Trends and interaction of HIV-1 and HIV-2 in Guinea-Bissau, West Africa : no protection of HIV-2 against HIV-1 infection

To study trends in the prevalence and incidence of HIV-1 and HIV-2 infections in Guinea-Bissau over the last 7 years, and to evaluate the protective effect of HIV-2 against HIV-1 infection. Prospective follow-up of a cohort of police officers in Guinea-Bissau, and sentinel surveillance of pregnant women in Bissau. Participants in the police cohort were tested regularly for antibodies to HIV and Treponema pallidum, and information about sexual risk behaviour and a history of sexually transmitted diseases was obtained. Simultaneously, pregnant women at the maternity wards at the National Hospital in Bissau were screened annually for HIV antibodies. To evaluate changes in prevalence and incidence of HIV in the police cohort, the study period was divided into three time strata with 2-3 years in each stratum. For the evaluation of a protective effect of HIV-2 on subsequent HIV-1 infection, two multivariate Poisson regression models were constructed, adjusting for different selected confounding variables. Between 1990 and 1997, 2637 police officers were included in the cohort study, 90.7% of whom were male. The overall prevalence of HIV-1 was 0.9%, of HIV-2 it was 9.7% and of HIV-1 and HIV-2 dual reactivity it was 0.5%. For pregnant women the prevalence rates were 0.9, 5.5 and 0.2% for HIV-1, HIV-2 and dual reactivity respectively. The prevalence of HIV-1 increased significantly whereas the prevalence of HIV-2 declined significantly during the study period, among both police officers and pregnant women. The total incidence of HIV-1 and HIV-2 was 0.74 and 0.83 per 100 person-years respectively in the police cohort. The incidence of HIV-1 increased slightly from 0.62 to 0.78 per 100 person-years (not significant), whereas the incidence of HIV-2 declined significantly from 0.90 to 0.35 per 100 person-years over the study period. Seven police officers seroconverted from HIV-2 to dual reactivity (1.22 per 100 person-years). The adjusted incidence ratio of acquiring HIV-1 infection among HIV-2-positive subjects compared with HIV-negative subjects was 1.65 [95% confidence interval (CI), 0.73-3.74] and 1.98 (95% CI, 0.80-4.87), depending on the confounding variables included. Our study shows an increasing prevalence of HIV-1 and a decreasing prevalence of HIV-2 in Guinea-Bissau. The incidence of HIV-2 declined significantly whereas the incidence of HIV-1 was relatively stable over the study period. No protective effect of HIV-2 against subsequent HIV-1 infection was obs