Author + information
- Stephen Williamsa,b,
- Robert K. DeLislea,b,
- Craig L. Hydea,b,
- Anders Malarstiga,b,
- Rachel Ostroffa,b and
- Sophie Weissa,b
Background: Early detection of adverse effects of novel therapies and their mechanistic understanding prior to large-scale clinical trials could prevent harm to patients, avert futile phase 3 studies, and improve research productivity. ILLUMINATE was an international multi-center placebo-controlled trial of the cholesterol ester protein inhibitor torcetrapib, given in the background of atorvastatin, involving 15,067 participants at high risk of cardiovascular events. The trial was terminated prematurely at a median follow-up of 550 days due to significant absolute excess in all-cause mortality (0.45%) and cardiovascular events (1.2%) associated with torcetrapib.
Methods: We performed a nested case-control study of paired baseline and 3-months archived plasma samples from 223 patients assigned to placebo and atorvastatin (A) and 249 patients assigned to torcetrapib and atorvastatin (T+A); half of each group had cardiovascular or mortality events. We applied a 9-protein risk score validated to predict stroke, MI, heart failure or death (JAMA 2016;315: 2532-2541), embedded within a larger aptamer based assay of 1,129 protein measurements.
Results: At baseline (after run-in atorvastatin optimization) 9-protein based risk predictions were similar in T+A vs. A (8% event probability within the median 550 day follow-up) and were higher for individuals with subsequent events than those without (p<0.01). At 3 months, the absolute predicted risk increased vs. baseline in T+A (+0.65%, p<0.01) but declined in A (-0.43%, p=0.056), a +1.08% across-group difference (p<0.01). Among the total of 1129 proteins measured, those associated with torcetrapib treatment (adjusted for false discovery rates) included known proteins involving CV risk, the pituitary-adrenal-renal axis, immunity and inflammation and mitochondrial function.
Conclusions: Adverse outcomes of torcetrapib were accurately predicted by a protein-based risk score within 3 months. A mechanistic proteomic survey informed potential biological correlates. Protein-based risk assessment models embedded within a broader assay may prove to be a useful combination for the evaluation of new therapies and to prevent harm to patients.
Poster Hall, Hall C
Saturday, March 18, 2017, 9:45 a.m.-10:30 a.m.
Session Title: Advances in Lipid Management
Abstract Category: 3. Acute and Stable Ischemic Heart Disease: Therapy
Presentation Number: 1203-306
- 2017 American College of Cardiology Foundation