Herbert B. Peterson, M.D.
After this lecture, the participant will be able to:
A 1997 report from the U.S. Collaborative Review of Sterilization 1 evaluated prospectively a total of 10,685 women who underwent tubal sterilisation in medical centers in nine U.S. cities. Women underwent one of six methods of tubal occlusion under study in 1978 to 1987 and were followed for up to 8 to 14 years. The six methods studied were bipolar coagulation, unipolar coagulation, silicone rubber band application, spring clip application, interval partial salpingectomy, and postpartum partial salpingectomy. Follow-up was completed for the study in 1994. The risk of pregnancy was assessed using cumulative life-table probabilities and proportional hazards analysis.
A total of 143 true sterilisation failures were identified in the study. The failure rates varied by age and method of tubal occlusion. The 10 –year life table cumulative probabilities of pregnancy for all ages combined (18-44 years) ranged from 7.5 pregnancies per 1000 procedures (for unipolar coagulation and postpartum partial salpingectomy) to 36.5 pregnancies per 1000 procedures (for spring clip application). In general, the failure rates for all methods were highest among women who were young at sterilisation. Among women age 18-27 years at sterilisation, the 10-year cumulative probabilities of pregnancy were as high as 54.3 pregnancies per 1000 procedures (for bipolar coagulation) and 52.1 per 1,000 procedures (for spring clip application). By contrast, the failure rates for women age 34-44 years at sterilisation ranged from 1.8 (for unipolar coagulation) to 18.7 (for interval partial salpingectomy) pregnancies per 1000 procedures.
As for the probability of pregnancy after sterilisation overall, the risk of ectopic pregnancy was influenced by age and method of tubal occlusion 2. Likewise, the proportion of pregnancies that were ectopic varied by method of occlusion. Bipolar coagulation was associated with the highest proportion (0.65) followed by interval partial salpingectomy (0.43), silicone rubber band application (0.29), postpartum partial salpingectomy (0.20), unipolar coagulation (0.17), and spring clip application (0.15). The 10-year cumulative probabilities of ectopic pregnancy for all ages combined (18-44 years) ranged from 1.5 (for postpartum partial salpingectomy) to 17.1 (for bipolar coagulation) pregnancies per 1,000 procedures. For women sterilised before age 30 years, the probabilities ranged from 1.2 (for postpartum partial salpingectomy) to 31.9 (for bipolar coagulation) pregnancies per 1,000 procedures.
The risks of pregnancy after sterilisation in the U.S. Collaborative Review of Sterilization are higher than those generally expected. One reason is that most prior studies of sterilisation failure have had only short - term follow-up. The failure rates at one and two year follow-up in the U.S. Collaborative Review of Sterilization were similar to rates reported from other studies and, thus, one of the major lessons learned from the collaborative study is that failures, including ectopic pregnancies, may occur many years after the procedure. In fact, ectopic pregnancies occurred in the 10th year of follow-up after all four laparoscopic methods of occlusion.2 In addition, however the study evaluated procedures that were performed primarily in teaching institutions in the 1970s and 1980s. It is plausible that failure rates have decreased over time with further experience. Support for this possibility is seen with a detailed analysis of pregnancies after bipolar coagulation from the collaborative study.3 The five- year cumulative probability of pregnancy for women who had 3 or more cm of fallopian tube coagulated in 1985-1987 (3.2 per 1,000 procedures) was much smaller than the comparable group sterilised in 1978-1982 (27.1 per 1,000 procedures).
Ever since 1951, when Williams et al first observed a higher than expected number of women with increased menstrual flow and intermenstrual spotting after sterilisation 4, the existence of a post-tubal sterilisation syndrome of menstrual abnormalities has been debated. A recent review of the evidence for such a syndrome stated that, " The only consistency in the articles reviewed is their inconsistency."5 The authors of the review concluded that, " In the end, there appears to be no clear-cut evidence of a post-tubal sterilization syndrome." A final analysis of the U.S.Collaborative Review of Sterilization has not been published but a preliminary report found no menstrual changes attributable to sterilisation at 1-2 years after the procedure 6, a finding that is largely consistent with the rest of the literature. The study did find some changes at five years after sterilisation that may have been attributable to ageing. The study found no differences in menstrual changes between those methods causing the most tissue destruction and those causing the least.
A 1997 report from the U.S. Collaborative Review Of Sterilization indicated that the 14 year cumulative probability of hysterectomy after tubal sterilisation was 17% 7. Women most likely to undergo hysterectomy were those who, at sterilisation, reported a history of uterine leiomyomata (RR 2.7; 95%CI 2.0, 3.7), endometriosis (RR 2.5; 95%CI 1.7, 3.9) and bleeding of more than 7 days during the menstrual cycle (RR 1.8; 95% CI 1.1, 2.8) – all common reasons for hysterectomy without a history of sterilisation. Of note, however, is that most women with these conditions at the time of sterilisation did not undergo hysterectomy within the follow-up period.
A 1998 report from the study compared women who underwent sterilisation with women whose husbands underwent vasectomy and found that the sterilised women were fourfold more likely to undergo hysterectomy in 5 years after the procedure (5-year cumulative probability 8% versus 2%, respectively).8 Three other studies found that women sterilised at a young age were more likely to undergo hysterectomy than nonsterilised women but that women older at sterilisation were no more likely than nonsterilised women to undergo hysterectomy. 8 By contrast, younger and older women in the collaborative study had similar risks. Speculation regarding increased risks for younger women includes the probability that young women who are not sterilised are more likely than their nonsterilised counterparts to want to preserve fertility. Because there is no evidence that tubal sterilisation actually causes problems that would lead to hysterectomy, it is highly likely that any increased risk in sterilised women is unrelated to sterilisation, per se.
A 1998 report from the U.S. Collaborative Review of Sterilization found that the probability of expressing regret at least once within 14 years after sterilisation was 20.3% for women sterilised at age 30 years or younger and 5.9% for those sterilised at ages greater than 30 years.9 The strongest risk factor for regret present at the time of sterilisation was young age. After adjustments for other risk factors, women sterilised at age 30 years or younger were twice as likely as older women to express regret (RR 1.9; 95% CI 188.8.131.52). Among the young women sterilised who expressed regret, nearly half (48%) requested information about sterilisation reversal. Young age at sterilisation has been a consistent predictor of regret in other studies.9
A 1998 systematic review of the literature on this question summarised findings from 14 studies published from 1985 to 1996.10 Although the summary risk estimate from these studies was 1.23 (95% CI 1.01,1.49) the authors concluded that there was evidence against a causal relationship between vasectomy and prostate cancer because of study biases that tend to overestimate the effect of vasectomy. These biases were discussed further in an accompanying editorial .11 On balance, the weight of the evidence is well summarised in an editorial from the US National Cancer Institute commenting on a US population-based study finding no effect of vasectomy,
" With the new information from this large population – based study, vasectomy appears either not to cause prostate cancer or to have only a relatively weak relationship to the disease."
Edited by Aldo Campana,