Objective: To determine the prevalence of HIV-1 infection and to identify the most important risk factors for infection.
Design: A cross-sectional population survey carried out in 1990 and 1991 in Mwanza Region, Tanzania.
Methods: Adults aged 15-54 years were selected from the region (population, 2 million) by stratified random cluster sampling: 2434 from 20 rural villages, 1157 from 20 roadside settlements and 1554 from 20 urban wards. Risk factor information was obtained from interviews. All sera were tested for HIV-1 antibodies using enzyme-linked immunosorbent assay (ELISA); sera non-negative on ELISA were also tested by Western blot.
Results: The response rate was 81%. HIV-1 infection was 1.5 times more common in women than in men; 2.5% of the adult population in rural villages, 7.3% in roadside settlements and 11.8% in town were infected. HIV-1 infection occurred mostly in women aged 15-34 years and men aged 25-44 years. It was associated with being separated or widowed, multiple sex partners, presence of syphilis antibodies, history of genital discharge or genital ulcer, travel to Mwanza town, and receiving injections during the previous 12 months, but not with male circumcision.
Conclusion: This study confirms that HIV-1 infection in this region in East Africa is more common in women than in men. The results are consistent with the spread of HIV-1 infection along the main roads. There is no evidence that lack of circumcision is a risk factor in this population.
PIP: Between August 1990 and February 1991, health workers took blood samples from, and trained interviewers spoke to, 5145 15-to-54-year-old adults living in either an urban area (1554), in a rural village (2434), or in a roadside settlement (1157) in the Mwanza Region of Tanzania to determine the prevalence of HIV-1 infection and its most significant risk factors. The prevalence rate of HIV-1 infection was highest in urban areas, lower in roadside settlements, and lowest in rural areas (11.8%, 7.3%, and 2.5%, respectively), suggesting that HIV-1 had spread along main roads. Yet, there were as many people infected with HIV-1 in rural areas of the Mwanza Region as there were in Mwanza town. Women were 1.2 (rural) to 1.7 (urban) times more likely to be infected with HIV-1 than were men, indicating greater efficiency of HIV-1 transmission from men to women than from women to men. HIV-1 infection peaked in the 15-to-34-year-old group in women and in the 25-to-44-year-old group in men. Separated, divorced, or widowed men and women were at increased risk of being HIV-1 infected, even when controlled for numerous factors (odds ratio (OR) = 3.4 and 1.6, respectively). This may have been an indication of multiple partners, since the question concerning multiple partners was vague. Other important risk factors for women and men were syphilis antibodies (OR = 1.7 and 1.85), history of genital discharge or chancroid (OR = 2, 1.6 and 2.7, 1.6), travel to Mwanza town (OR = 2.1 and 1.7), and receiving injections during the previous 12 months (OR = 1.5 and 1.9). There was no link between male circumcision and HIV-1 infection. In fact, there seemed to be a moderate protective effect (OR = 0.8). This effect may be even more likely since urban men, who were at greatest risk of HIV-1 infection (8.7% vs. 5.4% [roadside] and 2.4% [rural]), had the highest rate of circumcision (61% vs. 29% and 17%, respectively).