This was a single center, non-inferiority, two–arm randomized controlled trial. The trial was conducted between May17, 2017 and July 16, 2018 at the Mayo Clinic in Rochester, Minnesota and was approved by the Mayo Clinic institutional review board (17–000898). The study was registered on ClinicalTrials.gov, NCT03135431, on 27/04/2017.
Women were stratified by number of prior cesarean deliveries and Body Mass Index (BMI). Mean difference (post- vs. pre-procedure) in hemoglobin of participants undergoing cesarean non-emergent delivery with BPS via Pomeroy or Parkland method was compared to those undergoing cesarean delivery with BTS. Inclusion criteria included women 21 years of age and older who desired permanent contraception and had an obstetric indication for a cesarean delivery. Women were excluded if they needed an emergent or immediate cesarean delivery, their BMI was greater than or equal to 50 kg/m2, and or if they had a single ovary/fallopian tube.
The study coordinator approached women at their scheduled obstetric clinic visit to inform them about the trial. In addition, women who were admitted to the hospital for inpatient care, and whose cesarean delivery was not deemed an emergency, were informed of the trial by the study coordinator or one of the investigators. The informed consent was written in English; however, women who did not speak English were consented with the aid of an interpreter. After informed consent for the study was obtained, the participants were randomly assigned to either BTS or BPS using an computer based electronic dynamic allocation platform based on the Pocock-Simon algorithm according to the number of prior cesarean deliveries (first vs. repeat) and BMI (< 35 vs. ≥35 kg/m2). Women who were enrolled and randomized prior to delivery underwent their assigned sterilization method regardless of the urgency of their cesarean delivery. The participants were blinded to the operative intervention at the time of the randomization, and this was later disclosed to them after their cesarean delivery.
Following enrollment, a pre-operative complete blood count (CBC) was collected from participants at the time of their preoperative labs or upon hospital admission, but not to exceed 72 hours before delivery. On admission, women were asked if they wanted to proceed with permanent contraception and consent was obtained for cesarean delivery and the assigned procedure. In the operating room, the delivering team proceeded with the cesarean delivery as indicated. Once completed, the feasibility of the participants’ assigned contraceptive procedure was assessed. The start and stop time of the permanent contraceptive procedure was recorded by the nurse in the room and collected by the study coordinator. Participants assigned to BPS did so via Pomeroy or Parkland method based on the surgeon’s preference. BTS was standardized using the technique described by Hall, et al. (ACOG May 2017 Film Festival, A Novel, Safe, Low Cost Approach to Bilateral Salpingectomy at Cesarean Section) using clamps and suture (13). This method involves using cautery to skeletonize the mesosalpinx, leaving only perforating vessels which were then suture ligated. Surgeons to be performing the bilateral total salpingectomies were first demonstrated this surgical technique via a training module and they then had to perform one proctored bilateral total salpingectomy and be checked off as competent by another competent surgeon prior to performing the bilateral total salpingectomies on their own on participants. Estimated blood loss (EBL) was calculated for the combined cesarean and sterilization procedure using visual estimation and recorded per standard practice. A post-operative CBC was collected 24 to 48 hours after the procedure. Following the participants’ six week post-partum visits, their charts were reviewed for data collection including pre and post-operative hemoglobin, EBL, post-operative pain scores, postoperative length of stay, and adverse events (significant postoperative pain, anemia, need for blood product transfusion, cesarean hysterectomy, loss of one or both adnexa, and return to the operating room and or death). Pain scores were based on a Numeric Pain intensity Score of 0–10. Post-operative length of stay was calculated using date of discharge minus date of surgery. All abstraction was performed by a single study coordinator who used a hard copy of the electronic data capture system (REDCap) data instrument to record information from charts and then input them into REDCap electronically. She then performed quality checks intermittently by cross-checking REDCap entries every 2–3 data points in the hardcopy abstraction tool. In addition, post-operative complications and adverse events were abstracted from the chart.
Due to the lack on pre-existing studies at the time this trial was started, sample size was initially calculated based on previous studies that reported a mean (SD) drop in hemoglobin of 1.32 g/dl (0.94) following a repeat cesarean delivery and 1.40 g/dl (1.2) following a first cesarean delivery.7 Based on a review of indications of cesarean deliveries performed at the study site in 2015, we anticipated that over 75% of the cases would be repeat cesarean deliveries. As a non-inferiority study, we chose a threshold of a difference of .5 g/dl, approximately one-half a unit of packed cells, as a clinically important difference between BPS and BTS. Assuming that a difference of 0.5 g/dl in the drop in hemoglobin between study arms would be considered as equivalent, and assuming a common standard deviation of 1.1 based on previous studies of cesarean deliveries, the study would have 80% power to test for non-inferiority with 60 participants in each arm (120 total). Although .5 g/dl clinically is often negligible a conservative approach was used to ensure finding a difference if there was one. This calculation was based on a two-group 1-sided t-test with a type I error of 0.05. Approximately 1 year after our study commenced, data from a separate, retrospective study on salpingectomy at the time of cesarean delivery within the study site’s larger health system revealed the mean decrease in hemoglobin among women who underwent salpingectomy (n = 30) was 1.6 g/dl (SD 0.7) and 1.8 g/dl (SD 0.9) for women who underwent BPS (n = 43)(12). Based on this information, we conducted a post-hoc power analysis using a common standard deviation of 0.7 instead of 1.1 and the number of accrued participants at the one-year mark of study recruitment. Based on this calculation, our study was well powered (84%) to assess our primary aim with 38 participants (18 BTS, 20 BPS); therefore, we discontinued recruitment and completed the study with the accrued subjects.
The data monitoring safety plan (DSMP) included a review of charts after the first 10 patients and then quarterly for the duration of the trial. The clinical trial was to be placed on hold if an increase in adverse events were noted. Stopping rules included greater than 20 cases of mild (increased pain scores, mild anemia) or moderate (severe anemia not requiring transfusion) adverse events; greater than 3 cases of severe adverse events (severe anemia requiring transfusion, loss of one or both ovaries); and or any life threatening adverse event (cesarean hysterectomy, return to the operating room) or death. If the above criteria for discontinuation are met, the trial was to be put on clinical hold. A Data Monitoring and Safety committee will be convened to assess the causal linkage between the adverse events and the procedure. If causal linkage is established, the study would have been permanently stopped. Participants would still have been followed, with their permission, even if the study was discontinued. If no causal linkage was established the study would have been taken off hold.
No interim analysis was performed. The data were descriptively summarized using means and standard deviations or medians and interquartile ranges for the continuous variables, and frequencies and percentages for the categorical variables. The primary endpoint, change (post- minus pre-procedure) in hemoglobin was calculated using a one-sided two sample T-test for non-inferiority. Secondary endpoints were analyzed using the two-sample T-test, Wilcoxon Rank Sum, Fishers exact test, and chi-square test, as appropriate. Statistical analyses were performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC) and R 3.1.1 (R Foundation for Statistical Computing, Vienna, Austria).