In 2003, more than 1 million cesarean deliveries were performed in the United States, 27.6% of all deliveries, the highest ever.1 The primary cesarean delivery rate rose 13% between 2001 and 2003, and the percentage of women undergoing vaginal birth after cesarean declined from 23.0% to 10.6%.1 With more than one-quarter of all U.S. births occurring by cesarean delivery, small changes in surgical technique may significantly alter morbidity. Despite the frequency of cesarean deliveries, data are limited regarding the long-term impact of different surgical techniques.
Whether to close the parietal peritoneum at cesarean delivery has been debated, with short-term data failing to support peritoneal closure and virtually no long-term data available. In the short-term, nonclosure of the parietal peritoneum appears to decrease operating time without a difference in postoperative wound infection, dehiscence, endomyometritis, ileus, or length of hospital stay.2 Nonclosure of both the parietal and visceral peritoneum, compared with closure of both, has been associated with decreased operating time, no difference in postoperative pain requirements, and conflicting data on return of bowel function.3,4 Closure of the visceral peritoneum alone has been associated with increased operating time and more infectious morbidity.5
In gynecologic,6 general surgical,7 and animal8 studies, closure of the parietal peritoneum has been associated with increased adhesions. Based on these data, some have recommended abandonment of peritoneal closure in both gynecologic and obstetric surgery.9,10 The effect on adhesions of leaving the peritoneum open at cesarean delivery, however, has not been well-studied. Adhesion-related complications, including bowel obstruction, chronic pelvic pain, infertility, and difficult repeat surgery, are estimated to cost $1.2 billion annually.11 The paucity of data regarding the long-term effect of nonclosure of the peritoneum at cesarean delivery originally prompted Cochrane Database reviewers to conclude in 2000 that “data are insufficient to warrant a change in practice,”12 supporting continued closure of the peritoneum at cesarean delivery.
Given the unanswered questions regarding the effect of peritoneal nonclosure at cesarean delivery on adhesions, we sought to examine whether closure of the peritoneum at time of cesarean delivery has an effect on adhesion formation.
MATERIALS AND METHODS
To examine the relationship of parietal peritoneal closure at prior cesarean delivery to adhesions at the subsequent cesarean delivery, we designed a prospective cohort study of women undergoing first repeat cesarean delivery at Stanford Medical Center from 1996 to 2003. Our primary outcome of interest was the presence of adhesions at the time of repeat cesarean. Secondary outcomes included operative times of first and second surgery, length of time to bowel function, and operative complications.
All surgeons, including private and faculty attendings and residents, were asked to report the location and severity of adhesions by completing an adhesion score sheet immediately after performing surgery (see box “Adhesion Scoring Data Sheet”). Adhesions score sheets were included, during the study period, with each patient's cesarean paper work and were also available in the recovery room. All surgeries were performed by a resident who was directly supervised by an attending. Data were then abstracted from medical records of patients' first and second cesarean deliveries, labors, and postoperative courses. Abstracted data included patient gravity, parity, age, payor status, indication for cesarean delivery, ethnicity, wound and postoperative infections, chorioamnionitis, bowel function, previous closure of the peritoneum, bladder and rectus muscles, suture types, use of irrigation and antibiotics, type of anesthesia, operative complications, operative times, and estimated blood loss.
Patient records were reviewed from all cesarean deliveries for which an adhesion score sheet was filled out. Patients were excluded if, upon medical record review, they were found to have adhesions at the first cesarean delivery, postoperative wound infection or breakdown following first operation, unavailable first-operative note and course, intervening laparotomy or laparoscopy, other surgeries at first cesarean delivery such as incidental myomectomy or cystectomy, insulin-dependent diabetes mellitus, steroid-dependent disease, and permanent suture at first cesarean.
An adhesion frequency of 50% was assumed with parietal peritoneal closure and 25% when the parietal peritoneum was left open. To have 80% power to detect a 50% reduction in adhesions when the parietal peritoneum was left open, with an α of 0.05 and β of 0.2, 128 patients total, or 64 patients per arm, were required.
The interrater reliability (the extent to which 2 “raters” agree) of the adhesion scoring sheet was measured. At 5 cesarean deliveries, different attending-resident pairs were instructed not to discuss adhesions during surgery and independently to fill out the adhesion score sheet after surgery. Each of the 15 potential data points was assessed and compared between members of the pair (see box “Adhesion Scoring Data Sheet”). The assessment was judged as similar if each surgeon from the pair checked the same box on the data sheet and as different if only one checked a box.
All data were entered into a Stata 7.0 (StataCorp, College Station, TX) database. Univariate statistical tests using the χ2 test of proportions were considered statistically significant with P < .05. Multivariable logistic regression analysis was used to control for potential confounders. We received approval for this study from the Committee on Human Research at Stanford University Medical Center.
One hundred seventy-three patients met inclusion criteria. The parietal peritoneum was left open at first cesarean delivery in 106 patients and was closed in 67 patients. In the study population, 21% received public assistance, 14% had chorioamnionitis or endomyometritis at first cesarean delivery, 72% underwent labor before the first cesarean delivery, and 61% were white (Table 1). There were no significant differences with regard to peritoneal closure for payor status, A1 gestational diabetes, prior infection, age, or previous labor. There were significant differences in ethnicity between the 2 groups, with more white women having their parietal peritoneum closed.
When we examined the primary outcome, patients with prior parietal peritoneal closure had significantly fewer dense and filmy adhesions (52% versus 73%, P = .006, Table 2). By location, significantly fewer adhesions were seen between the omentum andfascia, omentum and uterus, and fascia and uterus. Bowel adhesions, and adhesions at other pelvic structures, were not significantly different based on whether the parietal peritoneum was closed. When filmy adhesions were excluded and we analyzed only adhesions judged as dense, prior parietal peritoneal closure was associated with significantly fewer dense adhesions overall (30% versus 45%, P = .043, Table 2). By location, parietal peritoneal closure was associated with significantly fewer dense omentum-to-fascia adhesions (Table 2).
Examining the rates of adhesions among the different potential confounders, we found that receipt of public assistance was significantly associated with adhesions at the second cesarean delivery; type 2 diabetes, previous infection, age, previous labor, and race were not (Table 3). When we examined the rate of adhesion formation by payor status, we found that peritoneal closure was associated with lower rates of adhesions in both subgroups. Among women with private insurance, peritoneal closure was associated with a 23% rate of adhesions, whereas no peritoneal closure was associated with a 44% rate of adhesions (P = .013). Among women with Medicaid insurance, peritoneal closure was associated with a 41% rate of adhesions, whereas no peritoneal closure was associated with a 76% rate of adhesions (P = .037). When controlling for potential confounders using multivariable logistic regression, we found that patients receiving public assistance had more adhesions overall, regardless of type of closure (odds ration [OR] 5.04, 95% confidence interval [CI] 1.55–16.40).
A multivariable logistic regression analysis was performed, which controlled for the following potential confounding variables: prior infection, prior labor, visceral peritoneal closure, rectus muscle closure, payor status, ethnicity, maternal age, and gestational diabetes. When the parietal peritoneum was closed at first cesarean delivery, dense and filmy adhesions were seen four-fifths less frequently at repeat cesarean delivery (OR 0.20, 95% CI 0.08–0.49), and dense adhesions were seen two-thirds less frequently (OR 0.32, 95% CI 0.13–0.79), as shown in Table 4. By location, parietal peritoneal closure was associated with fewer filmy and dense adhesions between the fascia and uterus and between the omentum and fascia, and fewer dense adhesions between the omentum and fascia.
When we examined the secondary outcomes, we found that parietal peritoneal closure did not significantly alter the length of the first or repeat cesarean deliveries. In patients undergoing their first cesarean delivery, the mean operating time was 47.0 minutes for those with parietal peritoneal closure and 46.9 minutes for those without parietal peritoneal closure (P = .515). In patients undergoing the repeat cesarean, the mean operating time was 54.70 minutes for those with prior parietal peritoneal closure and 53.46 minutes for those without prior parietal peritoneal closure (P = .665). Parietal peritoneal closure was associated with a slower return of bowel function. When the parietal peritoneum was closed, bowel function took more than 1 day to return 77% of the time versus 56% of the time when the parietal peritoneum was left open (P = .009, OR 0.31, 95% CI 0.13–0.74). When we assessed the adhesion score sheet, among 75 comparisons between 2 surgeons, the interrater reliability was 0.84.
In contrast to data from nonpregnant patients, parietal peritoneal closure at cesarean delivery appears to protect against abdominal adhesions. Dense and filmy adhesions were four-fifths less frequent following parietal peritoneal closure, and dense adhesions alone were two-thirds less frequent, independently of potential confounding variables. Without parietal peritoneal closure, the fascia was more frequently adherent to the omentum and uterus. Considering these findings, for every 10 cesareans performed with a peritoneal closure, at least 2 fewer women will have any adhesions at a subsequent surgery, and 1 fewer woman will have dense adhesions.
Our findings may be explained by the unique nature of the peritoneum, pregnancy, and the puerperium. Unlike other tissues, the peritoneum does not require apposition of tissue edges for closure.13,14 Instead, mesothelial cells migrate into a supportive matrix and rapidly initiate multiple sites of simultaneous repair, regardless of the size of the peritoneal defect, a process which may be impaired by suture ischemia.14 The anatomic and physiologic changes of pregnancy may alter this healing process. The enlarged postpartum uterus may disrupt the supportive matrix normally created and necessary for reperitonealization. Perhaps the presence of the uterus within the location of the supportive matrix alters mesothelial cell migration, resulting in disordered healing and adhesions of the surrounding structures, such as the fascia to the omentum and uterus.
Adhesions form when fibrinolysis is suppressed and fibrin persists. Fibrin is then infiltrated by fibroblasts, which ultimately organize fibrin bands into adhesions.15 Tissue ischemia is known to suppress fibrinolysis15 and provides an explanation for increased adhesions when the peritoneum is sutured among nonpregnant patients. The intraamniotic environment and physiologic changes of pregnancy may provide a mechanism to explain our findings. Fibrinolytic activity has been demonstrated in amniotic fluid and rises significantly beyond 37 weeks of gestation.16 Perhaps the suppression of fibrinolysis, normally seen with peritoneal suturing, is altered enough by amniotic fluid fibrinolytic activity or in an unknown manner by pregnancy-related changes, such as maternal volume expansion or the presence of inflammatory cytokines, to favor peritoneal closure to reduce adhesions. Further, a cesarean delivery is quite different from nonobstetric surgery. The substerile nature of the operative field, increased blood loss, and type of incision may further alter the healing process and affect adhesion formation.
Several studies in the literature examine the short-term effect of peritoneal closure,2–5 but data on longer-term effects are limited. To our knowledge, only one other study (Myers SA, Bennett TL. The incidence of significant adhesions at repeat cesarean section and the relationship to method of prior peritoneal closure [abstract]. Am J Obstet Gynecol 2002;187(6, part 2):S102 has directly assessed the relation between peritoneal closure at cesarean delivery and adhesions as judged at repeat cesarean delivery. Similar to our results, these authors (Myers and Bennett 2002) identified an association between decreased adhesions with prior peritoneal closure. Roset et al17 conducted an indirect assessment of adhesions and peritoneal closure by following patients previously randomized in a trial of peritoneal closure at cesarean delivery. They identified only 29 patients who had any subsequent abdominal surgery. Adhesions were described in operative reports among 14 patients total: 8 from the nonclosure group and 6 from the closure group. Although ambitious in scope, these data are too limited to generate conclusions.
Unlike other studies,2–4 our study did not identify a significant difference in operative time with peritoneal closure, a process that can be performed quickly. Like others,2,3 we did identify a delay in the return of bowel function beyond postoperative day 1 with peritoneal closure.
We identified possible underlying differences in adhesions based on both payor status and ethnicity, underscoring the importance of using multivariable analysis to control for confounding. Patients receiving public assistance may have more frequent pelvic infections, possibly explaining increased adhesions among this group. Further, there may be ethnicity-driven biological factors that influence adhesion formation. We lack sufficient numbers to address whether some groups are more prone to adhesions and whether they benefit more from peritoneal closure, but this should be relevant for future research.
Our study had several strengths and limitations. Rather than rely on operative dictations, we used a standardized adhesion score sheet that was filled out by surgeons immediately following surgery. Although the adhesion score sheet is a subjective tool, we assessed its interrater reliability and determined that it was acceptable. Our adhesion score sheet enabled us to determine, not only the presence of adhesions, but also their location and quality. To minimize confounding, we excluded patients with adhesions already present at the first cesarean delivery and patients who underwent intervening surgeries and included only patients undergoing a first repeat cesarean delivery.
Our study was limited by its nonrandomized nature. A prospective, randomized study would ideally answer our question; such a study would need to be quite large in scope and of long duration. Our study population comes from a tertiary care training program with tremendous patient heterogeneity and many surgeons at all levels of training. All residents were supervised by an attending physician, but differences in operating times may have reflected different levels of training. Although we excluded women with intervening surgeries, we do not know whether patients acquired sexually transmitted infections between surgeries. Our study was powered to examine the primary outcome of peritoneal adhesions with relation to peritoneal closure, but it was not well-powered to examine the variety of subgroups and secondary outcomes that we mention. To examine some of these outcomes, our study, which took several years, would need to be expanded 4- to 5-fold. Ultimately, whether to close the parietal peritoneum at cesarean delivery may vary by patient and depend on factors such as labor, blood loss, and ethnicity. Our study was not powered to make such distinctions, but this should be an area of future study.
With more than 1 delivery in 4 occurring by cesarean in the United States and with the cost of adhesion-related complications approaching $1.2 billion annually,11 further study of the long-term impact of surgical techniques is needed. Contrary to the existing literature among nonpregnant patients, we found that parietal peritoneal closure is associated with fewer intra-abdominal and pelvic adhesions. Our study suggests that simply extrapolating data from nonpregnant patients is not appropriate. Ideally, questions regarding the impact of closure methods at cesarean delivery on adhesion-related complications, such as difficult repeat surgery, abdominal pain, infertility, and small bowel obstruction, should be addressed in a large, prospective randomized trial.
1. Hamilton BE, Martin JA, Sutton PD. Births: preliminary data for 2003. Natl Vital Stat Rep 2004;53(9):1–17.
2. Pietrantoni M, Parsons MT, O'Brien WF, Collins E, Knuppel RA, Spellacy WN. Peritoneal closure or non-closure at cesarean. Obstet Gynecol 1991;77:293–6.
3. Irion O, Luzuy F, Beguin F. Nonclosure of the visceral and parietal peritoneum at cesarean section: a randomized controlled trial. Br J Obstet Gynaecol 1996;103:690–4.
4. Hull DB, Varner MW. A randomized study of closure of the peritoneum at cesarean section. Obstet Gynecol 1991;77:818–21.
5. Nagele F, Karas H, Spitzer D, Staudach A, Karasegh S, Beck A, et al. Closure or nonclosure of the visceral peritoneum at cesarean delivery. Am J Obstet Gynecol 1996;174:1366–70.
6. Tulandi T, Hum HS, Gelfand MM. Closure of laparotomy incisions with or without peritoneal suturing and second-look laparoscopy. Am J Obstet Gynecol 1988;158:536–7.
7. O'Leary DP, Coakley JB. The influence of suturing and sepsis on the development of post-operative peritoneal adhesions. Ann R Coll Surg Engl 1992;74:134–7.
8. Swanwick RA, Stockdale PHG, Milne FJ. Healing of parietal peritoneum in the horse. Br Vet J 1973;129:29–35.
9. Bamigboye AA, Hofmeyr GJ. Closure versus non-closure of the peritoneum at cesarean section (Cochrane Review). In: The Cochrane Library, Issue 4, 2003. Oxford: Update Software.
10. Tulandi T, Al-Jaroudi D. Nonclosure of the peritoneum: a reappraisal. Am J Obstet Gynecol 2003;189:609–12.
11. Ray NF, Larsen JW, Stillman RJ, Jacobs RJ. Economic impact of hospitalizations for lower abdominal adhesiolysis in the United States in 1988. Surg Gynecol Obstet 1993;176:271–6.
12. Wilkinson CS, Enkin MW. Peritoneal non-closure at caesarean section (Cochrane Review). In: The Cochrane Library, Issue 2, 2000. Oxford: Update Software.
13. Raftery AF. Regeneration of parietal and visceral peritoneum: a light microscopical study. Br J Surg 1973;60:293–9.
14. Duffy DM, DiZerega GS. Is peritoneal closure necessary? Obstet Gynecol Surv 1994;49:817–22.
15. Montz FJ, Shimanuki T, diZerega GS. Postsurgical mesothelial reepithelialization. In: DeCherney AH, Polan ML, editors. Reproductive surgery. Chicago (IL): Year Book Medical Publishers; 1987. p. 31–47.
16. Pschera H, Kjaeldgaard A, Larsson B. Fibrinolytic activity in amniotic fluid during late pregnancy. Acta Obstet Gynecol Scand 1986;65:417–20.
17. Roset E, Boulvain M, Irion O. Nonclosure of the peritoneum during caesarean section: long-term follow-up of a randomised controlled trial. Eur J Obstet Gynecol Reprod Biol 2003;108:40–4.