Translating the observed differences in interleukin-6 levels between some antiretroviral regimens into potential long-term risk of serious non-AIDS events: A modeling study

INTRODUCTION: Variable levels of systemic inflammation are observed in people with HIV (PWH), but the clinical significance of differences among antiretroviral therapy (ART) regimens on associated levels of inflammatory markers is unclear. Based on data from previous epidemiologic studies that defined the predicted change in risk of serious non-AIDS events (SNAEs)/death by changes in interleukin-6 (IL-6) and D-dimer, we modeled the effects of differences in these markers between specific ART regimens on the long-term risk of clinical outcomes. METHODS: We used a Markov model to compare the risk of SNAEs/death with differences in IL-6 and D-dimer levels associated with remaining on specific three-drug regimens versus switching to specific two-drug ART regimens over 5 years of treatment. We used IL-6 and D-dimer data based on trajectories over time from the randomized TANGO and observational AIR studies. Age at model entry was set at 39 years. The primary endpoint was the number needed to treat for one additional SNAE/death. RESULTS: Over 3 years, PWH on one of the three-drug regimens studied were predicted to spend 13% more time in the low IL-6 quartile and 11% less time in the high IL-6 quartile compared with those on one of the two-drug regimens. Over 3 years, the predicted mean number of SNAEs/deaths per 100 PWH was 6.58 for a three-drug regimen associated with lower IL-6 levels versus 6.90 for a two-drug regimen associated with higher IL-6 levels. The number needed to treat for one additional SNAE/death among PWH receiving a two-drug versus three-drug regimen for 3 years was 81. Approximately 7,500 participants would be required for a 5-year clinical study to evaluate the accuracy of the model. CONCLUSIONS: Our Markov model suggests that higher IL-6 levels associated with switching from specific three- to two- drug ART regimens may be associated with an increase in the risk of SNAEs/death. Clinical studies are warranted to confirm or refute these results.