Recent studies of HIV-2 have suggested an increased incidence and prevalenc
e among women older than 45 y compared with younger women. We therefore exa
mined whether this phenomenon applied generally to all 3 major retroviruses
, HIV-1, HIV-2 and HTLV-I, among women in Africa. We conducted a MedLine se
arch from 1987 to 1997, using the keywords Africa and HIV-1, HIV-2 or HTLV,
respectively. Community studies, national surveys and studies on professio
nal cohorts were selected. Age groups < 45/50 y were compared with the age
group with the lowest female/male prevalence ratio between 20 and 44 y of a
ge. Thirty-one studies had sufficient data to be included. The female/male
odds ratio (OR) for seropositivity was calculated for the old and the young
age groups, respectively, providing the ratio of odds ratios: OR (old)/OR
(young). Summary ratios for studies of all 3 retroviruses were estimated. I
n general we found a higher female/male prevalence ratio in the age group o
ver 45/50 y than in the younger age group. For HIV-1 the odds ratio was 1.8
2 times [95% confidence interval (CI) 1.19-2.79] higher in the old age grou
p than in the young group. For HIV-2 it was 1.97 [95% CI 0.95-4.08], and fo
r HTLV-I it was 2.02 [95% CI 0.99-4.14] times higher. For all 3 viruses com
bined, the ratio sas 1.88 [95% CI 1.36-2.61]. The few incidence studies of
HIV-1 and HIV-2 indicated a similar tendency. Since differential mortality
is unlikely to explain the pattern, the increase in the HIV-1, HIV-2 and HT
LV-I female/male prevalence ratio suggests that older women may have increa
sed exposure or susceptibility to all 3 retrovirus infections.