Objective:To identify risk factors associated with HIV incidence in a rural Ugandan population.
Methods:Men and women who seroconverted between 1990 and 1997 (cases) and seronegative subjects (controls) were drawn from a general population cohort of approximately 5000 adults in rural, southwestern Uganda. Information on risk factors was ascertained through a detailed interview and physical examination by clinicians who were blind to the study subjects' HIV status. All patients were interviewed within 2 years of their estimated date of seroconversion.
Results:Data were available on 130 men (37 cases, 93 controls) and 133 women (46 cases, 87 controls). There was a significantly higher risk of infection in men (odds ratio [OR], 6.51; 95% confidence interval [CI], 1.06-39.84) and women (OR, 4.75; 95% CI, 1.26-17.9) who were unmarried and in a steady relationship, and in men who were divorced, separated, or widowed (OR, 4.33; 95% CI, 1.32-14.25) compared with those who were married. There was a significantly higher risk of HIV infection in men (OR, 3.78; 95% CI, 1.20-11.93) and women (OR, 20.78; 95% CI, 2.94-141.2) who reported ≥5 lifetime sexual partners compared with those who reported at most 1 partner. For men, there was an increased risk of infection associated with receiving increasing numbers of injections in the 6 months prior to interview (p < .001 for trend). Women reporting sex against their will in the year prior to interview were at higher risk of infection (OR, 7.84; 95% CI, 1.29-47.86,; p = .020).
Conclusions:The strongest risk factor for HIV incidence in this rural Ugandan population is lifetime sexual partners. The increased risks found for women reporting coercive sex and men reporting injections require further investigation.
Address correspondence and reprint requests to Maria Quigley, MRC Tropical Epidemiology Group, London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, U.K.
Manuscript received October 4, 1999; accepted February 16, 2000.
© 2000 Lippincott Williams & Wilkins, Inc.