Efficacy and safety of alirocumab and evolocumab: a systematic review and meta-analysis of randomized controlled trials

Abstract

Aims

The effect of low-density lipoprotein cholesterol-lowering therapy with alirocumab or evolocumab on individual clinical efficacy and safety endpoints remains unclear. We aimed to evaluate the efficacy and safety of alirocumab and evolocumab in patients with dyslipidaemia or atherosclerotic cardiovascular disease.

Methods and results

We performed a review of randomized controlled trials (RCTs) comparing treatment with alirocumab or evolocumab vs. placebo or other lipid-lowering therapies up to March 2018. Primary efficacy endpoints were all-cause death, cardiovascular death, myocardial infarction (MI), and stroke. We estimated risk ratios (RR) and 95% confidence intervals (CI) using random effect models. We included 39 RCTs comprising 66 478 patients of whom 35 896 were treated with proprotein convertase subtilisin–kexin type 9 (PCSK9) inhibitors (14 639 with alirocumab and 21 257 with evolocumab) and 30 582 with controls. Mean weighted follow-up time across trials was 2.3 years with an exposure time of 150 617 patient-years. Overall, the effects of PCSK9 inhibition on all-cause death and cardiovascular death were not statistically significant (P = 0.15 and P = 0.34, respectively). Proprotein convertase subtilisin–kexin type 9 inhibitors were associated with lower risk of MI (1.49 vs. 1.93 per 100 patient-year; RR 0.80, 95% CI 0.74–0.86; I  ² = 0%; P < 0.0001), ischaemic stroke (0.44 vs. 0.58 per 100 patient-year; RR 0.78, 95% CI 0.67–0.89; I  ² = 0%; P = 0.0005), and coronary revascularization (2.16 vs. 2.64 per 100 patient-year; RR 0.83, 95% CI 0.78–0.89; I  ² = 0%; P < 0.0001), compared with the control group. Use of these PCSK9 inhibitors was not associated with increased risk of neurocognitive adverse events (P = 0.91), liver enzymes elevations (P = 0.34), rhabdomyolysis (P = 0.58), or new-onset diabetes mellitus (P = 0.97).

Conclusion

Proprotein convertase subtilisin–kexin type 9 inhibition with alirocumab or evolocumab was associated with lower risk of MI, stroke, and coronary revascularization, with favourable safety profile.

Introduction

Atherosclerotic cardiovascular disease (ASCVD) remains the leading cause of death worldwide.^1–5 Lipid-lowering therapy with statins, targeting low-density lipoprotein cholesterol (LDL-C) levels, was demonstrated to reduce the risk of ASCVD events in both primary- and secondary-prevention populations.^{4  ,  5} However, a substantial proportion of patients cannot tolerate statins, do not achieve a significant reduction in LDL-C levels despite use of high-intensity statin therapy or may remain at significant residual risk for ASCVD events despite being on maximally tolerated statin therapy.^5–7 Over a decade ago, proprotein convertase subtilisin–kexin type 9 (PCSK9) emerged as a therapeutic target to treat hypercholesterolaemia in humans.⁸ Proprotein convertase subtilisin–kexin type 9 plays a major role in cholesterol homeostasis by reducing the amount of functional LDL receptors on the plasma membranes thereby increasing the serum levels of LDL-C.⁸ In fact, carriage of loss-of-function PCSK9 alleles is associated with lower cholesterol levels and reduced risk of myocardial infarction (MI).^9–11

Recently, results from Phase 2 and Phase 3 randomized controlled trials (RCTs) investigating the efficacy and safety of injectable monoclonal antibodies that inhibit PCSK9 have been reported.^12–16 Across these studies, PCSK9 inhibition appeared to be associated with reductions in LDL-C levels by at least 60% alongside with lower rates of composite ASCVD events.^12–15 These encouraging results led to the expedited approval from the Food and Drug Administration (FDA) of these agents, as an adjunct to diet and maximally tolerated statin therapy for patients with familial hypercholesterolaemia and/or clinical ASCVD. However, the overall effect of PCSK9 inhibitors as a class and of the individual approved agents (evolocumab and alirocumab) on hard efficacy and safety endpoints remain uncertain. We therefore performed a systematic review and meta‐analysis of RCTs to examine the efficacy and safety of PCSK9 inhibitors currently available in clinical practice.

Methods

Research strategy and selection criteria

We conducted a systematic review of the literature according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines (Supplementary material online, Table S1).¹⁷ We searched PubMed/Medline, CENTRAL (Cochrane Central Register of Controlled trials), clinicaltrials.gov, and slides presentations from the latest international conferences for relevant abstracts and manuscripts published up to 13 March 2018. The following keywords were used: PCSK9 inhibitor, alirocumab, evolocumab, SAR236553, REGN727, or AMG145. Citations were screened at the title and abstract level and retrieved if considered relevant. The main inclusion criterion was a Phase 2 or 3 randomized trial comparing alirocumab or evolocumab to a control strategy (placebo and/or other lipid-lowering drugs) in adult patient with dyslipidaemia and/or established ASCVD. There was no restriction on follow-up and study size. Observational studies (including single-arm pilot studies), case reports, case series, meta-analyses, and studies with duplicate data were excluded from this analysis. Proprotein convertase subtilisin–kexin type 9 inhibitors not approved by the FDA, such as bococizumab, were not included in this study.¹⁶ The primary efficacy endpoints of interest were: all-cause and cardiovascular death, MI, and stroke, as per individual study definitions. Primary safety endpoints of interest were: study-drug discontinuation, neurocognitive adverse events, liver enzymes elevations, and rhabdomyolysis. Neurocognitive adverse events were defined as per single study criteria. Liver dysfunction was defined as an increase of alanine aminotransferase or aspartate aminotransferase as reported in each study. Rhabdomyolysis was defined as an elevation of serum creatine kinase above >10 time the upper reference limit. The study is registered in PROSPERO (CRD42018090768).

Data extraction

Two investigators not involved in any of the selected studies (P.G. and S.S.) independently screened each title and abstract, excluding duplicates and studies not meeting the inclusion criteria. Data were extracted using pre-specified data collection forms. The following relevant data were extracted: trial name, design, sample size, follow-up duration, type of control, and PCSK9 dosage. Baseline characteristics of the study population and the mean LDL-C at baseline and at the maximum time of follow-up available were extracted and entered in a pre-specified structured dataset. Efficacy and safety endpoints were collected at the longest available time of follow-up according to the intention-to-treat principle. The accuracy of the abstracted data was independently confirmed by two other investigators (B.V. and D.K.), and discrepancies were resolved by consensus. Risk of bias of the included studies was assessed according to the Cochrane Collaboration guidelines.

Statistical analysis

Study-level data were entered in a pre-specified structured dataset and analysed according to the intention-to-treat principle. Baseline characteristics across studies are reported with summary rate estimates. Exposure times with weighted incidence rates of adverse events per 100 patient-years of follow-up and corresponding incidence risk differences were analysed taking into account the variable follow-up times within each study. Risk ratios (RRs) and 95% confidence intervals (CI) were estimated using Mantel–Haenszel random-effect models according to DerSimonian and Laird. Fixed effect models for all efficacy and safety outcomes were also reported in the Supplementary material online, Appendix. Heterogeneity among trials for each outcome was estimated with χ² tests and quantified with I  ² statistics (with I  ² <50%, 50–75%, and >75% indicating low, moderate, and high heterogeneity, respectively). Publication bias and small study effect for the primary efficacy and safety endpoints were estimated via visual inspection of the funnel plot and with the Harbord test. Sensitivity analyses including only placebo-controlled trials, excluding trials with low statin use (≤20% overall statin use in the trial), evaluating secondary-prevention trials vs. other trials and according to LDL-C levels at baseline and follow-up were also performed. A P < 0.05 was set as the threshold for statistical significance. Analyses were conducted using STATA version 14.0 (Stata Corp., College Station, TX, USA) and Cochrane’s Review Manager (RevMan) version 5.3 (The Cochrane Collaboration, Copenhagen, Denmark).

Results

Baseline characteristics

The study selection flow diagram is illustrated in Supplementary material online, Figure S1. A total of 39 RCTs comprising 66 478 patients were included. Of these, 35 896 were treated with a PCSK9 inhibitor (14 639 with alirocumab and 21 257 with evolocumab), in addition to maximally tolerated statin therapy or other adjunct lipid-lowering therapies, and 30 582 were treated with placebo or control therapy. Out of 39 studies, 31 (79.4%) were placebo-controlled. Mean weighted follow-up time was 2.3 years across trials and 3.1 years and 1.7 years for alirocumab and evolocumab, respectively. The exposure time was 150 617 patient-years overall and 83 289 patients-years and 67 329 patients-years for alirocumab and evolocumab, respectively. Characteristics of the included RCTs are resumed in Supplementary material online, Tables S2 and S3. Out of 39 studies, in 7 (17.9%) statins were used in ≤20% of the study population either because of documented statin intolerance or assessment of PCSK9 inhibition as monotherapy (Supplementary material online, Table S2). Pooled estimates of baseline characteristics across trials overall and by study drug are displayed in Table 1. Mean pooled LDL-C levels at baseline were 128.9 mg/dL in the PCSK9 inhibitors arm and 126.6 mg/dL in the control arm. At the longest available follow-up time, mean pooled LDL-C levels were 52.6 mg/dL in the PCSK9 inhibitor arm and 121.6 mg/dL in the control arm.

Efficacy endpoints

The effect of PCSK9 inhibition on clinical efficacy endpoints is reported in Figures 1  and  2, and Supplementary material online, Figures S2–S8. Overall, there were no significant differences between PCSK9 inhibitors and control in all-cause death (1.03 vs. 1.15 per 100 patient-years; RR 0.89, 95% CI 0.75–1.04; I  ² = 13%; P = 0.15) and cardiovascular death (0.66 vs. 0.73 per 100 patient-years; RR 0.94, 95% CI 0.84–1.06; I  ² = 0%; P = 0.34). However, use of alirocumab, but not of evolocumab, was associated with lower risk of all-cause death compared with control using random-effect models (0.81 vs. 1.01 per 100 patient-years; RR 0.83, 95% CI 0.72–0.95; I  ² = 0%; P = 0.008). Of note, alirocumab was associated with a trend toward lower all-cause mortality after exclusion of the ODYSSEY-OUTCOMES trial (0.37 vs. 0.68 per 100 patient-years; RR 0.59, 95% CI 0.34–1.03; I  ² = 0%; P = 0.06). Compared with controls, use of PCSK9 inhibitors was associated with significant reductions in MI (1.49 vs. 1.93 per 100 patient-years; RR 0.80, 95% CI 0.74–0.86; I  ² = 0%; P < 0.0001), ischaemic stroke (0.44 vs. 0.58 per 100 patient-years; RR 0.78, 95% CI 0.67–0.89; I  ² = 0%; P = 0.0005), and coronary revascularization (2.16 vs. 2.64 per 100 patient-years; RR 0.83, 95% CI 0.78–0.89; I  ² = 0%; P < 0.0001). Individually, both evolocumab and alirocumab were associated with a reduction of the risks of MI, ischaemic stroke, and coronary revascularization (Figures 1  and  2 and Supplementary material online, Figures S4–S6). There were no significant differences between PCSK9 inhibitors and controls for the endpoints of unstable angina requiring hospitalization and heart failure-related hospitalizations (Figure 2 and Supplementary material online, Figures S7 and S8).

Safety endpoints

Clinical safety endpoints for PCSK9 inhibitors vs. control are reported in Figures 3  and  4 and Supplementary material online, Figures S9–S16. There were no significant differences in the risk of drug discontinuation between groups (1.26 vs. 1.07 per 100 patient-years; RR 1.04, 95% CI 0.95–1.15; I  ² = 0%; P = 0.40). No significant differences in the risk of neurocognitive adverse events (0.57 vs. 0.55 per 100 patient-years; RR 1.01, 95% CI 0.84–1.21; I  ² = 6%; P = 0.91), liver enzymes elevation (0.73 vs. 0.73 per 100 patient-years; RR 0.94, 95% CI 0.84–1.06; I  ² = 0%; P = 0.34), allergic reactions (2.05 vs. 1.83 per 100 patient-years; RR 1.04, 95% CI 0.97–1.12; I  ² = 0%; P = 0.29), haemorrhagic stroke (0.06 vs. 0.06 per 100 patient-years; RR 0.86, 95% CI 0.43–1.74; I  ² = 53%; P = 0.68), rhabdomyolysis (0.14 vs. 0.14 per 100 patient-years; RR 0.90, 95% CI 0.62–1.31; I  ² = 0%; P = 0.58), or new-onset of diabetes mellitus (1.92 vs. 1.93 per 100 patient-years; RR 1.00, 95% CI 0.93–1.07; I  ² = 0%; P = 0.97) were observed between PCSK9 inhibitors and controls. However, PCSK9 inhibitors were associated with higher injection site reactions (1.51 vs. 0.83 per 100 patient-years; RR 1.41, 95% CI 1.21–1.65; I  ² = 19%; P < 0.0001; Supplementary material online, Figure S16).

Bias assessment and sensitivity analyses

No evidence of publication bias or small study effect was found for both efficacy and safety outcomes (Supplementary material online, Figures S17 and S18). Internal bias assessment for each study is reported in Supplementary material online, Table S4. The effect of PCSK9 inhibitors on safety and efficacy outcomes remained consistent with the application of fixed effect models (Supplementary material online, Figures S2–S16). Results for the primary efficacy endpoints remained consistent after inclusion of only placebo-controlled trials (Supplementary material online, Table S5). Effect estimates for the primary efficacy outcomes remained consistent also after exclusion of trials with low statins use (defined as ≤20% overall statin use in the trial) (Supplementary material online, Table S6). The effect of PCSK9 inhibitors on the primary efficacy endpoints was consistent when evaluated across secondary-prevention trials only vs. other trials (Supplementary material online, Table S7), or across trials enrolling statins-intolerant patients vs. patients without statins intolerance (Supplementary material online, Table S8). Of note, a significant interaction was observed in terms of the reduction in all-cause death in trials with an average baseline LDL-C level >100 mg/dL vs. ≤100 mg/dL (Supplementary material online, Table S9). Consistently, a significant interaction was present in terms of the reduction in MI when a LDL-C level at follow-up ≤50 mg/dL was achieved, compared to LDL-C level >50 mg/dL (Supplementary material online, Table S10). Finally, no significant interactions were observed in terms of study-drug discontinuation and neurocognitive adverse events with the use of PCSK9 inhibitors and the achievement of low level of LDL-C at follow-up (i.e. ≤50 mg/dL, Supplementary material online, Table S11).

Discussion

In this large, comprehensive meta-analysis of RCTs, we investigated the efficacy and safety of the FDA-approved PCSK9 inhibitors, evolocumab and alirocumab, across a broad range of patients with hyperlipidaemia or established ASCVD. At a mean weighted follow-up time of 2.3 years, use of PCSK9 inhibitors was associated with lower risk of MI, ischaemic stroke, and coronary revascularization compared with controls (Take-home figure). While we did not observe a mortality benefit with use of PCSK9 inhibitors overall, alirocumab, but not evolocumab, was associated with lower risk of all-cause mortality compared with controls. There were no significant differences between PCSK9 inhibitors and the control group in terms of major safety endpoints including neurocognitive adverse events, rhabdomyolysis, liver enzymes elevations, new-onset diabetes mellitus, or allergic reactions. Results were consistent with restriction of the analysis to only placebo-controlled trials.

Inhibition of PCSK9 emerged as a key therapeutic target to lower LDL-C in humans. Inhibition of PCSK9 increases the extracellular membrane density of LDL receptors, thereby reducing the levels of circulating LDL-C.¹⁸ Recently the FDA approved two fully human, injectable, monoclonal antibodies that inhibit PCSK9 (evolocumab and alirocumab), for the treatment of adults with familial hypercholesterolaemia and/or clinical ASCVD who require additional lowering of LDL-C as an adjunct to diet and maximally tolerated statin therapy.¹⁹

In our large meta-analysis encompassing the totality of the evidence from RCTs investigating the efficacy and safety of the FDA-approved PCSK9 inhibitors, PCSK9 inhibition on top of maximally tolerated statin therapy or other adjunct lipid-lowering therapies was associated with an average absolute reduction in LDL-C levels of roughly 75.0 mg/dL from baseline compared with controls. Overall, use of PCSK9 inhibitors was associated with statistically significant relative risk reductions in MI by 20%, ischaemic stroke by 22%, and coronary revascularization by 17%. Of note, significant and meaningful benefits on clinical ischaemic endpoints were noted with both FDA-approved PCSK9 inhibitors, evolocumab and alirocumab. The observed effects of PCSK9 inhibitors on ASCVD events are in alignment with the mechanistic observations from the Global Assessment of Plaque Regression with a PCSK9 Antibody as Measured by Intravascular Ultrasound (GLAGOV) trial in which patients with angiographic coronary artery disease treated with statins, the addition of evolocumab, compared with placebo, resulted in greater reductions in total atheroma volume and plaque regression.²⁰ Given that ASCVD is in continuum across the coronary, cerebral, and peripheral circulation, the benefits of aggressive cholesterol-lowering therapy extend systemically with correspondent reductions in MI, ischaemic stroke, and possibly major adverse limb events as suggested by recent post hoc analyses from the FOURIER trial.^{21  ,  22}

Although post-marketing studies, observational registries and small RCTs have suggested that LDL-C lowering therapies may be associated with adverse outcomes such as neurocognitive impairment, rhabdomyolysis, or significant enzymes elevation,²³ in this large meta-analysis PCSK9 inhibitors had an excellent safety profile with no significant differences between controls and either evolocumab or alirocumab in terms of major safety endpoints. Particularly for neurocognitive adverse events, our findings are in line with the Evaluating PCSK9 Binding Antibody Influence on Cognitive Health in High Cardiovascular Risk Subjects (EBBINGHAUS) study in which, compared with placebo, evolocumab neither improved nor worsened neurocognition among 1974 patients enrolled in the FOURIER trial.²⁴ With enhanced statistical power, we corroborated these initial observations and further shed light on the safety profile of PCSK9 inhibitors. The only safety endpoint that occurred more frequently with use of PCSK9 inhibitors was injection site reactions; however, this was not paralleled by a greater risk of study-drug discontinuation.

The use of PCSK9 inhibitors overall was not associated with a significant benefit in terms of all-cause and cardiovascular death. However, alirocumab was associated with a statistically significant reduction in all-cause mortality by 17% using random-effect models compared with controls. This finding is mainly driven by the recently reported ODYSSEY-OUTCOMES trial in which the use of alirocumab was associated with a 15% relative reduction in all-cause mortality compared with placebo (P = 0.026) at a median follow-up time of 2.8 years.¹⁵ However, this finding should be considered hypothesis-generating given that, in the ODYSSEY-OUTCOMES trial, all-cause mortality was a secondary endpoint planned to be tested in a hierarchical fashion. Given that in this trial coronary heart disease death (which preceded all-cause death in the hierarchical endpoint) was not found to be statistically significant reduced (P = 0.38), the effect of alirocumab on all-cause mortality remains inconclusive. Although the results of our study do not firmly support a mortality benefit of PCSK9 inhibitors as a class, given the significant reductions in clinical ischaemic endpoints coupled with an excellent safety profile, it remains biologically plausible that a potential survival benefit with this class of therapies could be observed with longer follow-up times and within subset of patients at greater risk of ASCVD events.²⁵ The underlying differences of the study populations and duration of follow-up of the largest trials, FOURIER and ODYSSEY OUTCOMES, may explain the variance between the two agents on long-term mortality observed in our analysis. In addition, longer follow-up may be needed to observe a mortality benefit with use of PCSK9. In fact, in previous trials evaluating statins, a time of exposure of 5–6 years was needed to observe a mortality benefit with cholesterol-lowering therapy.^{26  ,  27}

Limitations

Our study has multiple limitations that need to be disclosed. First, the present findings are subject to the inherent limitations of the included RCTs due to study design, follow-up, definitions, and events ascertainment. Particularly, although all-cause and cardiovascular mortality are now commonly treated as competing risks, this may not have been the case in all included trials, which could further limit our results regarding these endpoints. Second, as we lack patient-level data we remained unable to perform time-to-event analyses and to evaluate the efficacy and safety of PCSK9 inhibition across different levels of baseline patient risk. Consistently, we were unable to further characterize the interplay between the effectiveness of PCSK9 inhibitions and clinical presentation. In addition, we could not directly evaluate the effect between the magnitude of LDL-C lowering and the proportional benefits on hard ischaemic endpoints observed with PCSK9 inhibitors. Third, although the majority of patient enrolled in the included studies were treated with high-intensity statin therapy some received other lipid-lowering therapies due to statin intolerance or other factors. Fourth, inclusion and exclusion criteria and study definitions across RCTs were not homogeneous despite an observed minimal heterogeneity for most of the analysed endpoints. Finally, the control group included a mixture of placebo-controlled and open-label studies; however, effect estimates for the primary efficacy and safety endpoints remained consistent with restriction of the analysis to only placebo-controlled trials.

Conclusions

Across a broad range of patients with hyperlipidaemia or established ASCVD, use of PCSK9 inhibitors significantly reduced the risk of MI, ischaemic stroke, and coronary revascularization. However, the effect of PCSK9 inhibitors on all-cause and cardiovascular mortality as a class remains inconclusive. No major safety issues associated with PCSK9 inhibition were observed. On the basis of this favourable benefit–risk ratio, the results of the present study support the use of PCSK9 inhibitors in clinical practice to mitigate residual ASCVD risk or to reduce LDL-C for patients who cannot tolerate statin therapy.

Conflict of interest: G.M. reports research grants to the Institution or consulting/lecture fees from Abbott, Amgen, Actelion, AstraZeneca, Bayer, Boehringer Ingelheim, Boston-Scientific, Bristol-Myers Squibb, Beth Israel Deaconess Medical, Brigham Women’s Hospital, Cardiovascular Research Foundation, Daiichi-Sankyo, Idorsia, Lilly, Europa, Elsevier, Fédération Française de Cardiologie, ICAN, Medtronic, Journal of the American College of Cardiology, Lead-Up, Menarini, MSD, Novo-Nordisk, Pfizer, Sanofi, Servier, The Mount Sinai School, TIMI Study Group, and WebMD. R.M. has received consulting fees from Abbott Laboratories; Abiomed (spouse); Boston Scientific; Bracco Group; Medscape/Web MD; Siemens Medical Solutions; The Medicines Company (spouse), PLx Opco, Inc,/dba PLx Pharma Inc. (scientific advisory board), Regeneron Pharmaceuticals Inc. (personal - no fee), Roivant Sciences, Inc.; Spectranetics/Phillips/Volcano Corporation (paid to the institution), Sanofi (consultant), janssen Pharmaceuticals, research funding to the institution from Astrazeneca; Bayer; Beth Israel Deaconess; BMS; CSL Behring; Eli Lilly/DSI; Medtronic; Novartis Pharmaceuticals; OrbusNeich; PLC/Renal Guard and Regeneron Pharmaceuticals, Inc.; Holds equity (<1%) in Claret Medical and Elixir Medical; is advisory/speaker for Medteligence (Janssen), advisory board for Bristol-Meyers Squibb (funding to institution), Data safety and monitoring board membership (paid to the institution) for Watermark Research Partners. G.D. has received grants (to the institution) from The Medicines Company (modest level), has served on advisory board for Abbott Vascular, Boston Scientific and Philips, has been a consultant for biosensor, Abiomed and declares research grant to the institution from PLC (Renalguard) and Osprey Medical. U.B. has received speaker honoraria from Boston Scientific and AstraZeneca. S.J.P. has served on Steering Committees or Data Monitoring Committees for trials sponsored by Abbott Vascular, Amirin, AstraZeneca, Bayer, Biosensors, Boehringer Ingelheim, Boston Scientific, Bristol-Myers Squibb, Edwards Lifesciences, Idorsia, Medtronic, Novartis, and Vifor. R.S.R. has received research grants from Akcea, Amgen, Astra Zeneca, Medicines Company, and Regeneron; has received consulting fees/honoraria from Akcea, Amgen, Kowa, and Pfizer; and has stock holdings in MediMergent, and royalties from UpToDate. All other authors have reported that they have no relationships relevant to the contents of this article to disclose.

References

Author notes