HIV incidence has started to tumble in one of the best-studied groups of people in Africa, the Conference on Retroviruses and Opportunistic Infections (CROI 2017) heard in Seattle. The annual infection rate has fallen 40% in the last four years, the conference was told. A combination of factors, including wider availability of antiretroviral therapy, increased male circumcision, and later age of sexual debut in young people, all appear to be contributing to this decline.
Presenter Mary Grabowski at the Johns Hopkins University, Baltimore confirmed that this is the first time a population-level decline in incidence in the Rakai cohort has been observed.
The Rakai Community Cohort Study, in a corner of southern Uganda, is probably the longest-studied group of people with, and at risk of, HIV in a low-income setting. This study, which dates from 1994-5 and was built upon an even older one dating from 1989, enrols every consenting adult aged 15-49 living in a 50-village territory and regularly ‘tops itself up’ with new members, keeping its participants at a level of about 12,000.
Rakai has hosted many pioneering HIV studies ranging from the first study to show conclusively that HIV treatment reduced infectiousness to one of the three landmark randomised studies of medical male circumcision for HIV prevention.
Rakai was chosen as a site because it was the area first and hardest hit by HIV in Uganda; in 1989 it had an HIV prevalence of over 20%, only surpassed since by the later hyper-epidemic in southern Africa in the late 1990s. HIV prevalence fell to 13% by the mid-1990s and has stayed pretty steadily at that level ever since. At the time the fall in prevalence was ascribed to the success of behaviour change programmes, but was probably more due to deaths among the first wave of people infected during the peak of HIV incidence in the mid-1980s.
Incidence – the rate of new infections among a population – is a better guide to whether HIV epidemics are truly growing or shrinking, though true incidence is quite difficult to establish as it is affected by testing rates. However in situations with an established HIV epidemic and high rates of testing, as in Rakai, new diagnoses are a reasonable surrogate for incidence. And in the last four years there is no doubt that it has decreased.
Data were collected between 1999 and 2016 in twelve surveys from 30 communities in the Rakai Community Cohort. Incidence was compared at seven particular time-points between 2004 and 2016 and compared with incidence prior to 2004. Over this period, 33,937 people took part; , including 17,870 initially HIV-negative persons and 16,067 with HIV. Of the HIV-negative people, 931 (5.2%) tested HIV-positive during 94,427 person-years of follow-up, an annual incidence of just under 1%.
Antiretroviral therapy (ART) started to become available in Rakai by 2004 and by 2016, 69% of people with HIV in the cohort were taking ART. This has led to an increase in the proportion of people in the community who have HIV and are virally suppressed; this proportion, including the undiagnosed and those not on ART, rose from 42% in 2009 to 75% by 2016 (and thus achieved the UNAIDS 90/90/90 target).
At the same time, the roll-out programme of medical male circumcision meant that the proportion of men who were circumcised rose from 15% in 1999 to 59% in 2016. MC coverage increased from 15% in 1999 to 59% by 2016.
A third significant change happened during the same period; young people started having sex later. The proportion of young people aged 15-19 who reported not yet starting sex rose from 30% to 55% over the 1999-2016 study period.
HIV incidence was a steady 1.17% a year over most of the study period, from 2000 to 2010. But after that it started to fall. By 2012 it was 0.8% a year and by 2016 it was 0.66% a year, a 42% decline. Incidence fell further in men (by 54%) than in women (by 32%). Grabowski said this might be due to men reaping the benefits both of ART and of medical male circumcision.
Grabowski confirmed that this is the first time a population-level decline in incidence in the Rakai cohort.
At the same time however, prevalence has stayed stuck at a remarkably steady 13%, exactly what it was in 1994. This is probably due to the fact that ART is prolonging people’s lives and so the net total of people with HIV in the population is staying the same.
A second study led by Oluwasolape M Olawore, at the Johns Hopkins University, looked at migrants to the Rakai area and compared HIV incidence in them versus permanent residents. Migrants were defined as a person “who moves to a new community with intention to stay”. Among 13,991 people in this study, one third (4571 people) met this definition.
Annual HIV incidence was 0.88% in permanent residents, 0.97% in long-term settled migrants and 1.6% in recently arrived migrants. After statistical adjustment, it was found that in women, recently arrived migrants were 60% more likely to become infected with HIV than permanent residents, but settled migrants no more likely. Recently arrived migrant men were 34% more likely to become infected, but this was not statistically significant.
This study confirms that in east Africa, as much as in other parts of the world, being a new arrival within a community is a risk factor for HIV in itself.
To assess the impact of combination HIV prevention (CHP) on HIV incidence, we measured long-term trends in HIV incidence based on observed seroconversion data in a prospective population-based cohort in Rakai, Uganda, and evaluated their associations with antiretroviral therapy use (ART), male circumcision (MC) scale-up, population-level viral load suppression, and sexual behaviors.
Between 1999 and 2016, data were collected in 12 surveys from 30 communities in the Rakai Community Cohort Study (RCCS), an open population-based longitudinal cohort of persons aged 15-49. Poisson regression was used to assess trends in HIV incidence, self-reported ART/MC coverage, population-level HIV viral load suppression (proportion of HIV-positive population with
Over the analysis period, 33,937 individuals participated in the RCCS, including 17,870 HIV-negative persons who contributed 94,427 person-years of follow-up and 931 incident HIV cases. ART was introduced in 2004 and by 2016 coverage was 69%. Increasing ART coverage was accompanied by significant changes in HIV viral load suppression rising from 42% in 2009 to 75% by 2016 among all HIV-positive persons (p
In this large prospective population-based study, HIV incidence significantly declined as ART and MC were scaled and sexual activity in young persons declined. These results provide empiric evidence that HIV control efforts utilizing combination interventions can have a substantial population-level impact
Mary K Grabowski, Gertrude Nakigozi, Fred Nalugoda, Thomas Quinn, Godfrey Kigozi, Ronald H Gray, David Serwadda, Maria Wawer, Steven J Reynolds, Larry W Chan1
Higher HIV prevalence is commonly observed among migrant populations in Sub-Saharan Africa. However the extent to which migration is a cause or consequence of HIV infection is largely unknown. Here, we use population-based, longitudinal data to assess the association between duration of residence since migration into a community and HIV incidence in Rakai District, Uganda.
We used prospective data from HIV-negative participants residing in thirty communities under continuous surveillance between 1999 and 2015 in the Rakai Community Cohort Study (RCCS), an open population-based census and cohort of adults aged 15-49 in rural south-central Uganda. Migrants were identified during census and classified as individuals who moved to a new community with intention to stay. Newly HIV-positive individuals were considered incident HIV cases if they had an HIV-negative test result at a prior survey. Poisson regression with generalized estimating equations was used to estimate incidence rate ratios (IRR) of HIV infection associated with years since arrival for migrants relative to long term-residents with adjustment for demographics, sexual risk behaviors, and calendar time.
HIV incidence was assessed among 13,991 HIV-negative individuals of whom 57% (n=8,049) were women and 34% (n=4571) were classified as migrants. Participants were followed for 85,654 person-years (pys) during which 802 incident HIV events were detected (n=313 in men; n=489 in women). Overall, incidence was 1.6/100pys in recent migrants (arrived 2 years), and 0.88/100pys among long-term residents. Among women, HIV incidence was significantly elevated in recent migrants relative to long-term residents before and after adjustment for potential confounders (IRR=1.86, 95%CI:1.43-2.41; adjIRR=1.60, 95%CI: 1.21-2.13) but not in non-recent migrants (IRR=0.89, 95%CI:0.71-1.12; adjIRR=0.97, 95%CI: 0.76-1.23). We observed no significant increases in HIV risk among recent (IRR=1.34, 95%CI:0.78-2.32; adjIRR=1.36, 95%CI:0.78-2.37) or non-recent migrant men (IRR=1.03, 95%CI:0.71-1.48; adjIRR=1.12, 95%C1: 0.77-1.64).
These data suggest that the earliest years after migration are associated with increased risk of HIV acquisition in women but not men in rural East Africa. These findings highlight the need for timely interventions targeted to migrant populations, particularly women, to reduce HIV incidence in Sub-Saharan Africa.
Oluwasolape M Olawore, Aaron Tobian, Fred Nalugoda, Ronald H Gray, Maria Wawer, Robert Ssekubugu, John Santelli, Larry W Chang, David Serwadda, Kate Grabowski