An in vitro model and case report that used gelatin sponge to restore amniotic fluid volume after spontaneous

premature rupture of the membranes

John M. OBrien, MD,a Brian M. Mercer, MD,b John R. Barton, MD,a and Douglas A. Milligan, MDa Lexington, Ky, and Cleveland, Ohio
OBJECTIVE: The purpose of this project was to study an in vitro model for plugging membrane defects with gelatin sponge and to develop a method with which to use this material to treat premature rupture of the membranes. STUDY DESIGN: Fetal membranes were fixed over the opening of a flask that was filled with saline solution and gelatin sponge. Defects of various sizes were created, and the usefulness of differing sizes of gelatin sponge to obstruct the defects was observed. This technique was then applied to a case of previable, spontaneous premature rupture of the membranes. RESULTS: Fifteen amniotomies were performed in the in vitro model. The gelatin sponge obstructed all defects less than 7 mm in length, when pieces up to 1 1 cm in dimension (n = 8 amniotomies) were used. For larger defects or those defects with a complex shape (such as cruciate), gelatin sponge was not effective at arresting fluid loss (n = 4 amniotomies). An inspection of larger gelatin sponge pieces, after instillation through a 12-gauge angiocatheter, revealed 36% (15 of 42 pieces) of 1 1 cm pieces remained intact. A case of spontaneous, previable premature rupture of the membranes was treated with this material. A favorable outcome was observed. CONCLUSION: Gelatin sponge is successful at arresting the egress of fluid through membrane defects when smaller defects are present. Complex or larger linear defects may not be treated by this method alone and necessitate adjuvant therapies. This therapeutic strategy can be applied to cases of previable, spontaneous premature rupture of the membranes. (Am J Obstet Gynecol 2001;185:1094-7.)

Key words: Premature rupture of membranes, gelatin sponge, embolization

No therapeutic method has yet been devised that reliably restores amniotic fluid volume after spontaneous premature rupture of the membranes (PROM). Such a treatment would be valuable, particularly for early midtrimester PROM with minimal residual amniotic fluid because of the marked morbidity and death associated with this condition.1-3 One strategy is to plug the endocervical canal or membrane defect. Theoretically, a substance could be instilled into the uterus that fills the membrane defect or endocervical canal, prevents or retards the loss of amniotic fluid, and does not readily degrade for the remainder of the pregnancy. The gelatin sponge, Gelfoam

(Pharmacia, Peapack, NJ), has been used in fetal surgery and is a candidate for such an obstructing agent.4 The purpose of this experiment was to determine whether gelatin sponge could act as an embolizing agent and to identify the size and shape of a membrane defect that could be successfully approached with this method alone. Further, our goal was to develop a therapeutic approach to spontaneous PROM and attempt to test this strategy in a case of previable PROM. Material and methods Fetal membranes were excised and rinsed with normal saline solution. Segments of fused amnion and chorion that measured approximately 10 10 cm were then fixed to the opening of a glass flask with rubber bands. The flask was filled with normal saline solution by way of an intravenous set-up through a sideport (Fig). Solubilized gelatin sponge (Gelfoam) pieces were added to the solution through a syringe that had been fitted with a catheter. Defects of various sizes were created in the membrane with differing gauge needles or a scalpel blade. The ability of varying sizes of gelatin sponge to obstruct the defects was observed. The size and character of residual pieces of gelatin sponge were also assessed.

From the Department of Maternal-Fetal Medicine, Perinatal Diagnostic Center, Central Baptist Hospital,a and the Department of MaternalFetal Medicine, MetroHealth Medical Center, Case Western Reserve University.b Presented at the Twenty-first Annual Meeting of the Society for MaternalFetal Medicine, Reno, Nev, February 5-10, 2001. Reprint requests: John M. OBrien, MD, Director, Perinatal Diagnostic Center, Central Baptist Hospital, 1740 Nicholasville Rd, Lexington, KY 40503. E-mail: jobrien@bhsi.com. Copyright 2001 by Mosby, Inc. 0002-9378/2001 $35.00 + 0 6/6/117684 doi:10.1067/mob.2001.117684

1094

Volume 185, Number 5 Am J Obstet Gynecol

OBrien et al 1095

After a demonstration of the potential for this technique to retard the loss of amniotic fluid, the institutional review board at Central Baptist Hospital approved a protocol and an informed consent process for select patients with previable PROM. Candidates for the intra-amniotic administration of gelatin sponge must meet the following inclusion criteria: (1) PROM must be definitively documented; (2) the gestational age must be <22 weeks at the time of the procedure; and (3) persistent, severe oligohydramnios must be observed, with a maximum vertical pocket of <1.5 cm. Exclusion criteria include labor, active vaginal bleeding, fetal anomalies, or signs of infection. Chorioamnionitis is defined as a temperature of >100.4F that is associated with uterine tenderness, an abnormal discharge, or both. This procedure consisted of 3 parts that included cerclage placement, amnioinfusion, and administration of the gelatin sponge. Patients were observed for signs of infection and labor as an inpatient for 3 to 7 days, depending on their symptoms, and were then discharged home for follow-up with referring physicians. Patients were asked to document whether they had persistent leakage, noted any onset of fever, and obtained a weekly assessment of the amniotic fluid volume. Results For the in vitro model, 15 amniotomies were performed. The gelatin sponge was effective in obstructing all defects <7 mm in length when the gelatin sponge was up to 1 1 cm in dimension (n = 8 amniotomies). When defects were equal to approximately 7 mm, incomplete obstruction of fluid loss was observed (n = 2 amniotomies). For larger defects or those defects with a complex shape (such as cruciate), the gelatin sponge was not effective at arresting fluid loss (n = 4 amniotomies). Finally, in 1 instance, smaller gelatin sponge pieces (5 5 mm) could not obstruct a 6 4mm defect. Attempts to instill larger pieces of gelatin sponge required larger gauge needles/Angiocaths (Becton Dickinson, Sandy, Utah). An inspection of larger gelatin sponge pieces after instillation into the flask through the largest gauge Angiocath (12-gauge) revealed that 36% (15/42 pieces) of 1 1cm pieces remained intact. A remarkable shredding of gelatin sponge was seen with the administration of even larger pieces (>1 1 cm), because only 17% (5/29 pieces) of those remained intact. Case report A 19-year-old primigravid woman was referred to our center at 20 weeks of gestation after an ultrasound examination at 18 weeks demonstrated oligohydramnios and a diagnostic amnioinfusion revealed spontaneous, previable PROM. Repeat amnioinfusions on 2 other occasions at 19 and 20 weeks of gestation demonstrated rapid loss of fluid, with residual fluid measurements being less than 1.0 cm in the vertical plane.

Figure. Schematic representation of a gelatin sponge plugging a membrane defect in the in vitro model.

The patient was counseled and consented to undergo the administration of gelatin sponge at 21 weeks of gestation when the maximum vertical pocket measured 0.8 cm. The procedure consisted of the initial placement of a McDonald cerclage. The patient was repositioned, and an amnioinfusion of 150 mL of normal saline solution was performed. Finally, 2 sheets of Gelfoam (size 100) were cut into 1 1cm blocks, solubilized in saline solution, and loaded into syringes. The administration of the gelatin sponge was accomplished by expressing the contents of the syringes through a 12-gauge Angiocath inserted transabdominally into the lower segment of the uterine cavity under ultrasound guidance. The patient was given prophylactic antibiotics of azithromycin and ampicillin and was observed as an inpatient for 1 week. The amniotic fluid volume improved, and no complaints of further leakage of fluid were elicited. Repeat pelvic examinations were avoided. Weekly amniotic fluid indices from 25 through 32 weeks of gestation revealed values from 3.2 to 7.2 cm. Biometry revealed appropriate chest growth and normal heart/thoracic circumference ratios. The gelatin sponge was visualized in the amniotic cavity at 32 weeks of gestation and was also suspected in the fetal stomach. The gelatin sponge appeared as small echogenic entities floating in the uterine cavity. Transvaginal scanning also showed the echogenic gelatin sponge in the endocervical canal and lower segment. Finally, during antepartum examination, a clubbed right foot was identified (which is a common deformation in the presence of midtrimester PROM) when screening was performed for anomalies.5 At 32 weeks, the patient was admitted for vaginal bleeding. Three days later she again complained of leaking fluid, and the amniotic fluid index decreased from 7.2 to 3.1 cm, which is consistent with the loss of the gelatin sponge plug and recurrent symptomatic PROM. She was observed as an inpatient until 35 weeks of gestation, with reassuring antepartum testing and resolution of her vaginal bleeding. She then underwent removal of the cerclage, induction with oxytocin, and a spontaneous vaginal delivery of a 2495-g male with Apgar scores of 7 at

1096 OBrien et al

November 2001 Am J Obstet Gynecol

1 minute and 9 at 5 minutes. No respiratory or gastrointestinal difficulties were observed. A clubbed foot was identified at birth. Pathologic examination of the placenta did not demonstrate an inflammatory response in the membranes. Follow-up examination of the newborn has revealed some restitution of the clubbed foot with conservative measures, but surgical intervention has been planned. Appropriate neurologic development has been documented. Comment The reported mortality rate associated with early midtrimester rupture of the membranes ranges from 45% to 90% in prospective series.1-3 Death, in these cases, is most often the result of previable delivery, pulmonary hypoplasia, or other sequelae of extreme prematurity. Because of the poor outcomes observed in previous series, we developed a therapeutic strategy for this condition. Initially, an animal model was attempted in the sheep; however, we were unsuccessful at reliably achieving rupture of membranes in this species, presumably because of the type of placentation. After developing a process and gaining experience with our in vitro work, we believed a relatively low-risk procedure could be justifiably offered to patients with previable PROM who were at greatest risk for pregnancy loss and who did not desire pregnancy termination. By our in vitro studies, we have demonstrated that gelatin sponge can plug defects in fetal membranes. This work has similarities to an animal model reported by Luks et al,4 in which gelatin sponge was used to maintain membrane integrity after fetoscopy. This group did not report any adverse fetal outcomes that were related to exposure to the gelatin sponge. Furthermore, the gelatin sponge has a long history in surgical procedures and has not been associated with marked inflammatory reactions or other adverse sequelae. The primary benefit of this material is the length of time it remains intact within the amniotic cavity. We observed this material for 12 weeks after administration within the uterine cavity. Gelatin sponge appears to remain intact longer than fibrin or other blood products, which have also been used as possible plugs.6-8 We remain mindful during our counseling, however, that gelatin sponge may adversely affect the fetus by being swallowed, which would cause intestinal obstruction, or could enter the respiratory tract, theoretically resulting in airway obstruction. Patient selection is of utmost importance in the evaluation of the effectiveness of a new technique. We have intentionally chosen those patients that we believe are at highest risk for a poor outcome on the basis of gestational age at PROM and the lack of residual amniotic fluid. We further do not wish to offer this procedure to those patients who have infection as the cause of membrane rupture; however, excluding these patients can be difficult, given the limitations of clinical signs, microbial culture of

amniotic fluid, and biochemical assessment of amniotic fluid. We currently also use time as a discriminating factor for infection and believe most patients without labor within 3 days after PROM may have other causes (such as abnormal membrane formation/growth) for this complication. These latter patients are those patients who might benefit from any therapeutic attempt at restoration of the amniotic fluid volume. The procedure we use to restore amniotic fluid volume has 3 parts. The first element is to perform a McDonald cerclage. Its purpose is to narrow the endocervical canal so that the cervix, in addition to the membranes, may act as a potential site for arresting fluid loss by the gelatin sponge. Complex or large defects in the membranes may not be successfully treated with the administration of gelatin sponge alone, as we demonstrated in our model. Furthermore, Quintero et al9 have shown by fetoscopy that the size of membrane defect may enlarge over time because of apoptosis or other degenerative changes. While the cerclage is being performed, the vagina is not irrigated nor is the endocervical canal instrumented or traversed. Therefore, we attempt to minimize procedurerelated infection and monitor outcomes to identify any association. The second step in our procedure is an amnioinfusion, which creates a potential space for the placement of a catheter and establishes a flow exiting the uterus. Locatelli et al10 have demonstrated that amnioinfusion alone may be therapeutic in cases of midtrimester PROM. In our case, however, the rapid loss of fluid after serial amnioinfusions and the early gestational age at PROM argued against this strategy alone being successful. In their series, Locatelli et al noted an 80% mortality rate with 60% of survivors having abnormal neurologic outcomes when PROM occurred at less than 26 weeks and amnioinfusion did not restore the fluid volume to a vertical fluid pocket to >2 cm for at least 48 hours. Finally after amnioinfusion, the gelatin sponge is administered in the lower uterine segment and is carried along with the exiting fluid to a narrowed site. We have termed such a therapeutic approach embolization. Retarding the egress of fluid anywhere along the pathway that leaves the uterus may improve the likelihood of normal pulmonary development and may reduce the risk for previable delivery in cases of PROM. In our case, visualization of the gelatin sponge in the endocervical canal suggests that the cervix can act as a site for obstructing fluid loss. This technique of using gelatin sponge plugging deserves further study under an approved, monitored protocol, and such a trial is underway at our institution. Any potential benefit we have achieved to date is only preliminary and cannot validate this approach. A trial aimed at testing the efficacy of any technique in women with PROM requires a clear description of eligible patients and a standardized therapy. We are attempting to iden-

Volume 185, Number 5 Am J Obstet Gynecol

OBrien et al 1097

tify appropriate patients and are evaluating the usefulness of our present strategy. Modifications may ultimately be necessary in our approach (such as fetoscopic-directed gelatin sponge administration or combining this embolization strategy with fibrin glue application to the endocervical canal to improve effectiveness).11 However, despite this uncertainty, as investigators develop methods to address this too frequently lethal complication of pregnancy, patients may be offered alternatives to current practices of termination and expectant treatment.
REFERENCES

5.

6.

7.

8.

1. Richards DS. Complications of prolonged PROM and oligohydramnios. Clin Obstet Gynecol 1998;41:817-26. 2. Xiao ZH, Andre P, Lacaze-Masmonteil T, Audibert F, Zupan V, Dehan M. Outcome of premature infants delivered after prolonged premature rupture of membranes before 25 weeks of gestation. Eur J Obstet Gynecol Reprod Biol 2000;90:67-71. 3. Kilbride HW, Yeast J, Thibeault DW. Defining limits of survival: lethal pulmonary hypoplasia after midtrimester premature rupture of membranes. Am J Obstet Gynecol 1996;175:675-81. 4. Luks FI, Depest JA, Peers KHE, Steegers EAP, van der Wildt B. Gelatin sponge plug to seal fetoscopy port sites: Tech-

9.

10.

11.

nique in ovine and primate models. Am J Obstet Gynecol 1999;181:995-6. Chistianson C, Huff D, McPherson E. Limb deformations in oligohydramnios sequence: effects of gestational age and duration of oligohydramnios. Am J Med Genet 1999;86:430-3. Quintero RA, Morales WJ, Allen M, Bornick PW. Arroyo J, LeParc G. Treatment of iatrogenic previable premature rupture of membranes with intra-amniotic injection of platelets and cryoprecipitate (amniopatch): preliminary experience. Am J Obstet Gynecol 1999;181:744-9. Young BK, Roque H, Abdelhak YE, Poiolek D, Demopulos R, Lockwood CJ. Minimally invasive endoscopy in the treatment of preterm premature rupture of the membranes by application of fibrin sealant. J Perinat Med 2000;28:326-30. Uchide K, Terada S, Hamasaki H. Intracervical fibrin instillation as an adjuvant to treatment for second trimester rupture of membranes. Arch Gynecol Obstet 1994;255:95-8. Quintero RA, Morales WJ, Kalter CS, Allen M, Mendoza G, Angel JL, et al. Transabdominal intra-amniotic endoscopic assessment of previable premature rupture of membranes. Am J Obstet Gynecol 1998;179:71-6. Locatelli A, Vergani P, Di Pirro G, Doria V, Biffi A, Ghidini A. Role of amnioinfusion in the management of premature rupture of the membranes at < 26 weeks gestation. Am J Obstet Gynecol 2000;183:878-82. Sciscione AC, Manley JS, Pollock M, Mass B, Shlossman PA, Mulla W, et al. Intracervical fibrin sealants: a potential treatment for early preterm premature rupture of the membranes. Am J Obstet Gynecol 2001;184:368-73.