Considerable public health resources have been dedicated to implementing HIV ‘treatment-as-prevention’ in an effort to reduce new infections. Although promising, no large-scale studies have yet evaluated the community-level impact of treatment-as-prevention on direct measures of HIV incidence among gay and bisexual men (GBM). This study assessed the temporal relationship between community viremia and HIV incidence among GBM living in New South Wales and Victoria, Australia’s most populous states.
For 2012-2017, we established a longitudinal cohort of HIV-positive (n=12,200) and HIV-negative (n=45,719) GBM using data from a targeted sentinel surveillance system of 49 sexual health clinics, general practices, community HIV-testing sites and hospitals. Among GBM with diagnosed HIV, annual prevalence of viremia was calculated for each patient’s last viral load test of a calendar year (≥200 RNA copies/mm3 ) while mathematical modelling was used to estimate the proportion of HIV-positive GBM living with undiagnosed HIV infection (assuming 100% viremia); these outcomes were combined to estimate ‘community viremia’. A correlation coefficient was calculated to assess the temporal relationship between community viremia and HIV incidence, which was directly measured among HIV-negative sentinel surveillance patients using the repeat testing method. To account for the introduction of HIV pre-exposure prophylaxis (PrEP) in 2016, the analysis was repeated for the 2012-2015 period only.
HIV viremia among diagnosed GBM decreased from 27.9% in 2012 to 3.7% in 2017 (p<0.001) while the proportion living with undiagnosed HIV decreased from 10.0% to 8.4% (p=0.01). As shown in Figure 1, annual community prevalence of HIV viremia decreased from 28.6% in 2012 to 12.8% in 2017 (p<0.001) while HIV incidence decreased from 0.88/100 person years in 2012 to 0.22/100 person years in 2017 (p<0.001). The correlation coefficient between annual community prevalence of viremia and HIV incidence from 2012 to 2017 was 0.94 (p<0.001) and for 2012 to 2015 was 0.90 (p<0.001).
Decreasing community viremia among GBM was strongly associated with decreasing HIV incidence, including before the implementation of PrEP. Our findings justify the significant investment in HIV treatment initiatives, highlighting that these should be sustained as key elements of HIV prevention.