Abstract Body

The SMART and START trials established continuous/immediate ART as the standard of care for HIV+ persons. The CD4 difference between treatment groups during follow-up in each trial was approximately 200 cells/µL. We hypothesized that treatment hazard ratios (HRs) were similar in each trial and the pooled analysis of the two studies would better estimate the effect of continuous/immediate ART on disease risk.

The drug conservation arm in SMART and the deferred ART arm in START were pooled as the ‘immune impairment group’ (IIG) and the viral suppression arm in SMART and the early ART arm in START as ‘immune preservation group’ (IPG). Endpoints assessed were AIDS or AIDS-death (AIDS), serious non-AIDS or non-AIDS-death (SNA), cardiovascular disease (CVD), cancer, and all-cause death. HRs from Cox models were obtained for IIG vs IPG for each study and heterogeneity assessed. For the pooled cohort we 1) obtained estimated HRs, 2) assessed mediators of IPG benefit from models adjusting for baseline demographics, CVD and cancer risk factors, and time-updated CD4 and HIV RNA, and 3) performed subgroup analyses based on baseline demographics, CVD and cancer risk factors, CD4 count and geographical region.

Among 10157 participants (median age 40y; 27% female; 51% MSM; median baseline CD4 634 cells/µL; 37% smokers), there were 123 AIDS, 244 SNA, 117 cancers, 103 CVD, 118 deaths, and 359 AIDS or SNA. Nadir median CD4 counts in SMART and START were 250 and 553 cells/µL, respectively. HRs for endpoints were similar for both trials without evidence of heterogeneity (Figure). Unadjusted pooled HRs (95% CI) of IIG vs IPG were for AIDS 3.60 (2.35, 5.51); SNA 1.59 (1.23-2.06); CVD 1.60 (1.07-2.37); cancer 1.86 (1.27-2.72); death 1.82 (1.25-2.65), and AIDS or SNA 2.03 (1.63-2.53). Adjustment for time-updated CD4 and HIV RNA attenuated the pooled HRs for SNA (1.35, 1.00-1.82) and death (1.21, 0.78-1.87). IIG was consistently associated with increased risk of the various endpoints across all subgroups investigated (interaction p>0.1), except that the risk of cancer associated with IIG was higher among those ≤ 35 years (HR 3.16 vs 1.86, interaction p=0.05).

Continuous/immediate ART use reduced the risk of AIDS and SNA events among HIV+ persons consistently in SMART and START despite a difference in nadir CD4 count of 300 cells/µL. Pooled treatment differences were similar across a number of subgroups except for age and cancer risk.