Sexual behaviour patterns and other risk factors for HIV infection in rural Tanzania: a case-control study

Abstract

Objective: To examine the association between HIV infection and patterns of sexual behaviour and other risk factors in a rural Tanzanian population in a case-control study, nested within a randomized trial of improved sexually transmitted disease treatment.

Methods: All HIV-positive patients from the baseline survey of the randomized trial were eligible as cases. Cases (n = 338) and controls (a random sample of one in eight HIV-negative persons; n = 1078) were interviewed about risk factors for HIV infection using a structured questionnaire.

Results: A significantly higher HIV prevalence was found among men and women not currently employed in farming [men: odds ratio (OR), 2.08; women: OR, 3.65], women who had travelled (OR, 3.27), educated women (OR, 4.51), and widowed/ divorced people compared with those currently married (men: OR, 3.10; women: OR, 3.54). Two spouse-related factors were significantly associated with HIV, even after adjustment for the sexual behaviour of the index case: HIV was more prevalent in men with younger spouses (P = 0.020 for trend) and in women married to men currently employed in manual work, office work or business (OR, 2.20). In women only, blood transfusions were associated with a higher HIV prevalence (OR, 2.40), but only a small population attributable fraction (4%). There was an increased HIV prevalence associated with increasing numbers of injections. Reported number of lifetime sexual partners was significantly associated with HIV infection (women: OR, 7.33 if > or = 10 lifetime partners compared with < or = 1; men: OR, 4.35 for > or = 50 compared with < or = 1). After adjustment for confounders, male circumcision was associated with a lower HIV prevalence (OR, 0.65; P = 0.11).

Conclusions: In these rural communities, many HIV infections occur through sexual transmission. Some people are at high risk of HIV infection through large numbers of sex partners, whereas some are at risk through their spouse or regular partner. The role of circumcision in HIV transmission is unclear. Commercial sex seems to play a negligible role in HIV transmission in these communities. Our results confirm marked heterogeneity in HIV risk, indicating the scope for risk reduction strategies.

PIP: In a baseline survey a cohort of 12,537 adults was enrolled, interviewed, and examined between November 1991 and December 1992 in the Mwanza Region of Tanzania using random cluster sampling. The study itself took place in May and June of 1993, and it consisted of 338 cases (149 men and 189 women) and 1078 controls (504 men and 574 women). The remainder of the analysis of men was restricted to the 149 cases and 394 controls 20-54 years old. The blood samples from consenting adults were tested for HIV antibodies by enzyme-linked immunosorbent assay (ELISA). A significantly higher HIV prevalence was found among men and women not currently employed in farming (men: odds ratio [OR] 2.08; women: OR 3.65), women who had traveled (OR 3.27), educated women (OR 4.51), and widowed/divorced people compared with those currently married (men: OR 3.10; women: OR 3.54). Two spouse-related factors were significantly associated with HIV even after adjustment for the sexual behavior of the index case: HIV was more prevalent in men with younger spouses (p = 0.020 for trend) and in women married to men currently employed in manual work, office work, or business (OR 2.20). In women only blood transfusions were associated with a 2-fold increased prevalence of HIV (OR 2.40), but only a small population-attributable fraction (4%). There was an increased HIV prevalence associated with increasing numbers of injections, even after adjustment for confounders. The reported number of lifetime sexual partners was significantly associated with HIV infection (women: OR 7.33 if or= 10 lifetime partners compared with or= 1; men: OR 4.35 for or= 50 compared with or= 1). After adjustment for confounders, male circumcision was associated with a lower HIV prevalence (OR 0.65; p = 0.11). Most HIV infections occurred through sexual transmission, although some were attributable to nonsterile injections. Since the large number of sexual partners was a major risk factor, intervention strategies should promote the reduction of partners and the use of condoms.