Reductions in Atherogenic Lipids and Major Cardiovascular Events: A Pooled Analysis of 10 ODYSSEY Trials Comparing Alirocumab to Control

Transcription

1 /CIRCULATIONAHA Reductions in Atherogenic Lipids and Major Cardiovascular Events: A Pooled Analysis of 10 ODYSSEY Trials Comparing Alirocumab to Control Running Title: Ray et al., Lipid reductions + CV events in ODYSSEY Kausik K Ray, MD, MPhil 1 ; Henry N Ginsberg, MD 2 ; Michael H Davidson, MD 3 ; Downloaded from by guest on April 8, 2018 Robert Pordy, MD 4 ; Laurence Bessac, MD 5 ; Pascal Minini, PhD 6 ; Robert H Eckel, MD 7 ; Christopher P Cannon, MD 8 1 Imperial Centre for Cardiovascular Disease Prevention, Department of Primary Care and Public Health, Imperial College, London, UK; 2 Columbia University, New York, NY; 3 Department of Medicine, University of Chicago Medicine, Chicago, IL; 4 Regeneron Pharmaceuticals, al Tarrytown, NY; 5 Sanofi, Paris, France; 6 Biostatistics stics and Programming, Sanofi, Chilly-Mazarin, France; 7 University of Colorado, Anschutz Medical Campus, Aurora, CO; 8 Harvard Clinical Research Institute, Boston, MA Address for Correspondence: Kausik K Ray, MD, MPhil Imperial Centre for Cardiovascular Disease Prevention Department of Primary Care and Public Health School of Public Health, Imperial College Reynolds Building, St Dunstan s Road London, W68RP, UK Tel.: 44 (0) Fax.: +44 (0) k.ray@imperial.ac.uk Journal Subject Terms: Coronary Artery Disease; Atherosclerosis; Vascular Disease; Pharmacology; Lipids and Cholesterol 1

2 /CIRCULATIONAHA Abstract Downloaded from by guest on April 8, 2018 Background A continuous relationship between reductions in low-density lipoprotein cholesterol (LDL-C) and major adverse cardiovascular events (MACE) has been observed in statin and ezetimibe outcomes trials, down to achieved levels of 54 mg/dl. However, it is uncertain whether this relationship extends to LDL-C levels <50 mg/dl. We assessed the relationship between additional LDL-C, non-high-density lipoprotein cholesterol (non-hdl-c), and apolipoprotein B100 (apob) reductions and MACE among patients within the ODYSSEY trials that compared alirocumab versus controls (placebo/ezetimibe), mainly as add on to maximally tolerated statin. Methods Data were pooled from 10 double-blind trials (6699 patient-years follow-up). Randomization was to alirocumab 75/150 mg every 2 weeks or control for weeks, added to background statin therapy in 8 trials. This analysis included 4974 patients (3182 alirocumab, 1174 placebo, 618 ezetimibe). In a post hoc analysis, the relationship between average on treatment lipid levels and percent reductions in lipids from baseline were correlated with MACE (coronary heart disease death, non-fatal myocardial infarction [MI], ischemic stroke, or unstable angina requiring hospitalization) using multivariable a analyses. Results Overall, 33.1% of the pooled cohort achieved average LDL-C <50 mg/dl ( % allocated to alirocumab, 6.5% allocated to ezetimibe, and 0% allocated to placebo). In total, 104 patients experienced MACE (median time to event: 36 weeks). For every 39 mg/dl lower achieved LDL-C, the risk of MACE appeared to be 24% lower (adjusted hazard ratio [HR] 0.76, 95% confidence c interval [CI]: ; P=0.0025). 5. Percent reductions in LDL-C from baseline were inversely correlated r with MACE rates (HR 0.71 [0.57 to 0.89] per additional a 50% reduction from baseline; P=0.003). 03). Materially ly similar strengths of association to those described d for LDL-C were observed with achieved ed non-hdl-c and apob levels ls or percentage reductions. Conclusions In a post hoc analysis alys from 10 ODYSSEY trials greater percentage reductions in LDL-C and lower on-treatment tment LDL-C were associated with a lower incidence iden of MACE, including very low levels of LDL-C (<50 mg/dl). These findings require further validation in the ongoing prospective ODYSSEY OUTCOMES trial. Clinical Trial Registration: clinicaltrials.gov identifiers: NCT , NCT , NCT , NCT , NCT , NCT , NCT , NCT , NCT , NCT Key-words: PCSK9; low-density lipoprotein cholesterol; apolipoprotein; cardiovascular events; MACE, ASCVD, Risk reduction, Alirocumab, ODYSSEY 2

3 /CIRCULATIONAHA Clinical Perspectives: What s new? Cardiovascular benefits of statins and add-on lipid-lowering therapy only extend to LDL- C levels of ~54 mg/dl. We investigated whether this relationship extends below 50 mg/dl using data from ODYSSEY trials of alirocumab (PCSK9 monoclonal antibody) versus placebo/ezetimibe. Downloaded from by guest on April 8, 2018 About half of alirocumab-treated patients achieved LDL-C <50 mg/dl. For each 39 mg/dl lower achieved LDL-C, MACE incidence fell by 24%, and 50% reductions in LDL-C from baseline reduced MACE by 29%. Similar associations were observed with non-high-density lipoprotein cholesterol and apolipoprotein B achieved levels and percentage reductions and MACE. Low LDL-C levels were not associated with excess treatment emergent adverse events. What are the clinical implications? These analyses provide further reassurance about the safety and cardiovascular benefit of achieving even further reductions in LDL-C with alirocumab beyond what was previously achieved with statins and ezetimibe. Limitations include the low number of events (104), the limited duration of the studies ( weeks) and the post hoc nature of this analysis. If the forthcoming outcomes trials of PCSK9 inhibitors such as alirocumab demonstrate additional reduction in MACE with further LDL-C reduction, then guideline committees may investigate lower LDL-C goals or a larger reduction in LDL-C from untreated baseline for those at highest risk of MACE. 3

4 /CIRCULATIONAHA Introduction Statins have been the cornerstone of global lipid modification guidelines for the prevention of major adverse cardiovascular events (MACE).1 With accumulating evidence, each subsequent iteration of clinical guidelines have extended statin indications and have recommended the achievement of progressively lower low-density lipoprotein cholesterol (LDL-C) goals2-4 or most recently percentage reductions in LDL-C using the most potent statins.5 Recently, the IMPROVE IT trial demonstrated that further reductions in LDL-C to a mean of 54 mg/dl with a Downloaded from by guest on April 8, 2018 non-statin lipid-lowering therapy (LLT) in statin-treated individuals provided additional reductions in MACE,6 supporting the concept that lower LDL-C levels are better, and consistent with observational data from several statin trials.7-9 Inhibitors of proprotein convertase subtilisin/kexin type 9 (PCSK9) offer the prospect of achieving even lower LDL-C levels than ezetimibe when added to statins. Data from the phase 3 ODYSSEY SEY trial program, ra in which average achieved LDL-C levels els of <50 mg/dl were e observed ed when alirocumab was added to statin therapy,10-12 allows lows for the exploration of changes in event rates at much lower LDL-C levels than previously analyzed. In addition, percent reduction in LDL-C has also been cited as being important in some guidelines 5, 13, 14 and as patients in the ODYSSEY trials have already had their LDL-C reduced by 30 50% with statins, these trials allow us to explore the relationship between further percentage reductions in LDL-C and risk. Therefore, we assessed whether the relationship between LDL-C and risk extends to very low levels of LDL-C (<50 mg/dl) in the ODYSSEY trial program. We hypothesized that, among patients already receiving maximally tolerated statin therapy, both lower achieved LDL-C and greater percentage reductions from baseline would translate into lower rates of MACE. Furthermore, as many guidelines increasingly recommend 4

5 /CIRCULATIONAHA non-high-density lipoprotein cholesterol (non-hdl-c) or apolipoprotein B100 (apob) as potential alternatives to LDL-C for assessing the efficacy of LLT, we provide similar analyses using these 4, 15, 16 lipid parameters for comparison. Methods Patient population The Phase 3 ODYSSEY trial designs have been reported previously. 6, 11, 12, Briefly, patients Downloaded from by guest on April 8, 2018 were enrolled if they had established atherosclerotic cardiovascular disease (ASCVD) or high CV risk such as heterozygous familial hypercholesterolemia (HeFH) with LDL-C inadequately controlled on their existing treatment (statin/other LLT/diet). The main exclusion criteria ia were baseline LDL-C <70 mg/dl for those with ASCVD and very high risk or <100 mg/dl for high risk patients without ASCVD at screening. Individuals with triglyceride levels >400 mg/dl were excluded (for further details on inclusion and exclusion criteria r see etable1, online ne Supplemental plem e Material). For the present analysis a data were pooled from 10 Phase 3 ODYSSEY SEY trials (efigure 1). Patients were randomized to receive alirocumab or control (placebo or ezetimibe) with doubleblind treatment periods of weeks. Six of the 10 studies, representing ~80% of the population, had a minimum study duration of 52 weeks. All study protocols were approved by the relevant local independent review boards, and all participating patients provided written, informed consent. Lipid measurements LDL-C was calculated using the Friedewald equation unless triglycerides were >400 mg/dl when it was determined by beta quantification. ApoB levels in serum were determined using immunonephelometry by a central laboratory (Medpace Reference Laboratories, Cincinnati, OH, 5

6 /CIRCULATIONAHA US, and Leuven, Belgium, except for the LONG TERM study, which used Covance Central Laboratories, Indianapolis, IN, USA, and Geneva, Switzerland) and non-hdl-c via subtraction of HDL-C from total cholesterol. MACE definitions MACE was defined as per the primary endpoint of the ODYSSEY OUTCOMES study: 23 coronary heart disease (CHD) death, non-fatal myocardial infarction [MI], ischemic stroke, or unstable angina (UA) requiring hospitalization. Cardiovascular events were adjudicated by a Downloaded from by guest on April 8, 2018 central Clinical Events Committee 12 (the same Committee is involved in the ODYSSEY OUTCOMES trial 23 ). UA cases considered here were limited to those with definite evidence of the ischemic condition, i.e., small proportion of UAs qualified (see definition of UA in the Supplemental Material). Statistical Analysis Baseline characteristics c ti cs and distribution tion of lipid parameters Baseline data were pooled for all randomized patients and presented stratified according to whether the studies were placebo-controlled or ezetimibe-controlled. For continuous variables, the data are reported as means and standard deviations or median and interquartile range if they were not normally distributed. The distribution of LDL-C, non-hdl-c and apob levels at baseline and average on treatment levels or average percentage reductions in these parameters during the study treatment period are depicted graphically and analyzed using descriptive statistics comparing treatment groups (alirocumab versus placebo, or versus ezetimibe). These include only patients in the safety population i.e. patients randomized and who received at least 1 dose or part of a dose of study treatment. 6

7 /CIRCULATIONAHA Lipid changes and risk of MACE Irrespective of treatment allocation patients were pooled into one overall cohort and the relationship between LDL-C and MACE during the treatment period was assessed using 1) achieved LDL-C levels during treatment and 2) percentage reductions in LDL-C from baseline. Average on-treatment LDL-C or the mean percentage reduction during the treatment period were determined from the area under the curve (using trapezoidal method), taking into account all Downloaded from by guest on April 8, 2018 LDL-C values up to end of treatment period or occurrence of MACE, whichever came first. The relationship between on treatment LDL-C and MACE was assessed using a multivariable Cox regression model with adjustment for age, gender, diabetes, prior history of MI or stroke, baseline LDL-C and smoking status as previously published. 24 Hazard ratios (HR) and 95% confidence intervals (CI) were calculated for every 39 mg/dl lower LDL-C to provide a comparison with the Cholesterol Treatment Trialist ist (CTT) meta-regression s line. 25 Similar ilar analyses alys were conducted for the percentage reduction in LDL-C from baseline and subsequent risk of MACE with HR and 95% CI expressed for each 50% reduction in LDL-C. To assess the shape of association, the adjusted rates of MACE and associated 95% CI were determined from a multivariate Poisson model and depicted graphically as a function of average LDL-C levels or average percentage reduction during treatment. Instead of using all available lipid measurements (average LDL-C levels or reductions) sensitivity analyses were conducted using only LDL-C values and percentage reduction at Week 4 and subsequent events after excluding events that occurred prior to Week 4. Finally, similar analyses to those described above were conducted for non-hdl-c or apob. All analyses were generated using SAS version 9.4 and all tests and confidence intervals were two-sided. 7

8 /CIRCULATIONAHA Safety analysis For safety, treatment-emergent adverse events (TEAEs) were defined as those events occurring from the first dose of study treatment and up to 70 days after the last dose. The principal analyses compared randomized treatment to alirocumab versus control. We also explored the strength of association between any TEAE and LDL-C levels or percentage reductions in LDL-C using multivariable logistic regression, after adjustment for the same covariates included in the MACE analyses. Data are reported as odds ratio (OR) and 95% CI per 39 mg/dl difference or 50% Downloaded from by guest on April 8, 2018 reduction in LDL-C. The shape of the association was also depicted graphically using adjusted TEAE rate and associated 95% CI plotted against average LDL-C levels or percentage LDL-C reductions derived from multivariate logistic regression models. Results Baseline characteristics c ti cs The baseline characteristics of individual studies are shown in etable2 in the online ne Supplemental material, and the pooled summary of placebo and ezetimibe comparator trials are shown in Table 1. In the placebo-controlled trials, 2318 patients were treated with alirocumab and 1174 were treated with placebo; in the ezetimibe-controlled trials, 864 were treated with alirocumab and 618 were treated with ezetimibe. Hence, a total of 4974 patients were included in the lipid and MACE analyses described below. Overall the average age was ~60 years, with patients being mostly white and having an average body mass index of ~30 kg/m 2. Approximately two thirds were male, one third had diabetes, two thirds had a prior history of ASCVD, and about one-fifth were smokers. One third of participants in the placebo-controlled trials had HeFH. 8

9 /CIRCULATIONAHA Baseline lipids Baseline lipid values for individual studies are reported in etable 3 in the Supplement. Pooled mean baseline LDL-C levels ranged from to mg/dl, non HDL-C between to mg/dl, and apob between to mg/dl (Table 1). As expected, the distribution of baseline lipids at randomization was fairly similar between alirocumab and the control groups (efigure 2 in the Supplement). Lipid levels during treatment Downloaded from by guest on April 8, % for alirocumab and 6.5% for ezetimibe arms, respectively. Thus when all patients were pooled, the overall distribution of each lipid parameter during treatment largely reflected the greater proportion of patients achieving very low levels of LDL-C, non-hdl-c and apob in the alirocumab group (efigure 3). Percent reductions in lipids from baseline Figure 2 depicts the distribution of the average percentage change from baseline in LDL-C, non- Figure 1 illustrates the distribution of different lipid parameters during treatment. The mean LDL- C levels achieved during treatment were: in placebo-controlled studies, 56.9 and mg/dl among those treated with alirocumab and placebo, and in ezetimibe-controlled studies, 64.0 and mg/dl with alirocumab and ezetimibe, respectively (Table 2; individual trial dataa in etable 4). Corresponding values for non-hdl-c and apob shown in Table 2. Similar results were obtained using Week 4 achieved ed lipids i instead of average levels throughout h ut the entire duration of the study (etable 5). Overall, 33.1% 3 of patients achieved an average LDL-C <50 mg/dl during treatment. In placebo-controlled studies, 52.6% of alirocumab and 0% of placebo treated patients achieved LDL-C levels <50 mg/dl. In the ezetimibe-controlled studies, the corresponding figures were 9

10 /CIRCULATIONAHA HDL-C, and apob during the trials. The average percentage change in LDL-C from baseline during treatment was -55.4% for alirocumab and +2.7% for placebo; and -48.1% with alirocumab and -18.0% with ezetimibe, respectively, in placebo- and ezetimibe-controlled trials (Table 2, individual trial data in etable 4). Corresponding results for non-hdl-c and apob are shown in Table 2 and etable 4. When Week 4 percentage reductions were assessed rather than the average reductions over the course of the trial similar values were observed (etable 5). The combined distribution plot (efigure 3) mainly reflected the greater reductions in lipids achieved with Downloaded from by guest on April 8, 2018 alirocumab versus the relatively modest reductions observed with ezetimibe and with placebotreated patients largely remaining unchanged from baseline. On treatment lipid levels and MACE A total of 104 first MACE were reported: 20 CHD deaths, 64 non-fatal MI, 16 ischemic strokes, and 4 unstable angina episodes occurred (median time to event: 36 weeks), among 4974 patients treated ted during a total of 6699 patient-years a of follow-up. low-up. A lower risk of MACE was observed ed with lower achieved ed LDL-C levels (Figure 3; adjusted HR 0.76, 95% CI: per 39 mg/dl lower achieved LDL-C; P=0.0025; Table 2). Similar il results were obtained using a single Week 4 LDL-C measurement instead of average levels throughout the trial (etable 5). In pairwise comparisons of LDL-C, non-hdl-c, and apob, there was a strong and significant correlation between levels of each lipid parameter (all correlation coefficients >0.9; P<0.0001; etable 6). As with achieved average LDL-C, lower (average) achieved non-hdl-c and apob levels were associated with a lower risk of MACE (Figure 3). A 39 mg/dl difference in LDL-C corresponds to a 42 mg/dl difference in non-hdl-c and 27 mg/dl difference in apob in the present pooled datasets. For each 42 mg/dl lower non-hdl-c the HR was 0.77 (CI

11 /CIRCULATIONAHA ; P=0.0056; Table 2). The corresponding HR for each 27 mg/dl lower apob was 0.72 (CI ; P=0.0002; Table 2). Percentage reductions in lipids and MACE LDL-C percent reduction was inversely correlated with MACE rates (Figure 3; HR 0.71 [ ] per additional 50% reduction in LDL-C; P=0.003; Table 2). Similarly the risk of MACE was also lower with greater percent reductions from baseline in both non-hdl-c (Figure 3, HR 0.71 [ ] per 50% reduction; P=0.0323) and apob (Figure 3, HR 0.68 [ ] per Downloaded from by guest on April 8, % reduction; P=0.0008; Table 2). Qualitatively similar results were observed using a single Week 4 measurement of LDL-C or non-hdl-c and Week 12 apob instead of using average values throughout the trial (etable 5) Safety Overall incidence of TEAEs, serious TEAEs, deaths, and discontinuations due to TEAEs were similar between alirocumab ab and control ol patients within the pools of placebo- and ezetimibe- etimi controlled studies (etable 7). A higher rate of injection site reactions, mostly mild in intensity and self-limiting, were observed with alirocumab compared with controls. Analyses comparing the relationship between a 39 mg/dl lower LDL-C and odds of any TEAE were not significant (OR 1.02, 95% CI: , efigure 4). Nor was there any significant association between a 50% lowering of LDL-C and odds of any TEAE (OR 1.02, 95% CI: ; efigure 4). Results Currently all global guidelines for ASCVD risk reduction focus on optimization of statin therapy as the first option for reducing LDL-C for those at high risk.4, 5, 13 The therapeutic limits of statins and the clinical scenarios in which they have been tested have therefore established the 11

12 /CIRCULATIONAHA boundaries of contemporary clinical guidelines and the recommendations they have set, whether that be a LDL-C goal or a percentage reduction in LDL-C. Based on randomized clinical trial data of intensive versus standard statin therapy7, 8, 25 knowledge of the distribution of LDL-C levels in general populations26, 27 and what it is achievable on average with the most potent statins, guidelines such as the updated Adult Treatment Panel III and those from the European Society of Cardiology/European Atherosclerosis Society recommended pragmatic goals for LDL-C of <70 mg/dl for those at highest ASCVD risk.2, 4 Downloaded from by guest on April 8, 2018 More recently, the ACC/AHA guidelines and the UK NICE guidelines have argued that most statin trials have tested whether a certain percentage reduction in LDL-C translates into a reduction in clinical events rather than a LDL-C goal, and thus recommended LDL-C L-C reductions of at least 30 50% for those at elevated risk, again based on what is achievable using the most potent statins.5, 13 Until recently, it was uncertain based on trials data whether consistently achieving LDL-C levels ls <70 mg/dl, or achieving ing further percentage reductions in LDL-C after maximizing m statins, ti would translate into a lower risk of MACE. The IMPROVE IT trial extended our evidence-base beyond statins, demonstrating that LDL-C levels on average 54 mg/dl with ezetimibe plus statins further reduced MACE as compared with the achievement of an LDL-C of ~70 mg/dl with statins alone.6 The risk reduction observed in the IMPROVE IT trial was also entirely consistent with the absolute reduction in LDL-C that would have been predicted by the CTT statin derived regression line, supporting the notion that LDL-C reduction by statins and ezetimibe confer similar benefits and that the real determinant of the relative risk reduction is the magnitude of the change in LDL-C rather than the mechanism by which this is achieved. If the findings of IMPROVE-IT are considered as a further percentage reduction in LDL-C, then they support the notion that starting 12

13 /CIRCULATIONAHA with a baseline LDL-C of ~70 mg/dl, a further 20% reduction in LDL-C translates into a 6 7% lower risk of MACE. However, the question remains as to whether the additional percentage reduction in LDL-C (of 50 60%) and even lower on-treatment LDL-C level (<50 mg/dl) that can be achieved by adding a PCSK9 inhibitor to a statin will be associated with a lower risk of MACE. The present analysis reports data from 10 randomized trials in the ODYSSEY trial program, providing information on 6699 patient-years of exposure and suggests that there is Downloaded from by guest on April 8, 2018 continuous relationship between average on-treatment LDL-C and MACE with no evidence of discernable attenuation even at low achieved levels of LDL-C (<50 mg/dl). Furthermore, for every additional 39 mg/dl lower LDL-C achieved with either alirocumab or ezetimibe (on top of maximally tolerated statins in most patients), there was a further 24% lower risk of MACE (HR 0.76, 95% CI: ). This is remarkably similar to the CTT point estimate of a 22% risk reduction (rate ratio 0.78, 95% CI ) for every 39 mg/dl reduction in LDL-C achieved e with statins.25 Similarly, we observed an inverse relationship with additional d percentage reductions in LDL-C and MACE, without evidence of attenuation of benefit for greater percentage reductions in LDL-C. In multivariable regression analyses, each 50% incremental reduction in LDL-C on top of statins was associated with a further 29% reduction in the risk of MACE (HR 0.71, 95% CI 0.57 to 0.89). There was a significant correlation between LDL-C and non-hdl-c, between LDL-C and apob and between non-hdl-c and apob (all P<0.0001). As with LDL-C, we observed a continuous relationship between lower achieved levels of both non-hdl-c and apob with lower rates of MACE (HR 0.77 [95% CI ] per 42 mg/dl difference and HR 0.72 [ ] 13

14 /CIRCULATIONAHA per 27 mg/dl difference, respectively), with no discernable evidence of attenuation at lower levels. Furthermore, greater percentage reductions in non-hdl-c and apob were also both associated with a lower risk of MACE with no evidence of attenuation of benefit. Our findings are consistent with epidemiological studies in statin naïve populations which have suggested a continuous relationship between LDL-C, non-hdl-c and apob and risk28 with no apparent attenuation of the relationship as well as the data from the statin trials where a continuous relationship between on-treatment LDL-C, non-hdl-c and MACE29 have been Downloaded from by guest on April 8, 2018 observed without evidence of a threshold. Although LDL-C continues to be the main target of lipid lowering strategies, levels of non-hdl-c and apob have been shown to more closely correlate with risk of MACE, as these parameters more accurately reflect the actual al number of circulating atherogenic particles or their cholesterol content, particularly in patients with elevated TGs that likely have elevations in non-ldl atherogenic particles.30 Our findings showed materially similar ilar benefit efit with reductions in each parameter in part due to the co-linearity ity of the parameters and the relatively small number of events. ents. Prior work by Robinson et al.31, 32 have demonstrated that the benefit of LLT is also related to the percentage reduction in LDL-C and non-hdl-c with consistent benefits between statins. While high-intensity statins have offered us the scope of a 50% reduction in LDL-C, the addition of a PCSK9 inhibitor to statin therapy offers us a further 50 60% reduction in these parameters on top of statins, or ~75 80% total reduction from the patient s untreated baseline LDL-C level. Consistent with this rationale is the recent 2016 ACC pathway for the addition of non-statin therapies for those individuals with high absolute risk and high LDL-C levels despite maximally tolerated statin therapy.14 Our observation that high-risk patients with LDL-C levels 14

15 /CIRCULATIONAHA between mg/dl, despite maximally tolerated statin, derive benefit from a further 50% reduction in LDL-C or lower achieved absolute levels lends support to the consensus statement. As we approach the possibility of achieving lower LDL-C levels consistently with PCSK9 inhibition, concerns have been raised about the potential safety of achieving very low levels of LDL-C (e.g. <50 mg/dl). The present Phase 3 clinical trial analyses provide further data regarding the overall safety of alirocumab and lower LDL-C levels not being associated with an increase in total adverse events (AEs). These findings add to earlier observations from high- Downloaded from by guest on April 8, 2018 intensity statin trials which have also failed to demonstrate any relationship adverse events and lower achieved LDL-C levels.24, 33, 34 Limitations The limitations of the present analysis merit consideration. Importantly, although adverse consequences of very low LDL-C were not identified in these trials, the long-term effects of very low levels of LDL-C L- induced d by PCSK9 inhibitors itors are unknown. n. Moreover, er, while these data are derived d from randomized controlled trials, the analyses are observational in nature and derived from a relatively ely small number of events and as such we cannot therefore exclude the potential for confounding as an explanation for the observed associations. We have attempted to take these into account via statistical adjustment and conducting sensitivity analyses using alternative methodology which have produced materially similar findings. Furthermore, the 10 studies pooled differed most notably in prevalence of HeFH, diabetes, age, prior history of MI or stroke and baseline LDL-C. Studies derived principally from HeFH patients tended to be a decade younger, have fewer patients with diabetes and higher baseline LDL-C levels. Similarly, in trials without background statin therapy baseline LDL-C was higher. Therefore, a potential critique of the analyses of achieved LDL-C and MACE could be that patients that achieve lower LDL-C had 15

16 /CIRCULATIONAHA lower baseline LDL-C and lower risk than patients at higher baseline LDL-C levels. However, we controlled for baseline LDL-C in all analyses and thus we do not believe that baseline LDL-C levels accounted for our findings. Furthermore, we also assessed percentage reductions in LDL-C which provided similar quantitative findings. Further reassurance of our methodology arises from the magnitude of the association observed between a 39 mg/dl difference in LDL-C and MACE, which is similar to the point estimate derived from the CTT meta-regression line and lies within its confidence intervals. For the safety analysis, we only looked at overall TEAE rates to Downloaded from by guest on April 8, 2018 maximize power, and more sophisticated analyses looking at specific TEAEs would be more informative in larger trials with longer follow up than the present pooled data. In summary, these analyses provide further reassurance about the safety and cardiovascular benefit of achieving even further reductions in LDL-C, non-hdl-c and apob beyond what was previously achieved with statins alone. The results of the large cardiovascular outcomes studies with PCSK9 inhibitors such as ODYSSEY OUTCOMES are assessing s ssin whether her PCSK9 inhibition with alirocumab on top of maximally ally ly tolerated statin therapy reduces MACE. If these trials demonstrate effectiveness of further LDL-C reduction, then guideline committees may investigate lower targets or a larger reduction in LDL-C from untreated baseline for those at highest risk of MACE. Acknowledgments The following individuals from the study sponsors were involved in critical review of the manuscript: Carol Hudson, BPharm, and William Sasiela, PhD (Regeneron Pharmaceuticals, Inc.); L. Veronica Lee, MD, and Michael Howard, MBA (Sanofi). Medical writing support was 16

17 /CIRCULATIONAHA provided by Rob Campbell, PhD, Prime Medica, Knutsford, UK, funded by Sanofi and Regeneron Pharmaceuticals, Inc. Sources of Funding This analysis was funded by Sanofi and Regeneron Pharmaceuticals, Inc. Disclosures Downloaded from by guest on April 8, 2018 KK Ray: Personal fees (data safety monitoring board) from AbbVie, Inc., consultant fees/honoraria from Aegerion, Algorithm, Amgen, AstraZeneca, Boehringer Ingelheim, Cerenis, Eli Lily and Company, Ionis Pharmaceuticals, Kowa, Medicines Company, MSD, Novartis, Pfizer, Regeneron, Reservlogix, Sanofi, and Takeda, and research grants from Kowa, Pfizer, and Regeneron. HN Ginsberg: Grants and personal fees from Sanofi, grants from Regeneron, eneron, consultant nt for Amgen and for Pfizer. MH Davidson: Consultant/advisory board fees from Merck, Sanofi, Regeneron, Amgen and Lipimedix. R Pordy: Employee and stockholder, Regeneron. L Bessac and P Minini: Employees and stockholders, Sanofi. RH Eckel: Personal fees from Sanofi/Regeneron, Ionis Pharmaceuticals, UniQure, and Merck. CP Cannon: Grants from Arisaph, AstraZeneca, Bristol-Myers Squibb, Boehringer Ingelheim, GlaxoSmithKline, Janssen, Merck and Takeda; consulting fees from Alnylam, Amgen, Arisaph, Boehringer Ingelheim, Boehringer Ingelheim/Eli Lilly, BMS, GlaxoSmithKline, Kowa, Merck, Takeda, Lipimedix, Pfizer, Regeneron and Sanofi. 17

18 /CIRCULATIONAHA References Downloaded from by guest on April 8, Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III). JAMA. 2001;285: Grundy SM, Cleeman JI, Merz CN, Brewer HB, Jr., Clark LT, Hunninghake DB, Pasternak RC, Smith SC, Jr., Stone NJ. Implications of recent clinical trials for the National Cholesterol Education Program Adult Treatment Panel III guidelines. Circulation. 2004;110: Wood D. JBS 2: Joint British Societies' guidelines on prevention of cardiovascular disease in clinical practice. Heart. 2005;91:v1-v Reiner Z, Catapano AL, De Backer G, Graham I, Taskinen MR, Wiklund O, Agewall S, Alegria E, Chapman MJ, Durrington P, Erdine S, Halcox J, Hobbs R, Kjekshus J, Filardi PP, Riccardi G, Storey RF, Wood D. ESC/EAS Guidelines for the management of dyslipidaemias: the Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS). Eur Heart J. 2011;32: Stone NJ, Robinson J, Lichtenstein e AH, Bairey Merz CN, Lloyd-Jones DM, Blum CB, McBride P, Eckel RH, Schwartz JS, Goldberg AC, Shero ST, Gordon D, Smith SC, Jr., Levy D, Watson K, Wilson PW ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines. J Am Coll Cardiol. 2014;63: Cannon non CP, Blazing MA, Giugliano RP, McCagg A, White JA, Theroux P, Darius H, Lewis BS, Ophuis TO, Jukema JW, De Ferrari ri GM, Ruzyllo lo W, De Lucca P, Im K, Bohula EA, Reist C, Wiviott SD, Tershakovec AM, Musliner TA, Braunwald E, Califf RM. Ezetimibe Added d d to Statin Therapy after Acute Coronary Syndromes. N Engl J Med. 2015;372: Cannon non CP, Braunwald ald E, McCabe CH, Rader DJ, Rouleau JL, Belder e R, Joyal SV, Hill KA, Pfeffer MA, Skene AM, Pravastatin or Atorvastatin E, Infection Therapy-Thrombolysis h in Myocardial Infarction I. Intensive versus moderate lipid lowering with statins after acute coronary syndromes. N Engl J Med. 2004;350: LaRosa JC, Grundy SM, Waters DD, Shear C, Barter P, Fruchart JC, Gotto AM, Greten H, Kastelein JJ, Shepherd J, Wenger NK, Treating to New Targets I. Intensive lipid lowering with atorvastatin in patients with stable coronary disease. N Engl J Med. 2005;352: Ridker PM, Danielson E, Fonseca FA, Genest J, Gotto AM, Jr., Kastelein JJ, Koenig W, Libby P, Lorenzatti AJ, MacFadyen JG, Nordestgaard BG, Shepherd J, Willerson JT, Glynn RJ, Group JS. Rosuvastatin to prevent vascular events in men and women with elevated C-reactive protein. N Engl J Med. 2008;359: Cannon CP, Cariou B, Blom D, McKenney JM, Lorenzato C, Pordy R, Chaudhari U, Colhoun HM. Efficacy and safety of alirocumab in high cardiovascular risk patients with inadequately controlled hypercholesterolaemia on maximally tolerated doses of statins: the ODYSSEY COMBO II randomized controlled trial. Eur Heart J. 2015;36: Kereiakes DJ, Robinson JG, Cannon CP, Lorenzato C, Pordy R, Chaudhari U, Colhoun HM. Efficacy and safety of the PCSK9 inhibitor alirocumab among high cardiovascular risk patients on maximally tolerated statin therapy: the ODYSSEY COMBO I study. Am Heart J. 2015;169: e

19 /CIRCULATIONAHA Downloaded from by guest on April 8, Robinson JG, Farnier M, Krempf M, Bergeron J, Luc G, Averna M, Stroes ES, Langslet G, Raal FJ, El Shahawy M, Koren MJ, Lepor NE, Lorenzato C, Pordy R, Chaudhari U, Kastelein JJ. Efficacy and safety of alirocumab in reducing lipids and cardiovascular events. N Engl J Med. 2015;372: National Institute for Health and Care Excellence. Cardiovascular disease: risk assessment and reduction, including lipid modification. NICE guidelines [CG181]. Available from: [Accessed 31 March 2016]. 14. Lloyd-Jones DM, Morris PB, Ballantyne CM, Birtcher KK, Daly DD, Jr., DePalma SM, Minissian MB, Orringer CE, Smith SC, Jr ACC Expert Consensus Decision Pathway on the Role of Non-Statin Therapies for LDL-Cholesterol Lowering in the Management of Atherosclerotic Cardiovascular Disease Risk: A Report of the American College of Cardiology Task Force on Clinical Expert Consensus Documents. J Am Coll Cardiol. 2016;68: Expert Dyslipidemia Panel of the International Atherosclerosis Society Panel m. An International Atherosclerosis Society Position Paper: global recommendations for the management of dyslipidemia--full report. J Clin Lipidol. 2014;8: Jacobson TA, Ito MK, Maki KC, Orringer CE, Bays HE, Jones PH, McKenney JM, Grundy SM, Gill EA, Wild RA, Wilson DP, Brown WV. National lipid association recommendations e o for patient-centered e ed management age e of dyslipidemia: de part 1--full report.. J Clin Lipidol. 2015;9: Bays H, Gaudet D, Weiss R, Ruiz JL, Watts GF, Gouni-Berthold I, Robinson on J, Zhao J, Hanotin C, Donahue S. Alirocumab as add-on to atorvastatin versus other lipid treatment strategies: ODYSSEY OPTIONS I randomized trial. J Clin Endocrinol Metab. 2015: Farnier M, Jones P, Severance R, Averna M, Steinhagen-Thiessen E, Colhoun HM, Du Y, Hanotin C, Donahue S. Efficacy and safety of adding alirocumab to rosuvastatin versus adding ezetimibe or doubling the rosuvastatin stat dose in high cardiovascular-risk ar-risk patients: the ODYSSEY OPTIONS II randomized trial. Atherosclerosis. 2016;244: Ginsberg HN, Rader DJ, Raal FJ, Guyton JR, Baccara-Dinet MT, Lorenzato C, Pordy R, Stroes E. Efficacy fica and Safety of Alirocumab in Patients with Heterozygous eroz Familial Hypercholesterolemia and LDL-C of 160 mg/dl or Higher. Cardiovasc Drugs Ther DOI: /s y [Epub ahead of print]. 20. Kastelein JJ, Ginsberg HN, Langslet G, Hovingh GK, Ceska R, Dufour R, Blom D, Civeira F, Krempf M, Lorenzato C, Zhao J, Pordy R, Baccara-Dinet MT, Gipe D, Geiger MJ, Farnier M. ODYSSEY FH I and FH II: 78 week results with alirocumab treatment in 735 patients with heterozygous familial hypercholesterolaemia. Eur Heart J. 2015;36: Moriarty PM, Thompson PD, Cannon CP, Guyton JR, Bergeron J, Zieve FJ, Bruckert E, Jacobson TA, Kopecky SL, Baccara-Dinet MT, Du Y, Pordy R, Gipe DA. Efficacy and safety of alirocumab versus ezetimibe in statin-intolerant patients, with a statin-re-challenge arm: The ODYSSEY ALTERNATIVE randomized trial. J Clin Lipidol. 2015;9: Roth EM, Taskinen MR, Ginsberg HN, Kastelein JJ, Colhoun HM, Robinson JG, Merlet L, Pordy R, Baccara-Dinet MT. Monotherapy with the PCSK9 inhibitor alirocumab versus ezetimibe in patients with hypercholesterolemia: results of a 24 week, double-blind, randomized Phase 3 trial. Int J Cardiol. 2014;176: Schwartz GG, Bessac L, Berdan LG, Bhatt DL, Bittner V, Diaz R, Goodman SG, Hanotin C, Harrington RA, Jukema JW, Mahaffey KW, Moryusef A, Pordy R, Roe MT, Rorick T, Sasiela WJ, Shirodaria C, Szarek M, Tamby JF, Tricoci P, White H, Zeiher A, Steg PG. Effect of alirocumab, a monoclonal antibody to PCSK9, on long-term cardiovascular outcomes following 19

20 /CIRCULATIONAHA Downloaded from by guest on April 8, 2018 acute coronary syndromes: Rationale and design of the ODYSSEY Outcomes trial. Am Heart J. 2014;168: Wiviott SD, Cannon CP, Morrow DA, Ray KK, Pfeffer MA, Braunwald E. Can lowdensity lipoprotein be too low? The safety and efficacy of achieving very low low-density lipoprotein with intensive statin therapy: a PROVE IT-TIMI 22 substudy. J Am Coll Cardiol. 2005;46: Baigent C, Blackwell L, Emberson J, Holland LE, Reith C, Bhala N, Peto R, Barnes EH, Keech A, Simes J, Collins R. Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet. 2010;376: Mindell J, Aresu M, Zaninotto P, Falaschetti E, Poulter N. Improving lipid profiles and increasing use of lipid-lowering therapy in England: results from a national cross-sectional survey Clin Endocrinol (Oxf). 2011;75: Kuklina EV, Yoon PW, Keenan NL. Trends in high levels of low-density lipoprotein cholesterol in the United States, JAMA. 2009;302: Di Angelantonio E, Sarwar N, Perry P, Kaptoge S, Ray KK, Thompson A, Wood AM, Lewington S, Sattar N, Packard CJ, Collins R, Thompson SG, Danesh J. Major lipids, apolipoproteins, pop o and risk of vascular a disease. JAMA. 2009;302: Boekholdt SM, Hovingh GK, Mora S, Arsenault BJ, Amarenco P, Pedersenen TR, LaRosa JC, Waters DD, DeMicco DA, Simes RJ, Keech AC, Colquhoun D, Hitman GA, Betteridge erid DJ, Clearfield MB, Downs JR, Colhoun HM, Gotto AM, Jr., Ridker PM, Grundy SM, Kastelein JJ. Very low levels of atherogenic lipoproteins and the risk for cardiovascular events: a meta-analysis of statin trials. J Am Coll Cardiol. 2014;64: Boekholdt SM, Arsenault BJ, Mora S, Pedersen TR, LaRosa JC, Nestel PJ, Simes RJ, Durrington P, Hitman GA, Welch KM, DeMiccoc DA, Zwinderman AH, Clearfield ld MB, Downs JR, Tonkin AM, Colhoun HM, Gotto to AM, Jr., Ridker PM, Kastelein JJ. J. Association on of LDL cholesterol, non-hdl cholesterol, erol, and apolipoprotein op B levels e with risk of cardiovascular cular events ents among patients ts treated with statins: s: a meta-analysis. JAMA. A. 2012;307: ; Robinson JG, Smith B, Maheshwari N, Schrott H. Pleiotropic effects of statins: benefit beyond cholesterol reduction? A meta-regression analysis. J Am Coll Cardiol. 2005;46: Robinson JG, Wang S, Smith BJ, Jacobson TA. Meta-analysis of the relationship between non-high-density lipoprotein cholesterol reduction and coronary heart disease risk. J Am Coll Cardiol. 2009;53: LaRosa JC, Grundy SM, Kastelein JJ, Kostis JB, Greten H. Safety and efficacy of Atorvastatin-induced very low-density lipoprotein cholesterol levels in Patients with coronary heart disease (a post hoc analysis of the treating to new targets [TNT] study). Am J Cardiol. 2007;100: Everett BM, Mora S, Glynn RJ, MacFadyen J, Ridker PM. Safety profile of subjects treated to very low low-density lipoprotein cholesterol levels (<30 mg/dl) with rosuvastatin 20 mg daily (from JUPITER). Am J Cardiol. 2014;114:

21 /CIRCULATIONAHA Table 1. Baseline characteristics Placebo-controlled trials Alirocumab (n=2324) Placebo (n=1175) Ezetimibe-controlled trials* Alirocumab (n=864) Ezetimibe (n=620) Age, years, mean (SD) 58.7 (11.6) 58.8 (11.4) 61.9 (9.4) 62.1 (9.5) Sex, male, n (%) 1415 (60.9) 712 (60.6) 581 (67.2) 388 (62.6) Race, white, n (%) 2139 (92.0) 1072 (91.2) 745 (86.2) 548 (88.4) BMI, kg/m 2, mean (SD) 30.1 (5.6) 30.3 (5.6) 30.2 (6.0) 30.0 (5.7) HeFH, n (%) 838 (36.1) 419 (35.7) 40 (4.6) 43 (6.9) Diabetes, n (%) 699 (30.1) 355 (30.2) 283 (32.8) 192 (31.0) ASCVD, n (%) 1615 (69.5) 834 (71.0) 651 (75.3) 411 (66.3) Downloaded from by guest on April 8, 2018 CHD 1454 (62.6) 766 (65.2) 611 (70.7) 390 (62.9) Ischemic stroke/tia 199 (8.6) 86 (7.3) 67 (7.8) 42 (6.8) PAD 97 (4.2) 56 (4.8) 33 (3.8) 19 (3.1) Current smoker, n (%) 453 (19.5) 231 (19.7) 146 (16.9) 118 (19.0) Statin intensity, n (%) High 1327 (57.1) 682 (58.0) 430 (49.8) 265 (42.7) Moderate 650 (28.0) 314 (26.7) 208 (24.1) 152 (24.5) Low 344 (14.8) 177 (15.1) 47 (5.4) 27 (4.4) No statins (20.6) 176 (28.4) Missing 3 (0.1) 2(0.2) 1(0.1) 0 Baseline lipids, mg/dl, mean (SD) LDL-C (46.3) (44.8) (51.5) (56.9) non-hdl-c (49.6) (48.5) (57.2) (64.1) apob (29.0) (28.5) (30.7) (32.2) Pooled data of all randomized patients from the 10 trials included in this analysis. *In combination with statins or not. Patients may be in more than one category. Atorvastatin 40 to 80 mg, rosuvastatin 20 to 40 mg or simvastatin 80 mg. Atorvastatin 20 to <40 mg, rosuvastatin 10 to <20 mg or simvastatin 40 to <80mg. Atorvastatin <20 mg, rosuvastatin <10 mg or simvastatin <40 mg. ASCVD, atherosclerotic cardiovascular disease; BMI, body mass index; CHD, coronary heart disease, HeFH, heterozygous familial hypercholesterolemia, non-hdl-c, non-high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; PAD, peripheral artery disease, SD, standard deviation, TIA, transient ischemic attack. 21

22 /CIRCULATIONAHA Table 2. Achieved lipid levels and percentage reductions during treatment and relationship to MACE Downloaded from by guest on April 8, 2018 Placebo-controlled trials Ezetimibe-controlled trials MACE versus average achieved level (pool of all patients from the trials) Average achieved, Alirocumab Placebo Alirocumab Ezetimibe Category N Hazard ratio (95% CI) P-value mg/dl (n=2318) (n=1174) (n=864) (n=618) LDL-C 56.9 (38.8) (43.9) 64.0 (42.4) (50.8) Per 39 mg/dl (0.63 to 0.91) difference non-hdl-c 82.0 (41.8) (47.2) 91.1 (45.6) (55.2) Per 42 mg/dl (0.65 to 0.93) difference apob 57.1 (29.1) (28.1) 64.9 (28.1) 89.2 (31.0) Per 27 mg/dl difference (0.60 to 0.86) Average % change MACE versus % change in average age level from baseline LDL-C (23.5) 2.7 (25.7) (23.8) (28.9) Per 50% (0.57 to 0.89) reduction non-hdl-c (20.7) 2.6 (21.1) (20.1) (20.5) Per 50% (0.52 to 0.97) reduction apob (22.8) 2.2 (24.0) (20.9) (20.1) Per 50% reduction (0.54 to 0.85) Lipids are mean (SD). Hazard ratio, 95% confidence ce interval (CI) and P-value l e determineded from a multivariate Cox model. Multivariate te analysis adjusted on baseline characteristics cter cs (age, sex, diabetes, es, prior history of MI/stroke, baseline LDL-C and smoking status). Average LDL-C during the treatment period determined from the area under the curve (using trapezoidal method), taking into account all LDL-C values up to end of treatment period or occurrence of MACE, whichever comes first. For patients with no postbaseline LDL-C, LDL-C at baseline was used. Note: 2 patients with missing baseline LDL-C and 3 with missing baseline apob were excluded from the multivariate analysis. Non-HDL-C, non-high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; MACE major adverse cardiovascular events. 22

23 /CIRCULATIONAHA Figure Legends Figure 1. Distribution of achieved levels of LDL-C (A), non-hdl-c (B) and apob (C) during treatment, stratified by control group.for patients with no post-baseline lipid measurement, baseline values were used. Non-HDL-C, non-high-density lipoprotein cholesterol; LDL-C, lowdensity lipoprotein cholesterol. Downloaded from by guest on April 8, 2018 Figure 2. Distribution of the percentage reductions in LDL-C (A), non-hdl-c (B) and apob (C) from baseline during treatment stratified by control group. For patients with no post-baseline lipid measurement, baseline values were used. Two patients with missing baseline LDL-C and three patients with missing baseline apob were excluded from the analysis.non-hdl-c, nonhigh-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol. Figure 3. Relationship between e on-treatment lipids and reductions in lipid levels with MACE. Panels A, B and C show adjusted MACE rate by achieved levels of LDL-C, L non-hdl-c, and apob, respectively, during follow up. Corresponding results for percent reductions are shown in panels D, E and F, respectively. Multivariate analysis adjusted on baseline characteristics (age, sex, diabetes, prior history of MI/stroke, baseline LDL-C and smoking status). Non-HDL-C, non-high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; MACE major adverse cardiovascular events. 23

24 Downloaded from by guest on April 8, 2018 Figure 1A

25 Downloaded from by guest on April 8, 2018 Figure 1B

26 Downloaded from by guest on April 8, 2018 Figure 1C

27 Downloaded from by guest on April 8, 2018 Figure 2A

28 Downloaded from by guest on April 8, 2018 Figure 2B

29 Downloaded from by guest on April 8, 2018 Figure 2C

30 Downloaded from by guest on April 8, 2018 Figure 3A

31 Downloaded from by guest on April 8, 2018 Figure 3B

32 Downloaded from by guest on April 8, 2018 Figure 3C

33 Downloaded from by guest on April 8, 2018 Figure 3D

34 Downloaded from by guest on April 8, 2018 Figure 3E

35 Downloaded from by guest on April 8, 2018 Figure 3F

36 Reductions in Atherogenic Lipids and Major Cardiovascular Events: A Pooled Analysis of 10 ODYSSEY Trials Comparing Alirocumab to Control Kausik K. Ray, Henry N. Ginsberg, Michael H. Davidson, Robert Pordy, Laurence Bessac, Pascal Minini, Robert H. Eckel and Christopher P. Cannon Downloaded from by guest on April 8, 2018 Circulation. published online October 24, 2016; Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 2016 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: Free via Open Access Data Supplement (unedited) at: Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Circulation is online at:

37 Dr. Carolyn Lam: Welcome to Circulation on the Run, your weekly podcast summary and backstage pass to the journal and its editors. I'm Dr. Carolyn Lam, associate editor from the National Heart Center and Duke National University of Singapore. Today we will be discussing the pooled analysis results of the 10 ODYSSEY Trials with important implications for the reduction of lipids in major cardiovascular events. But first, here's your summary of this week's journal. The first paper provides experimental data on vascular disease that brings into focus the critical roles of transcription factors such as GATA2 in the maintenance of endothelial cell function, as well as the role of selected micrornas as a novel player of vascular regulation. In this study by first author Dr. Hartman, corresponding author Dr. Thum from Hanover Medical School, and colleagues, authors used GATA2 gain and loss of function experiments in human umbilical vein endothelial cells to identify a key role of GATA2 as a master regulator of multiple endothelial functions, and this via microrna-dependent mechanisms. Global microrna screening identified several GATA2-regulated micrornas, including mir-126 and mir-221. GATA2 deficiency led to vascular abnormalities, whereas supplementation with mir-126 normalized vascular function. In a mouse model of carotid injury, GATA2 was reduced and systemic supplementation of mir-126-coupled nanoparticles enhanced mir-126 availability in the carotid artery and improved reendothelialization of injured carotid arteries in vivo. In summary, GATA2-mediated regulation of mir-126 and mir-221 has an important impact on endothelial biology. Thus, modulation of GATA2 and its targets mir-126 and mir-221 represents a promising therapeutic strategy for the treatment of vascular diseases. The next study is the first to show that current smokers from the general population have lower levels of circulating cardiac troponin I, a seemingly paradoxical observation given the known detrimental cardiovascular impact of cigarette smoking. First author Dr. Lyngbakken, corresponding author Dr. Omland, and colleagues from the University of Oslo used data from the large population-based HUNT study, in which cardiac troponin I was measured in 3,824 never smokers, 2,341 former smokers, and 2,550 current smokers. Current smokers had significantly lower levels of cardiac troponin I than never smokers and former smokers, an association that remains significant even after adjustment for potential confounders. The authors also found an association between increasing concentrations of troponin I and clinical endpoints, namely acute myocardial infarction, heart COTR134_24 Page 2 of 8

38 failure, and cardiovascular death in the total cohort. However, this association was attenuated in current smokers and was significantly weaker than in never or former smokers with a p for interaction of The prognostic accuracy of troponin I as assessed by C-statistics was lower in current smokers than in never smokers. Troponin I provided no incremental prognostic information to the Framingham Cardiovascular Disease risk score in the current smokers. Together, these results suggest that mechanistic pathways other than those involving subclinical myocardial injury may be responsible for the cardiovascular risk associated with current smoking. Future studies are needed to determine whether a lower cardiac troponin I threshold should be considered for exclusion of myocardial infarction in smokers or whether prognostic tools other than measurement of cardiac troponins should be utilized when evaluating risk of future events in current smokers. The next study contributes to our understanding of cardiomyocyte signaling in response to ischemic injury. In the study by first author Dr. [Wool 00:05:04], corresponding author Dr. [Ju 00:05:04] from Tongji University School of Medicine in Shanghai, and colleagues, authors sought to understand the role of low-density lipoprotein receptor-related proteins 5 and 6 as well as beta-catenin signaling in the heart. They did this using conditional cardiomyocyte-specific knockout mice who had surgically induced myocardial infarction. They found that deletion of lipoprotein receptor-related proteins 5 and 6 promoted cardiac ischemic insults. Conversely, deficiency of beta-catenin, a downstream target, was beneficial in ischemic injury. Interestingly, although both insulin-like growth factor-binding protein 4 and Dickkopf-related protein 1 are secreted betacatenin pathway inhibitors, the former protected the ischemic heart by inhibiting beta-catenin, whereas the latter enhanced the injury response mainly through inducing lipoprotein-related protein 5 and 6 endocytosis and degradation. These findings really add to our understanding of the beta-catenin signaling pathway in ischemic injury and suggests that new therapeutic strategies in ischemic heart disease may involve fine-tuning these signaling pathways. The next paper from the International Consortium of Vascular Registries is the first study allowing an assessment of variations in repair of abdominal aortic aneurysms in 11 countries over 3 continents represented by the Society of Vascular Surgery and European Society for Vascular Surgery. Dr. Beck from University of Alabama-Birmingham School of Medicine, and colleagues, looked at registry data for open and endovascular abdominal aortic aneurysm repair during 2010 to 2013, collected from 11 countries. These were Australia, Denmark, Hungary, Iceland, New Zealand, Norway, Sweden, Finland, Switzerland, Germany, and the United States. Among more than 51,000 patients, utilization of endovascular aortic repair for intact aneurysms varied from 28% in Hungary to 79% in the United States, and COTR134_24 Page 3 of 8

39 for ruptured aneurysms from 5% in Denmark to 52% in the United States. In addition to the between-country variations, significant variations were present between centers within each country in terms of endovascular aortic repair use and rate of small aneurysm repair. Countries that more frequently treated small aneurysms tended to use the endovascular approach more frequently. Octogenarians made up 23% of all patients, with a range of 12% in Hungary to 29% in Australia. In countries with a fee for service reimbursement systems, such as Australia, Germany, Switzerland, and the United States, the proportion of small aneurysms and octogenarians undergoing intact aneurysm repair was higher compared to countries with a population-based reimbursement model. In general, center-level variation within countries in the management of aneurysms was as important as variation between counties. Hence, this study shows that despite homogeneous guidelines from professional societies, there is significant variation in the management of abdominal aortic aneurysms, most notably for intact aneurysm diameter at repair, utilization of endovascular approaches, and the treatment of elderly patients. These findings suggest that there is an opportunity for further international harmonization of treatment algorithms for abdominal aortic aneurysms. This is discussed in an accompanying editorial entitled, Vascular Surgeons Leading the Way in Global Quality Improvement, by Dr. Fairman. The final paper from Dr. Gibson at Beth Israel Deaconess Medical Center and Harvard Medical School and colleagues, presents the results of the apoai event reducing in ischemic syndromes I, or AEGIS-I, trial, which was a multicenter, randomized, doubleblind, placebo-controlled dose-ranging phase 2b trial of CSL112, which is an infusible, plasma-derived apoai that has been studied in normal subjects and those with stable coronary artery disease, but now studied in the current study in patients with acute myocardial infarction. The trial showed that among patients with acute myocardial infarction, four weekly infusions of a reconstituted, infusible, human apoai, CSL112, was associated with a dose-dependent elevation of circulating apoai and cholesterol efflux capacity without adverse hepatic or renal outcomes. The potential benefit of CSL112 to reduce major adverse cardiovascular events will need to be assessed in an adequately powered phase 3 trial. Now for our future discussion. Today I am delighted to have with us Dr. Kausik Ray from Imperial College London, who's the first and corresponding author of a new paper regarding the pooled analysis of the 10 ODYSSEY Trials. To discuss it with us is Dr. Carol Watson, associate editor from UCLA. Kausik, just let me start by congratulating you on this paper. I believe this is the first data that allows us to look under the 50 mg/dl mark of LDL and really ask if the LDL MACE relationship extends below this level. Dr. Kausik Ray: Yes, the reason for looking at this is that the IMPROVE-IT trial really looked at people down to an average LDL cholesterol of about 54, and with the new COTR134_24 Page 4 of 8

40 PCSK9 inhibitors, which instead of giving you a 20% further reduction LDL, they give you the opportunity for a further 50 to 60% reduction. We actually get the chance to get people down to levels like 25 mg/dl, and the question is, does the benefit continue at that level? We did a pooled analysis of 10 of the ODYSSEY Trials, really in some ways to try and help predict what you might see in ODYSSEY outcomes, what you might see in the [Fuliay 00:12:00] trial, and to also manage expectations as well, because there's probably been a lot of hype around the two New England Journal papers about 50, 60% reductions of all potential reductions based on small numbers of events. So the question is, if you reduce LDL by 39 mg/dl, how might that reduce your risk, and is the relationship continuous? So those were the aims. Dr. Carolyn Lam: Dr. Kausik Ray: Dr. Carolyn Lam: Dr. Kausik Ray: Dr. Carolyn Lam: Dr. Kausik Ray: That's great, and maybe could you give us an idea of the number of patients you are looking at and the number of events? Yeah. In the 10 pool studies, we had just under 5,000 individuals, and we had just about 6,700 person years' worth of followup. In total, we had 104 first MACE events. To put this into context, it's about one third of the number of events that the first [framing 00:12:53] of analysis had. It's an observation analysis rather than randomized trial data, so you got to bear that in mind with the usual caveats that go with observational data. But the same endpoints that were adjudicated, this is [inaudible 00:13:10] heart disease death, non-fatal MI, ischemic stroke, and unstable angina requiring hospitalization. This is the same endpoint that is in the ODYSSEY Outcomes Trial, so it's interesting in that regard. Yeah, it sure is. So what's the bottom line? What did you find? What we found was that there was a continuous relationship all the way down to LDL cholesterol levels of about 25 mg/dl, that every 39 mg/dl lower on treatment LDL, your risk went down by about 24%. If you looked at [apo-like 00:13:48] approaching be on non-hdl cholesterol, again, you found the same continuous relationship with a similar point estimate for a similar standardized difference in LDL cholesterol. We also looked at many of the guidelines, talk about percentage reduction. We actually looked at percentage reductions. If you start with a baseline LDL of X and you achieve a 50% further reduction in LDL, how much further benefit does that give you? A 50% further reduction gave you a 29% further lower risk of MACE. So we didn't find any threshold or limit all the way down to LDLs of about 25. That's really a key, novel finding that you contributed, so congratulations once again. I suppose the question will always be, you're talking about relative risk reductions here. At such low levels, can you give us an idea of the absolute risk reductions? Yes. You've got to remember that the relative risk reductions are what you can apply to population differences. If you pick a high-risk patient population, you COTR134_24 Page 5 of 8

41 would expect to see a much bigger absolute risk reduction than maybe this study or another study. Similarly, if you pick a low-risk group, you are going to see a much smaller absolute benefit. I always try to advise a little bit of caution that if you basically look at the range... If you start with let's say an LDL of 150 and you go down to let's say an LDL of 25, you are talking about a 1.25% absolute risk reduction. Remember, these patients are possibly going to be a slightly lower risk than the ones that are recruited into the ODYSSEY Outcomes and into the [Fuliay 00:15:46] trial, for example. Dr. Carolyn Lam: Dr. Kausik Ray: I think you mentioned what I was going to just ask you about. This is observational. You had 104 events, and I suppose another limitation might be that your followup was two years at max, if I'm not wrong? What do you say about that, and are there plans for future analyses? Within the context of these studies, I think that the whole of this data will eventually become dwarfed by what we see with the big CDOTs, because you've got 18, 27,000 people, 3 years' worth of exposure and followups, so you are going to have many, many more events. That is a limitation, but I think what is interesting is that we know that the baseline LDL cholesterol level is around about 90 mg/dl. We don't actually know what the actual baseline... because the baseline [characters 00:16:43] haven't been published for ODYSSEY Outcomes, but the [Fuliays 00:16:46] around about 89. What it tells you is what the point estimate is likely to be. It's likely to be in the 24 to 32% ballpark because that's what your baseline LDL is and that's what we'd predict in the regression lines that we observed here. I think that we're not going to get many more events in these studies because largely the randomized period of followup is now over. Many of these people are now into open labels, extensions for safety, so we won't get many more events from this. In terms of, I think, the way people should maybe look at this is possibly as a taster for what's to come in the next 18 months or so. I think for the time being it answers two questions. Is lower likely to be better? And it is. I think the other question it tells is how might you get people down to LDLs below 50? One of the important things was that if you were just on statins, in this population, if you were recruited on the basis of a high baseline LDL, you got no additional people down to LDLs below 50. You got under 10% with add-on [inaudible 00:18:05], but you got around about 50% when you used the PCSK9 inhibitor as an add-on to existing therapy. It tells you about how to get to such low levels as well. I think that's the other key thing that it actually gives you. We did an analysis of safety [inaudible 00:18:23], and I think that's really important. Once you see the efficacy, or if you see the MACE events continue to go down... If you looked at treatment-emergent adverse events... and I completely take the fact that it's every side effect reported altogether, which may or may not be linked to LDL levels specifically, but when we did that, the COTR134_24 Page 6 of 8

42 relationship actually was just a horizontal line, so there was no relationship with percentage reduction or on treatment LDL, so it gave us a nice idea of both safety and efficacy that we might experience in the big outcome studies. Dr. Carolyn Lam: Dr. Carol : All right. Obviously the big outcome studies are going to be game changers, and I'd really love to invite [Carol Scotts 00:19:09] here, because there's a whole lot of other things that need to be considered if this becomes the case, isn't it? Carol, I really appreciated that you invited an editorial, and the editorial is by Neil Stone who entitles it, Looking Beyond Statins: Will the Dollars Make Cents? Please tell us about the discussions about this paper that occurred. I would again like to congratulate Dr. Ray on a fantastic paper, and I would like to reiterate exactly what he said. I think it really does give us some comfort about this class of medication and its relative safety. I think that's very important, because I can't tell you how many patients I get and how many referring physicians I get who worry when their patients come back with LDLs of 20 or below. I think that gave us some comfort, and I do also think it was very important to show that this would fall along the same regression line that statins perhaps would fall. As with all the caveats that Dr. Ray said, I agree with all of them, but I do say this is a tasty little taster, and I appreciate and congratulate you for publishing this. The editorial by Dr. Neil Stone was quite interesting. As you said, he subtitled it, Will the Dollars Make Cents? C E N T S or S E N S E, sort of a play on words there. Will the relative benefits that we can achieve with this class of medications make sense for the cost of these drugs? That's obviously a very separate issue from what was discussed in the manuscript, but it's something to think about. We understand that there are additional patients that will be helped if they can get their LDL down, and we hope that that will translate into the outcomes. Again, just as Dr. Ray mentioned, we will have to wait for the cardiovascular outcomes trials to be completed. When they are, if they do show the benefits that we hope, will their price point make them accessible to enough patients for this to be a widely applied, utilized therapy? Or will they not? That's part of what was discussed in Dr. Stone's editorial. Dr. Kausik Ray: When we were writing the manuscript and stuff like that, and we were doing this and everybody's like, "Oh, wow, look at the graphs." I said, "Look, we need to balance all of these bits and reassure... We've got an opportunity." So I suggested them giving those additional analyses, and you saw how big the online supplement was. There was a ton of work that we put into this, and to format it into a concise... I really want to just thank the editorial board for giving us the chance and actually being able to help us and work with us on this, because it's really important. I hope people look at all of those things because it will help people also that question the LDL. They all talk about the hypothesis COTR134_24 Page 7 of 8

43 and the safety of really low LDLs, and people come off statins as a result. I think this will help. Dr. Carolyn Lam: You're listening to Circulation on the Run. Thank you so much for being with us, and don't forget to tune in next week. COTR134_24 Page 8 of 8

44 Supplemental Material 1

45 Definition of unstable angina used in the ODYSSEY studies A diagnosis of an unstable angina (new ACS event without elevation in cardiac biomarkers) that meets the primary endpoint requires the following: Admission to hospital or emergency room (until at least next calendar day) with symptoms presumed to be caused by myocardial ischemia with an accelerating tempo in the prior 48 hours. and/or prolonged (at least 20 minutes) rest chest discomfort AND New high-risk ECG findings consistent with ischemia (or presumed new if no prior ECG available), as defined below: New or presumed new ST depression >0.5mm in 2 contiguous leads or T wave inversion >1mm in leads with prominent R wave or R/S >1 in 2 contiguous leads New or presumed new ST elevation at the J point in >2 contiguous leads >0.2mV in V2 or V3 in men or >0.15 mv in women in V2 or V3 or >0.1mV in other leads LBBB (new or presumed new) AND Definite contemporary* evidence of angiographically significant coronary disease as demonstrated by: 2

46 Need for coronary revascularization procedure (PCI or CABG) excluding those performed to treat only restenosis lesion(s) at previous PCI site(s) OR Angiographic evidence of at least 1 significant (> 70%) epicardial coronary stenosis not due to restenosis at previous PCI site. *The coronary revascularization procedure or the diagnostic angiography must have been performed during the hospitalization for that event. 3

47 etable 1. Key inclusion and exclusion criteria for trials included in this analysis Study Inclusion criteria LDL-C exclusion criteria FH I, FH II 1 Patients with HeFH not adequately controlled with a maximally-tolerated stable daily dose of statin * for 4 weeks prior to screening ± other LLT LDL-C <70 mg/dl (for patients with documented clinical ASCVD) or LDL-C <100 mg/dl (for patients with no documented ASCVD) HIGH FH 2 As above LDL-C <160 mg/dl. COMBO I, 3 COMBO II 4 LONG TERM 5 OPTIONS I, 6 OPTIONS II 7 ALTERNATIVE 8 MONO 9 Patients with hypercholesterolemia (non-fh), not adequately controlled with a maximally-tolerated stable daily dose of statin * for 4 weeks prior to screening (± other LLT in COMBO I; no other LLT allowed in COMBO II) Included both HeFH and non-fh patients, otherwise inclusion criteria as for the FH studies above Patients with HeFH or non-fh, receiving either atorvastatin 20 or 40 mg (OPTIONS I) or rosuvastatin 10 or 20 mg (OPTIONS II) ± other LLT (apart from ezetimibe as it was used as a comparator) Patients with HeFH or non-fh and documented statin intolerance and moderate to very high cardiovascular risk; patients were not receiving a statin but other LLT (apart from ezetimibe) were allowed Subjects with 10-year risk of fatal CV events of 1% and 5% based on the European Systematic Coronary Risk Evaluation (SCORE) 10 not receiving statin or other LLT (history of CHD or HeFH was an exclusion criterion for this study) LDL-C <70 mg/dl (for patients with documented clinical ASCVD) or LDL-C <100 mg/dl (for patients with no documented ASCVD) LDL-C <70 mg/dl. LDL-C <70 mg/dl (for patients with documented clinical ASCVD) or LDL-C <100 mg/dl (for patients with no documented ASCVD) LDL-C <70 mg/dl for very-high risk, LDL-C <100 mg/dl for moderate or high risk LDL-C <70 mg/dl or >190 mg/dl Exclusion criteria common to all studies Age <18 years Use of fibrates, other than fenofibrate 4

48 Fasting serum triglycerides >400 mg/dl Estimated glomerular filtration rate (egfr) <60 ml/min/1.73 m 2 Clinicaltrials.gov identifiers: LONG TERM, NCT ; FH I NCT ; FH II, NCT ; HIGH FH, NCT ; COMBO I, NCT ; COMBO II, NCT ; MONO, NCT ; OPTIONS I, NCT ; OPTIONS II, NCT ; ALTERNATIVE, NCT ASCVD, atherosclerotic cardiovascular disease; LLT, lipid-lowering therapy. * Maximally tolerated statin dose = the highest tolerable registered dose of daily statin currently administered to the patient, i.e. rosuvastatin 20 or 40 mg, atorvastatin 40 or 80 mg; or simvastatin 80 mg. Lower doses could be used e.g. in the case of intolerance or local practice according to the investigator s judgment. Inability to tolerate 2 or more statins because of muscle-related symptoms. Moderate risk = 10-year risk of fatal CV events of 1% and 5% (SCORE); high risk = SCORE 5%; egfr 30 to <60 ml/min/1.73 m 2 ; type 1 or 2 diabetes mellitus without target organ damage; or HeFH; very-high risk = CHD, ischemic stroke, peripheral artery disease, transient ischemic attack, abdominal aortic aneurysm, or carotid artery occlusion >50% without symptoms; carotid endarterectomy or carotid artery stent procedure; renal artery stenosis or renal artery stent procedure; or type 1 or type 2 diabetes mellitus with target organ damage. References 1. Kastelein JJ, Ginsberg HN, Langslet G, Hovingh GK, Ceska R, Dufour R, Blom D, Civeira F, Krempf M, Lorenzato C, Zhao J, Pordy R, Baccara-Dinet MT, Gipe D, Geiger MJ, Farnier M. ODYSSEY FH I and FH II: 78 week results with alirocumab treatment in 735 patients with heterozygous familial hypercholesterolaemia. Eur Heart J. 2015;36: Ginsberg HN, Rader DJ, Raal FJ, Guyton JR, Baccara-Dinet MT, Lorenzato C, Pordy R, Stroes E. Efficacy and Safety of Alirocumab in Patients with Heterozygous Familial Hypercholesterolemia and LDL-C of 160 mg/dl or Higher. Cardiovasc Drugs Ther. 2016;30: Kereiakes DJ, Robinson JG, Cannon CP, Lorenzato C, Pordy R, Chaudhari U, Colhoun HM. Efficacy and safety of the PCSK9 inhibitor alirocumab among high cardiovascular risk patients on maximally tolerated statin therapy: the ODYSSEY COMBO I study. Am Heart J. 2015;169: e913. 5

49 4. Cannon CP, Cariou B, Blom D, McKenney JM, Lorenzato C, Pordy R, Chaudhari U, Colhoun HM. Efficacy and safety of alirocumab in high cardiovascular risk patients with inadequately controlled hypercholesterolaemia on maximally tolerated doses of statins: the ODYSSEY COMBO II randomized controlled trial. Eur Heart J. 2015;36: Robinson JG, Farnier M, Krempf M, Bergeron J, Luc G, Averna M, Stroes ES, Langslet G, Raal FJ, El Shahawy M, Koren MJ, Lepor NE, Lorenzato C, Pordy R, Chaudhari U, Kastelein JJ. Efficacy and safety of alirocumab in reducing lipids and cardiovascular events. N Engl J Med. 2015;372: Bays H, Gaudet D, Weiss R, Ruiz JL, Watts GF, Gouni-Berthold I, Robinson J, Zhao J, Hanotin C, Donahue S. Alirocumab as add-on to atorvastatin versus other lipid treatment strategies: ODYSSEY OPTIONS I randomized trial. J Clin Endocrinol Metab. 2015: Farnier M, Jones P, Severance R, Averna M, Steinhagen-Thiessen E, Colhoun HM, Du Y, Hanotin C, Donahue S. Efficacy and safety of adding alirocumab to rosuvastatin versus adding ezetimibe or doubling the rosuvastatin dose in high cardiovascular-risk patients: the ODYSSEY OPTIONS II randomized trial. Atherosclerosis. 2016;244: Moriarty PM, Thompson PD, Cannon CP, Guyton JR, Bergeron J, Zieve FJ, Bruckert E, Jacobson TA, Kopecky SL, Baccara-Dinet MT, Du Y, Pordy R, Gipe DA. Efficacy and safety of alirocumab versus ezetimibe in statin-intolerant patients, with a statin-re-challenge arm: The ODYSSEY ALTERNATIVE randomized trial. J Clin Lipidol. 2015;9: Roth EM, Taskinen MR, Ginsberg HN, Kastelein JJ, Colhoun HM, Robinson JG, Merlet L, Pordy R, Baccara-Dinet MT. Monotherapy with the PCSK9 inhibitor alirocumab versus ezetimibe in patients with hypercholesterolemia: results of a 24 week, double-blind, randomized Phase 3 trial. Int J Cardiol. 2014;176: Reiner Z, Catapano AL, De Backer G, Graham I, Taskinen MR, Wiklund O, Agewall S, Alegria E, Chapman MJ, Durrington P, Erdine S, Halcox J, Hobbs R, Kjekshus J, Filardi PP, Riccardi G, Storey RF, Wood D. ESC/EAS Guidelines for the management of dyslipidaemias: the Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS). Eur Heart J. 2011;32:

50 efigure 1. Overview of studies included in this analysis ALI, alirocumab; CV, cardiovascular ; EZE, ezetimibe; HeFH, heterozygous familial hypercholesterolemia; LLT, lipid-lowering therapy; PBO, placebo; Q2W, every 2 weeks. ALI 75/150 indicates studies using a dose titration strategy, whereby ALI 75 mg Q2W was increased to 150 mg Q2W at Week 12 if LDL-C at Week 8 was 70 mg/dl (or, in OPTIONS I, OPTIONS II, and ALTERNATIVE, 70 or 100 mg/dl, depending on CV risk). ALI 150 indicates studiess where patients received 150 mg Q2W from the outset. In all but two studies, patients also received background statin therapy ± other LLT (EZE was not allowed as background LLT in EZE-controlled studies). The statin was at maximally tolerated dose in 6 studies (85% of all patients in the 10 trials). Under additional populations, 2 studies were performed without background statin (ALTERNATIVE, in patients with documented statin intolerance, and MONO, a monotherapy study performed without background LLT). OPTIONS studies included patients at high CV risk receiving moderate to highh doses of potent statins. *No other non-statin LLT allowed in COMBO II. Concomitant statin and doses were atorvastatin 20 or 40 mg in OPTIONS I and rosuvastatin 10 or 20 mg in OPTIONS II. 7

51 etable 2. Baseline characteristics in Phase 3 ODYSSEY trials (randomized population) Study Group Age, years, mean (SD) Male, n (%) Race, white n (%) BMI, kg/m 2, mean (SD) HeFH, n (%) Type 2 diabetes, n (%) ASCVD, n (%) Placebocontrolled LONG TERM (MTD statin) ALI 150 (n = 1553) 60.4 (10.4) 983 (63.3) 1,441 (92.8) 30.2 (5.7) 276 (17.8) 545 (35.1) 1,187 (76.4) PBO (n = 788) 60.6 (10.4) 474 (60.2) 730 (92.6) 30.5 (5.5) 139 (17.6) 269 (34.1) 612 (77.7) HIGH FH (MTD statin) ALI 150 (n = 72) 49.8 (14.2) 35 (48.6) 64 (88.9) 28.8 (5.2) 72 (100) 9 (12.5) 32 (44.4) PBO (n = 35) 52.1 (11.2) 22 (62.9) 30 (85.7) 28.9 (4.2) 35 (100) 6 (17.1) 22 (62.9) COMBO I (MTD statin) ALI 75/150 (n = 209) 63.0 (9.5) 131 (62.7) 170 (81.3) 32.6 (6.3) 0 94 (45.0) 179 (85.6) PBO (n = 107) 63.0 (8.8) 77 (72.0) 88 (82.2) 32.0 (7.1) 0 42 (39.3) 87 (81.3) FH I (MTD statin) ALI 75/150 (n = 323) 52.1 (12.9) 180 (55.7) 300 (92.9) 29.0 (4.6) 323 (100) 32 (9.9) 154 (47.7) PBO (n = 163) 51.7 (12.3) 94 (57.7) 144 (88.3) 30.0 (5.4) 163 (100) 25 (15.3) 81 (49.7) FH II (MTD statin) ALI 75/150 (n = 167) 53.2 (12.9) 86 (51.5) 164 (98.2) 28.6 (4.6) 167 (100) 7 (4.2) 63 (37.7) PBO (n = 82) 53.2 (12.5) 45 (54.9) 80 (97.6) 27.7 (4.7) 82 (100) 3 (3.7) 32 (39.0) Ezetimibecontrolled COMBO II (MTD statin) ALI 75/150 (n = 479) 61.7 (9.4) 360 (75.2) 404 (84.3) 30.0 (5.4) (30.3) 461 (96.2) EZE (n = 241) 61.3 (9.2) 170 (70.5) 206 (85.5) 30.3 (5.1) 0 76 (31.5) 224 (92.9) OPTIONS I (atorvastatin mg) ALI 75/150 (n = 104) 63.1 (10.2) 64 (61.5) 91 (87.5) 31.2 (6.9) 12 (11.5) 58 (55.8) 59 (56.7) EZE (n = 102) 64.8 (9.6) 67 (65.7) 91 (89.2) 31.2 (5.9) 4 (3.9) 45 (44.1) 66 (64.7) OPTIONS II (rosuvastatin mg) ALI 75/150 (n = 103) 59.9 (10.2) 59 (57.3) 87 (84.5) 31.0 (6.9) 14 (13.6) 37 (35.9) 60 (58.3) 8

52 Study Group Age, years, mean (SD) Male, n (%) Race, white n (%) BMI, kg/m 2, mean (SD) HeFH, n (%) Type 2 diabetes, n (%) ASCVD, n (%) EZE (n = 101) 61.8 (10.3) 57 (56.4) 88 (87.1) 31.1 (6.4) 14 (13.9) 44 (43.1) 63 (61.8) ALTERNATIVE (no statin) ALI 75/150 (n = 126) 64.1 (9.0) 70 (55.6) 117 (92.9) 29.6 (6.6) 14 (11.1) 36 (28.6) 71 (56.3) EZE (n = 125) 62.8 (10.1) 67 (53.6) 116 (92.8) 28.4 (4.9) 25 (20.0) 24 (19.2) 58 (46.4) MONO (no statin) ALI 75/150 (n = 52) 60.8 (4.6) 28 (53.8) 46 (88.5) 30.1 (5.9) 0 3 ( 5.8) 0 EZE (n = 51) 59.6 (5.3) 27 (52.9) 47 (92.2) 28.4 (6.7) 0 1 ( 2.0) 0 ALI = alirocumab; ASCVD = atherosclerotic cardiovascular disease; atorva = stoarvastatin; BMI = body mass index; EZE = ezetimibe; LDL-C = low-density lipoprotein cholesterol; MTD, maximally tolerated dose of statin; PBO = placebo; Q2W = every 2 weeks: rosuva = rosuvastatin; SD = standard deviation. 9

53 etable 3. Baseline lipid values in individual studies (randomized population) Trial LONG TERM HIGH FH COMBO I FH I FH II COMBO II OPTIONS I OPTIONS II ALTERNATIVE MONO Group ALI n=1550 PBO n=788 ALI n=72 PBO n=35 ALI n=207 PBO n=107 ALI n=322 PBO n=163 ALI n=167 PBO n=81 ALI n=479 EZE n=241 ALI n=104 EZE n=101 ALI n=103 EZE n=101 ALI n=126 EZE n=124 ALI n=52 EZE n=51 LDL-C (42.6) (41.4) (57.9) (43.4) (29.5) (35.3) (51.1) (46.8) (41.1) (41.4) (36.5) (34.1) (36.4) 99.7 (29.2) (30.0) (45.7) (72.7) (70.9) (27.1) (24.5) non HDL-C (46.6) (45.8) (58.8) (47.6) (34.0) (39.8) (54.6) (50.6) (44.8) (43.7) (40.4) (40.4) (39.3) (35.1) (37.1) (49.8) (80.4) (82.7) (0.3) (29.7) apob (27.7) (27.3) (32.0) (28.3) 90.8 (21.4) 91.4 (24.1) (30.8) (28.5) (27.4) (23.9) 94.3 (23.2) 93.5 (23.1) 93.1 (23.7) 86.5 (20.3) 95.8 (21.5) 95.1 (26.4) (39.5) (37.4) (18.4) (19.1) ALI, alirocumab; EZE, ezetimibe; non-hdl-c, non-high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; PBO, placebo. Lipids are means (SD), mg/dl. 10

54 efigure 2. Distribution of lipid levels at baseline in pooled treatment groups according to control used in the ODYSSEY trials and overall distributions for all groups combined 11

55 12

56 Two patients with missing baseline LDL-C and 3 patients with missing apo B weree excluded from the analysis. Non-HDL-C, non-high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol. 13

57 efigure 3. Distribution of average achieved lipid levels and percent reduction in lipids from baseline during treatment for all treatment groups combined 14

58 Non-HDL-C, non-high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol. 15

59 etable 4. Average achieved lipid levels and percentage reductions during treatment in individual trials (safety population) Trial LONG TERM HIGH FH COMBO I FH I FH II COMBO II OPTIONS I OPTIONS II ALTERNATIVE MONO Group ALI n=1550 PBO n=788 ALI n=72 PBO n=35 ALI n=207 PBO n=107 ALI n=322 PBO n=163 ALI n=167 PBO n=81 ALI n=479 EZE n=241 ALI n=104 EZE n=101 ALI n=103 EZE n=101 ALI n=126 EZE n=124 ALI n=52 EZE n=51 Average achieved, mg/dl LDL-C 49.8 (34.6) (39.0) (65.0) (47.0) 52.2 (25.6) (34.9) 76.3 (41.0) (49.1) 68.2 (35.9) (41.5) 53.5 (31.4) 82.4 (33.3) 58.7 (33.4) 78.1 (32.2) 67.9 (33.1) 89.3 (42.4) (68.0) (58.3) 68.3 (20.5) (22.6) non-hdl-c 75.1 (38.1) (42.7) (65.1) (50.2) 80.1 (31.2) (38.6) 99.7 (44.3) (52.4) 91.0 (39.2) (46.3) 80.6 (34.2) (38.2) 83.4 (36.8) (36.5) 93.1 (38.2) (44.3) (70.4) (64.2) 91.1 (20.3) (26.6) apob* 51.5 (28.4) (26.3) 90.5 (36.2) (30.1) 59.5 (20.3) 91.2 (25.8) 71.4 (27.0) (29.6) 64.4 (23.7) (23.4) 59.0 (22.5) 79.6 (24.2) 58.3 (25.9) 76.3 (25.7) 67.0 (27.3) 83.8 (26.5) 90.2 (38.0) (32.2) 65.9 (14.6) 92.8 (17.2) Average % change from baseline LDL-C

60 (22.1) (27.6) (27.4) (21.5) (23.7) (17.5) (23.8) (23.8) (22.2) (16.5) (24.6) (26.3) (27.1) (20.3) (24.7) (51.7) (19.0) (16.0) (13.0) (11.0) non-hdl-c (19.4) 2.0 (22.1) (24.4) -6.3 (20.0) (20.3) -0.5 (15.8) (21.4) 8.5 (20.7) (20.1) 4.6 (16.0) (20.8) (22.2) (23.5) (18.0) (21.4) (28.9) (14.5) (11.8) (11.3) (11.2) apob (22.6) 2.0 (26.4) (24.4) -5.6 (18.5) (20.3) 1.5 (18.8) (19.9) 6.2 (19.2) (18.2) 0.8 (12.3) (20.6) (19.8) (24.7) (20.5) (25.1) -9.2 (28.9) (15.9) (12.3) (13.4) (11.8) ALI, alirocumab; EZE, ezetimibe; non-hdl-c, non-high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; PBO, placebo. Lipids are means (SD), mg/dl. For patients with no post-baseline lipid values, baseline values were used. *Two patients with missing baseline and no post-baseline apob were excluded from the analysis. Two patients with missing baseline LDL-C were excluded from the analysis. 103 patients with missing baseline apob were excluded from the analysis. 17

61 etable 5. Week 4* achieved lipid levels and percentage reductions in individual trials (safety population) Trial Group LONG TERM ALI n=1550 PBO n=788 HIGH FH COMBO I FH I FH II COMBO II OPTIONS I OPTIONS II ALTERN. MONO MACE vs achieved level ALI n=72 PBO n=35 ALI n=207 PBO n=107 ALI n=322 PBO n=163 ALI n=167 PBO n=81 ALI n=479 EZE n=241 ALI n=104 EZE n=101 ALI n=103 EZE n=101 ALI n=126 EZE n=124 ALI n=52 EZE n=51 (pool of all patients from the trials) Achieved mg/dl Category N HR (95% CI) P- value LDL-C 47.9 (37.1) (42.5) (74.3) (51.5) 52.2 (33.0) (36.0) 77.8 (43.3) (53.8) 76.2 (37.8) (43.9) 52.4 (34.4) 76.6 (31.3) 50.4 (30.6) 72.7 (30.6) 62.5 (33.9) 79.0 (36.4) (69.8) (55.9) 68.0 (20.6) (24.7) Per 39 mg/dl 4972 difference 0.79 (0.67 to 0.94) non-hdl- C 71.0 (39.6) apob 48.8 (30.9) (46.5) (73.0) 99.5 (27.6) 84.4 (36.2) (57.6) (29.3) 78.3 (37.6) 61.4 (25.5) (41.3) 93.2 (31.7) (47.8) 75.8 (30.0) (57.5) (29.5) 97.4 (43.6) 73.6 (30.2) (46.8) (24.9) 78.2 (38.2) 58.6 (24.8) (35.7) 76.8 (22.8) 73.7 (32.9) 60.4 (26.7) 96.8 (36.8) 73.3 (23.2) 86.8 (39.7) 66.7 (27.9) (38.9) 81.1 (25.5) (79.2) 91.2 (37.9) (67.0) (33.1) 91.6 (23.1) 66.3 (18.4) (29.2) 92.1 (17.8) Per 42 mg/dl decrease Per 27 mg/dl decrease (0.68 to 0.96) 0.79 (0.67 to 0.94) % change from baseline MACE vs % change LDL-C (24.3) non-hdl- C (20.1) apob (2.54) 1.3 (30.1) 0.5 (24.0) 0.4 (30.5) (27.2) (23.9) (24.4) -9.4 (19.3) -9.0 (19.0) -9.1 (16.6) (25.3) (21.0) (22.9) -1.7 (25.0) -2.4 (20.8) 3.2 (21.1) (22.9) (20.9) (20.4) 3.5 (33.3) 4.0 (29.5) 2.9 (18.3) (20.2) (18.8) (20.1) 1.2 (16.0) 0.3 (15.2) -1.0 (15.1) (26.3) (22.5) (21.8) (23.7) (21.4) (19.9) (25.4) (21.7) (23.8) (22.6) (19.4) (16.8) (23.8) (20.5) (25.9) (46.4) (24.4) (26.4) (18.2) (15.0) (19.2) (11.8) (11.0) (13.5) (12.8) (10.5) (16.7) (12.2) (12.1) (13.9) 0.74 Per 50% reduction 4972 (0.60 to 0.90) 0.71 Per 50% reduction 4974 (0.54 to 0.94) 0.75 Per 50% reduction 4871 (0.61 to 0.92)

62 ALI, alirocumab; ALTERN., ALTERNATIVE; EZE, ezetimibe; non-hdl-c, non-high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; PBO, placebo. Lipids are means (SD), mg/dl. *Week 12 data shown for apob as no Week 4 values were collected. For patients with no lipid values at Week 4 (or Week 12 for apob), baseline values were used. 14 patients with missing baseline apob and no Week 12 apob were excluded from the analysis. Two patients with missing baseline LDL-C were excluded from the analysis. 103 patients with missing baseline apob were excluded from the analysis. 19

63 etable 6. Correlation between average LDL-C, non-hdl-c and apob during the treatment period Average LDL-C Average non-hdl-c Average apob Average LDL-C during the treatment period (P<0.0001) (P<0.0001) Average non-hdl-c during the treatment period (P<0.0001) (P<0.0001) Average apob during the treatment period (P<0.0001) (P<0.0001) - Pearson correlation coefficient Non-HDL-C, non-high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol. 20

64 etable 7. Safety summary for the 10 Phase 3 trials used in this analysis* Placebo-controlled trials Ezetimibe-controlled trials Alirocumab Placebo Alirocumab Ezetimibe n (%) (n=2318) (n=1174) (n=864) (n=618) TEAEs 1851 (79.9) 954 (81.3) 657 (76.0) 457 (73.9) Treatment-emergent SAEs 385 (16.6) 202 (17.2) 147 (17.0) 86 (13.9) TEAEs leading to death 16 (0.7) 13 (1.1) 6 (0.7) 9 (1.5) TEAEs leading to discontinuation 144 (6.2) 67 (5.7) 84 (9.7) 66 (10.7) TEAEs in 5% of patients Nasopharyngitis 291 (12.6) 142 (12.1) 52 (6.0) 41 (6.6) Injection site reaction 167 (7.2) 62 (5.3) 25 (2.9) 13 (2.1) Upper respiratory tract infection 162 (7.0) 94 (8.0) 62 (7.2) 40 (6.5) Influenza 147 (6.3) 63 (5.4) 37 (4.3) 23 (3.7) Urinary tract infection 128 (5.5) 65 (5.5) 21 (2.4) 25 (4.0) Back pain 123 (5.3) 70 (6.0) 33 (3.8) 26 (4.2) Diarrhoea 123 (5.3) 57 (4.9) 30 (3.5) 21 (3.4) Headache 119 (5.1) 64 (5.5) 43 (5.0) 24 (3.9) 21

65 Arthralgia 118 (5.1) 76 (6.5) 42 (4.9) 26 (4.2) Myalgia 111 (4.8) 46 (3.9) 62 (7.2) 48 (7.8) Accidental overdose 30 (1.3) 17 (1.4) 54 (6.3) 24 (3.9) SAE, serious adverse event; TEAE, treatment-emergent adverse event. *Placebo-controlled studies: Phase 3 (LONG TERM, FH I, FH II, HIGH FH, COMBO I); ezetimibecontrolled studies: Phase 3 (COMBO II, OPTIONS I, OPTIONS II, ALTERNATIVE, MONO). Originally published by Elsevier and reproduced under STM Permissions Guidelines from: Gaudet D, Watts GF, Robinson JG, et al. Effect of Alirocumab on Lipoprotein(a) Over 1.5 Years (From the Phase 3 ODYSSEY Program). Am J Cardiol. 2016;(epub ahead of print). DOI: 22

66 efigure 4. Adjusted rate of any TEAE by average LDL-C (B) percentage reduction in LDL- during treatmentt period: (A) achieved LDL-C (multivariate analysis adjusted on baseline characteristics; pool of Phase 3 during treatment, C from baselinee studies) A B LDL-C, low-density lipoprotein cholesterol; TEAE, treatment-emergen adverse event. 23

67 Lipid 161 LDL-C 을 50mg/dL 까지감소시켜도심혈관사건이계속감소한다 : 10 개의 ODYSSEY 연구분석 초록 LDL-C(low density lipoprotein cholesterol) [major adverse cardiovascular events(mace)], ezetimibe 54mg/dL. LDL-C 50mg/dL MACE., LDL-C, non-hdl-c(non-high density lipoprotein cholesterol) apob(apolipoprotein B100) MACE ODYSSEY. ODYSSEY alirocumab ( /ezetimibe). 10 [6,699 - (patients-year) ] alirocumab 75/150mg ( 2 ). 4,974 (3,182 alirocumab, 1,174, 618 ezetimibe ). (post hoc analysis), LDL-C (%) MACE(,,, ). 33.1% LDL-C 50mg/dL (alirocumab, %; ezetimibe, 6.5%;, 0%). 104 MACE (, 36 ), LDL-C 39mg/dL MACE 24% (adjusted HR, 0.76; 95% CI, ; P=0.0025). LDL-C MACE (HR, 0.71; 95% CI, per additional 50% reduction from baseline; P=0.003). Non-HDL-C apob LDL-C MACE. 10 ODYSSEY, LDL-C, LDL-C 50mg/dL. ODYSSEY OUTCOMES.

68 reductions in atherogenic lipids and Major cardiovascular events a Pooled analysis of 10 ODYsseY trials comparing alirocumab With control Editorial, see editorial, Circulation see. 2016;134: p 1944 BACKGROUND: A continuous relationship between reductions in low-density lipoprotein cholesterol (LDL-C) and major adverse cardiovascular events (MACE) has been observed in statin and ezetimibe outcomes trials down to achieved levels of 54 mg/dl. However, it is uncertain whether this relationship extends to LDL-C levels <50 mg/dl. We assessed the relationship between additional LDL-C, non highdensity lipoprotein cholesterol, and apolipoprotein B100 reductions and MACE among patients within the ODYSSEY trials that compared alirocumab with controls (placebo/ezetimibe), mainly as add-on therapy to maximally tolerated statin. METHODS: Data were pooled from 10 double-blind trials (6699 patient-years of follow-up). Randomization was to alirocumab 75/150 mg every 2 weeks or control for 24 to 104 weeks, added to background statin therapy in 8 trials. This analysis included 4974 patients (3182 taking alirocumab, 1174 taking placebo, 618 taking ezetimibe). In a post hoc analysis, the relationship between average on-treatment lipid levels and percent reductions in lipids from baseline were correlated with MACE (coronary heart disease death, nonfatal myocardial infarction, ischemic stroke, or unstable angina requiring hospitalization) in multivariable analyses. RESULTS: Overall, 33.1% of the pooled cohort achieved average LDL-C <50 mg/dl (44.7% 52.6% allocated to alirocumab, 6.5% allocated to ezetimibe, and 0% allocated to placebo). In total, 104 patients experienced MACE (median time to event, 36 weeks). For every 39 mg/dl lower achieved LDL-C, the risk of MACE appeared to be 24% lower (adjusted hazard ratio, 0.76; 95% confidence interval, ; P=0.0025). Percent reductions in LDL-C from baseline were inversely correlated with MACE rates (hazard ratio, 0.71; 95% confidence interval, per additional 50% reduction from baseline; P=0.003). Strengths of association materially similar to those described for LDL-C were observed with achieved non high-density lipoprotein cholesterol and apolipoprotein B100 levels or percentage reductions. CONCLUSIONS: In a post hoc analysis from 10 ODYSSEY trials, greater percentage reductions in LDL-C and lower on-treatment LDL-C were associated with a lower incidence of MACE, including very low levels of LDL-C (<50 mg/dl). These findings require further validation in the ongoing prospective ODYSSEY OUTCOMES trial. CLINICAL TRIAL REGISTRATION: URL: Unique identifiers: NCT , NCT , NCT , NCT , NCT , NCT , NCT , NCT , NCT , and NCT Circulation. 2016;134: DOI: /CIRCULATIONAHA Kausik K. Ray, MD, MPhil Henry N. Ginsberg, MD Michael H. Davidson, MD Robert Pordy, MD Laurence Bessac, MD Pascal Minini, PhD Robert H. Eckel, MD Christopher P. Cannon, MD Correspondence to: Kausik K. Ray, MD, MPhil, Imperial Centre for Cardiovascular Disease Prevention, Department of Primary Care and Public Health, School of Public Health, Imperial College, Reynolds Bldg, St. Dunstan s Rd, London, W6 8RP UK. k.ray@imperial.ac.uk Sources of Funding, see page Key Words: apolipoproteins cardiovascular diseases cholesterol, LDL risk 2016 The Authors. Circulation is published on behalf of the American Heart Association, Inc., by Wolters Kluwer Health, Inc. This is an open access article under the terms of the Creative Commons Attribution Non-Commercial- NoDervis License, which permits use, distribution, and reproduction in any medium, provided that the original work is properly cited, the use is noncommercial, and no modifications or adaptations are made. clinical Perspective What is new? Cardiovascular benefits of statins and add-on lipidlowering therapy extend only to low-density lipoprotein cholesterol (LDL-C) levels of 54 mg/dl. We investigated whether this relationship extends below 50 mg/dl using data from the ODYSSEY trials of alirocumab (proprotein convertase subtilisin/ kexin type 9 monoclonal antibody) versus placebo/ ezetimibe. About half of alirocumab-treated patients achieved LDL-C <50 mg/dl. For each 39 mg/dl lower achieved LDL-C, MACE incidence fell by 24%, and 50% reductions in LDL-C from baseline reduced MACE by 29%. Similar associations were observed with non highdensity lipoprotein cholesterol and apolipoprotein B achieved levels, percentage reductions, and MACE. Low LDL-C levels were not associated with excess treatment-emergent adverse events. What are the clinical implications? These analyses provide further reassurance about the safety and cardiovascular benefit of achieving even further reductions in LDL-C with alirocumab beyond what was previously achieved with statins and ezetimibe. Limitations include the low number of events (104), the limited duration of the studies ( weeks), and the post hoc nature of this analysis. If the forthcoming outcomes trials of PCSK9 inhibitors such as alirocumab demonstrate additional reduction in MACE with further LDL-C reduction, then guideline committees may investigate lower LDL-C goals or a larger reduction in LDL-C from untreated baseline for those at highest risk of MACE. statins have been the cornerstone of global lipid modification guidelines for the prevention of major adverse cardiovascular events (MACE). 1 With accumulating evidence, each subsequent iteration of clinical guidelines has extended statin indications and recommended the achievement of progressively lower lowdensity lipoprotein cholesterol (LDL-C) goals 2 4 or, most recently, percentage reductions in LDL-C with the most potent statins. 5 Recently, the IMPROVE-IT trial (Improved Reduction of Outcomes: Vytorin Efficacy International Trial) demonstrated that further reductions in LDL-C to a mean of 54 mg/dl with a nonstatin lipid-lowering therapy in statin-treated individuals provided additional reductions in MACE, 6 supporting the concept that lower LDL-C levels are better and consistent with observational data from several statin trials. 7 9 Inhibitors of proprotein convertase subtilisin/kexin type 9 (PCSK9) offer the prospect of achieving even lower LDL-C levels than ezetimibe when added to statins. Data from the phase 3 ODYSSEY trial program, in which average achieved LDL-C levels of <50 mg/dl were observed when alirocumab was added to statin therapy, allow the exploration of changes in event rates at much lower LDL-C levels than previously analyzed. In addition, percent reduction in LDL-C has been cited as being important in some guidelines, 5,13,14 and because patients in the ODYSSEY trials have already had their LDL-C reduced by 30% to 50% with statins, these trials allow us to explore the relationship between further percentage reductions in LDL-C and risk. Therefore, we assessed whether the relationship between LDL-C and risk extends to very low levels of LDL-C (<50 mg/dl) in the ODYSSEY trial program. We hypothesized that, among patients already receiving maximally tolerated statin therapy, both lower achieved LDL-C and greater percentage reductions from baseline would translate into lower rates of MACE. Furthermore, because many guidelines increasingly recommend non high-density lipoprotein cholesterol (non HDL-C) or apolipoprotein B100 (apob) as potential alternatives to LDL-C for assessing the efficacy of lipid-lowering therapy, we provide similar analyses using these lipid parameters for comparison. 4,15,16 MethODs Patient Population The phase 3 ODYSSEY trial designs have been reported previously. 6,11,12,17 22 Patients were enrolled if they had established atherosclerotic cardiovascular disease (ASCVD) or high cardiovascular risk such as heterozygous familial hypercholesterolemia with LDL-C inadequately controlled on their existing treatment (statin/other lipid-lowering therapy/diet). The main exclusion criteria were baseline LDL-C <70 mg/dl for those with ASCVD and very high risk and <100 mg/dl for high-risk patients without ASCVD at screening. Individuals with triglyceride levels >400 mg/dl were excluded (for further details on inclusion and exclusion criteria, see Table I in the onlineonly Data Supplement). For the present analysis, data were pooled from 10 phase 3 ODYSSEY trials (Figure I in the onlineonly Data Supplement). Patients were randomized to receive alirocumab or control (placebo or ezetimibe) with double-blind treatment periods of 24 to 104 weeks. Six of the 10 studies, representing 80% of the population, had a minimum study duration of 52 weeks. All study protocols were approved by the relevant local independent review boards, and all participating patients provided written informed consent. lipid Measurements LDL-C was calculated with the Friedewald equation unless triglycerides were >400 mg/dl, when it was determined by β quantification. ApoB levels in serum were determined from immunonephelometry by a central laboratory (Medpace Reference Laboratories, Cincinnati, OH, and Leuven, Belgium, except for the LONG TERM study, which used Covance Central

69 Laboratories, Indianapolis, IN, and Geneva, Switzerland) and non HDL-C via subtraction of HDL-C from total cholesterol. Mace Definitions MACE were defined as per the primary end point of the ODYSSEY OUTCOMES study 23 : coronary heart disease death, nonfatal myocardial infarction, ischemic stroke, or unstable angina requiring hospitalization. Cardiovascular events were adjudicated by a central Clinical Events Committee 12 (The same committee is involved in the ODYSSEY OUTCOMES trial 23 ). Unstable angina considered here was limited to those with definite evidence of the ischemic condition; that is a small proportion of unstable angina events qualified (see the definition of unstable angina in the online-only Data Supplement). statistical analysis Baseline Characteristics and Distribution of Lipid Parameters Baseline data were pooled for all randomized patients and presented stratified according to whether the studies were placebo controlled or ezetimibe controlled. For continuous variables, the data are reported as mean and SD or median and interquartile range if they were not normally distributed. The distribution of LDL-C, non HDL-C, and apob levels at baseline and the average on-treatment levels or average percentage reductions in these parameters during the study treatment period are depicted graphically and analyzed with descriptive statistics comparing treatment groups (alirocumab versus placebo or versus ezetimibe).these include only patients in the safety population, that is patients who were randomized and received at least 1 dose or part of a dose of study treatment. lipid changes and risk of Mace Regardless of treatment allocation, patients were pooled into 1 overall cohort, and the relationship between LDL-C and MACE during the treatment period was assessed with achieved LDL-C levels during treatment and percentage reductions in LDL-C from baseline. Average on-treatment LDL-C or the mean percentage reduction during the treatment period was determined from the area under the curve (using the trapezoidal method), taking into account all LDL-C values up to end of the treatment period or the occurrence of MACE, whichever came first. The relation between on-treatment LDL-C and MACE was assessed with a multivariable Cox regression model with adjustment for age, sex, diabetes mellitus, history of myocardial infarction or stroke, baseline LDL-C, and smoking status, as previously published. 24 Hazard ratios (HRs) and 95% confidence intervals (CI) were calculated for every 39 mg/dl lower LDL-C to provide a comparison with the CTT (Cholesterol Treatment Trialists Collaboration) meta-regression line. 25 Similar analyses were conducted for the percentage reduction in LDL-C from baseline and subsequent risk of MACE with HRs and 95% CIs expressed for each 50% reduction in LDL-C. To assess the shape of association, the adjusted rates of MACE and associated 95% CIs were determined from a multivariate Poisson model and depicted graphically as a function of average LDL-C levels or average percentage reduction during treatment. Instead of using all available lipid measurements (average LDL-C levels or reductions), we conducted sensitivity analyses with only LDL-C values and percentage reduction at week 4 and subsequent events after the exclusion of events that occurred before week 4. Last, analyses similar to those described above were conducted for non HDL-C or apob. All analyses were generated with SAS version 9.4, and all tests and CIs were 2 sided. safety analysis For safety, treatment-emergent adverse events (TEAEs) were defined as those events occurring from the first dose of study treatment up to 70 days after the last dose. The principal analyses compared randomized treatment with alirocumab and control. We also explored the strength of association between any TEAE and LDL-C levels or percentage reductions in LDL-C using multivariable logistic regression, after adjusting for the same covariates included in the MACE analyses. Data are reported as odds ratio and 95% CI per 39-mg/dL difference or 50% reduction in LDL-C. The shape of the association was also depicted graphically with adjusted TEAE rate and associated 95% CI plotted against average LDL-C levels or percentage LDL-C reductions derived from multivariate logistic regression models. results Baseline characteristics The baseline characteristics of individual studies are shown in Table II in the online-only Data Supplement, and the pooled summary of placebo and ezetimibe comparator trials is shown in Table 1. In the placebocontrolled trials, 2318 patients were treated with alirocumab and 1174 were treated with placebo; in the ezetimibe-controlled trials, 864 were treated with alirocumab and 618 were treated with ezetimibe. Hence, a total of 4974 patients were included in the lipid and MACE analyses described below. Overall, the average age was 60 years, with patients being mostly white and having an average body mass index of 30 kg/ m 2. Approximately two thirds were male; one third had diabetes mellitus; two thirds had a history of ASCVD; and about one fifth were smokers. One third of participants in the placebo-controlled trials had heterozygous familial hypercholesterolemia. Baseline lipids Baseline lipid values for individual studies are reported in Table III in the online-only Data Supplement. Pooled mean baseline LDL-C levels ranged from to mg/dl; non HDL-C, between and mg/dl; and apob, between and mg/dl (Table 1). As expected, the distribution of baseline lipids at randomization was fairly similar between the alirocumab and control groups (Figure II in the online-only Data Supplement). table 1. Baseline characteristics Placebo-controlled trials ezetimibe-controlled trials* alirocumab (n=2324) Placebo (n=1175) alirocumab (n=864) ezetimibe (n=620) Age, mean (SD), y 58.7 (11.6) 58.8 (11.4) 61.9 (9.4) 62.1 (9.5) Sex, male, n (%) 1415 (60.9) 712 (60.6) 581 (67.2) 388 (62.6) Race, white, n (%) 2139 (92.0) 1072 (91.2) 745 (86.2) 548 (88.4) BMI, mean (SD), kg/m (5.6) 30.3 (5.6) 30.2 (6.0) 30.0 (5.7) HeFH, n (%) 838 (36.1) 419 (35.7) 40 (4.6) 43 (6.9) Diabetes mellitus, n (%) 699 (30.1) 355 (30.2) 283 (32.8) 192 (31.0) ASCVD, n (%) 1615 (69.5) 834 (71.0) 651 (75.3) 411 (66.3) CHD 1454 (62.6) 766 (65.2) 611 (70.7) 390 (62.9) Ischemic stroke/tia 199 (8.6) 86 (7.3) 67 (7.8) 42 (6.8) PAD 97 (4.2) 56 (4.8) 33 (3.8) 19 (3.1) Current smoker, n (%) 453 (19.5) 231 (19.7) 146 (16.9) 118 (19.0) Statin intensity, n (%) High 1327 (57.1) 682 (58.0) 430 (49.8) 265 (42.7) Moderate 650 (28.0) 314 (26.7) 208 (24.1) 152 (24.5) Low 344 (14.8) 177 (15.1) 47 (5.4) 27 (4.4) No statins (20.6) 176 (28.4) Missing 3 (0.1) 2 (0.2) 1 (0.1) 0 Baseline lipids, mean (SD), mg/dl LDL-C (46.3) (44.8) (51.5) (56.9) Non HDL-C (49.6) (48.5) (57.2) (64.1) ApoB (29.0) (28.5) (30.7) (32.2) ApoB indicates apolipoprotein B100; ASCVD, atherosclerotic cardiovascular disease; BMI, body mass index; CHD, coronary heart disease; HeFH, heterozygous familial hypercholesterolemia; LDL-C, low-density lipoprotein cholesterol; non HDL-C, non high-density lipoprotein cholesterol; PAD, peripheral artery disease, and TIA, transient ischemic attack. Pooled data of all randomized patients from the 10 trials included in this analysis. *In combination with statins or not. Patients may be in >1 category. Atorvastatin 40 to 80 mg, rosuvastatin 20 to 40 mg, or simvastatin 80 mg. Atorvastatin 20 to <40 mg, rosuvastatin 10 to <20 mg, or simvastatin 40 to <80 mg. Atorvastatin <20 mg, rosuvastatin <10 mg, or simvastatin <40 mg. lipid levels During treatment Figure 1 illustrates the distribution of different lipid parameters during treatment. The mean LDL-C levels achieved during treatment were as follows: in placebocontrolled studies, 56.9 and mg/dl among those treated with alirocumab and placebo, and in ezetimibecontrolled studies, 64.0 and mg/dl in those treated with alirocumab and ezetimibe, respectively (Table 2; individual trial data in Table IV in the online-only Data Supplement). Corresponding values for non HDL-C and apob are shown in Table 2. Similar results were obtained when week 4 achieved lipids were used instead of average levels throughout the entire study duration (Table V in the online-only Data Supplement). Overall, 33.1% of patients achieved an average LDL- C <50 mg/dl during treatment. In placebo-controlled studies, 52.6% of alirocumab-treated and 0% of placebo-treated patients achieved LDL-C levels <50 mg/dl. In the ezetimibe-controlled studies, the corresponding figures were 44.7% for the alirocumab and 6.5% for the ezetimibe arm, respectively. Thus, when all patients were pooled, the overall distribution of each lipid parameter during treatment largely reflected the greater proportion of patients achieving very low levels of LDL-C, non HDL-C, and apob in the alirocumab group (Figure III in the online-only Data Supplement). Percent reductions in lipids From Baseline Figure 2 depicts the distribution of the average percentage change from baseline in LDL-C, non HDL-C, and apob during the trials. The average percentage change in LDL-C from baseline during treatment was 55.4% for

70 in non HDL-C and 27-mg/dL difference in apob in the present pooled data sets. For each 42 mg/dl lower non HDL-C, the HR was 0.77 (95% CI, ; P=0.0056; Table 2). The corresponding HR for each 27 mg/dl lower apob was 0.72 (95% CI, ; P=0.0002; Table 2). Percentage reductions in lipids and Mace LDL-C percent reduction was inversely correlated with MACE rates (Figure 3; HR, 0.71; 95% CI, per additional 50% reduction in LDL-C; P=0.003; Table 2). Similarly the risk of MACE was lower with greater percent reductions from baseline in both non HDL-C (Figure 3; HR, 0.71; 95% CI, per 50% reduction; P=0.0323) and apob (Figure 3; HR, 0.68; 95% CI, per 50% reduction; P=0.0008; Table 2). Qualitatively similar results were observed with the use of a single week 4 measurement of LDL-C or non HDL-C and week 12 apob instead of average values throughout the trial (Table V in the online-only Data Supplement). safety Overall incidences of TEAEs, serious TEAEs, deaths, and discontinuations as a result of TEAEs were similar between alirocumab and control patients within the pools of placebo- and ezetimibe-controlled studies (Table VII in the online-only Data Supplement). A higher rate of injection site reactions, mostly mild in intensity and selflimiting, was observed with alirocumab compared with controls. Analyses comparing the relationship between a 39 mg/dl lower LDL-C and odds of any TEAE were not significant (odds ratio, 1.02; 95% CI, ; Figure IV in the online-only Data Supplement), nor was there any significant association between a 50% lowering of LDL-C and odds of any TEAE (odds ratio, 1.02; 95% CI: ; Figure IV in the online-only Data Supplement). Figure 1. Distribution of achieved levels of low-density lipoprotein cholesterol (ldl-c; a), non high-density lipoprotein cholesterol (non hdl-c; B), and apolipoprotein B100 (apob; c) during treatment stratified by control group. Figure 1 continued. For patients with no postbaseline lipid measurement, baseline values were used. CI indicates confidence interval; and MACE, major adverse cardiovascular event. achieved LDL-C levels (Figure 3; adjusted HR, 0.76; 95% alirocumab and 2.7% for placebo in placebo-controlled trials and 48.1% with alirocumab and 18.0% with ezetimibe in ezetimibe-controlled trials (Table 2; individual trial data in Table IV in the online-only Data Supplement). Corresponding results for non HDL-C and apob are shown in Table 2 and Table IV in the online-only Data Supplement. When week 4 percentage reductions were assessed rather than the average reductions over the course of the trial, similar values were observed (Table V in the online-only Data Supplement). The combined distribution plot (Figure III in the online-only Data Supplement) reflected mainly the greater reductions in lipids achieved with alirocumab versus the relatively modest reductions observed with ezetimibe and with placebo-treated patients largely remaining unchanged from baseline. On treatment lipid levels and Mace A total of 104 first MACE were reported: 20 coronary heart disease deaths, 64 nonfatal myocardial infarctions, 16 ischemic strokes, and 4 unstable angina episodes occurred (median time to event, 36 weeks) among 4974 patients treated during a total of 6699 patient-years of follow-up. A lower risk of MACE was observed with lower CI, per 39 mg/dl lower achieved LDL-C; P=0.0025; Table 2). Similar results were obtained with the use of a single week 4 LDL-C measurement instead of average levels throughout the trial (Table V in the online-only Data Supplement). In pairwise comparisons of LDL-C, non HDL-C, and apob, there was a strong and significant correlation between levels of each lipid parameter (all correlation coefficients >0.9; P<0.0001; Table VI in the online-only Data Supplement). As with achieved average LDL-C, lower (average) achieved non HDL-C and apob levels were associated with a lower risk of MACE (Figure 3). A 39-mg/dL difference in LDL-C corresponds to a 42-mg/dL difference DiscussiOn At present, all global guidelines for ASCVD risk reduction focus on optimization of statin therapy as the first option for reducing LDL-C for those at high risk. 4,5,13 The therapeutic limits of statins and the clinical scenarios in which they have been tested have therefore established the boundaries of contemporary clinical guidelines and the recommendations they have set, whether an LDL-C goal or a percentage reduction in LDL-C. On the basis of randomized, clinical trial data of intensive versus standard statin therapy, 7,8,25 knowledge of the distribution of LDL- C levels in general populations, 26,27 and what is achievable on average with the most potent statins, guidelines such as the updated Adult Treatment Panel III and those from the European Society of Cardiology/European Atherosclerosis Society recommended pragmatic goals for LDL-C of <70 mg/dl for those at highest ASCVD risk. 2,4