Female sterilization is a common form of contraception in the United States, with more than 10 million women relying on this method.1,2 Interval sterilizations are those that do not occur immediately postpartum and can be performed either laparoscopically or hysteroscopically based on patient and health care provider factors and preferences. Hysteroscopic sterilization does not require general anesthesia or abdominal entry and can often be performed as an outpatient procedure.3 With this method, follow-up with a confirmatory hysterosalpingogram is recommended to ensure tubal occlusion; however, many women do not complete this procedure.3 Laparoscopic sterilization is immediately effective but incurs additional recovery time as well as risks associated with intraabdominal surgery and general anesthesia.3 Our previous research showed that hysteroscopic sterilization is associated with more menstrual dysfunction and future hysteroscopic procedures, but laparoscopic sterilization was associated with more subsequent pelvic pain and future intraabdominal surgery.4
All methods of sterilization carry the risk of failure. The average risk of pregnancy after sterilization is 0.13%, but varies based on procedure approach.5 For example, the 5-year risk of pregnancy after laparoscopic sterilization ranges from 2.3% for unipolar cautery to 16.5% for bipolar fulguration.3,5 Bayer's initial study of Essure hysteroscopic sterilization yielded a 5-year risk of pregnancy of 1.64%.3 There are no prospective studies directly comparing failure rates after laparoscopic and hysteroscopic interval sterilization and a paucity of data regarding hysteroscopic sterilization pregnancy outcomes outside the context of a clinical trial. Our prior retrospective study using the MarketScan database6 indicated higher rates of pregnancy after hysteroscopic sterilization than laparoscopic sterilization (1.02 pregnancies/100 person-years vs 0.88/100 person-years, P=.003), although this difference reversed when considering only women who completed confirmatory hysterosalpingogram after hysteroscopic sterilization (0.76 pregnancies/100 person-years vs 0.88/100 person-years, P=.031).4
There are limited data on the outcomes of pregnancies that occur after sterilization, and most studies focus on rates of ectopic pregnancy (17.1/every 1,000 laparoscopic fulguration procedures).3 Moreover, there has been controversy around hysteroscopic sterilization that has led to questions around its safety and outcomes. In 2015, a member of the U.S. Congress, in conjunction with a private consulting firm, publicized that more than 300 fetal deaths had occurred in women who had undergone hysteroscopic sterilization, which included losses before 20 weeks of gestation (including spontaneous and induced abortions) and those at greater than 20 weeks of gestation; the U.S. Food and Drug Administration reports cases of fetal deaths only past 20 weeks of gestation.7,8
The purpose of this retrospective cohort study was to investigate obstetric outcomes of patients who became pregnant after sterilization to assess whether there is a difference in pregnancy outcomes based on type of sterilization. Our primary outcome was to determine whether there is a difference in this population in obstetric outcomes between those who underwent hysteroscopic sterilization as compared with laparoscopic sterilization. Our secondary outcome was to determine specifically whether there were “negative” pregnancy outcomes such as increased rates of ectopic pregnancies or preterm births in women who had previously had hysteroscopic sterilization as compared with laparoscopic sterilization.
MATERIALS AND METHODS
We extracted data from the Truven Health Analytics MarketScan Commercial Claims Database (MarketScan).4,6 MarketScan includes employer-based health insurance claims from more than 350 payers for the continuum of outpatient and inpatient care for employees and their dependents through age 64 years. We searched for evidence of sterilization in all women included in the database for the years 2007–2013 (N=62.5 million). The data are statistically deidentified, and the Boston University institutional review board deemed this project not human subject research. Please see Figure 1 for a flowchart of participant exclusion.
We examined administrative claims for all women who had either a laparoscopic or hysteroscopic sterilization during the study period and who were included in the MarketScan database, indicating continuous insurance coverage with a single insurance plan for 12 months before and at least 12 months after the sterilization procedure. We did not include women who had postpartum sterilization because we were specifically comparing pregnancy outcomes between women with failed laparoscopic and hysteroscopic sterilizations. We did not include salpingectomy as a means of laparoscopic sterilization in our search because that procedure is also performed for other purposes (such as cancer risk reduction) and was not as common for sterilization during our time period. Among those women who had at least one claim indicating a pregnancy, we identified those pregnancies for which outcomes of live birth, stillbirth, spontaneous abortion, therapeutic abortion, ectopic pregnancies, or trophoblastic disease could be determined.9
Current Procedural Terminology (CPT) and International Classification of Diseases, 9th Revision (ICD-9) codes were used to identify women undergoing laparoscopic sterilization (CPT code 58,670 or 58,671 or ICD-9 procedure code 66.2x or 66.3x) and hysteroscopic sterilization (CPT code 58,565). We identified pregnancy outcomes based on a previously published coding algorithm9 supplemented by a manual chart review by two researchers (K.B. and K.O.W.; Box 1). This algorithm has been compared previously with actual outcomes and has been found to accurately predict outcomes.9 For outcomes that could have occurred during the first trimester, including spontaneous abortion, therapeutic abortion, and diagnosis of ectopic pregnancy or trophoblastic disease, we considered CPT and ICD-9 codes as valid if they occurred at least 4 weeks after hysteroscopic or laparoscopic sterilization. Codes for a live birth were considered valid if they occurred at least 6 months after the procedure to avoid pregnancy antecedent to or concurrent with sterilization. We also conducted a sensitivity analysis excluding women with both spontaneous abortion and therapeutic abortion because these diagnoses can be difficult to distinguish in medical coding.
Codes Used to Identify Live Birth in Ambiguous Outcome Charts
Very Likely Live Birth
- Examination of the lactating mother
- Breast pump
- Normal delivery
- Postpartum sterilization
Moderately Likely Live Birth
- Routine postpartum follow-up
- Twin gestation with fetal loss and retention of 1 fetus
- 1st-, 2nd-, or 3rd-degree perineal laceration
- High vaginal laceration
- Early onset of delivery, delivered
- Postpartum care and examination immediately after delivery
- Suppressed lactation
- Shoulder dystocia
- Vulvar and perineal hematoma, postpartum condition or complication
- Cord around neck with compression, complicating labor and delivery
- Other manually assisted delivery
We compared the baseline characteristics of the sample using a t test or Wilcoxon-Mann-Whitney test for continuous variables and a χ2 test for categorical variable. We calculated the raw rate per 100 person-years of each pregnancy outcome and compared outcome rates between women receiving hysteroscopic or laparoscopic sterilization using a Z-test. Cox proportional hazard models were used to compare the relative likelihood of each pregnancy outcome controlling for patient age, geographic region, urbanicity (metropolitan statistical area designation), comorbidities (measured by Elixhauser score, a comorbidity index validated for use with administrative databases10), and insurance type (preferred provider organization, health maintenance organization , point of service, and other). We investigated the robustness of our results by examining the Kaplan-Meier curves, comparing the raw hazard ratios with the adjusted hazard ratios, and verifying the proportional hazards assumptions using Schoenfeld residuals.11 Data were analyzed using SAS 9.4.
We identified 997 pregnancy outcomes in 817 women who underwent either laparoscopic or hysteroscopic sterilization between 2008 and 2012 (n=387 pregnancies/27,724 cases) or laparoscopic sterilization (n=610 pregnancies/42,391 cases). Figure 1 shows our exclusion process. Demographics for the 817 women with pregnancies after sterilization are listed in Table 1. Women who underwent hysteroscopic sterilization were slightly older (34.4 vs 33.3 years, P=.002) but otherwise similar to women undergoing laparoscopic sterilization. As previously reported, the cumulative rate of pregnancy was 1.02 pregnancies per 100 person-years for hysteroscopic sterilization and 0.88 pregnancies per 100 person-years laparoscopic sterilization (P=.003) with a rate of 0.76 pregnancies per 100 person-years for hysteroscopic sterilization if a hysterosalpingogram was completed (P=.031).4
The most common result of pregnancy was live birth (n=438 [60.1%]). Live birth was more common after hysteroscopic sterilization (0.27 live births/100 person-years) than laparoscopic sterilization (0.20 live births/100 person-years, P=.030) (Table 2), and results were similar in adjusted models (adjusted hazard ratio [HR] 1.32, 95% CI 1.09–1.60) (Table 2). After adjusting for potential confounders, women living in urban areas and younger women were more likely to have a live birth after sterilization (adjusted HR urban vs nonurban 1.45, 95% CI 1.10–1.92; adjusted HR 0.89 for each additional year of age, 95% CI 0.88–0.91). Among live births, there was no difference in the rates of preterm birth with a rate of 0.03 per 100 person-years for laparoscopic sterilization and 0.04 per 100 person-years for hysteroscopic sterilization (P=.398).
Spontaneous abortion occurred in 288 (36%) of pregnancies after sterilization. Spontaneous abortion rates were similar after hysteroscopic and laparoscopic sterilization (0.14 and 0.15/100 person-years after hysteroscopic and laparoscopic sterilization, respectively; Table 2). This difference was not statistically significant when comparing raw rates (P=.249) or in the hazards model (adjusted HR of hysteroscopic sterilization 0.88, 95% CI 0.69–1.14) (Tables 3 and 4). Older women were less likely to experience a spontaneous abortion after sterilization (adjusted HR 0.94 for each year of age, 95% CI 0.92–0.96), consistent with their overall lower risk of pregnancy. When we compared the effect of age in spontaneous abortion (adjusted HR 0.94) with live birth (age adjusted HR 0.89), age has a relatively higher effect on miscarriage.
Therapeutic abortion represented 24.1% of the outcomes; women were more likely to undergo a therapeutic abortion after hysteroscopic sterilization (0.12 therapeutic abortions/100 person-years) compared with laparoscopic sterilization (0.08 abortions/100 person-years) with an associated adjusted HR of 1.49 (95% CI 1.10–2.01). Women in the Northeast were also more likely than women in other regions to undergo therapeutic abortion. Our findings did not change in a sensitivity analysis excluding women with diagnoses of both spontaneous abortion and therapeutic abortion, suggesting that misclassification was low (Tables 5 and 6).
Ectopic pregnancies represented 11.1% of pregnancies and were less common after hysteroscopic compared with laparoscopic sterilization (0.01 vs 0.07 events/100 person-years; adjusted HR of 0.12 95% CI 0.05–0.29) and, consistent with all pregnancy outcomes, was less common among older women (adjusted HR 0.87 for each year of age, 95% CI 0.84–0.91). Stillbirth and trophoblastic disease were rare in this population; low sample size prevented us from fully exploring these outcomes with a Cox proportional hazard model. When comparing the raw rates, we found no evidence for a difference in rates of stillbirth (n=19, 2.6% of pregnancies, 0.01 stillbirths/100 person-years after both types of sterilization). Trophoblastic disease occurred in 15 women (2.1%) and was more commonly observed after laparoscopic sterilization (n=14, 0.01 events/100 person-years) than after hysteroscopic sterilization (n=1, less than 0.01/100 person-years; adjusted HR 0.11, 95% Cl 0.02–0.85).
Our evaluation of model fit suggested that the Cox proportional hazards model was a valid specification. Our raw and adjusted hazards were similar in direction and magnitude, and the Kaplan-Meier curves behaved as we would expect (Fig. 2). We evaluated the proportional hazards assumption by examining the Schoenfeld residuals and did not find evidence of a violation (Table 7).
Because preventing pregnancy is the reason women seek out sterilization, clinicians need to counsel patients about the risks of becoming pregnant and the likely outcomes of pregnancies that occur after different sterilization procedures. Among women becoming pregnant after either sterilization procedure, live birth was the most common pregnancy outcome and was more likely in the hysteroscopic sterilization group. Ectopic pregnancy was more common after laparoscopic sterilization. Therapeutic abortion was also more common among women who had hysteroscopic sterilization, possibly as a result of the higher likelihood of intrauterine, compared with tubal, pregnancies after this method. The total numbers for some outcomes are small and do not have enough power to detect a difference. However, our results do not support claims that hysteroscopic sterilization has adverse effects on subsequent pregnancy outcomes.
Our data confirm that there remains an elevated risk of ectopic pregnancy after sterilization, especially after laparoscopic procedures. The rate of ectopic pregnancy in the general population is 19.7 per 1,000 pregnancies.12 After sterilization, pregnancy is rare. The U.S. Collaborative Review of Sterilization, the largest prospective study of sterilization outcomes, found a rate of ectopic pregnancy after sterilization to be 7.3 per 1,000 pregnancies or 32.9% of their total pregnancy outcomes,5,13 higher than our rate of 11.1% of pregnancy outcomes. The overall higher rate of ectopic pregnancy after sterilization compared with the general population is expected considering the damage to the fallopian tube that would likely disrupt a subsequent blastocyst's passage into the uterus.
The finding that live births were the most common outcome is reflective of general unintended pregnancy outcomes. Finer et al14 found that 58% of unintended pregnancies end in birth, whereas 42% end in abortion.
Of note, we did not find a significant difference in stillbirth or spontaneous abortion between the two groups, suggesting that laparoscopic and hysteroscopic sterilizations have a similar effect on continuing pregnancies. However, our numbers were small for this outcome and may not be powered to detect a difference. The U.S. Collaborative Review of Sterilization study reported 14.7% of their pregnancy outcomes were spontaneous abortions, which is similar to a general population level of 10–20%.15 The rate of spontaneous abortions after sterilization in this study was 36%. One possible explanations for the apparently higher rate of spontaneous abortions among women after sterilization compared with the general population is the relatively older age of women seeking sterilization compared with those seeking reversible contraception or pregnancy. We found what appeared to be a slightly protective effect of increasing age on spontaneous abortion risk. Because older women are less likely to be pregnant, they are also less likely to have a miscarriage. Additionally, when we compare the effect of age in spontaneous abortion (adjusted HR 0.94) with live birth (age adjusted HR 0.89), age has a relatively higher effect on miscarriage; however, there may be a complex interplay between the effects of age on these various outcomes. An exploration of age effects on pregnancy outcomes for women with failed sterilization procedures by laparoscopic or hysteroscopic means is beyond the scope of the current article but is worth considering in future research.
One strength of this investigation is that we were able to compare outcomes of laparoscopic with hysteroscopic cases among the same pool of patients. Another strength is the large data set, which allows for analysis of a diverse, nationally representative group of women. Indeed, as a result of the breadth of MarketScan, our study included approximately half of all U.S. women with commercial insurance.
Limitations of the study include lack of data on race or ethnicity and lack of inclusion of publicly insured patients. The use of administrative data to classify medical outcomes also raises the possibility of misclassification of pregnancy outcomes resulting from the imperfect correlation between coding on billing claims and information recorded in clinical charts. This misclassification could be the result of imprecise billing codes, women exiting their commercial insurance plan, and instances in which the pregnancy was resolved outside of a medical setting (such as with miscarriages at home). Consequently, we were unable to determine the outcome of all pregnancy episodes. However, the algorithm we chose has been shown to accurately identify actual pregnancy episodes and accurately assess outcomes.9 We also supplemented the algorithm with manual review of the ICD-9 and CPT codes from individual patient charts. We compared women whose pregnancy outcomes were unknown with those with identifiable outcomes to explore whether missing pregnancy outcome was associated with other factors, but found no differences between groups (Table 8). A sensitivity analysis revealed that even if missing outcomes differed by sterilization method, the results would only change if the missing outcomes were ectopic pregnancies or trophoblastic disease. The likelihood of ectopic and trophoblastic disease being frequently missed in administrative data should be relatively low because they are both acute, serious conditions normally requiring hospital management. Outcomes more likely to be missing might be discrete, self-limited conditions and those in which patients might either manage at home or seek self-pay care such as spontaneous or therapeutic abortion as mentioned previously.
This investigation of pregnancy outcomes in women who have been previously sterilized has generated information that may factor into clinical care. Our results may be reassuring to women who may be concerned about a negative pregnancy outcome should either sterilization method fail. Furthermore, these data can help dispel myths about adverse pregnancy outcomes in women faced with an unintended pregnancy secondary to hysteroscopic sterilization failure. Future studies should prospectively look at outcomes in cohorts of sterilized women based on sterilization method, particularly in uninsured or Medicaid populations, to evaluate pregnancy outcomes in different populations. Overall, we found relatively small differences in pregnancy outcomes between pregnancies after laparoscopic or hysteroscopic sterilization with similar rates of pregnancy overall.
1. Jones JMW, Daniels K. Current contraceptive use in the United States, 2006–2010, and changes in patterns of use since 1995. National health statistics reports. Hyattsville (MD): National Center for Health Statistics; 2012.
2. Bartz D, Greenberg JA. Sterilization in the United States. Rev Obstet Gynecol 2008;1:23–32.
3. Benefits and risks of sterilization. Practice Bulletin No. 133. American College of Obstetricians and Gynecologists. Obstet Gynecol 2013;121:392–404.
4. Perkins RB, Morgan JR, Awosogba TP, Ramanadhan S, Paasche-Orlow MK. Gynecologic outcomes after hysteroscopic and laparoscopic sterilization procedures. Obstet Gynecol 2016;128:843–52.
5. Peterson HB, Xia Z, Hughesa JM, Wilcox LS, Tylor LR, Trussell J. The risk of pregnancy after tubal sterilization: findings from the U.S. Collaborative Review of Sterilization. Am J Obstet Gynecol 1996;174:1161–8.
6. Health research data for the real world: the MarketScan databases. 2014. Available at: http://truvenhealth.com/portals/0/assets/PH_11238_0612_TEMP_MarketScan_WP_FINAL.pdf
. Retrieved November 13, 2014.
7. Mezher M. Congressman says more than 300 fetal deaths related to Essure ahead of FDA announcement. Regulatory Focus, February 17, 2016.
9. Naleway AL, Gold R, Kurosky S, Riedlinger K, Henninger ML, Nordin JD, et al. Identifying pregnancy episodes, outcomes, and mother–infant pairs in the Vaccine Safety Datalink. Vaccine 2013;31:2898–903.
10. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care 1998;36:8–27.
11. Allison P. Survival analysis using SAS: a practical guide. Cary (NC): SAS Institute Inc; 2010.
12. Centers for Disease Control and Prevention (CDC). Ectopic pregnancy—United States, 1990–1992. MMWR Morb Mortal Wkly Rep 1995;44:46–8.
13. Peterson HB, Xia Z, Hughes JM, Wilcox LS, Tylor LR, Trussell J. The risk of ectopic pregnancy after tubal sterilization. U.S. Collaborative Review of Sterilization Working Group. N Engl J Med 1997;336:762–7.
14. Finer LB, Zolna MR. Declines in unintended pregnancy in the United States, 2008–2011. N Engl J Med 2016;374:843–52.
15. Wilcox AJ, Weinberg CR, O'Connor JF, Baird DD, Schlatterer JP, Canfield RE, et al. Incidence of early loss of pregnancy. N Engl J Med 1988;319:189–94.