Condom and nonoxynol-9 use and the incidence of HIV infection in serodiscordant couples in Zambia

We aimed to measure the effectiveness of latex condoms and of nonoxynol-9 [N-9] spermicides, in preventing HIV transmission in heterosexual serodiscordant couples in Lusaka. Each couple was examined at clinic visits scheduled at 3-month intervals for one year or more per couple, or until seroconversion or discontinuation. Couples were given condoms and their choice of 3 N-9 products and advised to use both at every intercourse. Sexual exposure was ascertained from coital logs that recorded coitus and barrier method use. HIV serological testing was done at each clinic visit (ELISA and Western blot if positive). One hundred and ten discordant couples were followed for a mean of 17.6 months. Seventy-eight per cent of coital episodes were protected by condoms, 85% by spermicides and 6.4% were unprotected. Fourteen seroconversions occurred (8.7 infections per 100 couple-years [c-y]). The rate was higher among seronegative men than seronegative women. Among couples who reported using condoms at every intercourse the infection rate was 2.3/100 c-y, compared with 10.7/100 c-y among couples using condoms less consistently (rate ratio [RR] 0.2; 95% confidence interval [CI] 0-1.6). Among couples who reported using N-9 at every intercourse, the seroconversion rate was 6.9/100 c-y; among couples who reported less than fulltime N-9 use, the rate was 8.9/100 c-y (RR 0.8; 95% CI 0.2-2.8). Among the subset of female seronegatives, the N-9 RR was 0.5 (95% CI 0.1-3.8). But when we calculated HIV rates according to N-9 consistency in coital acts when condoms were not used, there was no evidence of protection with higher N-9 use. Consistent use of latex condoms reduces the incidence of HIV infection, but the association between N-9 spermicides and HIV is less clear. The current study could not provide compelling data on the impact of N-9 spermicide use on risk of HIV infection. The study's small size, as well as the consistency of concurrent condom use, limited our inferences. Available spermicide products must be studied further.