorly, the esophagus is transposed to the anterior hiatus and released. If it springs back to the diaphragm or above, it is too short and should be lengthened. Preoperatively, a short esophagus was predicted in 16% of patients who met liberal preop criteria; this group represents 4% of our entire population. In another series, a 14% incidence of esophageal foreshortening was predicted by preoperative criteria, of which 9% (1.2% of the patient population) required esophageal lengthening with a thoracoscopic Collis gastroplasty [11]. With our technique, considerable time may be required to oversew the gastric staple lines. The staple line of the gastroplasty may not need to be oversewn; however, it adds additional security in preventing gastric leakage, since the endoscopic linear cutting staples are only 3.5 mm long as opposed to the 4.8-mm staple length generally used on the stomach. A running vertical mattress suture has been most effective for oversewing these staple lines. Two other minimally invasive techniques of performing Collis gastroplasty have been described [3, 11]. In the first technique, the usual subdiaphragmatic and mediastinal dissection for laparoscopic fundoplication is performed with laparoscopic visualization, followed by right thoracoscopy and placement of the linear stapler for the gastroplasty through a second port. This maneuver facilitates the placement of the stapler at the angle of His in the proper orientation. Crural closure and fundic wrap are then completed laparoscopically [11]. In the second technique, the entire procedure is performed with left thoracoscopic access. The gastroplasty is created using a noncutting linear stapler, followed by fundic wrap and reduction below the diaphragm [3]. Although these techniques recapitulate the standard open surgical techniques, they require single-lung ventilation, chest preparation, two video carts, thoracotomy privileges, and the increased pain associated with thoracoscopy. These disadvantages can be avoided if laparoscopic Collis gastroplasty is performed instead. Certainly, other traditional approaches for the performance of Collis gastroplasty should not be neglected. Although most surgeons perform this procedure through a left thoracotomy, the technique performed by Steichen through a laparotomy yields equivalent results [10]. Although our

preferred access is laparoscopic in patients without previous operation, in one patient we elected to perform Collis gastroplasty through a laparotomy because the patient had previously undergone open fundoplication. It is generally most expedient to perform redo surgery through a laparotomy if the first operation was performed through a laparotomy. Another very acceptable approach is to convert a laparoscopic Nissen to an open Collis-Nissen when esophageal foreshortening is discovered intraoperatively. It is more important that the operation be performed correctly than that laparoscopic access be maintained. Because we developed this technique in the cadaver lab and practiced multiple times in both the pig and human cadaver, it was not necessary for us to convert any of our cases. The surgical management of patients with esophageal foreshortening is a complex problem. In patients with extremely poor esophageal motility and a tight stricture, segmental esophagectomy is often the optimal therapy [8]. When esophageal motility is poor and the esophagus is short but there is no stricture, the Collis gastroplasty may be combined with a posterior partial fundoplication. Less definitive approaches, such as gastropexy with crural closure and mediastinal positioning of the fundoplication, result in unacceptably high recurrence rates and postoperative discomfort. Most authors agree that in order to obtain the best results, the fundoplication following gastroplasty must be placed below the diaphragm and under no tension [9]. We believe that our technique of laparoscopic Collis-Nissen provides an effective means of managing patients with a foreshortened esophagus while also offering the advantages of a minimally invasive approach.

References
1. Adler RH (1990) Collis gastroplasty: origin and evolution. Ann Thorac Surg 50: 839842 2. Collis JL (1957) An operation for hiatus hernia with short oesophagus. Thorax 12: 181188 3. Demos NJ, Kulkarni VA, Arago A (1994) Video-assisted transthoracic hiatal hernioplasty using stapled, uncut gastroplasty and fundoplication. Surg Rounds xx: 427436 4. Henderson RD, Marryatt GV (1985) Transabdominal total fundoplication gastroplasty to control reflux. A preliminary report. Can J Surg 28: 127129 5. Hunter JG, Champion JK (1996) Laparoscopic Nissen fundoplication. In: Endosurgery. Churchill Livingstone, New York & London 6. Hunter JG, Trus TL, Branum GD, Waring JP, Wood WC (1996) A physiologic approach to laparoscopic fundoplication for gastroesophageal reflux disease. Ann Surg 223: 673687

1060 7. Oddsdottir M, Laycock W, Champion K, Hunter JG (1995) Laparoscopic esophageal lengthening procedure [abstract]. Surg Endosc 9: 621 8. Pearson FG (1995) Peptic esophagitis, stricture, and short esophagus. In: Esophageal surgery. Churchill Livingstone, New York 9. Pearson FG, Todd TR (1987) Gastroplasty and fundoplication for complex reflux problems: long-term results. Ann Surg 206: 473481 10. Steichen FM (1986) Abdominal approach to the Collis gastroplasty and Nissen fundoplication. Surg Gynecol Obstet 162: 273275 11. Swanstrom LL, Marcus DR, Galloway GQ (1996) Laparoscopic Collis gastroplasty is the treatment of choice for the shortened esophagus. Am J Surg 171: 477481 12. Trus TL, Cornwell M, Waring JP, Galloway K, Hunter JG (1998) Patterns of failure and results of redo fundoplication. Surg Endosc (in press) 13. Watson DI, Jamieson GG, Devitt PG, Mitchell PC, Game PA (1995) Paraesophageal hiatus hernia: an important complication of laparoscopic Nissen fundoplication. Br J Surg 82: 521523

Surg Endosc (1998) 12: 10641072

Springer-Verlag New York Inc. 1998

Laparoscopic surgery for abdominal aortic aneurysms

Technical elements of the procedure and a preliminary report of the first 22 patients
J. K. Edoga, K. Asgarian, D. Singh, K. V. James, J. Romanelli, S. Merchant, D. Romano, B. Joostema, J. Street
Departments of Surgery, Anesthesia, and Nursing, Morristown Memorial Hospital, Morristown, NJ 07960, USA Received: 23 June 1997/Accepted: 11 December 1997

Abstract Background: Laparoscopic surgery for infrarenal aortic aneurysms is based on the principle of retroperitoneal exclusion of the aneurysm sac with aortofemoral or aortoiliac bypass. Methods: Of 22 patients who met the selection criteria, 20 successfully underwent laparoscopic aortic surgery at Morristown Memorial Hospital between February and October 1997. Technical elements and steps of this operation are described and illustrated. Results: Within 30 days of surgery, 2 patients died and 9 had various major and minor perioperative complications. As a group, the laparoscopic patients had less postoperative pain, needed fewer hours of ventilator support, had shorter intensive care unit (ICU) and hospital lengths of stay, and resumed diet and normal activity earlier than the historical norms for patients undergoing transabdominal or retroperitoneal aortic resections at the same institution. Conclusions: These early observations suggest that the laparoscopic treatment of infrarenal abdominal aneurysms may have several significant potential benefits. Long-term results and randomized prospective studies with patients matched by risk stratification will be needed to confirm these impressions. Key words: Laparoscopic Retroperitoneal Transabdominal Minimally invasive potential benefits

randomized prospective trial comparing the retroperitoneal with the transabdominal approach to the aorta for routine infrarenal aortic reconstruction. His findings demonstrated that the retroperitoneal approach was associated with

Fewer postoperative complications Shorter ICU stay Shorter hospital stay Decreased cost.

However, long-term incisional pain was clearly greater in the retroperitoneal group. In our small but rapidly growing experience, the laparoscopic technique for the treatment of abdominal aortic aneurysms appears to contribute the proven benefits of minimally invasive surgery while adding none of the current drawbacks and limitations of the endovascular approach [6, 7, 11]. Our laparoscopic procedure is based on principles of retroperitoneal aneurysm exclusion and bypass described by Shah and his associates in 1991 [9] and their long-term results reported by Resnikoff et al. in 1996 [8]. However, performance of his operation through laparoscopic ports and/or small videoscopically guided incisions has the potential benefits of

The posterolateral retroperitoneal approach to surgery on the abdominal aorta, as initially described by Williams et al. [12] and more recently by Darling and his colleagues [3], has been associated with several physiologic advantages when compared with the traditional transperitoneal procedure [2]. In 1995, Gregorio Sicard [10] published the first
Correspondence to: John K. Edoga, Northwest Surgery P.A., 95 Madison Avenue, Morristown, NJ 07960, USA

Decreased postoperative pain Reduced postoperative ventilator dependence Shorter ICU stay Reduced hospital length of stay Significantly reduced total cost Safe early resumption of normal activity [1].

Our experience with the laparoscopic approach has also demonstrated that the minimally invasive nature of the operation does not appear to positively influence the outcome in high-risk patients and that, as with the traditional open procedures, the risk of perioperative death is still largely determined by the number and severity of comorbid conditions [4, 5].

1065 Table 1. Patient demographics and initial outcome data AAA size (cm) 6.0 5.0 6.5 7.5 5.5 6.0 6.7 8.0 4.4 5.5 6.0 6.0 6.0 5.0 5.8 7.0 6.1 4.0 7.5 5.5 Total length of stay (days) 4 5 5 19 4 6 8 2 25 7 3 3 5 5 4 3 5 4 2 5

Patient 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
a

Date 2/3/97 2/6/97 2/17/97 2/20/97 3/3/97 3/3/97 3/4/97 3/6/97 3/10/97 3/17/97 3/20/97 4/7/97 4/10/97 5/29/97 6/9/97 7/7/97 8/7/97 8/14/97 9/4/97 10/2/97

Age 66 77 66 66 67 72 74 88 83 69 69 67 64 77 70 78 62 75 69 85

Sex M F M M F M M M M M F M M F M F M M M M

Time to extubation <24 h <24 h <24 h <24 ha <24 h <24 h 3 days 2 days 1 day <24 h <24 h <24 h <24 h <24 h <24 h 1 day 1 day <24 h <24 h <24 h

ICU days 1 2 1 19 1 1 4 2 4 1 1 1 1 4 2 1 3 1 1 1

Days to diet 0 0 0 19 0 1 6 2 4 1 0 2 0 1 1 1 1 1 0 1

Outcome Discharged Discharged Discharged Death Discharged Discharged Discharged Death Discharged Discharged Discharged Discharged Discharged Discharged Discharged Discharged Discharged Discharged Discharged Discharged

Reintubated after initial extubation.

Materials and methods

From February to October, 1997, 20 of 22 patients (Table 1) successfully underwent completely laparoscopic or laparoscopic-assisted exclusion of aortic aneurysm with aortoiliac or aortofemoral bypass at Morristown Memorial Hospital. This approach to surgery on aneurysms of the abdominal aorta was initially offered to all patients with infrarenal abdominal aortic aneurysms who were considered medically acceptable candidates for the open operation. Patients with symptomatic rupture of their aneurysms were excluded, as were patients who had renal or other visceral arterial disease requiring surgical correction. Our current practice, which began after review of our first 12 cases, is to segregate the high-risk patients into a separate group for treatment with axillobifemoral bypass followed by laparoscopic complete exclusion of the aneurysm sac. To date, we have operated on nine patients in this category under an IRB-approved protocol designed as a pilot program to validate a new global risk assessment and stratification system and to compare the results of this approach with those of stent-graft trials currently in progress. We limit this initial article to the 20 patients who have undergone laparoscopic aneurysm exclusion with aortoiliac or aortofemoral bypass either among the first 12 patients or in the group of 8 patients who underwent successful surgery after the change in the patient selection process mentioned earlier. Two patients who were converted to the open retroperitoneal approach after completion of the laparoscopic dissection have been excluded from the outcomes report because their postoperative recovery paralleled the historical norms seen in the open patients. The first such patient was converted because we could not correlate the anatomy as observed laparoscopically with the findings on his preoperative aortogram. In the second patient, the aortic neck was buckled severely to the right and could not, in our opinion, be laparoscopically dissected circumferentially without risking injury to the inferior vena cava or left renal vein. Because few open elective aneurysmectomies have been performed at this institution since the beginning of this program in February, 1997, all comparisons between the laparoscopic and open patient populations, by necessity, were made with historical controls neither matched by risk stratification nor necessarily cared for in similar fashion in the perioperative period. Patients who died early in the postoperative period have been excluded from the calculations of hospital or ICU length of stay. Four of the open patients died within 72 h after surgery, and one laparoscopic patient (#8) died on the second postoperative day. Of the historical controls, 20% were operated on and cared for perioperatively by the same surgeon (JKE) who has performed all of the laparoscopic procedures. Therefore, we also have attempted to make separate comparisons between this cohort of controls and the laparoscopic group to account for the

so-called surgeon factor. Scatograms showing these comparisons with a single surgeon series of open vs. laparoscopic aneurysm surgery at our institution are shown in Figs. 1A1D. To estimate the learning curve for the performance of this procedure, we have also studied the patient outcome comparisons, taking into consideration only the last 12 laparoscopic cases (Table 2).

Patients. Among the patients whose operations were successful were 15 men and 5 women. The age range was 62 to 88 years. Aneurysm size ranged from 4.0 to 8.0 cm in maximum transverse diameter, and all had met size and/or other criteria for surgical treatment. In 2 patients, the operation was considered urgent because symptoms of abdominal and back pain associated with finding of aneurysm tenderness were present preoperatively. One patient had an asymptomatic contained rupture discovered incidentally on a preoperative computed tomography (CT) scan. All aortic aneurysms were infrarenal in location. In 12 patients, the aneurysmal disease involved one or both common iliac arteries. Nine patients also had distal iliac arterial occlusive disease, 4 had femoral artery aneurysms, and 7 (5 males and 2 females) had common femoroarterial occlusive disease requiring endarterectomy. Two patients also had popliteal aneurysms. Patients who had aneurysms involving the hypogastric arteries were not offered this operative approach.

Preoperative evaluation and preparation. All patients referred for abdominal aortic aneurysm surgery had already undergone, at the least, an ultrasound examination of their abdominal aortas. Several also had abdominal computerized axial tomography (CAT) scans. If a CAT scan had not already been obtained, we requested one, not only to obtain additional information on the aortic aneurysm itself, but also to look for other aneurysms that may have gone undetected on abdominal ultrasound because of the obscuring effects of bowel gas. Biplane aortograms were obtained on all patients to evaluate the aorta and its branches for the following: Length of the infrarenal aorta not involved by the aneurysm (the aortic neck) Configuration of the iliac arteries Patency as well as size of and direction of blood flow in the inferior mesenteric artery Patency of the middle sacral artery Patency, number, and location of large lumbar arteries Presence of significant stenoses or aneurysms of renal and other visceral

1066

Figs. 1. AD Scatograms showing outcome comparisons between a single-surgeon series of open and laparoscopic aneurysmectomy patients. (), Laparoscopic AAAs; (), open AAAs.

arteries, which by our present criteria for patient selection could constitute contraindications to this surgical approach. In keeping with our practice with patients undergoing open operations for abdominal aortic aneurysms, we prescribed mechanical as well as antibiotic bowel preparation for all patients on the day before surgery. All patients were admitted in a fasting state on the morning of surgery except patients 4, 7, and 20, for whose cases we felt the comorbid conditions warranted preoperative hospitalization for extra preparation. Prophylactic parenteral antibiotics were given preoperatively in all cases. We routinely used intermittent compression boots for DVT prophylaxis.

Anesthesia protocol. General endotracheal anesthesia is induced with the patient in supine position. Selection of the most appropriate anesthetic agent is left to the discretion of the anesthesiologist. We use perioperative epidural analgesia only in patients with marginal pulmonary reserve to reduce the need for narcotic pain control. As in open surgery, we enhance diuresis during surgery with mannitol (0.5 gm/kg bolus) given intravenously once the retroperitoneal access is gained and followed by a continuous infusion of mannitol at 5 gm/h for the remainder of the operation. A low-dose (3 to 5 g/kg/min) infusion of dopamine is started in some cases and continued for 12 to 24 h after surgery if there is preoperative evidence of renal insufficiency such as an elevated serum creatinine level, or if the urine output decreases significantly after application of the aortic cross-clamp. Radial artery and pulmonary artery catheters are placed in all patients before induction of general anesthesia in keeping with our practice

for traditional open operations. In addition to continuous systemic and pulmonary arterial pressure readouts, frequent cardiac output and index measurements are recorded. Arterial blood gasses, hemoglobin levels, platelet counts, and plasma potassium determinations are checked regularly and corrective actions taken as indicated. Following administration of heparin, activated clotting time (ACT) measurements are performed every half hour, and additional heparin boluses are given as needed to maintain the ACT at twice the baseline value. Despite relatively prolonged cross-clamp times, in our early experience, the need for pressors and afterload reducers to maintain cardiac indices after aortic cross-clamping appears to have been dictated by the presence and severity of coronary artery disease and left ventricular dysfunction. In our experience, only 2 of the 20 patients (#4 and #9) have required this pharmacologic support of cardiac function after application of the aortic cross-clamp. Routine use of the cell-saver to save and wash shed blood for reinfusion has significantly reduced the need for blood bank transfusions.

Technical elements of the procedure

Patient positioning. The patients torso is lifted and rotated so that the left shoulder is elevated about 60 degrees from the horizontal. The left upper extremity is supported on an arm board on the right side of the table. The pelvis is left in as neutral a position as possible to facilitate surgical access to both femoral regions. (In very obese patients, the operating table has to

1067 Table 2. Outcome comparisons: laparoscopic vs open Laparoscopic (n 20) Age % Male AAA size X-Clamp time Anesthesia time ICU days Length of stay Days NPO Days on vent EBL OR crystalloid OR crystalloid/h OR colloid OR cell saver OR PRBCs OR plts OR FFP Hospital PRBCs Hospital plts Hospital FFP Mortality rate 72.2 (6288) 75.00% (15/20) 5.98 cm 2 h, 26 min 7 h, 48 min 2.45 days 6.20 days 2.15 days 0.50 days 1713 ml 7595 ml 977 ml/h 275 ml 844 ml 1.95 U 0.80 U 0.40 U 2.75 U 6.50 U 2.50 U 10.0% (2/20) Last 12 laparoscopic (n 12) 72.3 (6285) 75.00% (9/12) 5.73 cm 2 h, 5 min 7 h, 18 min 1.50 days 5.92 days 1.08 days 0.42 days 1825 ml 8333 ml 1170 ml/h 250 ml 727 ml 1.92 U 0.00 U 0.18 U 1.92 U 1.33 U 1.25 U 0.00% (0/12) Open (n 100) 71.6 (5291) 76.0% (76/100) 6.12 cm (3.912.0) 1 h, 36 min 5 h, 1 min 3.22 days 9.97 days 5.35 days 2.18 days 1387 ml 6524 ml 1288 ml/h 303 ml 719 ml 1.31 U 0.60 U 0.18 U 1.46 U 0.90 U 0.91 U 4.00% (4/100)

be banked severely to the left so the panniculus can be rolled away from the right side when the right groin incision is made.) The table is then extended approximately 30 degrees before the bean bag is molded and placed on suction to maintain the patient position. The left thigh is elevated on either pillows or a folded bean bag to relax the left psoas muscle. Generous padding is used to alleviate pressure in areas of bony prominences and along the edges of the actuated bean bag (Fig. 2).

Initial retroperitoneal access. Initial retroperitoneal access is gained through a 1.5-cm muscle-splitting incision in the left flank. This incision is located just posterior to the anterior axillary line halfway between the rib cage margin and the iliac spine. A retroperitoneal tunnel is then created by digital dissection and directed to pass posterior to the inferior pole of the left kidney. A kidney-shaped preperitoneal distension balloon OMSPDBS2 (Origin Medsystems, Menlo Park, California) is then placed completely within the tunnel and inflated under visual inspection by means of a zero-degree laparoscope inserted through the shaft of the distension balloon. Inflation is continued until a sufficient retroperitoneal space has been created and familiar structures (e.g., the psoas muscle with its broad tendon, the genitofemoral nerve along its medial border, the left iliac artery, and the crossing left ureter as well as the left gonadal vein) are clearly visible through the inflated balloon. The dissecting balloon is then removed, and the freshly created space is accessed with a blunt-tipped balloon trocar (OMS-T10BT or OMS-T10BTS (Origin Medsystems), depending on the thickness of the abdominal wall. The retroperitoneum is insufflated with CO2 up to a maximum cavity pressure of 15 mm of mercury. The CO2 flow rate is kept at a constant maximum to compensate for anticipated gas loss during insertion and manipulation of laparoscopic instruments. The insertion of additional trocars through the sites shown in Fig. 2 is carried out under videoscopic observation, the main operating ports being inserted first, then followed, in due course, by the insertion of the left kidney retraction port once the left kidney has been mobilized. The aortic cross-clamp port is created much later during the gasless phase of the operation.

Fig. 2. Depiction of the patient position on the operating table.

Vascular dissection. The vascular dissection is begun at the level of the left common iliac artery. The common iliac artery is mobilized by circumferential dissection, and if an aneurysm of the common iliac is present, the iliac bifurcation or the distal extent of the iliac aneurysm must be exposed. (Because stapling is used to exclude the iliac arteries, the iliac vein must be dissected free of the posteromedial arterial wall (Fig. 3) to avoid inadver-

tent inclusion of the vein in the staple line.) The left kidney is then mobilized and held rotated in a cephalad and counterclockwise direction with an extra-hand balloon retractor (Origin Medsystems) to expose the perirenal portion of the abdominal aorta. This rotated position of the left kidney held with an extra-hand balloon retractor can in some cases be maintained by attaching the handle of the balloon retractor to a laparoscopic utility belt (SMC Surg-Med Devices Inc., Morristown, NJ). The aortic neck is then dissected circumferentially by clipping and dividing the lumbar tributaries of the left renal vein, the surrounding lymphatics, and the lumbar arteries and veins. The left renal artery must be identified in its new location anterior to the aorta because of the rotation of the left kidney. The proximal 2 cm of the aneurysm is similarly freed anteriorly and posteriorly. Care is taken to minimize the manipulation of the aneurysm sac during this dissection to avoid potential distal embolization of thrombus or atheromata. We have used two approaches to expose the right iliac artery. If there is aneurysmal involvement of the right common iliac artery, we feel it is safer to use a counter incision in the right lower abdominal quadrant for a retroperitoneal exposure of the right common iliac artery and its branches. However, if the right iliac artery is normal, it can be safely exposed circumferentially simply by continuation of the dissection already begun in the left retroperitoneal space. The aortic bifurcation is identified, and the course of the right iliac artery is followed. We have been able to expose the

1068

Fig. 3. The left common iliac vein is fully separated from the posteromedial arterial wall. Fig. 4. A bifurcated graft in a graft server is introduced through the left groin incision and advanced to a location adjacent to the divided neck of the abdominal aortic aneurysm. entire right iliac artery videoscopically down to the right inguinal ligament, even in very obese patients. This method greatly facilitates subsequent passage of the crossover limb of the bifurcated graft. During creation of this tunnel, the first assistant maintains an upward lift of the posterior peritoneum with a Pennington grasper or similar 1 cm-wide device. The right ureter must be clearly identified not only to avoid injury, but also to make sure that the right limb of the graft passes posterior to the ureter. Right and left groin incisions are then made to expose the femoral arteries. In four patients who had normal external iliac arteries and no femoral artery disease, we stayed out of the groins and obtained retroperitoneal exposure of the external iliac arteries through incisions above the inguinal ligaments. The tunnels are then completed between these incisions and the left retroperitoneal operative space. Once completed, ease of subsequent access through the crossover tunnel into the right incision is ensured by placing a 12-in. Penrose drain within it.

The gasless phase Gasless laparoscopic techniques are used for the remainder of the operation because pneumoretroperitoneum is difficult to maintain when tunnels to the incisions used to expose the femoral or external iliac arteries have been completed and certainly cannot be maintained once the graft server is introduced into the operative cavity through the left outflow incision. To achieve a seamless transfer from the gas to the gasless phase, a sequence of steps must be executed. The rotation of the left kidney is maintained with the extra-hand retractor. CO2 is discontinued. The laparoscopic trocars are removed from the initial access and the two main operative incisions. The initial access incision is widened to 2 cm, the laparolift device (fan, fan-lift/retractor, or air-lift; Origin Medsystems) is inserted, and the mechanical lift is used to reacquire the same exposure obtained during the gas phase. The laparoscope is reintroduced through the same port as the lift. The main operative channels are then reaccessed with disposable rubber thoracoscopy access ports, which are cut longitudinally to permit the surgical instruments a greater range of movement. The aortic cross-clamp incision and tunnel is made through the 10th intercostal space and maintained with a thoracoscopy port. Heparin (100 U/kg IV bolus) is then given, and 3 to 4 min later the iliac arteries are stapled distal to all aneu-

rysms using endoscopic T.A. 30 stapling devices (U.S. Surgical Corp., Norwalk, CT). (We have found that the 3.5-mm staples are adequate for most vessels. When we use the 4.8-mm staples, we have shoed the devices with PTFE or bovine pericardium strips.) The aortic neck is crossclamped just distal to the renal arteries. We use a specially designed laparoscopic aortic cross-clamp (Scanlan International Catalogue #9909-912-13 St. Paul, MN). Previously identified large lumbar arteries are clipped. If the inferior mesenteric artery is patent but not dominant, it is ligated. At this point, the surgeon must ascertain the absence of pulsation in the aortic sac. After the stapling of the common iliac arteries and the application of the cross-clamp to the aortic neck, the aneurysm sac becomes floppy and can be easily and safely maneuvered to seek and ligate any remaining lumbar branches. The aneurysm sac is partially opened and evacuated. (Continued bleeding from the aneurysm sac beyond the usual initial flash is indicative of a missed lumbar branch, which should be sought out and clipped.) A portion of the aneurysm adjacent to the neck should be excised and the remaining collapsed sac closed with a laparoscopic T.A. 60 stapling device (U.S. Surgical Corp., Norwalk, CT). We have found that excising a portion of the aortic sac near the neck greatly facilitates the laparoscopic suturing required for the proximal aorta to graft anastomosis, especially when the aneurysm sac is very large or buckled to the left. The neck is then prepared for the graft anastomosis by circumferential incision and definition. Any needed endarterectomy is performed at this juncture. In our experience so far, only five patients have not required endarterectomy of the aortic neck. An appropriately sized bifurcated woven Dacron graft is selected and loaded into a graft server (SMC Surg-Med Devices). The graft server is introduced into the operative cavity through the left groin or suprainguinal incision and is positioned next to the incised, defined aortic neck (Fig. 4). The graft-to-aorta anastomosis is performed with Teflonpledgeted individual horizontal mattress sutures. We have used specially fashioned double armed 3-0 polypropylene sutures on ski-shaped needles (specially manufactured

1069

Fig. 5. AE The sequence of sutures, which we recommend for performance of the proximal aorta-to-graft anastomosis, is illustrated.

by U.S. Surgical Corp.), but Dacron, PTFE, or other nonabsorbable material sutures may be used according to operator preference. The sequence of suture placement that we espouse is shown in Figs. 5A5E and is designed to ensure that visibility for accurate placement of the following suture is always preserved. Because of the laparoscopic magnification and better visibility, suture placement can be extremely accurate and although the suturing is relatively slow, a water-tight anastomosis can usually be achieved on the first attempt. The graft server is then removed. The right limb of the bifurcated graft is passed into the right groin or suprainguinal incision through the previously created cross-

over tunnel in the retroperitoneum. Anastomoses to both femoral or external iliac arteries are performed in the usual fashion, and previously established declamping and flushing techniques are followed to prevent declamping shock or distal embolization of air or thrombus. Completion peritoneoscopy is recommended in all cases.

Results Patient demographics as well as recorded operative and initial outcome data are shown in Table 1. Comparison data

1070

with 100 historic controls and outcome trends are shown in Table 2. Complications Two patients died within 30 days of operation: patient 4 who had several preoperation risk factors and succumbed to multisystem organ failure several weeks postoperatively and patient #8 who developed cardiogenic shock (diffuse myocardial hypokinesis on echocardiogram) immediately after surgery, probably due to acute myocardial infarction. By our present patient selection criteria, neither of these two patients nor, in retrospect, patients 7 and 9 would have been operated on in this fashion. Instead, according to their global risk assessment scores, they would have been segregated into the high-risk group and treated by one of the two alternative procedures discussed earlier. After our adjustment of the patient selection process, there have been no deaths and only one complication in the group of patients undergoing laparoscopic aneurysm exclusion with aortofemoral or aortoiliac bypass. Nine patients developed perioperative complications Two patients, #1 and #2, complained of right lower extremity paresthesias without focal neurologic findings in the immediate postoperative period. Patient 1 also had significant edema of the right buttock region and myoglobinuria. In both cases, the symptoms resolved spontaneously within 48 h. Despite the relatively prolonged cross-clamp times early in our experience, none of the patients showed evidence of declamping shock or acidosis. No patients developed calf swelling or tenderness. We therefore concluded that pressure on the right gluteal region and the sciatic nerve due to the patient position during surgery and anesthesia, which early in our experience lasted more than 10 h, was responsible for this phenomenon observed in the first two patients. In addition to placing the operating table in its final position before molding and actuating the bean bag, which prevents formation of hard ridges within the contour of the bean bag, we now place extra padding under the buttocks and along the edges of the activated bean bag. No additional patients have experienced these postoperative symptoms since we instituted these precautions. The duration of the entire operation also has been considerably shortened. Several of these procedures have now been completed in less than 6 h, the anesthesia time usually lasting about 112 h longer than the operation time in each case. Patient 7 developed occlusive thrombosis of right graft limb and nonocclusive thrombi on the left side requiring transfemoral thrombectomy 1 day postoperatively. His operation had been performed without heparin because of a preexisting coagulopathy. He suffered no sequelae and had intact peripheral pulses at discharge and office follow-up. Patient 3 sustained an avulsion injury of the left ureter, which was treated with a left nephrectomy. A nephrectomy was selected rather than ureteral repair because a large normally functioning contralateral kidney was present on preoperative studies and partly because of the aortic grafts proximity to the area of ureteral injury. In our opinion, this was a traction injury related to continued anterior retraction

of the ureter when the kidney dropped posteriorly on loss of the anterior counterclockwise rotation of the kidney. We now maintain the anterior displacement of the left kidney and ureter on a single balloon retractor so that differential traction between the left kidney and its pedicle is much less likely to occur. In patients with bulky left kidneys, we make a small trap-door incision by joining two of the trocar sites and use a Thompson deaver retractor to securely maintain the left kidney in a rotated, elevated position. Anticipated laparoscopic mechanized retraction systems are designed to achieve the same results without the need for additional incisions. Two patients who had preoperative renal insufficiency developed worsening azotemia postoperatively. One of them, patient 4, who also had many other comorbid illnesses, required hemodialysis and subsequently died of multisystem organ failure. In patient 9 the blood urea nitrogen (BUN) and creatinine returned to baseline levels within 6 weeks of surgery. He did not require hemodialysis. Only patients with preexisting pulmonary disease required ventilator support after surgery. Of the 20 patients, 13 were extubated either in the operating room on completion of the procedure or in the recovery room. Five patients were weaned from ventilator support by the morning after surgery. All patients were extubated within 48 h of the operation except for patient 7 who made a return trip to the operating room for thrombectomy of his graft. Patient 4 required reintubation, then developed adult respiratory distress syndrome (ARDS) and subsequently succumbed to multisystem organ failure. Patient 4 also developed invasive Clostridium difficile enterocolitis, which could not be controlled with oral Flagyl and vancomycin and required total abdominal colectomy. We believe that the C. difficile infestation probably predated his surgery because he had been on outpatient treatment with oral antibiotics for at least 2 weeks for a flu-like ailment before he was hospitalized with symptoms of impending rupture of his large abdominal aneurysm. He also developed acalculous cholecystitis late in his hospital course. This was treated initially with percutaneous drainage, then with cholecystectomy at the time of his total abdominal colectomy. Patients 4, 7, 8, 9, and 12 had postoperative ileus sufficiently severe to delay resumption of regular diet by more than 24 h. In our open series, this would not have been reported as a complication, but in the patients who underwent laparoscopic aortic surgery, failure to tolerate oral intake within 24 h after surgery was considered a departure from the predicted recovery pathway. Patient 17 exhibited transient paraparesis in the postoperative period without specific findings on CT scan or magnetic resonance imagery (MRI). His multiple muscle-group weakness was attributed to cauda equina or spinal cord ischemia, possibly due to emboli. He recovered full function of his lower extremities after 6 weeks of physical therapy. Discussion: caveats, potential intraoperative complications, and lessons learned Violation of the peritoneal layer The previously described operation is heavily dependent on the maintenance of the created retroperitoneal space from

1071

which bowel and other viscera are effectively excluded by an intact peritoneal wall. Our animal laboratory experience would seem to indicate that a large rent in the peritoneum, especially during the initial retroperitoneal access or during the gas phase of the operation, is probably not repairable, and conversion to an open procedure would, in all likelihood, be required. A small, at times invisible, breach of the peritoneum may occur and would be manifested by increasing abdominal distension and contraction of the retroperitoneal space during CO2 insufflation. Respiratory acidosis also can theoretically result despite aggressive high minute volume ventilation because of a significant increase in CO2 absorption. If this occurs, the peritoneal cavity should be vented with a supraumbilical blunt-tipped trocar, and the operation can then be continued as previously described. Although we encountered this problem in the animal laboratory, it has not occurred in our clinical experience, probably because the peritoneal layer is much thinner and therefore more easily torn in the porcine model. Rents that occur in the peritoneal layer during the gasless phase of the procedure are not nearly as threatening to the videoscopic approach and frequently can be repaired.

distension balloon or transiently increasing the maximum cavity pressure to 30 mmHg frequently leads to a drier, more manageable operative field. Infusion of blood and blood products should be based on previously established criteria.

Surgical wound problems Due to the present technical limitations of the preceding approach, a bifurcated graft with an aortobilateral femoral bypass probably will be required in many cases. In following the technique described by Shah and his associates [9], some of the outflow anastomoses can be performed to the external iliac arteries above the inguinal ligaments, thereby avoiding the groin altogether, which we have been able to do in four of the last eight patients. Because groin wounds have traditionally been associated with a higher incidence of complications such as seromas, lymph leaks, and infection, the surgeon must exercise extra care in the management of these wounds during and immediately after surgery. All of the lymph-bearing tissue is mobilized medially as a single pad. Any divided lymph-bearing tissue must be ligated rather than cauterized, and closure of groin incisions must be watertight. Frequently, changed occlusive dressings are recommended, especially in obese patients. Any wound seroma and/or drainage should be managed aggressively, and wound reexploration in the operating room should be carried out immediately if the integrity of the wound closure is threatened. We have not experienced any groin wound problems while following the aforementioned recommendations.

Collateral organ damage Structure such as the ureters, the gonadal veins, the renal vasculature, the inferior vena cava, the genitofemoral nerve, and the sympathetic nerve chain, which are located within the operative field especially close to the aorta and iliac arteries, are subject to surgical injury unless they are clearly identified and their locations taken into consideration during each surgical maneuver. We have encountered an avulsion injury of the left ureter in one case when the left kidney was permitted to drop posteriorly while the renal pedicle was still being independently retracted forward. Since then, it has been our practice to retract the kidney and its pedicle as a single unit. A videoscopically directed small incision is made and used for the introduction of a Thompson deaver retractor in patients with bulky left perirenal fat. These small incisions do not appear to affect adversely the postoperative course of these patients.

Conclusion Our experience, although early and still quite limited, has demonstrated the feasibility of a completely videoscopic approach to the posterior retroperitoneal management of infrarenal abdominal aortic aneurysms. The initial outcome in our first group of patients would seem to indicate potential significant benefits for low- and moderate-risk patients compared with that of the traditional open transperitoneal or retroperitoneal procedures. In patients deemed to be at high risk, we now perform axillobifemoral bypass followed by laparoscopic exclusion of the aneurysm sac, or we recommend endovascular repair if the anatomic requirements are met. This approach, in our opinion, should virtually eliminate any warm ischemia time of the lower extremities and significantly reduce the usual increase in cardiac afterload and release of inflammatory mediators associated with infrarenal aortic cross-clamping, thus possibly improving the outcome in these high-risk patients. We are also confident that technologic developments in the field of videoscopic surgery, particularly in the area of mechanized retraction and automated suturing, will soon facilitate many crucial steps of this operation and lead to more widespread application. The next generation of endoscopic automatic suturing devices will speed up the proximal anastomosis and also make videoscopic distal aortic and common iliac artery anastomoses possible, thereby possibly limiting the need for groin incisions to only those patients with documented external iliac artery and/or femoral artery occlusive disease.

Bleeding Due to the limitations of suction and access during videoscopic operations, bleeding is more easily avoided than controlled. Bleeding may occur at any stage of the operation, but is most likely to be encountered during the exposure, ligation, and division of lumbar vessels required to expose the aortic neck. If bleeding occurs during the gas phase, venous injuries may constitute a greater risk to the patient owing to the possibility of gas embolism. Large venous injuries may therefore be best treated by discontinuation of CO2 insufflation and immediate conversion to an open procedure. Arterial bleeding, however, can usually be handled by temporary application of pressure followed by suture or endoclip ligation. In some patients, the retroperitoneum was quite sticky, which led to considerable oozing during the initial dissection. In such cases, we have found that prolonging the initial tamponade provided by the preperitoneal

1072 Acknowledgments. Illustrations were prepared by Grace Moore, Mark Palangio, and Delia V. Edoga. The authors owe a debt of gratitude to Mr. Che Edoga for the use of his computer skills in the initial design and verification of operative techniques and strategies and for the final format of the illustrations. We also thank Mr. Ken Zeiher and his associates at Origin Med Systems for providing us with laboratory facilities and a great deal of technical assistance. 6. Moore WS, Rutherford RB, for the EVT investigators (1996) Transfemoral endovascular repair of abdominal aortic aneurysm: results of the North American EVT phase 1 trial. J Vasc Surg 23: 543553 7. Nasim A, Thompson MM, Sayers RD, Bolia A, Bell PR (1996) Endovascular repair of abdominal aortic aneurysm: an initial experience. Br J Surg 83: 516519 8. Resnikoff M, Darling RC, Chang BB, Lloyd WE, Paty PSK, Leather RP, Shah DM (1996) The fate of the excluded abdominal aortic aneurysm sac: long-term follow up of 831 patients. J Vasc Surg 25: 851855 9. Shah DM, Chang BB, Paty PSK, Kaufman JL, Koslow AR, Leather RP (1991) Treatment of abdominal aortic aneurysm by exclusion and bypass: an analysis of outcome. J Vasc Surg 13: 1522 10. Sicard GA, Reilly JM, Rubin BG, Thompson RW, Allen BT, Flye MW, Schechtman KB, Young-Beyer P, Weiss C, Anderson CB (1995) Transabdominal versus retroperitoneal incision for abdominal aortic surgery: report of a prospective randomized trial. J Vasc Surg 21: 174183 11. White GH, May J, McGahan T, Yu W, Waugh RC, Stephen MS, Harris JP (1996) Historic control comparison of outcome for mathed groups of patients undergoing endoluminal versus open repair of abdominal aortic aneurysms. J Vasc Surg 23: 201212 12. Williams GM, Ricotta J, Zinner M, Burdick J (1980) The extended retroperitoneal approach for treatment of extensive atherosclerosis of the aorta and renal vessels. Surgery 88: 846855

References
1. Chen MH, DAngelo AJ, Murphy EA, Cohen JR (1996) Laparoscopically assisted abdominal aortic aneurysm repair: a report of 10 cases. Surg Endosc 10(12): 11361139 2. Creech O Jr (1966) Endoaneurysmorrhaphy and treatment of aortic aneurysm. Ann Surg 164: 935946 3. Darling RC, Shah DM, Chang BB, Paty PS, Leather RP (1996) Current status of the use of retroperitoneal approach for reconstructions of the aorta and its branches. Ann Surg 224: 501508 4. Huber TS, Flynn TC, Albright JL, Seeger JM (1995). Operative mortality rates after elective infrarenal aortic reconstructions. J Vasc Surg 22: 287294 5. Kazmers A, Jacobs L, Perkins A, Lindenauer SM, Bates E (1996) Abdominal aortic aneurysm repair in Veterans Affairs medical centers. J Vasc Surg 23: 191200

New technology
Surg Endosc (1998) 12: 10881090 Springer-Verlag New York Inc. 1998

Endoscopic ultrasound
Diagnostic and therapeutic uses
T. A. Knox
Division of Gastroenterology, New England Medical Center and Tufts University School of Medicine, 750 Washington Street, Boston, MA 02111, USA Received: 29 September 1997/Accepted: 29 December 1997

Abstract. Endoscopic ultrasound (EUS) is an emerging technique that can be used to visualize lesions in or adjacent to the gastrointestinal tract. We performed a review of current indications and capabilities of endoscopic ultrasound. EUS provides diagnostic information on submucosal lesions of the gastrointestinal tract. It is the most accurate method for local staging of esophageal and rectal cancer. It appears to be more sensitive in detecting common bile duct stones than endoscopic retrograde cholangiopancreatography. Furthermore, its use avoids the risk of pancreatitis. Lesions deep to the gastrointestinal tract, such as mediastinal nodes or pancreatic lesions, can be biopsied through the EUS endoscope. New therapeutic uses of EUS include aid in endoscopic drainage of pancreatic pseudocysts and celiac axis neurolysis for the treatment of pain from pancreatic cancer. EUS offers many new diagnostic capabilities to visualize and biopsy lesions in or adjacent to the gastrointestinal tract. Therapeutic uses are also emerging for this new technique. Key words: Endoscopic ultrasound Sonography

fiberoptic endoscope. Sonographic examination of the lesion can establish its exact location in relation to the layers of the intestinal wall and local structures. Sonographic texture and echogenicity help to identify the likely components of the lesionsuch as tumor, fat, or fluidand to determine blood flow. At the same time, new therapeutic applications are being introduced for the EUS scope.

Technology The current generation of EUS endoscopes provides a fiberoptic channel for the visual identification of gastrointestinal lesions and to maneuver the endoscope through the pylorus. The instrument has a suction and biopsy channel with an elevator to assist in control of the biopsy forceps or needle aspiration device. The sonographic transducer is mounted on the distal tip of the endoscope and attached to an external imaging processor with photographic capabilities. EUS examinations can now be recorded on videotape, photographs, or radiographic film. There are two different types of EUS endoscopethe Olympus model and the Pentax [5]. In addition, some types of probes are available that fit through the biopsy channel of a regular endoscope; however, these will not be discussed here. The Olympus EU-M20 (Olympus, Lake Success, NY, USA) provides radial imagingthat is, it provides a complete, circular ultrasound image in a thin plane perpendicular to the axis of the endoscope (Fig. 1). This is the easiest instrument for orienting the endoscopist in the gastrointestinal tract, and it provides good-quality images. The frequency of the ultrasound waves can be selected at 7.5 or 12 MHz. However, because the thin plane of ultrasound imaging is at right angles to the biopsy channel, obtaining accurate submucosal biopsies may be difficult. Because of the constant revolution of the ultrasound transducer at the tip of

External ultrasound scanning is limited in its ability to evaluate chest and abdominal lesions by the presence of air in the bowels and lungs, which prevents transmission of sound waves. In addition, its depth of penetration is also limited. However, the development in the 1980s of ultrasound transducers attached to the tips of endoscopes has obviated these problems. Endoscopic ultrasound (EUS) now allows the physician to inspect a lesion visually through the

Correspondence to: T. A. Knox

1089

a substitute for pathologic diagnosis by biopsy, which can be obtained through the same endoscope. Diagnostic uses EUS is ideally suited for the identification of submucosal lesions in the gastrointestinal tract and masses adjacent to the upper gastrointestinal tract. Common lesions such as leiomyomas and lipomas have characteristic sonographic patterns and typical locations in the five-layered wall structure of the gastrointestinal tract. In addition, unexplained external compression of the gastrointestinal tract from adjacent organs such as spleen, gallbladder, varices, or pancreatic processes can be identified. Cancer staging is greatly facilitated by EUS examination. Because of the ability of EUS to determine wall layer involvement of the esophagus and to image adjacent nodes, both T and N staging of esophageal cancer staging are much more accurate with EUS than with CT scan (T stage accuracy is 88% for EUS and 59% for CT scan; N stage accuracy is 74% for EUS and 54% for CT scan) [1, 2]. Similarly, EUS provides accurate staging of gastric adenocarcinoma and gastric lymphoma. It can be useful in distinguishing benign lesions from infiltrative lesions of the stomach, however, deep biopsies are still required to confirm the diagnosis. The staging of rectal cancer now depends on EUS examination either with the Olympus or Pentax scopes or with dedicated rectal probes. The staging of colon cancer, although feasible, is more difficult using the side-viewing EUS scopes and does not contribute to clinical management. EUS is becoming more valuable in the diagnosis of pancreaticobiliary diseases [1]. It can detect changes of chronic pancreatitis, small pancreatic pseudocysts and their relationship to the ducts, and small masses in the pancreas. It has been used to stage early pancreatic cancer and is more accurate than angiography in detecting vascular invasion by tumor. Neuroendocrine tumors can also be localized with greater accuracy with EUS (82%) than with angiography, external ultrasound, or CT scan. Recent work suggests that EUS is more sensitive and has a lower complication rate than endoscopic retrograde cholangiopancreatography (ERCP) in detecting choledocholithiasis before laparoscopic cholecystectomy. It is possible to biopsy submucosal lesions, mediastinal lymph nodes, and pancreatic lesions through the EUS scope. Real-time EUS guidance of the biopsy aspiration needle is also possible with the Pentax linear array endoscope [3]. Therapeutic uses The EUS endoscope has increased the safety and accuracy of a number of new therapeutic techniques. Endosonography can identify the penetrating vessel causing Dieulafoys lesions of the stomach and allow targeted hemostasis with electrocautery or sclerotic injection. EUS is also used to identify the structures between the stomach wall and a pancreatic pseudocyst before endoscopic drainage. The Pentax echoendoscope is particularly useful for this application because of its Doppler capability, which identifies vessels clearly. Celiac axis neurolysis, using the Pentax echoendo-

Fig. 1. Schematic drawings of radial scanner (Olympus type) and linear array (Pentax type) echoendoscopes scanning a submucosal lesion in the stomach. The images generated by the echoendoscopes are depicted in the shaded boxes at the right. (Illustration by D. Paquette, CMI, Educational Media Center, New England Medical Center, Boston, MA, USA.)

the endoscope, this instrument is subject to breakage unless it is handled very carefully. The Pentax FG-32UA (Pentax, Orangeburg, NY, USA) linear array or sector scanner is a somewhat more rigid instrument that has no moving parts. The linear array transducer at the distal tip provides a 100 sector image in a plane oriented along the length of the endoscope instead of perpendicular to it (Fig. 1). It scans at lower frequencies (5 or 7.5 MHz) than the Olympus. A biopsy needle can be introduced into the visual field and followed into a lesion under ultrasound guidance. In addition, the Pentax model has color Doppler capability for identifying blood flow and direction. However, the limited view makes learning, identifying lesions, and orienting the endoscope more difficult.

Normal mucosal layers The five layers of the intestinal wall structure are seen on EUS as alternating bright and dark rings: mucosa/balloon interface (bright), muscularis mucosa (dark), submucosa (bright), muscularis propria (dark), and serosa or adventitia (bright) [1]. Lesions such as leiomyomas, which are composed of muscular tissue and arise in the muscularis propria, will appear as isolated thickening of the dark (hypoechoic) muscularis propria layer. Cancer usually appears hypoechoic, but it destroys the wall layer structure. Careful identification of the wall layers and the echogenicity of the lesion aids in the diagnosis and staging of submucosal lesions and tumors. However, EUS should not be regarded as

1090

scope, with transgastric injection of absolute alcohol into the celiac plexus, has been shown to be a safe and effective method of pain control for patients with intraabdominal malignancy [4]. In conclusion, EUS is a new technique that has many valuable diagnostic and therapeutic applications. It is a difficult tool to master because it requires both advanced endoscopic and ultrasonographic skills. However, the new diagnostic and therapeutic advances in this areaparticularly those related to the diagnosis and staging of cancer and the management of pancreaticobiliary diseasesmake it an important new tool for the surgeon.

References
1. Rosch T, Classen M (1992) Gastroenterologic endosonography. Thieme, Stuttgart, New York 2. Tio TL, Coene PPLO, Schouwink MH, Tytgat GNJ (1989) Esophagogastric carcinoma: preoperative TNM classification with endosonography. Radiology 173: 411417 3. Wiersema MJ, Wiersema LM, Khusro Q, Cramer HM, Tao LC (1994) Combined endosonography and fine-needle aspiration cytology in the evaluation of gastrointestinal lesions. Gastrointest Endosc 40: 199206 4. Wiersema MJ, Wiersema LM (1996) Endosonography-guided celiac plexus neurolysis. Gastrointest Endosc 44: 656662 5. Zerbey AL, Lee MJ, Brugge WR, Mueller PR (1996) Endoscopic sonography of the upper gastrointestinal tract and pancreas. Am J Radiol 166: 4550

Surg Endosc (1998) 12: 10351038

Springer-Verlag New York Inc. 1998

Increased tumor establishment and growth after open vs laparoscopic bowel resection in mice
J. D. F. Allendorf, M. Bessler, K. D. Horvath, M. R. Marvin, D. A. Laird, R. L. Whelan
Department of Surgery, Columbia University College of Physicians and Surgeons and the Columbia Presbyterian Medical Center, 161 Fort Washington Avenue, New York, NY 10032, USA Received: 27 October 1997/Accepted: 19 January 1998

Abstract Background: Surgery can suppress immune function and facilitate tumor growth. Several studies have demonstrated better preservation of immune function following laparoscopic procedures. Our laboratory has also shown that tumors are more easily established and grow larger after sham laparotomy than after pneumoperitoneum in mice. The purpose of this study was to determine if the previously reported differences in tumor establishment and growth would persist in the setting of an intraabdominal manipulation. Methods: Syngeneic mice received intradermal injections of tumor cells and underwent either an open or laparoscopic cecal resection. In study 1, the incidence of tumor development was observed after a low dose inoculum; whereas in study 2, tumor mass was compared on postoperative day 12 after a high-dose inoculum. Results: In study 1, tumors were established in 5% of control mice, 30% of laparoscopy mice, and 83% of open surgery mice (p < 0.01 for all comparisons). In study 2, open surgery group tumors were 1.5 times as large as laparoscopy group tumors (p < 0.01), which were 1.5 times as large as control group tumors (p < 0.02). Conclusion: We conclude that tumors are more easily established and grow larger after open laparoscopic bowel resection in mice. Key words: Laparoscopy Surgery Mouse Tumor establishment Tumor growth Mouse mammary carcinoma

scopic techniques. Our laboratory has developed several small-animal models to study the immunological and oncological consequences of both laparoscopic and open surgery [3]. In rats, we compared immune function by serial delayed-type hypersensitivity (DTH) testing and demonstrated that cell-mediated immunity is better preserved after pneumoperitoneum than after sham laparotomy through postoperative day 5 [8]. Similar results were observed when rats underwent either an open or laparoscopic-assisted bowel resection [2]. Having documented postoperative differences in immune function and understanding that the immune system has been shown to play an integral role in defense against tumor establishment and proliferation for a number of malignancies, we hypothesized that there may be oncological benefits to minimally invasive surgery. Using a mouse model, we demonstrated that intradermal tumors are more easily established and grow larger after sham laparotomy than after pneumoperitoneum [1]. The goal of the present study is to determine if the differences in tumor establishment and growth observed after sham procedures would persist in the setting of an intraabdominal manipulation. Materials and methods Animals
Five- to six-week-old female C3H/He mice (Charles River Laboratories, Wilmington, MA, USA) were used for both experiments. These mice are immunocompetent and syngeneic to the mouse mammary carcinoma tumor line. All studies were performed under protocols approved by the Columbia University Institutional Animal Care and Use Committee in accordance with FDA regulations. The animals were acclimated to a climate- and light-cyclecontrolled environment for 24 h prior to investigations. Mice were fed standard laboratory rodent chow and tap water ad libitum.

Although the clinical benefits of minimally invasive surgery in terms of postoperative pain, length of hospital stay, and time to return to work have been well established [47], there may be additional physiological benefits to laparo-

Tumor cell line

Both studies involved the use of mouse mammary carcinoma (MMC) cells derived from the MC2 cell line [9]; they were obtained from Dr. J. Vaage

Correspondence to: R. L. Whelan

1036 of the Roswell Park Cancer Institute. Mouse mammary carcinoma is an immunogenic cell line that shows a plateau of maximal growth from 12 to 14 days after tumor cell inoculation. It is syngeneic to the C3H/He mouse strain. At a dose of 10,000 tumor cells (study 1), <10% of control mice are expected to develop a tumor nodule. At a dose of 1,000,000 tumor cells (study 2), >95% of control mice are expected to develop tumors. Eventually, 20% of these tumors are expected to spontaneously regress.

Tumor cell preparation and inoculation

On the day of operative intervention, tumor cells were prepared as a singlecell suspension for intradermal inoculation. MMC cells growing freefloating in RPMI 1640 medium supplemented with 10% fetal calf serum, 150 U/ml penicillin, and 150 mg/ml streptomycin were washed twice, counted, and resuspended in phosphate-buffered saline. In study 1, a suspension of 105 cells per ml was prepared, and mice were injected with 0.1 ml, for a total inoculum of 104 cells. In study 2, a suspension of 107 cells per ml was prepared, and mice were injected with 0.1 ml, for a total inoculum of 106 cells. Tumor cell viability was determined to be >95% by trypan blue exclusion. On the day of intervention, mice were restrained, shaved, and injected in the dorsal skin with 0.1 ml of tumor cell suspension prior to beginning the surgical interventions.

Fig. 1. Incidence of tumor establishment by postoperative day 30 after open versus laparoscopic cecectomy. *p < 0.01 versus control and open surgery, **p < 0.01 versus control and laparoscopy.

Study 2: C3H/He high-dose tumor inoculum

A total of 112 C3H/He mice received an intradermal inoculation of 106 MMC cells on the day of operative intervention. Animals were randomly assigned to one of three groups. As in study 1, anesthesia control mice underwent no procedure, laparoscopic resection group mice underwent a laparoscopic-assisted cecectomy, and open resection group mice underwent a cecectomy through a midline incision. Tumors were excised and weighed on postoperative day 12. This time point was chosen because it represents the beginning of the plateau phase in the normal growth curve of mouse mammary carcinoma. All data were collected in a blinded fashion (n 38 for anesthesia control and open resection, n 35 for laparoscopic resection).

Studies Study 1: C3H/He threshold-dose tumor inoculum

A total of 115 C3H/He mice received an intradermal inoculation of 104 MMC cells on the day of operative intervention. All animals were anesthetized by intraperitoneal injection of ketamine (50 mg/kg) and xylazine (5 mg/kg) in a total volume of 0.3 ml. After being anesthetized, the animals were randomly divided into one of three groups:

Statistics
Group 1. Anesthesia control animals underwent no procedure and were returned to their cages. In study 1, differences among groups were analyzed for statistical significance by chi-square using Yates correction for small numbers. In study 2, differences among groups were analyzed for statistical significance by ANOVA, followed by Students t-test to determine p values.

Group 2. Laparoscopic cecectomy group mice underwent a laparoscopicassisted cecal resection. The procedure was performed as previously described [5]. Briefly, the mouse was placed in Trendelenberg position and the abdomen was insufflated to a pressure of 35 mmHg with carbon dioxide gas through a 25-gauge cannula placed in the right upper quadrant. A 4-mm rigid laparoscope was inserted through a small incision created in the midline just caudal to the xiphoid, and a 2-mm operative port was created in the right lower quadrant. Under laparoscopic visualization, the cecum, which in mice is 1 to 2 cm in length, was grasped at its end and exteriorized. Extracorporeally, the cecum was ligated just distal to the ileocecal junction. The cecum was then resected and the stump was irrigated before being gently returned to the peritoneal cavity.

Results Study 1 By postoperative day 30, tumor nodules developed in 5% of control mice, 30% of laparoscopic resection group mice, and 83% of open resection group mice (p < 0.01 by chisquare for all comparisons), (Fig. 1).

Group 3. Open cecectomy group animals underwent a cecal ligation and resection through a 4-cm midline incision. The operative time was standardized to 20 min for both procedures (n 40 for anesthesia control and laparoscopic resection, n 32 for open resection).

Study 2 Tumor nodules developed in all mice. Tumor mass on postoperative day 12 showed a stepwise increase from the control group to the laparoscopic resection group to the open resection group. The open resection group tumors were 1.5fold larger than laparoscopic resection group tumors and more than twice as large as control group tumors (p < 0.01 by Students t-test for both comparisons). Laparoscopic resection group tumors were 1.5 times as large as control group tumors (p < 0.02 by Students t-test) (Fig. 2). (Control

Mice were assessed weekly by blinded palpation of the dorsal skin for the presence or absence of a tumor nodule. If a tumor nodule was palpated, the mouse was killed, and the dorsal skin was reflected to confirm the presence of a tumor nodule by direct observation. Weekly observations were necessary so as not to miss tumors that established and regressed in the 1st postoperative month. On postoperative day 30, the remaining mice were killed, the dorsal skin overlying the tumor cell injection site was reflected, and the presence or absence of a tumor nodule was determined by direct observation. Several random specimens were analyzed histologically to confirm the presence of tumor. All data were collected in a blinded fashion.

1037

Fig. 2. Tumor mass (SD) on postoperative day 12 after open versus laparoscopic cecectomy. *p < 0.02 versus control, **p < 0.01 versus control and laparoscopy.

75 68 mg, laparoscopic resection 115 68 mg, open resection 180 132 mg.) Morbidity and mortality Two mice in both the open and the laparoscopic resection groups were sacrificed intraoperatively for inadvertent injury to the bowel when gaining access to the peritoneum (2.8% vs 2.6% complication rate). There were no leaks from the site of the cecal ligation, and none of the mice died postoperatively. Discussion Previous work in our laboratory demonstrated that tumors are more easily established and grow larger after sham laparotomy than after pneumoperitoneum in mice [8]. These initial studies did not involve an intraabdominal procedure. In the current study, we overcame this hurdle by using a newly developed mouse model of laparoscopic bowel resection [5]. Using the same mouse strain and tumor line as in our previous studies, laparoscopic-assisted and open cecectomy were compared. As shown in Fig. 3, the long cecum of the mouse permits resection of 12 cm of bowel without requiring a technically challenging anastomosis to reestablish bowel continuity. The model proved to be safe, reliable, and economical. There were no postoperative leaks and the intraoperative complication rate was similar to that of the open procedure (2.6% vs 2.8%). Using this model, we determined tumor establishment (study 1) and tumor growth (study 2) after open and laparoscopic bowel resection in mice. In study 1, a threshold dose of tumor cells was injected into the dorsal skin immediately prior to surgical intervention. By postoperative day 30, tumors had developed in 83% of open resection group mice, 30% of laparoscopic resection group mice, and 5% of control group mice (p < 0.01 for all comparisons), (Fig. 1). These results are similar to those observed previously after sham interventions; significant differences between the laparoscopic and open groups persisted despite the addition of an intraabdominal procedure. Although not proven by this study, these results suggest that viable tumor cells remaining after resection would be better

Fig. 3. Normal anatomy of the mouse colon, ileum, and cecum.

able to form a metastatic tumor nodule after open resection than after laparoscopic-assisted resection. In study 2, a high dose of tumor cells was intradermally injected immediately prior to surgical intervention in order to study postoperative tumor growth. As expected, tumors developed in all animals and were excised and weighed on postoperative day 12. Open resection group tumors were 1.5 times as large as laparoscopic resection group tumors (p < 0.01), which were 1.5 times as large as control group tumors (p < 0.02) (Fig. 2). Despite the addition of an intraabdominal procedure, significant differences between groups persisted. These results suggest that tumor left behind at the time of surgery may grow less rapidly after laparoscopic resection than after open surgery. Thus, studies 1 and 2 demonstrate that the previously reported differences in tumor establishment and growth after sham procedures persisted in the setting of a bowel resection. While several authors are actively investigating what factors may contribute to the development of tumor nodules at trocar sites, our investigation aimed to understand how systemic postoperative physiology affects tumor establishment and growth. We accomplished this by studying tumor behavior at a site distant from the surgical manipulation. Our model was designed to separate the local effects of carbon dioxide pneumoperitoneum and surgical manipulation from the systemic effects of postoperative physiology. The mechanism of port site recurrence needs to be studied with the understanding that techniques designed to minimize surgical trauma should limit, not enhance, tumor establishment and growth. We conclude that tumors are more easily established and grow larger after open than after laparoscopic bowel resection in mice. Additional investigation is necessary to assess the behavior of other tumor lines and to identify the factors involved in the mechanism of differential tumor growth.
Acknowledgments. This investigation was made possible by generous support from the Association of Women Surgeons, the Ethicon division of Johnson and Johnson Corporation, and the United States Surgical Corporation.

1038

References
5. 1. Allendorf JDF, Bessler M, Kayton ML, Oesterling SD, Treat MR, Whelan RL (1995) Increased tumor establishment and growth after laparotomy versus laparoscopy in a murine model. Arch Surg 130: 649653 2. Allendorf JDF, Bessler M, Whelan RL, Trokel M, Laird DA, Terry MB, Treat MR (1996) Better preservation of immune function after laparoscopic-assisted vs. open bowel resection in a murine model. Dis Colon Rectum 39 (Suppl):S6772 3. Allendorf JDF, Bessler M, Whelan RL (1997) A murine model of laparoscopic-assisted intervention. Surg Endosc 11: 622624 4. Barkun JS, Barkun AN, Sampalis JS (1992) Randomized controlled trial 6. 7. 8.

9.

of laparoscopic versus mini cholecystectomy. Lancet 340: 1116 1119 Gadacz TR, Talamini MA (1991) Traditional versus laparoscopic cholecystectomy. Am J Surg 161: 336338 Grace PA, Quereshi A, Coleman J (1991) Reduced postoperative hospitalization after laparoscopic cholecytectomy. Br J Surg 78: 160162 Reddick EJ, Olsen DO (1989) Laparoscopic laser cholecytectomy. Surg Endosc 3: 131133 Trokel MJ, Bessler M, Treat MR, Whelan RL, Nowygrod R (1994) Preservation of immune response after laparoscopy. Surg Endosc 8: 13851387 Vaage J, Pepin K (1985) Morphological observations during developing concomitant immunity against a C3H/He mammary tumor. Cancer Res 45: 659666

Surg Endosc (1998) 12: 10731077

Springer-Verlag New York Inc. 1998

Laparoscopic approach to hydatid liver cysts

Is it logical? Physical, experimental, and practical aspects
A. Bickel,1 G. Daud,1 D. Urbach,2 E. Lefler,3 E. F. Barasch,1 A. Eitan1
1

Department of Surgery, Western Galilee Hospital, Nahariya, Post Office Box 21, 22100, Israel; and the Rappaport School of Medicine, the Technion, Israel Institute of Technology, Israel Faculty of Biomedical Engineering, Tel-Aviv University, Israel 3 Department of Microbiology, Western Galilee Hospital, Nahariya, Post Office Box 21, 22100, Israel
2

Received: 10 March 1997/Accepted: 20 November 1997

Abstract Background: In recent years attempts have been made to treat hydatid liver cysts laparoscopically. The purpose of this study was to evaluate different aspects of this approach and to examine whether a reasonable model could be developed. Methods: Three different subjects were analyzed. In the first, physical aspects related to transmembrane pressures were analyzed to demonstrate that evacuation of the cyst under pneumoperitoneum does not carry increased risk of spillage, and may even offer an advantage when the proper technique is used. In the second subject, an isolated liver model of a goat was used to study several techniques for evacuating hydatid cysts without spillage. This was tested qualitatively by demonstrating scolices in the fluid medium around the isolated liver after surgical manipulations. In the third subject, the implication of the technique was evaluated in human patients. Results: According to basic physical assumptions, the following conclusions were reached: (1) The increase in intracystic pressure is equal to or less than the increase in intraperitoneal pressure after pneumoperitoneum. (2) Aspiration of parasitic cysts by laparoscopic needle through a large cannula under vacuum or by sealing the cannula and adhering it to the liver by cyanoacrylate or fibrin glue was found to be very safe. Simple needle aspiration failed to prevent spillage. (3) A new transparent cannula 18 mm in diameter with a beveled tip was designed that enables good accessibility to liver cysts and safe evacuation even of huge and complex cysts. Conclusions: The novel technique to manage hydatid liver cysts, described in the study, is feasible, sensible, and safe.

The isolated goat liver containing hydatid cysts can be used as a reliable animal model to test new techniques in the future. Key words: Hydatid liver cyst Laparoscopic approach Echinococcal cyst Pneumoperitoneum

Presented at the Joint Euro-Asian Congress of Endoscopic Surgery, Istanbul, Turkey, 1721 June 1997 (3rd Asian-Pacific Congress of Endoscopic Surgery and 5th Annual Congress of the European Association for Endoscopic Surgery) Correspondence to: A. Bickel

In recent years, since the revolutionary development of laparoscopic surgery, there have been various attempts to treat hydatid cysts of the liver laparoscopically [1, 4, 12, 13, 15, 16, 19, 20]. In general, the main purpose of the surgical treatment is to neutralize the parasite and excise it together with the cystic germinal layers, while avoiding any spillage of the pressurized contents into the peritoneal cavity. The aims are to prevent anaphylaxis, dissemination of disease, and recurrence. Obviously, in cases of giant cysts, we prefer to avoid total excision of the pericyst. The laparoscopic approach may be imagined as the evacuation and removal of an elastic membrane through the external wall of a playball, which contains a noxious fluid under high pressure, through a small aperture in a surrounding sealed box. This must be accomplished without rupturing its wall and with no spillage. Although the exact physical properties of the ball (cyst) membranes are not defined, it seems to us that basic physical principles can be applied to analyze such a model, with implications concerning the laparoscopic approach in vivo. When a novel surgical technique is proposed, it should meet several criteria. First, a convincing argument should be made that the laparoscopic approach to hydatid cysts of liver is reasonable and does not carry a greater risk of spillage than the conventional open approach. In other words, there must be assurance that manipulations under pressure (pneumoperitoneum) are safe. Second, an experimental model should be implemented to test the efficacy and safety of various techniques. Third, the technique should be efficacious. This study was divided into three sections, and we shall

1074

apply the same force in the opposite direction according to Newtons third law (action equal to reaction): F2mem F1mem F. (3) Thus, F1in F2in, and the radius of the sphere remains constant. Because pressure is the force per unit surface area (P F/S), the pressure inside the sphere will not change; the pressure gradient will be less than before (Pperitoneum > Pcyst); and it will oppose leakage at the exact moment of puncture. To repeat the point, if the sphere wall is rigid, external pressure will decrease the pressure difference across the wall. If the cyst membrane is elastic, if stress is proportional to strain (linear or Hookes law region), and if the fluid inside it is incompressible (ideal fluid), we can apply Laplaces law that relates the pressure differences across a closed elastic membrane to the tension in the membrane [5, 7, 8]. Laplaces law for a spherical membrane dictates that Pin Pout 2/r, (4)

Fig. 1. The sphere as an idealization of a liver hydatid cyst and the net forces acting on its membrane. For further explanation see text.

expand individually on these points. As far as we know, the laparoscopic approach to treat hydatid cysts has never been thoroughly investigated before. The physical basis of the laparoscopic approach to parasitic liver cysts We use a basic heuristic model of an elastic sphere to demonstrate that the slight increase in intraperitoneal pressure during laparoscopy (15 mmHg) does not increase the risk of rupture or spillage of the high-pressure cyst during manipulations (puncture, evacuation, etc.). We propose the sphere as an idealization of a human liver hydatid cyst. To prevent incidental spillage, safety demands that the pressure gradient across the cyst membrane (Pcyst Pperitoneum) does not increase or even decrease (Pcyst < Pperitoneum) as we increase the intraperitoneal pressure to 15 mmHg. We have not measured the elastic properties of the cystic wall that protrude outside the liver. This wall is composed of a germinal layer and the pericyst, that might vary in thickness, stiffness, and tissue components. We therefore discuss different parameters that affect the model. Membrane characteristics Assuming that the cyst wall may have arbitrary elastic properties from rigid to high elasticity, we first consider the wall of the sphere to be stiff and incompressible. The net forces acting on the membrane perpendicular to its surface in equilibrium state are zero: Fin (Fmem + Fout) 0 (Fig. 1). (1)

where Pin and Pout are the pressures inside and outside the sphere, respectively; is the wall tension; and r is the radius of the sphere. Because we are dealing with an equilibrium state of the forces acting on the membrane (Fi 0), we can easily substitute P F/S in Laplaces equation (Eq. 4) and relate it to Eq. 1. Thus Fmem Fin Fout S 2/r, (5) where S is the surface area, and the forces due to the pressures are perpendicular to the surface at each point. During pneumoperitoneum, Fout increases, but the volume of the sphere, and hence its radius r, does not change (ideal fluid), and thus Fmem does not change either. The main point here is that equilibrium (Eq. 5) is maintained. Fin will increase by the same amount as Fout, and hence the pressure increases proportionately. Thus, the pressure gradient across the wall will remain the same. For example, if we increase the intraperitoneal pressure by 15 mmHg, the new pressure inside the cyst will increase by 15 mmHg also. Practically, an ordinary cyst wall has physical properties intermediate to those just described. Thus the pressure gradient across the wall probably decreases slightly (cystic P < outside P). The preceding description of the spheres membrane does not take into account the thickness of the cyst wall. If the pericyst has an elastic and compressible wall that can change its thickness, the model can be described as composed of many tiny springs perpendicular to its surface and acting with force F K X, (6)

Fin and Fout are derived from the pressure inside and outside the cyst. Fmem, the force derived from the surface membrane tension, is perpendicular and directed toward the cyst inside. The tension is the force per unit length. Increasing the intraperitoneal pressure (pneumoperitoneum) results in increased Fout that acts on the membrane. F2out F1out + F, (2)

where indexes 1 and 2 correspond to the situation before and after creation of pneumoperitoneum. F is the force created by the pneumoperitoneum. The stiff membrane will

where K is the spring constant, which may be variable, reflecting the different elasticities of individual cyst walls, and X is its length displacement that reflects the changes in cyst membrane thickness. An increase in the intraperitoneal pressure will be partially transformed into potential energy to compress those tiny springs that represent the sphere wall. Potential energy 1/2Km (Xm)2 Pout Y 4 r2 Xm P V, (7)

1075

where m stands for membrane, V is the change in volume of the cyst wall, and Y is the fraction of the cyst wall exposed to the peritoneum. Xm and Km are defined in Eq. 6 and attributed to the membrane. The external force applied on the compressible cyst wall will cause plastic deformity shared by the neighboring structures (i.e., the liver), so only part of it will be applied on the cyst content. Thus the cystic pressure increase (Pcyst) should be somewhat smaller than the peritoneal pressure increase (Pout), independent of fluid characteristics. In addition, the instantaneous external force (affecting the cyst contents) as pneumoperitoneum is applied will be reduced because there is a cushioning effect due to the compressible wall.

Fm(1)hepatic Fm(1)peritoneal Fm(1)peritoneal > Fm(2)peritoneal where Fm(1)hepatic and Fm(2)hepatic are the tension forces being applied by the cyst membrane on the hepatic side of the cyst, and Fm(1)peritoneal and Fm(2)peritoneal are the tension forces being applied by the cyst membrane on the peritoneal side of the cyst, before and after pneumoperitoneum, respectively. The force applied by the inside cyst medium on its wall, F(2)cyst, is in equilibrium with the sum of the external forces being applied on the cyst and by the cystic wall tension force on both sides of the cyst: F(2)cyst Fm(2)hepatic + F(2)hepatic F(2)cyst Fm(2)peritoneal + (F(2)peritoneal + F) (9) (10)

Cyst fluid characteristics When the cyst contains fluid that is noncompressible, the only parameter affecting the intracystic pressure is the cystic wall, which was already discussed earlier. However, when the cyst contains fluid that is slightly compressible (because it may contain small bubbles of gas, etc.), applying additional external force will reduce the volume and thus the radius of the cyst. When the radius of the cyst decreases, the elastic force Fmem acting toward the interior of the cyst, which can be obtained from Eq. 5, slightly reduces because its magnitude is directly proportional to the radius of the sphere (S is proportional to r2). Consequently, the increase in Fin will be slightly less than the increase in Fout, and thus the pressure gradient will decrease.

As the radius of the peritoneal part of the cyst becomes smaller, Fm(2)peritoneal is less than Fm(1)peritoneal (see Eqs. 4 and 5 regarding Laplaces law), and thus the increase in F(2)cyst is smaller than the increase in F(2)peritoneal (F). Consequentially, the pressure gradient that depends on the difference between F(2)peritoneal and F(2)cyst decreases. In conclusion, our model shows that creation of pneumoperitoneum does not increase the rupture risk of the hightension cyst. Pressure gradient is maintained as constant, or even as less than before pneumoperitoneum is applied. Practically, although the pressure gradient is still high at the moment of puncture (due to the high intracystic pressure), pneumoperitoneum does not carry increased risk of rupture. Surgical manipulations might become safer laparoscopically than with the usual open procedure when the proper technique is used. The experimental animal model

Extending the model In the preceding sections we described a model that depends on the membrane and the fluid characteristics of the cyst. Does such a schematic model really represent the situation in vivo? Is it possible to have a reduced pressure gradient due to pneumoperitoneum, even when considering an ideal fluid and elastic membrane (Pcystic < Pperitoneal)? The cyst is partially embedded in the liver, which is anatomically included within the peritoneal cavity, and apparently is exposed to the same pressures as the cyst. Actually, the situation in vivo is different. The liver drains through the hepatic veins to the inferior vena cava outside the peritoneal cavity and its high pressure environment. Additionally, pneumoperitoneum causes reduction in portal and hepatic blood flow [6, 9]. Consequently, during pneumoperitoneum, the liver becomes less congested and more compressible. Thus, the wall of the cyst, which is embedded within the liver, might be exposed to less external (hepatic) pressure and resistance. Before the pneumoperitoneum is created, the membrane tension is equal on both sides of the cyst wall, inside the liver, and outside toward the peritoneal cavity. When pneumoperitoneum is applied, the sphere is being pushed toward the liver, so the membrane facing the peritoneal cavity is less tense than the intrahepatic membrane, and a new steady state is created: Fm(2)hepatic > Fm(1)hepatic (8) In this section we suggest an experimental model for qualitatively evaluating (under atmospheric and slight hyperbaric pressure of 15 mmHg) diverse techniques for evacuating parasitic cysts without spillage, which is a requirement. Materials, methods and techniques A 25 25 40-cm sealed box with glass walls similar to an aquarium was built. It has a removable cover made of an elastic rubber membrane, which enables inflation of the box, introduction of laparoscopic cannulae, and surgical manipulations under high pressure atmosphere. All the goats in the Yarka village slaughterhouse (located in the center of an endemic echinococcal area in Western Galilee) underwent veterinarian examination immediately after slaughter. When hepatic cystic lesions were noticed, the liver was harvested, preserved in a refrigerator, and sent immediately (within 30 to 40 min) to our laboratory, where it was washed and put in a plastic bag containing normal saline, then placed inside the sealed box. We evaluated several techniques for aspiration and evacuation of the cysts: 1. Simple aspiration of the cyst by a fine laparoscopic needle. 2. Aspiration (with the same needle) of the cyst through a

1076

12-mm transparent cannula connected via a two-way stopcock to a suction apparatus. This creates a vacuum inside the cannula that enables the tip to grip the liver and cyst surface. With such a technique, any spillage around the needle is aspirated by the large-bore cannule. 3. After introduction and positioning of the large cannule, n-butyl-2-cyanoacrylate (Hystoacryl blue, B. Braun, Melsungen AG, Germany) or fibrin glue (Tisseel, Immune US, Inc.) was introduced through a thin silicone tube around the cannule tip to further seal its attachment to the liver, especially in cases of accidental suction failure. We searched for any sign of spillage associated with either technique. Detection of spillage After aspiration of the cyst but before detachment of the equipment, the liver was gently flooded with saline solution. The fluid in the plastic bag was collected and left overnight in a conic container to get a concentrated sediment, which was further concentrated and sent for parasitologic analysis. Any evidence of scolices was considered a failure of technique because this indicated spillage. The aspirated fluid was sent for analysis to validate the existence of scolices in these cysts. In parallel, spillage was detected by evacuating and filling the cysts with 20% hypertonic saline, transhepatically, through a long fine needle, and any change in electrolyte concentration was detected after manipulation in the fluid around the liver. We did not validate complete removal of the cyst contents in this model because it was not the aim of our work. In vivo, the most important and dangerous step is the initial puncture and aspiration of the cystic fluid. After the introduction of scolicidal solution and lowering of the intracystic pressure, the next surgical steps become relatively safe. Results The techniques were evaluated for the presence of absence of cystic leakage. Five echinococcal cysts were tested for spillage after aspiration through transparent cannule under vacuum. In five cysts we used cyanoacrylate with suction, and in two cysts cyanoacrylate alone. In two cysts we used fibrin glue together with suction. There was no evidence of spillage in any of these trials. Six cysts were tested for spillage after aspiration through needle alone, and scolices were detected in the saline sediment in three. Statistical analysis, performed by using the Fisher test, compared our technique (14 cysts, 0 spillage) to simple aspiration (6 cysts, 3 spillage). The difference was found to be significant (t 0.0175). Implementation of the technique for human patients A transparent cannula with a beveled tip was used because it has some advantages: (1) It allows the surgeon to inspect the process of evacuation inside the cannula. (2) Cannulae tips are beveled or cut straight. Beveled tips have an oval (elliptical) cross section and are usually cut at a 45-degree

angle (). The area of a straight tip A is r2, where r the radius. In the case of a 12-mm diameter tip, r 6 mm and A 113 mm2. For a beveled tip 12 mm in diameter, A r2/cos , and if 45 degrees, the area is 161 mm2. Because F P A, the beveled tip has an enlarged suction grip with respect to straight-cut round tips. (3) The beveled tip enables good surface grip to cysts located at a distance from the site of cannula introduction [3]. We occlude the holes at the tip of the cannula to enable creation of a force due to vacuum. Recently, we have developed a new larger cannula made of two transparent cannula 18 mm in diameter (Ethicon) connected firmly together in series to create one long tube, using a special adhesive tape and glue. Our new cannula has a 45-degree beveled tip and a special two-way stopcock. Besides its ability to reach cysts located under the diaphragm, the large bore enables us to insert various tools and do safe surgical manipulations inside the cannula, especially when simple aspiration is impossible due to numerous daughter cysts. Thus since 1992 we have been able safely to evacuate more than 30 giant hydatid cysts of the liver in human patients after informed consent. The third technique (using cyanoacrylate) was used once, and the glue was excised and removed after evacuation. The pericysts were managed according the preference of the surgeon. During follow-up of 5 years to 6 months, no recurrences were recorded.

Discussion Recently, a number of reports were published concerning the laparoscopic approach to operating on hydatid cysts of liver [1, 4, 12, 13, 15, 16, 19, 20]. In most of the operations, the technique employed to safeguard against spillage of cyst contents was not described. In one report, spillage was discussed. In this case, it was managed by instillation of scolicidal solution (cefrimide) into the peritoneal cavity [1, 13, 15, 16, 20]. In another study, lowering the peritoneal pressure was advised to prevent excessive cystic pressure elevation and leakage during manipulations [20]. In this study, we have demonstrated that, theoretically, creation of pneumoperitoneum does not increase the risk of spillage and may even enhance safety. The experimental animal model has established the rationality and safety of the techniques we used, which were successfully implemented in vivo. In practice, even in cases when the suction grip is inactivated temporarily, the pneumoperitoneum creates a pressure gradient that causes the cyst contents to move through the cannula toward the outside of the peritoneum, and not inward. Butyl cyanoacrylate and fibrin glue were employed for better sealing and adherence of the contact area between the tip of the cannula and the cyst surface. In practice, the introduction of the hystoacryl glue should be through silicone tubes, which will prevent its polymerization before it reaches the cannula tip. The internal use of those materials is well known and safe [2, 10, 14]. Because the glue is removed with the pericyst (after evacuation), there is no reason to avoid its use when necessary. In our opinion, simple needle aspiration is neither safe nor justified, even when the patient is being treated in addition by scolicidal medication (albendazole). Recently, it

1077

was reported that hydatid cysts of liver were successfully treated percutaneously under ultrasonographic (US) and computed tomographic (CT) guidance (needle aspiration and injection of scolicidal solution) [11, 17, 18]. According to our laboratory findings and clinical experience, in many cases cyst contents cannot be aspirated. It is not a simple fluid and may contain numerous daughter cysts. The risk of spillage during needle aspiration is not negligible. We therefore suggest that the laparoscopic approach to managing hydatid cysts as described in our study is efficacious, rational, and safe, and has some advantages: decreased leakage, better accessibility, and the capability of inspecting the cyst cavity, all in addition to the well-known minimalization of invasiveness associated with the laparoscopic approach. The experimental animal model in this study can be used as a reliable model to test other new techniques for managing hydatid cysts. References
1. Alper A, Emre A, Acarli K, Bilge O, Ozden I, Ariogul O (1996) Laparoscopic treatment of hepatic hydatid disease. J Laparoendosc Surg 6: 2933 2. Al Sammarani AY, Jessen K, Haque K (1987) Endoscopic obliteration of a recurrent tracheoesophageal fistule. J Pediat Surg 22: 993 3. Bickel A, Eitan A (1995) The use of a large transparent cannule with a beveled tip for safe laparoscopic management of hydatid cysts of liver. Surg Endosc 9: 13041305 4. Bickel A, Loberant N, Shtamler B (1994) Laparoscopic treatment of hydatid cysts of the liver: initial experience with a small series of patients. J Laparoendosc Surg 4: 127133 5. Cohesive forces in liquids (1984) In: Kane JW, Sternheim MM (eds) Physics. Wiley, New York, pp 361368

6. Diebel LN, Wilson RF, Dulchavsky SA, Saxe J (1992) Effect of increased intra-abdominal pressure on hepatic arterial, portal venous, and hepatic microcirculatory blood flow. J Trauma 33: 279283 7. Elasticity (1972) In: Sears FW, Zemmansky MW (eds) College physics. Addison-Wesley, Cambridge, MA, pp 229240 8. Elastic properties of materials (1984) In: Kane JW, Sternheim MM (eds) Physics. Wiley, New York, pp 200212 9. Eleftheriadis E, Kotzampassi K, Botsios D, Tzartinoglon E, Farmakis H, Dadonkis J (1996) Splachnic ischemia during laparoscopic cholecystectomy. Surg Endosc 10: 324326 10. Farouk R, Drew PJ, Qureshi A, Roberts AC, Duthie GS, Monson JRT (1996) Preliminary experience with butyl-2-cyanoacrylate adhesive in tension-free inguinal hernia repair. Br J Surg 83: 1100 11. Giorgio A, Tarantino L, Francica G, Mariniello N, Aloisio T, Soscia E, Pierri G (1992) Unilocular hydatid liver cysts: treatment with USguided, percutaneous aspiration and alcohol injection. Radiology 184: 705710 12. Katkhouda N, Fabiani P, Benizri E, Mouiel J (1992) Laser resection of liver hydatid cyst under videolaparoscopy. Br J Surg 79: 560561 13. Khoury G, Jabbour-Khoury S, Bikhazi K (1996) Results of laparoscopic treatment of hydatid cysts of liver. Surg Endosc 10: 5759 14. Little JM, Laver C, Hogg J (1977) Pancreatic duct obstruction with an acrylate glue: a new method for producing pancreatic exocrine atrophy. Surgery 81: 243249 15. Massoud WZ (1996) Laparoscopic excision of a single hepatic hydatid cyst. Int Surg 81: 913 16. Mompean AL, Paricio PP, Campos RR, Ayllon JL (1993) Laparoscopic treatment of a liver hydatid cyst. Br J Surg 80: 907908 17. Mueller PR, Dawson SL, Ferrucci JT, Nordi GL (1985) Hepatic echinococcal cyst: successful percutaneous drainage. Radiology 155: 627 628 18. Saremi F, McNamara TO (1995) Hydatid cysts of the liver: long-term results of percutaneous treatment using a cutting instrument. AJR Am J Roentgenol 165: 11631167 19. Sover M, Scapin S (1995) Laparoscopic pericystectomy of liver hydatid cyst. Surg Endosc 9: 11251126 20. Yucel O, Talu M, Unalmiser S, Ozadede S, Gurcan A (1996) Videolaparoscopic treatment of liver hydatid cysts with partial cystectomy and omentoplasty. Surg Endosc 10: 434436

Original articles
Surg Endosc (1998) 12: 10131016 Springer-Verlag New York Inc. 1998

The effect of contrast variation on task performance in video evaluation

J. Danis
Surgical Department, General Public Hospital, A-2410 Hainburg an der Donau, Hofmeister Strasse 70, Austria Received: 24 January 1997/Accepted: 18 September 1997

Abstract Background: Luminosity and brightness of the organs viewed on the screen are important for orientation during teleoperating. They can impair the precision of our orientation ability. Methods: The way in which organs scatter light during laparoscopic operation was detected and analyzed by means of reflection spectrophotometry with the Photoshop program package (ADOBE). The arithmetical skill was measured during different illumination conditions of the screen created randomly, using values of the liver, gallbladder, and omental fat. Results: Separate human tissues and organs could be divided on the basis of their luminosity into three basic groups: the luminous tissues such as fat, the stomach, and the bowel; medium-luminous organs such as diaphragm and gallbladder; and dark, mostly parenchymatous organs such as the liver and the spleen. The most luminous tissue illuminated the screen intensely. In experiments the hyperillumination impaired arithmetical skills of 15 volunteers. The errors appeared constantly and the time needed to count the generated points was significantly longer when 40% of intensively illuminated view was exceeded. Conclusions: A situation may occur wherein the imaging system lacks the capacity to transmit needed optical information. Intense brightness of the screen may suppress surgical skills during teleoperating. Critical limit was achieved in our experiment with 40% of the view illuminated by omentum fat. This limit can be important for robotic surgical systems such as automated robotic camera control. Key words: Organ luminosity Hyperillumination Surgical robotic systems Camera control Teleoperating Operative endoscopy

perception, the topographical (or pathological) area must be determined from the smallest change in gray and/or color balance [8, 11]. Balance, or contrast, of light and colors [15] is also important. Only the contrast of features permits 3D perception and paralaxis. Contrast changes, such as the brightness and the size of different areas projected onto the screen, appear continuously during endoscopic surgery. The aim of this paper is to demonstrate how the various contrasts affect precision during teleoperating.

Materials and methods

The luminosity of human tissues and organs was investigated by means of digital analysis. Representative pictures of the upper abdominal viscera were recorded during laparoscopic cholecystectomy in ten patients. The reflectivity of silver and chrome mirrors and of the white background created by a 1-cm layer of white paper (Kopier- und Laserpapier TCF SB KOP A480 80 GR/M2) was measured to establish reference values. The image processing system used during laparoscopic surgery was created using a 10-mm 0 Surgiview laparoscope (USSC) connected to a Endo Video V/MV 9665 (CIRCON) camera and to a universal control module Docutrol/MV 9753 (CIRCON) as well as to a SONY Trinitron color video monitor PVM 1442 QM and to a JVC HR S 5500 EG videorecorder. As a light source the WISAP Endo-illumination 250K/7720K was used (Fig. 1). The laparoscope was marked so that the objective could be kept constantly at a distance of 10 cm above the surface of the organ investigated. The light emitted by the laparoscope was constantly adjusted to the second grade of intensity. A second transmission line was connected to a PC computer (Pentium 166 MHz), enabling digital processing. The digitized signal of the pictures was recorded and analyzed by means of the Photoshop 3.05 (ADOBE) program package, which is commercially available. Representative pictures were converted to pixels and analyzed by means of a 255-grade scale for gray, red (R), green (G), and blue (B). The mean value as well as the standard deviation were calculated from the data acquired. During the second part of our experiment the effect the light conditions transmitted onto the screen had on our ability to orient was investigated. A model situation during human laparoscopic cholecystectomy was created by means of the PC computer (Pentium 100) with a 14-inch color screen. A background for the pictures was generated with the same light quality (ADOBE) as detected from the liver (Table 2); the same was done for the values of the omentum. The liver values background represented the dark portion of the view and the omentum values the bright one. Both parts of the background changed randomly at 10% intervals between a low of 10% and a high of 90%. Ten to 15 points with a diameter of 0.5 cm and

Spatial orientation on the video screen during endoscopic surgery is based on a paralactic comparison of the different landmarks of the area of interest. With two-dimensional

1014

Fig. 1. Arrangement of laparoscopic device in experiment.

the brightness of the values of the gallbladder (ADOBE) were randomly and synchronously generated onto the dark field. Fifteen volunteers, sitting in a dark room at a distanct of 50 cm from the screen, were asked to count the generated points as rapidly as possible (Fig. 1).

Results In Table 1 the reflection of the silver mirror was 240.89 1.08 out of a maximum value of 255 grades on the scale. The chrome mirror had 96.98% of the reflection of the silver mirror (%Ag). Lower were the values of the reflection from the white paper on the gray scale as well as on the RGB (red-green-blue) spectrum. These values were close to 85% of the Ag and to 88% of the Cr mirror reflection. In Table 2, the omental fat was the most luminous structure, with a yellow spectral shift that appeared on the screen during the viewing of the human laparoscopic cholecystectomy. The large bowel was, with 154.94 23.39 grades, the next most luminous organ. A significant shift to the red spectrum could be here observed. The stomach achieved a mean gray brightness of 142.36 17.49 out of a total 255 grades possible. Lower luminosity with 133.29 9.57 gray grades was detected on the small bowel, parietal peritoneum (126.59 13.97 gray grades on the scale), the diaphragm (119.28 31.38), and the gallbladder (100.20 29.55 gray grades). The spleen (51.11 14.37) and liver (50.19 7.97 gray grades) were observed to have the greatest contrasts as compared with the omental fat. A significant shift to the red spectrum with the reduction of the green and blue reflection could be also observed (spleen with the 98.10 and liver with the 90.78 red grades vs 30.52 and 26.90 green and 32.57 ad 33.92 blue grades, respectively) (Table 2). With 10% of the area of the screen illuminated by omentum it took a mean time of 0.28 0.05 s (n SD) to count one (from 10 up to 15) randomly generated point (Fig. 2). With 20% of the omentum area illuminated it was 0.31 0.04 s. This value was not significant compared with the first measured time value. With 30% of the area illuminated by omentum the time needed for counting one generated

point was 0.32 0.06 s. This time was not significantly different compared with the first value. At 40% of the illuminated area 15 volunteers needed 0.34 0.04 s for one point. This value was significantly different in comparison with the time for 10% of the omentum-illuminated area (p < 0.05). At 50% of the bright area, a value of 0.36 0.06 was also different (p < 0.01). At 60%, the value was 0.36 0.04 (p < 0.001 compared with 10% of the illuminated surface). Seventy percent of the illuminated area further prolonged the time needed for one generated point. The main value was significantly different at 0.39 0.08 s (p1070 < 0.01). With 80% and 90% of the area illuminated by omentum, the addition of one generated point took an average 0.40 0.07 and 0.40 0.06 s, respectively. Both values were significantly different compared with the first three values (p10 80,90 < 0.001). Assuming that 15 persons made 24 errors in counting the generated points, the error frequency was 17.78% for all the runs. The volunteers twice counted more than the actual number of generated points; 22 times they counted fewer points than the actual generated number. The number of errors was a constant of four for the areas illuminated 40% and more. One error appeared from 10% up to 30% of intensively illuminated area (Fig. 2).

Discussion Three basic principles influence orientation in threedimensional space, i.e., the three-dimensional location of the subject, the paralaxis principle, and the perception of features based on their luminosity. Predatory animals and carnivores have developed the most perfect mode of subject location. The perception of only light and darkness is characteristic of protozoa and other primitive forms of life as well as some sea creatures and animals living without light. Whenever the closer subjects scatter more light, the intensively illuminated subjects are then perceived to be closer. Similarly, the larger the subject is, as a rule it is more

1015 Table 1. Mirrors and white paper reference valuesa Absolute value 255 Ag (n SD) Cr (n SD) % Ag WP (n SD) % Ag % Cr
a

Gray 240.89 1.08 233.61 0.81 96.98 205.39 4.64 85.26 87.92

Red 241.98 0.87 234.94 0.34 97.10 207.95 5.01 85.94 88.50

Green 240.93 1.17 233.49 1.03 96.91 204.39 4.68 84.83 87.54

Blue 240.37 1.85 234.34 1.2 97.49 204.1 5.42 84.91 87.10

Pixels 3,596 5,698 22,348

WP, white paper.

Table 2. Human upper abdominal organs visible during laparoscopic cholecystectomy Absolute value 255 Fatomentum Large bowel Stomach Small bowel Peritoneum Diaphragm Gallbladder Spleen Liver Gray 158.12 14.67 154.94 23.39 142.36 17.49 133.29 9.57 126.59 13.97 119.28 31.38 100.20 29.55 51.11 14.37 50.19 7.97 Red 199.98 11.33 222.63 23.95 211.02 10.45 217.22 11.65 206.09 16.51 173.86 31.38 158.15 33.56 98.10 26.85 90.78 15.96 Green 145.39 16.86 127.96 25.82 112.41 21.61 97.03 9.73 94.79 13.86 94.26 31.78 97.91 29.46 30.52 9.45 26.90 5.58 Blue 114.11 16.08 118.68 23.82 118.02 20.79 99.51 9.97 80.56 14.22 105.81 31.36 84.10 30.73 32.57 10.37 33.92 6.84 Pixels 212,594 16,128 60,125 9,996 46,640 16,622 24,096 13,970 101,916

Fig. 2. Work performance based on the luminosity model situation.

luminous, and it is also perceived as closer. This archaic art of seeing on an absolutely black background occurs also in developed animals and in humans. The loss of sensitivity to absolute light level may result largely from the sensitivity adjustment of the eye to the highest light flow from the screen. A similar situation appears during laparoscopic surgery, when a part of the screen is illuminated by fat or another intensively shining structure. A dominant perception of the luminous field appears, and the skill in the dark part of the screen is then suppressed. This effect can be powered when the illuminated subject is close to the objective of the laparoscope. A situation may also occur where

the imaging system lacks the available capacity to transmit more needed optical information. The technical lights of the detector system are 4,000 4,000 pixels, which may be captured by the camera. However, most endoscopic cameras do not have such capacity. Technical transmission limits of the screen are fewer compared with those of a camera, i.e., 1,000 1,000 pixels [5]. The maximum luminosity of the screens used in medicine achieve 600 candela/m2 (cd/m2) and carry 1012 bits of information. However, they individually have a distinction of 2.55 linepairs/mm (Lp/mm) [1, 4, 7, 9, 13]. The endoscope used in our experiment illuminated a circle with a diameter of 10 cm. A diameter of 2 cm was illuminated with a homogeneous light strain, as could be detected by means of the RGB analysis. The signal transmission effectivity was 94.47% for the gray scale and 94.89%, 94.48%, and 94.26% for the red, green, and blue colors, respectively, as could be determined by silver mirror reflection (Table 1). After the data of upper abdominal organs were acquired, the virtual picture was generated with the background based on values of the human liver as well as the human omentum. Onto this picture the optimum amount of 1015 points was then generated with the luminosity of the gallbladder. Our preliminary investigation showed that more points were not suitable for such an experiment. Learning, picture sector perception, as well as physiological eye accommodation can affect the results [3, 10]. Also, the time integration of the pictures during the interactive dialogue on the screen improved the volunteers perception during other observations [2, 6, 14]. Similar picture integration appears during the laparoscopic operation by means of the continual changing of the camera view. A significantly longer period was needed to count the generated points when a more than 40% hyperillumi-

1016

nated area with omental values of luminosity was projected on the screen. Over that 40% of hyperilluminated area the errors appeared constantly during the count (Fig. 2). The 40% of hyperilluminated area in our experiment can be thus appreciated as the critical limit of hyperillumination. This could be determined by both criteria selected: The time needed to count the generated points increased significantly and the errors appeared constantly. However, our critical limit of 40% of hyperilluminated area remains to be confirmed by subsequent clinical observations for each kind of operation performed. Critical limits could be exceeded by any cameraman during the surgery. The hyperilluminated area on the screen can be reduced by means of the zooming method through the narrowing of the view angle or by changing the distance as well as the angle between the optical lens and area of interest [2]. Moreover, the recently developed and commercially distributed automated endoscopic system for optimum camera positioning [12] could be adjusted with a program warning system which would change the camera view if close-to-critical hyperillumination limits should be achieved.
Acknowledgments. I am indebted to N. Demuth, R. Stary, and M. Reiter for recording experimental measurements, to J. Danis, Jr., for computer program writing for the model situation, and to M. Visnovsky and M. Danis for computing the statistics and creating the graphics.

2. Danis J (1996) Theoretical basis for camera control in teleoperating. Surg Endosc 10: 804 3. Fraser RG, Pare JAP, Pare PD, Fraser RD, Genereux GP (eds) (1988) Perception in chest roentgenology. In: Diagnosis of diseases of the chest. 2nd ed. WB Saunders, Philadelphia, p 291 4. Greene RE, Oestmann JW (eds) (1992) Computed digital radiology in clinical practice. Thieme, New York 5. Kayser K (1996) Telemedizin. Wien Klin Wochenschr 108: 532 6. Kayser K, Fritz P, Drlicek M (1995) Aspects of telepathology in routine diagnostic work with specific emphasis on ISDN. Arch Anat Cytol Pathol 43: 216 7. Lee KR, Siegel EL, Templeton AW, Dwyler III SJ, Murphey MD, Wetzel LH (1991) State-of-the-art digital radiography. Radio-Graphics 11: 1013 8. Manneberg G, Witt H, Slezak P (1995) Quantitative endoscopic classification of esophagitis by means of computerized image processing. Part 1: theoretical background. Hepato-Gastroenterol 42: 135 9. Murphey MD (1994) Imaging aspects of new techniques in orthopedic surgery. Radiol Clin North Am 32: 201 10. Pa rtan G, Mosser H, Tekusch A, Mathiaschitz U, Augustin I, Hruby W (1994) Befundung von digitalen Intensiv- bzw. Bettlungenaufnahmen am Monitor vs. Hardcopy: eine klinische ROC-Studie. Fortschr Rontgenstr 161: 354 11. Rey JF, Albuisson M, Greff M, Monget JM (1988) Electronic video endoscopy. Preliminary results of imaging modification. Endoscopy 20: 8 12. Sackier JM, Wang Y (1994) Robotically assisted laparoscopic surgery. From concept to development. Surg Endosc 8: 63 13. Schaefer C, Prokom M (1993) Storage phosphor radiography of the chest. Radiology 186: 314

References
1. Buckwalder KA, Braunstein M (1992) Digital skeletal radiology. Am J Roentgenol 158: 1071

14. Seltzer SE, Judy PF, Adams DF (1995) Spiral CT of the chest: comparison of cine and film based viewing. Radiology 197: 73 15. Werner JS (1996) Aging through the eyes of Monet. AvH-Magazin 68: 3

Surg Endosc (1998) 12: 10461050

Springer-Verlag New York Inc. 1998

Combined laparoendoscopic colon resection and anastomosis using the no touch technique and fibrin glue
An experimental study
G. Ussia,1 S. Cuccomarino,1 B. Ravo,2 G. Galletti1
1 2

Department of Surgery, A. Valsalva, University of Bologna, Bologna, Italy American Hospital, Roma, Italy

Received: 15 September 1997/Accepted: 16 November 1997

Abstract Background: Laparoscopic colon anastomosis are technically demanding. A new technique for colon resection and anastomosis using a combined laparoendoscopic approach is presented. Methods: In 10 pigs, pneumoperitoneum was induced and 5 trocars were placed. A sigmoid segment was isolated; a vein stripper was inserted from the anus, and the head was secured with a tie; the segment was intussuscepted pulling the stripper out; 4 seromuscular sutures were placed at the anastomotic site and fibrin glue was spread all around; an electrical wire loop, introduced via a colonoscope, was used to resect the intussuscepted segment that was removed from the anus. Results: All animals but one survived until sacrifice at 30, 60, 90, and 120 days. Macroscopically, the anastomosis appeared well healed; microscopically, after 90 days, there was a complete restitutio ad integrum of the intestinal wall. Conclusions: This technique is feasible and quick; it could be used clinically in small tumors not removable endoscopically. Key words: Laparoscopy Colon anastomosis Colon resection Endoscopy Fibrin glue

The expansion of laparoscopic techniques in resective colon surgery has boosted discussions on methods of realizing visceral anastomosis and on postoperative complications. The major complication in colon resections is anastomotic dehiscence, which is correlated with increased morbidity and mortality, and is more frequent in low colorectal anastomoses, oncologic surgery, and anastomosis perCorrespondence to: G. Ussia, Strada Maggiore 77, 40125 Bologna, Italy

formed manually [3, 6, 8, 9, 13]. The main causes of anastomotic dehiscence are insufficient oxygen supply, elevated mechanical tension, suture types (mechanical, manual, or with biologic glues), and bacterial contamination of the anastomosis [7, 11, 17]. In particular, sutures may cause local changes in the blood supply, with possible transmural necrosis and diastasis of the anastomotic lips. Another important contributory factor is initial prevalence of collagen catabolism in the colonic submucosal layer on the new synthesis [4, 12], which is caused by increased activity of a proteolytic specific enzyme (collagenase) combined with the presence of bacteria or phlogosis [4, 20]. Several studies have shown that human fibrin glue favors anastomotic collagen synthesis by promoting the hemostatic process and reducing collagenase activity through a specific inhibitor, the aprotinin [16]. Moreover, fibrin glue induces local migration of macrophages that produce factors promoting angiogenesis and fibroblast proliferation [14], and it also confines the bacterial flora within the colonic lumen, thereby limiting local sepsis [1, 15]. Experimental studies on the usefulness of fibrin glue in colorectal anastomoses are controversial [5, 19]. We have previously reported on the feasibility of its use [10, 18]. Fibrin glue could be very useful in laparoscopic colon surgery, in which anastomosis is still technically demanding. In fact, anastomosis in this case is commonly performed outside the peritoneal cavity by means of a mini-laparotomy. This experimental study aims to evaluate the feasibility of using human fibrin glue for laparoscopic colon anastomosis.

Materials and methods

Ten female white pigs weighing 2530 kg, were used. Preoperative colon preparation the day before the operation consisted of administering a laxative infuse of cassia acutifolia. Two hours before surgery, antibiotic prophylaxis and two enemas of water and glycerine were also performed.

1047

Fig. 1. Schematic illustration of the surgical technique. A colonic segment was isolated for resection 1520 cm from the anus. A vein stripper was introduced from the anus and fixed at its proximal extremity. The colonic segment was intussuscepted by pulling out the vein stripper, and the anas-

tomotic lips were reciprocally oriented with three seromuscular stay sutures; 24 ml of human fibrin glue were spread all around the anastomosis. The colonic segment was resected using a wire loop introduced by means of a colonoscope.

Fig. 2. After the introduction of a vein stripper from the anus, the head of the vein stripper is tied at the proximal extremity of the resecting colon segment (A, B). By pulling out the vein stripper, the colon segment is intussuscepted (C). Then, after placing three seromuscular stay sutures, the fibrin glue is spread (D). A wire loop is introduced from the anus under colonoscopic vision and placed around the intussuscepted segment (E). Finally, the resected segment is extracted with the vein stripper (F). The surgical technique is schematically shown in Fig. 1. The animals were anesthetized with ketamine (0.6 mg/kg body weight) and with a mixture of Fluothane, N20, and 02 by orotracheal intubation. After pneumoperitoneum induction, five trocars were placed, and a 12- to 15-cm colonic segment was isolated 1520 cm from the anus. A vein stripper was introduced from the anus, and its head was fixed by a transfixed suture at the proximal extremity of the isolated colonic segment (Fig. 2A and B). The colonic segment was intussuscepted by pulling out the vein stripper (Fig. 2C). The lips of the intussusception were approximated by four seromucosal stay sutures. The anastomosis was spread with 24 ml of fibrin glue (Tisseel Immuno, Vienna), using a special endoscopic catheter inserted through a 5-mm trocar (Figs. 2D and 3). Five minutes later, the intussuscepted colonic segment was electroresected by a wire loop introduced with a colonoscope from the anus (Fig. 2E) and extracted with the

1048

Fig. 3. Fibrin glue is spread on the anastomosis using an endoscopic 5-mm catheter. Fig. 4. Extraction of the resected colon segment from the anus. Fig. 5. Colonoscopic view after resection of the colon segment. Fig. 6. Macroscopic aspect of the anastomosis on day 30. Fig. 7. Endoscopic control on day 60. The anastomosis is well healed.

vein stripper (Fig. 2F and 4). Finally, the anastomosis was checked both endoscopically (Fig. 5) and laparoscopically at 50 mmHg under water by air insufflation. All the animals survived the operation except for one, which died on day 5 due to intestinal occlusion. The others were sacrificed 30, 60, and 90 days, respectively, after surgery. After the animals were sacrificed, tissue specimens were taken and studied macroscopically and microscopically (Table 1).

Results At autopsy, the anastomotic area was adhesion-free and difficult to recognize from its appearance at day 30 (Fig. 6). In four animals, a moderate stenosis of anastomosis in the

absence of clinically relevant signs was observed both at autopsy and endoscopically (Fig. 7). In the animal that died on day 5, an anastomotic dehiscence was observed. Histologically, from day 30 the mucosal layer appeared well repaired, without edema, and with scarce inflammatory infiltration (Fig. 8). The muscular layer sent out myofilaments that appeared well consolidated by day 60 (Fig. 8B). On day 90, a complete restitutio ad integrum of all layers of the colonic wall was observed (Fig. 8C). In the histologic sections taken on day 30, the mucosal layer appeared to be consolidated in correspondence with the sutures, whereas the muscular layer was diastased with a consistent inflammatory infiltrate. In some sections, the

1049

Fig. 8. Microscopical aspects of the anastomosis. A On day 30, mucosal and submucosal layers are well repaired while the muscular layer is diastased. The inflammatory infiltration is scarce. B On day 60, the muscular layer is in the advanced phase of consolidation. C On day 90, there is a

complete restitutio ad integrum of the colonic wall. D Day 30 histologic section at left of a stay suture. The suture filaments are incorporated into a foreign body-like granuloma.

Table 1. Results Macroscopic aspect of the anastomosis Animal number RCMTC RCMTC RCMTC RCMTC RCMTC RCMTC RCMTC RCMTC RCMTC RCMTC 01 02* 03 04 09 00 08 07 05 06 Sacrifice day 7 5 30 30 60 60 90 90 90 90 Stenosis ++ ++ + + + + ++ Adhesions ++ + Dehiscence No Yes No No No No No Yes No No Mucose Repaired Not repaired Repaired Repaired Well repaired Well repaired Well repaired Well repaired Well repaired Well repaired Microscopic features Muscular Not repaired Not repaired Repairing In consolidation Repaired Repaired Well repaired Well repaired Restitutio ad integrum Restitutio ad integrum Phlogosis ++++ ++++ + + + +

* Dead for intestinal occlusion

sutures appeared to be incorporated into voluminous foreign body-like granulomas (Fig. 8D). Discussion Anastomotic dehiscence represents the major complication of resective colon surgery. Many studies aimed at determining better methods of anastomosis have not permitted definitive conclusions.

Fibrin glue is particularly useful for laparoscopic colon anastomoses, in which use of conventional suture techniques appears difficult. Our combined laparoendoscopic technique is quick and easy and does not contaminate the peritoneal cavity with the intraluminal colonic content. Our results demonstrate that this technique, which uses only four stay sutures to orient the anastomotic lips, produces good anastomotic strength, and favors anastomotic sealing by promoting collagen synthesis. Within 30 days, these anas-

1050

tomoses consolidated with a complete restitutio ad integrum of the colonic wall, whereas the use of conventional suture materials lead to an irregular postoperative course with a higher incidence of complications. The successful sealing seems to be related also to some characteristics of fibrin glue (10, 16, 18): Its hemostatic and adhesive action combines with its biostimulating properties, which supports fibroblast proliferation with its threedimensional network. Moreover, bacterial growth is much slower in fibrin glue than in blood clots [3]. Our experimental model might be advantageously used in segmental colon resections for small neoplastic lesions too distant from the anus for the use of staplers, or when endoscopic excision is dangerous or not possible due to the risk of colon perforation. References
1. Ahrendt GM, Gardner K, Barbul A (1994) Loss of colonic structural collagen impairs healing during intra-abdominal sepsis. Arch Surg 129(11): 11791183 2. Akyol AM, McGregor JR, Galloway DJ, George WD (1992) Early postoperative contrast radiology in the assessment of colorectal anastomotic integrity. Int J Colorectal Dis 7(3): 141143 3. Bo sch P (1982) Erfahrungen mit der Fibrin-Spongiosaplastik. In: Cotta H, Braun A (eds) Fibrinkleber in orthopa die und traumatologie. 4. Heidelberger Orthopa die-Symposium, Georg Thieme Verlag, pp 84 85 4. Cronin K, Jackson DS, Dunphy JE (1968) Changing bursting strength and collagen content of the healing colon. Surg Gynecol Obstet 126: 747753 5. Detweiler MB, Durastante V, Verbo A, Muttillo I, Piantelli M, Kobos JW, Antinori A, Granone P, Magistrelli P, Picciocchi A (1995) Sutureless anastomosis of the small intestine and the colon in pigs using an absorbable intraluminal stent and fibrin glue. J Invest Surg 8(2): 129140 6. Docherty JG, McGregor JR, Akyol AM, Murray GD, Galloway DJ (1995) Comparison of manually constructed and stapled anastomoses in colorectal surgery. West of Scotland and Highland Anastomosis Study Group. Ann Surg 221(2): 176184

7. Fielding LP, Steward-Brown S, Blesovsky L, Kearny G (1980) Anastomotic integrity after operations for large-bowel cancer: a multicentre study. Br Med J 281: 411414 8. Fingerhut A, Elhadad A, Hay JM, Lacaine F, Flamant Y (1994) Infraperitoneal colorectal anastomosis: hand-sewn versus circular staples. A controlled clinical trial. French Associations for Surgical Research. Surgery 116(3): 484490 9. Fingerhut A, Hay JM, Elhadad A, Lacaine F, Flamant Y (1995) Supraperitoneal colorectal anastomosis: hand-sewn versus circular staplesa controlled clinical trial. French Associations for Surgical Research. Surgery 118(3): 479485 10. Galletti G, Giardino R, Ussia G, Brulatti M, Sorrenti G, Galletti M (1986) Anastomosi del colon con colla di fibrina umana (Tissucol). Studio sperimentale nel maiale. Minerva Chir 15 41(21): 17591765 11. Hunt TK, Hawley PR, Hale J, Goodson W, Thakral KK (1980) Colon repair: the collagenous equilibrium. In: Hunt TK (ed) Wound healing and wound infection: theory and surgical practice. Appleton-CenturyCrofts, New York, pp 153159 12. Jiborn H, Ahonen J, Zederfeldt B (1980) Healing of experimental colonic anastomoses: III. Collagen metabolism in the colon after left colon resection. Am J Surg 139: 398405 13. Karanjia ND, Corder AP, Bearn P, Heald RJ (1994) Leakage from stapled low anastomosis after total mesorectal excision for carcinoma of the rectum. Br J Surg 81(8): 12241226 14. Leibovich SJ, Ross R (1976) A macrophage-dependent factor that stimulates the proliferation of fibroblasts in vitro. Am J Pathol 84: 501513 15. Petrelli NJ, Cohen HC, De Risi D, et al. (1982) The application of tissue adhesives in small-bowel anastomoses. J Surg Oncol 19:5961 16. Pohl J, Bruhn HD, Christophers E (1979) Thrombin and fibrin induced growth of fibroblasts: role in wound repair and thrombus organisation. Klin Wochenschr 57: 273277 17. Shandall A, Lowndes R, Young HL (1985) Colonic anastomotic healing and oxygen tension. Br J Surg 72: 606609 18. Ussia G, Ravo B, Farruggia F, Dal Monte N, Galletti G (1996) Laparoscopic colon resection and anastomosis with fibrin glue: experimental study on pigs. European Surgical Research: Clinical and Experimental Surgery. 31st Congress of the European Society for Surgical Research (ESSR), April 13, 1996, Southampton, England. 19. Van der Ham AC, Kort WJ, Weijma IM, Jeekel H (1993) Transient protection of incomplete colonic anastomoses with fibrin sealant: an experimental study in the rat. J Surg Res 55(3): 256260 20. Yamakawa T, Patin CS, Sobel S, Morgenstern L (1971) Healing of colonic anastomoses following resection for experimental diverticulitis. Arch Surg 103(1): 1720

Surg Endosc (1998) 12: 10431045

Springer-Verlag New York Inc. 1998

Postlaparoscopic small bowel obstruction

Rethinking its management
J. M. Velasco, V. L. Vallina, S. R. Bonomo, T. J. Hieken
Department of Surgery, Rush University and Rush North Shore Medical Center, 9600 Gross Point Road, Skokie, IL 60076, USA Received: 2 September 1997/Accepted: 2 February 1998

Abstract Background: Patients with early postoperative small bowel obstruction (SBO) are usually managed nonoperatively with nasogastric suction, intravenous fluids, and observation. The majority of early postoperative SBO resolve without an operation. Methods: We performed a retrospective review of patients who had been diagnosed with postlaparoscopic SBO at three Chicago area teaching hospitals. Results: The patients were initially managed nonoperatively for up to 7 days. However, all of them subsequently required an operation. In every case, the postlaparoscopic SBO was caused by the small bowel being incarcerated in a peritoneal defect created either by trocar placement or peritoneal incision for herniorrhaphy. Conclusion: In contradistinction to the approach used for early SBO after laparotomy, prompt operative intervention for postlaparoscopic SBO is recommended. Key words: Laparoscopy Postoperative small bowel obstruction

The initial management of patients diagnosed with early postoperative small bowel obstruction is usually nonoperative. These cases can be managed with nasogastric decompression, intravenous fluid administration, and observation. The rationale for this approach is that early postoperative small bowel obstruction is usually secondary to remodeling adhesions or inflammation [12, 16]. Because ischemic bowel is unlikely in these patients, an initial trial of bowel rest and decompression is employed routinely. If the patient fails to improve after 2448 h, reoperation is recommended. Some surgeons have advocated even longer periods of nonoperative management (14 days), stating that complete
Correspondence to: J. M. Velasco

resolution of the obstruction occurs in 73% of patients managed nonoperatively [4, 5, 12]. Laparoscopic procedures have become very common for the treatment of gallbladder disease, groin hernias, gastroesophageal reflux, and appendicitis [7]. Complication rates after operation by minimal access techniques have been comparable to those after traditional open techniques. In 1994, in a multicenter review of complications following laparoscopic cholecystectomy, Deziel reported an overall complication rate of <2% [6]. As minimal access techniques become more common for a growing variety of diseases, the number of procedure-specific complications can also be expected to increase. One such complication is small bowel obstruction secondary to bowel incarceration through defects in the peritoneal cavity created by trocar insertion. This problem has been described in isolated case reports, prompting the recommendation for routine closure of fascial defects >10 mm [2, 3, 6, 11, 14, 18, 19]. The laparoscopic repair of groin hernias is based on the placement of a mesh in the preperitoneal space via a transabdominal (TAPP) or extraabdominal preperitoneal route (TEPP) [11]. TAPP creates an additional peritoneal defect and consequently an additional site for potential intestinal herniation. Some case reports have specifically examined patients with early postlaparoscopic bowel obstruction, but few specific management guidelines have been proposed [13, 6, 11, 14, 15, 18, 19]. Our objective in this study was to formulate guidelines for the management of postlaparoscopic small bowel obstruction on the basis of our experience with this entity at three Chicago-area teaching hospitals: Rush-PresbyterianSt. Lukes Medical Center, a major tertiary care center; Illinois Masonic Medical Center, a 500bed inner city teaching hospital; and Rush North Shore Medical Center, a 300-bed community suburban teaching hospital. Materials and methods
We conducted a retrospective review of the charts of all patients with a provisional diagnosis of postoperative small bowel obstruction occuring

1044

Fig. 1. Small bowel loop herniated through peritoneal defect at trocar site. Fig. 2. Small bowel loops in peritoneal defect after TAPP herniorrhaphy. Note exposed mesh.

within 30 days of a laparoscopic procedure for the 6-year period ending May 1, 1996. Appropriate roentgenographic studies were routinely obtained, including serial abdominal films. Serum electrolytes and complete blood counts were checked on admission if the patient was readmitted, or at the time of diagnosis if the patient was already in the hospital. All patients were noted to exhibit nonspecific abdominal tenderness and pain, as well as distention. No patient was noted on admission or at time of diagnosis to have any focal tenderness at the trocar sites except normal incisional tenderness. Other physical findings varied (i.e., nausea, vomiting, obstipation, diarrhea, etc.). Patients with clinical and radiologic evidence of simple ileus were excluded. Chart review was conducted to ascertain patient demographic data, the initial operative procedure, the time interval between initial operation and small bowel obstruction, the clinical course during admission, and operative findings.

Results A total of five patients met the inclusion criteria. Their age ranged from 38 to 78 years. Three patients were status postTAPP herniorrhaphy; one had undergone a laparoscopic cholecystectomy and Nissen fundoplication; and the fifth patient had undergone a laparoscopic appendectomy. The five patients presented within 12 days of their laparoscopic procedure. All but one patient had roentgenographic confirmation of a small bowel obstruction. Patient 4 had a firm clinical diagnosis of small bowel obstruction, yet radiographs of the abdomen showed a nonspecific gas pattern. Patients 1, 2, and 5 had a small bowel obstruction secondary to herniation in the trocar site (Fig. 1). Patients 3 and 4 had a small bowel obstruction secondary to an internal hernia at the TAPP repair site (Fig. 2). All patients had undergone fascial closure of their trocar sites. At the time of reoperation, the fascia at the trocar site was intact in all cases. The factors prompting surgical intervention included fever, leukocytosis, and increasing and/or localizing abdominal tenderness. Patients were discharged within 3 to 6 days after reoperation with uneventful postoperative courses to date. These data are summarized in Table 1. Discussion Adhesive intestinal obstruction is the leading cause of early postoperative bowel obstruction [4, 5, 12]. Signs and symp-

toms of postoperative intestinal obstruction may be the result of paralytic ileus or mechanical obstruction. There is still some debate over the timing of operation in patients thought to have an adhesive small bowel obstruction, particularly following an operation. In general, the initial approach is nonoperative [12]. These patients may experience a short period of apparently normal intestinal function before the obstruction becomes evident. It is accepted that the majority of these patients will achieve complete resolution of their obstruction with nasogastric suction and that most of them respond within 2448 h [4, 5, 12]. However, strangulation is a risk in this setting [13]. One study suggested that an absence of fever, tachycardia, localized tenderness, and leukocytosis indicates a situation in which observation may be safe [16]. Our data suggest that early postlaparoscopic bowel obstruction may not be adequately managed by such an approach. In this series, every patient was initially treated nonoperatively along the conventional guidelines for postoperative small bowel obstruction, since resolution of the obstruction takes place in 73% of patients [4, 5, 12, 13, 17]. However, all of our postlaparoscopic patients required an operation to resolve their small bowel obstruction. Although adhesions were present in several patients, the cause of the obstruction in all cases was a Richter type of herniation through the peritoneal defect at the site of trocar insertion or through a defect in the peritoneum at the site of hernia repair [8, 11]. Nevertheless, physical findings suggestive of this condition, such as a mass or exceptional tenderness at the trocar site, were not noted in any of these patients. Adhesions did not appear to play a pivotal role in any of our cases even though they were present in both of the patients who presented with small bowel herniated through the preperitoneal flap after TAPP. Three of our patients were reoperated successfully by laparoscopic technique alone. Although these patients can be reexplored laparoscopically, trocar insertion should be done with an open technique. Once inspection of the abdominal cavity is performed, the operation may be carried out laparoscopically or may require celiotomy, depending on the findings and the surgeons experience. Because a Richter-type hernia is often difficult to detect

1045 Table 1. Clinical course of five patients with postlaparoscopic small bowel obstruction Interval from initial operation to presentation with SBO (days) 2 0 2 7 12 Interval from initial presentation with SBO to reoperation (days) 7 10 4 5 10b Time from reoperation for SBO until discharge (days) 3 3 6 4 5

Patient no. 1 2 3 4 5
a b

Initial procedure Transabdominal preperitoneal herniorrhaphy (TAPP) Laparoscopic cholecystectomy and Nissen fundoplication TAPP TAPP Laparoscopic appendectomy

Site of small bowel obstruction (SBO) 10-mm trocar site 11-mm trocar site Site of TAPP and trocar site Site of TAPP and trocar site 12-mm trocar site

Sex M F M M F

Age (yr) 45 78 68 38 70

Type of exploration for SBO Celiotomy Laparoscopy Laparoscopy Laparoscopya Celiotomy

Inadvertent enterotomy during laparoscopy; repaired laparoscopically. Patient admitted with SBO, treated for 7 days with conservative management, and discharged. Admitted 2 days later with SBO and reoperated in 24 h.

on physical examination, computed tomography scan (CT) and ultrasonographic examination of the abdomen have been proposed as aids to making this diagnosis [9, 17]. However, their use has not been universally accepted in clinical practice. The diagnostic value of barium studies in these circumstances is also controversial [12]. One of the patients in our series underwent upper and lower gastrointestinal studies with inconclusive results. Thus, our experience suggests that the diagnosis of postlaparoscopic SBO is largely clinical and that early operative treatment (within 48 h) is the best course. In an effort to prevent postlaparoscopic SBO, routine closure of trocar sites 10 mm is advisable [9]. Closure should be done under direct vision, and it should incorporate all layers of the abdominal wall in order to eliminate peritoneal defects. Our experience with the two patients who required abdominal exploration after TAPP herniorrhaphy has led us to use a continuous running suture and/or a stapled overlapping peritoneal flap technique to seal the peritoneum in these cases. The extraperitoneal approach to laparoscopic herniorrhaphy may avoid such complications, provided that peritoneal rents are searched for meticulously and repaired appropriately as required [7, 10]. Our data suggest that postlaparoscopic SBO is a distinct subcategory of SBO. For these patients, extended nonoperative management may be harmful. We believe that a high degree of suspicion for incarcerated hernia and a low threshold for reoperation should be exercised when a patient presents with symptoms and radiologic evidence of a small bowel obstruction within 2 weeks of laparoscopy. Reoperation on these patients should be considered if their symptoms do not resolve within 48 h. Certainly, we must avoid rushing patients with abdominal pain and vomiting after laparoscopy to the operating room before excluding other causes of symptoms, such as prolonged ileus or bile leakage. By the same token, we recommend that surgeons not persist with extended nonoperative management of postlaparoscopic SBO patients. References
1. Arregui ME, Navarrete J, Davis CJ, Castro D, Nagan RF (1993) Laparoscopic inguinal herniorrhaphy: techniques and controversies. Surg Clin North Am 73: 513527

2. Azurin DJ, Schuricht AL, Stoldt HS, Kirkland ML, Paskin DL, Bar AH (1995) Small bowel obstruction following endoscopic extraperitoneal-preperitoneal herniorrhaphy. J Laparoendosc Surg 5: 263266 3. Ballem RV, Kenny R, Guiliano M (1993) Small bowel obstruction following laser laparoscopic cholecystectomy: a case study. J Laparoendosc Surg 3: 313314 4. Bizer LS, Liebling RW, Delany HM, Lickman ML (1981) Small bowel obstruction: the role of nonoperative treatment in simple intestinal obstruction and predictive criteria for strangulation obstruction. Surgery 89: 407413 5. Brolin RE (1984) Partial small bowel obstruction. Surgery 95: 145 149 6. Deziel DJ (1994) Complications of cholecystectomy: incidence, clinical manifestations, and diagnosis. Surg Clin North Am 74: 809823 7. Fitzgibbons RJ Jr, Camps J, Cornet DA, Nguyen NX, Litke BS, Annibali R, Salerno GM (1995) Laparoscopic inguinal herniorrhaphy: results of a multicenter trial. Ann Surg 221: 313 8. Hass BE, Schrager RE (1993) Small bowel obstruction due to Richters hernia after laparoscopic procedures. J Laparoendosc Surg 3: 421423 9. Maio A, Ruchman RB (1991) CT diagnosis of postlaparoscopic hernia. J Comput Assist Tomogr 15: 10541055 10. McKernan JB, Laws HL (1993) Laparoscopic repair of inguinal hernias using a totally extra-peritoneal prosthetic approach. Surg Endosc 7: 2628 11. Patterson M, Walters D, Browder W (1993) Postoperative bowel obstruction following laparoscopic surgery. Am Surg 59: 656657 12. Pickleman J, Lee RM (1989) The management of patients with suspected early postoperative small bowel obstruction. Ann Surg 210: 216219 13. Quatromoni JC, Rosoff L Sr, Halls JM, Yellin AE (1980) Early postoperative small bowel obstruction. Ann Surg 191: 7274 14. Reissman P, Shiloni E, Gofrit O, Rivkind A, Durst A (1994) Incarcerated hernia in a lateral trocar sitean unusual early postoperative complication of laparoscopic surgery. Case report. Eur J Surg 160: 191192 15. Seror D, Feigin E, Szold A, Allweis TM, Carmon M, Nissan S, Freund HR (1993) How conservatively can postoperative small bowel obstruction be treated? Am J Surg 165: 121126 16. Stewardson RH, Bombeck CT, Nyhus LM (1978) Critical operative management of small bowel obstruction. Ann Surg 187: 189193 17. Stewart RM, Page CP, Brender J, Schwesinger W, Eisenhut D (1987) The incidence and risk of early postoperative small bowel obstruction: a cohort study. Am J Surg 154: 643647 18. Tsang S, Normand R, Karlin R (1994) Small bowel obstruction: a morbid complication after laparoscopic herniorrhaphy. Am Surg 60: 332334 19. Wegener ME, Chung D, Crans C, Chung D (1993) Small bowel obstruction secondary to incarcerated Richters hernia from laparoscopic hernia repair. J Laparoendosc Surg 3: 173176

Surg Endosc (1998) 12: 10781081

Springer-Verlag New York Inc. 1998

Optimizing laparoscopic splenectomy

Technical details and experience in 59 patients
A. Szold, B. Sagi, H. Merhav, J. M. Klausner
Division of Laparoscopic Surgery, Department of Surgery, Tel Aviv Sourasky Medical Center and The Sackler School of Medicine, Tel Aviv University, 6 Weizman Street, Tel Aviv 64239, Israel Received: 3 April 1997/Accepted: 28 October 1997

Abstract Background: Laparoscopic splenectomy (LS), like other advanced laparoscopic procedures, is still an evolving procedure. The indications for surgery, criteria for patient selection, and operative technique are not yet well defined. We have therefore modified the standard technique for performing LS in an attempt to optimize the procedure. Methods: Over the past 2 years, we have performed LS in 59 patients. The last 43 patients were operated using a standardized technique that we believe to be optimal. It includes the routine use of the right lateral position, operating through three trocars, the mass transection of the splenic vasculature with a vascular endoscopic stapler, and the use of a self-retaining retrieval bag. Results: The average operating time was 79 min. Average blood loss was 95 cc, and average postoperative hospitalization was 2.3 days. There was one intraoperative complication and one postoperative complication. These results are superior to those we achieved earlier in our own experience, as well as to similar series that have been published recently. Conclusions: In our experience, the use of this new technique resulted in relatively short procedures with low morbidity. We believe that these results justify the use of LS as the procedure of choice for elective splenectomy in patients with normal or moderately enlarged spleens. Key words: Laparoscopic splenectomy Spleen Hematological disorders

nical aspects of LS comprise a number of steps, each of which represents a different problem for the surgeon and may be tackled in several ways. These steps are: 1. 2. 3. 4. 5. 6. 7. Patient positioning Introduction of trocars Dissection of the spleen from surrounding organs Control of the splenic artery Division of the splenic hilum Division of the short gastric vessels Removal of the specimen

In order to improve on the technique, we have tried to modify not only the individual steps of the procedure but also the sequence in which they are performed. After experimenting with various techniques over the past 2 years, we are now performing the procedure in a way we believe is optimal. Our modifications allow us to perform LS quickly, with minimal blood loss. Patients and methods
Since April 1995, we have performed 59 LS for various hematological disorders. In the past 18 months, we have used a standardized approach in 43 patients. The demographic data, diagnosis, operative data, and postoperative course were recorded.

Surgical technique
The patient is anesthetized and a nasogastric tube is inserted. The patient is then placed in the right lateral position. The table is slightly extended in order to enlarge the space between the costal margin and the iliac crest. The surgeon stands on the patients right, with the assistant sitting cephalad to the surgeon. The screen is placed across the table (Fig. 1). This position allows for good visualization of the operative field with almost no need for any further retraction. In addition, exposure of the short gastric vessels is excellent. The first incision is made 3 cm from the costal margin, at the midclavicular line. The abdomen is insufflated through a Verres needle, and the first trocar is introduced through the incision. A 30 laparoscope is introduced, and the second trocar is placed close to the costal margin just below the xyphoid. This trocar has to be introduced under direct vision

Since the introduction of laparoscopic splenectomy (LS) [3, 4, 6], several modifications of the procedure have been proposed to make its performance safer and easier. The techPresented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), San Diego, California, USA, 2021 March 1997 Correspondence to: A. Szold

1079

Fig. 2. Transection of the splenic hilum using an endoscopic stapling device. Once the spleen has been dissected from the diaphragm, a large laparoscopic clamp can be used for upward retraction.

Results
Fig. 1. Patient positioning and trocar placement for elective laparoscopic splenectomy. Note that the assistant is sitting cephalad to the surgeon.

because it is placed very close to the left lobe of the liver. The laparoscope is moved to the xyphoid trocar, and a third trocar is introduced at the same distance from the costal margin, at the posterior axillary line. The two lateral ports are used for the surgeons hands (Fig. 1). The procedure is started by a thorough search for accessory spleens. Then, the splenic flexure of the colon is dissected down, using coagulating scissors. The lower pole of the spleen is almost always supplied by a separate branch of the left gastroepiploic artery. We use a clip to ligate it proximally, away from the spleen, and divide it after coagulating the distal side with bipolar forceps. We try to avoid the use of any clips close to the hilum, because clips in this area may interfere with the stapling device. If the patient is somewhat fat, the peritoneum overlying the anterior aspect of the hilum is incised carefully, so the fat can be pushed down to expose the blood vessels. Once the anterior aspect of the hilum is exposed, the left gutter is dissected and the spleen is separated from the kidney and the diaphragm, except for the last posterior attachments of the upper pole. At this point, the spleen remains attached almost entirely by the vascular supply, and the gap between the splenic hilum and tail of the pancreas is enlarged. The spleen can now be lifted, either by grasping it gently with a 10-mm ring clamp, or by placing one of the arms of the ring clamp on either side of the hilum and using it, open, as a retractor, to lift up the spleen (Fig. 2). An endoscopic stapling device (EZ 35, Ethicon, Cincinnati, OH, USA) is introduced through the most lateral port and used to divide the hilum. This maneuver usually requires one or two applications of the instrument. An additional application is used to divide the short gastric vessels. The last attachments to the diaphragm are incised with coagulating scissors. A laparoscopic bag is then introduced into the abdominal cavity. A bag with a self-retaining opening device (Endocatch II, U.S. Surgical, Norwalk, CT, USA) makes it much easier to place the spleen in the bag, so we use it routinely. After the spleen is in the bag, it is drawn through the most lateral port and removed under direct vision. While the opening of the big is held, the spleen is crushed using a regular ring clamp, with the aid of a powerful suction device. The bag is not very durable, so care must be taken not to damage it, lest splenic tissue be disseminated into the abdomen. After the spleen is withdrawn, the operative field is irrigated with lactated Ringers solution, and another inspection is made for bleeding or a missed accessory spleen. The trocars are withdrawn and the fascia closed with an absorbable suture. Finally, the nasogastric tube is removed, except in patients in whom the procedure was long, or when the dissection of the short gastric vessels was especially difficult.

A total of 43 patients were operated using this modified technique. Thirty-six patients were operated for idiopathic thrombocytopenic purpura (ITP), two for thrombotic thrombocytopenic purpura (TTP), two for autoimmune hemolytic anemia, one for hereditary spherocytosis, and two for splenic masses or cysts. The largest spleen found measured 16 cm. In eight patients (18%), an accessory spleen was found and removed. Average operating time was 79 min (range, 40130). Estimated blood loss was 95 cc (range, 0700); only one patient required a transfusion. Except for bleeding, there were no intraoperative complications. There were no conversions to laparotomy. There was one (2.3%) major complicationa subphrenic abscessand two (4.6%) minor onesone patient had an atelectasis, which was treated conservatively, and one was readmitted for fever. The fever responded to antibiotics, and no collection was seen on CT scan. Patients were discharged after 2.3 days (range, 19). No late postoperative complications were found on a mean follow-up of 11.3 months (range, 119). Discussion After experimenting with several different approaches, we found a technique that enables us to perform LS rapidly with very low morbidity rates. The reasoning behind this approach is discussed following the sequence of steps for performing LS. Patient positioning The first reports of LS suggested placing the patient in the supine or the modified lithotomy position [24, 6]. However, as experience was gained, more and more surgeons began advocating the right lateral position with the surgeon on the patients right [5, 7]. This position allows an excellent view of the left upper abdominal cavity while the stomach and bowel are naturally retracted by the gravity away

1080

from the working field, exposing the hilum of the spleen. In our series, we performed the procedure with a single assistant, who held the camera. We found that having this assistant sit on the surgeons left allows more freedom of movement for the surgeon. The only situation in which a supine position is preferable is when an additional procedure most commonly a cholecystectomyhas to be performed simultaneously (in patients with hereditary spherocytosis, for example). Reports describing the technique for the supine position argue that this position allows better control in case an emergency laparotomy is needed. It seems to us that the lateral position allows good control in laparotomy as well; it is, indeed, used by some surgeons for elective open splenectomy. We believe that this position is a far superior position for LS.

Introduction of trocars Most authors suggest the use of four to six trocars [26, 9]. They are used for the laparoscope, retractors, and the two instruments in the surgeons hands. However, our experience shows that, in most cases, the procedure can be performed with only three ports. Utilizing the right lateral position simplifies the retraction and enables the introduction of fewer trocars. We also found that with experience the spleen could be actually grasped and manipulated without injury, usually with a soft bowel clamp, thus necessitating less retraction.

infarction, and three patients developed pulmonary complications attributed to the embolization. Another strategy was to clip the splenic artery away from the spleen before beginning to mobilize the spleen, by opening the lesser sac, dissecting the artery, and placing a clip on it. This technique was previously reported in 17 patients [11], and performed by us. However, we later abandoned this part of the procedure for the following two reasons: First, with the patient in the right lateral position, which we believe is advantageous for most of the procedure, it is very hard to dissect the splenic artery in the lesser sac. Second, we have found that with good exposure of the hilum, it is possible to transect it safely with a stapling device, making proximal ligation of the artery unnecessary. We have also encountered problems caused by dissection of the splenic artery in the lesser sac, mostly in patients in whom the artery was found to be covered with pancreatic tissue, and the dissection consumed a substantial amount of operating time. Although several authors have argued that a separate ligation of the artery and vein is necessary [2, 10], we believe that the danger of creating an arteriovenous fistula by performing a mass transection of the artery and vein is only a theoretical one. In fact, we have observed no such complications in our patients. It seems to us that with correct use of the stapler, only one to three applications are needed to completely transect the vascular supply, and that the time and blood loss saved more than justify the price of the instrument.

Dissection of the spleen and division of the splenic hilum In 1995, Delaitre et al. first described the hanged spleen technique [5]. They reasoned that leaving the posterior attachments intact until the hilum is divided would facilitate exposure and transection of the vascular bundle. After experimenting with this technique, we have decided to modify it. We found that if the spleen is not detached from the diaphragm first, using a stapling device may be dangerous since the posterior arm of the instrument is introduced blindly, allowing potential injury to the posterior aspect of the hilar vessels. We therefore dissect all attachments, except for one posterior point, until the spleen floats, while the blood vessels are the last attachment to be transected. We use the posterior attachments of the upper pole as an anchor to facilitate exposure of the hilum and short gastric vessels.

Division of the short gastric vessels As with the vascular pedicle, suggested methods of dividing the short gastric vessels included the dissection of separate vessels and the use of clips, and the Ultrasonic coagulating shears device (Ultracision, Ethicon) [unpublished data]. In our patients, we continued to apply the stapler serially until all blood vessels were divided. As with the splenic artery and vein, we found that the serial application of the vascular stapler is safe and less time-consuming than a tedious dissection of separate blood vessels. Although it may be less technically demanding (and indeed less elegant), we strongly recommend the use of a vascular stapler for the division of all the blood supply to the spleen. In some patients the spleen has a concave configuration. In these patients, we found that dividing the short gastric vessels first, either with the use of the harmonic scalpel or bipolar coagulating forceps, makes it easier to place the spleen in a good position for the application of the stapling device on the hilum.

Control of the splenic artery In the evolution of LS, several methods of controlling the main splenic vessels were proposed. After performing several procedures with a substantial blood loss, Poulin et al. [10] advocated routine angiographic embolization of the splenic artery prior to surgery. In their series of 20 patients, blood loss was reduced as compared to their earlier experience. However, the embolization was not an innocent procedure. Several patients required narcotics for severe abdominal pain, one patient suffered from a segmental liver Removal of the specimen Placement of the spleen, especially an enlarged one, in a bag can be an extremely frustrating task. We found that using a self-retaining bag makes this task much easier than it might be with a simple plastic bag. The spleen can be grasped gently with a soft bowel clamp and placed in the bag. However, care must be taken to avoid any injury to the spleen

1081

and possible splenosis. We believe that it is unnecessary to make this procedure a laparoscopic-assisted one, or to make an incision merely to remove the specimen or manipulate it. We prefer to devote more time to crushing the spleen and removing it slowly with the aid of a powerful suction device. The results in this group of patients represent an improvement over those previously published. Most reports of LS from groups that have passed the initial-experience stage (with 1623 patients) [1, 2, 711] report operating times in the range of 150200 min. Although in a few cases, cholecystectomy was also performed and some patients had enlarged spleens, these operating times are still significantly longer than the one reported here. This cannot be attributed entirely to the additional procedure. We believe that our reduced operating time results from the fact that all the patients were treated by one team and from our modified operative technique. The most time-saving aspect of our technique is the fact that all the blood supply is transected with the use of a stapler. It is also possible that the small number of patients who had significant bleeding can be explained by our avoidance of the dissection of single vessels at the hilum, a manipulation that may occasionally incur major vascular injury. In several studies where the estimated blood loss is given, it is in the range of 200400 cc. The methods used to estimate the blood loss are not described. For our part, we have found it an extremely difficult task in laparoscopic procedures. For example, in the patient who was transfused, the initial blood loss was estimated at 300 cc; but his hemoglobin dropped significantly postoperatively, requiring blood transfusion. The estimated blood loss in our study reflects the fact that most patients (91%) had no significant bleeding. However, three of our patients suffered some vascular injury, with subsequent bleeding that had to be controlled. We found that the size of the resected spleen is also difficult to estimate during the operation. We do not routinely perform a preoperative CT scan or sonogram, and the only patients for whom we hesitate to attempt LS are those whose spleen is palpable well below the rib cage. We estimate the spleen span both before we start the dissection and during the operation, and we make an attempt to resect it laparoscopically in all cases. The weight of the resected organ is difficult to estimate because we remove the spleen in pieces, and some of the tissue volume is lost in the suction device. We believe that not finding an accessory spleen on CT scan or sonogram does not necessarily exclude its possible existence. Thus, we make most of our efforts to find an accessory spleen during the operation. The number of accessory spleens found in our patients (18%) is satisfactory compared to historical series of splenectomy for hematological disorders. In two patients with recurrent ITP 6 and 10 months following LS (one was referred, and the other was from our series), a missed accessory spleen was demonstrated and removed laparoscopically, with resolution of the disease. We conclude that LS, using the technique described, is a safe and efficient procedure and that it should be accepted as the procedure of choice for elective splenectomy in patients without massive splenomegaly. It is possible that in

time it will prove to be safer and more cost-effective than medical treatment in selected hematological disorders.

References
1. Brunt LM, Langer JC, Quasebarth MA, Whitman ED (1996) Comparative analysis of laparoscopic versus open splenectomy. Am J Surg 172: 596601 2. Cadiere GB, Verroken R, Himpens J, Bruyns J, Efira M, De Wit S (1994) Operative strategy in laparoscopic splenectomy. J Am Coll Surg 179: 668672 3. Carroll BJ, Phillips EH, Semel CJ, Fallas M, Morgenstern L (1992) Laparoscopic splenectomy. Surg Endosc 6: 183185 4. Cushieri A, Shimi S, Banting S, Velpen GV (1992) Technical aspects of laparoscopic splenectomy: hilar segmental devascularization and instrumentation. J R Coll Surg Edinb 37: 414416 5. Delaitre B (1995) Laparoscopic splenectomy: the hanged spleen technique. Surg Endosc 9: 528529 6. Delaitre B, Maignien B (1991) Laparoscopic splenectomy: one case ]Letter). Presse Med 44: 2263 7. Hashizume M, Sugimachi K, Kitano S, Shimada M, Baba H, Ueno K, Ohta M, Tomikawa M (1994) Laparoscopic splenectomy. Am J Surg 167: 611614 8. Katkhuda N, Waldrep DJ, Feinstein D, Soliman H, Stain SC, Ortega AE, Mouiel J (1996) Unresolved issues in laparoscopic splenectomy. Am J Surg 172: 585590 9. Phillips EH, Carroll BJ, Fallas MJ (1994) Laparoscopic splenectomy. Surg Endosc 8: 931933 10. Poulin EC, Thibault C, Mamazza J (1995) Laparoscopic splenectomy. Surg Endosc 9: 172177 11. Zamir O, Szold A, Matzner Y, Ben-Yehuda D, Seror D, Deutsch I, Freund HR (1996) Laparoscopic splenectomy for immune thrombocytopenic purpura. J Laparoendosc Surg 6: 301304

Discussion Dr. Roll: Are there contraindications to laparoscopic splenectomy? Dr. Szold: The size of the spleen is an important factor. Although we have removed several spleens that were in the range of 2025 cm, we presently do not do any spleens larger than 30 cm. Dr. Greenspun: Im wondering if you do any preoperative searches for accessory spleens? Any preoperative imaging studies for accessory spleens? Dr. Szold: We actually dont. We found that most of the available studies are not reliable enough for finding accessory spleens, except for the isotope scan. Since most of the accessory spleens are within the close vicinity of the spleen, they are usually overshadowed by the splenic mass. We had some patients whove had these tests done before they were referred to us, and had no accessory spleens seen on the scan, in whom we subsequently found accessory spleens at surgery. In our laparoscopic series we have found accessory spleens 14 percent of the time. This is equivalent to the results from large open series of patients. One of our open patients developed recurrent ITP and was shown to have a retained accessory spleen that was subsequently removed laparoscopically.

Proceedings
Surg Endosc (1998) 12: 10921103 Springer-Verlag New York Inc. 1998

General oncologic effects of the laparoscopic surgical approach

1997 Frankfurt international meeting of animal laparoscopic researchers
R. L. Whelan,1 J. D. F. Allendorf,1 C. N. Gutt,2 C. A. Jacobi,3 D. Mutter,4 H. R. Dorrance,5 M. Bessler,1 H. J. Bonjer6
1 2 3

Department of Surgery, Columbia University, New York, NY 10032, USA Klinik fr Allgemeiner und Abdominalchirurgie, Johann-Wolfgang-Goethe Universitt, Frankfurt, Germany Department of Surgery, University Hospital Charit, Berlin, Germany 4 Department of Surgery, University Hospital, Strasbourg, France 5 Department of Surgery, The Western Infirmary, Glasgow, United Kingdom 6 Department of Surgery, University Hospital Dijkrigt, Rotterdam, The Netherlands Received: 17 December 1997/Accepted: 2 January 1998

In March 1997 an international meeting regarding animal research in the field of laparoscopic surgery was held in Frankfurt, Germany. The Johann Wolfgang Goethe University was the host institution, and Drs. Carsten N. Gutt and H. Jaap Bonjer were the organizers of the meeting. Representatives of 8 countries and 16 laboratories attended this meeting. The purpose of this gathering was to bring together researchers with a common interest in laparoscopic surgery for an intensive 2-day meeting during which the results of animal studies from each laboratory would be presented and discussed. A total of 30 presentations were made during the meeting concerning the following topics: (1) physiology of open and laparoscopic approaches, (2) immunologic consequences of open and laparoscopic surgery, (3) general oncologic effects, and (4) port-site tumor metastases. In addition, a live rat surgery session was held, during which each research group had the opportunity to demonstrate the various surgical models used for their studies. A brief synopsis of the papers concerning the general oncologic impact of minimally invasive surgical techniques follows. This article is concerned with the impact of open and laparoscopic techniques on tumor growth outside the abdominal cavity (i.e., the systemic oncologic effects of abdominal surgical methods). Port-site tumor metastases are covered in a separate article and therefore are not discussed here. The goal of this article is to provide a brief summary of the relevant studies presented at the meeting. The reader is referred to the References for further information. Although most of the following discussion was presented at the Frankfurt meeting, mention is also made of other im-

portant studies in this field for the sake of completeness and clarity. It has been established in several rodent models that tumor growth is accelerated after laparotomy compared with anesthesia alone [5, 6, 9]. To determine the oncologic effects of minimally invasive abdominal surgical techniques, numerous investigators have performed animal studies over the past 5 years. The pertinent studies have been divided into four categories on the basis of the following study designs: (1) comparing sham laparotomy and pneumoperitoneum (pneumo), (2) comparing open and minimally invasive bowel resection, (3) comparing air, CO2, and helium pneumo, (4) looking at the rate of tumor cell proliferation. The end point of the first three categories of study was generally tumor size or volume, whereas the fourth category used an immunohistochemical assay.

Tumor studies comparing sham laparotomy and pneumoperitoneum The Columbia University group presented the results from a number of murine studies of this type. The mouse mammary carcinoma (MMC) MC-2 cell line was used for their initial studies, and the length of the procedures was standardized to 30 min. A full xyphoid to pubis incision was compared with a CO2 pneumo and anesthesia alone in these studies [1]. A suspension of tumor cells was injected intradermally into the dorsal skin of the animals on the day of surgery before the interventions. Both high- and low-dose experiments were carried out. In the high-dose study, all animals were expected to grow tumors, and the end point was tumor mass on postoperative day 12. Regarding the high-dose experiment results, the open group tumors were significantly larger than those of the

Correspondence to: R. L. Whelan

1093

pneumo and control groups. Of note, the pneumo group tumors, although smaller than those of the open groups, were significantly larger than the anesthesia control group lesions. In the low-dose study, only a small percentage of tumors were expected to develop in the control group. The incidence of tumor establishment 30 days after the interventions was determined for each of the three study groups. Significantly more tumors developed in the open group than in either the open or pneumo groups. There was no significant difference in the rate of tumor formation between the latter two groups. In a later study, the same investigators repeated the high-dose intradermal tumor cell injection study with the colon-26 mouse adenocarcinoma and the B-16 melanoma cell lines and again found the open group tumors to be significantly larger than the mean tumor size of both the pneumo and control groups [10]. In a final study by the Columbia group using the MMC tumor cell line, the size of intradermal tumors 12 days after sham laparotomy, pneumo, or anesthesia alone in athymic mice was determined and compared [2]. The athymic mice had no T-lymphocytes. A high-dose injection was used, which results in tumors in practically all animals. This study was prompted by the results of earlier rat studies that had documented significantly better cell-mediated immune function after pneumo than after full laparotomy [12]. The hypothesis of the nude mouse study was that the previously observed differences in tumor growth between the open and pneumo groups would be lost in athymic mice. In fact, this proved to be the case, although the tumors of both surgery groups were similarly larger than the anesthesia control group lesions. This suggested that postoperative immune function plays a role in limiting the rate of tumor growth in immunocompetent mice. However, the fact that the control group tumors were significantly smaller than those of the intervention groups suggested, furthermore, that other factors were also influencing tumor behavior after surgery. Mutter [8] from Strasbourg, France presented the results of a study that assessed tumor growth in the pancreas and at an intradermal extra-abdominal site after 2-h full laparotomy, CO2 pneumoperitoneum, or anesthesia alone. A pancreatic adenocarcinoma cell line was used for this study. A tumor cell suspension was injected into either the skin of the neck or the tail of the pancreas via a small preliminary laparotomy before the test intervention. The animals were killed 28 days after the procedures, and the tumors were excised, measured, and weighed. The results are interesting but difficult to interpret. At the pancreatic location, the control group tumors were significantly larger than the lesions of both the laparotomy and CO2 pneumo groups. There was no difference between the tumors of the latter two groups. Regarding the tumors in the neck, those of the control group were significantly larger than the lesions of the open group, yet significantly smaller than the tumors of the laparoscopy groups, which were the largest. These results would suggest that abdominal surgery confers a protective effect against tumor growth that is not in keeping with the published results of numerous other studies in this field. It may be that the lengthy surgical procedures in nonintubated rats resulted in respiratory acidosis or some other metabolic derangement that influenced tumor growth.

Tumor studies comparing laparoscopic-assisted and open bowel resection Three studies of this type have been carried out to date, and the results were presented at the Frankfort meeting. The Columbia University group determined and compared mouse mammary carcinoma tumor size in study animals 12 days after a high-dose day of surgery intradermal injection of tumor cells followed by either an open or laparoscopicassisted cecectomy. Tumor cell injections were again made in the dorsal skin of study mice. A CO2 pneumo was used in the laparoscopic group. A stepwise increase in tumor size was observed from the anesthesia control group to the laparoscopic-assisted group to the open cecectomy group. The open group tumors were significantly larger than those of the laparoscopic-assisted groups, which, in turn, were significantly larger than the tumors of the control groups [2]. In a second experiment by the Columbia group using the same cecectomy model, the incidence of tumor establishment was determined after a low-dose injection of tumor cells that results in tumors in a minority of control animals. The incidence of tumors was greatest in the open group (83%) followed by the laparoscopic-assisted group (30%) and the anesthesia control group (5%). Significant differences were found for all comparisons between these three groups [2]. The Dutch investigators from the University Hospital Dijkzigt presented the results of a study that compared tumor growth after open and laparoscopic-assisted small bowel resection in rats. The CC-531 colon adenocarcinoma tumor cell line was used for this study. A standardized small cube of tumor was implanted beneath the capsule of the kidney via a 2.5-cm abdominal incision several days before the small bowel procedure was carried out. A handsewn extracorporeal small bowel anastomosis was carried out for all animals. Twelve days later the tumors were excised and weighed. The open groups mean tumor size was significantly greater than the average tumor size of laparoscopicassisted and anesthesia control groups. There was no significant difference in mean tumor size between the latter two groups [3]. Tumor growth studies comparing different minimal-access exposure methods Jacobi presented the results of a rat study that determined and compared intradermal tumor growth after CO2 pneumo, helium pneumo, and anesthesia alone [7]. No laparotomy group was included in this study. The tumor cell injections were carried out on the day of surgery, and 5 weeks later the animals were killed and the tumors excised and weighed. The tumors of the CO2 pneumo group were significantly larger than those of the other two groups, which were of similar size. In vitro and ex vivo studies were also performed, and they demonstrated greater tumor growth in the CO2 groups. These results suggest that the CO2 pneumo has a systemic effect of encouraging tumor growth at extraabdominal sites. These findings are in keeping with several port-site studies that have demonstrated an association between higher port tumor rates and CO2 pneumo. In contrast, Southall [11] presented the results of a murine study that found no significant difference in intradermal tumor size

1094

when the impact of CO2 was compared with air pneumo. Of note, the tumors of a sham laparotomy group included in Southall et al.s study were significantly larger than the tumors both the air and CO2 pneumo groups. Dorrance [4] presented the results of a rat experiment that determined the influence of air, helium, and CO2 pneumo on the rate of pulmonary metastases after tail vein injection of tumor cells. There were no significant differences in the number of pulmonary metastases found 3 weeks after the procedures when the three different pneumo groups and an anesthesia control group were compared.

Tumor cell proliferation studies Allendorf [2] presented the results of a study that assessed tumor cell proliferation rates directly by determining the level of proliferating cell nuclear antigen (PCNA) via immunohistochemical staining of tumor sections. The PCNA is an intranuclear protein essential for DNA synthesis that is maximally expressed in the S-phase of the cell cycle. This assay has been documented to provide an accurate determination of cell proliferation. In this study, extra-abdominal intradermal mouse mammary carcinoma tumors were assessed 6 and 12 days after sham full laparotomy, CO2 pneumo, and anesthesia alone. The proliferation index was found to increase in a stepwise fashion from the anesthesia control to the CO2 pneumo to the sham laparotomy group. The proliferation index of the open groups was significantly higher than the index of the other two groups both 6 and 12 days after surgery. The CO2 pneumo PCNA index, although smaller than that of the open group, was significantly larger than that of the control groups at both time points.

Summary and conclusions The results from the majority of the reviewed studies support the hypothesis that abdominal surgery, performed via either a large incision or CO2 pneumoperitoneum, systemically encourages tumor growth in the postoperative period. A full laparotomy incision appears to have a significantly greater effect than a CO2 pneumoperitoneum on postoperative tumor growth. Whether the larger tumors observed in the surgical groups are the result of increased tumor cell proliferation or diminished tumor cell death remains unclear. There is some evidence pointing to both mechanisms. The loss of the postoperative tumor growth differences between the open and pneumo animals in the athymic mouse experiment suggests that cell-mediated immune function plays a role in tumor containment. The proliferation study results, however, suggest that other stimulatory influence(s) are also at work. Clearly, much research needs to be done regarding the etiology of these tumor growth differences. Other tumor cell lines need to be studied, and investigations regarding tumor growth in an intra-abdominal location need be performed as well. This body of research suggests that the manner in which the surgeon gains access to the abdominal cavity may have an impact on the propensity of tumor cells to implant, survive, and grow in the period immediately after surgery. If

true, this may be the most compelling justification for the use of minimally invasive techniques for the curative resection of malignancies. However, it remains to be proven that human tumors will demonstrate differences in tumor growth similar to those noted in some of these animal models. Furthermore, it is not at all clear that slight differences in tumor growth postoperatively will translate into significant differences in long-term survival or recurrence rates. At first glance, the existence of port-site tumors would appear to contradict totally the conclusions of many studies discussed in this synopsis. If laparoscopic methods are associated with decreased rates of tumor growth and establishment, then why do port-site tumors form? This is a complex issue calling for discussion that goes far beyond the scope of this article. However, several brief comments on this topic follow. The etiology of port tumors is unknown, although traumatization of the tumor during mobilization, resection, or removal is likely to play a significant role in the liberation of tumor cells from the primary. A relatively small protective benefit, in terms of slower tumor growth rates in laparoscopic patients, will likely not be sufficient to prevent a large inoculum of viable tumor cells in an abdominal wound from establishing a metastasis. Furthermore, as suggested earlier, the systemic effects on tumor growth may be different from the local (i.e., intra-abdominal or abdominal wound) effects. Finally, the true incidence of port tumors remains unknown. It has not been definitively established that the laparoscopic wound tumor incidence is significantly higher than the open rate, although this is the assumption of most surgeons. Several relatively large recently published laparoscopic series have reported port tumor incidences of 0 to 1.2%, which is in the same ballpark as the 0.6 1.0% abdominal wound tumor incidences mentioned in several open colectomy series. Clearly, much more research in this area is needed to understand port tumors better and to reconcile the port tumor results with the systemic tumor growth benefits that may be associated with minimally invasive methods. References
1. Allendorf JDF, Bessler M, Kayton M, Whelan RL, Treat MR (1995) Increased tumor establishment and growth after laparoscopy and laparotomy in a murine model. Archives Surg 130: 649653 2. Allendorf JDF, Bessler M, Whelan RL, Laird DA, Bertram A, Marvin M, Kim L, Horvath KA, Treat MR (1996) Differences in tumor growth afer open vs. laparoscopic surgery are lost in an athymic model and are associated with differences in tumor proliferative index. Surgical Forum 47: 150152 3. Bouvy ND, Marquet RL, Jeekel J, Bonjer HJ (1997) Laparoscopic surgery is associated with less tumour growth stimulation than conventional surgery: an experimental study. Brit J Surg 84: 358361 4. Dorrance HR, Oein K, ODwyer PJ Effects of laparoscopic insufflation on intraperitoneal tumour growth and distant metastases in an animal model. Presented at the Frankfurt meeting of animal laparoscopic researchers, March 1997, submitted for publication 5. Eggermont AM, Steller EP, Marquet RL, Jeekel J, Sugarbaker PH (1988) Local regional promotion of tumor growth after abdominal surgery is dominant over immunotherapy with interleukin-2 and lymphokine-activated killer cells. Cancer Detect Prevent 12: 421429 6. Goshima H, Saji S, Furata T, Taneumura H, Takao H, Kida H, Takahashi H (1989) Experimental study on preventative effects of lung

1095 metatastases using LAK cells induced from various lymphocytes special references to enhancement of lung metastasis after laparotomy stress. J Jap Surg Soc 90: 12451250 7. Jacobi CA, Sabat R, Bohm B, Zieren HU, Volk HD, Muller JM (1997) Pneumoperitoneum with carbon dioxide stimulates growth of malignant colonic cells. Surgery 121: 7278 8. Mutter D (1997) Effect of laparoscopy versus laparotomy on intraabdominal and subcutaneous pancreatic tumor in a rat model. Presented at the Frankfurt meeting of animal laparoscopic researchers, March 1997 9. Ratajczak HV, Lange RW, Sothern RB, Hagen KL, Vescei P, Wu J, Halberg F, Thomas PT (1992) Surgical influence on murine immunity and tumor growth: relationship of body temperature and hormones with splenocytes. Proc Soc Exp Biol Med 199: 432440 10. Southall JC, Lee SW, Allendorf JDF, Bessler MD, Whelan RL (1997) Colon adenocarcinoma and B-16 melanoma growth larger after laparotomy vs laparoscopy in a murine model. Dis Colon Rectum 40(6): A20 [abstract] 11. Southall JC, Lee SW, Bessler M, Allendorf JDF, Whelan RL (1998) The effect of peritoneal air exposure on postoperative tumor growth. Surg Endosc 12: 348350 12. Trokel MJ, Bessler M, Treat MR, Whelan RL, Nowygrod R (1994) Preservation of immune response after laparoscopy. Surg Endo 8(12): 13851388

Surg Endosc (1998) 12: 10511054

Springer-Verlag New York Inc. 1998

An improved technique for laparoscopic highly selective vagotomy using harmonic shears
N. Katkhouda,1 D. J. Waldrep,1 G. M. R. Campos,1 E. Tang,1 S. Offerman,1 A. P. Trussler,1 J. Gugenheim,2 J. Mouiel2
1

Division of Emergency Non-Trauma and Minimally Invasive Surgery, Department of Surgery, Healthcare Consultation Center, 1510 San Pablo Street, University of Southern California School of Medicine, Los Angeles, CA 90033, USA Department of Surgery, University of Nice School of Medicine, Nice, France Received: 9 July 1997/Accepted: 11 November 1997

Abstract Background: Results from classic highly selective vagotomy (HSV) are technique dependent because an incomplete operation will result in early recurrence of duodenal ulcer. Few reports describe laparoscopic completion of the procedure. All techniques use clips for division of neurovascular branches, making the laparoscopic approach tedious and thus the results, uncertain. Methods: Ten patients with intractable duodenal ulcer and negative Helicobacter pylori status underwent an extended HSV. All procedures were performed laparoscopically using a new surgical tool, the harmonic shears. Results: All procedures were completed laparoscopically and took approximately 1 h. There were no deaths and no postoperative complications. Patients were discharged the next day. Follow-up endoscopy at 2 months showed healing of duodenal ulcer in all cases, and postoperative acid secretion studies demonstrated a decrease in basal acid output (BAO) by 74% (8.2 meq/h to 2.16 meq/h) and maximal acid output (MAO) by pentagastrin stimulation by 79.2% (40 to 8.32). Conclusions: Harmonic shears expedite laparoscopic HSV. The operation can be taught safely, yields good results in early follow-up, and represents an acceptable option in patients with intractable duodenal ulcers who are H. pylori negative. Key words: Surgery-laparoscopic Vagotomy Duodenal ulcer Harmonic shears

Highly selective vagotomy (HSV) as the definitive treatment of peptic ulcer disease was first described by Griffith and Harkins [8] in 1957 and subsequently popularized by
Correspondence to: N. Katkhouda

Johnston [10]. The object of the procedure is to divide all vagal nerve fibers innervating the acid-producing cells of the stomach while preserving the terminal branches of the main vagal trunks and the nerves of Latarjet, thereby maintaining adequate antral motility. Numerous clinical studies have found this procedure to be both safe and effective when performed by experienced surgeons [6, 11, 12]. Improved skills have allowed minimal access techniques to assist or replace many open procedures. Yet, few reports have described laparoscopic completion of the traditional HSV [4, 7]. The procedure can be difficult and tedious because each branch of the anterior and posterior vagus nerve must be individually dissected, ligated, and divided. Success of HSV depends on meticulous technique because leaving a single fundic nerve branch intact will allow continued acid secretion in the corresponding gastric secretory zone, leading to early recurrence [3, 6]. In fact, many surgeons find laparoscopic HSV to be technically difficult and time consuming, especially in the obese patient. Therefore, alternative procedures have been advocated such as posterior truncal vagotomy with anterior seromyotomy as popularized by Taylor [15] and redefined laparoscopically by the senior author (NK) of this article [13]. Another technique is the posterior truncal vagotomy and anterior highly selective vagotomy described by Hill and Barker in 1978 [9] and performed laparoscopically by Zucker and Bailey [2]. This procedure was not validated in open surgery, and the laparoscopic reports are too scarce to contribute to its acceptance. These procedures represent attempts to circumvent the difficulties of traditional laparoscopic HSV. The ideal laparoscopic technique is one that is safely taught, reproduced by many, and yields the same or better results than open surgery. The recent development of ultrasonically activated coagulating shears (LCS, Ethicon Endosurgery, Cincinnati, OH) presents a new alternative to the tedious individual ligation of vagus nerve branches and associated vasculature [14, 15].

1052 The anterior leaf of the omentum is incised between the nerve and the stomach proximal to the crows foot, and the harmonic shears are introduced into this opening and positioned parallel to the nerve trunk (Fig. 2a). The lesser omentum is retracted using atraumatic graspers to ensure that the harmonic shears divide only the branches of the vagus nerve and not the trunk of Latarjet. There is no need to identify and ligate individual neurovascular bundles. The important technical aspect is to assure that the blunt jaws of the shears are applied and coapted beyond the vessel to avoid a partial welding that might lead to hemorrhage. The tissues tend to separate automatically when the welding process is complete, so no traction needs to be placed on these fragile structures. Four or five applications of the shears should suffice to clear the lesser curvature. This dissection then proceeds proximally to include the esophageal branches of the anterior vagus nerve. The fat pad of the cardioesophageal junction containing the anterior nerve trunk is then raised upward and to the right, with care taken to divide the criminal nerves of Grassi near the angle of His. The division of those aberrant branches of the main vagal trunks is of paramount importance to ensure complete vagotomy of the posterior fundus. The dissection then returns to the distal stomach. At this time, retracting the opened anterior leaf of the lesser omentum exposes the posterior vagal branches. The posterior leaf is incised, as before, between the stomach and posterior nerve of Laterjet, opening the lesser sac. Again, the harmonic shears are positioned parallel to, but away from, the vagal trunk, dividing all the branches of the posterior vagus nerve proximal to the crows foot. At the gastroesophageal junction, the main vagal trunks should be identified and retracted to avoid their injury. The esophagus is exposed circumferentially and an electrocoagulation hook is used to denervate smaller branches easily recognized by the magnification provided by the laparoscope. At the end of the procedure, the distal esophagus should be clear of all nerve fibers for a length of 6 to 8 cm while the two trunks and their terminal gastric branches are preserved. We agree with the concept of extended highly selective vagotomy [6] using the magnification provided by the laparoscope to divide the proximal right anterior gastroepiploic nerve along the greater curvature of the stomach. The nasogastric tube may be removed at the end of the procedure. The patient is encouraged to ambulate and resume a soft diet on the evening of surgery. Discharge is allowed the next day.

Fig. 1. Trocar placement for laparoscopic highly selective vagotomy (HSV).

Patients and methods

Between January, 1995 and April, 1996, 10 patients presented with chronic duodenal ulcers. All were male with a median age of 46 years (range, 36 to 64 years). All patients had intractable symptoms despite adequate medical therapy for a median duration of 3.5 years (range, 2 to 10.5). Two patients had a history of upper gastrointestinal bleeding. None had a history of symptoms suggestive of gastric outlet obstruction such as vomiting or weight loss. All patients had duodenal ulcer disease documented by upper gastroduodenal endoscopy and negative biopsies for Helicobacter pylori. Serum gastrin levels were normal in all patients. No patient was on nonsteroidal anti-inflammatory drug (NSAID) therapy before presentation. Patients gave informed consent in all cases. Senior surgical residents under the guidance of attending faculty performed the procedures. Followup at 2 months included upper endoscopy, antral biopsies, and acid secretion studies. Basal and pentagastrin-stimulated gastric acid secretion was analyzed before and 2 months after surgery. Acid output was calculated by multiplying the volume and acid concentration of 15-min samples. Basal acid output was expressed in millimoles of hydrochloric acid per hour. The maximal acid output induced by pentagastrin stimulation was calculated from the two 15-min samples for which the acid output was the highest.

Results All procedures were completed laparoscopically in a median operative time of 55 min (range, 50 to 72). No intraoperative complications were noted. Patients were observed for 24 h and discharged the next day on a soft diet. There was no postoperative morbidity. All patients experienced resolution of symptoms by 2 months, which correlated to healing ulcers documented by endoscopy. Antral biopsies remained negative for H. pylori. Acid secretion studies for the 10 patients, demonstrated a decrease in basal acid output (BAO) of 74% and a decrease in pentagastrinstimulated maximal acid output (MAO) of 79.2% (Table 1). No patient developed symptoms of gastric stasis. One patient reported intermittent diarrhea, which resolved with medical therapy. All patients remained asymptomatic at 6 months after surgery. Discussion Classic HSV has proven to be effective in the treatment of chronic duodenal ulcer with both a low recurrence rate and minimal morbidity in experienced hands. A recurrence rate of 16.5% at 25-year follow-up has been reported [12]. Limited access techniques may offer the patient not only a shorter and more comfortable postoperative course, but possibly fewer long-term complications associated with laparotomy such as adhesion formation. Laparoscopic HSV has been performed using standard hemostatic clips (Fig. 2b). However, the dissection can be tedious and difficult, gen-

Operative technique
The patient is placed in the inverted Y position with the operating surgeon standing between the legs. The video monitor is positioned at the patients shoulder. Pneumoperitoneum is established by the Veress needle technique at the umbilicus. The positioning of the trocars is depicted in Fig. 1. The procedure begins by the elevation of the left lobe of the liver using a fan retractor. Endoscopic babcock-type clamps provide lateral traction on the greater curvature of the stomach. The lesser omentum is carefully inspected to identify several anatomic landmarks: the avascular aspect of the lesser omentum crossed by the hepatic branch of the anterior vagus nerve; the terminal branch of the anterior vagal nerve, the nerve of Laterjet, which runs parallel to the lesser curvature; and the terminal crows foot.

1053

Fig. 2. The techniques of neurovascular ligation in highly selective vagotomy (HSV): (A) laparoscopic technique using harmonic shears, (B) laparoscopic technique using clips.

Table 1. Effect of HSV on gastric acid secretion BAO (meq/h) Preoperative Postoperative 8.2 (313) 2.16(1.24.1) Pentagastrin stimulated MAO (meq/h) 40 (2861) 8.32(5.812.7)

Results presented as median values and range. BAO, basal acid output; MAO, maximal acid output.

erating poor results due to incomplete vagotomies and leading surgeons to seek alternative techniques such as the posterior truncal vagotomy and anterior lesser-curve seromyotomy. This technique is less familiar to surgeons in North America and thus has limited universal acceptance. The introduction of the harmonic shears may change many surgeons attitudes toward laparoscopic HSV. The concept of the shears is based on the transmission of high frequency mechanical energy to a transducer in the handpiece that creates a longitudinal vibration of 55,000 times per second in one jaw of the clamp to effect a 50100 m excursion. Rapid acceleration of the tissues within the clamp breaks the hydrogen bonds of structural protein, creating a coagulum that seals off coapted vessels up to 6 mm in diameter. Harmonic ligation of the short gastric vessels during Nissen fundoplication has been routinely successful [14, 15]. An additional feature of the harmonic shears is a lateral thermal spread of only 0.75 to 1.5 mm, in contrast to standard electrocautery, which is approximately 3 mm [1]. Necrosis of the lesser curve has been reported as a complication of traditional HSV, and one cause is thought to be extensive use of electrocautery. The limited heat generated by the harmonic shears may prove important in minimizing inadvertent thermal injury to the stomach or the nerve trunks. The other major advantage of the harmonic shears is that branches of the vagus need not be dissected individually.

Four or five applications of the clamp complete the dissection of each leaf of the lesser omentum, thereby greatly reducing the operating time. In one series, 10 patients underwent laparoscopic HSV using standard clips [4]. The median operating time for these experienced laparoscopic surgeons was 110 min (range, 85 to 205) or twice the median operative time in our series of cases performed by surgical residents under faculty guidance. The results of the acid secretion studies in our series showing a reduction of 79.2% in pentagastrin-stimulated MAO at 2 months compare favorably with the results obtained after open HSV, ranging from 59% [12] to 70% in Johnston and Wilkinsens original article [10]. This confirms that the use of a new HSV technique did not adversely affect the outcome and achieved adequate vagal denervation. Although our experience is limited to a small number of patients, we believe that use of the ultrasonically activated harmonic shears significantly enhances the laparoscopic performance of HSV by overcoming extended operating times, tedium, and teaching difficulties attributed to other techniques. Also, there are now fewer patients with peptic ulcer disease who require surgical treatment, limiting the number of cases available to improve the learning curve for any surgical modality. All these factors serve to increase the chances of recurrence because results for HSV are highly technique dependent. The routine use of the laparoscopic harmonic shears overcomes many technical disadvantages, reduces the learning curve, and may further optimize the results for HSV. Although the patients in this series were observed for 24 h postoperatively, we believe this technique is well suited as an outpatient procedure. We did not perform a cost comparison analysis for this procedure, but a reduction in hospital costs can be expected when reduced operating room time combined with the cost of a standard clip applier and refills are considered [14, 15]. In conclusion, this article reports technical feasibility

1054

and good early results. Whether this laparoscopic technique of HSV will produce comparable results to its open counterpart will depend on longer follow-up of a larger number of patients. References
1. Amaral JF (1994) The experimental development of an ultrasonically activated scalpel for laparoscopic use. Surg Laparosc Endosc 4: 9299 2. Bailey RW, Flowers JC, Graham SM, Zucker KA (1991) Combined laparoscopic cholecystectomy and selective vagotomy. Surg Laparosc Endosc 1: 4546 3. Blackett RL, Johnston D (1981) Recurrent ulceration after highly selective vagotomy for duodenal ulcer. Br J Surg 68: 705710 4. Dallemagne B, Weerts JM, Jehaes C, Markiewicsz S, Lombard R (1994) Laparoscopic highly selective vagotomy. Br J Surg 81: 554 556 5. Dixon M (1993) Acid, ulcers and H. pylori. Lancet 342: 384385 6. Donahue PE, Richter HM, Liu KJM, Anan K, Nyhus LM (1993) Experimental basis and clinical application of extended highly selective vagotomy for duodenal ulcer. Surgery 176: 3948 7. Gordon J, Josephs LG (1994) Laparoscopic highly selective vagotomy: definitive therapy for peptic ulcer disease. Int Surg 353356

8. Griffith CA, Harkins HN (1957) Partial gastric vagotomy: an experimental study. Gastroenterology 32: 96102 9. Hill GL, Barker MC (1978) Anterior highly selective vagotomy with posterior truncal vagotomy: a simple technique for denervating the parietal cell mass. Br J Surg 65: 702705 10. Johnston D, Wilkinson AR (1970) Highly selective vagotomy without a drainage procedure in the treatment of duodenal ulcer. Br J Surg 57: 289296 11. Johnston GW, Spencer EF, Wilkinson AJ, Kennedy TL (1991) Proximal gastric vagotomy: follow-up at 1020 years. Br J Surg 78: 2023 12. Jordan PH, Thornby J (1994) Twenty years after parietal cell vagotomy or selective vagotomy antrectomy for treatment of duodenal ulcer. Ann Surg 220: 283296 13. Katkhouda N, Mouiel J (1991) A new technique of surgical treatment of chronic duodenal ulcer without laparotomy by videocoelioscopy. Am J Surg 16: 361364 14. Laycock WS, Trus TL, Hunter JG (1996) New technology for the division of short gastric vessels during laparoscopic Nissen fundoplication. Surg Endosc 10: 7173 15. Swanstrom LL, Pennings JL (1995) Laparoscopic control of short gastric vessels. J Am Coll Surg 181: 347351 16. Taylor TV, Gunn AA, Macleod DAD, MacLennan I (1982) Anterior lesser curve seromyotomy with posterior truncal vagotomy in the treatment of chronic duodenal ulcer. Lancet ii: 846848, 1957; 32: 96102

Surg Endosc (1998) 12: 10201024

Springer-Verlag New York Inc. 1998

Normal T lymphocyte and monocyte function after minimally invasive surgery

I. B. Brune, W. Wilke, T. Hensler, H. Feussner, B. Holzmann, J.-R. Siewert
Department of Surgery, Klinikum rechts der Isar, Technical University Munich, Ismaninger Strasse 22, D-81675 Mu nchen, Germany Received: 14 April 1997/Accepted: 20 October 1997

Abstract Background: The aim of this study was to evaluate immune defense mechanisms after laparoscopic (LCHE) and open cholecystectomy (CHE), particularly with regard to monocyte and T-lymphocyte function. Methods: In a prospective study, we evaluated the following immunological data from 27 patients (21 women, six men; mean age, 47.2 years) submitted to elective LCHE and 14 patients (seven women, seven men; mean age, 60.8 years) undergoing elective CHE: T-lymphocyte proliferation (stimulated by SEA, SEB, TSST-1 with antigen presentation by patient monocytes), expression of cell surface molecules on monocytes (HLA-DR, CD80, L-Selectin), CD4+ T lymphocytes (HLA-DR, CD25, ICAM-1, L-Selectin), and granulocytes (L-Selectin). Blood samples were collected preoperatively and on postoperative days 1 and 67. Statistical analysis was performed using the Mann-Whitney U test for paired samples. Results: HLA-DR on monocytes significantly decreased after LCHE during the early postoperative course but returned to preoperative levels within 1 week. After CHE, significant downregulation of HLA-DR expression persisted throughout the whole observation period. This decrease, however, did not alter the antigen-presenting capacity of monocytes in both groups. Moreover, the APC-independent proliferative capacity of T lymphocytes was unimpaired. CD25 expression was significantly increased on postoperative day 1 after CHE but not after LCHE. Expression of HLA-DR, ICAM1, and L-Selectin on CD4+ T cells was not altered in either group. CD80 on monocytes and L-Selectin on monocytes and granulocytes remained unchanged after both procedures. Conclusions: HLA-DR surface molecules on monocytes that are required for antigen presentation were significantly decreased in both groups; they returned to normal within 1 week after LCHE but not after CHE. However, the antigenpresenting capacity of monocytes remained normal in both

groups. T-cell stimulation, reflected by an increase of CD25 expression, was observed only after CHE, not after LCHE. We therefore conclude that LCHE interferes less with immune defense than CHE; however, the clinical relevance of the changes noted after the open operation remains to be determined. Key words: Laparoscopy Immunology Adverse effects Monocytes T lymphocytes

Correspondence to: I. B. Brune, B. Holzmann

Surgical stress has been shown to be associated with postoperative alterations in host immune functions [10, 12, 18, 25]. Moreover, the immunologic response appears to correlate with the severity of trauma [8]. With the development of laparoscopic surgery, the benefit of reduced access trauma became clinically evident. Early reports concentrated on the assessment of serum levels of the primary acute-phase stimulator interleukin-6 (IL6), which correlate with the extent of tissue trauma [6, 15, 21, 24, 27]. It was demonstrated that systemic IL-6 was significantly lower after laparoscopic procedures than after open surgery. As laparoscopic procedures evolved, they were used more often in patients with malignant disease. Since tumor recurrence and therefore the prognosis of patients with malignant diseaseis related to immune function [10], the possibility that laparoscopy has an adverse effect on immunologic defense must be evaluated thoroughly before minimally invasive procedures can be used with curative intent for cancer [22]. The relation of monocyte and T-lymphocyte function to clinical outcome after conventional surgery has been described extensively [5, 12, 19]. Reduced cellular proliferation of T lymphocytes has been shown to be one of the main functional alterations of the adaptive immune system after trauma, burn [11], and major surgery [18], and it seems to be associated with postoperative complications. The presence of major histocompatibility complex class II (MHCII), particularly HLA-DR expression in monocytes, is required for antigen presentation to T lymphocytes. The loss of cell-

1021

surface HLA-DR molecules on monocytes, which has been noted after major surgery and trauma, appears to correlate with the infection rate after surgery [5, 18, 19, 32]. On CD4+ T lymphocytes, the expression of HLA-DR was temporarily increased after major surgery, indicating the presence of elevated numbers of activated T cells in the circulation [32]. The expression of ICAM-1 is induced by inflammatory cytokines and has been shown to increase cell-cell interactions during immune responses [30]. It is related to T-cell mediated hypersensitivity reactions of the skin [29]. ICAM-1 also mediates the aggregation of T lymphocytes [26]. L-Selectin is a cell-surface adhesion molecule expressed by lymphocytes, monocytes, and granulocytes [31]. It mediates leukocyte rolling along vascular endothelium. It also plays an important role in lymphocyte homing and the recruitment of leukocytes to nonlymphoid tissues during inflammation. CD25 is the subunit of the IL-2 receptor and is mainly expressed on activated T lymphocytes. CD25 expression increases with T-cell stimulation [14]. CD80 has been demonstrated to be an important costimulator molecule for antigen presentation to T cells. It is mainly expressed by activated monocytes [1] and B cells. We set out to evaluate and compare the influence of laparoscopy and open surgery on the host defense system. We examined T-lymphocyte proliferation in the presence or absence of the patients antigen-presenting cells (APC) as well as the surface expression of activation markers and adhesion molecules on monocytes, T lymphocytes, and granulocytes after laparoscopic and conventional cholecystectomy in a prospective study. The immunologic parameters analyzed in our study are listed in Table 1. Materials and methods Patients
In a prospective, nonrandomized study approved by the ethics committee of the Technical University of Munich, we analyzed the immune functions of 27 patients undergoing laparoscopic and 14 patients submitted to open cholecystectomy at our surgical department. Patients with common bile duct stones, acquired or inherited immunodeficiencies, acute inflammatory cholesystitis, and patients receiving medication that interferes with the immunologic response (such as immunosupressive drugs administered posttransplantation or steroid or nonsteroid anti-inflammatory drugs) were excluded from the study. Similarly, patients in whom the operation had to be converted to laparotomy or who developed postoperative complications were excluded. In the laparoscopic group, there were 21 women and six men with a mean age of 47.2 years (range, 2978 years). The operation was performed via a standardized technique using four trocars and a pneumoperitoneum of 14 mmHg. Mean operative time was 68 min (range, 25135; SD 23.8). The group who underwent the open operation consisted of seven women and seven men with a mean age of 60.8 years (range, 3385 years). The procedure was carried out through a right subcostal incision. Mean operative time was 75.7 min (range, 55130; SD 20.5). Blood samples from all patients were collected preoperatively (pre-OP) and on postoperative days 1 (post-OP1) and 67 (post-OP2) and tested for the immunologic parameters.

Table 1. Immunologic parameters before and after laparoscopic and open cholecystectomy T-lymphocyte proliferation stimulated by cross-linking of CD3 and CD28 receptors stimulated by SEA, SEB, TSST-1 (presented by patients monocytes) Cell surface molecules on monocytes: HLA-DR, CD80, L-Selectin T lymphocytes (CD4+): HLA-DR, CD25, ICAM-1, L-Selectin granulocytes: L-Selectin

this study were directed against HLA-DR (B8.12.2), L-Selectin (DREG56), CD80 (MAB104), CD25 (B1.49.9), ICAM-1 (84/110), and CD14 (RMO52). Antibodies and isotype-matched control immunoglobulins were purchased from Immunotech (Hamburg, Germany). For threecolor flow cytometry analysis, CD4 antibody FK3 conjugated with PerCP was obtained from Becton Dickinson (Heidelberg, Germany).

Isolation of peripheral blood T cells, monocytes, and granulocytes

Human PBMC (peripheral blood mononuclear cells) were isolated from 25 ml heparinized blood using Ficoll-metrizoate density gradient centrifugation. Isolated PBMC were washed twice with PBS, and the total cell count was determined. PBMC were plated in six-well tissue culture plates (Nunc, Rothskilde, Denmark) in a total of 2 ml PBS and incubated for 1 h at 37C to remove adherent cells. Nonadherent cells were collected, washed with PBS, and resuspended in RPMI 1640 medium. T cells were enriched by depletion of B cells and monocytes from nonadherent cells using immunomagnetic beads coated with CD14 (RMO 52) and CD 19 (AB 1) antibodies according to the manufacturers instructions (Dynal, Oslo, Norway).

Proliferative response of peripheral blood T-cells

To determine the proliferative response of circulating T cells, two distinct protocols were applied. In one set of experiments, T-cell proliferation was stimulated by cross-linking of CD3 and CD28 receptors. T-cell receptor stimulation in the absence of a costimulus fails to induce T-cell activation [1, 4, 14, 20, 23, 28]. CD 28 costimulation in combination with anti-CD3 has been shown to induce optimal activation of T lymphocytes [1, 7, 13]. Round-bottom 96-well tissue culture plates (Nunc) were coated with CD3 antibody by incubating 50 ml of a 20 mg/ml solution for 1 h at 37C. To each well, 2 104 enriched T cells were added, together with CD28 antibody and goatanti-mouse immunoglobulin (5 mg/ml each). Alternatively, T-cell proliferation was induced by adding 20 ng/ml of each bacterial superantigen (staphylococcal enterotoxin A [SEA], staphylococcal enterotoxin B [SEB], toxic shock syndrome toxin-1 [TSST-1]) to unfractionated PBMC (2 104/well). The concentrations of superantigen used were found to result in maximal stimulation of T-cell proliferation. T cells were stimulated for 5 days. DNA synthesis was measured by adding 18.5 Kbq/well of [3H]-thymidine (Amersham Buchler, Braunschweig, Germany) 24 hs before termination of the experiment. Time course experiments had revealed that under the experimental conditions used, plateau levels of [3H]-thymidine incorporation were reached after 57 days of stimulation. The samples from triplicate cultures were collected onto glass fiber filters (Bibby Dunn, Asbach, Germany). Radioactivity was determined using a Matrix 96 Beta Counter (Packard Instruments, Frankfurt, Germany).

Immunofluorescence staining and flow cytometry analysis

Expression of cell surface molecules on circulating T cells, monocytes, and granulocytes was analyzed by two- and three-color immunofluorescence staining. Saturating concentrations of monoclonal antibodies were incubated with 100 ml of heparinized blood. For control, isotype-matched mouse immunoglobulins were included in each experiment. Red blood cells were lysed by adding 10 vol of FACS lysing solution (Becton Dickinson). Cells were washed with PBS and fixed in PBS containing 2% paraformaldehyde. Fluorescence was analyzed on an EPICS XL cytometer

Cell cultures and antibodies

Cell cultures were performed in RPMI 1640 medium containing 7% heatinactivated fetal calf serum, 100 mg/ml streptomycin, and 100 U/ml penicillin (Biochrom, Berlin, Germany). Murine monoclonal antibodies used in

1022

Fig. 1. Expression of HLA-DR on monocytes before and after laparoscopic cholecystectomy. The significant decrease on postoperative day 1 remained functionally compensated. After 1 week, preoperative levels were recovered.

Fig. 2. Expression of HLA-DR on monocytes after open cholecystectomy was significantly decreased on the 1st postoperative day and did not return to preoperative levels within 1 week.

(Coulter Corporation, Hialeah, FL, USA). Instrument calibration was performed daily using Calibrite Beads (Becton Dickinson) according to the recommendations of the manufacturer. T cells and monocytes were identified by forward and side-scatter profile and positive staining with antibodies to the TCR ab heterodimer or CD14, respectively. Granulocytes were identified by forward and side-scatter profile.

Statistical analysis
Data were analyzed using the Mann-Whitney U test for paired samples. Results are presented as mean SEM. The level of significance was set at p < 0.05.

Results To evaluate the effect of minimally invasive surgery on host defense mechanisms and compare it with changes after open surgery, circulating CD14+ monocytes were analyzed for expression of MHC class II molecules and the capacity to present antigens. The results in Fig. 1 show that after laparoscopic cholecystectomy, expression of HLA-DR on monocytes significantly decreased on the 1st postoperative day, but it returned to preoperative levels after 1 week. After open cholecystectomy, HLA-DR expression on monocytes was also significantly downregulated on the 1st postoperative day, but expression remained low throughout the whole observation period (Fig. 2). In spite of reduced HLA-DR levels, the capacity of circulating monocytes to present several independent superantigens (SEA, SEB, TSST-1) to T lymphocytes and to stimulate T-cell proliferation was not affected in both groups [data not shown]. Thus, reduced HLA-DR expression after laparoscopic or open cholecystectomy did not affect the antigen-presenting capacity of monocytes. The leukocyte adhesion molecule L-Selectin mediates lymphocyte recirculation and migration of phagocytes to sites of inflammation. In response to cell activation, LSelectin is rapidly lost from the cell surface by proteolytic cleavage. L-Selectin expression on monocytes and granulocytes was not affected by laparoscopic or open cholecystectomy throughout the entire observation period. Similarly, monocyte expression of CD80 was unaltered. These data

Fig. 3. Expression of CD25 on T lymphocytes increased in the early postoperative period after open cholecystectomy and returned to preoperative levels within 1 week.

therefore indicate that adhesion molecule expression and activation state of circulating phagocytes were not affected by laparoscopic or open cholecystectomy. On T lymphocytes, cell activation is associated with an increased expression of MHC class II molecules, CD25, and adhesion receptors such as ICAM-1, whereas L-Selectin is downregulated. Three-color immunofluorescence analysis demonstrated that cell surface expression of HLA-DR, ICAM-1, and L-Selectin on circulating CD4+ (helper) T lymphocytes was not altered by laparoscopic or open cholecystectomy. In contrast, CD25 increased significantly on postoperative day 1 after open cholecystectomy (p 0.02) and returned to preoperative levels after 1 week (Fig. 3). After the laparoscopic operation, CD25 did not change significantly (Fig. 4). To define the effects of laparoscopic and open cholecystectomy on T-cell function in more detail, we examined the proliferative response after ligation of the CD3 component of the T-cell antigen receptor and the CD28 costimulatory receptor. For the experiments, PBMC were depleted of monocytes and B cells to exclude the possibility that the patients APC might affect the results. T-cell proliferation after CD3CD28 ligation remained unchanged in both groups [data not shown].

1023

Fig. 4. Expression of CD25 on T lymphocytes did not show significant changes after laparoscopic cholecystectomy.

Discussion The capacity of circulating monocytes to present antigens and thereby stimulate T cells requires expression of MHCII molecules such as HLA-DR on the cell surface [19, 32]. Downregulation of HLA-DR on monocytes by major surgery and trauma has been described [11, 12, 19] and seems to be associated with an increase of postoperative infectious complications [2, 5, 19, 32]. Recently, we have shown that after major conventional surgery, HLA-DR on monocytes is reduced by >60% of preoperative values and is still significantly decreased on postoperative days 810 [18]. Kloosterman et al. [21] observed a significant reduction of HLADR on monocytes after conventional but not after laparoscopic cholecystectomy. In our study, HLA-DR on monocytes was significantly reduced on the 1st postoperative day after laparoscopic cholecystectomy, but it returned to normal within 67 days (Fig. 1). After open cholecystectomy, downregulation of HLA-DR expression on monocytes was also significant on postoperative day 1 (p 0.0002) and remained significant after 1 week (p 0.002) (Fig. 2). To assess the functional significance of suppressed HLA-DR expression on monocytes, we evaluated the antigen-presenting capacity of monocytes and their ability to stimulate T cells. For this purpose, T cells were stimulated with superantigens (SEA, SEB, TSST-1) presented by the patients monocytes. Superantigens are recognized by T cells expressing the appropriate V segment of the T-cell receptor b subunit (statistically, an average of 5% of all T cells in humans) independently of the clonal specifity of the antigen receptor. Moreover, superantigens are not endocytosed or proteolytically processed; instead, they bind directly to MHC class II molecules, therefore allowing for direct analysis of T-cell costimulation and HLA-DR loading [17]. Our results showed that the proliferative response of T cells stimulated by superantigens was not impaired after either laparoscopic or open cholecystectomy. We therefore conclude that the loss of HLA-DR, in spite of being significant, does not affect the antigen-presenting capacity of monocytes. A possible explanation for this finding can be found in studies that have shown that loading of as few as 0.1% of MHC II molecules with antigenic peptide or superantigen is sufficient to stimulate T-cell proliferation [3, 9, 16].

There is evidence that the proliferative response of T lymphocytes stimulated by mitogen is reduced after major surgery [11]. Similarly, T-cell proliferation induced by ligation of the T-cell receptor/CD3 complex and the CD28 receptor is severely impaired after major conventional surgery [18]. Ligation of CD28 induces an important costimulatory signal for T-cell proliferation [4]. In our study, we analyzed T-cell proliferation after antibody-mediated crosslinking of CD3 and CD28 receptors without the presence of APC. In contrast to the effects observed after major surgery, we found that T-cell proliferation was not significantly impaired after either open or laparoscopic cholecystectomy. T-cell stimulation is associated with an increased expression of surface receptors, including HLA-DR and CD25, whereas L-Selectin expression on T cells decreases with activation. Early after major conventional surgery, the number of circulating T cells expressing HLA-DR was found to be increased, suggesting that activation of T lymphocytes occurs as a consequence of major surgery [18, 32]. Our results revealed that expression of HLA-DR, ICAM-1, and L-Selectin on T cells remained unchanged by laparoscopic and open cholecystectomy, indicating that CD4+ T cells were not fully activated. Also, L-Selectin and CD80 expression on monocytes, as well as L-Selectin on granulocytes, was not affected by either procedure. CD25 expression increased significantly in the early postoperative period after open cholecystectomy, but it remained unchanged after the laparoscopic operation. These findings indicate that T-cell activation did not occur after the laparoscopic procedure, but several markers of T-cell activation were detected following major surgery. In addition, selective upregulation of CD25 following open cholecystectomy may reflect partial T-cell activation. In summary, we observed a significantly reduced HLADR expression on monocytes after both laparoscopic and open cholecystectomy. While the downregulation still persisted 1 week after the open operation, HLA-DR expression returned to preoperative levels within 1 week after the laparoscopic procedure. However, in both groups, the antigenpresenting capacity of monocytes remained unimpaired by the HLA-DR depression. After open cholecystectomy, a significant upregulation of CD25 expression on T cells occurred, reflecting partial activation of these cells. This effect could not be observed after the minimally invasive procedure. We therefore conclude that even though the clinical relevance of the immunologic changes after open cholecystectomy remains to be determined, the laparoscopic operation seems to interfere less with immune defense mechanisms than the open procedure. Whether more severe immunologic compromise occurs as a result of more extended laparoscopic procedures has to be determined in future studies. References
1. Allison JP (1994) CD28-B7 interactions in T-cell activation. Curr Opin Immunol 6: 414419 2. Appel SH, Wellhausen ST, Montgomery R, DeWeese RC, Polk HC (1989) Experimental and clinical significance of endotoxin dependent HLA-DR expression on monocytes. J Surg Res 5: 4749 3. Bhardwaj N, Young JW, Nisanian AJ, Baggers J, Steinman RM (1993) Small amounts of superantigen, when presented on dendritic cells, are sufficient to initiate T-cell responses. J Exp Med 178: 633642

1024 4. Bluestone JA (1995) New perspectives of CD28-B7mediated T-cell costimulation. Immunity 2: 555559 5. Cheadle WG, Hershman MJ, Wellhausen SR, Polk HC (1991) HLADR expression on peripheral blood monocytes correlates with surgical infection. Am J Surg 161: 639645 6. Cho JM, LaPorta AJ, Clark JR, Schofield MJ, Hammond SL, Mallory PL (1994) Response of serum cytokines in patients undergoing laparoscopic cholecystectomy. Surg Endosc 8: 13801384 7. Clements JL, Winslow G, Donahue C, Cooper SM, Allison JP, Budd RC (1993) Costimulation via CD28 induces activation of a refractory subset of MRL LPR/LPR lymphocytes. Int Immunol 5: 14511460 8. Cruickshank AM, Fraser WD, Burns HJG, Van Damme J, Shenkin A (1990) Response in serum interleukin 6 in patients undergoing elective surgery of varying intensity. Clin Sci 79: 161165 9. Demotz S, Grey HM, Sette A (1990) The minimal number of class II MHC-antigen complexes needed for T-cell activation. Science 249: 10281030 10. Eilber FR, Morton DL (1970) Impaired immunologic reactivity and recurrence following cancer surgery. Cancer 25: 362367 11. Faist E, Kupper TS, Baker CC, Chaudry ICH, Dwyer J, Baue AE (1986) Depression of cellular immunity after major surgery: its association with posttraumatic complications and its reversal with immunomodulation. Arch Surg 121: 10001005 12. Faist E, Mewes A, Strasser T, Walz A, Alkan S, Baker C, Ertel W, Heberer G (1988) Alteration of monocyte function following major injury. Arch Surg 123: 287292 13. Giese T, Allison JP, Davidson WF (1993) Functionally anergic 1pr and gld B220+ TCR a/b DN cells proliferate and secrete IL-2 in response to co-stimulation in vitro with antibodies to CD28 and TCR. J Immunol 151: 597609 14. Guinan EC, Gribben JG, Boussiotis VA, Freeman GJ, Nadler LM (1994) Pivotal role of the B7: CD28 pathway in transplantation tolerance and tumour immunity. Blood 84: 32613282 15. Goodale RL, Beebe DS, McNevin MP, Boyle M, Letourneau JG, Abrams JH, Cerra FB (1993) Hemodynamic, respiratory and metabolic effects of laparoscopic cholecystectomy. Am J Surg 166: 533537 16. Harding CV, Unanue ER (1990) Quantitation of antigen-presenting cell MHC class II/peptide complexes necessary for T-cell stimulation. Nature 346: 574576 17. Herman A, Kappler JM, Marrack P, Pullen AM (1991) Superantigens: mechanisms of T-cell stimulation and role in immune responses. Ann Rev Immunol 9: 745772 18. Hensler T, Hecker H, Heeg K, Heidecke CD, Bartels H, Barthlen W, Wagner H, Siewert JR, Holzmann B (1997) Distinct mechanisms of immunosuppression as a consequence of major surgery. Infect Immunol 65: 22832291 Hershman MJ, Cheadle WG, Wellhausen SR, Davidson PF, Polk HC (1990) Monocyte HLA-DR antigen expression characterizes clinical outcome in the trauma patient. Br J Surg 77: 204207 June CH, Bluestone JA, Nadler LM, Thompson CB (1994) The B7 and CD28 receptor families. Immunol Today 15: 321331 Kloosterman T, von Blomberg ME, Borgstein P, Cuesta MA, Scheper RJ, Meijer S (1994) Unimpaired immune function after laparoscopic cholecystectomy. Surgery 115: 424428 Kruitwagen RF, Swinkels BM, Keyser GG, Doesburg WH, Schijf CCH (1996) Incidence and effect on survival of abdominal wall metastasis at trocar or puncture sites following laparoscopy or paracentesis in women with ovarian cancer. Gynecol Oncol 60: 233237 Linsley PS, Ledbetter JA (1993) The role of the CD28 receptor during T cell responses to antigen. Annu Rev Immunol 11: 191212 Maruszynski M, Pojda Z (1995) Interleukin 6 levels in the monitoring of surgical trauma: a comparison of serum IL6 concentrations in patients treated by cholecystectomy via laparotomy or laparoscopy. Surg Endosc 9: 882885 Park SK, Brody JL, Wallace HA, Blakemore WS (1971) Immunsupressive effect of surgery. Lancet 1: 5355 Patarroyo M, Prieto J, Rincon J, Timonen T, Lundberg C, Lindbom L, Asjo B, Gahmberg CG (1990) Leukocyte cell adhesion: a molecular process fundamental in leukocyte physiology. Immunol Rev 114: 67 108 Roumen RM, van Meurs PA, Kuypers HH, Kraak WA, Sauerwein RW (1992) Serum IL6 and CRP responses in patients after laparoscopic or conventional cholecystectomy. Eur J Surg 158: 541544 Schwartz RH (1990) A cell culture model for T lymphocyte clonal anergy. Science 248: 13491356 Springer TA (1990) Adhesion receptors of the immune system. Nature 356: 425434 Springer TA (1994) Traffic signals for lymphocyte recirculation and leukocyte migration: the multiple step paradigm. Cell 76: 301314 Tedder TF, Steeber DA, Pizcueta P (1995) L-selectindeficient mice have impaired leukocyte recruitment into inflammatory sites. J Exp Med 181: 22592264 Wakefield CH, Carey PD, Foulds S, Monson RT, Guillou PJ (1993) Changes in major histocompatibility complex class II expression on monocytes and T-cells of patients developing infection after surgery. Br J Surg 80: 205209

19. 20. 21. 22.

23. 24.

25. 26.

27. 28. 29. 30. 31. 32.

Surg Endosc (1998) 12: 10251030

Springer-Verlag New York Inc. 1998

Cardiopulmonary responses to experimental venous carbon dioxide embolism

K. L. Mayer, H. S. Ho, K. A. Mathiesen, B. M. Wolfe
Department of Surgery, University of California Davis Medical Center, 4301 X Street, Room 2310, Sacramento, CA 95817-2214, USA Received: 26 March 1996/Accepted: 29 October 1997

Abstract Background: Although the low-flow CO2 insufflation rate used to initiate pneumoperitoneum may reduce the severity of potential venous embolism, its safety is not established. Methods: Anesthetized pigs were ventilated with room air at a fixed minute ventilation. After 1 h of baseline, they were intravenously infused with CO2 at the rate of 0.3, 0.75, or 1.2 ml/kg/min for 2 h (n 5 for each group), followed by 1 h of recovery. Results: All animals experienced pulmonary hypertension, depressed stroke volume, hypoxemia, hypercarbia, and acidemia during intravenous CO2 infusion. They had systemic hypertension at the low rate and hypotension at the highest rate of infusion. End-tidal CO2 levels briefly decreased, then increased in all cases. In the highest rate group, three of the five animals (60%) died at 50, 65, and 100 min of infusion. These three animals had severe hypotension and hypoxemia, with visible coronary gas embolism. There was no patent foramen ovale at necropsy in any animals. Conclusions: The low-flow insufflation rate exceeds the fatal rate of continuous intravenous CO2 infusion. End-tidal CO2 levels were increased in venous CO2 embolism, not decreased as seen in venous air embolism. Severe hypoxemia and hypotension are predictors of potentially fatal cases. Key words: Laparoscopy Venous CO2 embolism Hypoxemia Hypotension

Establishing the initial pneumoperitoneum with the Veress needle is a risky step during laparoscopic surgery. Besides

Presented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES) and the 5th World Congress of Endoscopic Surgery at Philadelphia, Pennsylvania, 1317 March 1996 Correspondence to: H. S. Ho

potential visceral injury, an unintentional vascular cannulation may result in massive hemorrhage or fatal gas embolism. Venous gas embolism is a rare complication of gaseous pneumoperitoneum, with an incidence of 0.002% to 0.6% in the gynecological patients [13, 22, 23]. Although CO2 is the current gas of choice for laparoscopy, partly because of its theoretical low risk of gas embolism, fatal incidents involving this gas have occurred [2, 7, 16, 23]. In a recent case series from France, two of the seven patients who had venous CO2 embolism died [7]. More importantly, six of these patients developed problems within minutes of creation of pneumoperitoneum using the low-flow rate of CO2 insufflation. There is a potentially false sense of security that using the low-flow rate for insufflation may reduce the risk or the severity of venous CO2 embolism. Furthermore, the rare incidence of venous CO2 embolism may contribute to its misdiagnosis or delay in diagnosis unless the index of suspicion is high. Clinically, the low-flow rate set for CO2 insufflation into the peritoneal cavity is 1.5 l/min or 25 ml/kg/min for an average 60-kg person. If unrecognized venous cannulation occurs during placement of the Veress needle, such a bolus injection of CO2 into the venous system can be fatal. In dogs, the LD-50 dose for a CO2 venous bolus injection is 25.0 1.7 ml/kg [9]. Therefore, it is important to identify simple, noninvasive but readily available clinical parameters that may signal venous CO2 embolism during the initiation of pneumoperitoneum. In this study, we compared the cardiopulmonary responses to three rates of intravenous infusion of CO2 in pigs. We also evaluated the changes in end-tidal CO2, pulse oximetry, arterial blood gases, and vital signs for their ability to predict clinically significant venous CO2 embolism. We aim to address the following issues: (1) Are cardiopulmonary responses to venous CO2 embolism dose dependent? (2) Can clinically significant venous CO2 embolism be detected or predicted with noninvasive, routine monitoring methods? and (3) Is the low-flow insufflation rate for initiation of pneumoperitoneum truly safe?

1026

Materials and methods Animal preparation

For this study, 15 pigs were used with protocols for care and use approved by the Animal Use and Care Administrative Advisory Committee of the University of California at Davis School of Medicine. Animals were fasted but had free access to water for 24 h before experiments. Pigs were premedicated with an intramuscular injection of atropine sulfate (0.044 mg/ kg) and ketamine hydrochloride/zolazepam hydrochloride (4 mg/kg). They were then endotracheally intubated, placed in a supine position, and ventilated with room air using a respiration pump with a nonrebreathing circuitry (Model 607, Harvard Apparatus Co., Inc., Millis, MA). General anesthesia was maintained with continuous intravenous (IV) infusion of pentobarbital sodium (0.10.4 mg/kg/min) for the rest of the experiment. Next, catheters were placed for cardiopulmonary monitoring. Silastic venous and arterial lines (0.062 inch ID) were inserted via a left femoral cutdown. The arterial line was advanced into the abdominal aorta for continuous measurement of the mean arterial pressure and arterial blood gas analyses, and the venous line was advanced into the inferior vena cava for infusion of gas. A pulmonary artery thermodilution catheter (Edwards Critical Care Division, Irvine, CA) was placed via a right internal jugular vein cut-down and connected to a multichannel monitor and recorder (7754 System, Hewlett Packard, Palo Alto, CA). All pressure transducers were positioned at the level of the right atrium and calibrated with a mercury manometer.

Experimental protocol
After surgical preparation, all animals were observed until hemodynamics were stable. Minute ventilation was adjusted to obtain normal arterial blood gases, and was then fixed throughout the remainder of the experiment. The animals temperatures were maintained at a range of 36.6 to 37.9C with a heating pad. Once stability was established, baseline data were collected at 15-min intervals for 1 h. The animals were divided into three groups to receive a 2-h continuous IV infusion of medical-grade CO2 at one of three rates (0.3 ml/kg/min, 0.75 ml/kg/min, or 1.2 ml/kg/min) via the femoral venous catheter. The infusion rate was controlled by a regulator connected to a 120-mm manostat flowmeter (Fisher Scientific, Santa Clara, CA). Hemodynamic and ET-CO2 measurements were obtained every 5 min during the infusion period. Temperature-corrected arterial blood gas analysis was performed every 15 min. At the end of the infusion, the animals were observed for an additional 1 h of recovery with data collected every 15 min. At necropsy, the coronary vessels were inspected for visible gas embolism, and cardiectomies were performed to look for any patent foramen ovale. Fig. 1. Pulmonary hypertension developed within a few minutes of IV CO2 infusion. There was a within-group difference between baseline values and all data points during IV CO2 infusion in all three groups: up to 75 min for the 0.3-ml/kg/min group, 120 min for the 0.75-ml/kg/min and 45 min for the 1.2 ml/kg/min groups, respectively. The changes in end-tidal CO2 level were similar in all three rates of infusion. ET-CO2 briefly decreased within minutes of infusion, then steadily increased significantly above baseline.

of infusion. At necropsy, all three dead pigs had visible evidence of coronary arterial gas embolism. None of the 15 pigs had evidence of a patent foramen ovale.

Evidence of pulmonary arterial embolism Data analysis

All data are reported as mean SEM. Analysis of variance (ANOVA) for repeated measurements was performed to examine changes within groups and between groups. A p value less than 0.05 was considered significant. When results were significant for within-group changes, Dunnetts multiple range post hoc test was performed to identify the times at which the values differed from baseline. When there were significant differences among the three groups, post hoc contrast tests were performed to identify the differences. For the high infusion rate group (1.2 ml/kg/min), the data were analyzed and presented for only the first 45 min of infusion due to the death of three pigs. Data from these three animals were analyzed as a separate group to compare with the data from the two animals in the same group that survived the infusion period. All analyses were done using computer software (Instat Instant Biostatistics, GraphPad Software, San Diego, CA and SuperANOVA, Abacus Concepts, Inc., Berkeley, CA).

Results All pigs infused with CO2 at the rate of 0.30 ml/kg/min or 0.75 ml/kg/min survived. Three of the five pigs infused with CO2 at the rate of 1.2 ml/kg/min died at 50, 65, and 100 min

Pulmonary hypertension developed immediately in all animals during IV CO2 infusion (Fig. 1). Within 10 min of infusion, the mean pulmonary arterial pressure (MPAP) increased almost threefold in all three groups, from 12.0 0.3 mmHg to 30.8 1.5 mmHg in the 0.3-ml/kg/min group (p < 0.01), from 11.4 1.3 mmHg to 30.4 3.7 mmHg in the 0.75-ml/kg/min group (p < 0.01), and from 13.2 1.2 mmHg to 37.0 3.7 mmHg in the 1.2-ml/kg/min group (p < 0.01). This pulmonary hypertension induced by IV infusion of CO2 remained for more than 60 min into the 2-h infusion period. In the high infusion rate group, MPAP began a steadily and sharply declining course after the initial peak in the animals that died, whereas MPAP values were similar to those of the other groups in the two animals that survived. The rate of MPAP normalization was similar in all surviving animals after cessation of IV CO2 infusion. Within the first few minutes of infusion, all animals experienced an approximate 10% reduction in end-tidal carbon dioxide levels (ET-CO2). However, this reduction in ET-CO2 was immediately followed by a increase toward

1027

Fig. 2. Hypercarbia also developed immediately on IV CO2 infusion, with similar peaks in all three groups. There was a within-group difference in all groups between baseline and the rest of the experiments. The changes in arterial pH reflected the developing hypercarbia (* denotes p < 0.05 for difference between groups, ANOVA for repeated measurements).

baseline ET-CO2 levels as the CO2 infusion continued, so that by 15 min the animals actually had an elevated ET-CO2 level. By the end of the 1-h recovery period, ET-CO2 levels remained markedly above baseline values in all groups. The two pigs that survived the highest rate of IV CO2 infusion had levels of ET-CO2 as high as 60 mmHg. The three pigs that died had erratic patterns after the initial changes in ET-CO2 levels: One pig had a rapid increase, another a modest increase, and the last a steady decline.

Fig. 3. As the rate of IV CO2 infusion was increased from 0.30 to 1.20 ml/kg/min, the responses in MAP changed from hypertension to hypotension (* denotes p < 0.05 for difference between groups, ANOVA for repeated measurements). The two surviving pigs in the high-dose group remained hypertensive, in contrast to hypotension in the dying pigs. Hypoxemia is severe and independent of the IV CO2 infusion rates. There was a within-group difference in all three groups between oxygenation at baseline and that following IV CO2 infusion. These differences persisted during the recovery period.

Acidbase changes with venous carbon dioxide embolism Hypercarbia and acidemia developed immediately on IV infusion of CO2. Furthermore, there was a graduated dose response to the increasing rates of infusion (Fig. 2). The PaCO2 levels rose rapidly during the first 45 min of CO2 infusion, with the 0.3-ml/kg/min group level increasing from 39.8 1.9 mmHg to 54.2 1.2 mmHg (p < 0.01), the 0.75-ml/kg/min group level increasing from 36.7 1.1 mmHg to 59.9 0.5 mmHg (p < 0.01), and the 1.2-ml/kg/ min group level increasing from 38.1 1.4 mmHg to 72.9 2.7 mmHg (p < 0.01). At 45 min of infusion, the PaCO2 reached a plateau and remained elevated in all surviving pigs. During the recovery, hypercarbia improved in all pigs that survived, yet it did not return to baseline even after 1 h. The peak in PaCO2 was similar in all pigs in the highinfusion-rate group regardless of whether they survived.

The severity of acidemia reflected the rise in arterial CO2 tension in all three groups. Severe hypoxemia developed immediately after initiation of IV infusion of CO2 in all three groups (Fig. 3). The partial pressure of arterial oxygen tension (PaO2) decreased to less than 50 mmHg within the first 15 min of IV CO2 infusion. All animals remained hypoxemic until death or until the IV CO2 infusion was stopped. The pigs that died in the high-dose group experienced no worse hypoxemia than those that survived. During the recovery period, the surviving pigs had some improvement in oxygenation, but still remained hypoxemic at approximately 55 mmHg after 1 h. Hemodynamic responses to venous carbon dioxide embolism There was no difference in hemodynamics within or among the three groups during the baseline period. All animals experienced significant cardiopulmonary changes in response to IV CO2 infusion (see Fig. 3). Animals infused with CO2 at the rate of 0.3 ml/kg/min experienced moderate systemic hypertension, with the mean arterial pressure (MAP) rising from 118.6 5.6 mmHg to 154.8 9.7 mmHg (p < 0.01), which persisted through the remainder of the experiment. Systemic hypertension did not develop in animals receiving IV CO2 at the rate of 0.75 ml/kg/min: The

1028 Table 1. Changes in hemodynamics and oxygenation during IV infusion of CO2 Infusion Baseline t0 HR (beats/min) 0.30 ml/kg/min 0.75 ml/kg/min 1.20 ml/kg/min CI (mL/kg/min) 0.30 ml/kg/min 0.75 ml/kg/min 1.20 ml/kg/min CVP (cmH2O) 0.30 ml/kg/min 0.75 ml/kg/min 1.20 ml/kg/min O2 Saturation (%) 0.30 ml/kg/min 0.75 ml/kg/min 1.20 ml/kg/min 5 min 10 min 15 min 30 min 60 min 90 min 120 min Recovery 60 min

104 6 107 7 106 7 106 13 110 4 105 7 1.4 0.9 0.8 0.6 0.4 0.5 96.3 0.4 96.9 0.2 96.1 0.6

109 6 113 8 121 12 98 15 95 6 86 7 3.4 0.7 2.0 0.8 3.4 0.8 82.0 5.4 85.6 1.0 65.2 2.7

114 6 114 10 131 12 107 15 100 5 95 7 3.4 0.9 2.2 0.8 4.0 0.8 75.9 6.5 81.1 4.3 49.7 6.1

121 6 119 12 140 15 117 15 98 6 99 10 3.4 0.9 1.8 0.9 3.6 1.4 71.2 5.4 76.7 5.2 61.1 5.5

134 5 122 12 137 16 119 13 101 7 109 8 2.8 1.0 1.8 1.0 3.0 1.4 64.9 6.0 65.1 1.5 47.9 5.1

143 5 130 11 131 21 132 18 115 4 92 20 1.2 0.9 2.2 1.0 3.3 1.9 70.5 2.2 64.5 2.3 50.5 9.0

151 6 134 12 135 31 131 21 109 9 105 46 0.8 0.7 2.0 0.7 2.0 2.5 71.9 2.3 62.5 2.9 54.8 3.7

153 9 130 17 163 41 151 25 100 17 111 6 0.4 1.1 2.0 1.3 1.0 1.0 72.9 2.9 58.0 6.0 50.7 9.7

149 16 116 12 148 53 157 42 88 12 119 15 0.2 1.5 1.2 1.1 0.0 0.0 83.3 3.9 88.9 2.4 87.4 4.8

MAP remained essentially unchanged from 120.4 8.4 mmHg at baseline to only as high as 126.8 16.6 mmHg during the first hour of infusion. The animals receiving the high infusion rate (1.2 ml/kg/min) quickly developed systemic hypotension, with MAP decreasing from 111 10.5 mmHg to 74.6 26.4 mmHg (p < 0.05). The two pigs that survived this high infusion rate were hypertensive to a MAP of 180 mmHg, unlike the pigs that died, which showed a steady decline in blood pressure shortly after initiation of CO2 infusion. All animals experienced tachycardia with IV CO2 infusion. However, it was transient in the high-infusion-rate group: Only one of the surviving pigs remained tachycardic, the other had a heart rate that steadily returned to baseline. The difference between the responses to the 0.3 ml/kg/min and 0.75 ml/kg/min infusion rates became evident after 90 min of infusion when the slopes of the changes in heart rate began to diverge (Table 1). The stroke volume immediately decreased in all groups on creation of venous CO2 embolism and remained depressed until the end of the experiments. Although the stroke volume showed similar decreases among the three groups, the cardiac index reflected the trend of the heart rate (see Table 1). The cardiac index was elevated throughout the experiment in the low-dose group, reflecting the observed tachycardia and modest depression in stroke volume, and it remained essentially unchanged in the other groups. All three groups had an immediate slight increase in central venous pressure (CVP). The two pigs that received high doses of IV CO2 but survived had a return to baseline of their CVP after cessation of CO2 infusion, whereas the three animals that died had an elevated CVP that persisted until death.

severe and sometimes fatal venous CO2 embolism have occurred despite its high solubility in plasma [2, 7, 16, 23]. Venous gas embolism is an inherent risk of laparoscopic surgery, especially during the initiation of pneumoperitoneum or the dissection of highly vascularized structures. If a large amount of CO2 enters the vasculature in a gaseous form, a detrimental or fatal cardiopulmonary collapse may result.

Physiology of venous carbon dioxide embolism When injected into the vasculature, CO2 displaces rather than mixes with the blood because the gas initially holds an interface with blood [14]. Venous CO2 embolism may induce physiologic changes by the gas lock phenomenon in which the CO2 bubbles impair blood flow. When this gas lock is located in the pulmonary arterial tree, obstruction of the right ventricular outflow may result. Bolus injections of CO2 into the abdominal aorta resulted in the gas being seen in the inferior vena cava 10 sec later, and imaging studies of the heart and lungs revealed that the gas was cleared in one pass. Occasionally, though, the gas-lock phenomenon persists, and retained or trapped CO2 is detected in the abdominal aortic aneurysm as long as 24 h after the initial injection [14]. Such gas trapping, when taking place in the right ventricle or the pulmonary circulation, may diminish the blood flow into the left ventricle, resulting in systemic hypotension [1, 20]. In a nearly fatal case of venous gas embolism during hysteroscopy, open heart surgery revealed a gas lock phenomenon in the right ventricle obstructing the superior vena cava [8]. When more than 65% of the pulmonary vascular bed is obstructed by the CO2 bubbles, dead space increases, and pulmonary blood flow from the right to the left side of the heart decreases, leading to hypotension [20]. The CO2 bubbles also obstruct gas exchange at the alveolararterial interface, which leads to physiologic shunting, the degree depending on the severity of the pulmonary embolism [19]. Hypotension and physiologic shunting leads to hypoxemia. In the study by Khan et al. [15], dynamic pulmonary com-

Discussion One advantage of CO2, the current gas of choice for pneumoperitoneum, is its high solubility in plasma [17]. The high solubility ensures rapid absorption of CO2 by the circulation and its elimination by the lungs. However, cases of

1029

pliance and total lung airflow resistance remained unchanged with a 30-ml bolus injection of pure CO2 in dogs. Therefore, bronchoconstriction alone cannot account for the observed hypoxemia. In addition, hypercarbia and accompanying acidemia shift the hemoglobin-dissociation curve to the right, unloading oxygen at the tissue level and further driving down the arterial tension of oxygen. In our animal model of venous CO2 embolism, hypoxemia and hypercarbia are observed responses that are independent of dosage. There was no attempt to correct hypoxemia during this experiment because we were studying the consequences of undetected venous gas embolism, not the therapeutic maneuvers to correct them. Furthermore, hypoxemia does not appear to be the sole cause of demise in our experimental animals because all the pigs experienced hypoxemia as low as 40 mmHg (see Fig. 2). We found that different rates of IV CO2 infusion have different effects on arterial blood pressure, perhaps depending on the interaction between the pharmacologic effects of the hypercarbia and the ventricular outflow obstruction from the CO2 bubbles. At a low infusion rate (0.3 ml/kg/ min), systemic hypertension occurred immediately and persisted, indicating a direct catecholamine response to hypercarbia. In contrast, animals with the high CO2 infusion rate (1.2 ml/kg/min) were markedly hypotensive, suggesting that the cardiodepressant effects of CO2 on the myocardium and the mechanical obstruction caused by the CO2 bubbles predominated over the CO2 stimulation of the autonomic nervous system. The absence of any CVP decrease in our animals suggests right ventricular outflow obstruction rather than hypovolemia as a cause for the observed hypotension. In our experiment, when the rate of CO2 infusion was between these two extremes (i.e., at 0.75 ml/kg/min), the MAP remained unchanged, further supporting the concept of competing catecholamine stimulant effects versus cardiodepressant effects of hypercarbia [12]. The marked tachycardia was possibly due to the catecholamine response to hypercarbia. Although the stroke volume was reduced, tachycardia resulted in a slight increase in cardiac output. Butler and Hill [3] found that under normal conditions, the lung is an excellent filter for bubbles, and the cutoff diameter is less than 22 m. However, bubbles did escape entrapment when the lungs were overloaded with gas from a bolus of approximately 30 ml. Graff et al. [11] also found that precipitous hypotension occurred with an IV CO2 dose of approximately 30 ml/kg. Using Doppler imaging to detect breakthrough of arterial bubbles across the pulmonary circulation, Vik et al. [20] showed that the infusion rate required for breakthrough of air bubbles to occur was only 0.1 ml/kg/min. Systemic hypotension occurred immediately after air bubbles were introduced into the left heart, which may indicate a reduced flow into the left heart, a release of vasoactive substances, or a reflex mechanism responsible for the hypotension [4, 20]. Pigs, as reported, have a 33% incidence of patent foramen ovale, and when pulmonary air embolism was induced, pulmonary hypertension was more severe and systemic hypotension more pronounced in pigs with patent foramen ovale [21]. At necropsy, none of our pigs had patent foramen ovale, including the three pigs that had coronary gas embolism. Mortality associated with venous CO2 embolism is reportedly due to trapped gas in the coronary vessels [10].

Necropsy findings in our study supported this observation: All three animals that died during the infusion period had visible coronary gas embolism. The three possible mechanisms of death are gas entering and occluding the coronary circulation, trapped gas in the ventricle itself resulting in failure of the cardiac pumping action, and cerebral arterial embolism. Myocardial ischemia was not documented in this study, and it remains unclear whether the deaths were due specifically to myocardial ischemia or to pulmonary vascular obstruction and subsequent cardiogenic shock.

Clinical implications and methods of detecting venous CO2 embolism Early diagnosis of venous gas embolism is difficult, and the only conclusive evidence of such an event is aspiration of gas from the vasculature or the heart, which is not always practical. Although cardiac dysrhythmias are classic symptoms of venous gas embolism, it would be difficult to differentiate cardiac arrhythmias induced by venous embolism from those caused by hypercarbia from CO2 pneumoperitoneum. Scott and Julian [18] reported a 17% incidence of arrhythmias in 100 consecutive patients who received CO2 pneumoperitoneum, compared with a 4.4% incidence in 45 consecutive patients who received N2 pneumoperitoneum. Durant et al. [9] reported that the ECG changes in pulmonary air embolism were T-wave inversion, marked depressions of ST segments in leads II and III, and some degrees of AV block and nodal rhythm. In the recent French report, ECG signs of right heart failure were present in only one of the two fatal cases, and a right bundle branch block was observed in a surviving patient [7]. Therefore, it appears that ECG changes are neither specific nor sensitive enough to be useful in detecting early venous CO2 embolism. Elevated pulmonary arterial pressure (MPAP) was found to be a sensitive indicator of venous gas embolism [6], but this method requires pulmonary arterial catheterization. We noted an immediate three fold increase above baseline in MPAP on IV infusion of CO2, particularly in the high-infusion-rate group. The rate of increase in MPAP was closely related to the rate of infusion, as previously reported by Verstappen et al. [19]. The MPAP reached a plateau thereafter, possibly indicating that a balance between the rate of infusion and the rate of elimination into the alveolar space had been reached, as proposed by Vik et al. [20]. The reliability of the ET-CO2 levels in detecting gas embolism has been in question [5, 6]. Air embolism caused a momentary increase followed by a progressive decrease in ET-CO2 [6], and inert gas embolism also produced a progressive decrease in ET-CO2 levels [17]. For venous CO2 embolism, we observed the opposite effect, because the pigs also excreted the CO2 being infused. Furthermore, in the three pigs that died, there was no consistent pattern of ETCO2 that would have been helpful in detecting fatal cases of venous CO2 embolism. Severe hypercarbia may alert the anesthesiologist to the possibility of venous CO2 embolism. The surviving pigs had elevations of PaCO2 similar to those of the pigs that died. Thus it appears that severe hypercarbia and its associated acidemia were not the sole cause of the pigs demise. Because minute ventilation was fixed throughout the CO2 in-

1030

fusion period, our animals did not have the capability to compensate for the metabolic changes induced by the gas embolism. Part of the infused CO2 not eliminated by the lungs may be dissolved in the plasma and deposited in the muscular skeleton. This amount of CO2 must be excreted long after CO2 infusion, and it may account for the prolonged hemodynamic effects observed during the recovery period. Immediate hypoxemia may be the result of venous gas embolism, but other potential causes must be excluded first: endotracheal tube malposition, pneumothorax, and hypovolemia. It is most likely a result of a mismatched ventilation perfusion ratio distribution, or physiologic shunting due to the CO2 bubbles obstructing the alveolararterial interface [19]. In our study, this was the most consistent sign of clinically significant venous CO2 embolism, and it was easily detected with pulse oximetry readings. In general, moderate hypertension is expected with CO2 pneumoperitoneum [12]. Hypotension, when associated with severe hypoxemia, was a predictor of potentially fatal venous CO2 embolism. In summary, the low-flow insufflation rate of CO2 is potentially lethal if the Veress needle enters the venous system. Our data showed 60% mortality for a continuous IV CO2 infusion rate of 1.2 ml/kg/min, which is equivalent to a rate of 72 ml/min for a 60-kg person. That volume is only 5% of the volume of CO2 gas that may be infused in 1 min at the low-flow rate into a vein unintentionally cannulated by the Veress needle. Therefore, any patient experiencing sudden or profound oxygen desaturation during low-flow CO2 insufflation must be suspected of having venous CO2 gas embolism. Pneumoperitoneum must be immediately stopped, and the patient must be managed as if venous CO2 embolism is present. Hypotension and hypoxemia are ominous signs, suggesting a potentially fatal case of gas embolism. End-tidal CO2 level increases in response to venous CO2 embolism, in contrast to the expected decrease associated with venous air embolism. Pulmonary hypertension requires pulmonary arterial catheterization for detection. Intraoperative transesophageal echocardiography monitoring is useful in documenting or detecting intracardiac gas, but it is not practical as a routine monitoring device. As a result of this experiment, we propose that frequent and careful monitoring of blood pressure and pulse oximetry readings should be a routine practice during the initiation of CO2 pneumoperitoneum for laparoscopic surgery. Both of these monitoring methods are inexpensive, noninvasive, and readily available, and we found them to be quite useful and practical in detecting potential venous CO2 embolism during the establishment of pneumoperitoneum.

References
1. Adornato DC, Gildenberg PL, Ferrario CM, Smart J, Frost AM (1978) Pathophysiology of intravenous air embolism in dogs. Anesthesiology 49: 120127 2. Beck DH, McQuillan PJ (1994) Fatal carbon dioxide embolism and severe haemorrhage during laparoscopic salpingectomy. Br J Anaesth 72: 243245 3. Butler BD, Hills BA (1979) The lung as a filter for microbubbles. J Appl Physiol 47: 537543 4. Butler BD, Hills BA (1985) Transpulmonary passage of venous air emboli. J Appl Physiol 59: 543547 5. Brampton WJ, Watson RJ (1990) Arterial to end-tidal carbon dioxide tension difference during laparoscopy: magnitude and effect of anaesthetic technique. Anaesthesia 45: 210214 6. Byrick RJ, Kay JC, Mullen JB (1989) Capnography is not as sensitive as pulmonary artery pressure monitoring in detecting marrow microembolism: studies in a canine model. Anesth Analg 68: 94100 7. Cottin V, Delafosse B, Viale JP (1996) Gas embolism during laparoscopy: a report of seven cases in patients with previous abdominal surgical history. Surg Endosc 10: 166169 8. Diakun TA (1991) Carbon dioxide embolism: successful resuscitation with cardiopulmonary bypass. Anesthesiology 74: 11511153 9. Durant TM, Long J, Oppenheimer MJ (1947) Pulmonary (venous) air embolism. Am Heart J 33: 269281 10. Geoghegan T, Lam CR (1953) The mechanism of death from intracardiac air and its reversibility. Ann Surg 138: 351359 11. Graff TD, Arbegast NR, Phillips OC, Harris LC, Frazier TM (1959) Gas embolism: a comparative study of air and carbon dioxide as embolic agents in the systemic venous system. Am J Obstet Gynecol 78: 259265 12. Ho HS, Saunders CJ, Gunther RA, Wolfe BM (1995) Effector of hemodynamics during laparoscopy: CO2 absorption or intra-abdominal pressure? J Surg Res 59: 497503 13. Hynes SR, Marshall RL (1992) Venous gas embolism during gynaecological laparoscopy. Can J Anaesth 39: 748749 14. Kerns SR, Hawkins IF, Sabatelli FW (1995) Current status of carbon dioxide angiography. Radiol Clin North Am 33: 1529 15. Khan MA, Alkalay I, Suetsugu S, Stein M (1972) Acute changes in lung mechanics following pulmonary emboli of various gases in dogs. J Appl Physiol 33: 774777 16. Moskop Jr JR, Lubarsky DA (1994) Carbon dioxide embolism during laparoscopic cholecystectomy. South Med J 87: 414415 17. Roberts MW, Mathiesen KA, Ho HS, Wolfe BM (1997) Cardiopulmonary responses to intravenous infusion of soluble and relatively insoluble gases. Surg Endosc 11(4): 341346 18. Scott DB, Julian DG (1972) Observations on cardiac arrhythmias during laparoscopy. Br Med J 1:411413 19. Verstappen FTJ, Bernards JA, Kreuzer F (1977) Effects of pulmonary gas embolism on circulation and respiration in the dog. II. Effects on respiration. Pflugers Arch 368: 97104 20. Vik A, Brubakk AO, Hennessy TR, Jenssen BM, Ekker M, Slordahl SA (1990) Venous air embolism in swine: transport of gas bubbles through the pulmonary circulation. J Appl Physiol 69: 237244 21. Vik A, Jenssen BM, Brubakk AO (1992) Paradoxical air embolism in pigs with a patent foramen ovale. Undersea Biomed Res 19: 361374 22. Wadhwa RK, McKenzie R, Wadhwa SR, Katz DL, Byers JF (1978) Gas embolism during laparoscopy. Anesthesiology 48: 7476 23. Yacoub OF, Cardona Jr I, Coveler LA, Dodson MG (1982) Carbon dioxide embolism during laparoscopy. Anesthesiology 57: 533535

1102

Port site metastases in laparoscopic surgery

First workshop on experimental laparoscopic surgery, Frankfurt, 1997
H. J. Bonjer,1 C. N. Gutt,2 G. Hubens,3 L. Kra henbu hl,4 S. H. Kim,5 N. D. Bouvy,1 L. N. L. Tseng,1 V. Paolucci,2 6 7 R. Whelan, C. A. Jacobi
1 2

Department Department 3 Department 4 Department 5 Department 6 Department 7 Department

of of of of of of of

Surgery, University Hospital Dijkzigt, Dr. Molewaterplein 40, 3015 GD, Rotterdam, the Netherlands Surgery, Johann-Wolfgang Goethe University, Frankfurt, Germany Surgery, University Hospital, Antwerpen, Belgium Visceral and Transplantation Surgery, Inselspital, Bern, Switzerland Colorectal Surgery, The Cleveland Clinic Foundation, USA Surgery, Columbia University, New York, NY 10032, USA Surgery, University Hospital Charite , Berlin, Germany

Received: 17 December 1997/Accepted: 2 January 1998

Metastatic disease at trocar wounds after laparoscopic surgery is one of the most important factors that precludes wide employment of laparoscopic techniques to resect malignant disease. Various reports of port-site metastases after laparoscopic resection of colorectal cancer, diagnostic laparoscopy for digestive cancer, or laparoscopic removal of gallbladders with occult cancer have been published [19]. In the first years of laparoscopic colorectal surgery, incidences of port-site metastases as high as 21% were quoted [19]. These high rates were probably due to poor surgical technique during early experience because more recent studies of larger populations of patients have documented incidences of port-site metastases as low as 2% [1, 13, 14]. Recurrence of colorectal cancer in abdominal wounds after conventional, open surgery for colorectal malignancy, has always been suggested to be rare. An incidence of 0.69% is frequently quoted on the basis of a study by Hughes [8]. However, Gunderson and Sosin [5] reported a higher incidence of wound recurrence after open curative resection and colorectal cancer. At routine relaparotomy 3 months after initial colorectal surgery, 3.3% of patients had colorectal cancer in the abdominal wall. Two-thirds of these recurrences had not been discovered at physical examination, probably because they were located at the fascial level. Therefore, it remains unclear whether abdominal wall metastases occur more frequently after laparoscopic than open surgery until randomized clinical trials have been completed. The pathogenesis of port-site metastases has not been unraveled. Development of port-site metastases requires the presence of viable cancer cells at the trocar site. This situation is very likely to occur when a malignant tumor is removed through a narrow incision of the abdominal wall. Such direct implantation of tumor cells has been confirmed by Clair et al. [4], who encountered an abdominal wall metastasis at the extraction site scar after laparoscopic removal of a gallbladder with unsuspected carcinoma. In an experimental study with rats, Bouvy et al. [3] recorded increased tumor growth at the extraction site of a lump of CC 531 colon cancer, which had been placed in the peritoneal henbu hl et cavity for 20 min while CO2 was insufflated. Kra

Correspondence to: H. J. Bonjer

al. [12] implanted a 1-mm3 cube of Morris Hepatoma 3924 A in the left liver lobe of rats. Tumor growth at the trocar site used to remove either a biopsy or the entire tumor was greater than at the other trocar sites. Protection of the extraction wound was not used by Bouvy or Kra henbu hl. Considering these data, it appears justifiable to state that the extraction site is at risk for abdominal wall metastases. Protection of the specimen or the extraction site should be recommended, and the size of the extraction site should allow atraumatic passage of the specimen. Instead of direct implantation of tumor cells during extraction of a tumor, free viable cancer cells can implant at trocar sites. Umpleby [18] has shown that free viable cancer cells occur in the peritoneal cavity in 70% of all patients with colorectal cancer. This observation provides a basis for justifying the cell suspension model that has been employed by the majority of experimental laparoscopy researchers. In the cell suspension model, cultured cancer cells are injected in the peritoneal cavity before or at the time of a surgical procedure. After an interval of time, the experimental animals are killed and tumor growth is assessed at autopsy. The length of this interval is determined by the number of injected tumor cells and their biologic behavior. Interpretation of experimental studies employing cell suspensions is difficult because different numbers of varying cancer cell lines have been used. Intraperitoneal injection of cancer cells results in spread out irregular growth. Therefore, accurate assessment of peritoneal tumor growth is difficult. Most authors have performed semiquantitative analysis by two observers. Steller et al. [16] have shown that such an analysis is reproducible. It has been suggested that growth of cancer cells at port sites is augmented by insufflating gas into the peritoneal cavity because the turbulence increases transportation of cancer cells to the port sites. In an experimental study with rats using colon cancer cell suspensions, Bouvy et al. [3] recorded absence of tumor deposits at trocar sites when gasless laparoscopy was used, whereas tumor growth was found at trocar sites after CO2 insufflation. These findings were in accordance with a similar experimental study by Hubens et al. [7]. However, Kra henbu hl et al. [12] using a solid liver tumor model found more post-site metastases after laparoscopic resection of the tumor than after open resection. It remains questionable whether ex-

1103

periments using cell suspensions can be extrapolated to the clinical situation because the concentration of intraperitoneal cancer cells appears greater than in the clinical situation. Another hypothesis is that CO2 has a stimulating effect on tumor growth. Jacobi et al. [9] assessed in vitro growth of colon adenocarcinoma by exposing the tumor cells to either CO2 or helium. Tumor growth was significantly less when helium was insufflated. Bouvy et al. [2] studied the growth of a colon cancer that had been implanted underneath the renal capsule before peritoneal insufflation of either CO2 or air while gasless laparoscopy was performed as a control. Renal tumor growth was less after gasless laparoscopy than after CO2 or after insufflation. However, a difference of tumor growth between the CO2 or air group was not found. Further studies are needed to determine the trophic potentials of various insufflating agents. Another concern about gas insufflation during laparoscopy for cancer is leakage of gas through and around trocars. This leakage might result in an accumulation of tumor cells at the trocar site, which has been described as the chimney effect by Kazemier et al. [10]. Tseng et al. [17] observed increased tumor growth at a leaking trocar site when cancer cells were injected intraperitoneally in rats having a pneumoperitoneum. However, the size of the leaking trocar site was twice that of the nonleaking trocar site. Therefore, the increased tumor growth could be partially due to a difference in size of the wounds. Kim et al. [11] assessed port-site metastases in a 5-mm wound in rats that had either no insufflation of gas or insufflation of either CO2 or air. Tumor deposits at the trocar sites did not differ in size among these three groups. Hubens et al. [6] also could not prove existence of the chimney effect in studies in rats. Therefore, the role of the chimney effect remains unclear. The impact of tissue trauma on tumor growth at the trocar site has been studied by Tseng et al. [17] in a rat study. Greater port-site metastases were encountered when trocar sites were crushed with a surgical clamp before trocar insertion. These findings are in accordance with earlier observations that surgical trauma promotes tumor growth [15]. Therefore, it appears preferable to use trocars that can be inserted and secured with minimal tissue trauma. Possibly, trocars with a balloon on both the peritoneal and cutaneous sides provide an atraumatic and gas-tight fixation. Local application of cytotoxic or antiadherence agents at trocar sites after a laparoscopic procedure is another option for preventing port-site metastases. Jacobi et al. [9] studied peritoneal tumor growth after application of heparin, taurolidin or a combination of the two agents. Tumor growth was the least when a combination of taurolidine and heparin was used. Use of other agents such as iodine, chlorate, and chemotherapeutics to prevent port-site metastases requires further study. Future experimental research should focus on the estab-

lishment of a reproducible solid tumor model that represents the clinical situation of colorectal cancer better than the cell suspension model. Local application of agents at the trocar site and the use of alternative insufflating gases and gasless laparoscopy deserve further study to determine their role in prevention of port-site metastases. References
1. Ballantyne GH (1995) Laparoscopic-assisted colorectal surgery: review of results in 752 patients. Gastroenterol 3: 7589 2. Bouvy ND, Giuffrida MC, Tseng LNL, Steyerberg EW, Marquet RL, Jeekel J, Bonjer HJ. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt 3. Bouvy ND, Marquet RL, Jeekel J, Bonjer HJ (1996) Ann Surg 224: 694701 4. Clair DG, Lautz DB, Brooks DC (1993) Rapid development of umbilical metastases after laparoscopic cholecystectomy for unsuspected gallbladder carcinoma. Surgery 113: 355358 5. Gunderson LL, Sosin H (1974) Areas of failure found at reoperation (second or symptomatic look) following curative surgery for adenocarcinoma of the rectum: clinicopathologic correlation and implications for adjuvant therapy. Cancer 34: 12781292 6. Hubens G, Pauwels M, Eyskens E Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt 7. Hubens G, Pauwels M, Hubens A, Vermeulen P, van March E, Eyskens E (1996) The influence of a pneumoperitoneum on the peritoneal implantation of free intraperitoneal colon cancer cells. Surg Endosc 10: 809812 8. Hughes ES, McDermott FT, Polglase Al, Johnson WR (1983) Tumor recurrence in the abdominal wall scar tissue after large bowel cancer surgery. Dis Colon Rectum 26: 571572 9. Jacobi CA, Ordemann J, Bo hm B, Zieren HU, Sabat R, Mu ller JM. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt 10. Kazemier G, Bonjer HJ, Berends FJ, Lange JF (1995) Port-site metastases after laparoscopic colorectal surgery for cure of malignancy (letter). Br J Surg 82: 11411142 11. Kim SH, Casillas S, Wilsom JW, Dietz DW, Vladisavljevic A. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt 12. Kra henbu hl L, Baer HU, Renzulli P, Feodorovici MA, Scha fer, Bu chler MW. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt 13. Lacy AM, Garcia-Valdecas JC, Pique JM, Delgado S, Campo E, Bordas JM, Taura P, Grande L, Fuster J, Pacheco JL (1995) Short-term outcome analysis of a randomized study comparing laparoscopic vs open colectomy for cancer. Surg Endosc 9: 11011105 14. Lord SA, Larach SW, Ferrara A, Williamson PR, Lago CP, Lube MW (1996) Laparoscopic resections of colorectal carcinoma: a three-year experience. Dis Colon Rectum 39: 148154 15. Murthy SM, Goldschmidt RA, Rao LN, Ammirati M, Buchmann T, Scanlan EF (1989) The influence of surgical trauma on experimental metastasis. Cancer 64: 20352044 16. Steller E Ph (1988) Enhancement and abrogation: modifications of host immune status influence IL-2 and LAK cell immunotherapy. Thesis, Erasmus University, Rotterdam, The Netherlands 17. Tseng LNL, Berends FJ, Bouvy ND, Marquet RL, Kazemier G, Bonjer HJ. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt 18. Umpleby HC, Fermor B, Symes MO, Williamson RCN (1984) Viability of exfoliated colorectal carcinoma cells. Br J Surg 71: 659663 19. Wexner SD, Cohen SM (1995) Port-site metastases after laparoscopic colorectal surgery for cure of malignancy. Br J Surg 82: 295298

1099

Peritonitis and adhesions in laparoscopic surgery

First workshop on experimental laparoscopic surgery, Frankfurt, 1997
C. A. Jacobi,1 L. Kra henbu hl,2 C. Blo chle,3 H. J. Bonjer,4 C. N. Gutt5
1 2 3

Department Department Department 4 Department 5 Department

of of of of of

Surgery, University Hospital Charite , Humboldt-University of Berlin, Schumann Strae 20/21, D-10098, Berlin, Germany Visceral and Transplantation Surgery, Inselspital, Bern, Switzerland Surgery, University Hospital Eppendorf, Hamburg, Germany Surgery, University Hospital Dijkzigt, Rotterdam, Netherlands Surgery, Johann-Wolfgang-Goethe University, Frankfurt, Germany

Received: 17 December 1997/Accepted: 2 January 1998

Laparoscopic surgery is increasingly used for intraabdominal diseases complicated by inflammatory processes and peritonitis. In prospective randomized trials [1, 8, 17] laparoscopic appendectomy appeared to be superior even to open appendectomy in terms of postoperative complications and recovery, and even laparoscopic management of perforated peptic ulcers has been reported to be simple and followed by a short recovery of the patients [20]. Although successful treatment of appendicitis, perforated peptic ulcer, and diverticulitis has been reported with low morbidity [1, 6, 8, 17, 20], the influence of pneumoperitoneum and elevated intraperitoneal pressure on bacteraemia, endotoxaemia, and systemic inflammation still remains unclear. Prospective clinical trials comparing open and laparoscopic procedures are needed to evaluate the benefit of each surgical procedure for patients with intraperitoneal inflammation. Besides clinical observations, experimental studies have tried to elucidate the influence of laparoscopy in peritonitis on bacterial translocation and endotoxaemia in different animal models. However, regarding the effects of pneumoperitoneum on physiologic changes and systemic inflammation during sepsis, only a few controversial experimental data exist. One theoretical concern in most of these studies was that elevated intra-abdominal pressure may promote bacteremia and systemic inflammatory response during laparoscopic surgery because it has been demonstrated that elevated intra-abdominal pressure leads to increased patency of lymphatic openings through which intra-abdominal elements and fluids are drained from the peritoneal cavity [19, 21]. Most of the authors used an animal model in which peritonitis was caused by intraperitoneal inoculation of Escherichia coli [3, 4, 7, 16]. Although the concentrations of E. coli were comparable between the different studies, the results were controversial. Gurtner et al. [7] found that pneumoperitoneum of 12 mmHg in rabbits did not increase bacteremia or endotoxemia, whereas Eleftheriadis et al. [3] demonstrated that elevated intra-abdominal pressure (15 mmHg) leads to intestinal ischemia, oxygen-free radial production, and increased bacterial translocation in rats. Sare et al. [16] and Evasovich et al. [4] also used rat models with intraperitoneal injection of E. coli, and both found a significant increase of bacteremia in the carbon dioxide group compared at least with that of the control group. Therefore, studies performed in rats and rabbits perhaps cannot be

Correspondence to: C. A. Jacobi

compared concerning intraperitoneal inflammatory diseases and perioperative bacterial translocation. It is also questionable whether these experimental models using inoculation of E. coli alone are similar to the clinical situation and spectrum of bacterial species commonly detected in blood cultures of patients with abdominal sepsis [22, 23]. Furthermore, laparoscopy was not compared with laparotomy in most of these studies [3, 4], although laparotomy is the only alternative approach to laparoscopic treatment of abdominal septic diseases. In two other studies, different sepsis models of rats were used to imitate the clinical situation [2, 9]. In the experiments of Bloechle et al. [2], peritonitis was induced by gastric ulceration (instillation of 2-ml ethanol 2%) followed by a defined gastrostomy of 2 mm in rats. The researchers found a significant increase in extent and severity of peritonitis and bacteremia after laparoscopy compared with that of the control group when the interval between gastric perforation and pneumoperitoneum lasted more than 12 h. Although this model seems to be similar to the clinical situation, indicating that laparoscopic techniques may be harmful under these conditions, the authors did not mention whether gastric perforation caused the same amount of intraperitoneal bacterial species in each animal. Unfortunately, a comparison between laparoscopy and laparotomy was not performed, although laparotomy is the only alternative approach to laparoscopic treatment of abdominal septic diseases. Therefore, it remains unclear whether laparoscopy or laparotomy should be used in the treatment of inflammatory diseases in this model. Jacobi et al. [9] also used another model of peritonitis in rats to analyze the influence of laparoscopy and laparotomy on bacterial translocation and endotoxaemia. In this experiment a nonfatal peritonitis was induced by intraperitoneal injection of standardized stool suspension with 42 different bacterial species. It was demonstrated by Lorenz et al. [13] that this model is very similar to the clinical situation. Intraabdominal pus, abscess formations, adhesions, and granulocytic infiltration of the organs were found in this model. Bacteremia and endotoxemia were significantly higher 1 h after laparotomy and laparoscopy than in the control group, but aerobic and anaerobic bacteria were found only in the laparotomy group. The number of abscesses was also significantly higher after laparotomy than after laparoscopy and in the control group in these experiments [9]. Besides elevating intraperitoneal pressure, the use of carbon dioxide also might influence bacterial growth and

1100

translocation. In an experimental study in rats Jacobi et al. [10] also evaluated the effects of pneumoperitoneum with carbon dioxide and an alternative gas, helium, using the same peritonitis model as mentioned before. Interestingly, they found a significant decrease of systemic inflammation after laparoscopy with helium compared with carbon dioxide and even with the control group. Thus, it appears that a laparoscopic approach with alternative gases such as helium may also be beneficial in patients with intra-abdominal infection. Further experimental and prospective randomized clinical trials are needed to analyze the pathomechanisms of bacterial translocation and systemic inflammation during laparotomy and laparoscopy. Alternative procedures such as laparoscopy with different gases or gasless laparoscopy also must be included in these evaluations. Regarding the controversial results of the different experimental studies, a uniform experimental setup must be developed to make experimental investigations comparable concerning the influence of laparoscopy and laparotomy on systemic inflammation during peritonitis. The incidence and extent of postoperative adhesion formation is related to intraperitoneal inflammatory reaction and has been investigated in laparoscopic surgery in the last few years. Intraperitoneal inflammation and peritoneal healing seem to enhance postoperative adhesions as shown by Pados et al. [15]. Clinical observations in gynecologic patients demonstrated that laparoscopic surgery is associated with less severe adhesion formation [14]. However, comparative scientific data are lacking due to restricted possibilities for repeated analysis. Only a few experimental studies have been performed investigating the effect of laparoscopic and open surgical procedures on postoperative adhesion formations [5, 11, 12, 18]. Although the animals differed in all four experimental studies, three authors found less intraperitoneal adhesion formations after laparoscopic approach in their specific models [11, 12, 18]. Kra henbu hl et al. [12] used a small animal model of the rat comparing a laparoscopic and an open fundoplication. Rats with open fundoplication developed significantly more adhesions than rats with the laparoscopic procedure. Furthermore, laparoscopic surgery induced more parietal adhesions, whereas open surgery was associated with more visceral adhesions in this model. In dogs, Tittel et al. [18] also found fewer adhesions after cecal resection, deserozation of 2 cm2 of the abdominal wall, and resection of the omentum majus in the laparoscopic group than in the open group. Jorgensen et al. [11] demonstrated a similar incidence of peritoneal adhesion formation after laparoscopic and open standardized deserozation of the cecum and peritoneal surface in rabbits. However, laparoscopy was associated with a lower incidence of wound adhesions. In contrast to these results, Fowler [5] could not find any significant difference in postoperative adhesions when open and laparoscopic pelvic lymphadenectomies were compared in pigs. But in these experiments 80% (8 to 10) of the laparoscopy group and 56% (5 to 9) of the laparotomy group had no adhesions, which may explain the difficulty in demonstrating a statistical significance between the two groups. The pathomechanisms that explain the findings of the different studies are yet unknown. Kra henbu hl et al. [12]

investigated intraperitoneal TNF- levels, which might be a factor in the formation of adhesions, but did not find a difference between laparoscopy and laparotomy. Again, further experimental and prospective randomized clinical trials are needed to analyze the pathomechanisms of intraperitoneal adhesions after laparotomy and laparoscopy. Alternative procedures, such as gasless laparoscopy and the intraperitoneal instillation of different agents, should also be included in these evaluations. Regarding experimental investigations, a uniform experimental setup is needed to make results concerning this topic comparable. References
1. Attwood SE, Hill AD, Murphy PG, Thorton J, Stephens RB (1992) A prospective randomized trial of laparoscopic versus open appendectomy. Surgery 112: 497501 2. Blo chle C, Emmermann A, Treu H, Achilles E, Mack D, Zornig C, Broelsch CE (1995) Effect of a pneumoperitoneum on the extent and severity of peritonitis induced by gastric ulcer perforation in the rat. Surg Endosc 9: 898901 3. Eleftheriadis E, Kotzampassi K, Papanotas K, Heliadis N, Sarris K (1996) Gut ischemia, oxidative stress, and bacterial translocation in elevated abdominal pressure in rats. World J Surg 20: 1116 4. Evasovich MR, Clark TC, Horattas MC, Holda S, Treen L (1996) Does pneumoperitoneum during laparoscopy increase bacterial translocation? Surg Endosc 10: 11761179 5. Fowler JM, Hartenbach EM, Reynolds HT, Borner J, Carter JR, Carlson JW, Twiggs LB, Carson LF (1994) Pelvic adhesion formation after pelvic and extraperitoneal lymphadenectomy: comparison between transperitoneal laparoscopy and extraperitoneal laparotomy in a porcine model. Gynecol Oncol 55: 2528 6. Geis WP, Kim HC (1995) Use of laparoscopy in the diagnosis and treatment of patients with surgical abdominal sepsis. Surg Endosc 9: 178182 7. Gurtner GC, Robertson CS, Chung SCS, Ling TKW, IP SM, Li AKC (1995) Effect of carbon dioxide pneumoperitoneum on bacteraemia and endotoxaemia in an animal model of peritonitis. Br J Surg 82: 844848 8. Hansen JB, Smithers BM, Schache D, Wall DR, Miller BJ, Menzies BL (1996) Laparoscopic versus open appendectomy: a prospective randomized trial. World J Surg 20: 1721 9. Jacobi CA, Ordemann J, Bo hm B, Zieren HU, Volk HD, Lorenz W, Halle E, Mu ller JM (1997) Does laparoscopy increase bacteraemia and endotoxaemia in a peritonitis model? Surg Endosc 11: 235238 10. Jacobi CA, Ordemann, Sabat R, Volk HD, Lorenz W, Halle E, Mu ller JM. Data presented at the first Workshop of Experimental Laparoscopic Surgery 1997, Frankfurt, Germany 11. Jorgensen JO, Lalak NJ, Hunt DR (1995) Is laparoscopy associated with lower rate of postoperative adhesions than laparotomy? A comparative study in the rabbit. Aust N Z J Surg 65: 342344 12. Kra henbu hl L, Scha fer M, Kuzinkovas V, Renzulli P, Baer HU, Bu chler M. Data presented at the first Workshop of Experimental Laparoscopic Surgery 1997, Frankfurt, Germany 13. Lorenz W, Reimund KP, Weitzel F, Celik I, Kurnatowski M, Schneider C, Mannheim W, Heiske A, Neumann K, Sitter H, Rothmund M (1994) Granulocyte colony-stimulating factor prophylaxis before operation protects aganist lethal consequences of postoperative peritonitis. Surgery 116: 925934 14. Lundorff P, Hahlin M, Kallfelt B, Thorburn J, Lindblom B (1991) Adhesion formation after laparoscopic surgery in tubal pregnancy: a randomized trial versus laparotomy. Fertl Steril 55: 911915 15. Pados GA, Devroey P (1992) Adhesions. Curr Opin Obstet Gynecol 4: 412418 16. Sare M, Yesilada O, Gurel M, Balkaya M, Yologlu S, Fiskin K (1997) Effects of CO2 insufflation on bacterial growth in rats with Escherichia coli- induced experimental peritonitis. Surg Laparosc Endosc 7(1): 3841. 17. Tate J, Dawson JW, Chung SC, Lau WY, LI AK (1993) Laparoscopic versus open appendectomy: prospective randomised trial. Lancet 342: 633637

1101 18. Tittel A, Schippers E, Treutner KH, Anuroff M, Polivoda M, Ottinger A, Schumpelick V (1994) Langenbecks Arch Chir 379: 9598 19. Tsilibary EC, Wissig SL (1983) Lymphatic absorption from the peritoneal cavity: regulation of patency of mesothelial stomata. Microvasc Res 25: 2239 20. Urbano D, Rossi M, De Simone P, Berloco P, Alfani D, Cortesini R (1994) Alternative laparoscopic management of perforated peptic ulcers. Surg Endosc 8: 12081211 21. Walker AP, Condon RE (1989) Peritonitis and intraabdominal abscesses. In: Schwartz S, Ed., Principles of surgery, 5th ed. McGrawHill, New York: 14591489 22. Walker AP, Krepel CJ, Gohr CM, Edmiston CE (1994) Microflora of abdominal sepsis by locus of infection. J Clin Microbiol 32: 557558 23. Zubkov MN, Menshikov DD, Gugutsidze EN, Chegin VM, Vasina TA (1995) Microbiologic diagnosis of mixed anaerobic and aerobic surgical infections. Antibiot Khimioter 40: 4650

1096

Metabolism and immunology in laparoscopy

First workshop on experimental laparoscopic surgery, Frankfurt, 1997
C. N. Gutt,1 C. Kuntz,2 Th. Schmandra,1 A. Wunsch,2 P. Heinz,3 N. Bouvy,4 M. Bessler,5 Ph. Sa nger,1 J. Bonjer,4 5 6 5 J. Allendorf, C. A. Jacobi, R. Whelan
1 2

Department Department 3 Department 4 Department 5 Department 6 Department

of of of of of of

Surgery, Surgery, Surgery, Surgery, Surgery, Surgery,

Johann-Wolfgang-Goethe University, Theodor-Stern-Kai 7, D-60590, Frankfurt, Germany University of Heidelberg, Germany Hospital Nordwest, Frankfurt, Germany University Hospital Dijkzigt, Rotterdam, the Netherlands Columbia University, New York, NY 10032, USA University Hospital Charite, Berlin, Germany

Received: 17 December 1997/Accepted: 2 January 1998

Surgery has been shown to result in a physiologic stress response and postoperative immunosuppression. The human and animal organisms answer to such an injury is characterized by an early rise of the serum level for stress hormones and a drop in the immunoresponse conveyed by the cells. This causes a reduction in lymphocyte and macrophage interactions; a decreased activity of natural killer cell, lymphocyte, and neutrophil chemotaxis; and delayedtype hypersensitivity (DTH) responses. The force and duration of the stress reaction is considered to be proportional to the severity of the injury. There is a correlation between a reduced perioperative cell-conveyed immunoresponse and an increased risk of postoperative infection and metastatic tumor spreading [12, 14, 15, 16, 18, 21]. Laparoscopic surgery is associated with less postoperative pain and earlier return to normal activity than conventional surgery. Furthermore, immune function of the patient may be better preserved using the laparoscopic approach [10, 17]. Nevertheless, further experimental studies about metabolism and immunology are necessary to support suggested advantages. The rat has been used extensively in studies of immunity and cancer is increasingly used for laparoscopic research [2, 3, 7, 8]. Surgery alters the stress and immune system of healthy rats. To evaluate the stress and immunologic response of laparoscopic and open surgery Kuntz [13] performed laparoscopy-assisted and conventional colon resection in rats. Immediately before and after surgery as well as 1 and 7 days postoperatively serum levels of corticosterone, neoptrine, and IL 1- were measured. At the end of the operation, significant differences were found between the laparoscopic and conventional groups with regard to corticosterone, neopterine, and IL 1-. One week after the operation, the stress and immune parameters were back to normal levels in all except the IL 1- group, but the recovery indicated by body weight was different according to the kind of applied procedure. Seven days after open surgery, the body weight was significant lower than after laparoscopic surgery. The body weight supposedly reflects anabolic state. Another representative parameter of anabolism is insulin-like growth factor I (IGF-I), an analog of insulin. IGF is important for normal human growth and development and

Correspondence to: C.N. Gutt

may play an important role in the propagation of malignant cells. IGF-I migrates to the extravascular tissues to promote metabolism, growth, and regeneration. Decrease of serum IGF levels can be regarded as a parameter of the extent of surgical trauma. To assess metabolic consequences of laparoscopic-assisted and open small bowel resection Bouvy et al. [5] investigated serum levels of IGF-I. Furthermore, with a subrenal capsule assay (SRC), growth of subrenal implanted tumor was measured. Serum IGF-I levels at postoperative days 1 and 2 were higher in the laparoscopy group than in the laparotomy group. In the SRC, less tumor growth was observed after laparoscopic bowel resection. To investigate the cell-mediated immune function, Whelan [22] measured the effects of CO2 insufflation and laparotomy on susceptibility to bacterial infection by skin pustule sizes after intradermal injection of Staphylococcus aureus 502A in rats. In this study, laparotomy was compared with pneumoperitoneum. Immune function also was assessed by measuring DTH responses to phytohemagglutinin (PHA), a nonspecific T-cell mitogen. After 48 h the pustule size in the laparotomy group was significantly larger than in the insufflation group. Similarly, the DTH was significantly smaller in the laparotomy group. In a similar study Bessler [4] measured the skin pustule sizes after intradermal injection of S. aureus 502A in rats undergoing laparoscopy-assisted cecal resection and open cecal resection. In the laparoscopy group, a 4 to 6-mmHg carbon dioxide pneumoperitoneum and 4-port technique was used to pull the cecum outside for extracorporeal resection. The rats were inoculated both preoperatively and 1 day after the operation. Animals undergoing laparoscopyassisted cecal resection had both significantly smaller and more rapidly healing pustules than their open counterparts [4, 22]. Jacobi et al. [11] evaluated TNF-alpha and IL-10 plasma levels in rats after insufflation with either CO2 or helium (8 mmHg, 30 min). TNF-alpha plasma levels were significantly decreased and IL-10 plasma levels significantly increased in the CO2 group compared with the helium and control groups in the postoperative course. Allendorf et al. [1] investigated cell-mediated immune function after laparoscopy-assisted and open bowel resection in rats by DTH responses to keyhole limpet hemocyanin (KLH) and phytohemagglutinin (PHA). Laparoscopyassisted resection group (CO2-pneumoperitoneum, 4 to 6-

1097

mmHg) responses differed from those of controls in only two of eight postoperative measures but they were significantly greater than responses to open resection. In a second study, Allendorf et al. measured the effects of incision length and exposure method for cecal resection with respect to postoperative immune function as assessed by DTH reactions. Rats underwent laparotomy (7 cm), minilaparotomy (3.5 cm), or laparoscopy (4 ports). The laparoscopy-assisted resection group responses were larger than open group responses, but not significantly different from those of an anesthesia control. The laparotomy group response was smaller than that of an anesthesia control. The mini-laparotomy group showed no significant difference in comparison with any other group. Numerous experimental and clinical investigations indicate that the mononuclear phagocytes system (MPS) has a relevant function in terms of physiologic defense against tumor metastases and bacterial infection. Consequently, a point of major interest is the influence of surgical techniques on the MPS function. Heinz and Gutt [9] examined the phagocytosis activity of the rats MPS during conventional fundoplication and laparoscopic fundoplication using a pneumoperitoneum and gasless laparoscopy. The MPS function was evaluated by an intravascular carbon clearance test. The fastest carbon elimination half-life was found in the gasless laparoscopy group. By way of contrast, there was a significant increase of carbon half-life in the laparoscopy group using a pneumoperitoneum. Even the open group caused less MPS depression [6, 9]. To measure alterations of hemodynamics in the portal vein during increasing intra-abdominal pressure due to CO2 insufflation, Schmandra and Gutt [20] extended the rat model by implantation of a perivascular flow probe. By use of this technique, an adequate flowmetry of the portal vein was possible. The installation of a CO2 pneumoperitoneum with elevating intra-abdominal pressure led to a linear decrease of portal venous flow. Wunsch [23] measured the influence of different gases used for insufflation on the pH inside the abdominal cavity and in subcutaneous tissue. After a control period without insufflation, rats were insufflated subsequently with room air, CO2 and helium with a pressure of 3 mmHg. Insufflation with helium and room air produced no significant changes in the pH value. Significant acidosis inside the abdominal cavity and in subcutaneous tissue were measured during CO2 insufflation. Nitric oxide (NO), a unique biological mediator, regulates different physiologic and pathophysiologic reactions. After stimulation by infection or trauma, NO causes vasodilatation and hypotension, increases microcirculation, inhibits platelet aggregation, and has an antiproliferative effect. Twenty-four hours after laparoscopic and open Nissen fundoplication in rats, Sa nger and Gutt [19] examined three isoforms: iNOS, ncNOS, and ecNOS. The iNOS showed a significant increased activity in liver, duodenum, and peritoneum 24 h after laparotomy. After laparoscopy with pneumoperitoneum (60 min, 8 mmHg), an increased activity was seen in the duodenum and perritoneum. In the gasless laparoscopy group only the peritoneum showed a slightly increased activity of the iNOS. Postoperative recovery and immune function varies with the degree of surgical trauma. Therefore, rats that under-

went laparoscopic operation showed quicker recovery than laparotomized rats [13]; procedures done through small incisions resulted in preservation of postoperative immune function [1]; and resistance to bacterial dermal infection and cell-mediated immune function were better preserved after pneumoperitoneum than after laparotomy [22]. Laparotomy stimulated the iNOS 24 h after Nissen fundoplication more than the laparoscopic approach. But the total invasiveness of a surgical procedure does not depend alone on the size of the incision used for access. Especially in the use of complex surgical techniques, the duration of a procedure, the local tissue trauma, and general effects of the operation method must also be considered. Insufflation of CO2 gas to establish a pneumoperitoneum causes partly known local and systemic side effects to the organism. Depending on pressure and duration, this technique can be disadvantegous. Local defense and immune mechanisms might be suppressed by the altered pH in the subcutaneous tissue. Macrophages, for example, show less phagocytotic activity under acidic conditions. Increased blood flow after CO2 insufflation might provide tumor cells with better chances of growing [23]. The phagocytosis activity can be decreased by gas insufflation during surgical procedures [6, 9]. Increased intra-abdominal pressure of a pneumoperitoneum compromises splanchnic blood flow and the afferent circulation of the liver. Portal blood flow plays an important role in liver function and cell-conveyed immunoresponse [20]. In further studies, all factors of invasiveness under a certain surgical method must be considered. For example, the pressure and duration of a pneumoperitoneum and the surgical trauma during laparoscopic procedures in the rat should be comparable to a realistic human situation. Tissue reactions and changes in local organ function seem to be of certain interest. References
1. Allendorf JD, Bessler M, Whelan RL, Trokel M, Laird DA, Terry MB, Treat MR (1996) Better preservation of immune function after laparoscopic-assisted vs. open bowel resection in amurine model. Dis Colon Rectum 39: S67S72 2. Berguer R, Gutt CN (1994) Laparoscopic colon surgery in a rat model: a preliminary report. Surg Endosc 8: 11951197 3. Berguer R, Gutt CN, Stiegman GV (1993) Laparoscopic surgery in the rat: description of a new technique. Surg Endosc 7: 345347 4. Bessler M. Cell-mediated immune function after open and laparoscopic surgery (II). Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt 5. Bouvy ND, Marquet RL, Jeekel J, Bonjer HJ (1997) Laparoscopic surgery is associated with less tumour growth stimulation than conventional surgery: an experimental study. Brit J Surg 84: 358361 6. Gutt CN, Heinz P, Kaps W, Paolucci V (1997) The phagocytosis activity during conventional and laparoscopic operations in the rat. Surg Endosc 11: 899901 7. Gutt CN, Held S, Heller K, Paolucci V (1996) A small animal model for laparoscopic microsurgery training. Min Invasive Ther Allied Technol 5: 302306 8. Gutt CN, Riemer V, Brier C, Berguer R (1997) Standardized technique of laparoscopic surgery in the rat. Dig Surg 15: 135139 9. Heinz P, Gutt C. The phagocytosis activity during conventional and laparoscopic operations in rats. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt 10. Horgan PG, Fitzpatrick M, Course NF, et al. (1992) Laparoscopy is less immunotraumatic than laparotomy. Min Invasive Ther 1: 241 11. Jacobi CA, Wenger F, Sabat R, Volk T, Ordemann J, Mu ller JM. The

1098 impact of laparoscopy with cabon dioxide versus helium on immunological function and tumor growth in a rat model. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt Keller SE, Weiss JM, Schleifer SJ, Miller NE, Stein M (1983) Stressinduced suppression of immunity in adrenalectomized rats. Science 221: 13011304 Kuntz C. Differences in stress and immune response in rats undergoing laparoscopic colon resection or open surgery. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt Lennard TWJ, Shenton BK, Borzotta A et al. (1985) The influence of surgical operations on components of the human immune system. Br J Surg 72: 771 Pollock RE, Lotzova E (1987) Surgical-stress-related suppression of natural killer cell activity: a possible role in tumor metastasis. Nat Immun Cell Growth Regul 6: 269278 Ramos JM, Gupta S, Anthone GJ, Ortega AE, Simons AJ, Beart RW Jr (1994) Laparoscopy and colon cancer. Is the port site at risk? A preliminary report. Arch Surg 9: 897899 Redmond HP, Watson RW, Houghton T, Londran C, Watson RG, Bouchier-Hayes D (1994) Immune function in patients undergoing open vs laparoscopic cholecystectomy. Arch Surg 12: 12401246 Saba TM, Antikatzides TG (1976) Decreased resistance to intravenous tumor cell challenge during periods of reticuloendothelial depression following surgery. Br J Cancer 34: 381386 Sa nger Ph, Gutt C. Nitric oxide synthase activity in laparoscopic vs. open surgery in rats. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt Schmandra Th, Gutt C. Portal vein flow during CO2-pneumoperitoneum in rats. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt Walsh DC, Wattchow DA, Wilson TG (1993) Subcutaneous metastases after laparoscopic resection of malignancy. Aust N Z J Surg 7: 563565 Whelan R. Cell-mediated immune function after open and laparoscopic surgery (I). Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt Wunsch A. Influence of different gases used for insufflation on the pH of subcutaneous tissue. Data presented at the First Workshop on Experimental Laparoscopic Surgery 1997, Frankfurt

18.

12. 13.

19.

20.

14. 15. 16. 17.

21.

22.

23.

Surg Endosc (1998) 12: 10311034

Springer-Verlag New York Inc. 1998

Does the ultrasonically activated scalpel release viable airborne cancer cells?
C. C. Nduka, N. Poland, M. Kennedy, J. Dye, A. Darzi
Academic Surgical Unit, Imperial College School of Medicine at Saint Marys, Norfolk Place, London W2 3PG, United Kingdom Received: 6 June 1997/Accepted: 20 November 1997

Abstract Background: Viable cancer cells may implant at distant sites and cause tumor recurrence. One possible mechanism is the inadvertent exfoliation of viable tumor cells during dissection. The ultrasonically activated scalpel (UAS) uses ultrasonic energy to disrupt tissues by cavitation and produces a dense cloud of cellular debris that may contain viable cells. This study aimed to investigate the viability of airborne cells released during cancer dissection using the UAS and electrosurgery. Methods: Flank tumors (n 8) measuring 1 cm3 were induced in male WAG rats by subcutaneous injection of 2 106 CC531s colon cancer cells. Dissection was performed in cutting mode using the maximum power output of the respective devices. Electrosurgery was performed using a standard monopolar electrosurgical unit and a needle probe, and ultrasonic dissection was performed with the Harmonic Scalpel utilising the open surgical handset and the hooked spatula tip. The smoke plume was aspirated by a vacuum pump and bubbled under Hanks balanced salt solution to trap particulate matter. The viability of the cellular material was blindly assessed with the trypan blue test and by in vitro culture. The morphology of the cellular debris was studied by examination of cytospin preparations. Results: Large quantities of cellular debris was trapped in the plume from both devices. However, no viable cells were isolated, nor did in vitro cell growth occur with either device. Examination of the debris from the UAS demonstrated a characteristic mixture of amorphous forms and very few morphologically intact cells. The cauterized tumor produced charred cells and tissue fragments. Conclusions: In conclusion, this study demonstrates that viable airborne cancer cells are not released after tumor ablation with the UAS or electrosurgery.
Presented at the Meet the Experts poster session at the Society of American Gastrointestinal Endoscopic Surgeons (SAGES) Annual Scientific Session, San Diego, California, USA, 2122 March 1997 Correspondence to: A. Darzi, Minimal Access Surgery Unit, Saint Marys Hospital, Praed Street, London, W2 1NY United Kingdom

Key words: Ultrasonically-activated scalpel Harmonic Scalpel Electrosurgery Laparoscopy Airborne cancer cells Port-site metastases

Although recurrence of cancer at a surgical incision is a well-recognized problem [1, 15], the numerous reports of port-site recurrence after laparoscopic surgery [3] has revived interest in the mechanisms of iatrogenic cancer spread. One possible mechanism of intraoperative tumor spread is the inadvertent release of airborne cancer cells during tumor resection. However, for this mechanism to be feasible, it must be demonstrated that any cancer cells released are viable. Interestingly, it has been demonstrated that use of the pulsed neodymium laser may result in the airborne spread of viable tumor cells, whereas the use of CO2 laser and electrosurgery do not [2, 13]. An increasingly used dissection modality in laparoscopic surgery [5, 6, 14] is the ultrasonically activated scalpel (UAS), also known as the Harmonic Scalpel. This device employs ultrasonic energy to disrupt tissues by cavitation and produces a dense cloud of cellular debris that may contain viable cells. This study investigated the morphology and viability of airborne cells released during cancer dissection by use of the UAS.

Materials and Methods

The cell line used was CC531s, a 1,2-dimethylhydrazine-induced moderately differentiated colon cancer transplantable in syngeneic WAG rats. The cells were grown in vitro in RPMI 1640 (Dutch Modification) supplemented with 10% fetal bovine serum, 20 mM HEPES, 12 mM NaHCO3, 4 mM L-glutamine, 50 IU/ml penicillin, and 50 g/ml streptomycin. The cells were maintained in at 37C in a 5% CO2 incubator and passaged every 3 days. For injection, the cancer cell monolayers were disaggregated with 0.25% trypsin-EDTA for 3 min at 37C and adjusted to 2 106cells/ ml in RPMI. All cell culture reagents were obtained from Sigma-Aldrich Co. Ltd. (Poole, UK) unless otherwise indicated. Inbred WAG rats weighing 350400 g were obtained from the Department of Comparative Biology at Charing Cross and Westminster Medical School, London and bred under pathogen-specific free laboratory condi-

1032

Fig. 1. Schematic diagram of the laboratory setup. The flank tumor is ablated, and the cellular debris aspirated by a suction system positioned 5 mm above the tumor surface. The cells are trapped under culture medium before isolation and cytologic examination. Fig. 2. Scanning electron micrograph (SEM) of normal CC531s colon cancer cell morphology. The cells were mounted fixed from a suspension of cells and demonstrate characteristic microvilli covering the cell membrane (4000).

tions (temperature 2022C, relative humidity 60%, 12-h light cycle). Flank tumors measuring 1 cm3 were induced in eight male WAG rats 3 weeks after subcutaneous injection of the cancer cells. The animals were anesthetised with 3% halothane by mask and killed by intracardiac injection of pentobarbitone at the end of the procedure. The flanks were shaved and cleaned with alcohol, and using aseptic technique the subcutaneous tumor nodules were exposed. The nodules were bisected for ablation with either the ultrasonic scalpel or electrosurgery. Ultrasonic dissection was performed with the Harmonic Scalpel using the open surgical handset and the hooked spatula tip at a power setting of 5. As a negative control, electrosurgical dissection was performed with a standard monopolar electrosurgical unit and needle probe in cutting mode at 60 W. The smoke plume was aspirated by a vacuum pump (5 l/min) with the suction funnel positioned 5 mm from the tumor surface to ensure that no cells escaped. The cellular debris was transported with 60 cm of plastic tubing and bubbled under 20 ml of Hanks balanced salt solution (HBBS) or serumfree RPMI to trap particulate matter (Fig. 1). After each experiment the tubing was rinsed to collect any cells trapped in the tubing. A suspension of cultured CC531s cells acted as a positive control. Isolation of the cellular material was achieved by centrifugation (5 min at 2,000 rpm) and three washes in RPMI culture medium. The Trypan blue (0.4% in HBBS) dye exclusion test was performed using an Improved Neubauer Hemocytometer (Sigma) in a blinded manner by an independent observer (J.D.) to assess the number and percentage viability of the trapped cells. The ability of the cells to grow was assessed by in vitro culture in conditions described earlier. The cells were seeded at 106 cells/ml in 25cm2 culture flasks. The culture medium was changed on days 2 and 5, and growth was monitored on an inverted microscope for 1 week.

revealed that no viable cells were isolated from either device. In contrast, the control cells had a viability exceeding 95%. This was confirmed by a total absence of in vitro cell growth in the treated group compared with rapid proliferation of the control cells. Examination of the debris from the UAS by light microscopy demonstrated a characteristic mixture of subcellular amorphous aggregates and very few fragments that, on the basis of their size, could be morphologically intact cells. The cauterized tumor produced debris that could be arbitrarily divided into three categories: (a) completely carbonized particles without morphologic characteristics of cells, (b) charred cells occurring solitarily and in clumps, and (c) coagulated tissue fragments. Electron microscopy confirmed the nonviable nature of the cellular debris. Ultrasonic cavitation resulted in the formation of acellular vesicles, whereas electrosurgery caused loss of cell surface microvilli and transmembrane perforations (Figs. 24).

Morphological examination
The morphology of the cellular debris was studied by examination of cytospin preparations and by scanning electron microscopy. For the cytospins, 1 106 cells in 1.5 ml RPMI were spun in a Cytospin (Shandon, UK) at 850 rpm for 5 min onto glass slides. The cells were fixed in 95% alcohol and stained with Papanicolaou stain for examination by light microscopy. Preparation of specimens for scanning electron microscopy was performed according to standard protocols. Briefly, the cellular debris was mounted on Thermanox (ICN, Pharmaceuticals Inc., California, USA) cover slips, fixed in 3% glutaraldehyde-PBS and freeze-dried in acetone. The specimens were then cut, mounted, and gold-plated for scanning.

Discussion In certain circumstances, the very act of removing a tumor may result in cancer dissemination [1, 15]. As new technologies are incorporated into the surgical armamentarium, it is vital that safety evaluation remain integral to a surgeons assessment of an instruments utility. Several forms of energy have been employed for dissection in endoscopic surgery: electrosurgery, lasers, and ultrasonic energy. Two forms of ultrasonic dissectors are in use: the ultrasonic cavitational aspirator (UCA) and the ultrasonically activated scalpel (UAS) also known as the Harmonic Scalpel. Both devices operate on the basis of mechanical wave energy by the application of electromagnetic energy to either piezoelectric (electostrictive) or magnetostrictive transducers. In contrast to the UCA, which acts primary as a tissue dissector, the UAS can be used for both cutting and coagulation.

Results Large quantities of cellular debris (>1 107 cell debris/ml) were trapped in the plume from both devices after ablation of each tumor. However, the trypan blue dye exclusion test

1033

Fig. 3. SEM of airborne cancer cells released after tumor ablation by electrosurgery. The cells have formed aggregates, and the few morphologically intact cells present demonstrate transmembrane perforations and loss of microvilli. Fig. 4. SEM of airborne cellular debris released by the Harmonic Scalpel. The debris consists of amorphous forms that are subcellular in size and demonstrate multiple cell surface defects and loss of microvilli.

plication of heat to tissues, but the effects of the UAS on tissues differs greatly according to its mode of action. Three main mechanisms are in operation to varying degrees: cavitation, heat generation, and protein denaturation [9, 12]. The process of cavitation arises from the creation, expansion, and implosion of cavities in liquids [7]. Mechanical oscillations of the instrument tip cause internal tissue pressures to rise and fall rapidly. Once the cellular pressure falls below the vapor pressure of cellular fluid, vapor-filled cavities form within the cells. It is the force generated by the expansion and contraction of these that result in tissue dissection. The heat generated by the UAS is a result of internal tissue friction caused by the high-frequency vibrations. In addition, coagulation is achieved partly by denaturing proteins through mechanical disruption of tertiary hydrogen bonds [12]. The ultrasonic scalpel has been proven useful in a number of clinical settings [5, 14]. It is particularly useful because it does not deliver electrical energy to tissue, so there is no danger from stray currents or capacitative coupling, and there is no nerve stimulation, electrical interference, requirement for a grounding pad, nor production of potentially toxic smoke as occurs with electrosurgery [10]. Furthermore, experimental studies have shown that the UAS causes less peritoneal adhesion formation [8] and less thermal damage to tissues [4]. However, although it is known that the pulsed neodymium laser may result in the airborne spread of viable tumor cells, no studies were previously performed to assess the use of the ultrasonic scalpel in cancer surgery. One consequence of applying the UAS to tissues is the generation of a visible plume of cellular debris. Given that the UAS has been shown to produce less thermal damage [8, 9], we hypothesised that some of this airborne debris contained viable cells. However, morphologic examination showed that the cellular debris is composed largely of amorphous cell fragments liberated as a result of the cavitational effect described earlier. In conclusion, this study demonstrates that although large numbers of cellular particles are released after tumor ablation with the ultrasonic scalpel, viable airborne cancer cells are not released by the procedure.
Acknowledgments. The authors thank Dr. Peter Kuppen for donation of the CC531s cell line and Mr. Morris Gross for his technical assistance. This work was supported by a grant from Dr. Carmel Coulter on behalf of the Special Trustees of Saint Marys Hospital.

The UAS is composed of a generator, handpiece, and blade applicator. The generator is a microprocessorcontrolled high-frequency power supply device that delivers an alternating current to the handpiece. The latter houses the acoustic mount and ultrasonic transducer, which is composed of piezoelectric crystals sandwiched between metal cylinders. Application of current across the transducer results in vibrations at its harmonic frequency of 55.5 Hz. These vibrations are transmitted along the handpiece to the instrument tip where they cause longitudinal displacements of 80 m [2]. The geometry of the instrument tip is largely responsible for the relative proportions of cutting or coagulation achieved on application to tissues. With the hookspatula tip, the inner radius is sharp and allows optimum cutting, whereas the outer edge is blunt and provides coagulation. Electrosurgery and lasers cause their effects by the ap-

References
1. Amaral JF (1994) The experimental development of an ultrasonically activated scalpel for laparoscopic use. Surg Laparosc Endosc 4: 9299 2. Amaral JF (1994) Ultrasonic dissection. End Surg 2: 211213 3. Amaral JF (1994) Ultrasonic energy in laparoscopic surgery. In: Szabo Z, Kerstein MD, Lewis JE (eds) Surgical Technology International III, Universal Medical Press Inc, San Francisco, pp 153161 4. Amaral JF (1995) Laparoscopic cholecystectomy in 200 consecutive patients using an ultrasonically activated scalpel. Surg Laparosc Endosc 5: 25562 5. Boddy SAM, Ramsay JWA, Carter S, Webster PJR, Levison DA, Whitfield HN (1987) Tissue effects of an ultrasonic scalpel for clinical surgical use. Urol Res 15: 49:52 6. Brandes WW, White WC, Sutton JB (1946) Accidental transplantation of carcinoma in the operating room. Surg Gynecol Obstet 82: 212

1034 7. DesCoteaux JG, Picard P, Poulin EC, Baril M (1996) Preliminary study of electrocautery smoke particles produced in vitro and during laparoscopic procedures. Surg Endosc 10: 152158 8. Hambley R, Hedba PA, Abell E, et al. (1988) Wound healing of skin incisions produced by ultrasonically vibrating knife, scalpel, electrosurgery, and carbon dioxide laser. J Dematol Surg Oncol 14: 1213 1217 9. Hoye RC, Ketcham AS, Riggle GC (1967) The airborne dissemination of viable tumour cells by high energy neodynium laser. Life Sci 6: 119125 10. Laycock WS, Trus TL, Hunter JG (1996) New technology for the division of short gastric vessels during laparoscopic Nissen fundoplication. Surg Endosc 10: 7173 11. Nduka CC, Monson JRT, Menzies-Gow N, Darzi A (1994) Abdominal wall metastases following laparoscopy. Br J Surg 81: 648652 12. Oosterhuis JW, Verschueren RCJ, Eibergen R, Oldhoff J (1982) The viability of cells in the waste products of CO2-laser evaporation of Cloudman mouse melanomas. Cancer 49: 6167 13. Suslick BH (1989) The chemical effects of ultrasound. Sci Am 2: 6268 14. Thomas CG Jr (1961) Tumour cell contamination of the surgical wound: experimental and clinical observations. Ann Surg 153: 697 705 15. Tulandi T, Chan KL, Arseneau J (1994) Histopathological and adhesion formation after incision using ultrasonic vibrating scalpel and regular scalpel in the rat. Fertil Steril 61: 548550

News and notices

Surg Endosc (1998) 12: 11041107 Springer-Verlag New York Inc. 1998

New Address for the European Association for Endoscopic Surgery (E.A.E.S.)
Effective January 1, 1997, the correspondence, telephone, and fax numbers of the E.A.E.S. office are: E.A.E.S. Office, c/o Mrs. Ria Palmen Luchthavenweg 81 Unit 1.42 5657 EA Eindhoven The Netherlands or: P.O. Box 335 5500 AH Veldhoven The Netherlands Tel: +31 40 2525288 Fax: +31 40 2523102

laparoscopic solid organ surgery, transanal endoscopic microsurgery, and advanced techniques. Additionally, tailor-made courses are available for single surgeons or groups of surgeons to fit their requirements. Courses consist of didactic construction, review of videos, and experience in our superbly-equipped laboratory. We utilize training boxes, phantom abdomens as well as in vivo animal models. Course fees depend on length and complexity of instructions. In addition, attendees may observe live surgery and attend other educational activities at the George Washington University. For further details and brochure please contact: Debbie Moser Washington Institute of Surgical Endoscopy George Washington University Department of Surgery 2150 Pennsylvania Avenue, N.W. Suite 6-B Washington, DC 20037, USA Tel: 202-994-5441 Fax: 202-944-0567

Volunteer Surgeons Needed Northwestern Nicaragua Laparoscopic Surgery Teaching Program, Leon, Nicaragua
Volunteer surgeons are needed to tutor laparoscopic cholecystectomy for this non-profit collaboration between the Nicaraguan Ministry of Health, the National Autonomous University of Nicaragua, and Medical Training Worldwide. The program consists of tutoring general surgeons who have already undergone a basic laparoscopic cholecystectomy course. Medical Training Worldwide will provide donated equipment and supplies when needed. For further information, please contact: Medical Training Worldwide Ramon Berguer, MD, Chairman Tel: 707-423-5192 Fax: 707-423-7578 e-mail: berguer.r@martinez.va.gov

Fellowship in Minimally Invasive Surgery George Washington Medical Center Washington, DC USA
A one-year fellowship is being offered at the George Washington University Medical Center. Interested candidates will be exposed to a broad range of endosurgical Education and Research Center. Active participation in clinical and basic science research projects is also encouraged. For further information, please contact: Debbie Moser Tel: 202-994-8425 or, send curriculum vitae to: Dr. Jonathan M. Sackier Director, Washington Institute of Surgical Endoscopy George Washington University Medical Center Department of Surgery 2150 Pennsylvania Avenue, N.W. 6B-417 Washington, DC 20037, USA

Courses at the Washington Institute of Surgical Endoscopy

We are delighted to offer a variety of regular courses in laparoscopic techniques year round. In addition, special arrangements can be made for individual tuition in all minimally invasive disciplines. CME credit is available and course fees depend on the instruction offered. For further information, please contact: Debbie Moser Washington Institute of Surgical Endoscopy 2150 Pennsylvania Avenue, N.W. Suite 6-B Washington, DC 20037, USA Tel: 202-994-8425, or 1-888-8WISEDOC Fax: 202-994-0567

Essentials of Laparoscopic Surgery Surgical Skills Unit University of Dundee Scotland, UK

Under the direction of Sir A. Cuschieri the Surgical Skills Unit is offering a three-day practical course designed for surgeons who wish to undertake the procedures such as laparoscopic cholecystectomy. This intensely practical program develops the necessary operating skills, emphasizes safe practice, and highlights the common pitfalls and difficulties encountered when starting out. Each workshop has a maximum of 18 participants who will learn both camera and instrument-manipulation skills in a purposebuilt skills laboratory. During the course there is a live demonstration of a laparoscopic cholecystectomy. The unit has a large library of operative videos edited by Sir Cuschieri, and the latest books on endoscopic surgery are on display in our Resource area. Course fee including lunch and course materials is $860.

Courses at the Washington Institute of Surgical Endoscopy

Courses are available in difficult cholecystectomy and common duct exploration, laparoscopic antireflux surgery, laparoscopic colorectal surgery,

1105 For further details and a brochure please contact: Julie Struthers, Unit Co-ordinator Surgical Skills Unit Ninewells Hospital and Medical School Dundee DD1 9SY Tel: +44 382 645857 Fax: +44 382 646042 tailored to one or two procedures at the participants request. Practical skills are developed using training simulators and anaesthetised pigs. Course fees: $A300 ($US225) for the basic course and $A1,600 ($US1,200) for the advanced course. For further details and brochure, please contact: Dr. D. I. Watson or Professor G. G. Jamieson The Royal Adelaide Centre for Endoscopic Surgery Department of Surgery Royal Adelaide Hospital Adelaide SA 5000 Australia Tel: +61 8 224 5516 Fax: +61 8 232 3471

Advanced Endoscopic Skills Surgical Skills Unit University of Dundee Scotland, UK

Each month Sir A. Cuschieri Surgical Skills Unit offers a 412 day course in Advanced Endoscopic Skills. The course is intensely practical with hands on experience on a range of simulated models. The program is designed for experienced endoscopic surgeons and covers advanced dissection techniques, extracorporeal knotting techniques, needle control, suturing, internal tying technique, stapling, and anastomotic technique. Individual workstations and a maximum course number of 10 participants allows for personal tuition. The unit offers an extensive collection of surgical videos and the latest books and publications on endoscopic surgery. In addition, participating surgeons will have the opportunity to see live advanced laparoscopic and/or thoracoscopic procedures conducted by Sir Cuschieri and his team. The course is endorsed by SAGES. Course fee including lunch and course materials is $1850. For further details and a brochure please contact: Julie Struthers, Unit Co-ordinator Surgical Skills Unit Ninewells Hospital and Medical School Dundee DD1 9SY Tel: +44 382 645857 Fax: +44 382 646042

Advanced Laparoscopic Suturing and Surgical Skills Courses MOET Institute San Francisco, CA, USA
Courses are offered year-round by individual arrangement. The MOET Institute is accredited by the Accreditation Council for Continuing Medical Education (ACCME) to provide continuing medical education for physicians and designates these CME activities for 2040 credit hours in Category 1 of the Physicians Recognition Award of the American Medical Association. These programs are also endorsed by the Society of Gastrointestinal Endoscopic Surgeons (SAGES). For further information, please contact: Wanda Toy, Program Administrator Microsurgery & Operative Endoscopy Training (MOET) Institute 153 States Street San Francisco, CA 94114, USA Tel: (415) 626-3400 Fax: (415) 626-3444

The Practical Aspects of Laparoscopic Fundoplication Surgical Skills Unit University of Dundee Scotland, UK
A three-day course, led by Sir A. Cuschieri, designed for experienced laparoscopists wishing to include fundoplication in their practice. The course covers the technical details of total and partial fundoplication using small group format and personal tuition on detailed simulated models. There will be an opportunity to observe one of these procedures live during the course. Maximum course number is six. Course fee including lunch is $1850. For further details and a brochure please contact: Julie Struthers, Unit Co-ordinator Surgical Skills Unit Ninewells Hospital and Medical School Dundee DD1 9SY Tel: +44 382 645857 Fax: +44 382 646042

International Course in Laparoscopic Colorectal Surgery September 24, 1998 Trondheim, Norway Course Directors: R. Mrvik, MD, R. Bergamaschi, MD Host Chairman: H.E. Myrvold, MD Guest Faculty: S.D. Wexner, MD
For further information, please contact: National Center for Advanced Laparoscopic Surgery Trondheim University and Regional Hospital 7006 Trondheim, Norway Tel: +47-73-999888 Fax: +47-73-999889

Tenth International Conference of the Society for Minimally Invasive Therapy September 35, 1998 London, England
Host Chairman: Mr. J. Wickham For further information, please contact: The Society for Minimally Invasive Therapy 2nd Floor, New Guys House Guys Hospital St. Thomas Street London, SE1 9RT, England Tel: +44 (0)171 955 4478 Fax: +44 (0)171 955 4477 email: j.wickham@umds.ac.uk

Courses at the Royal Adelaide Centre for Endoscopic Surgery

Basic and Advanced Laparoscopic Skills Courses are conducted by the Royal Adelaide Centre for Endoscopic Surgery on a regular basis. The courses are limited to six places to maximize skill development and tuition. Basic courses are conducted over two days for trainees and surgeons seeking an introduction to laparoscopic cholecystectomy. Animal viscera in simulators is used to develop practical skills. Advanced courses are conducted over four days for surgeons already experienced in laparoscopic cholecystectomy who wish to undertake more advanced procedures. A wide range of procedures are included, although practical sessions can be

1106

Current Trends in Colon and Rectal Surgery September 2426, 1998 Naples, Italy
Current Trends in Colon and Rectal Surgery, sponsored by the University of Naples, George Washington University, and the Cleveland Clinic, will be held September 2426, 1998, at the Sorento Palace Conference Center in Naples, Italy. For further information, please contact: Office of Continuing Medical Education George Washington University Medical Center 2300 K Street, NW Washington, DC 20037, USA Tel: (202) 994-4285 or Dottore Vincenzo Landolfi Second University Universita BDI Napoli Cattedra di Chirurgia dell Apparoato DiGenrente Primary PoliclinicoPiazza Miraglia 3 80138 Napoli, Italy Tel: 011-39-81-566-5279 Fax: 011-39-81-459-137 e-mail: comvplan@syren

September 16, 1998, Pelvic Floor Workshop: Testing, Patient Simulation and Evaluation, will be held at Midway Health Service in St. Paul. The fee is $350 and accreditation is 6 hours in Category 1 credit towards the AMA Physicians Recognition Award. September 16, 1998, The 5th Annual Workshop, Molecular Biology of Colorectal Cancer, will be held at the University of Minnesota, Minneapolis campus. The fee is $175 and accreditation is 7 hours in Category 1 credit towards the AMA Physicians Recognition Award. September 1719, 1998, The 61st Annual Course, Principles of Colon and Rectal Surgery, will be held at the University of Minnesota, Minneapolis campus. The fee is $500 and accreditation is 21 hours in Category 1 credit towards the AMA Physicians Recognition Award. For further information contact Continuing Medical Education, University of Minnesota, Suite 107, 615 Washington Avenue SE, Minneapolis, MN 55414, USA Tel: (612) 626-7600 Toll free: 1-800-776-8636 Fax: (612) 626-7766 CME Home Page Location of the Internet: Address is: http://www.cee. umn.edu:80/cme/

Third Congress of the International Federation for the Surgery of Obesity (IFSO) September 35, 1998 Holiday Inn Crowne Plaza, Bruges, Belgium (12th International Symposium on Obesity Surgery)
International papers and symposia covering the field of laparoscopic and open bariatric surgery. For further information, please contact: Third Congress of the IFSO Secretariat Post Office Box 80 B-8310 Bruges Sint Kruix, Belgium or J.W.M. Greve, M.D. Fax +31-43-387-5473 email: jgreve@shbe.azm.nl or George S.M. Cowan, Jr., M.D. Professor of Surgery University of Tennessee 956 Court Avenue Suite A 212 Memphis, TN 38163, USA Tel: (901) 448-6781 Fax: (901) 448-4688 email: bpitts@utmem1.utmem.edu

Cours Europeen de Chirurgie Laparoscopique (European Course on Laparoscopic Surgery) under the auspices of E.A.E.S. University Hospital Saint-Pierre (U.L.B.) November 17November 20, 1998 Brussels, Belgium
Course Director: G.B. Cadiere, MD Universite Libre de Bruxelles (U.L.B.) Department of G.I. Surgery University Hospital Saint-Pierre The course will include live demonstrations and interactive dialogue with the operating surgeons and a video forum with discussions of videotapes, technical details, and pitfalls. Topics include functional gastric surgery (Nissen-Toupet-gastroplasty), colon (colectomy-rectopexy), hernia (trans-/preperitoneal approach, balloon), retroperitoneoscopy, splenectomy, needle surgery, biliary surgery, and new technologies. Surgeons include J. Bruyns, G.B. Cadiere, J. Himpens, J. Leroy, and M. Vertruyen. The official language for the May course is French with simultaneous translation provided into English. The official language for the November course is English with no simultaneous translation. Internet site: http://www.LAP-SURGERY.com For further information, contact: Scientific Information Mrs. Solange Izizaw C.H.U. Saint-Pierre Service de Chirurgie Digestive Rue Haute 322 B-1000 Bruxelles, Belgium email: coelio@resulb.ulb.ac.be or Administrative Secretariat Conference Services S.A. Avenue de lObservatoire 3, bte 17 B-1180 Bruxelles, Belgium email: conference.services@skynet.be

The Division of Colon and Rectal Surgery of the University of Minnesota Medical School will sponsor the following workshops and course this fall.
September 15, 1998, The 7th Annual Workshop, Endorectal Ultrasonography, will be held at Midway Health Service in St. Paul. The fee is $650 and accreditation is 7 hours in Category 1 credit towards the AMA Physicians Recognition Award.

1107

Thirteenth International Workshop on Therapeutic Endoscopy December 13, 1998 Hong Kong
The Chinese University of Hong Kong and the Hong Kong Society of Digestive Endoscopy will hold the 13th International Workshop on Therapeutic Endoscopy on December 13, 1998. The workshop, intended for experienced endoscopists interested in endoscopic therapy, consists of 3 days of live demonstration of advanced techniques of therapeutic endoscopy. There will also be poster presentations by the participants. This year we have invited Peter Cotton (USA), Naotaka Fujita (Japan), K.L. Goh (Malaysia), Yoshiki Hiki (Japan), Kees Huibregtse (The Netherlands), Joseph Leung (USA), Charles Lightdale (USA), and Jose Sollano (Philippines) to be our international faculty. For further information, please write to: Prof. Sydney Chung Endoscopy Centre Prince of Wales Hospital The Chinese University of Hong Kong Shatin, N.T., Hong Kong Tel: (852) 2632-2233 Fax: (852) 2635-0075

Organizing Committee: Arbeitsgruppe fu r Endoskopische Chirurgie. For further information, contact: Congress Secretariat Netzibodenstrasse 34 Post Office Box 1527 4133 Pratteln, Switzerland. Tel: ++41 61 811 47 70 Fax: ++41 61 811 47 75

Call for Abstracts SAGES 1999 Annual Meeting Society of American Gastrointestinal Endoscopic Surgeons March 2427, 1999 San Antonio, TX, USA
Abstract Deadline: Oral and Poster Abstracts, September 11, 1998; video submissions: September 18, 1998. For further information or to obtain an abstract form, contact: SAGES Program Committee Society of American Gastrointestinal Endoscopic Surgeons Suite 3000 2716 Ocean Park Boulevard Los Angeles, CA 90405, USA Tel: (310) 314-2404 Fax: (310) 314-2585 email: SAGESMail@AOL.com

7th Congress of the European Society for Gynaecological Endoscopy December 69, 1998 Beaulieu Convention Centre Lausanne, Switzerland

Surg Endosc (1998) 12: 10391042

Springer-Verlag New York Inc. 1998

Port site metastases and recurrence after laparoscopic colectomy

A randomized trial
A. M. Lacy,1 S. Delgado,1 J. C. Garc a-Valdecasas,1 A. Castells,2 J. M. Pique ,2 L. Grande,1 J. Fuster,1 1 1 1 E. M. Targarona, M. Pera, J. Visa
1 2

Department of Surgery, Hospital Clinic, University of Barcelona, Villarroel 170, 08036 Barcelona, Spain Department of Gastroenterology, Hospital Clinic, University of Barcelona, Villarroel 170, 08036 Barcelona, Spain

Received: 4 November 1997/Accepted: 12 January 1998

Abstract Background: This study was performed to prospectively assess the impact of the laparoscopic approach to the patterns of port site metastases (PSM) and recurrence rate (RR) of resected colon carcinomas as compared with conventional colectomies. Methods: All patients were included in a prospective randomized trial comparing laparoscopic-assisted colectomy (LAC) versus open colectomy (OC) for colon cancer. The randomization was stratified for localization of the lesion. Patients with metastasic disease at the time of the surgery were excluded. Follow-up in the outpatient clinic was done every 3 months for a minimum of 12 months. Endpoints for the study were metastasis at port site and laparotomy incision as well as recurrence rate. Results: Of 91 segmental colectomies performed from November 1993 to January 1996, there were 44 LAC and 47 OC. Patient data were similar in both groups (age, sex, Dukes stage, type of operation). Mean follow-up was 21.4 months, with a range of 13 to 41 months. There were no wounds or PSM in those series. RR was similar for both groups. For LAC, it was five of 31 (16.1%); for OC, it was six of 40 (15%). Conclusions: The laparoscopic approach has a recurrence rate similar to that for open procedures for colon cancer. However, additional follow-up of these patients is needed before we can determine whether or not the laparoscopic approach influences overall survival. Key words: Laparoscopic surgery Colorectal surgery Colon cancer Recurrence Colectomy Laparoscopic colorectal surgery

The recurrence rate in colorectal surgical patients who undergo curative resection is 2030%; >90% of recurrences are detected within the 3 years after curative resection [4]. Laparoscopic-assisted colectomy is still a controversial operation in cases of malignant disease. Since the first reports appeared concerning expanding indications for the use of the minimal access surgery, numerous cases of tumor recurrence at the trocar sites after laparoscopic resections have been reported. Although follow-up on such cases has been limited, patterns of recurrence from previous studies indicate that 80% of recurrences occur within 1 year. Interval time to port site metastases (PSM) and recurrence rate (RR) after resection of colon cancer is variable. The incidence of PSM seems to be higher in laparoscopic operations, but the interval to wound recurrence is <12 months. Moreover, recurrence is seen during the first 2 years after surgery, with a peak at 612 months. The aim of this study was to analyze the impact of the laparoscopic approach to the patterns of PSM and RR of resected colon carcinomas and compare our findings with those for conventional colectomies.

Materials and methods

This study began in November 1993 and was continued through January 1996; it included 91 patients from a single institution (The Hospital Clinic of the University of Barcelona, Spain). All patients were included in a prospective randomized trial comparing laparoscopic-assisted colectomy (LAC) to open colectomy (OC) for colon cancer. Patients were fully informed of the investigative nature of the procedure and known controversies. The inclusion criteria, preoperative evaluations and preparations, and operative procedures have been previously published [11]. The randomization was stratified for localization of the lesion. Patients with metastasic disease (Dukes D2 or stage IV patients) at the time of the surgery were excluded from this study. The open and laparoscopic groups were under the supervision of the same surgical team at the same hospital. All patients included in Dukes B2 or C were included in the same postoperative chemotherapy protocol (5-fluoracil plus levamisole). Follow-up in the outpatient clinic was done by the attending surgeon in

Presented at the annual meeting of the Society of American Gastrointestinal Endoscopic Surgeons (SAGES), Philadelphia, Pennsylvania, USA, 1317 March 1996 Correspondence to: A. M. Lacy

1040 Table 1. Demographic data LAC No. patients Sex (M/F) Age (yr) Follow-up (mo) (mean SD) OC Table 2. Pathologic stages according to the Astler-Coller classification Dukes stage A B1 B2 C1 C2 D1 LAC 4 2 12 0 8 5 OC 2 3 21 2 10 2

31 40 24/20 23/24 71.2 9.2 69.7 10.2 21.4 11.5 (range, 1341)

LAC, laparoscopic-assisted colectomy; OC, open colectomy. LAC, laparoscopic-assisted colectomy; OC, open colectomy.

each group. The open group was followed closely according to a standard format every 3 months for a minimum of 12 months. The LAC group was evaluated more frequently (once a month) to detect possible trocar site implants. Survival data were calculated according to the following categories: alive with no evidence of the disease, alive with evidence of disease, or decreased (in relation or not with the colon carcinoma). Endpoints for the study were metastasis at port site and laparotomy incision as well as recurrences rates (local or distant). Examination of these data revealed comparable types and staging of resections in each group. However, Dukes D patients were more numerous in the LAC group. Patients suspected of adjacent organ invasion during the laparoscopic approach were converted immediately to an open approach. Local or regional tumor recurrence was defined as a tumor recurrence in the previous operative field after a presumed curative resection. Systemic or distant metastasis included liver, lung, disseminated peritoneal metastasis and other metastases, such as bone or retroperitoneal metastasis.

shows the incidence of recurrence and mortality in relation to the initial stage at operation.

Discussion Surgery continues to be the best therapeutical option for colorectal cancer. Adenocarcinoma of the colon is associated with significantly more hepatic and intraabdominal recurrences, whereas carcinoma of the rectum has more locoregional and pulmonary recurrences. Advances beyond exteriorization began to appear at the end of the 19th century. The antibiotic era brought more advances. Dogma abounds with respect to the technical aspects of surgery for colon, and rectal cancer (high vascular ligation) should be evaluated by controlled trials. However, few prospective and randomized trials have been done to evaluate the importance of these techniques. Firmly established techniques include resection of lymphatic drainage of tumors, en bloc resection of invaded structures, and obtaining 2-cm margins of rectal cancers. One recent paper discussed such other methods as radical lymph node dissection, luminal ligation, oophorectomy, and the no-touch technique [10]. Despite the paucity of irrefutable scientific data to support many of the described surgical techniques, differences in outcome between surgeons suggest that technique is important. There is a great need for randomized prospective trials to evaluate the multitude of techniques now in use for the surgical treatment of colorectal cancer. The application of laparoscopy to the field of general surgery has brought a new and different approach to old intraabdominal pathologies. Minimal access surgery has been extended to a wide variety of diseases, including colorectal malignancies. Several studies have demonstrated the feasibility of laparoscopic colectomy. However, concern still exists about the oncological safety of treating malignancies by laparoscopy, even though several studies have shown that this approach does not compromise the number of lymph nodes resected or the margins of resection, as compared with the standard open procedure [2, 5, 6]. Therefore it may be the unknown recurrence rate and the absence of a critical assessment of the balance of risks versus benefits that limits the widespread use of the laparoscopic approach in malignant colorectal disease. The American Society of Colon and Rectal Surgeons and the Society of American Endoscopic Surgeons have suggested that laparoscopic surgery for colorectal malignancies should be reconsidered in the light of appropriate prospective randomized studies. One of the questions about the validity of laparoscopic

Results Of 91 segmental colectomies, 20 cases were excluded due to distant metastases (liver and/or lung). Of the 71 patients left for analysis of results, 31 patients were allocated to the LAC group and 40 patients to the OC group. The demographic data and staging of tumors are comparable for both groups (Table 1). Mean follow-up was 21.4 11.5 months, with a range of 1341 months. Postoperative staging of the tumor was done following the Astler-Coller classification (Table 2). There were no wound or port site recurrences in either group. The recurrence rate was similar in both groups. For LAC, it was five of 31 (16.1%); for OC, it was six of 40 (15%). Deaths occurred in nine of 11 patients (81.8%) who had recurrences. All deaths in the follow-up period were cancer related. In the open group, four patients died at 7, 9, 10, and 24 months, respectively, after resection. Two patients (two of six, or 33%) of this group with Dukes B2 tumor died at 7 and 24 months, respectively, postoperatively with local and distant recurrence. The other two died (two of six, or 33%) with distant disease only; they were classified as Dukes stages C2 and D1. Two patients (two of six, or 33%) were alive with no evidence of the disease after resection of the local recurrence at the anastomosis level. In the laparoscopic group, all patients (five) with recurrence died in the follow-up period at 7, 10, 14, 14, and 16 months, respectively. All of them (100%) died with distant metastases; four were classified as Dukes D1 and one as C2. The four patients who were included in the Dukes D1 stage were started laparoscopically and converted immediately to an open procedure when we noted that the tumors had invaded other organs (as described by our protocol). Table 3

1041 Table 3. Recurrence and mortality rates Laparoscopic colectomy Classification Stage Stage Stage Stage Stage Stage A B1 B2 C1 C2 D1
c

Open colectomy Death

b

No. 0 0 0 0 1 (DR) 4 (DR)

Time

No. 0 1 2 0 2 1 (LR) (LR + DR) (LR; DR) (DR)

Timea 13 6, 21 6, 9 9

Deathb alive 7, 24 9, alive 10

9 3, 8, 8, 15

14 7, 10, 14, 16

LR, local recurrence; DR, distal recurrence. a Time of recurrence in months. b Time of death in months. c Classification of Dukese modified by Astler-Coller.

treatment for colorectal cancer is the recurrence rate after curative resections, since the pattern of recurrence after laparoscopic treatment is different. Moreover, a number of case reports indicate that laparoscopic resection of colon carcinoma may alter both the pattern and incidence of cancer recurrence and survival. Most of the reports dealing with colorectal open surgery have not mentioned recurrence at the abdominal wall or the scar. The precise incidence of port site metastases after laparoscopic colorectal resection for cancer is still unknown. Most of the initial reports must be considered anecdotal, and the metastases are probably due to inappropriate surgical technique. However, a review of the literature by Wexner and Cohen showed that by 1994, 33 patients had been reported with PSM after LAC for colorectal carcinoma [15]. The incidence of port site implantation following LAC has been variously estimated at from 0 to 21% [3, 15, 16]. A paper from the previous laparoscopic era, published by Hughes et al. [8] concerning abdominal wall recurrence after conventional surgery drew new attention because if found a recurrence of 1%. Although the length of followup is still limited in laparoscopic surgery, previous studies have suggested that 80% of recurrences occur within 1 year. Furthermore, a significant and increasing number of recurrences at the port-site have been reported during the last years, with a total incidence of 4%. Nevertheless, it is not yet clear whether abdominal or wall recurrence is a peculiar complication of laparoscopy, or whether it was simply underreported in the series of standard open procedures. In November 1993, the ethical committee of our institution approved a plan to include patients with colon carcinoma in the first randomized protocol comparing laparoscopic to open colon resections reported in the literature [11]. Our preliminary results of RR and PSM in patients with a follow-up period of >1 year are similar to those of several other large series of laparoscopic-assisted resections, which have reported some recurrence and no port site metastases at all [12, 14]. After curative resection of colorectal carcinoma, recurrence rates range 570%, depending on the site and stage of the primary tumor. Recent series found isolated local recurrence in 119% of patients, local and systemic recurrence in 516%, and systemic disease alone in 1222%. In our study, recurrence was similar in both groups (15%). Anastomotic recurrence represents a 35% risk. The condition may have originated outside the intestinal wall

(regional recurrence), with tumor seeding and spreading inward to reach the mucosal surface. Some authors have suggested that the laparoscopic resection of colon cancer has more anastomotic recurrences than conventional surgery. However, in our study, patients operated upon using the laparoscopic approach did not have anastomotic recurrences. Moreover, in two patients included in the open group, anastomotic recurrence was diagnosed as regional recurrence at relaparotomy. Another important prognostic factor is that the incidence of local recurrence from colorectal cancer correlates directly with the initial stage at operation. In the LAC group, all patients with recurrence were staged as Dukes C2 or D1 at operation. However, if we exclude patients with Dukes D1 stage, only one patient had recurrent disease (disseminated metastases) in the laparoscopic group. By contrast, five patients in the open group had recurrences. These results are not statistically significant; however, the inclusion of more patients should confirm that the laparoscopic approach is no different from the conventional one. The pathophysiologic mechanisms involved in neoplastic recurrences after laparoscopy are unknown. Since the first reported case of a port site metastasis by Alexander et al. in 1993 [1], this complication and the recurrence pattern after LAC has been subjected to an intense investigation. Several factors appear to be related to this worrisome complication. Surgical trauma, which is a major factor in oncological surgery, has been related to the recurrence pattern after both laparoscopic and open treatment of malignancies [13]. Surgery itself can promote neoplastic metastasespossibly secondary to the hematogenous spread of neoplastic cells or by direct implantation during the procedure. Therefore, we must question whether there are specific factors related to laparoscopy that cause the development of recurrences at trocar sites. Nduka et al. [13] have identified three factors that may lead to an increased implantation rate: exfoliation of malignant cells following excessive manipulation by laparoscopic instruments, increased contact between the malignant cells and skin incisions, and the presence of pneumoperitoneum. According to Jones et al. [9], insufflation of the abdominal cavity with carbon dioxide causes enhanced tumor uptake at the laparotomy incision and trocar sites, since the pressure gradient necessary for the creation of pneumoperitoneum seems to be sufficient for seeding tumor cells into port sites

1042

when carbon dioxide leaks around the trocars. Furthermore, in a recent experimental study, Hubens et al. [7] suggested that implantation occurs when free cancer cells, which circulate in the peritoneal cavity due to the pneumoperitoneum, are forced outward through the trocars at the moment of deflation, when the trocars are withdrawn. Nevertheless, it is still too early to draw any definitive conclusions about the ultimate cause of port site metastasis following laparoscopic resection of colon malignancies. Our preliminary data are not in agreement with studies that have shown an increased recurrence rate after laparoscopic resection of colon carcinoma. However, our experience is still rather limited. The definitive results of our investigation, which should be available in the near future, may well reveal different patterns of recurrence. Although the technique yields an adequate tissue resection, concern about trocar site tumor recurrences has led to a general consensus that the laparoscopic procedure should be done in a prospective investigational protocol setting for the treatment of curable carcinoma. References
1. Alexander RJT, Jaques BC, Mitchell KG (1993) Laparoscopically assisted colectomy and wound recurrence [letter]. Lancet 341: 249250 2. Falk PM, Beart RW, Wexner SD, Thorson AG, Jagelman DG, Lavery IC, Johansen OB, Fitzgibbons RJ (1993) Laparoscopic colectomy: a critical appraisal. Dis Colon Rectum 36: 2834 3. Franklin M, Rosenthal D, Dorman J, Glass J, Norem R, Diaz A (1996) Prospective comparison of open vs. laparoscopic colon surgery for carcinoma: five year results. Dis Colon Rectum 39: S3546 4. Hida J, Yasutomi M, Shindoh K, Kitaoka M, Fujimoto K, Leda S,

5. 6.

7.

8.

9.

10. 11.

12.

13. 14. 15. 16.

Machidera N, Kubo R, Morikawa E, Inufusa H, Watatani M, Okuno K (1996) Second-look operation for recurrent colorectal cancer based on carcinoembryonic antigen and imaging techniques. Dis Colon Rectum 39: 7479 Hoffman G, Baker J, Fitchett C, Vansant J (1994) Laparoscopicassisted colectomy. Initial experience. Ann Surg 219: 732743 Hoffman G, Baker J, Doxey J, Hubbard G, Ruffin W, Wishner J (1996) Minimally invasive surgery for colorectal cancer. Initial follow-up. Ann Surg 223: 790798 Hubens G, Pauwels M, Hubens A, Vermeulen P, Van Marck E, Eyskens E (1996) The influence of a pneumoperitoneum on the peritoneal implantation of free intraperitoneal colon cancer cells. Surg Endosc 10: 809812 Hughes E, McDermott F, Polglase A, Johnson W (1983) Tumor recurrence in the abdominal wall scar tissue after large-bowel cancer surgery. Dis Colon Rectum 26: 571572 Jones D, Guo W, MK R (1995) Impact of pneumoperitoneum on trocar site implantation of colon cancer in hamster model. Dis Colon Rectum 38: 11821188 Ko hler L, Eypasch E, Paul A, Troidl H (1997) Myths in management of colorectal malignancy. Br J Surg 84: 248251 Lacy AM, Garc a Valdecasas J, Pique J, Delgedo S, Campo E, Bondas JM, Taur P, Graude L, Fuster J, Pacheco JL, Visa J (1995) Short-term outcome analysis of a randomized study comparing laparoscopic vs open colectomy for colon cancer. Surg Endosc 9: 11011105 Lord S, Larach S, Ferrara A (1996) Laparoscopic resection for colorectal carcinomaa three year experience. Dis Colon Rectum 39: 148154 Nduka C, Monson T, Menzies-Gow N, Darzi A (1994) Abdominal wall metastases following laparoscopy. Br J Surg 81: 648652 Ortega A, Beart R, Steele G (1995) Laparoscopic bowel surgery registry: preliminary results. Dis Colon Rectum 38: 681686 Wexner S, Cohen S (1995) Port site metastases after laparoscopic colorectal surgery for cure of malignancy. Br J Surg 82: 295298 Whelan R, Sellars G, Allendorf J, Laird D, Bessler MD, Nowygrod R, Treat MR (1996) Trocar site recurrence is unlikely to result from aerosolization of tumor cells. Dis Colon Rectum 39: S713

Surg Endosc (1998) 12: 10171019

Springer-Verlag New York Inc. 1998

Chemical composition of smoke produced by high-frequency electrosurgery in a closed gaseous environment

An in vitro study
C. Hensman,1 D. Baty,2 R. G. Willis,2 A. Cuschieri1
1

Department of Surgery and Surgical Skills Unit, Ninewells Hospital and Medical School, University of Dundee, Dundee, Tayside DD1 9SY, Scotland, United Kingdom 2 Department of Molecular and Cellular Pathology, Ninewells Hospital and Medical School, University of Dundee, Dundee, Tayside DD1 9SY, Scotland, United Kingdom Received: 23 September 1997/Accepted: 3 December 1997

Abstract Background: High-frequency (HF) electrocoagulation and cutting procedures produce smoke by high-temperature pyrolysis of tissues. As distinct from the experience of conventional surgery, electrosurgical smoke is produced in a closed gaseous environment during laparoscopic operations. As a result, toxic chemicals may be absorbed into the circulation. The effects of this absorption are not known. Furthermore, the chemical composition of electrosurgical smoke produced in an anoxic environment may be different from that produced in air. Methods: Smoke was produced in vitro by HF electrocutting of fresh porcine liver in helium, CO2, and air-saturated closed environments. Smoke samples were collected and analyzed by gas chromatographymass spectrometry (GCMS). Results: The chemical constituents of electrosurgical smoke produced in air, CO2, and helium were similar. To date, 21 chemicals, some highly toxic, have been identified in the electrosurgical smoke produced in a closed environment. These consist of hydrocarbons, nitriles, fatty acids, and phenols. Conclusions: Electrosurgical smoke produced in a closed environment contains several toxic chemicals. The effects of these on cell viability, macrophage, and endothelial cell activation are not known but are being investigated. Meantime, measures to reduce smoke and evacuate it during endoscopic surgery are advisable. Key words: High-frequency electrocoagulation Electrosurgical smoke

The use of high-frequency (HF) electrosurgery is standard practice both in conventional and minimal access surgery (MAS). Concerns about the electrical hazards [1] of HF electrosurgery have overshadowed the risks associated with the generation of smoke from high-temperature pyrolysis of tissues. Electrosurgical smoke contains charred debris, cellsize fragments, breathable aerosols, and various complex organic chemicals [2]. The toxic, infective, and mutagenic risks of electrosurgical smoke generated during conventional open surgery are well documented [3, 7]. Use of smoke evacuation devices and safety precautions against inhalation by theater staff have been advocated during open surgery [6]. The situation is different in MAS because smoke is produced in a closed CO2 environment from which it is evacuated by standard suction devices. Two considerations arise from this situation. First, electrosurgical smoke produced in a CO2 environment may have a different composition. Second, toxic chemicals generated inside the peritoneal or thoracic cavities are likely to be absorbed into the circulation with systemic effects [8]. Materials and methods Apparatus and smoke production
Electrosurgical smoke was produced in vitro by the application of HF current to fresh porcine liver (100 g) placed in an aluminium foil dish inside a sealed chamber consisting of a SEP-T-VAC 1.2 suction canister (Sherwood Medical, Tullamore, Ireland). The electrical energy was delivered by a microprocessor-controlled generator (ValleyLab Force FX, Boulder, Colorado, USA) connected to an insulated electrosurgical hook knife (Storz, Tuttlingen, Germany) used to cut the liver tissue during a 15-min period at a constant power setting of 40 W. Within the chamber, electrosurgical smoke was produced under three atmospheric conditions: air, helium and CO2. These gases were introduced into the container via a regulator valve and flow meter to maintain a

Correspondence to: A. Cuschieri

1018 Table 1. List of chemicals identified in electrosurgical smoke by GC-MS Hydrocarbons 2,3-Dihydro indene 1 Decene 1 Undecene Ethynyl benzene Ethyl benzene Toluene Nitriles 3-Butenenitrile Benzonitrile 2-Propylene nitrile Amines Pyrrole 6-Methyl indole Indole Aldehydes 2-Methyl propanol 3-Methyl butenal Furfural Benzaldehyde Miscellaneous 2-Methyl furan2,5-dimethyl furan Methyl pyrazine Hexadecanoic acid 4-methyl phenol

Packard 5970, Palo Alto, USA). The gases released were separated by chromatography, and their chemical constitution was investigated by obtaining their mass spectra. Compound identification was achieved by library mass spectral data comparison with data obtained from the gaseous samples. When a preliminary identification had been made, pure samples of those compounds thought to be present in the smoke were analyzed to obtain reference chromatographic and mass spectral data under the same conditions.

Results The chromatographic traces of the compounds desorbed from electrosurgical smoke samples were very complex. The majority of peaks appeared in the trace during the first 15 min of analysis. The computerized spectral matching system generated more than 35 candidate compounds as possibly present in the smoke. This was based on 80% chemical matching criteria. By comparing the chromatographic and mass spectral properties of the reference compounds with those found in the electrosurgical smoke samples, we were able positively to identify 21 compounds as being present in the smoke produced during the experiments. The chemical constituents identified in smoke produced in air, CO2, and helium were similar in composition (Table 1). The compounds identified to date consist of hydrocarbons, nitriles, fatty acids, and phenols. Other chemicals also thought to be present are currently undergoing final identification.

Fig. 1. Schematic representation of apparatus used for smoke production in vitro. A, gas cylinder (helium, CO2, air); B, flow regulator; C, 1.5-l container (Sep T-Vac, Sherwood Tullamore, Ireland); D, electrosurgical generator (Valley Lab Force FX, Boulder, CO, USA); E, analytical tube containing adsorbent; F, tap; G, hook instrument (Storz Tuttlingen, Germany).

constant flow rate of 2.0 l/min. Samples of smoke generated in each of these three environments were collected and passed across adsorbent tubes for analysis by GC-MS (Fig. 1).

Sample collection
Before use, Tenax adsorbent tubes (Pampisford, Cambridge, UK) were heat cleaned in a helium atmosphere at 250C and then checked for contaminants by being taken through the analytical procedure. Care was taken not to handle the sample tubes directly. They were picked up with forceps and paper tissues. Once a sufficient flow of gas from the smoke chamber was established, the Teflon-sealed ends of a Tenax adsorbent tube were pierced, and the tube was then attached to the outlet of the smoke chamber by a short length of silicon tubing. The Teflon tape left on the adsorbent tube prevented contact with the silicon tubing and contamination of the ends of the adsorbent tube. Flow through the adsorbent tube was induced by partially closing the valve (F). With a total gas flow of 2.0 l/min through the smoke chamber, the flow through the adsorbent tube amounted to 300 ml/min. The adsorbent tube was then detached from the system, and the ends were sealed with Teflon tape before it was stored in tightly capped tubes at 4C before GC-MS analysis. Control samples were obtained in a similar fashion from the air or gas (helium, CO2) perfused chamber in the absence of smoke generation.

Discussion This study identified 21 highly toxic and carcinogenic chemicals in smoke produced by electrocutting of porcine liver in a closed environment. Some of these chemicals are similar to those reported after high-temperature pyrolysis of meat [5]. The results demonstrate that the atmosphere in which the smoke is produced (air vs. CO2 or helium) does not affect the nature of the toxic chemicals produced. However, the absorption of these toxic chemicals via the serosal membranes into both the portal and systemic circulation is a potential hazard that has been largely overlooked. Such absorption has been documented in a report of increased carboxyhaemoglobin and methaemoglobin levels in patients after prolonged laparoscopic operations [8]. The absorption of these chemicals in electrosurgical smoke may be enhanced by the vasodilatory effect of CO2 on the serosal

Gas chromatographymass spectrometry analysis

Tenax adsorbent tubes previously exposed to electrosurgical smoke were desorbed by heat into a gas chromatographymass spectrometer (Hewlett

1019

vasculature and the positive pressure pneumoperitoneum used in laparoscopic surgery. We have demonstrated in other experiments that electrosurgical smoke in a 1/20 dilution produced in the manner just described is cytotoxic to cultured cells [4]. The effects of high electrosurgical smoke dilutions (1/1001/1000) on the cellular components of the immune system and vascular endothelium are currently being investigated. The smoke problem encountered in MAS can be addressed in two ways. The first concerns the use of microprocessor-controlled electrosurgical generators that by sensor electronic feedback from the tissueelectrosurgical probe interface adjust the power output necessary to achieve the intended task [1] with minimal charring and smoke production. The second measure, currently unavailable, consists of an endoscopic smoke detection and evacuation system linked electronically to the CO2 insufflator to ensure rapid evacuation of any generated smoke without loss of the pneumoperitoneum. Such a system would have to collect the smoke-filled gas in a sealed container to avoid contamination of the operating theater environment.

References
1. Cuschieri A, Haag R (1993) Recent advances in high-frequency electrosurgery: development of automated systems. J Roy Coll Surg Edinb 38: 354364 2. DesCouteaux JG, Picard P, Poulin C, Baril M (1996) Preliminary study of electrocautery smoke particles produced in vitro and during laparoscopic procedures. Surg Endosc 19: 152158 3. Gatti JE, Bryant CJ, Noone RB, Murphy JB (1992) The mutagenicity of electrocautery smoke. Plastic Reconstr Surg 89: 781784 4. Hensman C, Newman EL, Shimi S, Cuschieri A (199?) The cytotoxicity of electrosurgical smoke produced in an anoxic environment. Am J Surg (in press) 5. Ott DE (1993) Smoke production and smoke reduction in endoscopic surgery: preliminary report. Endosc Surg Allied Tech 1: 230233 6. Recommended practices for endoscopic minimal access surgery: AORN standards and recommended practices for perioperative nursing. Association of Operating Room Nurses, Denver, Colorado, 1995, III, pp 169174 7. Sagar PM, Meagher S, Sobczak S, Wolff BG (1996) Chemical composition and potential hazards of electrocautery smoke. Br J Surg 83: 17921793 8. Thiebaud HP, Knize MG, Kuzminicky PA, Felton JS, Hsieh DPJ (1994) Mutagenicity and chemical analysis of fumes from cooking meat. Agric Food Chem 42: 15021510