JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL. 75, NO. 2, 2020 ª 2020 THE AUTHORS. PUBLISHED BY ELSEVIER ON BEHALF OF THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION. THIS IS AN OPEN ACCESS ARTICLE UNDER THE CC BY-NC-ND LICENSE (http://creativecommons.org/licenses/by-nc-nd/4.0/). ORIGINAL INVESTIGATIONS Effect of Alirocumab on Lipoprotein(a) and Cardiovascular Risk After Acute Coronary Syndrome Vera A. Bittner, MD, MSPH,a Michael Szarek, PHD,b Philip E. Aylward, BM, BCH, PHD,c Deepak L. Bhatt, MD, MPH,d Rafael Diaz, MD,e Jay M. Edelberg, MD, PHD,f Zlatko Fras, MD,g,h Shaun G. Goodman, MD,i,j Sigrun Halvorsen, MD,k,l Corinne Hanotin, MD,m Robert A. Harrington, MD,n J. Wouter Jukema, MD, PHD,o Virginie Loizeau, MS,m Patrick M. Moriarty, MD,p Angèle Moryusef, MD,f Robert Pordy, MD,q Matthew T. Roe, MD, MHS,r,s Peter Sinnaeve, MD,t,u Sotirios Tsimikas, MD,v Robert Vogel, MD,w Harvey D. White, DSC,x Doron Zahger, MD,y Andreas M. Zeiher, MD,z Ph. Gabriel Steg, MD,aa,bb Gregory G. Schwartz, MD, PHD,w for the ODYSSEY OUTCOMES Committees and Investigators ABSTRACT BACKGROUND Lipoprotein(a) concentration is associated with cardiovascular events. Alirocumab, a proprotein con- vertase subtilisin/kexin type 9 inhibitor, lowers lipoprotein(a) and low-density lipoprotein cholesterol (LDL-C). OBJECTIVES A pre-specified analysis of the placebo-controlled ODYSSEY Outcomes trial in patients with recent acute coronary syndrome (ACS) determined whether alirocumab-induced changes in lipoprotein(a) and LDL-C independently predicted major adverse cardiovascular events (MACE). METHODS One to 12 months after ACS, 18,924 patients on high-intensity statin therapy were randomized to alirocumab or placebo and followed for 2.8 years (median). Lipoprotein(a) was measured at randomization and 4 and 12 months thereafter. The primary MACE outcome was coronary heart disease death, nonfatal myocardial infarction, ischemic stroke, or hospitalization for unstable angina. RESULTS Baseline lipoprotein(a) levels (median: 21.2 mg/dl; interquartile range [IQR]: 6.7 to 59.6 mg/dl) and LDL-C [corrected for cholesterol content in lipoprotein(a)] predicted MACE. Alirocumab reduced lipoprotein(a) by 5.0 mg/dl (IQR: 0 to 13.5 mg/dl), corrected LDL-C by 51.1 mg/dl (IQR: 33.7 to 67.2 mg/dl), and reduced the risk of MACE (hazard ratio [HR]: 0.85; 95% confidence interval [CI]: 0.78 to 0.93). Alirocumab-induced reductions of lipoprotein(a) and corrected LDL-C independently predicted lower risk of MACE, after adjustment for baseline concentrations of both lipoproteins and demographic and clinical characteristics. A 1-mg/dl reduction in lipoprotein(a) with alirocumab was associated with a HR of 0.994 (95% CI: 0.990 to 0.999; p ¼ 0.0081). CONCLUSIONS Baseline lipoprotein(a) and corrected LDL-C levels and their reductions by alirocumab predicted the risk of MACE after recent ACS. Lipoprotein(a) lowering by alirocumab is an independent contributor to MACE reduction, which suggests that lipoprotein(a) should be an independent treatment target after ACS. (ODYSSEY Outcomes: Evaluation of Cardiovascular Outcomes After an Acute Coronary Syndrome During Treatment With Alirocumab; NCT01663402) (J Am Coll Cardiol 2020;75:133–44) © 2020 The Authors. Published by Elsevier on behalf of the American College of Cardiology Foundation. This is an open access article under the CC BY-NC-ND license Listen to this manuscript’s (http://creativecommons.org/licenses/by-nc-nd/4.0/). audio summary by Editor-in-Chief Dr. Valentin Fuster on JACC.org. From the aDivision of Cardiovascular Disease, University of Alabama at Birmingham, Birmingham, Alabama; bState University of New York, Downstate School of Public Health, Brooklyn, New York; cSouth Australian Health and Medical Research Institute, Flinders University and Medical Centre, Adelaide, South Australia, Australia; dBrigham and Women’s Hospital Heart & Vascular Center and Harvard Medical School, Boston, Massachusetts; eEstudios Cardiológicos Latinoamérica, Instituto Cardiovascular de Rosario, Rosario, Argentina; fSanofi, Bridgewater, New Jersey; g Division of Medicine, Department of Vascular Medicine, ISSN 0735-1097 https://doi.org/10.1016/j.jacc.2019.10.057 134 Bittner et al. JACC VOL. 75, NO. 2, 2020 Lipoprotein(a) and Cardiovascular Events JANUARY 21, 2020:133–44 Preventive Cardiology Unit, University Medical Centre Ljubljana, Ljubljana, Slovenia; hMedical Faculty, University of Ljubljana, Ljubljana, Slovenia; iCanadian VIGOUR Centre, University of Alberta, Edmonton, Alberta, Canada; jSt. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada; kDepartment of Cardiology, Oslo University Hospital, Oslo, Norway; lUniversity m n of Oslo, Oslo, Norway; Sanofi, Paris, France; Stanford Center for Clinical Research, Department of Medicine, Stanford University, Stanford, California; oDepartment of Cardiology, Leiden University Medical Center, Leiden, the Netherlands; pDivision of Clinical Pharmacology, University of Kansas Medical Center, Kansas City, Kansas; qRegeneron Pharmaceuticals Inc., Tarrytown, New York; rDuke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina; sDivision of Cardiology, Department of Medicine, Duke University School of Medicine, Durham, North Carolina; tDepartment of Cardiovascular Medicine, University Hospitals Leuven, Leuven, Belgium; uUniversity of Leuven, Leuven, Belgium; vDivision of Cardiovascular Medicine, University of California San Diego, La Jolla, California; wDivision of Cardiology, University of Colorado School of Medicine, Aurora, Colorado; xGreen Lane Cardiovascular Services Auckland City Hospital, Auckland, New Zealand; ySoroka University Medical Center, Faculty of Health Sciences, Ben Gurion University of the Negev, Beer Sheva, Israel; zDepartment of Medicine III, Goethe aa University, Frankfurt am Main, Germany; Assistance Publique-Hôpitaux de Paris, Hôpital Bichat, Université de Paris, FACT bb (French Alliance for Cardiovascular Trials), INSERM U1148, Paris, France; and the National Heart and Lung Institute, Imperial College, Royal Brompton Hospital, London, United Kingdom. The ODYSSEY OUTCOMES trial was funded by Sanofi and Regen- eron Pharmaceuticals. Dr. Bittner serves on the Executive Steering Committee of the ODYSSEY OUTCOMES trial (Sanofi), as National Coordinator for STRENGTH (AstraZeneca), DalGene (Dalcor), CLEAR (Esperion), and as site investigator for ORION IV (The Medicines Company), all contracted through the University of Alabama at Birmingham; has previously served as site investigator for ARTEMIS (Astra Zeneca) and COMPASS (Bayer Healthcare), both contracted through the University of Alabama at Birmingham; and has consulted for Sanofi. Dr. Szarek has served as a consultant or on Advisory Boards (or both) for CiVi, Resverlogix, Baxter, Esperion, and Regeneron Pharmaceuticals. Dr. Aylward has received research support, speaker fees, and served on Advisory Boards for Sanofi, Amgen, AstraZeneca, CSL, Bayer, Novartis, and Boehringer Ingelheim. Dr. Bhatt has served on Advisory Board for Cardax, Cereno Scientific, Elsevier Practice Update Cardiology, Medscape Cardiology, PhaseBio, and Regado Biosciences; has served on the Board of Directors for Boston VA Research Institute, Society of Cardiovascular Patient Care, and TobeSoft; has served as Chair for American Heart Association Quality Oversight Committee; has served on Data Monitoring Committees for Baim Institute for Clinical Research (formerly Harvard Clinical Research Institute, for the PORTICO trial, funded by St. Jude Medical, now Abbott), Cleveland Clinic (including for the ExCEED trial, funded by Edwards), Duke Clinical Research Institute, Mayo Clinic, Mount Sinai School of Medicine (for the ENVISAGE trial, funded by Daiichi-Sankyo), and the Population Health Research Institute; has received honoraria from American College of Cardiology (Senior Associate Editor, Clinical Trials and News, ACC.org; Vice-Chair, ACC Accreditation Committee), Baim Institute for Clinical Research (formerly Harvard Clinical Research Institute; RE-DUAL PCI clinical trial steering committee funded by Boehringer Ingelheim; AEGIS-II executive committee funded by CSL Behring), Belvoir Publications (Editor-in-Chief, Harvard Heart Letter), Duke Clinical Research Institute (clinical trial steering committees, including for the PRONOUNCE trial, funded by Ferring Pharmaceuticals), HMP Global (Editor-in-Chief, Journal of Invasive Cardiology), Journal of the American College of Cardiology (Guest Editor; Associate Editor), Medtelligence/ ReachMD (CME steering committees), Population Health Research Institute (for the COMPASS operations committee, publications committee, steering committee, and USA national co-leader, funded by Bayer), Slack Publications (Chief Medical Editor, Cardi- ology Today’s Intervention), Society of Cardiovascular Patient Care (Secretary/Treasurer), and WebMD (CME steering committees); has other relationships with Clinical Cardiology (Deputy Editor), NCDR-ACTION Registry Steering Committee (Chair), and VA CART Research and Publications Committee (Chair); has received research funding from Abbott, Afimmune, Amarin, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Chiesi, CSL Behring, Eisai, Ethicon, Ferring Pharmaceuticals, Forest Laboratories, Fractyl, Idorsia, Ironwood, Ischemix, Lilly, Medtronic, PhaseBio, Pfizer, PLx Pharma, Regeneron, Roche, Sanofi, Synaptic, and The Medicines Company; has received royalties from Elsevier (Editor, Cardiovascular Intervention: A Companion to Braunwald’s Heart Disease); has served as site co-investigator for Biotronik, Boston Scientific, CSI, St. Jude Medical (now Abbott), and Svelte; has served as a trustee for American College of Cardiology; and has performed unfunded research for FlowCo, Merck, Novo Nordisk, and Takeda. Dr Diaz has received research grants from Sanofi, DalCor Pharmaceuticals, Population Health Research Institute, Duke Clinical Research Institute, the TIMI group, Amgen, Cirius, Montreal Health Innovations Coor- dinating Center, and Lepetit; and has received personal fees, as a member of the Executive Steering Committee, from Amgen and Cirius. Dr. Edelberg was an employee of Sanofi during the conduct of the study and is now an employee of Myokardia. Dr. Fras serves as National Coordinator for the ODYSSEY OUTCOMES Trial (Sanofi); and has received research support and speaker/ consulting fees from Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Krka, Novo Nordisk, Pfizer, and Sanofi. Dr. Goodman has received research grant support (e.g., steering committee or data monitoring committee) and/or speaker/consulting honoraria (e.g., advisory boards) from Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, CSL Behring, Daiichi-San- kyo/American Regent, Eli Lilly, Esperion, Ferring Pharmaceuticals, GlaxoSmithKline, HLS Therapeutics, Janssen/Johnson & Johnson, Merck, Novartis, Novo Nordisk A/C, Pfizer, Regeneron, Sanofi, and Servier; and has received salary support from the Heart and Stroke Foundation of Ontario/University of Toronto (Polo) Chair, Canadian Heart Research Centre and MD Primer, Canadian VIGOUR Centre, Duke Clinical Research Institute, New York University Clinical Coordinating Centre, and PERFUSE. Dr. Halvorsen has received consultant fees and speaking honoraria from AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Merck, Novartis, Pfizer, and Sanofi. Dr. Hanotin is an employee of Sanofi. Dr. Harrington has received research grants from Apple, CSL, Sanofi, AstraZeneca, Portola, Janssen, Bristol-Myers Squibb, Novartis, and The Medicines Company; has served as a consultant or on the Advisory Board for Amgen, Bayer, Gilead, MyoKardia, and WebMD; and is on the Board of Directors (unpaid) for AHA and Stanford HealthCare. Dr. Jukema has received research grants from the Netherlands Heart Foundation, the Inter- university Cardiology Institute of the Netherlands, and the European Commission Seventh Framework Programme; and has received research support from Amgen, Astellas, AstraZeneca, Daiichi-Sankyo, Lilly, Merck-Schering-Plough, Pfizer, Roche, and Sanofi. Ms. Loizeau is an employee of and holds shares in Sanofi. Dr. Moriarty has received research grants from and/or has consulted for Regeneron, Sanofi, Amgen, Esperion, Kaneka, Gemphire, Ionis, Novartis, Stage II Innovations, Kowa, Akcea, Pfizer, FH Foundation, University of Penn, and Aegerion; and is a speaker for Regeneron, Sanofi, Amgen, and Amarin. Dr. Moryusef is an employee of Sanofi. Dr. Pordy is an employee of Regeneron Pharmaceuticals. Dr. Roe has received research grant funding from JACC VOL. 75, NO. 2, 2020 Bittner et al. 135 JANUARY 21, 2020:133–44 Lipoprotein(a) and Cardiovascular Events L ipoprotein(a), a genetically determined low- (but not United States) guidelines suggest ABBREVIATIONS density lipoprotein particle that contains that lipoprotein(a) is a potential target for AND ACRONYMS apolipoprotein(a) and apolipoprotein B-100 treatment if concentrations are $50 mg/dl ACS = acute coronary moieties, is believed to possess pro-atherogenic, (6,7). syndrome pro-thrombotic, pro-inflammatory, and pro- Although observational data with apher- CI = confidence interval oxidative properties (1). High levels of lipoprotein(a) esis suggest a possible benefit of lip- HR = hazard ratio have been associated with incident cardiovascular oprotein(a) lowering on cardiovascular IQR = interquartile range disease in most, but not all, population-based epide- outcomes (8), no randomized data to date LDL-C = low-density miological analyses (1,2) and in patients with estab- indicate that medications that lower lip- lipoprotein cholesterol lished coronary heart disease (3,4). Moreover, oprotein(a) reduce cardiovascular risk MACE = major adverse Mendelian randomization analysis supports a linear through that mechanism. Niacin reduces lip- cardiovascular events relationship between lipoprotein(a) concentration oprotein(a) by 15% to 25% but does not nonHDL-C = nonLhigh- and incident coronary heart disease in the general reduce death or ischemic cardiovascular density lipoprotein cholesterol population (5). Based on available data, European events (9,10), and anacetrapib lowers PCSK9 = proprotein convertase subtilisin/kexin type 9 Sanofi, AstraZeneca, Patient Centered Outcomes Research Institute, Ferring Pharmaceuticals, Myokardia, Familial Hypercholes- terolemia Foundation, and Bayer; and has received consulting fees or honoraria from AstraZeneca, Amgen, Cytokinetics, Eli Lilly, Roche-Genentech, Janssen Pharmaceuticals, Regeneron, Novo Nordisk, Pfizer, Sanofi, Signal Path, and Elsevier Publishers. Dr. Sinnaeve has received personal fees from Sanofi, Abbott, Boehringer Ingelheim, Bristol-Myers Squibb, Pfizer, and Amgen; has received personal fees and nonfinancial support from AstraZeneca and Merck Sharp Dohme; and has received grants and personal fees from Daiichi-Sankyo and Bayer. Dr. Tsimikas has served as a consultant to Boston Heart Diagnostics; is a co-inventor and receives royalties from patents owned by UCSD on oxidation-specific antibodies and of biomarkers related to oxidized lipopro- teins; has a dual appointment at UCSD and Ionis Pharmaceuticals; and is a co-founder of and has an equity interest in Oxitope, Inc. and Kleanthi Diagnostics, LLC; (Although these relationships have been identified for conflict of interest management based on the overall scope of the project and its potential benefit to Oxitope and Kleanthi Diagnostics LLC, the research findings included in this particular publication may not necessarily relate to the interests of Oxitope, Inc. and Kleanthi Diagnostics, LLC. The terms of this arrangement have been reviewed and approved by the University of California, San Diego in accordance with its conflict of interest policies.). Dr. Vogel has received grants and personal fees from Sanofi. Dr. White has received grant support paid to the institution and fees for serving on a steering committee for the ODYSSEY OUTCOMES trial (Evaluation of Cardiovascular Out- comes After an Acute Coronary Syndrome During Treatment With Alirocumab) from Sanofi and Regeneron Pharmaceuticals, for the ACCELERATE study (A Study of Evacetrapib in High-Risk Vascular Disease) from Eli Lilly, for the STRENGTH trial (Outcomes Study to Assess Statin Residual Risk Reduction With EpaNova in High CV Risk Patients With Hypertriglyceridemia) from Omthera Pharmaceuticals, for the SPIRE trial (The Evaluation of Bococizumab [PF-04950615; RN 316] in Reducing the Occurrence of Major Cardiovascular Events in High Risk Subjects) from Pfizer USA, for the HEART-FID study (Randomized Placebo-Controlled Trial of FCM as Treatment for Heart Failure With Iron Deficiency) from American Regent, for the CAMELLIA-TIMI study (A Study to Evaluate the Effect of Long-term Treatment With BELVIQ [Lorcaserin HC] on the Incidence of Major Adverse Cardiovascular Events and Conversion to Type 2 Diabetes Mellitus in Obese and Overweight Subjects With Cardiovascular Disease or Multiple Cardiovascular Risk Factors) from Eisai Inc, for the dal-GenE study (Effect of Dalcetrapib vs Placebo on CV Risk in a Genetically Defined Population With a Recent ACS) from DalCor Pharma UK Inc., for the AEGIS-II study from CSL Behring, for the SCORED trial (Effect of Sotagliflozin on Cardiovascular and Renal Events in Patients With Type 2 Diabetes and Moderate Renal Impairment Who Are at Cardiovascular Risk) and the SOLOIST-WHF trial (Effect of Sotagliflozin on Cardiovascular Events in Patients With Type 2 Diabetes Post Worsening Heart Failure) from Sanofi-Aventis Australia Pty. Ltd., and for the CLEAR Outcomes Study (Evaluation of Major Cardiovascular Events in Patients With, or at High Risk for, Cardiovascular Disease Who Are Statin Intolerant Treated With Bempedoic Acid [ETC-1002] or Placebo) from Esperion Therapeutics Inc.; has served on the Advisory Board for Acetelion, Sirtex, and Genentech, Inc. (an affiliate of F. Hoffmann-La Roche Ltd., “Roche”; Lytics Post-PCI Advisory Board at European Society of Cardiology); and has received lecture fees from AstraZeneca. Dr. Zahger serves as National Coordinator for the ODYSSEY OUTCOMES trial and the SCORED trial, both funded by Sanofi; and has consulted for Bayer, AstraZeneca, Boehringer Ingelheim, NovoNordisk, and Sanofi. Dr. Zeiher has received fees for serving on a steering committee for the ODYSSEY OUT- COMES trial from Sanofi; and has received Advisory Board and speaker fees from Sanofi, Amgen, Boehringer Ingelheim, Bayer, Novartis, Pfizer, AstraZeneca, and Vifor. Dr. Baccara-Dinet is an employee of and holds shares in Sanofi. Dr. Steg has received grants and nonfinancial support (co-chair of the ODYSSEY OUTCOMES trial; as such he received no personal fees, but his insti- tution has received funding for the time he has devoted to trial coordination, and he has received support for some travel related to trial meetings) from Sanofi; has received research grants and personal fees from Bayer (Steering Committee MARINER, grant for epidemiological study), Merck (speaker fees, grant for epidemiological studies), Sanofi (co-chair of the ODYSSEY OUTCOMES trial; cochair of the SCORED trial; consulting, speaking), Servier (Chair of the CLARIFY registry; grant for epidemiological research), and Amarin (executive steering committee the REDUCE-IT trial [Disease Reduction of Cardiovascular Events With Icosapent Ethyl– Intervention Trial]; consulting); has received personal fees from Amgen, Bristol-Myers Squibb, Boehringer Ingelheim, Pfizer, Novartis, Regeneron Pharmaceuticals, Lilly, and AstraZeneca; and has a European application number/patent number, issued on October 26, 2016 (No. 15712241.7), for a method for reducing cardiovascular risk. Dr. Schwartz has received research grants to the University of Colorado from Resverlogix, Roche, Sanofi, and The Medicines Company; and is co-inventor of pending U.S. patent 62/806313 “Methods for Reducing Cardiovascular Risk,” assigned in full to University of Colorado. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Manuscript received June 17, 2019; revised manuscript received October 7, 2019, accepted October 28, 2019. 136 Bittner et al. JACC VOL. 75, NO. 2, 2020 Lipoprotein(a) and Cardiovascular Events JANUARY 21, 2020:133–44 lipoprotein(a) by 25% with only modest cardiovascu- myocardial infarction, fatal or nonfatal ischemic lar benefits, which are likely explained by other ef- stroke, or unstable angina that required hospitaliza- fects on the lipid profile (11). Proprotein convertase tion. Secondary endpoints considered in the present subtilisin/kexin type 9 (PCSK9) inhibitors lower lip- analysis were coronary heart disease death or oprotein(a) concentrations by approximately 25% nonfatal myocardial infarction, fatal or nonfatal (12,13) and reduce cardiovascular events (14–16), but ischemic stroke, cardiovascular death, and all-cause it is uncertain whether, and to what extent, reduction death. Unstable angina was not considered individu- of lipoprotein(a) contributes to this benefit, inde- ally because of a small number of events. All end- pendent of the concurrent reduction of low-density points were adjudicated by a committee blinded to lipoprotein cholesterol (LDL-C). treatment assignment and lipid levels under the In a pre-specified analysis of the ODYSSEY Out- auspices of the Duke Clinical Research Institute. comes (ODYSSEY Outcomes: Evaluation of Cardio- MEASUREMENT OF LIPOPROTEINS. Lipoprotein(a) vascular Outcomes After an Acute Coronary mass was measured once at randomization, at Syndrome During Treatment With Alirocumab) trial, 4 months, and at 12 months at COVANCE Central we tested the hypotheses that baseline lipoprotein(a) Laboratories (Los Angeles, California) using an auto- predicted recurrent major adverse cardiovascular mated immunoturbidimetric assay on a Siemens BNII events (MACE) following an index acute coronary (Siemens, Healthcare Diagnostics, Malvern, Pennsyl- syndrome (ACS) in patients who received intensive vania) validated against the International Federation statin therapy. We also examined if the decrease in of Clinical Chemistry and World Health Organization lipoprotein(a) concentration with treatment using the standards (18). The interassay coefficient of variation PCSK9 inhibitor, alirocumab, was associated with a ranged from 3.1% to 4.8%, depending on the lip- decreased risk of MACE, independent of the concur- oprotein(a) concentration. Apolipoprotein(a) size rent reduction of LDL-C. heterogeneity had only a moderate effect on lip- SEE PAGE 145 oprotein(a) recovery with this assay. LDL-C was calculated by the Friedewald formula (19), except METHODS when triglycerides were >400 mg/dl (4.52 mmol/l) or when the Friedewald-calculated LDL-C was <15 mg/dl PATIENTS. Details of the ODYSSEY OUTCOMES trial (0.39 mmol/l). In these cases, LDL-C was measured by design and results have been previously published beta-quantification. (16,17). In brief, the trial included 18,924 patients age Calculated or directly measured LDL-C levels 40 years or older who experienced an ACS 1 to include cholesterol contained in lipoprotein(a), which 12 months before randomization and who had a corresponds to approximately 30% of the lip- LDL-C level of $70 mg/dl (1.81 mmol/l), nonhigh- oprotein(a) mass (20,21). To account for this and density lipoprotein cholesterol (nonHDL-C) level derive an estimate of cholesterol contained in LDL of $100 mg/dl (2.59 mmol/l), or an apolipoprotein B particles, we calculated corrected LDL-C (referred to level of $80 mg/dl on high-intensity statin therapy herein as LDL-C corr) using the formula (21): (atorvastatin 40 to 80 mg daily, rosuvastatin 20 to LDL-Ccorr ¼ LDL-C  0:3  lipoproteinðaÞ mass 40 mg daily, or the maximum tolerated dose of either). Study inclusion was not based on lip- Similarly, to derive an estimate of cholesterol car- oprotein(a) concentrations. The trial was approved by ried in all apolipoprotein-Bcontaining particles the institutional review board of each site, and all other than lipoprotein(a), corrected nonHDL-C was patients provided informed consent. calculated using the relationship: TREATMENTS. Patients were randomly assigned to non  HDL-Ccorr ¼ ðtotal cholesterol  HDL-CÞ  0:3 treatment with alirocumab 75 mg subcutaneously  lipoproteinðaÞ mass every 2 weeks or matching placebo. Among patients assigned to alirocumab, the dose was blindly STATISTICAL ANALYSIS. Lipoprotein(a), LDL-C, and increased to 150 mg in patients who did not achieve nonHDL-C distributions were assessed for the an LDL-C level of <50 mg/dl (1.29 mmol/l). Placebo overall population and by treatment group at base- was blindly substituted for alirocumab in patients line and at months 4 and 12 (4 weeks) after who had 2 consecutive LDL-C measurements randomization. If a patient had multiple values of <15 mg/dl (0.39 mmol/l). within each of these periods, the last value was ENDPOINTS. The primary endpoint (MACE) was a analyzed. Missing values were imputed by pre- composite of coronary heart disease death, nonfatal specified methods. JACC VOL. 75, NO. 2, 2020 Bittner et al. 137 JANUARY 21, 2020:133–44 Lipoprotein(a) and Cardiovascular Events Baseline characteristics were assessed by lip- change in lipoprotein(a) and MACE was modified by oprotein(a) quartile and compared across quartiles by adjustment for the simultaneous change in LDL-C corr. chi-square tests for categorical variables and Kruskal- Similarly, a comparison of models 2 and 3B indicated Wallis tests for continuous variables. Relationships whether the relationship between the change in lip- between baseline lipoprotein(a) and endpoint events oprotein(a) and MACE was modified by adjustment in the placebo group were determined by Cox pro- for the simultaneous change in all other apolipopro- portional hazards models using the baseline lip- tein-Bcontaining lipoproteins. Effects are summa- oprotein(a) quartile as the predictor variable. We rized by HRs per 1-mg/dl reduction and the observed constructed unadjusted models and models that median reduction in lipoprotein(a) (all models) adjusted for demographic and clinical variables (age, and in LDL-Ccorr or nonHDL-C corr (models 3A, 3B, sex, race, geographic region, body mass index, 4A, and 4B) at month 4. The predicted absolute smoking history, diabetes, time from index ACS to reduction in the risk of MACE with alirocumab randomization) and baseline LDL-Ccorr. p Values were attributable to lowering of lipoprotein(a) and to computed for linear trends across baseline lip- simultaneous lowering of LDL-C corr (model 3A) or oprotein(a) quartiles. A spline analysis of degree 3 nonHDL-Ccorr (model 3B) 4 years after randomiza- (piecewise cubic curve) of the relationship between tion was calculated at the 25th (6.7 mg/dl), 50th continuous baseline lipoprotein(a) and MACE in the (21.2 mg/dl), and 75th (59.6 mg/dl) percentiles of placebo group was performed, setting the hazard ra- baseline lipoprotein(a), using the relationships be- tio (HR) to 1.00 at the overall baseline median tween the variables and baseline lipoprotein(a) (21.2 mg/dl) concentration of lipoprotein(a) with described in the Online Appendix. Attribution for natural cubic basis and 3 knots, located at the overall each parameter was based on its contribution to 25th percentile (6.7 mg/dl), median (21.2 mg/dl), and the predicted absolute risk reduction relative to 75th percentile (59.6 mg/dl). The p value for the no change in lipoprotein(a) and LDL-C corr or spline effect was based on the score test. nonHDL-Ccorr . Heterogeneity in the relative and absolute effects All analyses were conducted by an independent of alirocumab treatment on MACE were assessed academic statistical team at the State University of according to baseline lipoprotein(a) quartile. To New York Downstate School of Public Health using assess the former, we constructed a Cox propor- SAS version 9.4 (SAS Institute, Cary, North Carolina). tional hazards model with baseline lipoprotein(a) RESULTS quartile, treatment, and their interaction as pre- dictors, as well as a baseline hazard stratification BASELINE CHARACTERISTICS. The distribution of by geographic region. To assess the latter, absolute baseline lipoprotein(a) was highly skewed, with a risk reductions with alirocumab treatment, quanti- median of 21.2 mg/dl (interquartile range [IQR]: 6.7 to fied as differences in observed incidences between 59.6 mg/dl) (Online Figure 1); 16% of patients had the treatment groups, were compared across baseline minimum value for the assay of 2.0 mg/dl. Baseline lipoprotein(a) quartiles using a Gail-Simon test (22). characteristics of the patients by lipoprotein(a) quar- To determine the association between modification tile are shown in Online Table 1. Participants in the of lipoprotein(a) levels by alirocumab treatment and upper lipoprotein(a) quartiles were more likely to be MACE, the relationships between the change in lip- women, black, and from North America but less likely oprotein(a) from baseline to month 4 and the risk of to smoke or have diabetes. LDL-C and nonHDL-C MACE after month 4 were described using Cox pro- concentrations and the percentage of patients treated portional hazards models in the alirocumab group. with high-intensity statin were highest in the highest The following models were developed: a model quartile of lipoprotein(a). Conversely, LDL-Ccorr and without covariates (model 1); a model adjusted for nonHDL-Ccorr decreased across increasing quartiles baseline lipoprotein(a) (model 2); a model addition- of lipoprotein(a). Participants in the higher lip- ally adjusted for either baseline LDL-Ccorr and change oprotein(a) quartiles were more likely to have from baseline to month 4 in LDL-C corr (model 3A) or had blinded up titration of alirocumab and less likely baseline nonHDL-C corr and change from baseline to to have had blinded substitution of placebo month 4 in nonHDL-Ccorr (model 3B); and a model for alirocumab. adjusted for all variables in model 3, as well as the previously mentioned demographic and clinical var- BASELINE LIPOPROTEIN(A), CARDIOVASCULAR iables with either LDL-Ccorr (model 4A) or nonHDL- EVENTS, AND MORTALITY IN THE PLACEBO GROUP. C corr (model 4B). A comparison of models 2 and 3A Median follow-up was 2.8 years (IQR: 2.3 to 3.4 years). indicated whether the relationship between the The relationship between baseline lipoprotein(a) 138 Bittner et al. JACC VOL. 75, NO. 2, 2020 Lipoprotein(a) and Cardiovascular Events JANUARY 21, 2020:133–44 F I G U R E 1 Baseline Lipoprotein(a) Quartile as a Predictor of Events in the Placebo Group p < 0.0001* p < 0.0001* 12.5 10.0 Incidence (%) 7.5 ns ns ns 5.0 2.5 0.0 MACE CHD Death Ischemic CV Death All-Cause + Nonfatal MI Stroke Death Quartile 1 Quartile 2 Quartile 3 Quartile 4 (<6.7 mg/dl) (6.7 to <21.2 mg/dl) (21.2 to <59.6 mg/dl) (≥59.6 mg/dl) Incidence of major adverse cardiovascular events (MACE) and secondary endpoints are shown by baseline lipoprotein(a) quartiles for the placebo group. MACE and coronary heart disease (CHD) death and nonfatal myocardial infarction (MI) incidence were greater at higher lipoprotein(a) levels. There were no significant relationships between lipoprotein(a) quartile and ischemic stroke, cardiovascular (CV) death, or all-cause death. *p < 0.0001 for linear trend across quartiles in unadjusted Cox proportional hazards model for MACE and coronary heart disease death and nonfatal myocardial infarction. p > 0.05 for other endpoints. quartile and incidence of events in the placebo group to 0.93; p < 0.001) with an absolute risk reduction of is shown in Figure 1, and modeling of this relationship 1.6%. There was no statistically significant interaction is shown in Online Table 2. The occurrence of MACE between treatment and baseline lipoprotein(a) quar- and coronary heart disease death and/or nonfatal tile on the relative risk of MACE (pinteraction ¼ 0.55) myocardial infarction increased significantly from the (Figure 3, left). In contrast, absolute risk reduction in lowest to the highest lipoprotein(a) baseline quartile. MACE with alirocumab was several-fold higher in the In unadjusted Cox proportional hazards models, upper quartiles (2.3% and 2.1%) than in the lower participants in the highest, compared with the lowest, quartiles of baseline lipoprotein(a) (0.4% and 1.4%, baseline quartile of lipoprotein(a) were 46% and 54% respectively), but there was evidence that all were more likely to experience MACE and nonfatal positive (pinteraction ¼ 0.0011) (Figure 3, right). The myocardial infarction and/or coronary heart disease numbers of patients needed to treat with alirocumab death, respectively. These relationships were for a median of 2.8 years to prevent 1 event were 238, numerically stronger after adjustment for baseline 69, 43, and 49 in quartiles 1, 2, 3, and 4 of baseline LDL-Ccorr . There were no significant relationships lipoprotein(a), respectively. between baseline lipoprotein(a) quartile and ischemic Effect of alirocumab on lipoprotein(a) levels. Online stroke, cardiovascular death, or all-cause death. Figure 2 shows the medians and IQRs of lip- Spline analysis of continuous baseline lipoprotein(a) oprotein(a) concentrations by baseline quartile of and the HR for MACE (Figure 2) indicated a relatively lipoprotein(a) and treatment group. Baseline distri- linear relationship between baseline lipoprotein(a) butions of lipoprotein(a) were similar in both and the risk of MACE. treatment groups. At month 4, lipoprotein(a) con- Effect of alirocumab on MACE stratified by centrations were significantly lower in the alirocumab baseline lipoprotein(a) q u a r t i l e . Relative and group than in the placebo group, with levels absolute treatment effects on MACE stratified by remaining stable at month 12. Figure 4 shows the baseline lipoprotein(a) quartile are shown in Figure 3. absolute change from baseline to month 4 in lip- Overall, the HR for MACE (alirocumab and/or oprotein(a), LDL-C, and LDL-Ccorr in the alirocumab placebo) was 0.85 (95% confidence interval [CI]: 0.78 (Figure 4A) and placebo groups (Figure 4B). Overall, JACC VOL. 75, NO. 2, 2020 Bittner et al. 139 JANUARY 21, 2020:133–44 Lipoprotein(a) and Cardiovascular Events F I G U R E 2 Spline Analysis of Continuous Baseline Lipoprotein(a) and the Primary Endpoint Hazard Ratio (MACE) in the Placebo Group 1.50 1.25 MACE HR 1.00 0.75 0 25 50 75 100 125 Baseline Lp(a) (mg/dl) Hazard ratio (HR) set to 1.00 at the overall baseline median (21.2 mg/dl) concentration of lipoprotein(a). p < 0.0001 for spline effect. This reflects spline of degree 3 (piecewise cubic curve) with natural cubic basis and 3 knots, located at overall 25th percentile (6.7 mg/dl), median (21.2 mg/dl), and 75th percentile (59.6 mg/dl). The p value for spline effect is based on the score test. The blue dotted line indicates the upper bound of 95% confidence intervals (CIs) and the red dotted line indicates the lower bound of 95% CIs. Lp(a) ¼ lipoprotein(a); other abbreviation as in Figure 1. the median relative and absolute changes in lip- lipoprotein(a) quartiles, there was a substantial oprotein(a) from baseline to month 4 in the alir- gradient of median absolute change that ranged from ocumab group were 23% (IQR: 47% to 0%) 0 mg/dl in quartile 1 to 20.2 mg/dl in quartile 4. Most and 5.0 mg/dl (IQR 13.5 to 0 mg/dl), respectively. patients in quartile 1 had no change in lipoprotein(a) Although the relative change in lipoprotein(a) with from baseline to month 4, whereas >80% of patients alirocumab treatment was similar across baseline in each of the other quartiles had decreases. Changes F I G U R E 3 Relative and Absolute Treatment Effect on MACE by Baseline Lipoprotein(a) Quartile MACE Incidence Absolute Risk Alirocumab Placebo HR (95% CI) Reduction (95% CI) Subgroup n/N (%) n/N (%) (pinteraction = 0.55) (pinteraction = 0.0011) Quartile 1 211/2,327 (9.1) 228/2,403 (9.5) 0.95 (0.79, 1.15) 0.4% (–1.2, 2.1) Quartile 2 219/2,438 (9.0) 239/2,293 (10.4) 0.85 (0.71, 1.03) 1.4% (–0.3, 3.1) Quartile 3 212/2,356 (9.0) 269/2,373 (11.3) 0.79 (0.66, 0.94) 2.3% (0.6, 4.1) Quartile 4 261/2,341 (11.2) 316/2,393 (13.2) 0.83 (0.70, 0.98) 2.1% (0.2, 3.9) Overall 903/9,462 (9.5) 1,052/9,462 (11.1) 0.85 (0.78, 0.93) 1.6% (0.7, 2.4) 0.70 0.85 1.0 3.2% 1.6% 0% Alirocumab Placebo Alirocumab Placebo Better Better Better Better Incidence of MACE is shown for the alirocumab and placebo groups stratified by baseline quartile of lipoprotein(a) and for the overall population. The Forest plots depict relative and absolute risk reduction with alirocumab compared with placebo. For relative risk reduction, there was no significant interaction by baseline lipoprotein(a) quartile, but absolute treatment effect was significantly greater in the 2 highest lipoprotein(a) quartiles. Abbreviations as in Figures 1 and 2. 140 Bittner et al. JACC VOL. 75, NO. 2, 2020 Lipoprotein(a) and Cardiovascular Events JANUARY 21, 2020:133–44 F I G U R E 4 Absolute Change From Baseline to Month 4 in LDL-C, LDL-C corr , and Lipoprotein(a) Levels A B 20 40 Absolute Change (mg/dl) Absolute Change (mg/dl) 0 20 Median (Q1, Q3) Median (Q1, Q3) –20 0 –40 –20 –60 –40 –80 –60 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Baseline Lp(a) Quartile Baseline Lp(a) Quartile Patients within quartile Patients within quartile with Lp(a): with Lp(a): Increase 10.0% 11.8% 16.2% 14.6% Increase 28.1% 42.6% 45.2% 37.9% No change 59.3% 0.3% 0.3% 0.3% No change 54.0% 0.7% 0.4% 0.6% Decrease 30.7% 87.9% 83.5% 85.1% Decrease 17.9% 56.7% 54.4% 61.5% LDL-C LDL-Ccorr Lp(a) LDL-C LDL-Ccorr Lp(a) Absolute change from baseline to month 4 in low-density lipoprotein cholesterol (LDL-C), low-density lipoprotein cholesterol corrected for cholesterol content in lipoprotein(a) (LDL-Ccorr), and lipoprotein(a) levels by baseline lipoprotein(a) quartile in (A) alirocumab group and (B) placebo group. Medians and interquartile ranges of absolute changes from baseline to month 4 are shown stratified by baseline lipoprotein(a) quartile for 3 lipid parameters: LDL-C (blue), LDL-Ccorr (red), and lipoprotein(a) (green). The percentages below the x-axis summarize the percent of patients within each baseline lipoprotein(a) quartile with an increase, no change, or decrease in lipoprotein(a) from baseline to month 4. Q ¼ quartile. in LDL-C were similar in all lipoprotein(a) quartiles; was strongly correlated with the change from baseline however, accounting for the cholesterol content in to month 4 in lipoprotein(a) and was weakly corre- lipoprotein(a), the change in LDL-C corr diminished lated with the change in LDL-C corr and nonHDL-C corr slightly in the upper lipoprotein(a) quartiles, with an (Online Figure 3). There were no systematic changes overall median change of 51.1 mg/dl (IQR: 67.2 in lipoprotein(a) levels over time in the placebo group to 33.7 mg/dl) (Figure 4A). Baseline lipoprotein(a) (Figure 4B). T A B L E 1 Relationship of Changes in Lipoprotein(a) and LDL-C From Baseline to Month 4 to MACE After Month 4 in the Alirocumab Group Change HR (95% CI) per HR (95% CI) for Model Model Adjustments Parameter 1-mg/dl Decrease Observed Median Decrease p Value 1 None Lp(a) 0.998 (0.9931.002) 0.988 (0.9671.009) 0.2730 2 Baseline Lp(a) Lp(a) 0.993 (0.9890.998) 0.968 (0.9480.989) 0.0027 3A Baseline Lp(a), baseline LDL-Ccorr, change Lp(a) 0.994 (0.9900.999) 0.972 (0.9510.992) 0.0081 from baseline to month 4 in LDL-Ccorr LDL-Ccorr 0.996 (0.9940.998) 0.807 (0.7200.904) 0.0002 3B Baseline Lp(a), baseline nonHDL-Ccorr, Lp(a) 0.994 (0.9900.998) 0.972 (0.9510.992) 0.0078 change from baseline to month 4 in NonHDL-Ccorr 0.997 (0.9950.998) 0.819 (0.7340.914) 0.0004 nonHDL-Ccorr 4A Baseline Lp(a), baseline LDL-Ccorr, change Lp(a) 0.994 (0.9900.998) 0.973 (0.9530.992) 0.0071 from baseline to month 4 in LDL-Ccorr, LDL-Ccorr 0.995 (0.9930.997) 0.780 (0.6960.874) <0.0001 demographic and clinical characteristics 4B Baseline Lp(a), baseline nonHDL-Ccorr, Lp(a) 0.994 (0.9900.998) 0.973 (0.9530.992) 0.0064 change from baseline to month 4 in NonHDL-Ccorr 0.996 (0.9940.998) 0.802 (0.7170.897) 0.0001 nonHDL-Ccorr, demographic and clinical characteristics Observed median decreases for lipoprotein(a) [Lp(a)], low-density lipoprotein cholesterol corrected for lipoprotein(a) cholesterol (LDL-Ccorr), and non-high density lipoprotein cholesterol corrected for lipoprotein(a) cholesterol (non-HDL-Ccorr) were 5.0 mg/dl, 5.1 mg/dl, and 57.1 mg/dl, respectively. CI ¼ confidence interval; HR ¼ hazard ratio. JACC VOL. 75, NO. 2, 2020 Bittner et al. 141 JANUARY 21, 2020:133–44 Lipoprotein(a) and Cardiovascular Events Effect of alirocumab-induced changes in lipoprotein(a) and changes in LDL-C corr (Central Illustration A) or LDL-Ccorr on outcomes. Table 1 shows the results of the nonHDL-Ccorr (Central Illustration B) for patients in sequential Cox proportional hazards models related the 25th, 50th, and 75th percentiles of baseline lip- to the change in lipoprotein(a) on alirocumab treat- oprotein(a). Consistent with Figure 3 (right), the pre- ment to the risk of MACE, with concurrent adjust- dicted 4-year absolute risk reduction was greater for ment for LDL-C corr or nonHDL-Ccorr . The analysis higher percentiles of baseline lipoprotein(a). At the included 9,245 patients, 710 of whom had a MACE 25th percentile, lipoprotein(a) reduction accounted event. Online Table 3 shows the modeling for coro- for a small fraction of the predicted 2.29% absolute nary heart disease death and/or nonfatal myocardial reduction in MACE with alirocumab, whereas at the infarction, ischemic stroke, cardiovascular death, or 75th percentile of baseline lipoprotein(a), lip- all-cause death. In unadjusted models, no significant oprotein(a) reduction accounted for 25% of the pre- relationship was found between the change in lip- dicted 3.27% absolute reduction in risk of MACE with oprotein(a) and the risk of MACE (model 1). After alirocumab (Central Illustration A). Findings were adjustment for baseline lipoprotein(a), a significant similar when changes in nonHDL-C corr were relationship of reduction in lipoprotein(a) with a considered (Central Illustration B). Thus, among pa- lower risk of MACE was apparent (model 2). This was tients with low baseline lipoprotein(a), reduction of because higher baseline lipoprotein(a) was associated lipoprotein(a) with alirocumab contributed minimally with both greater cardiovascular risk and greater to the reduction in MACE. In contrast, among patients reduction in lipoprotein(a) on alirocumab treatment. with high baseline lipoprotein(a), reduction of lip- Therefore, accounting for the former exposed the oprotein(a) with alirocumab contributed substantially relationship of the latter to the risk of MACE. Impor- to the reduction of MACE, although the effect of tantly, further adjustment for baseline concentration reducing LDL-C corr (or nonHDL-C corr) remained pri- and change in concentration (baseline to month 4) of mary. The randomized treatment predicted 4-year either LDL-C corr or nonHDL-C corr did not attenuate absolute risk reduction after month 4, based on the relationship of change in lipoprotein(a) to risk of 18,487 patients and 1,576 events, was 2.34%, and the MACE (comparison of model 2 with models 3A and 3B, treatment HR was 0.81 (95% CI: 0.73 to 0.89). respectively). In models adjusted for LDL-C corr or nonHDL-C corr (models 3A and 3B), a 1-mg/dl DISCUSSION decrease in lipoprotein(a) was associated with HRs for MACE of 0.994 (95% CI: 0.990 to 0.999) and 0.994 Among patients with recent ACS who received (95% CI: 0.990 to 0.998), respectively. In these intensive or maximum-tolerated statin treatment, models, a 1-mg/dl decrease in LDL-Ccorr or nonHDL- baseline lipoprotein(a) levels were predictive of C corr was associated with HRs for MACE of 0.996 MACE, nonfatal myocardial infarction or coronary (95% CI: 0.994 to 0.998) and 0.997 (95% CI: 0.995 to heart disease death, and cardiovascular death, inde- 0.998), respectively. This indicated that reductions in pendent of baseline LDL-C corr. Baseline lipoprotein(a) lipoprotein(a) and LDL-C corr (or nonHDL-C corr) with level did not predict ischemic stroke or all-cause alirocumab treatment independently contributed to death. For patients in the upper 2 quartiles of base- the reduced risk of MACE. Further adjustment for line lipoprotein(a), alirocumab was a particularly demographic and clinical variables had minimal ef- efficient intervention that required treatment of 43 to fects on the relationships (models 4A and 4B). 49 patients for a median of 2.8 years to prevent 1 The magnitude of lipoprotein(a) change with alir- MACE. ocumab treatment increased with baseline lip- Alirocumab produced a median 23% reduction in oprotein(a) concentrations. For example, patients at lipoprotein(a). The absolute reduction in lip- the 25th, 50th, and 75th percentiles of the baseline oprotein(a) was directly related to the baseline con- lipoprotein(a) distribution had expected changes centration. A novel observation from this analysis in lipoprotein(a) with alirocumab treatment was that the reductions of lipoprotein(a) and of 1.6, 4.8, and 13.4 mg/dl, respectively. In turn, LDL-Ccorr (or nonHDL-Ccorr) by alirocumab were greater lipoprotein(a) reduction with alirocumab independently associated with the absolute reduction treatment was associated with greater contribution in risk of MACE. The relative contribution of lip- to the reduction in risk of MACE. The Central oprotein(a) reduction to reduced risk of MACE was Illustration shows the contributions to the predicted negligible when baseline lipoprotein(a) concentration MACE absolute risk reduction with alirocumab was low but became substantial when baseline lip- attributable to changes in lipoprotein(a) and to oprotein(a) concentration was high. Nonetheless, 142 Bittner et al. JACC VOL. 75, NO. 2, 2020 Lipoprotein(a) and Cardiovascular Events JANUARY 21, 2020:133–44 C E NT R AL IL L U STR AT IO N Relative Contributions of Changes in Concentrations of Corrected Low-Density Lipoprotein Cholesterol, Corrected Non-High-Density Lipoprotein Cholesterol, and Lipoprotein(a) to the Absolute Reduction in Major Adverse Cardiovascular Events in the Alirocumab Group A B 3.5 3.5 3.27 3.04 3.0 3.0 0.82 Predicted 4-Year Absolute Predicted 4-Year Absolute 2.54 0.81 2.5 2.5 2.34 Risk Reduction, % Risk Reduction, % 2.29 0.27 0.09 2.12 0.26 2.0 2.0 0.09 1.5 1.5 2.27 2.45 2.23 2.20 2.03 2.08 1.0 1.0 0.5 0.5 0 0 25th 50th 75th 25th 50th 75th Baseline Lp(a) Percentile Baseline Lp(a) Percentile Lp(a) Corrected LDL-C Lp(a) Corrected Non-HDL-C Bittner, V.A. et al. J Am Coll Cardiol. 2020;75(2):133–44. Based on models with adjustments for baseline levels shown in Table 1 (Models 3A and 3B), A shows the absolute contributions of reductions in lipoprotein(a) [Lp(a)] and low-density lipoprotein cholesterol corrected for the cholesterol in Lp(a) (LDL-Ccorr) to the predicted 4-year absolute reduction in major adverse cardiovascular events (MACE) at the 25th, 50th, and 75th percentile of baseline Lp(a) concentration (Model 3A), while B shows the corresponding data for reductions in non-high- density lipoprotein cholesterol corrected for the cholesterol in Lp(a) (non-HDL-Ccorr) (Model 3B). The absolute contribution of Lp(a) reduction to reduced risk of MACE was minimal when baseline Lp(a) concentration was low but was substantial when baseline Lp(a) concentration was high. The expected baseline levels at the 25th, 50th, and 75th percentiles of baseline Lp(a) are 87.3 mg/dl, 84.3 mg/dl, and 76.4 mg/dl, respectively for LDL-Ccorr and 118.2 mg/dl, 114.7 mg/dl, and 105.4 mg/dl, respectively, for non-HDL-Ccorr. The expected reductions at the 25th, 50th, and 75th percentiles of baseline Lp(a) are 1.6 mg/dl, 4.8 mg/dl, and 13.4 mg/dl, respectively, for Lp(a), 51.1 mg/dl, 50.5 mg/dl, and 48.9 mg/dl, respectively, for LDL-Ccorr, and 57.1 mg/dl, 56.2 mg/dl, and 53.9 mg/dl, respectively, for non-HDL-Ccorr. reduction of MACE remained predominantly attrib- atherosclerosis (as in ACS) than in healthy pop- utable to reduction of LDL-C corr (or nonHDL-Ccorr ) ulations. Lipoprotein(a) was purported to have a role across the range of baseline lipoprotein(a) in thrombosis and atherosclerosis (25). Both processes concentrations. are involved in the pathogenesis of ACS. Because of These novel observations added to evidence from the propensity of lipoprotein(a) to bind to fibrin in the epidemiological (1,2) and genetic (5,23,24) studies injured vascular wall (2), outcomes after ACS may be that lipoprotein(a) is an independent and causal particularly sensitive to its concentration. contributor to the risk of coronary heart disease and Our finding that lipoprotein(a) was a prognostic supported the hypothesis that interventions specif- marker in a statin-treated coronary heart disease ically aimed at reducing lipoprotein(a) have the po- population was consistent with a recent meta- tential to reduce cardiovascular risk through that analysis of 7 statin trials (26). In contrast, 2 trials mechanism. among patients with ACS found no association be- Our data indicate a greater benefit of lipoprotein(a) tween baseline lipoprotein(a) and MACE, but enrolled reduction than that estimated in a Mendelian patients with lower baseline lipoprotein(a) levels randomization analysis relating genetically deter- than those measured in the present study (3,4). mined lipoprotein(a) levels in healthy individuals to Niacin and cholesteryl ester transfer protein in- the risk of incident coronary heart disease (5,24). This hibitors reduced lipoprotein(a) by 20% to 25%; how- might be the case if lipoprotein(a) was a more ever, trials with these agents did not show reduction important risk factor in patients with advanced in MACE (10,11,27). A potential benefit of JACC VOL. 75, NO. 2, 2020 Bittner et al. 143 JANUARY 21, 2020:133–44 Lipoprotein(a) and Cardiovascular Events lipoprotein(a) reduction with niacin or cholesteryl treatment, and possibly, general adherence (i.e., to ester transfer protein inhibitors was perhaps miti- other evidence-based cardiovascular therapies and gated by other, undesirable effects of the drugs lifestyle modifications), which, in turn, might affect (27,28). The FOURIER (Further Cardiovascular Out- prognosis. However, changes in LDL-C or nonHDL-C comes Research with PCSK9 Inhibition in Subjects would be similarly reflective of adherence, and with Elevated Risk) trial (29) compared the PCSK9 adjustment for those variables should account for any inhibitor evolocumab with placebo in patients with effect of adherence in the present analysis. Finally, stable atherosclerotic cardiovascular disease and results in patients with ACS who received intensive demonstrated reduction of lipoprotein(a) and reduc- statin therapy might not be generalizable to tion in MACE similar in magnitude to the present other populations. analysis. A regression analysis of treatment group CONCLUSIONS differences in lipoprotein(a) at week 48 by baseline decile found a correlation between greater differ- There is strong evidence that elevated lipoprotein(a) ences in lipoprotein(a) and risk of coronary events contributes to the incidence of coronary heart dis- after adjustment for LDL-C. Our findings extend those ease, but no treatment has yet been proven to reduce of the FOURIER trial by demonstrating, for the first coronary risk through a reduction in lipoprotein(a). time, that patient-level pharmacological lowering of The ODYSSEY OUTCOMES trial was not designed lipoprotein(a) relatively early after the initiation of specifically to enroll and treat patients with high therapy was associated with reduced risk of subse- lipoprotein(a). However, our observations suggest quent MACE, independent of concurrent reductions that reduction of lipoprotein(a) contributed to the of LDL-C corr or nonHDL-C corr . reduction of cardiovascular risk with alirocumab The predicted MACE 4-year absolute risk re- therapy, independent of the concurrent reduction of ductions corresponding to joint changes in lip- other atherogenic lipoproteins. Therefore, lip- oprotein(a) and LDL-C corr (or nonHDL-Ccorr) with oprotein(a) is both a prognostic factor and a poten- alirocumab shown in the Central Illustration varied tially important independent treatment target among around the randomized treatment risk reduction of patients with recent ACS. 2.34%. Although this is >2.0%, the previously re- ported risk reduction in the ODYSSEY OUTCOMES ACKNOWLEDGMENTS The authors thank the pa- trial (16), numerical correspondence of these risk re- tients, study coordinators, and investigators who ductions were not necessarily expected, because the participated in this trial. Sophie Rushton-Smith, Ph.D. analysis in Table 1 and the Central Illustration (MedLink Healthcare Communications, London) pro- considered MACE beginning after the month 4 vided editorial assistance in the preparation of the assessment for each patient, whereas the overall manuscript (limited to editing for style, referencing, analysis of the trial considered MACE beginning at and figure and table editing) and was funded by randomization. Alirocumab had no apparent effect on Fondation Assistance Publique  Hôpitaux de Paris, MACE through month 4; therefore, effects after Paris, France. month 4 were greater than the overall effects during the trial. ADDRESS FOR CORRESPONDENCE: Dr. Vera Bittner, University of Alabama at Birmingham, 701 19th Street STUDY LIMITATIONS. The cholesterol content in lip- South - LHRB 310, Birmingham, Alabama 35294. oprotein(a) particles is variable. Correction of LDL-C E-mail: vbittner@uab.edu. Twitter: @gabrielsteg, or nonHDL-C by 30% of lipoprotein(a) mass is thus @DLBHATTMD, @VeraBittner2. an approximation of the contribution from choles- terol in lipoprotein(a). Lipoprotein(a) mass, as PERSPECTIVES measured in this study, correlates imperfectly with molar concentration of lipoprotein(a) because mass is COMPETENCY IN MEDICAL KNOWLEDGE: Baseline levels influenced by apolipoprotein(a) isoform size. At high of lipoprotein(a) and LDL-C and reductions by alirocumab are lipoprotein(a) mass, molar concentration is under- associated with the risk of MACE in patients after recent ACS. estimated, and vice versa (30). However, to the extent such effects were present, they would have biased TRANSLATIONAL OUTLOOK: Further studies are needed to our study toward the null. Furthermore, the magni- determine whether lipoprotein(a) is an important treatment tude of lipoprotein(a) lowering by alirocumab is not target after recent ACS. affected by apolipoprotein(a) size (28,31). Changes in lipoprotein(a) might reflect adherence to study 144 Bittner et al. JACC VOL. 75, NO. 2, 2020 Lipoprotein(a) and Cardiovascular Events JANUARY 21, 2020:133–44 REFERENCES 1. Tsimikas S. A test in context: lipoprotein(a): 12. Sabatine MS, Giugliano RP, Wiviott SD, et al. mendelian randomisation analysis. Lancet Dia- diagnosis, prognosis, controversies, and emerging Efficacy and safety of evolocumab in reducing betes Endocrinol 2017;5:524–33. therapies. J Am Coll Cardiol 2017;69:692–711. lipids and cardiovascular events. N Engl J Med 24. Lamina C, Kronenberg F, Lp(a)-GWAS-Con- 2015;372:1500–9. 2. Nordestgaard BG, Langsted A. Lipoprotein (a) sortium. Estimation of the required lipoprotein(a)- as a cause of cardiovascular disease: insights from 13. Robinson JG, Farnier M, Krempf M, et al. Effi- lowering therapeutic effect size for reduction in epidemiology, genetics, and biology. J Lipid Res cacy and safety of alirocumab in reducing lipids coronary heart disease outcomes: a Mendelian 2016;57:1953–75. and cardiovascular events. N Engl J Med 2015;372: randomization analysis. JAMA Cardiol 2019;4: 1489–99. 575–9. 3. Schwartz GG, Ballantyne CM, Barter PJ, et al. Association of lipoprotein(a) with risk of recurrent 14. Sabatine MS, Giugliano RP, Keech AC, et al. 25. Thanassoulis G. Using genetics to plan future ischemic events following acute coronary syn- Evolocumab and clinical outcomes in patients with randomized trials of lipoprotein(a) lowering-how drome: analysis of the dal-Outcomes randomized cardiovascular disease. N Engl J Med 2017;376: much reduction, for how long, and in whom? clinical trial. JAMA Cardiol 2018;3:164–8. 1713–22. JAMA Cardiol 2019;4:513–4. 4. O’Donoghue ML, Morrow DA, Tsimikas S, et al. 15. Ridker PM, Revkin J, Amarenco P, et al. Cardio- 26. Willeit P, Ridker PM, Nestel PJ, et al. Baseline Lipoprotein(a) for risk assessment in patients with vascular efficacy and safety of bococizumab in high- and on-statin treatment lipoprotein(a) levels for established coronary artery disease. J Am Coll risk patients. N Engl J Med 2017;376:1527–39. prediction of cardiovascular events: individual Cardiol 2014;63:520–7. patient-data meta-analysis of statin outcome tri- 16. Schwartz GG, Steg PG, Szarek M, et al. Alir- als. Lancet 2018;392:1311–20. 5. Burgess S, Ference BA, Staley JR, et al. Asso- ocumab and cardiovascular outcomes after acute ciation of LPA variants with risk of coronary coronary syndrome. N Engl J Med 2018;379: 27. Tall AR, Rader DJ. Trials and tribulations of disease and the implications for lipoprotein(a)- 2097–107. CETP inhibitors. Circ Res 2018;122:106–12. lowering therapies: a Mendelian randomization 17. Schwartz GG, Bessac L, Berdan LG, et al. Effect 28. Parish S, Hopewell JC, Hill MR, et al. Impact of analysis. JAMA Cardiol 2018;3:619–27. of alirocumab, a monoclonal antibody to PCSK9, apolipoprotein(a) isoform size on lipoprotein(a) on long-term cardiovascular outcomes following lowering in the HPS2-THRIVE study. Circ Genom 6. Nordestgaard BG, Chapman MJ, Ray K, et al. acute coronary syndromes: rationale and design of Precis Med 2018;11:e001696. Lipoprotein(a) as a cardiovascular risk factor: cur- rent status. Eur Heart J 2010;31:2844–53. the ODYSSEY outcomes trial. Am Heart J 2014; 29. O’Donoghue ML, Fazio S, Giugliano RP, et al. 168:682–9. Lipoprotein(a), PCSK9 inhibition, and cardiovas- 7. Grundy SM, Stone NJ, Bailey AL, et al. 2018 18. Gaudet D, Watts GF, Robinson JG, et al. Effect cular risk. Circulation 2019;139:1483–92. AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/ APhA/ASPC/NLA/PCNA guideline on the man- of alirocumab on lipoprotein(a) over $1.5 years 30. Tsimikas S, Fazio S, Ferdinand KC, et al. NHLBI agement of blood cholesterol: a report of the (from the phase 3 ODYSSEY program). Am J Car- working group recommendations to reduce American College of Cardiology/American Heart diol 2017;119:40–6. lipoprotein(a)-mediated risk of cardiovascular Association Task Force on Clinical Practice 19. Friedewald WT, Levy RI, Fredrickson DS. Esti- disease and aortic stenosis. J Am Coll Cardiol Guidelines. J Am Coll Cardiol 2019;73:e285–350. mation of the concentration of low-density lipo- 2018;71:177–92. 8. Roeseler E, Julius U, Heigl F, et al. Lipopro- protein cholesterol in plasma, without use of the 31. Enkhmaa B, Anuurad E, Zhang W, Yue K, Li CS, tein apheresis for lipoprotein(a)-associated car- preparative ultracentrifuge. Clin Chem 1972;18: Berglund L. The roles of apo(a) size, phenotype, diovascular disease: prospective 5 years of 499–502. and dominance pattern in PCSK9-inhibition- follow-up and apolipoprotein(a) characteriza- 20. Kinpara K, Okada H, Yoneyama A, Okubo M, induced reduction in Lp(a) with alirocumab. J Lipid tion. Arterioscler Thromb Vasc Biol 2016;36: Murase T. Lipoprotein(a)-cholesterol: a significant Res 2017;58:2008–16. 2019–27. component of serum cholesterol. Clin Chim Acta 9. HPS2-THRIVE Collaborative Group, Landray MJ, 2011;412:1783–7. KEY WORDS acute coronary syndromes, Haynes R, et al. Effects of extended-release niacin 21. Yeang C, Witztum JL, Tsimikas S. ’LDL- alirocumab, low-density lipoprotein with laropiprant in high-risk patients. N Engl J C’ ¼ LDL-C þ Lp(a)-C: implications of achieved cholesterol, major adverse Med 2014;371:203–12. ultra-low LDL-C levels in the proprotein con- cardiovascular events, proprotein convertase vertase subtilisin/kexin type 9 era of potent LDL-C subtilisin/kexin type 9 inhibition 10. AIM-HIGH Investigators, Boden WE, lowering. Curr Opin Lipidol 2015;26:169–78. Probstfield JL, et al. Niacin in patients with low HDL cholesterol levels receiving intensive statin 22. Gail M, Simon R. Testing for qualitative in- A PPE NDI X For a complete list of the therapy. N Engl J Med 2011;365:2255–67. teractions between treatment effects and patient ODYSSEY OUTCOMES committee members, subsets. Biometrics 1985;41:361–72. 11. HPS3/TIMI55–REVEAL Collaborative Group, investigators, and contributors, an expanded Bowman L, Hopewell JC, et al. Effects of anace- 23. Saleheen D, Haycock PC, Zhao W, et al. Apo- Methods section, and supplemental figures trapib in patients with atherosclerotic vascular lipoprotein(a) isoform size, lipoprotein(a) con- and tables, please see the online version of disease. N Engl J Med 2017;377:1217–27. centration, and coronary artery disease: a this paper.