How to discontinue beta blockers

Abstract

β-blockers have been among the first medications shown to improve outcomes after acute myocardial infarction (AMI). With the advent of reperfusion therapy and other secondary-prevention medications, their role has become uncertain, and large-scale experience after AMI in the contemporary era is lacking. In particular, the effect of stopping β-blockers in patients initially treated after AMI is unknown.

Methods and Results:

Using the French healthcare databases, 73 450 patients (<80 years of age), admitted for AMI in 2007 to 2012, without acute coronary syndrome (ACS) in the previous 2 years and no evidence of heart failure, having received optimal treatment with myocardial revascularization and all recommended medications in the 4 months after index admission, and not having discontinued β-blockers before 1 year, were followed for 3.8 years on average. β-Blocker discontinuation was defined as 4 consecutive months without exposure. If β-blocker treatment was resumed later on, follow-up was stopped. Both the risk of the composite outcome of death or admission for ACS and the risk of all-cause mortality were assessed in relation with β-blocker discontinuation during follow-up. Adjusted hazard ratios were estimated using marginal structural models accounting for time-varying confounders affected by previous exposure. A similar analysis was performed with statins. Of 204 592 patient-years, 12 002 (5.9%) corresponded to discontinued β-blocker treatment. For β-blocker discontinuation, the adjusted hazard ratio for death or ACS was 1.17 (95% confidence interval, 1.01–1.35); for all-cause death, the adjusted hazard ratio was 1.13 (95% confidence interval, 0.94–1.36). In contrast, for statin discontinuation, the adjusted hazard ratios for death or ACS and for all-cause death were 2.31 (95% confidence interval, 2.01–2.65) and 2.57 (95% confidence interval, 2.19–3.02), respectively.

Conclusions:

In routine care of patients without heart failure, revascularized and optimally treated after AMI, discontinuation of β-blockers beyond 1 year after AMI was associated with an increased risk of death or readmission for ACS, while statistical significance was not reached for the association with all-cause mortality. A contemporary randomized clinical trial is needed to precise the role of β-blockers in the long-term treatment after AMI.

See Editorial by Ko and Jackevicius

• In the contemporary era of reperfusion therapy and modern secondary-prevention medications, the role of β-blockers in improving outcomes after acute myocardial infarction has become uncertain.

• The effect of stopping β-blockers in patients without heart failure beyond the early phase of myocardial infarction therapy is still being debated.

• The impact of discontinuing β-blockers in patients <80 years, with no recent history of heart failure or acute coronary syndrome, and having received optimal management after acute myocardial infarction was assessed in a large cohort study using the French healthcare databases.

• This study found that discontinuation of β-blockers beyond 1 year after myocardial infarction may increase risk of death or readmission for acute coronary syndrome.

• Statistical significance was not reached for the association with all-cause mortality.

Considerable progress has been made in the treatment of acute myocardial infarction (AMI) and coronary artery disease in the past 30 years, leading to substantial decrease in early mortality.1,2 β-Blockers have been among the first classes of medications to document a reduction in early and long-term mortality in patients with AMI.3 This has led to the highest level of recommendation for their prescription in patients with AMI.4 By extension, β-blockers are considered first-line therapy for angina relief in chronic coronary artery disease.5 Most of the trials documenting the efficacy of β-blockers after AMI, however, were performed before the role of reperfusion therapy (fibrinolysis and primary percutaneous coronary intervention [PCI]) had been demonstrated and before the widespread use of secondary-prevention medications, such as statins or antiplatelet agents.

Because of the lack of evidence from recent randomized trials, the level of recommendation for β-blockers after AMI in patients without left ventricular dysfunction has been downgraded in the most recent European Society of Cardiology guidelines,6 whereas the American College of Cardiology Foundation/American Heart Association guidelines still strongly recommend β-blockers in all patients with ST-segment–elevation myocardial infarction (STEMI),7 this divergence reflecting the current uncertainty on their intrinsic usefulness in patients receiving otherwise optimal treatment.8 Similar discrepancies between guidelines exist for non-STEMI.9,10

In addition, the duration of many of the trials assessing the benefit of β-blockers after AMI was short, with few trials having follow-up periods >2 years,11 and the trials essentially demonstrated the benefits of short-term β-blocker treatment after AMI. Whether the benefit of β-blockers is confined to the first months after myocardial infarction, or whether protracting β-blocker therapy beyond the first year is also beneficial is a yet unanswered question. Recently, cohort studies have challenged the usefulness of long-term β-blocker therapy in patients with stable (or stabilized) coronary artery disease,12,13 whereas others have suggested a benefit in early post-AMI patients,14,15 with no impact on outcomes beyond 1 year,16 leading to interrogations as to the possibility of discontinuing β-blocker therapy in patients with preserved left ventricular function.17,18

We, therefore, performed a study to evaluate the impact of β-blocker discontinuation beyond the acute stage in a cohort of patients hospitalized for AMI in 2007 to 2012, who had no evidence of heart failure, were treated optimally (with myocardial revascularization and receiving recommended secondary-prevention medications: combined antiplatelet agents, statins, and angiotensin-converting enzyme inhibitors [ACE-Is] or angiotensin receptor blockers [ARBs]), and who had received β-blocker for at least 1 year after AMI, using the French healthcare databases. For comparison and validation purposes, a similar analysis was performed to assess clinical events after discontinuation of statin therapy.

Methods

Data Source

The French national health insurance covers the entire French population (65.5 million inhabitants in 2013). Its information system (Système National d’Information Inter-Régimes de l’Assurance Maladie) contains information on each health spending reimbursement, as well as demographic data, such as age, sex, area of residence, or vital status, and is linked to the French hospital discharge database19 and has been used previously for clinical research, particularly in the field of metabolic and cardiovascular diseases.20–22 Hospital discharge diagnoses are recorded using the International Classification of Diseases, TenthRevision (ICD-10), some codes having been further refined.17 Patients with severe chronic conditions are entitled to full reimbursement for a long-term disease, which is also recorded as ICD-10 code. The present study was conducted in patients covered by the general scheme of health insurance (75% of the population), with comprehensive data available for the whole period considered. The study was approved by the French data protection agency (Commission Nationale de l’Informatique et des Libertés), but the data access permission policy prohibits making the data set publicly available. All databases used in this study only contained anonymous patient records.

Study Population

All patients admitted to hospital for AMI (ICD-10 code I21-22 with refinement indicating initial management) in 2007 to 2012, aged 25 to 79 years at admission, with myocardial revascularization after AMI, and receiving optimal secondary-prevention therapy, ie, combined β-blocker, statin, ACE-I or ARB, and antiplatelet therapy, were eligible (Figure 1). If there was >1 admission for a given patient during the inclusion period, the index admission was selected randomly. Time was divided into calendar months, month 1 referring to the month after the month of index admission (Figure I in the Data Supplement).

Figure 1. Flow chart of patient selection. Time was divided into calendar months. Month m=0 referred to the month of index admission. ACE-I indicates angiotensin-converting enzyme inhibitor; ACS, acute coronary syndrome; AMI, acute myocardial infarction; ARB, angiotensin II receptor blocker; CABG, coronary artery bypass graft; and PCI, percutaneous coronary intervention. *If there was >1 admission for a given patient, the index admission was selected at random. †Four consecutive months without any drug delivery or withdrawal from the general health insurance scheme.

We excluded (1) patients with hospital discharge diagnosis of acute coronary syndrome (ACS; ICD-10 code, I21-I24) in the 2 years before index admission or with hospital discharge or long-term disease diagnosis of heart failure (I50) up to the end of month 4; (2) patients without PCI or coronary artery bypass graft between start of index admission and end of month 4; (3) patients not exposed to each of the following drug classes between months 1 and 4: β-blockers, statins, ACE-Is or ARBs, and antiplatelet agents; and (4) patients having discontinued β-blockers before 1 year.

Exposure

Exposure to β-blockers was modeled as a time-varying variable and in a first step measured on a daily time scale (days covered by the prescription) as described in detail in the Data Supplement. In a second step, exposure was calculated on a monthly time scale: for each month, exposure was recorded if there was at least 1 day of exposure in this month. Finally, β-blocker treatment was considered to be discontinued when there were 4 consecutive months without exposure. Exposures to statins, ACE-Is or ARBs, and antiplatelet agents, respectively, as well as discontinuation of statin treatment were measured in the same way.

Outcome Variables and Follow-Up

Two outcomes were assessed: first, a composite of all‐cause mortality or hospital admission for ACS (ICD‐10 code for primary or secondary hospital discharge diagnosis, I21‐24); and second, all‐cause mortality.

Follow-up started at month 13 (ie, only patients surviving at 12 months were included) and continued until (1) outcome, (2) resumption of β-blocker treatment after discontinuation, (3) loss to follow-up (4 consecutive months without any drug delivery or withdrawal from the general health insurance scheme), (4) month 59 (ie, patients were followed ≤5 years after index admission), or (5) December 2014, whichever occurred first.

Covariates

Baseline covariates included age at index admission, sex, deprivation index,23 type of index AMI according to ICD codes, drug use in the year before index admission, inpatient treatments between start of the index admission and end of the fourth month after index admission (PCI, coronary artery bypass graft, pacemaker, or cardioverter-defibrillator implantation), as well as the following hospital discharge diagnoses in this period: ischemic stroke, cardiac arrest, ventricular tachycardia, atrial fibrillation, tobacco use, and obesity. In addition, cardiovascular drug use was assessed for the 4 months after index admission (calcium channel blockers, nitrates or other vasodilator agents, other lipid-lowering agents, other antiarrhythmic agents, loop diuretics, other diuretics, and oral anticoagulants).

Time-varying covariates were assessed monthly during follow-up. They covered all the abovementioned inpatient treatments and hospital discharge diagnoses, plus heart failure, as well as use of all the abovementioned cardiovascular drugs, plus statins, ACE-Is or ARBs, and antiplatelet agents. Comorbidities were identified by hospital discharge and long-term disease diagnoses and prescriptions of specific drugs (Data Supplement), consultations with a cardiologist in private practice were searched for, and total length of all hospital admissions for at least 1 night were calculated.

Statistical Analysis

A weighted Cox proportional hazards model (a so-called marginal structural Cox model) was used to estimate the hazard ratio (HR) of β-blocker discontinuation versus continuation, controlling for baseline and time-varying confounding.24–26 Each patient’s contribution to the risk set for a given month t was weighted by the inverse of the probability of the treatment history he/she actually had, given his/her covariate history. To correct also for potentially informative censoring at treatment resumption, the patient’s initial weight was multiplied by the inverse probability of remaining uncensored up to month t, given his/her covariate history. The resulting weights are called inverse probability of treatment and censoring weights. Details, including information on weight truncation, are provided in Figure II in the Data Supplement. All baseline and time-varying covariates listed in Tables 1 and 2 were used for weight calculation, plus the calendar year of index admission. Using the same weights, the inverse probability of treatment and censoring weights method was also applied to calculate adjusted cumulative incidence curves.26 Their 95% confidence bands were obtained by bootstrap with 200 replications.

Predefined subgroup analyses were performed, according to sex, prior use of β-blockers before index admission, type of myocardial infarction, presence of asthma or chronic obstructive pulmonary disease at baseline, and type of β-blockers: vasodilating β-blockers (carvedilol, nebivolol, labetalol et celiprolol) and conventional β-blockers (others).

Sensitivity analyses focused on the impact of weight truncation and the role of the different time-varying covariates in adjustment. In a first complementary analysis, a conventional multivariate Cox analysis was performed, including all baseline and time-varying covariates listed in Tables 1 and 2, plus the year of index admission. In a second complementary analysis, patients were not censored when they resumed the treatment after discontinuation, but they were continued to be considered as having discontinued treatment, and estimation was based on inverse probability of treatment weights only (no censoring weights).

Reported P values are 2 sided, and results were considered statistically significant for P<0.05. All statistical analyses were performed with SAS software, version 9.2 (SAS Institute, Inc, Cary, NC). The SAS program was based on codes published previously.24,26

Results

A total of 73 450 patients were included in the study (Figure 1). Mean duration from index admission to end of follow-up was 3.8 years (range, 13–59 months). Baseline and time-varying characteristics are summarized in Tables 1 and 2. A total of 4381 patients died or were admitted for ACS, corresponding to a crude incidence rate of 2.1 per 100 patient-years. For death, the number of events was 2739, corresponding to a crude incidence rate of 1.3 per 100 patient-years.

β-Blocker Exposure and Predictors of Discontinuation

Total follow-up time was 2 455 104 patient-months, of which 144 024 (5.9%) were time with discontinued β-blocker treatment.

The β-blockers prescribed during follow-up were predominantly cardioselective (95.9% of the total number of pills). Bisoprolol was the most commonly prescribed (46.7%), followed by atenolol (20.4%), acebutolol (14.4%), nebivolol (8.3%), metoprolol (5.2%), carvedilol (1.8%), sotalol (1.7%), and celiprolol (0.7%).

The most important factors related with β-blocker discontinuation were time varying: use of calcium channel blockers (adjusted odds ratio [OR], 2.4; 95% confidence interval, 2.3–2.6) and other antiarrhythmic agents (adjusted OR, 1.8; 95% confidence interval, 1.6–2.1), as well as a pacemaker implantation (adjusted OR, 1.8; 95% confidence interval, 1.2–2.8) and a prolonged total length of hospital admissions; conversely, time-varying exposure to statins (adjusted OR, 0.30; 95% confidence interval, 0.28–0.32), ACE-Is or ARBs (adjusted OR, 0.49; 95% confidence interval, 0.46–0.52), and antiplatelet agents (adjusted OR, 0.29; 95% confidence interval, 0.25–0.34), cardiodefibrillator implantation (adjusted OR, 0.44; 95% confidence interval, 0.21–0.93), and coronary artery bypass graft (adjusted OR, 0.53; 95% confidence interval, 0.34–0.82), as well as β-blocker use in the year before index admission (adjusted OR, 0.54; 95% confidence interval, 0.51–0.58) were associated with a lower likelihood of discontinuation. The estimated exposure model is fully described in Table 3, as well as the estimated model of treatment resumption. The same analyses were performed for statin therapy (Table I in the Data Supplement).

Among the 10 409 patients who discontinued β-blocker treatment, 4432 (42.6%) resumed treatment during follow-up. Among these 4432 patients, 1647 (37.2%) resumed in the first month after the 4-month period without exposure, 816 (18.4%) in the second month, and 487 (11.0%) in the third month.

Clinical Events After β-Blocker Discontinuation

The crude HR for death or ACS, comparing β-blocker discontinuation versus continuation, was 1.22 (95% confidence interval, 1.09–1.37), whereas the adjusted HR was 1.17 (95% confidence interval, 1.01–1.35). The adjusted cumulative incidence rate of the composite outcome event at month 59 (ie, 5 years after index admission) was 9.4% (95% confidence interval, 8.4%–10.4%) for β-blocker discontinuation versus 8.0% (95% confidence interval, 7.7%–8.3%) for continuation (Figure III in the Data Supplement). The crude and adjusted HR for death were 1.24 (95% confidence interval, 1.08–1.44) and 1.13 (95% confidence interval, 0.94–1.36), respectively.

Subgroup analyses for the composite outcome showed no important effect modification according to sex or previous use of β-blockers before the index episode, but the estimated effect of β-blockers discontinuation was modified by the type of AMI (no effect in patients with nontransmural AMI, or transmural AMI not located to the anterior or inferior walls) and the presence of asthma/chronic obstructive pulmonary disease at baseline (statistical significance was not reached in patients without asthma/chronic obstructive pulmonary disease at baseline; Figure 2). The adjusted HR for death or ACS was 1.07 (95% confidence interval, 0.77–1.49) for vasodilating β-blockers and 1.16 (95% confidence interval, 0.99–1.37) for conventional β-blockers.

Figure 2. Estimated effect of β-blocker discontinuation in patient sub-groups. The outcome was a composite of all-cause mortality or acute coronary syndrome. The estimated effects were adjusted for confounding because of all baseline and time-varying covariates listed in Tables 1 and 2, plus the year of index acute myocardial infarction (AMI). CI indicates confidence interval; and COPD, chronic obstructive pulmonary disease.

Sensitivity and Complementary Analyses

The results were not sensitive to the thresholds used for weight truncation (Table II in the Data Supplement): for the composite outcome, truncation at the 0.1% and 99.9% percentiles resulted in an HR of 1.16 (95% confidence interval, 0.99–1.36), whereas truncation at the 1% and 99% percentiles yielded an HR of 1.18 (95% confidence interval, 1.03–1.35). Beside exposure to statins, to ACE-Is or ARBs, and to antiplatelet agents, the covariates with the most important impact on adjustment were consultation with a cardiologist, prolonged total length of hospital admissions, and calcium channel blocker use (Table III in the Data Supplement). The other covariates had little impact.

Estimated with a conventional Cox model, the HR for the composite outcome was 1.09 (95% confidence interval, 0.97–1.23; Figure IV in the Data Supplement), and the HR for mortality was 1.04 (95% confidence interval, 0.90–1.21). Without censoring at treatment resumption, total follow-up time was 2 543 148 patient-months, of which 232 068 (9.1%) were time with discontinued β-blocker treatment, including time with β-blocker resumption, and 4541 patients experienced the composite outcome. This analysis yielded a crude HR of 1.15 (95% confidence interval, 1.04–1.27) and an adjusted HR of 1.07 (95% confidence interval, 0.95–1.20) when using inverse probability of treatment weighting.

Comparative Analysis of Clinical Events After Statin Discontinuation

The study’s approach was also applied to statin exposure, considering patients not having discontinued statins before 1 year and yielded an adjusted HR of 2.31 (95% confidence interval, 2.01–2.65) for the composite outcome and 2.57 (95% confidence interval, 2.19–3.02) for all-cause mortality (Figure V in the Data Supplement). The HRs estimated from a conventional Cox model for the composite outcome were 1.93 (95% confidence interval, 1.71–2.17; Figure IV in the Data Supplement) and 2.03 (95% confidence interval, 1.76–2.33), respectively.

Discussion

In patients <80 years of age with AMI, no evidence of heart failure or ACS in the previous 2 years, and receiving optimal medical and revascularization therapy, stopping β-blockers beyond 1 year after the acute episode was associated with an increased risk of a composite outcome of death or readmission for ACS and was not significantly associated with increased all-cause mortality. The increased risk was not found in patients with nontransmural myocardial infarction, and it was not statistically significant in patients without asthma/chronic obstructive pulmonary disease at baseline. The sensitivity analyses showed consistent results. These findings sharply contrasted with the observation of a >2-fold increase in risk of events after stopping statin therapy.

Rate of β-Blocker Discontinuation

The rate of discontinuation of β-blockers was low (14.2% of all patients, 9.5% among all patients still present at month 59). An additional 5.4% who had received β-blockers during the first 4 months had stopped their treatment during the first year and were, therefore, excluded from the study. Large variations in adherence to β-blockers have been reported, depending on the population selected, clinical context, country, health insurance coverage, and way adherence was assessed (proportion of days covered versus discontinuation). Thus, in a large UK database in primary care, Kalra et al27 found that β-blocker discontinuation increased from 27% at 1 year to 50% at 3 years; the population, however, included patients with heart failure or angina, in addition to patients with myocardial infarction. In a large cohort of patients hospitalized for AMI in 2001 in the United States, 35% of the patients with β-blockers at 30 days were poorly adherent at 1 year.28 A similar figure (27%) was observed in a large population of elderly patients with post-AMI in Ontario, but the proportion of patients with very low adherence rates (<40% of days covered) was only 9%.29 Our figure for discontinuation of β-blockers compares with data from the Canadian Acute Coronary Syndromes Registry II,30 which documented a discontinuation rate of 10% between discharge and 1 year for β-blockers, and with those of a survey in Switzerland, in which 9.5% of patients on β-blockers after AMI had stopped their treatment at 1 year.31 Likewise, in the 2005 FAST-MI, 11% of the patients with β-blockers at discharge had stopped their medication at 1 year.16 In fact, beyond the potential role of health insurance coverage in France (all secondary-prevention medications after AMI being fully reimbursed), the low discontinuation rate in our population is also likely explained by the selection criteria we used: patients with myocardial revascularization at the acute stage, receiving full optimal therapy at discharge and in the months immediately following, and <80 years of age.

Prognostic Impact of β-Blockers in the Era of Reperfusion Therapy and Other Evidence-Based Therapies

Most of the trials documenting the benefit of β-blockers have been performed before the era of reperfusion therapy and the benefit was observed in high-risk subgroups.32 In a recent meta-analysis of β-blocker trials in AMI, Bangalore et al11 found that β-blockers reduced mortality and improved outcomes in the prereperfusion era but had no such effect in the most recent period. The results with β-blockers in the most recent period, however, were mostly driven by those of the COMMIT (Clopidogrel and Metoprolol in Myocardial Infarction Trial),33 in which 45% of the patients had no reperfusion therapy and none were treated with primary PCI; also, the primary outcome of the trial was at 28 days.

Several observational studies have assessed the relationship between β-blocker prescription, adherence or discontinuation after AMI, and clinical outcomes. Twenty years ago, in 1994 to 1995, the Cooperative Cardiovascular Project showed that prescription of β-blockers at discharge was associated with improved survival, particularly in high-risk groups, and in patients with nontransmural infarction; at that time, statins were virtually not used, and primary PCI was uncommon.34 The GRACE showed that early (≤24 hours from admission) treatment with β-blockers in patients with non-STEMI included from 1999 to 2004 was associated with lower hospital and 6-month mortality, but no information beyond that time is available.35 More recently, in a Korean registry of patients admitted between 2004 and 2009, β-blocker prescription at discharge was associated with lower mortality at 3 years.15 A dose–response-type adherence–mortality association was observed for β-blockers in 31 455 elderly AMI survivors included between 1999 and 2003 from the Ontario reimbursement database.13 A more pronounced association was observed with statin therapy. In the PREMIER Registry,36 early discontinuation of β-blockers (1 month after AMI) was associated with a 2-fold increase in mortality at 1 year (adjusted HR, 1.96; 95% confidence interval, 1.10–3.45). Conversely, the large REACH Registry, assessing outcomes of 14 043 patients with previous myocardial infarction during 44 months of follow-up, showed no significant impact of late discontinuation of β-blockers.12 In patients with a recent AMI (in the past year), the risk of cardiovascular death, myocardial infarction, or stroke was numerically lower with β-blocker therapy (HR, 0.79; 95% confidence interval, 0.60–1.04), with a significantly reduced risk of cardiovascular events, including hospitalization or revascularization (HR, 0.77; 95% confidence interval, 0.64–0.92).

Recently, the FAST-MI 2005 registry found that persistence with β-blockers therapy during the first year was not associated with higher 5-year mortality (adjusted HR, 1.19; 95% confidence interval, 0.65–2.18).16 However, the latter analysis was based on a relatively small number of patients (after propensity score matching, 95 patients without β-blockers were compared with 277 patients with β-blockers). Finally, a recent study based on 90 869 Medicare beneficiaries of ≥65 years of age, who survived >180 days after AMI hospitalization in 2008 to 2010 and followed ≤18 months, suggested limited additional benefit for β-blockers in patients who were adherent to statins and ACE inhibitors/ARBs.37 Also, better clinical outcomes were observed with vasodilating β-blockers than conventional β-blocker therapy in patients with AMI based on multicenter Korean Acute Myocardial Infarction Registry data.38

Direct comparisons with these studies are difficult because of methodological issues, including important differences concerning the definition of study populations, healthcare setting, design, and length of follow-up. However, overall, these data suggest a benefit of β-blockers, particularly in the early phase of therapy and in high-risk patients, with smaller or even absent effect beyond 1 year.

The results from our large-scale cohort study are in keeping with a potential benefit of β-blockers beyond 1 year of the AMI, albeit clearly less pronounced than that of statin therapy: the adjusted HRs of β-blocker discontinuation versus continuation were 1.17 (95% confidence interval, 1.01–1.35) and 1.13 (95% confidence interval, 0.94–1.36) for the composite outcome of death or readmission for ACS and for all-cause death, respectively.

Strengths and Limitations

Observational studies with time-varying exposure are particularly challenging, both with respect to the study design and the analytical method. The use of marginal structural models is a major strength of our study because they can appropriately correct for time-varying confounders affected by previous exposure and for informative censoring. This is not the case for conventional multivariate Cox models which, in particular, remove indirect effects mediated by these confounders, and it is, therefore, not surprising that the 2 approaches resulted in somewhat different effect estimates (for the composite outcome, the conventional model yielded an HR of 1.09). Marginal structural models are being increasingly applied in pharmacoepidemiologic studies,39 including studies on β-blockers and statins.40,41 However, they rely on the assumption of no unmeasured confounding—an assumption we discuss below.

With our study design, time-related bias, in particular immortal time bias, was avoided by (1) defining the study population based on information on a fixed time period (up to month 12 after index admission) and starting follow-up at a time point after this fixed time period (month 13), and (2) using a time-dependent exposure variable whose value at month m is defined by information up to m-1.42,43 Prevalent users may be a concern in study design.44 For the population, we studied prevalent user bias is per se limited because prior users and nonusers experienced AMI, and both were treated by β-blockers after AMI. Moreover, only less than one quarter of the patients had received β-blockers before the index episode, and subgroup analyses showed that the estimated effect of β-blocker discontinuation differed only slightly between prevalent users (HR, 1.14; 95% confidence interval, 0.87–1.51) and new users (HR, 1.20; 95% confidence interval, 1.02–1.42).

This real-life study experiences some fundamental limitations common to observational studies using healthcare databases. First, treatment was defined by the prescriptions filled-in, and we cannot formally exclude that some patients may have bought the medications but not actually taken them; such a bias would result in an underestimated difference between treatment discontinuation and treatment continuation. The dose of β-blockers could not be documented, and, therefore, we have no means to determine whether the patients still on β-blockers received an optimal dose; use of suboptimal doses might have attenuated the between-group difference; conversely, however, our study reflected the use of β-blockers, such as observed in real life. Second, detailed medical or psychosocial characterization of the patients is limited; thus, we could not separate between STEMI and non-STEMI because the ICD-10 classification only classifies infarctions into transmural and nontransmural infarctions; also, left ventricular ejection fraction was not available. In this regard, it is noteworthy that bisoprolol—a medication preferentially used in patients with left ventricular dysfunction—was the most widely used β-blocker; however, if some patients with poor left ventricular function were in fact included in our cohort, the bias thus created would tend to favor the efficacy of β-blockers and not the reverse. Third, the precise reasons for β-blockers or statins discontinuation were unknown. However, important cardiovascular risk factors, cardiovascular drug use, and comorbidities were searched for both at baseline and continuously during follow-up, and their association with treatment discontinuation was taken into account by the chosen analytical method. Of note, some risk factors increased the probability of β-blocker discontinuation, for example, male sex, whereas others decreased it, for example, obesity, diabetes mellitus, and cancer. For statin discontinuation, the direction of some associations was reversed, for example, diabetes mellitus and cancer increased the probability of statin discontinuation, whereas male sex decreased it.

Because of the observational nature of our study, however, residual confounding cannot be excluded. Adherence is a complex human behavior influenced by environmental and other factors only partially measured in medico-administrative databases, and some of these factors may be associated with death or untoward clinical events. In particular, only a minority of the patients (14.2% of patients and 5.9% of patient-months) stopped their β-blocker treatment, suggesting that this population may have specificities and some of them may not have been captured in our databases. Statin discontinuation was still rarer than β-blocker discontinuation (10.4% of patients and 3.1% of patient-months). Residual confounding might, therefore, explain why the benefit observed in our analyses of statin therapy was larger than expected: continuation of statin therapy was associated with a >50% reduction in events (including mortality) compared with statin discontinuation—a magnitude of effect much greater than findings in randomized statin trials after myocardial infarction (≈20% reduction in events).45,46 However, the comparison of the potential bias in the analyses of β-blockers and statins should be made with caution.

Finally, treatment discontinuation was defined as at least 4 consecutive months without exposure, and we, therefore, could not measure a potential short-term deleterious impact of β-blocker discontinuation. Because the follow-up of patients began 1 year after index AMI, the results of our study cannot be generalized to this first year.

Conclusions

The results of this study based on medico-administrative data suggest a potential benefit associated with maintaining β-blocker therapy beyond 1 year after AMI in patients <80 years, with no recent history of heart failure or ACS, and receiving optimal management, including both myocardial revascularization and all recommended secondary-prevention medications at the acute stage. However, unmeasured confounding might explain this result in light of the overestimated benefit of statin therapy, found in this study using a similar methodology, compared with that reported in randomized trials. Pending the results of contemporary randomized clinical trials assessing the role of β-blockers in this setting, observational studies from multiple sources, when carefully designed and analyzed, can provide useful information and contribute, together, to update guidelines on the use of β-blockers therapy after AMI.

Dr Danchin has received research grants from Amgen, AstraZeneca, Bayer, Boehringer-Ingelheim, Daiichi-Sankyo, Eli-Lilly, Glaxo-Smith-Kline, Merck Sharp & Dohme, Novartis, Pfizer, and Sanofi. He has received consulting or lecture fees from Amgen, AstraZeneca, Bayer, Boehringer-Ingelheim, Bristol-Myers Squibb, Daiichi-Sankyo, Eli-Lilly, GlaxoSmithKline, Merck Sharp & Dohme, Novo-Nordisk, Roche, Servier, and Sanofi. Drs Alla, Maura, Neumann, and Weill are employees of the French National Health Insurance and have no conflicts of interest with the pharmaceutical industry.

Footnotes

References

• 1. Jernberg T, Johanson P, Held C, Svennblad B, Lindbäck J, Wallentin L; SWEDEHEART/RIKS-HIA. Association between adoption of evidence-based treatment and survival for patients with ST-elevation myocardial infarction.JAMA. 2011; 305:1677–1684. doi: 10.1001/jama.2011.522.CrossrefMedlineGoogle Scholar
• 2. Puymirat E, Simon T, Steg PG, Schiele F, Guéret P, Blanchard D, Khalife K, Goldstein P, Cattan S, Vaur L, Cambou JP, Ferrières J, Danchin N; USIK USIC 2000 Investigators; FAST MI Investigators. Association of changes in clinical characteristics and management with improvement in survival among patients with ST-elevation myocardial infarction.JAMA. 2012; 308:998–1006. doi: 10.1001/2012.jama.11348.CrossrefMedlineGoogle Scholar
• 3. Freemantle N, Cleland J, Young P, Mason J, Harrison J. beta Blockade after myocardial infarction: systematic review and meta regression analysis.BMJ. 1999; 318:1730–1737.CrossrefMedlineGoogle Scholar
• 4. Van de Werf F, Bax J, Betriu A, Blomstrom-Lundqvist C, Crea F, Falk V, Filippatos G, Fox K, Huber K, Kastrati A, Rosengren A, Steg PG, Tubaro M, Verheugt F, Weidinger F, Weis M; ESC Committee for Practice Guidelines (CPG). Management of acute myocardial infarction in patients presenting with persistent ST-segment elevation: t he Task Force on the management of ST-segment elevation acute myocardial infarction of the European Society of Cardiology. Eur Heart J. 2008; 29:2909–2945. doi: 10.1093/eurheartj/ehn416.CrossrefMedlineGoogle Scholar
• 5. Montalescot G, Sechtem U, Achenbach S, Andreotti F, Arden C, Budaj A, Bugiardini R, Crea F, Cuisset T, Di Mario C, Ferreira JR, Gersh BJ, Gitt AK, Hulot JS, Marx N, Opie LH, Pfisterer M, Prescott E, Ruschitzka F, Sabaté M, Senior R, Taggart DP, van der Wall EE, Vrints CJM, Zamorano JL, Achenbach S, Baumgartner H, Bax JJ, Bueno H, Dean V, Deaton C, Erol C, Fagard R, Ferrari R, Hasdai D, Hoes AW, Kirchhof P, Knuuti J, Kolh P, Lancellotti P, Linhart A, Nihoyannopoulos P, Piepoli MF, Ponikowski P, Sirnes PA, Tamargo JL, Tendera M, Torbicki A, Wijns W, Windecker S, Knuuti J, Valgimigli M, Bueno H, Claeys MJ, Donner-Banzhoff N, Erol C, Frank H, Funck-Brentano C, Gaemperli O, Gonzalez-Juanatey JR, Hamilos M, Hasdai D, Husted S, James SK, Kervinen K, Kolh P, Kristensen SD, Lancellotti P, Maggioni AP, Piepoli MF, Pries AR, Romeo F, Rydén L, Simoons ML, Sirnes PA, Steg PG, Timmis A, Wijns W, Windecker S, Yildirir A, Zamorano JLTask Force Members; ESC Committee for Practice Guidelines; Document Reviewers. 2013 ESC guidelines on the management of stable coronary artery disease: the Task Force on the management of stable coronary artery disease of the European Society of Cardiology.Eur Heart J. 2013; 34:2949–3003. doi: 10.1093/eurheartj/eht296.CrossrefMedlineGoogle Scholar
• 6. Steg PG, James SK, Atar D, Badano LP, Blömstrom-Lundqvist C, Borger MA, Di Mario C, Dickstein K, Ducrocq G, Fernandez-Aviles F, Gershlick AH, Giannuzzi P, Halvorsen S, Huber K, Juni P, Kastrati A, Knuuti J, Lenzen MJ, Mahaffey KW, Valgimigli M, van’t Hof A, Widimsky P, Zahger D; Task Force on the Management of ST-Segment Elevation Acute Myocardial Infarction of the European Society of Cardiology (ESC). ESC Guidelines for the management of acute myocardial infarction in patients presenting with ST-segment elevation: t he Task Force on the management of ST-segment elevation.Eur Heart J. 2012; 33:2569–2619. doi: 10.1093/eurheartj/ehs215.CrossrefMedlineGoogle Scholar
• 7. O’Gara PT, Kushner FG, Ascheim DD, Casey DE, Chung MK, de Lemos JA, Ettinger SM, Fang JC, Fesmire FM, Franklin BA, Granger CB, Krumholz HM, Linderbaum JA, Morrow DA, Newby LK, Ornato JP, Ou N, Radford MJ, Tamis-Holland JE, Tommaso CL, Tracy CM, Woo YJ, Zhao DX, Anderson JL, Jacobs AK, Halperin JL, Albert NM, Brindis RG, Creager MA, DeMets D, Guyton RA, Hochman JS, Kovacs RJ, Kushner FG, Ohman EM, Stevenson WG, Yancy CW; American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. 2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines.Circulation. 2013; 127:e362–e425. doi: 10.1161/CIR.0b013e3182742cf6.LinkGoogle Scholar
• 8. Rossello X, Pocock SJ, Julian DG. Long-term use of cardiovascular drugs: challenges for research and for patient care.J Am Coll Cardiol. 2015; 66:1273–1285. doi: 10.1016/j.jacc.2015.07.018.CrossrefMedlineGoogle Scholar
• 9. 2007 Writing Committee Members; Presidents and Staff; American College of Cardiology Foundation; American College of Cardiology Foundation/American Heart Association; American Heart Association. 2012 ACCF/AHA focused update incorporated into the ACCF/AHA 2007 guidelines for the management of patients with unstable angina/non-ST-elevation myocardial infarction: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines.Circulation. 2013; 127:e663–e828. doi: 10.1161/CIR.0b013e31828478ac.AbstractGoogle Scholar
• 10. Roffi M, Patrono C, Collet JP, Mueller C, Valgimigli M, Andreotti F, Bax JJ, Borger MA, Brotons C, Chew DP, Gencer B, Hasenfuss G, Kjeldsen K, Lancellotti P, Landmesser U, Mehilli J, Mukherjee D, Storey RF, Windecker S, Baumgartner H, Gaemperli O, Achenbach S, Agewall S, Badimon L, Baigent C, Bueno H, Bugiardini R, Carerj S, Casselman F, Cuisset T, Erol Ç, Fitzsimons D, Halle M, Hamm C, Hildick-Smith D, Huber K, Iliodromitis E, James S, Lewis BS, Lip GY, Piepoli MF, Richter D, Rosemann T, Sechtem U, Steg PG, Vrints C, Luis Zamorano J; Management of Acute Coronary Syndromes in Patients Presenting Without Persistent ST-Segment Elevation of the European Society of Cardiology. 2015 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: Task Force for the Management of Acute Coronary Syndromes in Patients Presenting without Persistent ST-Segment Elevation of the European Society of Cardiology (ESC).Eur Heart J. 2016; 37:267–315. doi: 10.1093/eurheartj/ehv320.CrossrefMedlineGoogle Scholar
• 11. Bangalore S, Makani H, Radford M, Thakur K, Toklu B, Katz SD, DiNicolantonio JJ, Devereaux PJ, Alexander KP, Wetterslev J, Messerli FH. Clinical outcomes with β-blockers for myocardial infarction: a meta-analysis of randomized trials.Am J Med. 2014; 127:939–953. doi: 10.1016/j.amjmed.2014.05.032.CrossrefMedlineGoogle Scholar
• 12. Bangalore S, Steg G, Deedwania P, Crowley K, Eagle KA, Goto S, Ohman EM, Cannon CP, Smith SC, Zeymer U, Hoffman EB, Messerli FH, Bhatt DL; REACH Registry Investigators. β-Blocker use and clinical outcomes in stable outpatients with and without coronary artery disease.JAMA. 2012; 308:1340–1349. doi: 10.1001/jama.2012.12559.CrossrefMedlineGoogle Scholar
• 13. Rasmussen JN, Chong A, Alter DA. Relationship between adherence to evidence-based pharmacotherapy and long-term mortality after acute myocardial infarction.JAMA. 2007; 297:177–186. doi: 10.1001/jama.297.2.177.CrossrefMedlineGoogle Scholar
• 14. Andersson C, Mérie C, Jørgensen M, Gislason GH, Torp-Pedersen C, Overgaard C, Køber L, Jensen PF, Hlatky MA. Association of β-blocker therapy with risks of adverse cardiovascular events and deaths in patients with ischemic heart disease undergoing noncardiac surgery: a Danish nationwide cohort study.JAMA Intern Med. 2014; 174:336–344. doi: 10.1001/jamainternmed.2013.11349.CrossrefMedlineGoogle Scholar
• 15. Choo EH, Chang K, Ahn Y, Jeon DS, Lee JM, Kim DB, Her SH, Park CS, Kim HY, Yoo KD, Jeong MH, Seung KB. Benefit of β-blocker treatment for patients with acute myocardial infarction and preserved systolic function after percutaneous coronary intervention.Heart. 2014; 100:492–499. doi: 10.1136/heartjnl-2013-305137.CrossrefMedlineGoogle Scholar
• 16. Puymirat E, Riant E, Aissaoui N, Soria A, Ducrocq G, Coste P, Cottin Y, Aupetit JF, Bonnefoy E, Blanchard D, Cattan S, Steg G, Schiele F, Ferrières J, Juillière Y, Simon T, Danchin N. β blockers and mortality after myocardial infarction in patients without heart failure: multicentre prospective cohort study.BMJ. 2016; 354:i4801. doi: 10.1136/bmj.i4801.CrossrefMedlineGoogle Scholar
• 17. Kezerashvili A, Marzo K, De Leon J. Beta blocker use after acute myocardial infarction in the patient with normal systolic function: when is it “ok” to discontinue?Curr Cardiol Rev. 2012; 8:77–84.CrossrefMedlineGoogle Scholar
• 18. Danchin N, Laurent S. Coronary artery disease. Are β-blockers truly helpful in patients with CAD?Nat Rev Cardiol. 2013; 10:11–12. doi: 10.1038/nrcardio.2012.159.CrossrefMedlineGoogle Scholar
• 19. Tuppin P, Rudant J, Constantinou P, Gastaldi-Ménager C, Rachas A, de Roquefeuil L, Maura G, Caillol H, Tajahmady A, Coste J, Gissot C, Weill A, Fagot-Campagna A. Value of a national administrative database to guide public decisions: from the système national d’information interrégimes de l’Assurance Maladie (SNIIRAM) to the système national des données de santé (SNDS) in France.Rev Epidemiol Sante Publique. 2017; 65(suppl 4):S149–S167. doi: 10.1016/j.respe.2017.05.004.CrossrefMedlineGoogle Scholar
• 20. Neumann A, Weill A, Ricordeau P, Fagot JP, Alla F, Allemand H. Pioglitazone and risk of bladder cancer among diabetic patients in France: a population-based cohort study.Diabetologia. 2012; 55:1953–1962. doi: 10.1007/s00125-012-2538-9.CrossrefMedlineGoogle Scholar
• 21. Maura G, Blotière PO, Bouillon K, Billionnet C, Ricordeau P, Alla F, Zureik M. Comparison of the short-term risk of bleeding and arterial thromboembolic events in nonvalvular atrial fibrillation patients newly treated with dabigatran or rivaroxaban versus vitamin K antagonists: a French nationwide propensity-matched cohort study.Circulation. 2015; 132:1252–1260. doi: 10.1161/CIRCULATIONAHA.115.015710.LinkGoogle Scholar
• 22. Neumann A, Maura G, Weill A, Ricordeau P, Alla F, Allemand H. Comparative effectiveness of rosuvastatin versus simvastatin in primary prevention among new users: a cohort study in the French national health insurance database.Pharmacoepidemiol Drug Saf. 2014; 23:240–250. doi: 10.1002/pds.3544.CrossrefMedlineGoogle Scholar
• 23. Rey G, Jougla E, Fouillet A, Hémon D. Ecological association between a deprivation index and mortality in France over the period 1997 - 2001: variations with spatial scale, degree of urbanicity, age, gender and cause of death.BMC Public Health. 2009; 9:33. doi: 10.1186/1471-2458-9-33.CrossrefMedlineGoogle Scholar
• 24. Hernán MA, Brumback B, Robins JM. Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men.Epidemiol Camb Mass. 2000; 11:561–570.CrossrefMedlineGoogle Scholar
• 25. Cole SR, Hernán MA. Constructing inverse probability weights for marginal structural models.Am J Epidemiol. 2008; 168:656–664. doi: 10.1093/aje/kwn164.CrossrefMedlineGoogle Scholar
• 26. Westreich D, Cole SR, Tien PC, Chmiel JS, Kingsley L, Funk MJ, Anastos K, Jacobson LP. Time scale and adjusted survival curves for marginal structural cox models.Am J Epidemiol. 2010; 171:691–700. doi: 10.1093/aje/kwp418.CrossrefMedlineGoogle Scholar
• 27. Kalra PR, Morley C, Barnes S, Menown I, Kassianos G, Padmanabhan S, Gupta S, Lang CC. Discontinuation of beta-blockers in cardiovascular disease: UK primary care cohort study.Int J Cardiol. 2013; 167:2695–2699. doi: 10.1016/j.ijcard.2012.06.116.CrossrefMedlineGoogle Scholar
• 28. Kramer JM, Hammill B, Anstrom KJ, Fetterolf D, Snyder R, Charde JP, Hoffman BS, Allen LaPointe N, Peterson E. National evaluation of adherence to beta-blocker therapy for 1 year after acute myocardial infarction in patients with commercial health insurance.Am Heart J. 2006; 152:454.e1–454.e8. doi: 10.1016/j.ahj.2006.02.030.CrossrefGoogle Scholar
• 29. Jackevicius CA, Li P, Tu JV. Prevalence, predictors, and outcomes of primary nonadherence after acute myocardial infarction.Circulation. 2008; 117:1028–1036. doi: 10.1161/CIRCULATIONAHA.107.706820.LinkGoogle Scholar
• 30. Bagnall AJ, Yan AT, Yan RT, Lee CH, Tan M, Baer C, Polasek P, Fitchett DH, Langer A, Goodman SG; Canadian Acute Coronary Syndromes Registry II Investigators. Optimal medical therapy for non-ST-segment-elevation acute coronary syndromes: exploring why physicians do not prescribe evidence-based treatment and why patients discontinue medications after discharge.Circ Cardiovasc Qual Outcomes. 2010; 3:530–537. doi: 10.1161/CIRCOUTCOMES.109.919415.LinkGoogle Scholar
• 31. Bally K, Buechel RR, Buser P, Tschudi P, Martina B, Zeller A. Discontinuation of secondary prevention medication after myocardial infarction - the role of general practitioners and patients.Swiss Med Wkly. 2013; 143:w13896. doi: 10.4414/smw.2013.13896.MedlineGoogle Scholar
• 32. Viscoli CM, Horwitz RI, Singer BH. Beta-blockers after myocardial infarction: influence of first-year clinical course on long-term effectiveness.Ann Intern Med. 1993; 118:99–105.CrossrefMedlineGoogle Scholar
• 33. Chen ZM, Pan HC, Chen YP, Peto R, Collins R, Jiang LX, Xie JX, Liu LS; COMMIT (Clopidogrel and Metoprolol in Myocardial Infarction Trial) Collaborative Group. Early intravenous then oral metoprolol in 45,852 patients with acute myocardial infarction: randomised placebo-controlled trial.Lancet. 2005; 366:1622–1632. doi: 10.1016/S0140-6736(05)67661-1.CrossrefMedlineGoogle Scholar
• 34. Gottlieb SS, McCarter RJ, Vogel RA. Effect of beta-blockade on mortality among high-risk and low-risk patients after myocardial infarction.N Engl J Med. 1998; 339:489–497. doi: 10.1056/NEJM199808203390801.CrossrefMedlineGoogle Scholar
• 35. Emery M, López-Sendón J, Steg PG, Anderson FA, Dabbous OH, Scheuble A, Eagle KA; GRACE Investigators. Patterns of use and potential impact of early beta-blocker therapy in non-ST-elevation myocardial infarction with and without heart failure: the Global Registry of Acute Coronary Events.Am Heart J. 2006; 152:1015–1021. doi: 10.1016/j.ahj.2006.08.024.CrossrefMedlineGoogle Scholar
• 36. Ho PM, Spertus JA, Masoudi FA, Reid KJ, Peterson ED, Magid DJ, Krumholz HM, Rumsfeld JS. Impact of medication therapy discontinuation on mortality after myocardial infarction.Arch Intern Med. 2006; 166:1842.CrossrefMedlineGoogle Scholar
• 37. Korhonen MJ, Robinson JG, Annis IE, Hickson RP, Bell JS, Hartikainen J, Fang G. Adherence tradeoff to multiple preventive therapies and all-cause mortality after acute myocardial infarction.J Am Coll Cardiol. 2017; 70:1543–1554. doi: 10.1016/j.jacc.2017.07.783.CrossrefMedlineGoogle Scholar
• 38. Chung J, Han J, Kim YJ, Kim CJ, Ahn Y, Chan Cho M, Chae SC, Chae I, Chae JK, Seong I, Yang H, Park K, Kang H, Koo B, Jeong MH, Kim H; Investigators for Korea Acute Myocardial Infarction Registry (KAMIR). Benefit of vasodilating β-blockers in patients with acute myocardial infarction after percutaneous coronary intervention: Nationwide Multicenter Cohort Study.J Am Heart Assoc. 2017; 6:e007063. doi: 10.1161/JAHA.117.007063.LinkGoogle Scholar
• 39. Yang S, Eaton CB, Lu J, Lapane KL. Application of marginal structural models in pharmacoepidemiologic studies: a systematic review.Pharmacoepidemiol Drug Saf. 2014; 23:560–571. doi: 10.1002/pds.3569.CrossrefMedlineGoogle Scholar
• 40. Shen L, Shah BR, Reyes EM, Thomas L, Wojdyla D, Diem P, Leiter LA, Charbonnel B, Mareev V, Horton ES, Haffner SM, Soska V, Holman R, Bethel MA, Schaper F, Sun JL, McMurray JJ, Califf RM, Krum H. Role of diuretics, β blockers, and statins in increasing the risk of diabetes in patients with impaired glucose tolerance: reanalysis of data from the NAVIGATOR study.BMJ. 2013; 347:f6745. doi: 10.1136/bmj.f6745.CrossrefMedlineGoogle Scholar
• 41. Lavikainen P, Helin-Salmivaara A, Eerola M, Fang G, Hartikainen J, Huupponen R, Korhonen MJ. Statin adherence and risk of acute cardiovascular events among women: a cohort study accounting for time-dependent confounding affected by previous adherence.BMJ Open. 2016; 6:e011306. doi: 10.1136/bmjopen-2016-011306.CrossrefMedlineGoogle Scholar
• 42. Suissa S. Immortal time bias in pharmaco-epidemiology.Am J Epidemiol. 2008; 167:492–499. doi: 10.1093/aje/kwm324.CrossrefMedlineGoogle Scholar
• 43. Suissa S, Azoulay L. Metformin and the risk of cancer: time-related biases in observational studies.Diabetes Care. 2012; 35:2665–2673. doi: 10.2337/dc12-0788.CrossrefMedlineGoogle Scholar
• 44. Danaei G, Tavakkoli M, Hernán MA. Bias in observational studies of prevalent users: lessons for comparative effectiveness research from a meta-analysis of statins.Am J Epidemiol. 2012; 175:250–262. doi: 10.1093/aje/kwr301.CrossrefMedlineGoogle Scholar
• 45. Gutierrez J, Ramirez G, Rundek T, Sacco RL. Statin therapy in the prevention of recurrent cardiovascular events: a sex-based meta-analysis.Arch Intern Med. 2012; 172:909–919. doi: 10.1001/archinternmed.2012.2145.CrossrefMedlineGoogle Scholar
• 46. Baigent C, Keech A, Kearney PM, Blackwell L, Buck G, Pollicino C, Kirby A, Sourjina T, Peto R, Collins R, Simes R; Cholesterol Treatment Trialists’ (CTT) Collaborators. Efficacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins.Lancet. 2005; 366:1267–1278. doi: 10.1016/S0140-6736(05)67394-1.CrossrefMedlineGoogle Scholar