Peritoneal Drainage as Primary Management of Perforated NEC in

the Very Low Birth Weight Infant

By Lori J. Morgan, Stephen J. Shochat, and Gary E. Hartman
Stanford, California

@Advances in perinatal and neonatal care in the past tion for a formal laparotomy. In the reported experi-
decade have produced a change in the population of infants
ence with peritoneal drainage,2-4 a significant number
with perforated necrotizing enterocolitis (NEC) treated at our
institution: the majority are now of very low birth weight
of infants (27% to 46%) have had complete resolu-
(VLBW, <l,OOO g). Peritoneal drainage has been reported as tion of NEC perforation with drainage as the only
an initial resuscitative procedure for unstable infants who operative intervention. The purpose of this report is
have complicated NEC. Initial success with peritoneal drain- to review the results of an aggressive use of peritoneal
age prompted us to adopt an aggressive approach to its use
drainage as the definitive operative procedure in
in this patient population. Since 1987, peritoneal drainage
has been the primary treatment for most infants weighing
infants with complicated NEC.
less than 1,500 g who have perforation, and for unstable
infants weighing more than 1,500 g. Perforation was docu-
MATERIALS AND METHODS
mented by pneumoperitoneum or aspiration of meconium by
paracentesis. Intestinal resection was performed in most The records of 49 consecutively treated infants requiring opera-
infants weighing more than 1,500 g and in those for whom tive treatment of complicated NEC were reviewed; the infants had
drainage was ineffective. Twenty-nine infants with low or been treated at the Stanford University Medical Center between
VLBW (mean gestational age, 27 weeks; mean birth weight, 1984 and 1992. The mean birth weight was 1,350 g and the mean
994 g) were treated with one or two drains in the right lower gestational age was 29 weeks. One infant had previously undergone
quadrant. Broad spectrum antibiotics were continued until laparotomy with primary anastomosis at another institution. All
all drains were removed, usually within 10 to 14 days. others underwent surgical therapy at our Center. Twenty-nine
Nasogastric suction was continued until patency of the infants were treated with initial peritoneal drainage under local
gastrointestinal (GI) tract was confirmed by a nonionic upper anesthesia; the other 20 were treated with initial laparotomy. The
GI series. Six (21%) infants died, although one of the deaths hospital records were reviewed, and the following data were
occurred 5 months after drainage; the patient had chronic abstracted: (1) birth weight, (2) gestational age, (3) age at
lung disease and an intact GI tract. Seventeen of the 23 (74%) diagnosis, (4) age at operation, (5) indication for operation, (6)
survivors required no further operative procedure, and 6 type of operation, (7) amount of blood required during operation
(26%) required laparotomy and resection because drainage and subsequent 24 hours, (8) mean airway pressure and (9)
had been ineffective. Peritoneal drainage provided definitive inspired oxygen concentration at operation, (10) repeat operative
treatment in 18 of 29 (62%) infants in this series. The low procedure, (11) duration ofventilation. (12) duration of parenteral
mortality rate and the successful treatment without lapa- nutrition, (13) age at achieving full oral feeding, (14) length of
rotomy in nearly two thirds of the infants suggest that bowel remaining, (15) oxygen use at discharge, and (16) survival.
peritoneal and systemic host defenses and wound healing Associated conditions such as patent ductus arteriosus (PDA),
are significantly different in the VLBW infant. These differ- retinopathy of prematurity (ROP), and neurological injury were
ences indicate the need to reevaluate treatment strategies also recorded.
for this patient population. Peritoneal drainage was the treatment of choice for infants
Copyright o 1994 by W.B. Saunders Company weighing less than 1,500 g and for selected infants weighing more
than 1,500 g who were hemodynamically unstable (requiring
INDEX WORDS: Necrotizing enterocolitis, peritoneal drain- inotropic support) or unstable (requiring frequent hand bagging or
age. ventilatory manipulation) despite maximal ventilatory support.
Drainage was performed with muscle relaxation and local anesthe-

T
sia in the intensive care nursery. A right lower quadrant incision
HE ROLE of peritoneal drainage in the manage- was used unless an air/fluid cavity was identified radiographically
ment of infants with complicated necrotizing in another quadrant. Gas and fluid were evacuated, and the
enterocolitis (NEC) remains undefined. Since the peritoneal cavity was occasionally irrigated with 10 to 20 mL of
initial report by Ein et al, this technique has been normal saline. One or two /J-inch Penrose drains were inserted and
used to stabilize low birth weight infants in prepara- managed with sterile technique. The drains were gradually re-
moved with 7 to 10 days unless meconium or significant purulent
material was still draining. A second drainage procedure was
performed in seven infants because of inadvertent drain removal
From the Division of Pediattic Surgery, Department of Surgery, (2) reaccumulation of pneumoperitoneum (2) development of
Stanford University Medical Center, Stanford, CA. enterocutaneous fistula (2) or evidence of intraabdominal abscess
Presented at the 24th Annual Meeting of the American Pediatric by ultrasound (1).
Surgical Association, Hilton Head, South Carolina, May 15-18, 1993. A formal laparotomy was performed in the early phase of the
Address reprint requests to Gaty E. Hartman. MD, Division of study (1984 to 1987) as the initial procedure of choice. Subse-
Pediattic Surgery Department of Surgery, Stanford University Medical quently (1987 to 1992) laparotomy was reserved for infants who
Center, Stanford, CA 94305. weighed more than 1.500 g and those for whom peritoneal drainage
Copyright o 1994 by W.B. Saunders Company was ineffective. Determination of failure of peritoneal drainage
0022-346819412902-0033$03.00/O was based on reaccumulation of pneumoperitoneum or develop-

310 JournalofPediatricSurgery, Vol29,No 2 (February),1994:

pp 310-315
PERITONEAL DRAINAGE AND NEC 311

ment of intestinal obstruction or enterocutaneous fistula after one Table 1. Postoperative Deaths
(3 infants) or more (3 infants) drainage procedures. Birth Time of
All laparotomies were performed in the operating room, with Weight Death Cause of
the infants under general anesthesia. Perforated and necrotic (9) Procedure (POD) Death

bowel was resected, and a proximal enterostomy and mucous 820 Drain 166 Respiratory
fistula were created at the lateral margins of the incision. For 1,150 Drain 1 Sepsis
infants with small bowel enterostomies, closure was performed 550 Drain 25 Candida sepsis
before discharge from the intensive care nursery. Those with 922 Drain 11 Candida sepsis
colostomies underwent closure at approximately 1 year of age. 682 Drain 3 Sepsis
Survival was determined at the time of examination; the fol- 960 Drain 25 Candida sepsis
low-up period ranged from 4 months to 9 years (Fig 1). Comparison 610 Laparotomy 1 Sepsis
between the laparotomy group and the drainage group, as well as 810 Laparotomy 62 Sepsis
comparison between infants treated with drainage alone and those
treated with intial drainage followed by laparotomy, was per- Abbreviation: POD, postoperative day.

formed using the two-tailed t test; a P value of < .05 was considered
significant.
RESULTS
Peritoneal Drainage
Twenty-nine infants underwent peritoneal drain-
age. There were six (21%) deaths (Table 1). Twenty-
six infants had a birth weight of less than 1,500 g. Two
of the deaths were within 72 hours of drainage; one of
these patients had no perforation but had diffuse
mucosal necrosis, noted during the autopsy. Three of
the deaths occurred 225 days after drainage. One
resulted from chronic lung disease 4 months after
drainage (full enteral feeding had been tolerated),
and two were caused by fungal sepsis. The sixth death
was also caused by fungal sepsis, 11 days after
peritoneal drainage.
Seven (24%) of the infants treated with peritoneal
drainage underwent another drainage procedure an
average of 16 days after the initial procedure (Table
2). In three, a subsequent laparotomy was required
because of failure to improve or persistent pneumo-
peritoneum. The fourth patient, who had enterocuta-
neous fistula and fungemia, presumably would have
required resection had he survived. The other three
(43%) infants who underwent a second drainage
procedure healed without further intervention. In a
comparison of patients with successful versus unsuc-
cessful drainage, no significant differences were noted
in gestational age, birth weight, respiratory status, or
indication for operation.
Three other infants in the primary drainage group
underwent laparotomy without a second attempt at
drainage (Table 2). Therefore, a total of six (21%)
infants required laparotomy after one or more at-
tempts at peritoneal drainage. All these infants sur-
vived. Two of the six were found to have diagnoses
other than NEC. One had an unrecognized incarcer-
ated inguinal hernia, and the other had a small bowel
atresia. When the six infants requiring laparotomy
were compared with the remainder of the drainage
group, there were no significant differences in birth
weight, gestational age, age at operation, FIO, at
Table 2. Patients in Whom Further Operative Treatment Was
Required After Initial Peritoneal Drainage

No. of No. of
Patient Drainage Drainage
COMPLICATED NEC NO. Procedures Days Indication Outcome

29 Inadvertent No further
removal surgery
14 Inadvertent No further
removal surgery
LAPAROTOMY PERITONEAL DRAIN
3 2 27 Abscess (US) No further
0 9
surgery
4 2 14 Fistula Died
5 3 14 Obstruction Laparotomy
(Incarcerated
SURVIVORS SURilVORS hernia)
18 (90%) 23 (79%) 6 21 Pneumoperi- Laparotomy
toneum
7 27 Fistula Laparotomy
8 7 Obstruction Laparotomy
(atresia)
9 1 12 Fistula Laparotomy
DRAIN ONLY LAPAROTOMY (intraoperative
17 (74%) 6 (21 %I arrest)
10 18 Fistula Laparotomy
Fig 1. Diagram of patient distribution and results.
312 MORGAN, SHOCHAT, AND HARTMAN

operation, or total number of ventilator days. The Table 4. Indication for Initial Operation

infants requiring laparotomy had a significant in- Primary Primary

Laparotomy Drainage
crease (P < .03) in total operative blood replacement
Pneumoperitoneum 9 25
(111 li 17 mL/kg) and in age at full enteral feeding
Obstruction 6 0
(112 v 83 days). Overall, 17 of the 23 survivors (74%)
Sepsis/+TAP 4 2
had healing of the gastrointestinal tract and resolu- Abdominal abscess 1 2
tion of peritoneal contamination with drainage alone.

Primary Laparotomy by 1 year of age. Three patients had enterocutaneous

Twenty (20) infants were treated with primary
laparotomy. There were two (10%) deaths (Table l),
both caused by overwhelming sepsis. Of the 18
survivors, six had closure of the enterostomy at 1 year
of age. The others underwent closure before dis-
charge at a mean age of 80 days. Compared with the
infants treated by peritoneal drainage, those treated
with laparotomy had a significantly greater birth
weight and gestational age, and a lower oxygen
requirement (Table 3). In the laparotomy group,
there was also a higher incidence of intestinal obstruc-
tion or stricture as an indication for the initial
procedure (Table 4). Many of these infants had
survived the acute NEC episode, but strictures had
developed at sites of presumed circumferential necro-
sis or perforation. There were no wound infections or
dehiscences in either group.
Late Complications
Four infants, all in the drainage group, required
additional operative procedures. In two, scrotal ab-
scesses developed, which required local incision and
drainage. Two infants had abdominal wall hernias at
fistulae, one died of fungal sepsis, and two required
laparotomy with enterostomy. One of these patients
suffered an unexplained arrest during elective enter-
ostomy closure. This patient has the most severe
neurological impairment in the series. At the time of
study (4 months to 9 years) eight (28%) patients in
the peritoneal drainage group were considered devel-
opmentally delayed as were six (30%) in the lapa-
rotomy group.
Survival by Weight
Infants with a birth weight of less than 1,000 g had
higher mortality than those whose weight was more
than 1,000 g (Table 5). The very low birth weight
(VLBW) infants had a slightly lower mortality with
peritoneal drainage, but the small number of infants
treated with laparotomy precludes valid statistical
analysis of this difference. Between 1,000 and 1,500 g,
there was a slight (not statistically testable) advantage
in survival with primary laparotomy. Survival for
infants greater than 1,500 g was 100%.
DISCUSSION

the drain site that required closure. Another infant NEC affects 1% to 2% of infants admitted to
who had an incisional hernia experienced resolution neonatal intensive care units, with 20% to 40%
after compression bandaging for 6 months. requiring operative intervention. Complicated NEC
The majority of infants in both groups required is now the most common surgical emergency in the
prolonged total parenteral nutrition (TPN), which newborn.5 During the past 3 decades there has been a
averaged 76 days from birth in the drainage group definite trend toward lower birth weight and younger
and 52 days from birth in the laparotomy group. None gestational age in infants requiring operative treat-
of the infants in this series required home parenteral ment for NEC6- Concurrently, the mortality rate for
nutrition. Five, all in the drainage group, required infants requiring operation has improved, although it
oxygen supplementation at home, and all were weaned remains substantial (25% to 40%); infants weighing
less than 1,000 g have the highest mortality rates
Table 3. Comparison of Primary Laparotomy and Drainage Groups (35% to 5O%).J
Primary Primary The pathogenesis of NEC remains undefined. Al-
Laparotomy Drarnage t test
In = 20) (n = 29) (PI
though there are a number of similar theories, none
effectively explains the observed variations in patient
Birth weight(g) 1,654 t 203 994 ? 65 < .OOl
Gestational age (wk) 32 t 1.1 27 r 0.5 < ,001
population or the frequency and virulence of the
Age at diagnosis (d) a + 1.7 14 + 2.4 < .04
Age at operation (d) 32 + 7.0 17 ? 2.6 NS Table 5. Survival by Birth Weight and Treatment
Fio? at operation .2a + .02 .49 * .05 < ,001
No. of
No. of ventilator days 25 ? 6.0 60 ? 9.0 < ,005
Birth Weight Patients Laparotomy Drainage
No. of TPN days 52 + a.0 76 -t 5.2 < .05
Survival 90% 79% <l,OOOg 22 60% (3/5) 71% (12/17)
1,000-l ,500 g 13 100% (4/4) 89% (a/9)
NOTE. Data are expressed as mean 2 standard error. Data for all
> 1,500 g 15 100% (12/12) 100% (313)
other categories were not significant (NS).
PERITONEAL DRAINAGE AND NEC
disease. Surgical management has been patterned
after therapies used for adults who have intestinal
ischemia or perforation caused by mesenteric isch-
emia or low flow states. In practice, there is consider-
able variation in the indications for operation6 the
role of primary anastomosis,* and the use of perito-
neal drainage in the management of complicated
NEC.
Traditional operative treatment of complicated
NEC has produced improvements in mortality but is
associated with significant morbidity. Intestinal resec-
tion with a significant incidence of short bowel
syndrome (23%) and total abdominal wound dehis-
cence (10% to 15%)i4 are two of the more obvious
complications. The effect of a major laparotomy with
significant blood loss, volume shifts, and stress re-
sponse on the developing respiratory and central
nervous systems is unknown and less obvious.
Simple drainage of the peritoneal cavity appears to
violate established surgical principles of the manage-
ment of intestinal ischemiaiperforation and intraab-
dominal sepsis. In addition, the premature infant is
considered to be relatively immune-incompetent and
to have a limited ability to compartmentalize intraperi-
toneal contamination, in part because of the undevel-
oped omentum. The fact that a significant number of
patients in this (62%) and other series (27% to
46%)-4 were successfully treated with peritoneal
drainage alone requires us to reevaluate our under-
standing of the pathophysiology and treatment of
NEC, particularly in the VLBW infant. This group of
infants appears to have unusual responses in terms of
peritoneal and systemic host-defense mechanisms as
well as gastrointestinal healing.
The unique features of fetal wound healing de-
scribed in experimental models may play a role in the
healing of gastrointestinal ischemia and perforation
in this setting. Fetal wound healing without scarring
or inflammation is well documented for humans and
appears to be related to intrinsic differences in fetal
cells and the extracellular matrix. Adzick and
Longakeri noted that this type of wound healing
begins to diminish at the start of the third trimester
and does not occur once gestational maturity is
attained. These findings would suggest that if the
bowel would heal, it would be less likely to produce
significant scarring or stricture in the less mature
infants.
The ability of the fetal peritoneal cavity to resorb
infarcted bowel and sterile meconium is well demon-
strated by intestinal atresia, which is generally consid-
ered a consequence of vascular insufficiency. Al-
though in utero ischemia is not comparable to
postnatal infarction of colonized bowel, there are
parallels that suggest an interaction between ischemic
313
tissue, bacteria, and fetal host defenses that may be
useful in designing studies to identify the critical
factors in successful host defense. The ability of the
very premature infant to mount a significant inflamma-
tory cytokine response and its role in this setting is
speculative but deserving of investigation.
Although peritoneal drainage was effective for the
majority of the low birth weight infants in this series,
it did not eliminate the need for intestinal resection.
The six infants ultimately requiring laparotomy were
not distinguishable in terms of size, gestational age,
or features of their respiratory or gastrointestinal
status. However, they did require a longer drainage
period (averaging 21 days before laparotomy) than
did those who were successfully treated with drainage
alone. The two errors in diagnosis and the require-
ment for prolonged drainage suggest that laparotomy
should be considered if the perforation is not con-
trolled. Perhaps this determination could be made
earlier than 21 days; however, retrospective review of
the cases in this series did not identify factors that
would distinguish responders from nonresponders.
Of the six deaths, the four infants who died of
progressive sepsis conceivably could have benefited
from laparotomy with more aggressive intraperito-
neal debridement. However, these infants demon-
strated such hemodynamic and respiratory instability
that it was believed they were unlikely to survive a
major operative procedure.
The results of this review suggest that primary
peritoneal drainage under local anesthesia can pro-
duce survival rates comparable to those of lapa-
rotomy (Tables 3 and 5) and may provide definitive
therapy in the majority of VLBW infants with perfo-
rated NEC. Whereas a second drainage procedure
will be successful in approximately 40%, a significant
number of infants whose first drainage was unsuccess-
ful will require laparotomy. It appears that overall
mortality and gastrointestinal morbidity are improved
with use of drainage, but it is unclear whether the
increase in respiratory dysfunction in the drainage
group is related to the mode of treatment or the
degree of prematurity and respiratory dysfunction at
the time of perforation. The morbidity and mortality
of laparotomy in this group would presumably be
significant, although it was not tested in this study.
The use of initial peritoneal drainage does not pre-
clude subsequent laparotomy. Given the excellent
survival rate and the high probability of avoiding
laparotomy and intestinal resection, we believe that
peritoneal drainage should be the initial form of
treatment for low birth weight infants with perforated
NEC, particularly those whose weight is less than
1,000 g.
314
REFERENCES
1. Ein SH, Marshall DG, Girvan D: Peritoneal drainage under
local anesthesia for perforations from necrotizing enterocolitis. J
Pediatr Surg 12:963-967, 1977
2. Ein SH, Shandling B, Wesson D, et al: A 13-year experience
with peritoneal drainage under local anesthesia for necrotizing
enterocolitis perforation. J Pediatr Surg 25:1034-1037, 1990
3. Cheu HW, Sukarochana K, Lloyd DA: Peritoneal drainage
for necrotizing enterocolitis. J Pediatr Surg 23557-561, 1988
4. Robertson JFR, Azmy AF, Young DG: Surgery for necrotiz-
ing enterocolitis. Br J Surg 74:387-389, 1987
5. Kosloske AM, Musemeche C: Necrotizing enterocolitis of the
neonate. Clin Perinatol 16:97-111, 1989
6. Santulli TV, Schullinger JN, Heird WC, et al: Acute necrotiz-
ing enterocolitis in infancy: A review of 64 cases. Pediatrics
55:376-387, 1975
7. GNeill JA, Stahlman MT. Meng HC: Necrotizing enterocoli-
tis in the newborn: Operative indications. Ann Surg 182274-279,
1975
8. Ricketts RR: Surgical therapy for necrotizing enterocolitis.
Ann Surg 200:653-657, 1984
9. Cikrit D, Mastandrea J, West KW, et al: Necrotizing Entero-
colitis: Factors affecting mortality in 101 surgical cases. Surgery
96:648-655, 1984
Discussion
WI. Pokomy (Houston, TX): Although the authors
emphasize the 79% survival after primary drainage,
survival after laparotomy in their hands was 24 of 26,
or 93%, and the overall survival was 88%, which is
truly remarkable.
I have a few questions and comments. Since your
stated indication for operation was free air or feces
on peritoneal tap, were there infants with extensive
disease, without perforation, who died of sepsis,
without operation ? If so, what was the mortality
among the nonoperated NEC patients?
In our experience with over 200 patients operated
on for acute NEC, we have not found a correlation
between the patients weight and the chance of
survival; that is, after surgical intervention, infants
weighing 800 g seem to do as well as those weighing
1,700 g, so why did you use 1,500 g as a cutoff? Could
it be that the larger babies do not tolerate peritonitis
as well as the smaller ones?
Six of the 29 patients with primary drainage subse-
quently underwent laparotomy. What did you use for
your decisions and the timing?
Do you believe, as I think you infer in your report,
that operative intervention with resection and enter-
ostomy or anastomosis is actually detrimental to the
survival of these infants? Or, is it simply unnecessary?
The average length of time to total enteral feedings
was 83 days, which seems very long. Certainly, after
primary anastomosis, the vast majority of our infants
MORGAN, SHOCHAT, AND HARTMAN
10. Ricketts RR, Jerles ML: Neonatal necrotizing enterocolitis:
Experience with 100 consecutive surgical patients. World J Surg
14:600-605, 1990
1 I. Grosfeld JL, Cheu H, Schlatter M, et al: Changing trends in
necrotizing enterocolitis, experience with 302 cases in two decades.
Ann Surg 214:300-307,199l
12. Cooper A, Ross RJ, ONeill JA, et al: Resection with
primary anastomosis for necrotizing enterocolitis: A contrasting
view. J Pediatr Surg 23:64-68, 1988
13. Takamatsu H, Akiyama H, Ibara S, et al: Treatment for
necrotizing enterocolitis perforation in the extremely premature
infant (weighing < 1000 g). J Pediatr Surg 27:741-743, 1992
14. Angel C, Daw S, Phillipe P, et al: Pig in pouch: A technique
for the management of complete wound dehiscence after lapa-
rotomy for neonatal necrotizing enterocolitis. J Pediatr Surg
27~67-69, 1992
15. Longaker MT, Adzick NS: The biology of fetal wound
healing: A review. Plast Reconstr Surg 87:788-798, 1991
16. Adzick NS, Longaker MT: Scarless fetal healing, therapeu-
tic implications. Ann Surg 215:1-7, 1992
are fed much earlier than this. Could you comment on
that length?
Sixty-nine percent of our patients have free air, and
30% present with pneumoperitoneum. Many of these
babies have a simple ileal perforation that can be
closed primarily. Are these patients included in your
series, and was simple drainage performed?
We continue to perform either an enterorrhaphy or
resection of well-demarcated necrotic segments with
primary anastomosis in nearly 50% of our patients. In
this group the survival remains at 90%; and if deaths
after 30 days are excluded, the survival is closer to
95%, regardless of their weight. These patients typi-
cally have an uneventful postoperative course. Ap-
proximately 35% in our experience undergo resection
and enterostomy, with a 71% survival. Unfortunately,
15% of our patients require only drainage or lapa-
rotomy, and in our hands the mortality has been 71%.
I agree with your conclusion that wound healing
and peritoneal and systemic host defenses are signifi-
cantly different in the very low birth weight infant,
and that these differences indicate the need to
reevaluate treatment strategies in this patient popula-
tion.
S.H. Ein (Toronto, Canada): I read this abstract
with great interest. I remember very well when I first
presented this idea to APSA in Mexico, about 10 or
15 years ago. All the diarrhea we were supposed to
PERITONEAL DRAINAGE AND NEC
get came down on me. The main concern at that time
was that this drain procedure not be considered the
definitive procedure. However, this was never our
intention nor was it ever the intention of Donny
Marshall, its originator. We have always considered it
to be a temporizing minor surgical intervention and
strongly advised that it should stay that way. More-
over, we suggested that this be done almost entirely
for preemies under 1,000 g, and we have realized that
as the weight of infants increases, the good results
begin to decrease.
Our results fall into three groups. The first group
heals up, and the GI tract is forever patent. The
second group heals up, but a stricture develops 6
weeks to 3 months later, requiring further surgery.
The third group continues to go down hill and
requires a laparotomy usually within 24 hours. The
finding in this group is a long piece of necrotic bowel,
and many of these babies die. If I lump our first two
groups together, I too will get about a 66% success
rate. Nonetheless, 1 still caution you to consider this
drain procedure as an initial temporizing procedure,
and watch the preemie closely to see what happens.
T. C. Moore (Palos Verdes, CA): My experience here
parallels that of the authors. In the last 17 infants,
from non-crack-addict mothers, there has been no
mortality and no gut loss from resection or enteros-
tomy necrosis and slough.
Furthermore, there has been no intestinal stric-
ture. Here again, my experience parallels that of the
authors. The absence of intestinal diversion may have
contributed to the absence of stricture formation.
This patch, drain, and wait approach is modeled on
the experiment of nature in which extensive gut
necrosis from a vascular catastrophe in utero may
result in only a limited atresia (or atresias) or may
eventuate only in cases of meconium peritonitis with
no intestinal obstruction. Broviac hyperalimentation
functions as the placenta for nutrition, and the drains
simulate a sterile in utero peritoneal cavity, as inflam-
mation-induced angiogenesis, good angiogenesis, leads
to a maximum of gut preservation, large and small.
This approach also has been used by me with
375
success in midgut volvulus with extensive intestinal
ischemic damage and necrosis.
G.E. Hartman {response): In evaluating our report,
it is important to acknowledge that it is nonrandom-
ized, retrospective, and uncontrolled. It represents
our initial experience with an unconventional therapy
that has been surprisingly effective.
No patients were excluded from operative treat-
ment, although three with a rapidly fulminant course
(12 to 24 hours) died without operation. Dr Pokorny
has not found a correlation between survival and
birth weight. The higher mortality in infants weighing
less than 1,000 g in our study is similar to the findings
in recent reports by Ricketts and Grosfeld.
The timing and definition of failure of peritoneal
drainage is evolving as we gain experience. During
the period of this study, we considered recurrent
pneumoperitoneum, persistent intestinal obstruction,
and development of enterocutaneous fistula as signs
of failure of drainage. Once peritoneal drainage was
determined to be ineffective, a standard laparotomy
with intestinal resection and enterostomy was per-
formed.
From our data it would appear that laparotomy as
initial therapy may be unnecessary in the majority of
infants weighing less than 1,000 g. It is premature to
state that laparotomy is detrimental in this patient
population. A direct comparative study of primary
laparotomy and peritoneal drainage is needed to
determine the relative effectiveness of these two
modalities. Such a study will decipher whether the
long times to full enteral feeding and extubation in
our study are a result of the treatment or the
associated conditions in this population.
Dr Ein has described this technique and has
cautioned against exactly what we have reported
today. The surprising results of our initial experience
suggest that the very low birth weight infant may have
pathophysiological responses different from those of
older, more mature infants. We believe the treatment
strategies in this population need to be reevaluated
and that peritoneal drainage should be considered as
primary therapy for perforated NEC in the very low
birth weight infant.