Four drug programs tested drugs that raise “good” HDL cholesterol by blocking a protein called CETP, each in a large trial built to show fewer heart attacks and deaths. Pfizer’s torcetrapib was stopped in 2006 because more patients died. Roche’s dalcetrapib was stopped for futility. Eli Lilly’s evacetrapib was stopped for lack of efficacy. Only anacetrapib, tested in the Oxford-run REVEAL trial with Merck funding, reduced major coronary events, by 9%, alongside an 18% lower non-HDL cholesterol as well as higher HDL. Merck then decided not to file it for approval.
This page sets out what the trial registries, company releases and papers say, and what a reader of the evidence should ask. Every figure was read from the cited source. No dollar loss is stated for the class, because none of the sources read gives one; company annual reports were not searched for program costs.
| Jul 2004 | The registry lists a July 2004 start for Pfizer’s ILLUMINATE, a trial of torcetrapib plus atorvastatin in high-risk patients. The registry lists Pfizer as sponsor and 15,067 patients. | ILLUMINATE registry NCT00134264 |
| Dec 2, 2006 | Pfizer stops all torcetrapib trials and ends development, after the independent safety monitoring board recommended terminating ILLUMINATE because of an imbalance of mortality and cardiovascular events. | Pfizer press release, Dec 2 2006 |
| 2007 | The ILLUMINATE paper reports HDL up 72.1% and LDL down 24.9%, but also a 5.4 mm Hg rise in systolic blood pressure and higher aldosterone. Cardiovascular events were more frequent (hazard ratio 1.25, 95% CI 1.09 to 1.44) and so were deaths from any cause (hazard ratio 1.58, 95% CI 1.14 to 2.19). The authors describe an increased risk of mortality and morbidity of unknown mechanism, with evidence of an off-target effect. | Barter et al., NEJM 2007, PMID 17984165 |
| Apr 2008 | The registry lists an April 2008 start for Roche’s dal-OUTCOMES, testing dalcetrapib in 15,871 patients after a recent acute coronary syndrome. The registry lists Hoffmann-La Roche as sponsor. | dal-OUTCOMES registry NCT00658515 |
| Jun 2011 | The registry lists a June 2011 start for REVEAL, testing anacetrapib in 30,449 adults. The registry lists the University of Oxford as lead sponsor; the paper says it was funded by Merck and others. | REVEAL registry NCT01252953 |
| May 17, 2012 | A genetics study (mendelian randomisation) finds that a gene variant raising HDL cholesterol was not associated with risk of heart attack (odds ratio 0.99, 95% CI 0.88 to 1.11), and neither was a 14-variant HDL genetic score (odds ratio 0.93, 95% CI 0.68 to 1.26). The authors say the data challenge the idea that raising HDL will always lower heart-attack risk. The study is not specific to CETP. | Voight et al., Lancet 2012, PMID 22607825 |
| 2012 | dal-OUTCOMES result. HDL rose 31 to 40% on dalcetrapib versus 4 to 11% on placebo, with minimal effect on LDL. At a planned interim analysis with 71% of the projected events, the safety monitoring board recommended stopping for futility. Event rates were 8.0% versus 8.3% (hazard ratio 1.04, 95% CI 0.93 to 1.16). | Schwartz et al., NEJM 2012, PMID 23126252 |
| Oct 2012 | The registry lists an October 2012 start for Eli Lilly’s ACCELERATE, testing evacetrapib in 12,092 patients. The registry lists Eli Lilly as sponsor and the status as terminated, with the reason “insufficient efficacy.” | ACCELERATE registry NCT01687998 |
| 2017 | ACCELERATE result. HDL rose 133.2% and LDL fell 31.1%. After 1,363 of 1,670 planned events the monitoring board recommended stopping for lack of efficacy. Event rates were 12.9% versus 12.8% (hazard ratio 1.01, 95% CI 0.91 to 1.11). Funded by Eli Lilly. | Lincoff et al., NEJM 2017, PMID 28514624 |
| 2017 | REVEAL result. Over a median 4.1 years on intensive statin therapy, the primary outcome occurred in 10.8% on anacetrapib versus 11.8% on placebo (rate ratio 0.91, 95% CI 0.85 to 0.97, P = 0.004). HDL was 43 mg per dL higher (104%) and non-HDL cholesterol 17 mg per dL lower (−18%). No significant differences in death, cancer or other serious adverse events. | HPS3/TIMI55-REVEAL, NEJM 2017, PMID 28847206 |
| Oct 11, 2017 | Merck says it will not submit applications for regulatory approval for anacetrapib, and that its clinical profile “does not support regulatory filings.” | Merck release, Oct 11 2017 |
| 2017 | A genetics study of variants in the CETP gene finds a CETP score linked to higher HDL, lower LDL, lower apolipoprotein B (apoB) and a lower risk of major vascular events (odds ratio 0.946, 95% CI 0.921 to 0.972). Combined with a statin-gene score, the extra benefit was attenuated and not significant. The authors conclude the benefit of lowering LDL may depend on the matching reduction in apoB. | Ference et al., JAMA 2017, PMID 28846118 |
Enrollment is the sum of the four registry counts: 15,067, 15,871, 30,449 and 12,092. The trials differ in drug, population and design, so the total is only an indicator of scale.
None of these is a finding that a company’s decision was wrong. The method behind them, tracing each load-bearing claim to its primary source and tagging each link supported, partial, contradicted or missing, is described on the main page.
Hindsight is cheap. What an audit can do is show, before the money moves, which claims rest on a marker, which on a trial, and which on a single genetics study.