Antiretroviral treatment as prevention (TasP) in Denmark could nearly eliminate the HIV epidemic among men who have sex with men (MSM), according to predictions of a Bayesian back-calculation model. But researchers caution that their findings depend on "an exceptionally high" viral suppression rate in Denmark and may not readily apply to sub-Saharan Africa, according to the study.
TasP emerged as an HIV prevention option when the randomized HPTN 052 trial found that people who began immediate combination antiretroviral therapy (cART) at a high CD4 count had a 96% lower chance of transmitting HIV to their steady sex partner than did people who deferred cART. Scientists believe the elimination threshold for an HIV epidemic is one new infection per 1000 people. Researchers estimated the impact of TasP on reaching that threshold among MSM in Denmark.
Because CD4 counts offer one indicator of time since HIV acquisition, the researchers built a back-calculation model including four CD4-stratified stages: (1) ≥500 cells/mm3, (2) 350-500 cells/mm3, (3) 200-350 cells/mm3 and (4) <200 cells/mm3. They applied the model to data from the Danish HIV Cohort Study, which includes data on all MSM diagnosed with HIV in Denmark since 1995. The goal was to estimate HIV incidence in MSM over time, as well as the hidden epidemic, that is, the number of MSM with undiagnosed HIV infection.
The analysis showed that since 2002 the number of MSM diagnosed with HIV at a high CD4 count rose, while the number diagnosed at a low CD4 count fell. The fairly abrupt increase in CD4 count at diagnosis in 2002 implied that the time between infection and diagnosis began dropping at that point. Estimates of undiagnosed MSM indicated that the hidden epidemic in MSM fell substantially since potent combination antiretroviral therapy (cART) first saw wide use in 1996. The model estimated that by 2013 only 21% of all HIV-infected MSM in Denmark remained undiagnosed. That percentage means Denmark had only an estimated 617 MSM with undiagnosed HIV in 2013. Most of these 617 men with undiagnosed HIV were in the two earlier stages of infection.
The researchers defined potential HIV transmitters as (1) undiagnosed MSM, (2) diagnosed MSM not on antiretroviral therapy or (3) diagnosed MSM on treatment but with a detectable viral load. As numbers in all three possible transmitter groups waned over the past 20 years, potential MSM transmitters in Denmark fell by almost two-thirds, from 2218 men in 1996 to 819 men in 2013. By 2013, most potential transmitters were men with undiagnosed HIV.
The model determined that annual HIV incidence among MSM in Denmark has been declining for almost two decades -- from 117 per year in 1996 (95% Bayesian credible interval [BCI] 94 to 140) to 75 per year in 2013 (95% BCI 20 to 117). Thus the model provides "strong supportive evidence" that annual HIV incidence has been dropping since cART arrived in 1996. The strong correlation between increasing antiretroviral coverage and decreasing HIV incidence since 1996 indicates that TasP is having an impact on the HIV epidemic among MSM in Denmark.
Using an estimate of 55,000 MSM in Denmark in 2009, the researchers set HIV incidence among MSM at 1.4 new infections per 1000 men in 2013 (95% BCI 0.4 to 2.1). That value approaches the World Health Organization (WHO) elimination threshold of one per 1000.
The model also indicated that risky sex has increased among MSM in the past two decades. The researchers calculated five new HIV infections for every 100 potential MSM transmitters in 1996; that rate almost doubled by 2013.
The authors note that favorable model predictions of declining HIV incidence among MSM in Denmark reflect high antiretroviral coverage in the country and a 98% viral suppression rate. They propose that "TasP will reduce HIV epidemics in countries where the treatment programmes are exceptional, as in Denmark." But they caution that "the WHO's global elimination strategy is unlikely to succeed" if treatment programs do not attain this level of success in countries with high HIV prevalence.