Decline in the prevalence of HIV-1 infection in young women in the Kagera region of Tanzania.

In northwestern Tanzania, a population-based survey of HIV-1 infection in the Kagera region in 1987 demonstrated a high prevalence (24.2%) in adults of Bukoba town, whereas it was lower (10.0%) in the surrounding rural district of Bukoba. In 1993 and 1996, population-based cross-sectional studies were carried out in urban and rural Bukoba districts, respectively, to monitor the time trend in the prevalence of HIV-1 infection in the region. In both studies, a multistage cluster sampling technique was adopted in selecting study individuals. Consenting individuals between 15 and 54 years of age were interviewed using a structured questionnaire. Following individual counseling, blood samples were drawn and tested for HIV infection using enzyme-linked immunosorbent assay (ELISA) antibody detection tests. The overall age-adjusted HIV-1 seroprevalence in urban Bukoba decreased from 24.2% (134 of 553) in 1987 to 18.3% (118 of 653) in 1993 (p = .008). The age-adjusted gender-specific prevalence declined significantly in women, from 29.1% (95 of 325) to 18.7% (74 of 395; p = .0009). Except for men > or = 35 years of age, whose prevalence appeared to have an upward trend between the two studies, all other age groups in both genders had a downward trend; this finding was most significant in women between 15 and 24 years of age (from 27.6% to 11.2%; p = .0004). For the rural population, the overall prevalence decreased from 10.0% (54 of 538) in 1987 to 6.8% (118 of 1728) in 1996 (p = .01). Except for rural women between 15 and 24 years of age whose prevalence decreased from 9.7% (12 of 124) to 3.1% (12 of 383; p = .002), other age groups in the rural populations showed no change in prevalence. Ongoing interventions in this area leading to behavioral change may have contributed to this observation. An incidence study is under way to confirm this observation and to investigate the factors that are responsible for the decline in the HIV-1 prevalence.