The aim of the four communitybased studies from Guinea-Bissau presented in this thesis was to describe trends in the spread of HIV-1 and HIV-2 from 1987 to 1996, to evaluate the current spread and to identify potential risk determinants currently associated with the three retroviral infections. For HIV-2 with emphasis on old age.
Between 1993 and 1996, we conducted HIV surveys in three population-based samples from our study area. Population "A" consisted of adults (>14 years of age) from 92 houses which had already been followed with three previous surveys in 1987, 1989 and 1992. Population "B" consisted of adults from 212 new randomly selected houses from the study area. Population "C" consisted of older subjects (>49 years of age) from the entire study area. For the surveys in population "A" and “B” we also included screening of antibodies to HTLV-I and HTLV-II.
Sex and age-specific incidences of HIV-1 and HIV-2 infection were measured in an early (January 1987-June 1991) and a late calendar period (July 1991December 1996) among adult subjects who had participated in at least two of the four surveys conducted in population "A". Changes in HIV-2 prevalences over time were evaluated comparing test data from the initial survey in population "A" with those from the recent survey in population "B". Risk determinants of HIV-1 , HIV-2 and HTLV-I infection were studied in univariate and multivariate models using questionnaire information collected during the surveys in 1993-96.
The overall annual incidence of 0.34% was slightly higher among men than among women but did not differ between younger and older subjects. The HIV-1 incidence increased more than 10-fold from the early (0.04%) to the late calendar period (0.64%), and was 10 times higher for subjects already infected with HIV-2. Among subjects from the incidence study has participated in the 1996 survey seroconversion rates were significantly elevated in subjects already HIV-2 infected, in subjects from two distinct ethnic groups and in subjects with a history of STD-related symptoms. From 1987 to 1996 there was an increase in the prevalence of HIV-1 from 0.0% to app. 2.5%.
The overall annual HIV-2 incidence of 0.54% tended to be higher among women (0.72%) than among men (0.31%), and was higher among older subjects, though significantly so only among women. The overall HIV-2 incidence was unchanged at app. 0.55% in the early and the late calendar period. However this trend differed between the two sexes. The incidence increased slightly among women, whereas it declined markedly among men from 0.66% in the early period to 0.00 in the late period. Similar tendencies in sex and agespecific prevalences of HIV-2 infection were observed when comparing surveys from 1987 and 1996. The prevalence remained unchanged in women (app. 8.5%) whereas a significant decrease from 9.1% to 4.7% was observed in men. In 1996, an elevated prevalence of HIV-2 infection was associated with female gender, increasing age, increasing number of sexual partners, history of hospitalisation, and not believing in the existence of HIV/AIDS. Male circumcision was a protective factor. In population "C" (>49 years of age), risk determinants of HIV-2 infection were female gender, being a tenant, belonging to two distinct ethnic groups from the northwestern part of the country and having children with more than one partner.
In 1996, the prevalence of HTLV-1 infection was 3.6%, increasing from 1.8% in the youngest individuals to 9.8% in subjects older than 44 years of age. The prevalence was two-fold higher among women (4.6%) than among men (2.2%). Among risk factors studied, HTLV-I infection was significantly associated with being female, increasing age and HIV-2 seropositivity. Elevated prevalences were also found in HIV-1 infected women, and in men who had received blood transfusion.
The studies show that HIV-1, though still lowprevalent, is spreading faster in Bissau than in other capitals of the subregion; and that HIV-1, during the latest five years, has had a four-fold increased incidencetoprevalence ratio as compared to HIV-2. The 10-fold higher incidence among HIV-2 infected individuals could reflect high-risk behaviour in this group, but also increased susceptibility of HIV-1 infection due to HIV-2induced immunosuppression.
The declining male HIV-2 prevalences observed especially among older subjects is likely to reflect a cohort effect originating from the extended liberation war prior to independence in 1974. Higher levels of promiscuity during this period could have induced the unequaled high prevalences of infection in Guinea-Bissau. A gradual disappearance of the generation which was active during the war may thus have affected the lower male prevalences observed in 1996. With female spouses being younger than male spouses, the emergency of a declining trend in the female HIV-2 prevalence may be only a matter of time.
The HTLV-I prevalence of 3.6% observed in this study is higher than in most other countries in the subregion. The gradual increase in prevalence observed with age closely resembles the age distribution of HIV-2 and suggests that also HTLV-I infection may have been endemic in Guinea-Bissau for an extended period. In men, there was a significant association between HTLV-I infection and blood transfusion, but this was based on a very low number of transfusion recipients. Therefore, with the low prevalences observed among adolescents, HTLV-I infection is likely to be transmitted predominantly via sexual contacts. We failed to demonstrate significant associations to indicators of sexual activity. This probably reflects the low sexual transmissibility of HTLV-I.