Tuuli, Methodius G. MD, MPH; Odibo, Anthony O. MD, MSCE; Fogertey, Patricia MSN; Roehl, Kimberly MPH; Stamilio, David MD, MSCE; Macones, George A. MD, MSCE
Cesarean delivery is the most common major surgical procedure in women worldwide. In the United States, almost 33% of the 4 million annual births are cesarean deliveries.1 In performing this procedure, it is imperative that evidence-based techniques are utilized to minimize complications and optimize outcomes. Recently, the case has been made for simplifying the procedure by eliminating unnecessary and potentially harmful steps.2
Creation of the bladder flap has been an integral part of the standard cesarean delivery without evidence of benefit. The practice dates back to the era before antibiotics, when its creation and subsequent closure was thought to protect the peritoneal cavity from spreading intrauterine infection. Its creation was also intended to enable the surgeon to gain access to the lower uterine segment while minimizing injury to the bladder. Yet, evidence on the role of the bladder flap at cesarean delivery is limited.3,4
We conducted a randomized controlled trial to test the hypothesis that omission of the bladder flap in primary and repeat cesarean deliveries shortens operating time without significantly increasing intraoperative and postoperative complications.
MATERIALS AND METHODS
This was a randomized controlled trial comparing the effects of omitting the bladder flap from the standard cesarean delivery technique. The study was approved by the Washington University Human Research Protection Office and was registered with the clinical trials registry (ClinicalTrials.gov # NCT00918996).
The study was conducted at Barnes Jewish Hospital, a tertiary teaching hospital. Women undergoing primary or repeat cesarean delivery at gestational age of 32 weeks of gestation or more were eligible. Women undergoing emergent cesarean deliveries, planned vertical uterine incision, and those with previous laparotomies besides cesarean deliveries were excluded. Eligible patients scheduled for cesarean deliveries were approached by the research nurse or other members of the research team for consent to participate in the study. For unscheduled patients, consent was sought once the decision was made to perform a cesarean delivery. Enrolled patients were stratified into primary and repeat cesarean deliveries and randomly allocated to the bladder flap group or no bladder flap group using computer-generated random numbers concealed in opaque envelops. Patients were blinded to their group allocation. They were not told at any point during surgery or follow-up whether they were allocated to the bladder flap group or no bladder flap group. They also were not informed whether the allocated procedure was actually performed.
Before entering the operating room, the surgeon opened a sealed envelope containing the randomization number and the patient's allocation. The technique of cesarean delivery was similar in the two groups, except for omission or creation of the bladder flap. In the no bladder flap group, a low-transverse uterine incision was made approximately 1 cm above the vesico-uterine peritoneal fold without dissection and formation of a bladder flap. Patients underwent the procedure based on their allocation unless intraoperative findings as determined by the surgeon made the alternative procedure medically necessary.
The circulating nurse documented intraoperative information on a data collection form: incision time, delivery time, fascial closure time, skin closure time, type of uterine incision, creation or omission of the bladder flap, and any intraoperative complications. The research nurse abstracted demographic information and antepartum, intrapartum, and delivery information from patients' charts.
On the first postoperative day, patients were asked to score their pain on a scale of 0 (no pain) to 10 (worst pain) immediately before removal of their indwelling Foley catheter. A urine sample was collected directly through the Foley catheter before its removal and was tested with dipsticks for microhematuria. Postoperative hemoglobin levels were measured and compared with preoperative values to obtain the changes in hemoglobin.
At 4–6 weeks postoperatively, patients were contacted by the research nurse by telephone and asked about their postoperative course, including pain score on a scale of 0 to 10, symptoms of infection, and visits to the emergency department or their physician's office for postoperative complications. At the patients' 4- to 6-week postpartum visits, clean void urine samples were collected and tested with dipsticks for blood, leukocyte esterase, and nitrite. Reflex urine cultures were performed for samples suspicious for urinary tract infection.
Because there was no previous published data on omitting the bladder flap at repeat cesarean deliveries, an independent three-member data safety and monitoring board conducted one planned interim analysis after half of the patients undergoing repeat cesarean deliveries were evaluated. None of the investigators was involved in the interim analysis. The significance level for the interim analysis was calculated using the O'Brien-Fleming stopping boundaries.5
We estimated the required sample size a priori based on the primary outcome of total operating time. The sample size was estimated separately for primary and repeat cesarean deliveries. For primary cesareans, we estimated that a total of 128 patients (64 in each group) would be needed for 80% power to detect a 5-minute difference in total operating time, based on a mean duration (±standard deviation) of 45 (±10) minutes (two-sided test and α of 0.05). For repeat cesarean deliveries, we anticipated a larger savings of time; however, to be conservative, we estimated that 130 patients (65 in each group) would provide 80% to detect, at least, a 5-minute difference in total operating time (two-sided test and nominal α of 0.047, incorporating the one interim analysis in this group and using the Obrien-Fleming stopping rule to adjust for type I error5). Therefore, the total sample size was 258 (64×2+65×2).
Analysis was based on the intention-to-treat principle, in which patients were analyzed in the groups to which they were randomized whether the assigned procedure was performed. Selected baseline characteristics and most outcome measures were compared between the two groups using the unpaired Student t test for normally distributed continuous variables and the χ2 test or Fisher exact test for categorical variables as appropriate. Continuous variables that were not normally distributed as assessed by the Kolmogorov-Smirnov test were logarithmically transformed. Data not normally distributed even after logarithmic transformation were compared using the Mann-Whitney U test.
We tested for interaction between the creation or omission of the bladder flap and primary or repeat cesarean delivery on total operating time using a linear regression model with an interaction term. In the absence of significant interaction, we performed combined analysis of both primary and repeat cesarean deliveries. We also performed prespecified stratified analysis for primary and repeat cesarean deliveries. To account for the interim analysis, we considered tests with P<.047 as significant. Analysis was performed using Stata 11.
A total of 259 women were enrolled from March 2010 to May 2011 (Fig. 1). One woman in the bladder flap group was lost to follow-up, yielding 131 patients in the bladder flap group and 127 in the no bladder flap group. Of the 131 patients assigned to the bladder flap group, 108 (82.4%) received the bladder flap, whereas 23 (17.6%) did not. Reasons for not creating the bladder flap were presence of scar tissue (n=13), need for a vertical uterine incision (n=1), and other reasons (n=9). Among the 127 women allocated to the no bladder flap group, the bladder flap was omitted in 113 (89.0%), whereas 14 (11.0%) had bladder flaps created. The most common reason was presence of scar tissue (n=9).
Women assigned to the bladder flap or no bladder flap group were similar regarding baseline demographic, pregnancy, and operative characteristics (Table 1). Specifically, the experience level of the primary surgeon was similarly distributed between the two groups. Approximately half of the patients in each group had primary cesarean deliveries. The most common indications for cesarean delivery were elective repeat, failure to progress, fetal intolerance of labor, and malpresentation. These were evenly distributed between the two groups. Most cesarean deliveries were low transverse, and the skin was most commonly closed with staples.
Because we noted no significant interaction between creation or omission of the bladder flap and primary or repeat cesarean delivery on the primary outcome and no significant effect size for either stratum of cesarean delivery type, we present results for the combined analysis (primary and repeat cesarean deliveries). Duration from skin incision to delivery was shorter among women assigned to the no bladder flap group compared with those in the bladder flap group (median time 9 [range 1–43] compared with 10 [range 2–70] minutes; P=.04; Table 2). There was no difference in median total operating time between the two groups (51 [range 18–124] minutes compared with 51 [range 16–178]; P=.10). No bladder injuries occurred in either group, and there were no significant differences in estimated blood loss (800 [range 300–2,000] compared with 800 [range 400–3,500] mL), change in hemoglobin level (−1.7 [range −6.5 to 1.5] compared with −1.6 [range −4.4 to 1.2] g/dL), postoperative day 1 microhematuria (54.7% compared with 44.5%; P=.14), postoperative day 1 pain score (median 4 [range 0–10] compared with 3 [range 0–9]; P=.69), hospital days (median 4 [range 2–35] compared with 4 [range 2–35]; P=.71), endometritis (3.2% compared with 2.3%; P=.72), or urinary tract infection (1.6% compared with 0.8%; P=.62). Pain score and microhematuria also were not significantly different at 4–6 weeks postpartum.
When analysis was stratified into primary or repeat cesarean deliveries, similar findings were noted with the exception of incision-to-delivery time, which was significantly shorter in the no bladder flap group among primary cesarean deliveries but not among repeat cesarean deliveries (Table 3).
We found that omission of the bladder flap during cesarean delivery was associated with shorter incision-to-delivery time, whereas total operating time was unchanged. We also found no associated increase in intraoperative or postoperative complications.
The only other published randomized trial on short-term effects of omitting the bladder flap at cesarean delivery reported benefits, including a decreased incision to delivery interval, total operating time, blood loss, microhematuria, and analgesia requirement when the bladder flap was omitted.6 That study was limited to white women undergoing primary cesarean delivery. We included both primary and repeat cesarean deliveries and enrolled a more diverse group of patients, making our results potentially more generalizable. Inclusion of both primary and repeat cesarean deliveries also enabled us estimate whether the benefits or risks of omitting the bladder flap differ for primary and repeat cesarean deliveries. Another strength of our study is the randomized design that resulted in two groups that were similar regarding baseline characteristics. Follow-up was also thorough, with only one patient lost to follow-up for the primary outcome.
Despite these strengths, there are limitations that should be considered when interpreting our results. The nature of the intervention did not permit blinding of the surgeons. The absence of blinding means operator behavior could have been modified by knowing whether or not a bladder flap was performed. We followed-up patients only until 4–6 weeks postoperatively for complications. Although this is an improvement on the previous study in which outcomes were reported only until hospital discharge,6 long-term effects of omitting the bladder flap were not assessed in this present study. However, we speculate that the separation of tissue through a natural plane rather than traumatically creating a bladder flap will be associated with fewer adhesions. This is consistent with results of a recent study by Malvasi et al7 that reported decreased adhesions and submesothelial fibrosis at subsequent cesarean delivery when the bladder flap was omitted. We plan surveillance of women in our trial to assess long-term effects of omitting the bladder flap, including adhesions and outcomes of trial of labor. Patients who subsequently undergo repeat cesarean delivery in our hospital within the next 5 years will be assessed for duration of the subsequent surgery as well as for the location and severity of intraperitoneal adhesions.
Finally, our study, like the other trial,6 was not powered to assess the effect of omitting the bladder flap on bladder injury. Although there were no cases of bladder injury, the low incidence of bladder injury during cesarean delivery means an impractically large sample size would be required to detect any differences.6,8,9 However, other reports have linked difficulty encountered while creating the bladder flap, rather than its omission, as the risk factor for bladder injury at cesarean delivery.8,9
The choice of operating time as the primary outcome deserves comment. It is difficult to select an appropriate short-term primary outcome for a study evaluating the effect of omitting a step that is hypothesized to be unnecessary, but not necessarily beneficial or harmful. Although it would have been informative to power the study to estimate the effect of omitting the bladder flap on significant intraoperative and postoperative morbidity, it is not biologically plausible to link creation or omission of the bladder flap to most of such outcomes at cesarean delivery. Bladder injury that could be directly linked to the bladder flap in the short-term occurs infrequently at cesarean delivery, and we would require an impractically large sample size to detect any differences. We therefore chose operating time as the primary short-term outcome with the hypothesis that omitting the bladder flap will decrease operating time without increasing complications. Reduced operating times have been associated with reduced postoperative complications at cesarean delivery.10 The duration difference of 1 minute observed in our study is a difference in median duration, a summary statistic that can mask important variability and differences in subpopulations. Of note, the observed incision-to-delivery durations were quite variable with ranges as high as 43 and 70 minutes in the no bladder flap and bladder flap groups, suggesting that actual duration differences could be much greater for some patients.
Unlike in the previous trial,6 we noted no differences in total operating times. This may be the result of the multiplicity of variables that affect total operating time and the wide variation in total operating times noted in our study. Because our study was conducted in a teaching hospital and cesarean deliveries were performed by surgeons of varying levels of experience, there was a wide distribution of operating times. These were not normally distributed even after logarithmic transformation, precluding use of parametric statistical methods for analysis and making it more difficult to detect significant differences. In addition, we noted a significant crossover effect, in which 17.6% of patients randomized to the no bladder flap group had a bladder flap and 11.0% of those randomized to the bladder flap group did not have a bladder flap. This would tend to bias our results toward the null hypothesis of no difference. However, our results were unchanged when women were analyzed “as treated.” The average body mass index (BMI, calculated as weight (kg)/[height (m)]2) of women in our trial was more than 40, with 88.4% obese (BMI more than 30). Whereas BMI was not reported in the earlier trial,6 the high BMI of the women in our study may have contributed to the overall duration of surgery and reduced the effect of omitting the bladder flap.
In conclusion, although not powered to detect differences in rare outcomes such as bladder injury, results of this randomized controlled trial suggest that omission of the bladder flap at primary and repeat cesarean deliveries does not increase intraoperative or postoperative complications. It may shorten the interval from skin incision to delivery, albeit by only a median time of 1 minute. These findings, together with results of another study suggesting increased adhesions, may argue against the practice of routinely creating the bladder flap at cesarean deliveries in the absence of a specific indication for bladder dissection.