Reduced fertility associated with HIV: the contribution of pre-existing subfertility

BACKGROUND: HIV-1 infection is associated with lower fertility among women in sub-Saharan Africa and this association is not explained by the frequency of sexual intercourse, illness, knowledge of HIV status or infection with other sexually transmitted diseases. Women with fertility problems are at increased risk of marital instability and, therefore, HIV infection; consequently, pre-existing subfertility among HIV-infected women may contribute to the association. OBJECTIVE: This study examines the relationship between HIV-1 infection and the incidence of recognised pregnancy and the role of low gravidity prior to seroconversion in rural Uganda. METHODS: A group of 176 women (80 HIV infected and 96 uninfected) were enrolled into an HIV-1 natural history cohort and invited to attend 3-monthly clinic appointments. Data from clinic visits were analysed to assess the independent effects of HIV infection and age, lactation, illness, reported frequency of sexual intercourse and sexually transmitted diseases (STD) on the risk of pregnancy in the following 3 months. The number of previous pregnancies was recorded at enrolment, and the effect of gravidity was examined for the subgroup of women who were uninfected at enrolment or who enrolled within 2 years of their estimated seroconversion date. RESULTS: During follow-up, 124 pregnancies were observed in 83 women beginning in the 3 months following 47 (7.0%) of 669 visits made by HIV-infected women and 77 (9.5%) of 812 visits by HIV-negative women (P = 0.12). Adjusting for age, lactation, illness, STD and the reported frequency of sexual intercourse, the estimated reduction in the risk of pregnancy associated with HIV infection was 47% 95% confidence interval (CI) 18-66. Pre-existing low gravidity was strongly associated with a reduced incidence of pregnancy (odds ratio 0.39; CI 0.19-0.81). Additionally, adjusting for low gravidity reduced the estimate of the effect of HIV infection by almost a half, to 25% (95% CI-57-29). CONCLUSION: Low gravidity prior to seroconversion accounts for almost 50% of the observed association between HIV infection and lowered incidence of pregnancy, after adjusting for age, lactation, illness, STD and the frequency of sexual intercourse.