Skip navigation

Correspondence On Contraceptives

Dear Bandolier
I agree that comparative failure rates are hard to come by. Trussell's 1995 paper is a good place to start, but it is entirely from a US perspective, and they are somewhat deprived of modern contraceptives. They do not have the levonorgestrel-releasing IUS, Mirena (TM), whose failure rate is only 0.2 percent - readers need to note that the progesterone-T referred to in your (his) table is the Progestasert, with a failure rate 10 times higher - nor the new GyneFIX implantable IUD, nor Persona, nor (until 1997) the Filshie clip for female sterilization.
Failure rate percentages are usually gross rates for failures among 100 women in the first year of use, with discontinuations for other reasons taken into account. Given a first-year infertility rate of 15 percent, the 1,500 out of any 10,000 who would not become pregnant are included among the successes. This applies to your example of Persona, so a claim of 94 percent effective means 600 pregnancies in the first year out of 10,000 women using the method "perfectly" (with abstinence in the "red phase" each cycle).
Trussell's paper unfortunately predates an extremely important new study on female sterilisation - I would say the most significant of the last ten years - namely Peterson et al (Am J Obstet Gynaecology, 1996; 174:1161-70). This reports the findings from the US Collaborative Review of Sterilization (CREST). The cumulative 10 year probability of pregnancy among 10,685 women followed up for 8 to 14 years was 1.8 percent, far higher than previously believed, since earlier studies were either smaller or with no more than two years follow-up. One third of the 143 sterilisation failures were extra-uterine. The Hulka spring clip and bipolar coagulation had particularly high rates at ten years of 3.6 and 2.5 percent, with the failure rate after silicone (Falope) band application similar to the mean of all the methods (1.8 percent).
The new finding of greatest interest was: whatever the failure rate of a given method at two years, one could predict from the data at least 50 percent and in some cases 100 percent more failures by 10 years. This has to mean recanalisation, rather than poor surgery, since among the late failures beyond year 2 the method had been 100 percent effective for the first 2 years. The main laparoscopic method in the UK is the Filshie clip. Its one-year failure rate is fortunately far better than the spring clip, around 0.3 percent, but the CREST study means women should now be given a life-time failure risk of about 0.5 percent or 1 in 200.
CREST is a study to change practice. For a start, vasectomy after two negative sperm counts is a whole order of magnitude more effective (better than 1 in 2000 according to Oxford's Elliot-Smith Clinic, Philp et al, BMJ 1984;289:77-79). Moreover the best intrauterine methods - not the Nova T/Novagard, but the Copper T 380S, the new GyneFIX (TM) and Mirena (TM) - should now be seen as having entirely comparable efficacy to female sterilization, while retaining reversibility. Prevalence studies show the UK as `the sterilization capital of Europe' with, additionally, almost the lowest rate of usage of IUDs and IUSs: a position now difficult to justify even on efficacy grounds, leave alone (with soaring divorce rates) on societal grounds!
Yours sincerely

Professor of Family Planning and Reproductive Health
Department of Gynaecology, University College London
Medical Director, Margaret Pyke Centre

previous story in this issue