Comparative efficacy of postpartum sterilisation with the titanium clip versus partial salpingectomy: a randomised controlled trial

Authors


MI Rodriguez, World Health Organization, 20 Avenue Appia, 1211 Geneva, Switzerland. Email rodriguezmar@who.int

Abstract

A multicentre trial was conducted to compare efficacy of postpartum sterilisation with clip and partial salpingectomy. Life-table estimates of pregnancy probabilities were compared through 2 years of follow up. A significantly different risk of pregnancy between the clip and partial salpingectomy groups was observed. Nine pregnancies were observed in the clip group versus two in the partial salpingectomy group. Cumulative probability of pregnancy through 2 years was 0.017 with the clip and 0.004 [correction added after online publication 25 September 2012; 0.044 has been replaced with 0.004] for partial salpingectomy (P < 0.04). Equivalent efficacy of the clip compared with partial salpingectomy has not been demonstrated in postpartum women.

Introduction

Postpartum sterilisation is one of the most common methods of contraception globally.1. Annually in the USA, sterilisation follows approximately 8% of all births.2 Rates are believed to be less in the UK, because of the popularity of vasectomy.3 Use of postpartum sterilisation internationally, where access to long-acting reversible contraception is limited, is even higher in many countries. The postpartum period extends up to 6 weeks, but a postpartum sterilisation procedure is typically performed within 48 hours of a delivery.4 Postpartum partial salpingectomy has one of the lowest failure rates of all sterilisation techniques, with a 1-year failure rate of 0.6 per 1000 and a 10-year failure rate of 7.5 per 1000.4

An assortment of techniques is used for tubal occlusion, with excellent efficacy and safety records.1 Partial salpingectomy with the modified Pomeroy or Parkland techniques are two common methods used in postpartum sterilisation.5 Partial salpingectomy is a simple, safe procedure, where a segment of fallopian tube is ligated and removed bilaterally.6 Partial salpingectomy can be more challenging in obese women, or in women with extensive intra-abdominal adhesions.4 The use of occlusive devices, such as a ring or clip placed via laparoscopy, are commonly used methods for interval sterilisation (defined as tubal sterilisation at least 6 weeks following a birth), but have also been used in the postpartum period.5 The use of an occlusive device, rather than transecting and ligating with suture, has the advantages of speed and ease of procedure.

The titanium clip (FemCare-Nikomed, Romsey, UK) is the most commonly used method for laparoscopic interval tubal sterilisations in the USA and the UK.6 Its use for sterilisation was first described in 1981 in a prospective cohort study of 540 British women.5 Pregnancy outcomes and complications were assessed over 5–17 months. The titanium clip has a silicone lining that is applied to the isthmic region of the tube.5 The clip is applied with a unique applicator, and must encompass the entire tube for effective compression. As necrosis ensues, the silicone lining maintains tubal collapse. Applicators have been developed for use in both laparoscopic surgery and with mini-laparotomy.

The largest, long-term study of sterilisation is the US Collaborative Review of Sterilization (CREST) trial, which included postpartum sterilisation by mini-laparotomy, and interval sterilisation using various laparoscopic techniques and mini-laparotomy. Although the titanium clip was not available at the time of the study, the efficacy of the titanium clip has been well established.5,6 However, with regard to postpartum use, data to support routine use of the clip are less clear.3 Only eight postpartum women were included in the original efficacy trial, and their outcomes were not reported separately.5 Additional small studies have supported its efficacy postpartum, and advocates tout its speed and ease of use.7–9 The device is available, and widely marketed for postpartum use. However, no study with the statistical strength to determine equivalent efficacy rates of the postpartum use of the titanium clip versus partial salpingectomy, has been done.

Given the physiological differences occurring in the fallopian tubes with pregnancy, such as oedema and hypertrophy, i.e. swelling and enlargement of the fallopian tubes, results taken from the interval population should not be directly extrapolated to postpartum women.

Given the paucity of data on postpartum efficacy of the titanium clip, a randomised controlled trial comparing titanium clip use with partial salpingectomy was conducted. Our objective was to compare efficacy over 24 months of the titanium clip with the standard Pomeroy approach in women undergoing postpartum sterilisation.

Methods

FHI 360, formerly known as Family Health International, conducted a multi-centre, international trial to compare the efficacy of titanium clip use with the Pomeroy technique in postpartum sterilisation. The primary outcome was pregnancy rates over 24 months following sterilisation. Data from this trial were reported to the Food and Drug Administration as part of the device’s approval process in the USA.

Studies were conducted at four independent centres in Thailand, Taiwan, Panama and the Philippines. The centre in Panama completed two studies during the period. Identical protocols were followed at all centres, and regular monitoring and an audit of each centre were performed. Approval by an Institutional Review Board was obtained before study initiation at each site.

Contraceptive counselling was offered to women as part of routine antenatal care. Women electing sterilisation, were screened for study eligibility. Women were eligible for study inclusion if they were legally able to consent to sterilisation, were 21 years of age or older, had a normal physical and pelvic examination, and were within 42 days postpartum. Exclusion criteria were: legally incapable of consenting, had severe, pre-existing systemic disease, profound anaemia (haematocrit < 30%), anticipated concurrent surgery other than curettage, limited accessibility to centre for follow up, caesarean section for this pregnancy, or request for sterilisation for reason other than voluntarily limiting family size.

A total of 1400 women were enrolled between April 1984 and June 1989 (see Supplementary material, Figure S1). At admission to the study, a medical history and physical examination (including pelvic examination) were performed. After consent, directly before surgery, assignment to titanium clip or Pomeroy technique was performed by random allocation, using an off-site computer-generated code that was unavailable to study staff. Although the surgeon could not be blinded to the procedure used, the investigator in charge of follow up was. The woman was admitted to the study by the investigator who performed the surgery. All surgeries were performed following a vaginal delivery. Sterilisations occurred within 72 hours of delivery. All procedures were performed by an infra-umbilical mini-laparotomy incision that was 1–2 cm in length. Fallopian tubes were visually identified, and elevated to the level of the incision with forceps. Tubal occlusion occurred under direct visualisation. The Mark VI model of the Filshie clip was used for all women randomised to clip sterilisation. Pomeroy method was the standard of care, and so was picked as the comparator. Sterilisation by Pomeroy method was accomplished by identification of a 2-cm length of tube. The tube was lifted with forceps, creating a partial loop, which was then ligated with suture. The tubal segment was then excised with scissors, and haemostasis was assured at the base. Both types of surgery were performed under local, epidural or spinal anaesthesia, at the surgeon’s discretion.

Women were asked to not use other forms of contraception, and were scheduled to return to clinic for follow-up 1, 6, 12 and 24 months following sterilisation. A complete physical examination, including a pelvic examination, was performed at each follow up. All suspected pregnancies were evaluated with an examination and laboratory studies as necessary. If women missed their clinic visit, they were contacted by phone, and requested to return to the clinic. If unable to do so, women were asked about menstrual irregularities, pregnancy symptoms or any adverse symptoms.

Planned enrolment was 300 women per site. No formal power calculation was included before study initiation. We conducted a retrospective power calculation. Analysis was based on intention to treat. The life-table method was used to estimate the gross cumulative life-table pregnancy rate by treatment group. One-month intervals were used for life-table analyses. The intrauterine pregnancy survival curves were compared using the log-rank test. For the purposes of tracking participants, personal contacts with participants completed at least 684 days since randomisation were counted as 24-month visits. Women were terminated from the study analysis for pregnancy, a serious adverse reaction, or if the investigator determined that continuation was detrimental to the physical or mental health of the woman. For life-table analyses of pregnancy, the estimated date of conception was used as the event date for women who became pregnant. Date of last clinic, home or phone/letter contact was used as the censoring date for women who did not become pregnant.

All statistical analyses were performed using Statistical Analysis System (SAS) software, version 5.18 (SAS Institute Inc., Cary, NC, USA).

Results

Fourteen hundred women were enrolled in the study with 698 randomly assigned to receive the titanium clip and 702 to partial salpingectomy (see Supplementary material, Figure S1). Demographic variables, including obstetric and contraceptive history, were comparable between groups at baseline (see Supplementary material, Table S1). Mean age of participants in the study was 30.1 years. Women in both groups had a median of three live births. Contraceptive history in the months before conception of the pregnancy preceding sterilisation was similar in the study and control groups, with most women in both groups relying on either an oral contraceptive or no method. Type of follow up was similar between groups. At 1 month, 92.3% of the titanium clip group and 88.9% of the Pomeroy group were seen in clinic, at home, or reached by phone. At 6 months, this number had dropped to just over 70% for both groups.

Thirteen technical failures and random allocation errors occurred. Nine pregnancies occurred in the titanium clip group, and two in the partial salpingectomy group. All pregnancies were noted to be intrauterine. None of the pregnancies occurred in women with technical failures or random allocation errors.

At 1 year, 43% of the titanium clip group could not be reached by clinic or home visit, letter or phone call. In the Pomeroy group, 41.9% were not reached at 1 year following sterilisation. By study end (24 months), 51% of the titanium clip group and 51% of the Pomeroy group had been lost to follow up.

Given the number of women present at the end of the study (348 in each group), the power to detect a significant difference in the pregnancy rates between the two treatment groups, of the size observed, with the two-sided log-rank test (at α = 0.05) is 53%. Cumulative pregnancy rates were significantly different between the two groups (Table 1). Through 24 months, the estimate of the pregnancy probability for titanium clip group was 0.017, compared with 0.004 in the Pomeroy group (P = 0.04).

Table 1. Gross life-table pregnancy rates by centre and sterilisation method
CentreTitanium clipPartial salpingectomy P value
n Cumulative pregnancy probabilitySE n Cumulative pregnancy probabilitySE
  1. NA, not available. Greenwood’s formula was used to estimate standard errors (SEs) so the estimates are expected to be biased downward.

Centre 0075
1 month134.50.000NA132.50.000NA0.32
6 months118.00.000NA114.00.000NA
12 months109.00.0180.012108.50.000NA
24 months79.00.0280.01681.00.0110.011
Centre 0600
1 month124.50.000NA133.00.000NA0.15
6 months98.00.000NA103.00.000NA
12 months86.50.0220.01684.00.000NA
24 months53.00.0220.01645.00.000NA
Centre 0781
1 month99.50.000NA100.00.000NA0.32
6 months93.50.000NA95.50.0100.010
12 months88.00.000NA86.00.0100.010
24 months68.50.000NA64.50.0100.010
Centre 0832
1 month296.50.000NA299.00.000NA0.04
6 months290.50.000NA296.00.000NA
12 months265.00.0070.005267.00.000NA
24 months147.50.0150.008157.50.000NA
Total
1 month655.00.000NA664.50.000NA0.04
6 months6000.000NA608.50.0020.002
12 months548.50.0110.004545.50.0020.002
24 months348.00.0170.006348.00.0040.003

Discussion

This study represents the largest, randomised, controlled trial that evaluates the efficacy of titanium clip use compared with partial salpingectomy in postpartum sterilisation. Although the pregnancy rates in both groups indicate that either method provides substantial protection against pregnancy, the data suggest that use of the titanium clip is significantly less effective than partial salpingectomy at preventing pregnancy in the postpartum population.

Our study is not without significant limitations. This trial was conducted over 20 years ago, when standards for scientific research were considerably different. For example, a power calculation is now a routine part of study design. All of the study centres were in the developing world, which may affect the generalisability of findings. Additionally, efforts are made to minimise loss to follow up rates to below 5%, to reduce bias in modern trials. This is a serious limitation in the data that we present. At 24 months, our loss to follow-up rate was high, at just over 50% in both groups. Although loss to follow up was evenly distributed between both study and control group, this high risk of a loss to follow up means that definitive conclusions should not be drawn from these data.

Additional limitations include that follow up was limited to 24 months following sterilisation. The CREST study provides evidence that risk of pregnancy persists for at least 10 years following sterilisation, especially among women sterilised before 28 years of age. We were unable to stratify analyses by nursing status, which affects fertility and may limit generalisability. Importantly, these limitations would probably result in an underestimate of the pregnancy rate to be expected in the population over the participants’ reproductive life times. Regardless of the limitations in this trial, it remains the only randomised trial, conducted to date, with sufficient power to compare efficacy rates of the titanium clip with partial salpingectomy in the postpartum population. This data set, with all its flaws, represents the highest level of evidence for use of the titanium clip postpartum, a practice for which the device is widely marketed, and used.

Previous reports of the titanium clip have provided mixed results among postpartum cases.10 Data from three of the four centres in this trial have been published separately.8,11 Yan et al.8 reported the results for 200 women cared for in Taiwan. The data for women sterilised by both clip or partial salpingectomy were reported. Outcomes for women sterilised by clip only were reported for women from the Philippines (n = 148) and Panama (n = 297) in a subanalysis that focused on whether timing of the surgery (before or after 48 hours postpartum) affected complication rates.11 Neither of these studies had adequate statistical power to compare efficacy. Graf et al.9 reported no pregnancies in a series of 282 women, but their results might have been a result of better than average surgical skills, or simply due to chance. Whereas Oligbo et al.12 reported data on too small a population to identify a statistically significant difference in pregnancy rates, although the 1.5% pregnancy rate in his clip cases is similar to the 1.7% cumulative rate in our data. Interestingly, Oligbo et al.’s series only included sterilisations at time of caesarean delivery. This suggests that the failure rate observed with the clip might be the result of a mechanism other than misplacement of the device. Importantly, none of these studies have been powered to evaluate efficacy.

Sterilisation is requested by women who are certain of their decision to not have additional pregnancies. Women requesting sterilisation should be offered a method with known efficacy. To establish equivalence of the titanium clip with partial salpingectomy in postpartum use, a well-designed clinical trial is needed. The trial must include an appropriate comparison group, adequate numbers to assess statistical difference or non-inferiority, and a lower loss to follow up than achieved in this study. In the absence of such a trial, the equivalence of postpartum sterilisation with the titanium clip compared with partial salpingectomy is unknown.

Disclosure of interest

The authors report no conflicts of interest.

Contributions to authorship

All authors reviewed the study protocol and results, and discussed analyses. MIR and DS wrote the paper. AS contributed commentary on the statistical methods, and provided the retrospective power calculation.

Details of ethics approval

An institutional review board was completed at all study sites.

Funding

Funding for the original clinical trial was obtained from the US Agency for International Development (USAID) under cooperative agreements AID/DPE-0537-C-1028-00, AID/DPE-0537-A-00-4047- 00, AID/DPE-3041-A-00-0043-00 and AID/CCP-3079-A-00-5022-00. Femcare Limited, the manufacturer of the titanium clip system (Filshie Clip System) supplied the clips and applicators at no cost. Per USAID policy on cost-recovery at the time the study was conducted, FHI 360 negotiated a royalty agreement with FemCare Ltd. FHI 360 receives a royalty from Femcare Ltd. on clips sold to private sector, non-public institutions in the USA. This paper was prepared with partial support from FHI 360 and from the World Health Organization.

Acknowledgements

We are grateful to the principal investigators at study sites: Dr Suporn Koetsawang, Dr Ruben Apelo, Dr Jaw-Shong Yan and Dr Jorge Lasso de la Vega. The opinions expressed are those of the authors and not necessarily those of FHI, USAID or the World Health Organization.

Ancillary