Results of three national HIV household surveys spanning six years, presented at the 17th Conference on Retroviruses and Opportunistic Infections (CROI) in San Francisco, indicate that HIV prevalence has stabilised at 11% in South Africa, with signs that the rate of infection is falling in younger age groups.
Five and a half million South Africans live with HIV, almost a quarter of the total in sub-Saharan Africa.
The household surveys were conducted in 2002, 2005 and 2008, and another is planned for 2011. They collected data on HIV status, sociodemographic characteristics and behavioural factors of adults and children at 1000 sites throughout the country.
The first survey tested saliva specimens for HIV while the second and third surveys tested dried blood spot specimens.
Nearly 8500 people were surveyed in 2002 and 16,000 and 15,000 respectively in 2005 and 2008.
Overall HIV prevalence has stabilised, with the surveys showing an HIV prevalence of 11.4% in 2002 and 10.9% in the other two years.
Among children aged 2 to 14, HIV prevalence decreased significantly from 5.6% in 2002 (95% confidence interval [CI], 3.7-7.4) to 2.5% in 2008 (95% CI, 1.9-3.5), probably due to increasing coverage of prevention of mother-to-child transmission.
HIV prevalence has stabilised, or is possibly starting to go down, in young people aged 15 to 24: it was 9.3% in 2002 and 8.7% in 2008. It increased in the core group of adults aged 15 to 49 from 15.5% in 2002 to 16.2% in 2005 and 16.8% in 2008. In neither of these age groups, however, were the changes statistically significant.
South Africa also now has the largest HIV treatment programme in the world, with an exponential expansion in the numbers of people taking treatment from nearly 33,000 in January 2005 to 744,000 in March 2009. Rather than ask people directly, the 2005 and 2008 surveys screened the blood specimens of HIV-positive people for antiretroviral drugs. In 2008, antiretroviral drugs were detected in 16.6% of specimens.
The researchers calculated that the effect of highly active antiretroviral treatment (HAART) at present was to increase HIV prevalence due to lower mortality: they estimated that without treatment prevalence among adults aged 15 to 49 would have been 1.7% lower in 2008, i.e. 15.2% instead of 16.9%.
HIV incidence was calculated using two methods. One, which calculated the incidence in young people by extrapolating directly from new infections and estimating the percentage that were recent, found that annual incidence in 18 year olds had declined from 1.8% in 2005, the peak incidence year, to 0.8% in 2008.
The second method of calculating incidence involved observing prevalence in sequential age groups. For instance, if prevalence is compared in 15 to 25 year olds in 2002 versus 18 to 28 year olds in 2005 and 21 to 31 year olds in 2008, an estimate can be made of the number of new infections that were occurring per year. The fact that three surveys have now been done then allows for this cumulative incidence to be compared with the same age groups in subsequent surveys.
This measure yielded an estimate of annual incidence of 2.0% in 2002-05 compared with 1.3% in 2005-08 in 15 to 49 year olds. This 35% decline did not reach statistical significance. However incidence declined more dramatically in young women aged 15 to 24, among whom there was a statistically significant decline in estimated annual incidence from 5.5% in 2003-05 to 2.2% in 2005-08.
Two trends are probably responsible for these estimated changes in incidence. Among survey respondents aged 15 to 49, reported condom use at last sex increased significantly from 31.3% in 2002 to 64.8% in 2008. There were particularly significant increases among women aged 15 to 49 and people aged over 50.
At the same time, the proportion of people who had ever been tested for HIV increased significantly from 25% to 56% between 2003 and 2008, while the proportion tested in the last twelve months increased from 12% in 2005 to 25% in 2008.
No changes were found in levels of other risk factors such as intergenerational sex or multiple partners.
Presenter Thomas Rehle commented that the observed figures for HIV prevalence masked two different trends, increasing coverage of ARVs and increased condom use, and these needed to be disentangled to unearth actual trends. A much higher coverage of ARVs over a longer period would have to happen for HIV treatment to have a positive effect on incidence, he added.
He recommended incorporating the testing of dried blood spots for antiretroviral drugs into routine surveillance as a means of estimating antiretroviral uptake levels.
Rehle T et al. Trends in HIV prevalence, incidence, and risk behaviors among children, youth, and adults in South Africa, 2002 to 2008. Seventeenth Conference on Retroviruses and Opportunistic Infections, San Francisco, abstract 37, 2010.