Background: To assess the effects of HIV-1 and other sexually transmitted infections on pregnancy, we undertook cross-sectional and prospective studies of a rural population in Rakai district, Uganda.
Methods: 4813 sexually active women aged 15-49 years were surveyed to find out the prevalence of pregnancy by interview and selective urinary human chorionic gonadotropin tests. The incidence of recognised conception and frequency of pregnancy loss were assessed by follow-up. Samples were taken to test for HIV-1 infection, syphilis, and other sexually transmitted diseases.
Findings: At time of survey 757 (21.4%) of 3544 women without HIV-1 infection or syphilis were pregnant, compared with 46 (14.6%) of 316 HIV-1-negative women with active syphilis, 117 (14.2%) of 823 HIV-1-positive women with no concurrent syphilis, and 11 (8.5%) of 130 women with both syphilis and HIV-1 infection. The multivariate adjusted odds ratio of pregnancy in HIV-1-infected women was 0.45 (95% CI 0.35-0.57); the odds of pregnancy were low both in HIV-1-infected women without symptoms (0.49 [0.39-0.62]) and in women with symptoms of HIV-1-associated disease (0.23 [0.11-0.48]). In women with concurrent HIV-1 infection and syphilis the odds ratio was 0.28 (0.14-0.55). The incidence rate of recognised pregnancy during the prospective follow-up study was lower in HIV-1-positive than in HIV-1-negative women (23.5 vs 30.1 per 100 woman-years; adjusted risk ratio 0.73 [0.57-0.93]). Rates of pregnancy loss were higher among HIV-1-infected than uninfected women (18.5 vs 12.2%; odds ratio 1.50 [1.01-2.27]). The prevalence of HIV-1 infection was significantly lower in pregnant than in non-pregnant women (13.9 vs 21.3%).
Interpretation: Pregnancy prevalence is greatly reduced in HIV-1-infected women, owing to lower rates of conception and increased rates of pregnancy loss. HIV-1 surveillance confined to pregnant women underestimates the magnitude of the HIV-1 epidemic in the general population.
PIP: A cross-sectional, prospective study conducted in Uganda's rural Rakai District indicated pregnancy prevalence is substantially reduced in women infected with HIV. A total of 4813 women 15-49 years of age who had at least one sexual partner in the previous year and were unaware of their HIV status were enrolled. 953 women (19.8%) were HIV-positive and 446 (9.3%) were diagnosed with active syphilis. 931 women (19.3%) became pregnant during the study period (1989-92). The pregnancy rate was 21.4% among women with no serologic evidence of HIV or syphilis compared with 14.6% among HIV-negative women with active syphilis and 8.5% among women infected with both HIV and syphilis. The pregnancy rate also was significantly higher among the 833 asymptomatic HIV-infected women (14.3%) than the 120 with clinical symptoms (7.5%). Symptomatic HIV-1 infection in male partners did not account for the lower pregnancy rate in HIV-positive women. After controls for age, marital status, gravidity, contraceptive use, lactation, subfertility, and time since last intercourse, the adjusted odds ratio of pregnancy among all HIV-positive women compared to women without HIV or syphilis was 0.45 (95% confidence interval, 0.35-0.57). Among the 3340 women who were not pregnant at baseline and were locatable, the pregnancy rate during follow-up was 23.5 per 100 woman-years among HIV-positive women and 30.1 per 100 woman-years among those without HIV or syphilis. Rates of pregnancy loss were higher among HIV-infected women (18.5%) than HIV-negative women (12.2%). The prevalence of HIV-1 infection was significantly lower in pregnant than nonpregnant women (13.9% and 21.3%, respectively). These findings indicate that, if HIV surveillance is confined to pregnant women, the prevalence of HIV-1 among women of reproductive age will be seriously underestimated.