Long-term Safety and Efficacy of Achieving Very Low Levels of Low-Density Lipoprotein Cholesterol : A Prespecified Analysis of the IMPROVE-IT Trial | Acute Coronary Syndromes | JAMA Cardiology | JAMA Network
[Skip to Navigation]
Figure 1.  Distribution of Achieved Calculated Low-Density Lipoprotein Cholesterol (LDL-C) Level at 1 Month Among Patients Who Did Not Have a Primary Efficacy or Prespecified Safety Event Prior to the Sample
Distribution of Achieved Calculated Low-Density Lipoprotein Cholesterol (LDL-C) Level at 1 Month Among Patients Who Did Not Have a Primary Efficacy or Prespecified Safety Event Prior to the Sample

The median LDL-C level was 56 mg/dL (interquartile range, 43-70 mg/dL). To convert LDL-C to millimoles per liter, multiply by 0.0259.

Figure 2.  Median Low-Density Lipoprotein Cholesterol (LDL-C) Level During the Trial Stratified by Achieved LDL-C Level at 1 Month
Median Low-Density Lipoprotein Cholesterol (LDL-C) Level During the Trial Stratified by Achieved LDL-C Level at 1 Month

The time-weighted mean LDL-C concentrations after randomization were 34.4, 48.3, 63.3, and 79.9 mg/dL for the 4 groups (<30, 30-49, 50-69, and ≥70 mg/dL, respectively) as identified by LDL-C level at 1 month. QE indicates qualifying event (at time of admission); R, randomization. To convert LDL-C to millimoles per liter, multiply by 0.0259.

Figure 3.  Safety Events by Achieved Low-Density Lipoprotein Cholesterol (LDL-C) Level at 1 Month
Safety Events by Achieved Low-Density Lipoprotein Cholesterol (LDL-C) Level at 1 Month

The adjusted risks are shown for prespecified safety events by the LDL-C level achieved at 1 month, with the group achieving the highest LDL-C level (≥70 mg/dL) as the referent. To convert LDL-C to millimoles per liter, multiply by 0.0259. ALT indicates alanine aminotransferase; AST, aspartate aminotransferase; CK, creatine kinase; HR, hazard ratio; ULN, upper limit of normal.

Table 1.  Baseline Characteristics by Achieved LDL-C Level at 1 Month
Baseline Characteristics by Achieved LDL-C Level at 1 Month
Table 2.  Safety Events by Achieved LDL-C Level at 1 Montha
Safety Events by Achieved LDL-C Level at 1 Montha
1.
Cannon  CP, Blazing  MA, Giugliano  RP,  et al; IMPROVE-IT Investigators.  Ezetimibe added to statin therapy after acute coronary syndromes.  N Engl J Med. 2015;372(25):2387-2397.PubMedGoogle ScholarCrossref
2.
Kritchevsky  SB.  Dietary lipids and the low blood cholesterol-cancer association.  Am J Epidemiol. 1992;135(5):509-520.PubMedGoogle ScholarCrossref
3.
Neaton  JD, Blackburn  H, Jacobs  D,  et al; Multiple Risk Factor Intervention Trial Research Group.  Serum cholesterol level and mortality findings for men screened in the Multiple Risk Factor Intervention Trial.  Arch Intern Med. 1992;152(7):1490-1500.PubMedGoogle ScholarCrossref
4.
Tirschwell  DL, Smith  NL, Heckbert  SR, Lemaitre  RN, Longstreth  WT  Jr, Psaty  BM.  Association of cholesterol with stroke risk varies in stroke subtypes and patient subgroups.  Neurology. 2004;63(10):1868-1875.PubMedGoogle ScholarCrossref
5.
Berry  PH, MacDonald  JS, Alberts  AW,  et al.  Brain and optic system pathology in hypocholesterolemic dogs treated with a competitive inhibitor of 3-hydroxy-3-methylglutaryl coenzyme A reductase.  Am J Pathol. 1988;132(3):427-443.PubMedGoogle Scholar
6.
Cannon  CP, Giugliano  RP, Blazing  MA,  et al; IMPROVE-IT Investigators.  Rationale and design of IMPROVE-IT (IMProved Reduction of Outcomes: Vytorin Efficacy International Trial): comparison of ezetimbe/simvastatin versus simvastatin monotherapy on cardiovascular outcomes in patients with acute coronary syndromes.  Am Heart J. 2008;156(5):826-832.PubMedGoogle ScholarCrossref
7.
Robinson  JG, Rosenson  RS, Farnier  M,  et al.  Safety of very low low-density lipoprotein cholesterol levels with alirocumab: pooled data from randomized trials.  J Am Coll Cardiol. 2017;69(5):471-482.PubMedGoogle ScholarCrossref
8.
Hsia  J, MacFadyen  JG, Monyak  J, Ridker  PM.  Cardiovascular event reduction and adverse events among subjects attaining low-density lipoprotein cholesterol <50 mg/dl with rosuvastatin: the JUPITER Trial (Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin).  J Am Coll Cardiol. 2011;57(16):1666-1675.PubMedGoogle ScholarCrossref
9.
Friedewald  WT, Levy  RI, Fredrickson  DS.  Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge.  Clin Chem. 1972;18(6):499-502.PubMedGoogle Scholar
10.
Blazing  MA, Giugliano  RP, de Lemos  JA,  et al.  On-treatment analysis of the Improved Reduction of Outcomes: Vytorin Efficacy International Trial (IMPROVE-IT).  Am Heart J. 2016;182(182):89-96.PubMedGoogle ScholarCrossref
11.
Morris  PB, Ballantyne  CM, Birtcher  KK, Dunn  SP, Urbina  EM.  Review of clinical practice guidelines for the management of LDL-related risk.  J Am Coll Cardiol. 2014;64(2):196-206.PubMedGoogle ScholarCrossref
12.
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(11):1682-1689.PubMedGoogle ScholarCrossref
13.
Sabatine  MS, Giugliano  RP, Wiviott  SD,  et al; Open-Label Study of Long-Term Evaluation against LDL Cholesterol (OSLER) Investigators.  Efficacy and safety of evolocumab in reducing lipids and cardiovascular events.  N Engl J Med. 2015;372(16):1500-1509.PubMedGoogle ScholarCrossref
14.
Robinson  JG, Farnier  M, Krempf  M,  et al; ODYSSEY LONG TERM Investigators.  Efficacy and safety of alirocumab in reducing lipids and cardiovascular events.  N Engl J Med. 2015;372(16):1489-1499.PubMedGoogle ScholarCrossref
15.
Thompson  PD, Panza  G, Zaleski  A, Taylor  B.  Statin-associated side effects.  J Am Coll Cardiol. 2016;67(20):2395-2410.PubMedGoogle ScholarCrossref
16.
Giugliano  RP, Sabatine  MS.  Are PCSK9 inhibitors the next breakthrough in the cardiovascular field?  J Am Coll Cardiol. 2015;65(24):2638-2651.PubMedGoogle ScholarCrossref
17.
Luo  L, Yuan  X, Huang  W,  et al.  Safety of coadministration of ezetimibe and statins in patients with hypercholesterolaemia: a meta-analysis.  Intern Med J. 2015;45(5):546-557.PubMedGoogle ScholarCrossref
18.
Amarenco  P, Bogousslavsky  J, Callahan  A  III,  et al; Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) Investigators.  High-dose atorvastatin after stroke or transient ischemic attack.  N Engl J Med. 2006;355(6):549-559.PubMedGoogle ScholarCrossref
19.
Athyros  VG, Tziomalos  K, Karagiannis  A, Wierzbicki  AS, Mikhailidis  DP.  Aggressive statin treatment, very low serum cholesterol levels and haemorrhagic stroke: is there an association?  Curr Opin Cardiol. 2010;25(4):406-410.PubMedGoogle ScholarCrossref
20.
Baigent  C, Blackwell  L, Emberson  J,  et al; Cholesterol Treatment Trialists’ (CTT) Collaboration.  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(9753):1670-1681.PubMedGoogle ScholarCrossref
21.
LaRosa  JC, Grundy  SM, Kastelein  JJ, Kostis  JB, Greten  H; Treating to New Targets (TNT) Steering Committee and Investigators.  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(5):747-752.PubMedGoogle ScholarCrossref
22.
Boekholdt  SM, Hovingh  GK, Mora  S,  et al.  Very low levels of atherogenic lipoproteins and the risk for cardiovascular events: a meta-analysis of statin trials.  J Am Coll Cardiol. 2014;64(5):485-494.PubMedGoogle ScholarCrossref
23.
Wiviott  SD, Cannon  CP.  The safety and efficacy of achieving very low LDL-cholesterol concentrations with high dose statin therapy.  Curr Opin Lipidol. 2006;17(6):626-630.PubMedGoogle ScholarCrossref
24.
Wiviott  SD, Cannon  CP, Morrow  DA, Ray  KK, Pfeffer  MA, Braunwald  E; PROVE IT-TIMI 22 Investigators.  Can low-density 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(8):1411-1416.PubMedGoogle ScholarCrossref
25.
Leeper  NJ, Ardehali  R, deGoma  EM, Heidenreich  PA.  Statin use in patients with extremely low low-density lipoprotein levels is associated with improved survival.  Circulation. 2007;116(6):613-618.PubMedGoogle ScholarCrossref
26.
Hochholzer  W, Giugliano  RP.  Lipid lowering goals: back to nature?  Ther Adv Cardiovasc Dis. 2010;4(3):185-191.PubMedGoogle ScholarCrossref
27.
Dietschy  JM, Turley  SD.  Thematic review series: brain lipids: cholesterol metabolism in the central nervous system during early development and in the mature animal.  J Lipid Res. 2004;45(8):1375-1397.PubMedGoogle ScholarCrossref
28.
Zhao  Z, Tuakli-Wosornu  Y, Lagace  TA,  et al.  Molecular characterization of loss-of-function mutations in PCSK9 and identification of a compound heterozygote.  Am J Hum Genet. 2006;79(3):514-523.PubMedGoogle ScholarCrossref
29.
Brown  MS, Goldstein  JL.  A receptor-mediated pathway for cholesterol homeostasis.  Science. 1986;232(4746):34-47.PubMedGoogle ScholarCrossref
Original Investigation
May 2017

Long-term Safety and Efficacy of Achieving Very Low Levels of Low-Density Lipoprotein Cholesterol : A Prespecified Analysis of the IMPROVE-IT Trial

Author Affiliations
  • 1Cardiovascular Division, Department of Medicine, Brigham and Women’s Hospital, TIMI Study Group, Boston, Massachusetts
  • 2Duke Clinical Research Institute, Durham, North Carolina
  • 3Policlinico San Matteo, Pavia, Italy
  • 4Merck & Co, Kenilworth, New Jersey
JAMA Cardiol. 2017;2(5):547-555. doi:10.1001/jamacardio.2017.0083
Key Points

Question  What are the safety and efficacy outcomes with longer-term achievement of very low levels of low-density lipoprotein cholesterol (LDL-C) as compared with less intensive LDL-C level lowering in patients after acute coronary syndrome?

Findings  In a prespecified analysis of 15 281 patients enrolled in IMPROVE-IT, patients achieving an LDL-C level of less than 30 mg/dL at 1 month after acute coronary syndrome had a similar safety profile (and numerically the lowest rate of cardiovascular events) over a 6-year period as compared with patients achieving higher LDL-C concentrations.

Meaning  Intensive lipid-lowering therapy following acute coronary syndrome to levels well below current guideline recommendations is safe and more effective than the current target of 70 mg/dL.

Abstract

Importance  In the Improved Reduction of Outcomes: Vytorin Efficacy International Trial, intensive low-density lipoprotein cholesterol (LDL-C)–reducing therapy with ezetimibe/simvastatin compared with simvastatin alone was associated with a significant reduction in cardiovascular events in 18 144 patients after acute coronary syndrome. The safety of very low LDL-C levels over the long-term is unknown.

Objective  To assess the safety and clinical efficacy of achieving a very low (<30 mg/dL) level of LDL-C at 1 month using data from the Improved Reduction of Outcomes: Vytorin Efficacy International Trial.

Design, Setting, and Participants  This prespecified analysis compared outcomes in patients stratified by achieved LDL-C level at 1 month in the Improved Reduction of Outcomes: Vytorin Efficacy International Trial and adjusted for baseline characteristics during 6 years’ median follow-up. Patients were enrolled from October 26, 2005, to July 8, 2010, and the data analysis was conducted from December 2014 to February 2017.

Main Outcomes and Measures  Safety end points included adverse events leading to drug discontinuation; adverse muscle, hepatobiliary, and neurocognitive events; and hemorrhagic stroke, heart failure, cancer, and noncardiovascular death. Efficacy events were as specified in the overall trial.

Results  Among the 15 281 patients included in the study, 11 645 (76.2%) were men and the median age was 63 years (interquartile range, 56.6-70.7 years). In these patients without an event in the first month, the achieved LDL-C values at 1 month were less than 30 mg/dL, 30 to 49 mg/dL, 50 to 69 mg/dL, and 70 mg/dL or greater in 6.4%, 31%, 36%, and 26% of patients, respectively. Patients with LDL-C values less than 30 mg/dL (median, 25 mg/dL; interquartile range, 21-27 mg/dL) at 1 month were more likely randomized to ezetimibe/simvastatin (85%), had lower baseline LDL-C values, and were more likely older, male, nonwhite, diabetic, overweight, statin naive, and presenting with a first myocardial infarction. After multivariate adjustment, there was no significant association between the achieved LDL-C level and any of the 9 prespecified safety events. The adjusted risk of the primary efficacy composite of cardiovascular death, major coronary events, or stroke was significantly lower in patients achieving an LDL-C level less than 30 mg/dL at 1 month (adjusted hazard ratio, 0.79; 95% CI, 0.69-0.91; P = .001) compared with 70 mg/dL or greater.

Conclusions and Relevance  Patients achieving an LDL-C level less than 30 mg/dL at 1 month had a similar safety profile (and numerically the lowest rate of cardiovascular events) over a 6-year period compared with patients achieving higher LDL-C concentrations. These data provide reassurance regarding the longer-term safety and efficacy of the continuation of intensive lipid-lowering therapy in very higher-risk patients resulting in very low LDL-C levels.

Trial Registration  clinicaltrials.gov Identifier: NCT00202878

Introduction

The Improved Reduction of Outcomes: Vytorin Efficacy International Trial (IMPROVE-IT) demonstrated that the combination of ezetimibe, 10 mg, and simvastatin that achieved a median low-density lipoprotein cholesterol (LDL-C) level of 54 mg/dL (to convert to millimoles per liter, multiply by 0.0259) improved cardiovascular outcomes in patients with a recent acute coronary syndrome (ACS) compared with simvastatin monotherapy (median LDL-C level, 70 mg/dL).1 These clinical benefits were achieved without an excess in adverse safety events. However, there are concerns that lowering cholesterol to very low levels may be disadvantageous because cholesterol has a central role in hepatic bile production, is a main component of cell membranes and intracellular structures, and serves as a precursor in the biosynthesis of some vitamins, steroids, and sex hormones. Early epidemiologic studies showed an association between low cholesterol level and increased risk for cancer, intracranial hemorrhage, and death.2-4 Furthermore, studies in canine models raised concerns that supratherapeutic doses of statins may cause brain and optic pathology.5

Our goal was to assess the safety and clinical efficacy of a very low (<30 mg/dL) achieved LDL-C level at 1 month using data from IMPROVE-IT.

Methods

IMPROVE-IT1,6 was a randomized, double-blind, placebo-controlled trial that enrolled 18 144 patients hospitalized within the preceding 10 days with ACS with a LDL-C level of 50 to 100 mg/dL (if taking a prior prescription lipid-lowering therapy) or 50 to 125 mg/dL (otherwise). Patients were randomized to receive either ezetimibe, 10 mg, plus simvastatin, 40 mg, or matching placebo plus simvastatin, 40 mg, once daily, and they were followed up for a median of 6 years (interquartile range, 4.3-7.1 years). Key exclusion criteria were creatinine clearance less than 30 mL/min/1.73 m2 (to convert to milliliters per second per square meter, multiply by 0.0167), active liver disease, clinical instability, or treatment with a lipid-lowering regimen more potent than simvastatin, 40 mg. All patients provided written informed consent and the protocol was approved by ethics committees at each participating center. The study was approved by the Brigham and Women’s Hospital institutional review board. Data analyses for this current study were conducted from December 2014 to February 2017.

Prespecified safety end points included abnormal elevations of liver enzyme and creatine kinase levels, myopathy, rhabdomyolysis, adverse hepatobiliary events, and cancer. Additional prespecified safety end points included adverse events leading to study drug discontinuation, heart failure leading to hospitalization, and noncardiovascular death. A post hoc analysis explored the relationship between cataract-related adverse events and achieved LDL-C level in light of the recent report of an association between an LDL-C level less than 25 mg/dL and increased incidence of cataracts.7 Investigators reported other adverse events as verbatim terms that were then mapped to the Medical Dictionary for Regulatory Activities preferred terms8 with overreading by trained coders. Neurocognitive events were identified by 2 of the authors (R.P.G. and M.A.B.) blinded to treatment assignment and all LDL-C measurements prior to performing the current analysis, and included the following event terms: altered state of consciousness, amnesia, aphasia (not related to a stroke or transient ischemic attack), aphonia, cognitive disorder, dementia, altered or depressed consciousness, encephalopathy, lethargy, or impairment of memory.

The primary efficacy end point was a composite of cardiovascular death, myocardial infarction, unstable angina requiring hospitalization, coronary revascularization after 30 days, and stroke (both hemorrhagic and ischemic). Efficacy end points (except revascularization), muscle-related events, and cancer were adjudicated by independent committees who were unaware of treatment assignment using prespecified definitions.1

Follow-up visits occurred at 30 days, 4 months, and every 4 months thereafter. Blood specimens were drawn at the visits through 12 months and annually thereafter for patients who continued treatment with the study drug, and were sent to a central core laboratory for analysis. Patients who prematurely stopped taking the study drug had a final blood draw at the time of discontinuation and thereafter were followed up for clinical events. The LDL-C values were calculated using the Friedewald et al9 equation; however, if the triglyceride levels exceeded 400 mg/dL (to convert to millimoles per liter, multiply by 0.0113), LDL-C and high-density lipoprotein cholesterol levels were measured directly using beta-quantification. If the LDL-C level was greater than 79 mg/dL on 2 consecutive visits, the dose of simvastatin was increased to 80 mg in a double-blind fashion. Uptitration to 80 mg was stopped in June 2011, consistent with the revised US labeling regarding the use of simvastatin, 80 mg, as per the US Food and Drug Administration. There was no change in study drug administration for low LDL-C levels.

Statistical Analyses

Patients who had LDL-C assessed at 1 month and who did not experience a primary efficacy or prespecified safety event prior to the 1-month visit were included in this analysis (n = 15 281). Patients were divided into 4 prespecified subgroups (<30 mg/dL, 30-49 mg/dL, 50-69 mg/dL, and ≥70 mg/dL) at 1 month, regardless of treatment assignment. Independent predictors of achieving a low LDL-C level were identified by multivariable logistic regression of baseline variables. These independent predictors served as covariates to generate multivariate adjusted hazard ratios (HRs) (using Cox proportional hazard models) or odds ratios (using logistic regression models) for the outcome of interest as described in the prespecified study statistical analysis plan. Adjusted trend P values were calculated using Cox proportional hazard regression by testing the coefficient of the LDL groups or the Cochran-Armitage trend test of proportions among LDL groups, as appropriate. Kaplan-Meier estimates at 7 years were compared using the log-rank test. A P value less than .05 was considered evidence of a statistically significant effect and no adjustments were made for assessment of multiple end points. Interaction terms between achieved LDL-C level at 1 month and randomized treatment were evaluated to assess for effect modification. The baseline LDL-C concentration was considered the value obtained at the time of the presenting ACS event (ie, at or within 24 hours after admission) as measured by the local hospital laboratory. Analyses were performed with SAS version 9.4 (SAS Institute).

Sensitivity Analysis

We performed a sensitivity analysis of the safety end points in patients who had received at least 1 dose of study drug, only counting events that occurred between the first dose through the last dose plus 30 days (excluding events >30 days after the last dose).10 Safety events in this analysis were compared using the Cochran-Armitage trend test of proportions among LDL-C groups for adverse events leading to study drug discontinuation, elevated liver enzymes, myopathy, and neurocognitive events. Adjusted Cox proportional hazards regression with testing of the coefficient of the LDL-C group was used to calculate the trend P values for end points with adjudicated time to event data (hemorrhagic stroke, hospitalization for heart failure, noncardiovascular death, and cancer).

Results

Of the 18 144 patients who were randomized in IMPROVE-IT, 1255 (6.9%) experienced a primary efficacy (n = 526) and/or prespecified safety event (n = 870) in the first month after randomization (ie, prior to the assessment of the LDL-C level at 1 month) and were excluded from this analysis (eTable 1 in the Supplement). A total of 1608 patients (8.9%) did not provide a lipid specimen at the 1-month visit. Thus, 15 281 patients (84.2% of the patients randomized) were included in the current analysis. The patients were followed up for a median of 6 years (range, 3.9-8.6 years). The distribution of LDL-C levels at 1 month is shown in Figure 1; the median LDL-C level was 56 mg/dL (interquartile range, 43-70 mg/dL). There were 971 patients (6.4%) who achieved an LDL-C level less than 30 mg/dL at 1 month. Among these 971 patients, the median LDL-C level at 1 month was 25 mg/dL (interquartile range, 21-27 mg/dL), and their time-weighted average LDL-C level after randomization was 34 mg/dL (interquartile range, 27-44 mg/dL) over a median of 6 years’ follow-up (Figure 2). Similarly, the time-weighted mean LDL-C concentrations for the other 3 groups were 48 mg/dL, 63 mg/dL, and 80 mg/dL for the 4780, 5504, and 4026 patients who had 1-month LDL-C levels of 30 to 49 mg/dL, 50 to 69 mg/dL, and 70 mg/dL or greater, respectively.

Across the 4 groups of achieved LDL-C level at 1 month ordered from lowest to highest, the proportion of patients randomized to ezetimibe/simvastatin were 85%, 72%, 44%, and 22%, respectively. Continuation of study drug during follow-up was significantly lower in the group achieving an LDL-C level of 70 mg/dL or greater at 1 month for each time point, while compliance was similar across the remaining 3 groups (<30 mg/dL, 30-49 mg/dL, and 50-69 mg/dL) who achieved an LDL-C level less than 70 mg/dL (eFigure 1 in the Supplement). Specifically in the group who achieved a 1-month LDL-C level less than 30 mg/dL, 68% were still taking randomized study drug at the time of median follow-up.

Baseline characteristics varied according to the LDL-C level achieved at 1 month (Table 1). On univariate analysis, the group who achieved an LDL-C level less than 30 mg/dL were more likely male, nonwhite, had higher body mass index, and more likely to have preexisting diabetes, but less likely to be a smoker or to have had a prior myocardial infarction, percutaneous coronary intervention, or coronary artery bypass graft surgery compared with the other 3 groups. Patients achieving an LDL-C level less than 30 mg/dL were also the least likely to have been treated with statin prior to their qualifying ACS event, and they had the lowest median LDL-C level at baseline of the 4 groups (85 mg/dL, 93 mg/dL, 96 mg/dL, and 97 mg/dL from the lowest to highest achieved LDL-C at 1 month). Independent predictors of achieving an LDL-C level at 1 month of less than 30 mg/dL, in order of model contribution, included randomization to ezetimibe, lower LDL-C level at baseline, history of diabetes, higher body mass index, and no prior coronary artery bypass graft (see eTable 2 in the Supplement for the complete list).

Baseline characteristics by randomized treatment for each of the 4 groups of achieved LDL-C level at 1 month are shown in eTable 3 in the Supplement. There were several significant differences in baseline characteristics when stratified by both achieved LDL-C level and treatment group as would be expected given the nonrandomized nature of the 4 groups identified by achieved LDL-C level 1 month after randomization.

Safety

Adverse events leading to drug discontinuation after 1 month were infrequent (8.9% during the 6-year median follow-up), with no difference across the groups stratified by achieved LDL-C level at 1 month (9.5%, 9.4%, 8.5%, and 8.8% for the 4 groups from lowest to highest LDL-C level, P for trend = .21; Table 2 and Figure 3). The rate of new, worsening, or relapsing malignancies reported in patients at lower levels of achieved LDL-C levels was higher in the unadjusted analysis (9.0%, 8.6%, 8.7%, and 7.5%; unadjusted P for trend = .04) for the 4 groups from lowest to highest achieved LDL-C level, but this was no longer statistically significant after accounting for baseline characteristics (adjusted P for trend = .14; Figure 3). Older age was the strongest confounding variable, correlating both with the risk for cancer and the likelihood of achieving an LDL-C level less than 30 mg/dL at 1 month.

There were no differences in the other 7 prespecified safety end points (Table 2 and Figure 3), including serious muscle events, elevation in aminotransferases levels greater than 3 times the normal level (2.2% for LDL-C level <30 mg/dL and 1.8%-2.1% in the other 3 groups), gall-bladder adverse events (3.6% for LDL-C <30 mg/dL and 3.2%-3.6% in the other 3 groups), or neurocognitive events (2.1% for LDL-C <30 mg/dL and 2.3%-2.9% in the other 3 groups), hemorrhagic stroke (0.3% for LDL-C <30 mg/dL and 0.4%-0.9% in the other groups), heart failure requiring hospitalization (4.6% for LDL-C <30 mg/dL and 3.4%-4.2% in the other 3 groups), or noncardiovascular death (5.8% for LDL-C <30 mg/dL and 4.9%-5.6% in the other 3 groups). With regard to neurocognitive events, there were no differences across the groups in either the rates of shorter-term (eg, transient global amnesia) or longer-lasting (eg, dementia) events (Table 2). In addition, there were no differences in the adjusted risk for cataract-related adverse events across the groups of achieved LDL-C at 1 month (<30 mg/dL: odds ratio, 1.12; 95% CI, 0.78-1.62; 30-49 mg/dL: odds ratio, 1.20; 95% CI, 0.96-1.50; and 50-69 mg/dL: odds ratio, 1.08; 95% CI, 0.86-1.34) compared with the referent group with LDL-C level of 70 mg/dL or greater.

Interaction analyses demonstrated no evidence of effect modification by achieved LDL-C level at 1 month on the relationship between treatment (ezetimibe vs placebo) and any of the 9 prespecified safety events or the post-hoc exploratory end point of cataract-related adverse events. No excess in safety events in either randomized treatment arm when analyzed separately was observed at progressively lower achieved LDL-C levels at 1 month (eTable 4 in the Supplement).

Efficacy

When the primary composite efficacy end point was analyzed across the 4 groups in ascending order of achieved LDL-C at 1 month, the adjusted HRs favored lower achieved LDL-C groups (HRs of 1.0, 0.82, 0.80, and 0.79 for the groups of achieved LDL-C level at 1 month of >70, 50-69, 30-49, and <30 mg/dL, respectively; P for trend < .001; eFigure 2 in the Supplement), with the numerically lowest HR present in the group who achieved an LDL-C level less than 30 mg/dL at 1 month. In this group, the unadjusted Kaplan-Meier rate of the primary efficacy end point at 7 years was 31.9% compared with 36.0% in patients who achieved an LDL-C level of 70 mg/dL or greater at 1 month (adjusted HR, 0.79; 95% CI, 0.69-0.91; P < .001). Similarly, the adjusted HRs for each of the 3 secondary composite end points were significantly reduced for each of the 3 groups that achieved an LDL-C level less than 70 mg/dL and were numerically the lowest for the group who achieved an LDL-C level less than 30 mg/dL at 1 month (eFigure 2 in the Supplement).

Sensitivity Analysis

In the sensitivity analysis limited to safety events that occurred while taking treatment or within 30 days after discontinuation, the results were qualitatively similar. There were no differences in the 9 prespecified safety end points by achieved LDL-C level at 1 month (eTable 5 in the Supplement). Analysis of safety stratified by treatment group during the period while taking treatment did not reveal any evidence for effect modification by treatment on the safety profile across the achieved LDL-C subgroups (eTable 6 in the Supplement).

Discussion

In this prespecified analysis of patients from IMPROVE-IT based on achieved LDL-C level at 1 month, we found that patients with an LDL-C level less than 30 mg/dL had a similar safety profile over a median follow-up of 6 years compared with patients who achieved a higher LDL-C level at 1 month. In addition, multiple efficacy outcomes were less frequent among patients who achieved LDL-C values progressively below 70 mg/dL, with numerically the lowest risk in the less than 30 mg/dL group. These findings are unique in that they represent the longest follow-up in one of the largest randomized trials of intensive lipid-lowering therapies, achieving LDL-C levels well below guideline recommendations of less than 70 to 77 mg/dL.11

Patients in other placebo-controlled trials of statins12 or proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitors13,14 added to background statin who achieved LDL-C levels less than 30 mg/dL had far shorter median follow-up (11-24 months) compared with IMPROVE-IT (6 years). An analysis of patients from the Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER) reported significant increases in physician-reported hematuria (HR, 2.10; 95% CI, 1.39-3.19), hepatobiliary disorders (HR, 1.77; 95% CI, 1.15-2.73), and insomnia (HR, 1.59; 95% CI, 1.03-2.48) in patients randomized to rosuvastatin, 20 mg, daily who achieved an LDL-C level less than 30 mg/dL compared with 30 mg/dL or greater. It is important to note that JUPITER was a placebo-controlled trial, whereas all patients in IMPROVE-IT were taking background simvastatin. Whereas statin-associated muscle symptoms and incident diabetes are known risks associated with intensive statin therapy,15 it is difficult to disentangle whether the other adverse events reported in JUPITER were related to the specific statin studied, and/or the degree of LDL-C lowering, or were due to the study design in which the control arm did not receive any statin.

In 2 studies of PCSK9 inhibitors (added to background statin) from intermediate-term (11-18 months) lipid-lowering trials, the only adverse events that were reported in significant excess in both studies were neurocognitive events, yet these were rare (<1%) and did not appear related to the achievement of very low LDL-C concentrations with treatment. Longer-term safety data from much larger studies with PCSK9 inhibitors are eagerly awaited.16 In contrast, no increased risk for any of the aforementioned adverse events were observed in the randomized findings from the overall population of IMPROVE-IT,1 nor were they observed in the subgroup who achieved an LDL-C level less than 30 mg/dL at 1 month, whether the analysis was conducted in the intention-to-treat or treatment populations. In addition, unlike the recent report7 with alirocumab that demonstrated an association between an achieved LDL-C level less than 25 mg/dL and a greater risk in the incidence of cataracts, we observed no similar relationship in our analysis.

Because ezetimibe was given to 85% of the patients who achieved the lowest LDL-C category of less than 30 mg/dL at 1 month, the present findings add to the totality of the evidence supporting a favorable safety profile of this agent when combined with statin as compared with statin alone.17 Several important negative findings are worthy of mention, including the lack of association between low LDL-C level and cancer, hemorrhagic stroke, and noncardiovascular death after multivariate adjustment. Each of these adverse events had been reported in excess in unadjusted observational analyses of patients with low LDL-C levels.2-4 In our data, we also found that the unadjusted rate of cancer was higher in those achieving the lowest LDL-C level in the intention-to-treat analysis (but not in the treatment analysis). However, this finding was no longer significant after adjustment for differences in baseline characteristics (including the 3-year difference in age). Thus, the prior observational data2-4 may represent associations without causality or even possibly reverse causality in patients with systemic illness that result in very low LDL-C concentration, rather than causally related events due to lower LDL-C achieved with lipid-lowering therapy. Although an increase in intracerebral hemorrhage was reported in patients with acute ischemic stroke treated with atorvastatin, 80 mg, compared with placebo in the Stroke Prevention by Aggressive Reduction in Cholesterol Levels Study,18 this risk appears to have been limited to patients with a history of intracerebral hemorrhage prior to the qualifying ischemic stroke.19 In both the Cholesterol Treatment Trialist second meta-analysis20 and IMPROVE-IT,1 hemorrhagic stroke was numerically (but not significantly) more frequent in patients receiving more intensive lipid therapy; thus, given the low incidence of hemorrhagic stroke in these populations, a modest increased risk for hemorrhagic stroke with therapies that achieve very low LDL-C levels may be present.

The findings in this analysis that patients who achieve low LDL-C values at 1 month have a reduced adjusted risk for future cardiovascular events is consistent with several prior analyses of patients who achieved LDL-C levels with progressively lower cut points, including LDL-C values below 64 mg/dL,21 less than 50 mg/dL,8,22 and less than 40 mg/dL.23,24 In addition, a propensity-matched analysis from the Veterans Administration Palo Alto Health Care System in 6107 consecutive patients with baseline LDL-C levels less than 60 mg/dL found that patients treated with a statin exhibited a significantly lower mortality, even when the baseline LDL-C level was less than 40 mg/dL prior to initiation of statin therapy.25

To our knowledge, although no large randomized trial to date (including IMPROVE-IT) has titrated therapy to achieve 2 different LDL-C levels, our findings show a lower cardiovascular event rate in patients who achieve LDL-C values less than 70 mg/dL at 1 month, which is currently the most aggressive guideline-recommended target.11 This does not address the important question of what the ideal range of LDL-C level is to prevent atherosclerosis. Several lines of evidence, including studies in veterinary medicine, anthropology, and maternal-fetal medicine, point to an LDL-C target below 70 mg/dL. Many mammals (including primates) have LDL-C levels well below 70 mg/dL, including several mammals with total cholesterol level less than 75 mg/dL.26 Modern hunter-gatherer societies and aboriginal cultures across the globe have mean total cholesterol levels between 100 and 130 mg/dL.26 In humans (as well as pigs and sheep), the mean total cholesterol in utero is less than 100 mg/dL and at birth is approximately 50 mg/dL.27

Although rare, individuals with compound heterozygote loss-of-function PCSK9 mutations and LDL-C levels less than 20 mg/dL have been reported, are apparently healthy, and have had healthy offspring.28 Because the LDL-C receptor optimally binds LDL at a concentration of 2.5 mg/dL, and there is a 10-to-1 gradient between the LDL-C concentration in the plasma and interstitial fluid, Brown and Goldstein29 have speculated that a plasma concentration of LDL-C of 25 mg/dL would be adequate for normal cellular function.

Limitations

We acknowledge limitations of this analysis. Results in patients participating in this randomized clinical trial, 34% treated with statin prior to ACS, may differ in a general patient population. Because we stratified patients based on the LDL-C level achieved at 1 month, these results do not apply to patients who develop rapid intolerance to statins. Although the trial was large, some events were rare, resulting in limited power. Last, early discontinuation rates of lipid therapy may bias our results toward the null.

Conclusions

In conclusion, we observed that the achievement of very low LDL-C concentration at 1 month in a cohort of 971 patients followed up for a median of 6 years experienced a generally similar safety profile as those with higher LDL-C levels, and that achievement of LDL-C levels below currently recommended targets was associated with even further numerical reductions in cardiovascular events. These observational data support the use of intensive lipid lowering in very high-risk patients following an ACS to prevent future ischemic events.

Back to top
Article Information

Corresponding Author: Robert P. Giugliano, MD, SM, TIMI Study Group, 350 Longwood Ave, 1st Floor Offices, Boston, MA 02115 (rgiugliano@partners.org).

Accepted for Publication: January 5, 2017.

Published Online: March 14, 2017. doi:10.1001/jamacardio.2017.0083

Author Contributions: Dr Giugliano had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Giugliano, Blazing, Cannon, Braunwald.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Giugliano, Park.

Critical revision of the manuscript for important intellectual content: Wiviott, Blazing, De Ferrari, Murphy, White, Tershakovec, Cannon, Braunwald.

Statistical analysis: Giugliano, Park, Murphy.

Obtained funding: Cannon, Braunwald.

Administrative, technical, or material support: Giugliano, Wiviott, Blazing, Cannon.

Supervision: Blazing, Murphy, Cannon, Braunwald.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Giugliano has received grant support from Amgen and Merck (to his institution) and honoraria from the American College of Cardiology, Amgen, Bristol-Myers Squibb, CVS Caremark, Daiichi Sankyo, GlaxoSmithKline, Merck, Pfizer, and Sanofi. Dr Wiviott has received grant support from Merck, Sanofi-Aventis, AstraZeneca, Bristol-Myers Squibb, Eisai, Arena, and Eli Lilly/Daiichi-Sankyo and honoraria from AstraZeneca, Bristol-Myers Squibb, Eisai, Arena, Eli Lilly/Daiichi-Sankyo, Sanofi-Aventis, Boehringer Ingelheim, Aegerion, Angelmed, Janssen, Xoma, ICON Clinical, and Boston Clinical Research Institute. Dr Blazing has served as an advisory board member for Merck and has provided consulting for AstraZeneca and Novartis. Dr de Ferrari has received grants and/or personal fees from Merck, Amgen, Boston Scientific, and Sigma Tau. Dr Park has received grant support from Merck. Ms Murphy has received grant support and honoraria from Merck. Dr Tershakovec is an employee and stockholder of Merck. Dr Cannon has received grants and/or personal fees from Accumetrics, Arisaph, AstraZeneca, Boehringer Ingelheim, GlaxoSmithKline, Janssen, Daiichi-Sankyo, Merck, Bristol-Myers Squibb, CSL Behring, Essentialis, Kowa, Takeda, Lipimedix, Pfizer, Regeneron, Sanofi, Alnylam, and Amgen. Dr Braunwald has received grant support from Merck (to his institution), Duke University, AstraZeneca, Novartis, Daiichi-Sankyo, and GlaxoSmithKline and honoraria from The Medicines Company, Medscape, Menarini International, and Theravance. He provided uncompensated consultancies and lectures for Novartis. No other disclosures were reported.

Funding/Support: The Improved Reduction of Outcomes: Vytorin Efficacy International Trial was supported by Merck & Co Inc.

Role of the Funder/Sponsor: The sponsor provided funding for the study; participated in the design and conduct of the study; collection, management, and analysis of the data; and confirmed statistical analyses that were primarily performed by the TIMI Study Group. The sponsor’s participation was under direction of the executive committee. Dr Tershakovec (a Merck employee) was a minority part of that committee.

References
1.
Cannon  CP, Blazing  MA, Giugliano  RP,  et al; IMPROVE-IT Investigators.  Ezetimibe added to statin therapy after acute coronary syndromes.  N Engl J Med. 2015;372(25):2387-2397.PubMedGoogle ScholarCrossref
2.
Kritchevsky  SB.  Dietary lipids and the low blood cholesterol-cancer association.  Am J Epidemiol. 1992;135(5):509-520.PubMedGoogle ScholarCrossref
3.
Neaton  JD, Blackburn  H, Jacobs  D,  et al; Multiple Risk Factor Intervention Trial Research Group.  Serum cholesterol level and mortality findings for men screened in the Multiple Risk Factor Intervention Trial.  Arch Intern Med. 1992;152(7):1490-1500.PubMedGoogle ScholarCrossref
4.
Tirschwell  DL, Smith  NL, Heckbert  SR, Lemaitre  RN, Longstreth  WT  Jr, Psaty  BM.  Association of cholesterol with stroke risk varies in stroke subtypes and patient subgroups.  Neurology. 2004;63(10):1868-1875.PubMedGoogle ScholarCrossref
5.
Berry  PH, MacDonald  JS, Alberts  AW,  et al.  Brain and optic system pathology in hypocholesterolemic dogs treated with a competitive inhibitor of 3-hydroxy-3-methylglutaryl coenzyme A reductase.  Am J Pathol. 1988;132(3):427-443.PubMedGoogle Scholar
6.
Cannon  CP, Giugliano  RP, Blazing  MA,  et al; IMPROVE-IT Investigators.  Rationale and design of IMPROVE-IT (IMProved Reduction of Outcomes: Vytorin Efficacy International Trial): comparison of ezetimbe/simvastatin versus simvastatin monotherapy on cardiovascular outcomes in patients with acute coronary syndromes.  Am Heart J. 2008;156(5):826-832.PubMedGoogle ScholarCrossref
7.
Robinson  JG, Rosenson  RS, Farnier  M,  et al.  Safety of very low low-density lipoprotein cholesterol levels with alirocumab: pooled data from randomized trials.  J Am Coll Cardiol. 2017;69(5):471-482.PubMedGoogle ScholarCrossref
8.
Hsia  J, MacFadyen  JG, Monyak  J, Ridker  PM.  Cardiovascular event reduction and adverse events among subjects attaining low-density lipoprotein cholesterol <50 mg/dl with rosuvastatin: the JUPITER Trial (Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin).  J Am Coll Cardiol. 2011;57(16):1666-1675.PubMedGoogle ScholarCrossref
9.
Friedewald  WT, Levy  RI, Fredrickson  DS.  Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge.  Clin Chem. 1972;18(6):499-502.PubMedGoogle Scholar
10.
Blazing  MA, Giugliano  RP, de Lemos  JA,  et al.  On-treatment analysis of the Improved Reduction of Outcomes: Vytorin Efficacy International Trial (IMPROVE-IT).  Am Heart J. 2016;182(182):89-96.PubMedGoogle ScholarCrossref
11.
Morris  PB, Ballantyne  CM, Birtcher  KK, Dunn  SP, Urbina  EM.  Review of clinical practice guidelines for the management of LDL-related risk.  J Am Coll Cardiol. 2014;64(2):196-206.PubMedGoogle ScholarCrossref
12.
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(11):1682-1689.PubMedGoogle ScholarCrossref
13.
Sabatine  MS, Giugliano  RP, Wiviott  SD,  et al; Open-Label Study of Long-Term Evaluation against LDL Cholesterol (OSLER) Investigators.  Efficacy and safety of evolocumab in reducing lipids and cardiovascular events.  N Engl J Med. 2015;372(16):1500-1509.PubMedGoogle ScholarCrossref
14.
Robinson  JG, Farnier  M, Krempf  M,  et al; ODYSSEY LONG TERM Investigators.  Efficacy and safety of alirocumab in reducing lipids and cardiovascular events.  N Engl J Med. 2015;372(16):1489-1499.PubMedGoogle ScholarCrossref
15.
Thompson  PD, Panza  G, Zaleski  A, Taylor  B.  Statin-associated side effects.  J Am Coll Cardiol. 2016;67(20):2395-2410.PubMedGoogle ScholarCrossref
16.
Giugliano  RP, Sabatine  MS.  Are PCSK9 inhibitors the next breakthrough in the cardiovascular field?  J Am Coll Cardiol. 2015;65(24):2638-2651.PubMedGoogle ScholarCrossref
17.
Luo  L, Yuan  X, Huang  W,  et al.  Safety of coadministration of ezetimibe and statins in patients with hypercholesterolaemia: a meta-analysis.  Intern Med J. 2015;45(5):546-557.PubMedGoogle ScholarCrossref
18.
Amarenco  P, Bogousslavsky  J, Callahan  A  III,  et al; Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL) Investigators.  High-dose atorvastatin after stroke or transient ischemic attack.  N Engl J Med. 2006;355(6):549-559.PubMedGoogle ScholarCrossref
19.
Athyros  VG, Tziomalos  K, Karagiannis  A, Wierzbicki  AS, Mikhailidis  DP.  Aggressive statin treatment, very low serum cholesterol levels and haemorrhagic stroke: is there an association?  Curr Opin Cardiol. 2010;25(4):406-410.PubMedGoogle ScholarCrossref
20.
Baigent  C, Blackwell  L, Emberson  J,  et al; Cholesterol Treatment Trialists’ (CTT) Collaboration.  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(9753):1670-1681.PubMedGoogle ScholarCrossref
21.
LaRosa  JC, Grundy  SM, Kastelein  JJ, Kostis  JB, Greten  H; Treating to New Targets (TNT) Steering Committee and Investigators.  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(5):747-752.PubMedGoogle ScholarCrossref
22.
Boekholdt  SM, Hovingh  GK, Mora  S,  et al.  Very low levels of atherogenic lipoproteins and the risk for cardiovascular events: a meta-analysis of statin trials.  J Am Coll Cardiol. 2014;64(5):485-494.PubMedGoogle ScholarCrossref
23.
Wiviott  SD, Cannon  CP.  The safety and efficacy of achieving very low LDL-cholesterol concentrations with high dose statin therapy.  Curr Opin Lipidol. 2006;17(6):626-630.PubMedGoogle ScholarCrossref
24.
Wiviott  SD, Cannon  CP, Morrow  DA, Ray  KK, Pfeffer  MA, Braunwald  E; PROVE IT-TIMI 22 Investigators.  Can low-density 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(8):1411-1416.PubMedGoogle ScholarCrossref
25.
Leeper  NJ, Ardehali  R, deGoma  EM, Heidenreich  PA.  Statin use in patients with extremely low low-density lipoprotein levels is associated with improved survival.  Circulation. 2007;116(6):613-618.PubMedGoogle ScholarCrossref
26.
Hochholzer  W, Giugliano  RP.  Lipid lowering goals: back to nature?  Ther Adv Cardiovasc Dis. 2010;4(3):185-191.PubMedGoogle ScholarCrossref
27.
Dietschy  JM, Turley  SD.  Thematic review series: brain lipids: cholesterol metabolism in the central nervous system during early development and in the mature animal.  J Lipid Res. 2004;45(8):1375-1397.PubMedGoogle ScholarCrossref
28.
Zhao  Z, Tuakli-Wosornu  Y, Lagace  TA,  et al.  Molecular characterization of loss-of-function mutations in PCSK9 and identification of a compound heterozygote.  Am J Hum Genet. 2006;79(3):514-523.PubMedGoogle ScholarCrossref
29.
Brown  MS, Goldstein  JL.  A receptor-mediated pathway for cholesterol homeostasis.  Science. 1986;232(4746):34-47.PubMedGoogle ScholarCrossref
×