Treatment as prevention is starting to work in parts of Africa

This article is more than 12 years old. Click here for more recent articles on this topic

A longitudinal study from KwaZulu Natal province in South Africa is the first study from the global south to relate an increase in the proportion of adults on HIV treatment to a fall in HIV incidence, the 19th Conference on Retroviruses and Opportunistic Infections was told last week.

The study found evidence of a threshold effect; incidence started to fall once the proportion of all adults diagnosed with HIV in the area who were on treatment exceeded 30%.

Meanwhile, a study that took place in a week-long intensive health campaign in Uganda, as well as studies from areas as diverse as San Francisco and Swaziland, documented large increases in the proportion of people with HIV who are on treatment.

Falls in incidence in KwaZulu Natal

National surveys in South Africa have found evidence of significant falls in HIV incidence in recent years, but have related this to behavioural change rather than treatment. In the study presented at CROI, of a rural area of northern KwaZulu Natal centred on the mining town of Somkhele (Tanser), the researchers found a relationship between HIV treatment and a fall in infections.


longitudinal study

A study in which information is collected on people over several weeks, months or years. People may be followed forward in time (a prospective study), or information may be collected on past events (a retrospective study).

detectable viral load

When viral load is detectable, this indicates that HIV is replicating in the body. If the person is taking HIV treatment but their viral load is detectable, the treatment is not working properly. There may still be a risk of HIV transmission to sexual partners.

They made use of a population-based HIV surveillance survey that has sampled 10,000 adults a year from 2004 onwards, by identifying 16,558 people who had taken at least two HIV tests during this period in order to gauge incidence rates. They then compared these data to individually linked data from the district-based HIV treatment and care programme.

Adult HIV prevalence in the area is high – 24%. The rate of new infections peaks at 8% a year in women in their early 20s and 5% a year in men in their late 20s. HIV testing rates are also high; researchers estimate that only 30% of the HIV-positive population is undiagnosed, a low proportion for Africa, and 75% of HIV-negative adults who have tested for HIV have done so more than once.

Since 2004, there has been a huge scale-up of HIV treatment, with 20,000 patients starting antiretroviral therapy since then, and by 2001 more than 40% of all adults diagnosed with HIV were on antiretroviral therapy (ART), and over 60% with a baseline CD4 count below 350 cells/mm3. HIV treatment at this CD4 threshold was only introduced in August 2011; previous to this it was 200 cells/mm3.

HIV incidence between 2004 and 2011 averaged 2.64% a year but was lower after 2009, when for the first time more than 30% of the diagnosed population was on ART. It was 3.0 to 3.5% 2007-09 but fell to 2.5% in 2010 and 2.0% in 2011.

After adjusting for HIV prevalence in the immediate area and demographic and behavioural variations, the researchers found that for every 10% increase in the proportion of adults on ART, the HIV incidence rate fell by 17%. Incidence was 40% lower when over 30% of the adult population was treated than when fewer than10% were. 

Viral load in Western Uganda

In Uganda, Makerere University, in partnership with UCSF and North Carolina Universities in the US, conducted a so-called “high throughput community-wide health campaign” conducted over five days in the rural parish of Kakyerere near the town of Mbarara in western Uganda (Jain). The population of Kakyerere is 6300, and during the campaign researchers managed to test 4343 (72%) of them for HIV, 2282 adults and 1826 children.

Fingerprick tests were used for HIV status and the samples of those testing HIV-positive were further tested for viral load, CD4 count and the presence of efavirenz and nevirapine as indicators of being on ART.

One hundred and eighty-nine people tested positive for HIV, 179 adults (7.8%) and ten children (0.5%). Of the adults, 46% were new diagnoses.

Viral load was determined in 174 HIV positive adults (92%). The researchers determined that the mean ‘community viral load’ (CVL) amongst adults was 64,000 copies/ml.

Thirty-seven per cent of adults had an undetectable viral load and of these 88% had detectable efavirenz or nevirapine in their blood; 83% of adults who were prescribed ART had undetectable viral load, as did 10% of adults not prescribed ART.

The viral load in adults not on ART was influenced by a few adults with very high viral loads, probably in early HIV infection: whereas the median viral load in adults off treatment was 19,048 copies/ml; the mean was 100,319 copies/ml. Apart from being on ART, the only other association with lower viral load was being married; married adults had on average a viral load that was 45% lower than unmarried adults. Encouragingly, greater distance from the local health centre was not associated with a higher viral load.

Early treatment in San Francisco

San Francisco Health Authority was the first body in the world to take the decision to offer treatment to all people diagnosed with HIV regardless of CD4 count, and it claims that considerable falls in HIV diagnoses seen in recent years (from 681 in 2994 to 434 in 2010) are largely due to this policy. In a study presented at CROI reviewing people diagnosed with a CD4 count over the US DHHS HIV treatment guidelines of 500 cells/mm3, Truong and colleagues observed that 89% of this section of the HIV-positive population was put on treatment in 2010, compared with 31% in 2004. With a shorter period between diagnosis and ART initiation, the mean CD4 drop before starting ART reduced from 135 in 2004 to 54 in 2010. The survey found evidence of considerable inequality in ART access, however, with patients starting ART at CD4 counts over 500 cells/mm3 more likely to be well-off white gay men who were diagnosed by private providers.

National testing and treatment survey in Swaziland

Finally, considerable progress is being made towards ART coverage in the country with the highest HIV prevalence in the world, Swaziland, where 26% of the adult population has HIV. Researchers from the country’s Ministry of Health conducted a survey of randomly sampled households in the first half of 2011(Nkambule), using rapid anonymous HIV testing and asking participants about a variety of behavioural and demographic factors.

HIV prevalence peaked at 54% in women aged 30 to 34 and 48% in men aged 35 to 39; the age at peak prevalence has shifted five years older in the last five years, indicating fewer infections amongst youth.

Seventy-two per cent reported having tested for HIV and self-reported HIV prevalence amongst those who tested was 28% – very near the actual figure, which was 31%. There was, however, high prevalence amongst those who had not tested recently: 28% who had never tested had HIV, as did 13% who had previously tested negative, and 48% of men and 32% of women who had HIV did not know it.

Of those who had HIV exactly one-third was taking ART, but the researchers comment that: “While HIV testing and ART services seem accessible, major efforts are needed to expand access.”


Tanser F et al. Effect of ART coverage on rate of new HIV infections in a hyper-epidemic, rural population: South Africa. 19th Conference on Retroviruses and Opportunistic Infections (CROI), Seattle, abstract 136LB, 2012. View the abstract on the official conference website.

Jain V et al. Assessment of community viral load using a fingerprick-based blood collection method: Uganda. 19th Conference on Retroviruses and Opportunistic Infections (CROI), Seattle, abstract 143LB, 2012. View the abstract on the official conference website.

Truong HH et al. Dramatic improvements in early ART initiation reveal a new disparity in treatment. 19th Conference on Retroviruses and Opportunistic Infections (CROI), Seattle, abstract 139, 2012. View the abstract on the official conference website.

Nkambule R et al. Estimating HIV prevalence from the Swaziland HIV incidence measurement survey: Swaziland. 19th Conference on Retroviruses and Opportunistic Infections (CROI), Seattle, abstract 142, 2012. View the abstract on the official conference website.