Background:  Prior analyses from the D:A:D study showed an increased risk of MI with cumulative use of PI, but not NNRTI. In addition, current/recent use of abacavir (ABC) and didanosine (ddI), but not stavudine (d4T), zidovudine (ZDV), or lamivudine (3TC) were also associated with MI risk. Sufficient follow-up has now accrued to allow us to explore additional associations between specific drugs and MI risk.

Methods:  The D:A:D collaboration includes data on 33,308 patients from 11 cohorts followed prospectively; 580 developed an MI over 178,835 person-years. Poisson regression, adjusted for demographics, cardiovascular risks, and use of other ARV, assessed the impact of cumulative (/year) or recent (current/in last 6 months) use of ARV with >30,000 person-years of exposure (table). Further analyses included adjustment for latest measures of lipids, metabolic parameters, CD4 and HIV-RNA to identify possible mechanisms for any associations. In the main analyses, exposure to indinavir (IDV) and saquinavir (SAQ) was calculated regardless of concomitant use of ritonavir, but sensitivity analyses also explored separate associations with IDV/r and SAQ/r.

Results:  There were no statistically significant associations between recent or cumulative use of tenofovir (TDF), dideoxycytidine (ddC), ZDV, d4T, or 3TC and MI risk. As before, recent exposure to ABC or ddI was associated with an increased risk of MI. No association was seen with cumulative exposure to nevirapine (NVP), efavirenz (EFV), nelfinavir (NFV), or SAQ. Cumulative (but not recent) exposure to IDV or lopinavir/ritonavir (LPV/r) was associated with an increased risk of MI (relative rate [RR]:  1.12.and 1.13/year, respectively). These increased risks were attenuated slightly (1.08 [1.02 to 1.14] and 1.09 [1.01 to 1.18], respectively) after adjustment for lipids but were not altered further after adjustment for other metabolic parameters; neither association was affected by adjustment for CD4 or HIV RNA. In sensitivity analyses, RR/year for IDV/r (22,185 person-years) and SAQ/r (24,727 person-years) were 1.14 (1.02 to 1.28), and 1.06 (0.97 to 1.14), respectively.

Conclusions:  Of the ARV considered, only IDV, LPV/r, ddI and ABC were associated with a significantly increased risk of MI. Additional follow-up will allow us to more definitively establish which of the specific (ritonavir-boosted) PI are most associated with MI risk. As with any observational study, our findings must be interpreted with caution given the potential for confounding.