PCSK9 inhibitors: what we know, what we should have understood, and what is to come.

Abstract

Currently, two PCSK9 (proprotein convertase subtilisin/kexin type 9) inhibitors are available for clinical use: evolocumab and alirocumab. Both are fully human monoclonal antibodies and both were studied in very large, randomized, double-blind, placebo-controlled, cardiovascular outcomes trials, which were entitled FOURIER and ODYSSEY Outcomes, respectively. Both trials enrolled very highrisk patients with ‘hard’ established atherosclerotic cardiovascular disease: prior myocardial infarction (MI), non-haemorrhagic stroke, or symptomatic peripheral artery disease in the case of FOURIER, and prior MI in ODYSSEY Outcomes. In both trials, the main objectives were achieved, and the PCSK9 inhibitors were shown to significantly reduce the risk of the primary endpoints. Specifically, in FOURIER, evolocumab reduced the risk of a broad quintuple primary endpoint (a composite consisting of cardiovascular death, MI, stroke, hospitalization for unstable angina, or coronary revascularization) by 15% and the harder key secondary endpoint of cardiovascular death, MI, or stroke by 20%. In ODYSSEY Outcomes, alirocumab reduced the risk of death from coronary heart disease, MI, ischaemic stroke, or unstable angina requiring hospitalization by 15%. More specifically, in both trials, there were clear, significant reductions in the risk of MI (27% with evolocumab and 14% with alirocumab), ischaemic stroke (25% and 27%, respectively), and coronary revascularization (22% and 12%, respectively). In this issue of the European Heart Journal, Guedeney and colleagues present a meta-analysis of the effect of evolocumab and alirocumab on cardiovascular outcomes, combining data from 39 phase II or phase III randomized controlled trials and totalling 66 478 patients. Although they included a large number of trials, the sheer size of FOURIER and ODYSSEY Outcomes means that those two trials completely dominated their analysis and accounted for >95% of the events. Thus, not surprisingly, the meta-analysed effect of PCSK9 inhibition on MI, ischaemic stroke, and coronary revascularization (20, 22, and 17% relative risk reductions, respectively) are quite close to what one might estimate from simply averaging the effect observed in FOURIER and ODYSSEY Outcomes. Likewise, no major safety issues emerged in either FOURIER or ODYSSEY Outcomes, and thus no major safety issues were seen in the meta-analysis, including a risk ratio (RR) of 1.00 [95% confidence interval (CI) 0.93– 1.07] for new-onset diabetes. Guedeney et al. note that there was no significant effect of PCSK9 inhibition on cardiovascular death (RR 0.94, 95% CI 0.84–1.06). Some people had expected to see a reduction in cardiovascular mortality. To that end, in the Cholesterol Treatment Trialists Collaboration of statin trials, there was a 12% reduction in cardiovascular mortality for each 1 mmol/L reduction in LDL cholesterol (LDL-C). In FOURIER and ODYSSEY Outcomes, treatment with evolocumab and alirocumab resulted in 1.38 and 1.24 mmol/L reductions in LDL-C, respectively, at 1 year. Based on those numbers, should we have expected 15–16% reductions in cardiovascular mortality in these trials? I think not. We also know from the Cholesterol Treatment Trialists Collaboration that there is a lag before LDL-C lowering translates into a clinical benefit (Take home figure). For the composite of major vascular events, the lag tends to be 6 months or so, translating into the benefit from LDL-C lowering in the first year being about half of what is seen in subsequent years. The lag for a reduction in fatal outcomes seems even longer. 4S and LIPID are two statin trials that published cumulative incidence curves for the successful reduction of a pre-specified fatal primary endpoint (all-cause mortality and death from coronary heart disease, respectively). For both trials, the event curves appear to start to diverge at 1.5 years. Faced with intense pressure to provide cardiovascular outcomes data, the sponsors of both trials (Amgen and Sanofi & Regeneron) opted to have larger trials with shorter durations in order to accrue events more quickly. In FOURIER, the median duration of follow-up was only 2.2 years; in ODYSSEY Outcomes, it was a bit longer at 2.8 years.