Vaginal spermicides have been popular in the United States for more than 40 years, but the evidence of their efficacy in preventing pregnancy is unreliable. Reported pregnancy rates range from zero to 60 per 100 woman-years of use. Data are based on older studies that fall short of modern standards for design, conduct or evaluation. Raymond and colleagues studied the efficacy and acceptability of two popular spermicides in an international study of 765 women.

The study enrolled healthy women 18 to 35 years of age from eight sites in North America, South America and Africa. The women had no history of subfertility, had a low risk of sexually transmitted disease and wished to use spermicides as their only method of contraception. All patients were assessed by history, physical examination, pelvic examination, wet preparation of vaginal secretions and urine pregnancy test before entering the study. Exclusion criteria included contraindications to nonoxynol-9 use, recent pregnancy, evidence of vaginal infection or conditions in which pregnancy would be contraindicated. Eligible patients were randomly assigned to use either a film or a foaming tablet preparation of nonoxynol-9. All patients were instructed in method use and were provided with diaries to record menstrual periods, frequency of intercourse and spermicide use. Diaries were reviewed and other data collected at follow-up visits at four, 12, 20 and 28 weeks. Pelvic examination, wet preparations and urine pregnancy tests were performed at the end of the study and when clinically indicated. Patients were questioned about product use and acceptability at the four-week visit and at the end of the study.

The 382 women using the film preparation did not differ in any important variables from the 383 women assigned to use the foaming tablet. Over 90 percent of the women in each group reported using the spermicide as instructed. The cumulative probability of pregnancy was 28 percent in the tablet group and 25 percent in women using the film, a statistically insignificant difference. Further statistical analyses showed that previous experience with contraception and frequency of intercourse were not related to probability of pregnancy. The only variable that was predictive of pregnancy was the desire to have more children.

The authors conclude that neither form of spermicide was effective in preventing pregnancy in these sexually active women. Compliance was reported to be high, but verification was impossible, and compliance may have been over-reported. Without contraceptive use, the cumulative six-month probability of pregnancy in this population was estimated to be 61 percent. The spermicides could, therefore, play some role in reducing overall pregnancy rates, but for women strongly motivated to avoid pregnancy, there are better means of contraception.