Background: Expansion of HIV surveillance systems in sub-Saharan Africa is leading to downward adjustments to the size of the AIDS epidemic. However, only analysis of surveillance data from the same populations over time can provide insight into trends of HIV prevalence. We have used data from the same antenatal clinics to document recent empirical trends.
Methods: We collated data from antenatal clinics on HIV prevalence between 1997 and 2003. Data were obtained from 140?000 pregnant women attending more than 300 antenatal clinics in 22 countries in sub-Saharan Africa. Additionally, long-term trend data are available for 57 urban areas and provinces.
Findings: Median HIV prevalence in 148 antenatal clinic sites in southern Africa increased from 21.3% (IQR 11.5-28.2%) in 1997/98 to 23.8% (15.6-29.2%) in 2002. At more than half the sites (58%) an increase of at least one-tenth was noted, but at a fifth of sites, prevalence dropped by at least one-tenth. In eastern Africa, median HIV prevalence decreased from 12.9% (7.0-16.9%) in 1997/98 to 8.5% (5.3-13.0%) in 2002, with prevalence rising in four (7%) sites, but falling at 25 (43%) sites. In west Africa, median HIV prevalence was 3.5% (2.2-5.9%) and 3.2% (2.3-6.1%) for 1997/98 and 2002, respectively, with reductions and increases in prevalence being noted in equal proportions. The long-term trends in urban areas in sub-Saharan Africa show a similar pattern, with increasing evidence of stabilisation during the past 2-3 years compared with the previous decade.
Interpretation: Evidence from surveillance of mostly urban antenatal clinic attendees indicates that the growth in the AIDS epidemic in sub-Saharan Africa has levelled off since the late 1990s but only eastern Africa shows a decline in HIV prevalence. Very large differences persist between subregions. Workers planning a response to the AIDS epidemic must take more careful consideration of these variations to allow locally appropriate responses to the epidemic.