Summary | PCSK9 inhibition & CV events: Review of recent and upcoming hard endpoint outcome trials
Antibodies against PCSK9 have been tested in three major CV outcome trials; the FOURIER trial3, the ODYSSEY trial, and SPIRE 1 and 2, of which the SPIRE trials were prematurely stopped in 2016. In SPIRE, treatment with bococizumab was compared to placebo in high risk prevention or secondary prevention patients. The trial was stopped due to antibody development against the drug, which attenuated in the therapeutic effect.24
In the ODYSSEY trial,1,2 dose titration of the drug was possible with up-titration to reach the prevailing LDL-c target level or down-titration when the achieved LDL-c level was ≤ 25 mg/dL. For patients who achieved extremely low levels (≤ 15 mg/dL) the intake of drug was stopped. There was no dose titration in the FOURIER trial.
Prof. Sever speculated that although the baseline risk of the patient populations in these 2 trials was different, similar baseline lipid profiles and in all probability, similar percentage reduction in CV events for similar baseline LDL-c levels, similar reduction in LDL-c and similar duration of follow-up will be seen in these trials. The FOURIER outcomes were compatible with those in a meta-analysis of statin trials produced by the Cholesterol Treatment Trialists Collaboration (CTTC), and Sever thinks the same results will be observed in ODYSSEY.
The FOURIER trial3 was stopped prematurely after a follow-up of 2.2 years when the prespecified number of primary and secondary endpoints had been reached. 27564 patients were randomized at a 1:1 ratio to evolocumab or placebo in 49 countries. Importantly, there was a placebo run-in period in this trial during which statin therapy was optimized,1 which was necessary for many patients. Safety assessment focused on the development of new-onset diabetes, since this had historically been an issue with statin treatment. All events in the trial were adjudicated independently by the TIMI Study Group (Boston, MA, USA). A small number of patients discontinued the drug, withdrew consent or were lost to follow-up. Baseline characteristics showed that ~80% of patients had MI, ~20% stroke, and 13% symptomatic PAD. About 2/3 of patients were on high intensity statin therapy at baseline, and 1/3 on moderate intensity in this high risk CV population.
Despite the observed reduction in the number of MI and stroke events with evolocumab, no reduction in CV death was seen.
The achieved baseline LDL-c levels were 92 mg/dL in both groups, and also other lipids levels were equally balanced. A 60% reduction in LDL-c levels was observed with treatment with evolocumab, which was maintained throughout the duration of the trial. The 3-year Kaplan Meier curve showed a relative risk reduction of 15% with evolocumab for the primary endpoint (HR 0.85) and a 2% absolute risk reduction. The number needed to treat (NNT) was 15 patients in 3 years. The relative risk reduction of the secondary endpoint was 20% (HR 0.80), a 2% absolute risk reduction, and the same NNT.
Despite the observed reduction in the number of MI and stroke events with evolocumab, no reduction in CV death was seen. In a number of early statin trials of secondary prevention, for example the 4S25 and LIPID26 trials, no divergence of the curves for the first two years following randomization was seen. This indicates that what we see in FOURIER may have been predicted based on earlier trials.
Prof. Sever warns for the dangers of subgroup analyses3; loss of randomization, smaller number of patients in each subgroup, and frequently inadequate power to detect drug effects in each subgroup. The only appropriate statistical test is the test of heterogeneity and he states that there does not appear to be a subgroup in the FOURIER trial that achieved a greater benefit from evolocumab than the whole trial population.
In view of the apparent delay in the development of the maximal protective CV effect, a landmark analysis was undertaken on the secondary endpoints CV death, MI and stroke, after PCSK9 inhibition at 1 year follow-up and 12-36 months follow-up; the event reduction after 1 year was substantially less than the reduction after 2 and 3 years. Landmark analysis for only MI and stroke showed again a substantially lower reduction in year 1 compared to year 2 and 3.
Focusing on the year-2 analyses of the trials when maximal benefits in terms of relative risk reduction were seen, there was little difference between the results of FOURIER and those predicted by the CTCC.27 Therefore, the observed results of FOURIER were exactly what was predicted for the duration of the trial and the extent of LDL-c lowering.
Another important objective in FOURIER was to establish whether PCSK9 inhibition in those achieving extremely low levels of LDL-c was safe and whether the benefits of evolocumab were consistent across the whole achieved LDL-c range. There was a single and clear relationship with no attrition of benefit at the low end of the curve.3 There were no safety issues among those achieving very low LDL-c levels and no difference in the number of adverse events between evolocumab and placebo groups.
Finally, although the trial was stopped, the SPIRE 2 trial also showed a relative risk reduction in CV events in patients taking bococizumab as compared to those randomized to placebo. Thus, two trials have demonstrated CV benefits of treatment with monoclonal antibodies directed against PCSK9.