Abstract 1694: Torcetrapib: Predictors of Increased Clinical Risk in the Illuminate Study
Torcetrapib (T), an experimental cholesterol ester transfer protein inhibitor (CETPi) raised HDL-C and lowered LDL-C. The ILLUMINATE study was a randomized, double-blind trial in 15,067 patients testing the clinical benefit of T on a background of atorvastatin (A). The A dose (10 – 80 mg) was determined during the trial run-in to provide LDL-C control. The trial was terminated prematurely for imbalanced all-cause mortality (ACM) (Hazard Ratio [HR] = 1.58, p = 0.006) and major cardiovascular events (MCVE) (HR = 1.25, p = 0.001) in T. These findings and effects of T on BP, electrolytes and aldosterone have been previously reported. Post hoc analyses were performed to further understand the effects of T. Evidence for risk was limited to the T/A 10 mg vs A 10 mg subgroup comparison (n = 6,492) for both ACM (HR = 2.68, p < 0.0001) and MCVE (HR = 1.41, p = 0.002). Point estimates of risk for both endpoints lowered as the A dose increased; none of the 95% confidence intervals for the higher dose subgroups excluded unity. The baseline factors were also balanced in the treatment subgroups. Multivariate Cox PH models showed that, in addition to T, factors significantly associated with ACM risk were: increasing age, prior CHF or stroke, and high CRP; and for MCVE: lower Apo A-I, and history of either smoking, hypertension, CHF, MI, or stroke. Changes in the on-treatment predictors of risk were also studied. The changes in measured markers due to T were not different between the 10 mg A and A > 10 mg subgroups. Specifically, post-randomization changes in HDL-C, LDL-C, blood pressure, electrolytes or aldosterone were of the same magnitude in the subgroups. The HR for T remains unexplained. There is no clinical evidence to support that lipid, blood pressure or RAAS played a contributing role. The post hoc finding of an inverse relationship between HR (for ACM and MCVE) and background A dose, coupled with the a lack of a corresponding pattern of change in on-trial markers of lipids, BP or RAAS is intriguing. It suggests that an effect of A at higher doses may have mitigated the risk of T which is unexpalined by typical biomarkers.