Hundreds of thousands of US women opt for surgical sterilization procedures each year to permanently prevent pregnancy. Laparoscopic sterilization (LS), a form of tubal ligation, is the most common permanent contraceptive procedure for women outside the postpartum period. In 2002, Essure, a hysteroscopic alternative in which coils are placed in the fallopian tubes, was FDA approved and marketed until 2019. Other forms of hysteroscopic sterilization (HS) are being developed and are expected to come to market. This makes information on the real-world comparative effectiveness and safety of HS relative to LS important, especially for the thousands of US women who currently have HS coils in their bodies.
We compared the real-world effectiveness and safety of HS and LS among non-postpartum women.
We analyzed claims data from women aged 18 to 50 years who underwent Medicaid-funded sterilization procedures in California between January 1, 2008, and August 31, 2014. The primary study outcome was sterilization effectiveness as measured by claims related to pregnancy after HS or LS. We also examined use of prescription contraception after sterilization, which might indicate that sterilization was not effective. Secondary outcomes included need for a repeated sterilization procedure, claims for pelvic pain, abnormal uterine bleeding, pelvic inflammatory disease (PID), abdominal pain and gastrointestinal (GI) symptoms, nonabdominal pain, and rates of additional surgical procedures (eg, hysterectomy). At the request of patient partners, to contextualize early findings, we also investigated whether autoimmune disorders, fatigue, or alopecia were more common after HS or LS. We used propensity-weighted, generalized estimating equation models to control for demographic and clinical factors that might affect outcomes of interest. After the study began, at the request of patient partners, we made additional comparisons of HS and LS with intrauterine contraceptive devices (IUDs) containing either copper or progestin (specifically, levonorgestrel), as data allowed.
We identified 5906 women who underwent HS and 23 965 women who underwent LS. Rates of pregnancy were higher than expected after surgical sterilization, regardless of approach. In unadjusted analysis, in the first year postprocedure, those who underwent HS had higher rates of pregnancy (3.26%) than those who underwent LS (2.61%), but in propensity-weighted adjusted models, there was no significant difference between groups (adjusted incidence rate ratio [aIRR], 1.06; 95% CI, 0.85-1.26). In subsequent years, rates of pregnancy were lower among those who had undergone HS than those who had undergone LS; 5 years postprocedure, unadjusted cumulative rates of pregnancy were 6.26% after HS and 7.22% after LS (propensity-weighted aIRR, 0.76; 95% CI, 0.62-0.90). At 5 years postprocedure, 7.80% of women who had undergone LS had additional claims for contraception, compared with 7.32% of women who had undergone HS. Although few women who underwent HS (23.1%) had claims for the recommended hysterosalpingogram (an outpatient X-ray test that uses dye injected through the cervix to outline the uterus) to confirm tubal occlusion, rates of pregnancy within 1 year after HS were similar whether or not women had the recommended hysterosalpingogram (aIRR, 1.13; 95% CI, 0.71-1.81). Repeated sterilization procedures were more common after HS (5.43%) than after LS (2.25%); in propensity-weighted covariate-adjusted models, women who had an HS were more than twice as likely to undergo a repeated attempt at sterilization within 1 year (aIRR, 3.48; 95% CI, 2.69-4.27) and within 5 years (aIRR, 2.32; 95% CI, 1.84-2.79). Pelvic pain claims were significantly less common after HS than after LS up to 2 years postprocedure (25 days compared with 38 days with 1 or more claim per 100 woman-years of observation, aIRR 0.79 (95% CI 0.66-0.94)). Claims for abdominal pain and other GI symptoms were significantly less common after HS than after LS up to 1 year postprocedure (45 days compared with 66 days with claims per 100 woman-years of observation, aIRR 0.80 (95% CI 0.68-0.93)). Pelvic and abdominal pain claims were most common among women who had claims for these conditions before undergoing their procedure. Women younger than 45 years were more likely than older women to have pelvic pain claims after either HS or LS. Latina women were less likely than White women to have claims related to pelvic pain after sterilization procedures. Claims for PID were less common after HS than after LS up to 2 years postprocedure. There were no differences in claims for nonabdominal pain after HS compared with LS. However, abnormal uterine bleeding claims were more common after HS than after LS up to 12 months postprocedure, and most common 6 months postprocedure (48 days compared with 24 days with a claim per 100 woman-years of observation), with no significant differences between methods after 1 year postprocedure. Procedural complications (eg, need for transfusion) on the day of surgical sterilization and need for additional surgical procedures (eg, hysterectomy) were less common among those who had undergone HS than among those who had undergone LS. Few women had claims related to autoimmune disorders (<1%), fatigue (<1%), or alopecia (≤0.08%) after HS or LS. When we compared outcomes for HS and LS with outcomes in the first year after placement of levonorgestrel IUDs (n = 35 704 women) and copper IUDs (n = 23 628 women), we found that unadjusted rates of pregnancy within 1 year were similar for women who had a levonorgestrel IUD placed (2.40%) or had undergone LS (2.64%), higher for women who had a copper IUD placed (2.99%), and highest for women who had undergone HS (3.34%). In fully adjusted models, compared with LS, rates of pregnancy within 1 year of procedure were higher with HS (aIRR, 1.24; 95% CI, 1.05-1.47), similar with copper IUD placement, and lower with levonorgestrel IUD placement (aIRR, 0.73; 95% CI, 0.65-0.83). Furthermore, women who had undergone HS or received an IUD were less likely than those who had undergone LS to have claims for pelvic pain or PID within 1 year postprocedure. For women with multiple days of claims for symptoms after having their contraceptive procedure, we counted the number of days on which 1 or more claim was submitted per individual during each time period of interest. Women who had undergone LS had an average of 76.93 days of claims compared with 29, 22, and 18.87 days of claims after HS, levonorgestrel IUD placement, or copper IUD placement, respectively, per 100 woman-years of observation in unadjusted models. Women who had undergone HS were more likely to have claims for abnormal uterine bleeding (aIRR, 1.48; 95% CI, 1.41-1.56) compared with those who had undergone LS. Rates of abdominal and nonabdominal pain in the year after placement of either kind of IUD were also lower than those after undergoing LS or HS.
Compared with LS, HS offered some advantages (eg, women were less likely to experience a pregnancy within 5 years) and some disadvantages (eg, women were more likely to need to undergo a second sterilization procedure). For women prioritizing effectiveness and decreased pelvic and abdominal pain over the following year, IUD placement may be preferable to a sterilization procedure. Women desiring permanent contraception via female sterilization should be informed that 5-year pregnancy rates may be as high as 6% after HS or LS.
Claims data likely underrepresent patients' experiences of pain and other symptoms. Few women were continuously enrolled in Medicaid. Although California funds abortion services for California citizens enrolled in Medicaid, because the federal government rarely provides abortion services, the federal data set we analyzed does not include claims for abortion, potentially undercounting poststerilization pregnancies.
