Chen J, Radford MJ, Wang Y, Marciniak TA, Krumholz HM. Are β-Blockers Effective in Elderly Patients Who Undergo Coronary Revascularization After Acute Myocardial Infarction?. Arch Intern Med. 2000;160(7):947-952. doi:10.1001/archinte.160.7.947
Although randomized clinical trials have demonstrated that β-blocker therapy is effective in reducing mortality after acute myocardial infarction (AMI), many of these studies excluded patients who undergo coronary revascularization. However, the clinical practice guidelines established by the American College of Cardiology and the American Heart Association recommend that β-blocker therapy be considered for patients who underwent successful revascularization after AMI.
Using data from the Cooperative Cardiovascular Project, we compared the initiation of β-blocker therapy at discharge in patients aged 65 years or older who underwent coronary artery bypass surgery (CABG) or percutaneous transluminal coronary angioplasty (PTCA) during their hospitalization for AMI with that of patients who did not undergo revascularization. We then examined whether β-blocker therapy was associated with lower 1-year mortality between revascularized and nonrevascularized groups.
After excluding patients with contraindications to β-blocker therapy, 84 457 patients remained in the study sample. Of these, 8482 patients underwent CABG, and 13 997 patients underwent PTCA. After adjusting for demographic and clinical factors, we found that these patients were less likely to initiate β-blocker therapy after CABG (odds ratio [OR], 0.44; 95% confidence interval [CI], 0.41-0.47) or PTCA (OR, 0.89; 95% CI, 0.85-0.93) relative to the nonrevascularized group. After adjusting for potential confounders, β-blockers were significantly associated with lower 1-year mortality in patients who underwent CABG (hazard ratio [HR], 0.70; 95% CI, 0.55-0.89) or PTCA (HR, 0.86; 95% CI, 0.74-1.00), similar to that of the nonrevascularized group (HR, 0.83; 95% CI, 0.80-0.87).
Therapy after AMI with β-blockers appears to be as effective in reducing 1-year mortality for elderly patients who have undergone CABG or PTCA as for a nonrevascularized group. Our findings suggest that routine use of β-blockers should be considered for patients who undergo revascularization after AMI.
MANY RANDOMIZED clinical trials have demonstrated the efficacy of β-blockers in reducing mortality and reinfarction after acute myocardial infarction (AMI).1 However, several of these trials, including the β-Blocker Heart Attack Trial,2 the Göteborg Metoprolol Trial,3 and the Norwegian Timolol Study,4 excluded patients who had undergone coronary revascularization after presenting with AMI. As a result, there are few data available on the effectiveness of β-blockers after AMI in patients undergoing revascularization.
The clinical practice guidelines for AMI established by the American College of Cardiology (ACC) and the American Heart Association (AHA) acknowledge that evidence of benefit is lacking. The ACC-AHA guidelines say that, "although no study has determined if long-term β-adrenoceptor blocker therapy should be administered to survivors of MI who subsequently have successfully undergone revascularization, there is no reason to believe that these agents act differently in patients who have undergone revascularization."5 Given the substantial number of patients referred for revascularization after AMI, there is a need to assess whether β-blocker therapy is effective in these patients.
To evaluate this issue, we examined whether elderly patients who undergo revascularization were less likely to be prescribed β-blocker therapy at discharge for AMI, and if prescription of β-blockers at discharge was independently associated with improved 1-year survival. To address these questions, we used data from the Cooperative Cardiovascular Project (CCP), a Health Care Financing Administration initiative designed to examine the patterns of care and to improve the outcomes of Medicare beneficiaries with AMI.
Subjects in the CCP were identified from hospital bills in the Medicare National Claims History File (form UB-92) that included claims submitted for patients treated under fee-for-service plans but not under Medicare-managed care contracts.6 Patients discharged from acute care hospitals in all 50 states and Puerto Rico with a principal discharge diagnosis code of AMI (International Classification of Diseases, Ninth Revision, Clinical Modification code 410)7 were initially selected, except those indicating a subsequent, nonacute treatment of AMI (fifth digit of International Classification of Diseases, Ninth Revision, Clinical Modification code, 2). Sampling of the CCP population was primarily conducted for approximately 8 months for each state from February 1, 1994, through July 31, 1995, except for the CCP pilot study states (Alabama, Connecticut, Iowa, and Wisconsin), which were sampled from August 1 through November 30, 1995. Predefined variables were abstracted from copies of hospital records. Data reliability was monitored through monthly random reabstractions, with overall agreement for treatment variables averaging more than 95%.8
The study sample was restricted to patients aged 65 years or older with a confirmed AMI who were discharged alive from the hospital. We defined a confirmed AMI as a discharge diagnosis of AMI accompanied by chart documentation of creatine kinase–MB fraction of greater than 0.05, lactate dehydrogenase (LDH) level of greater than 1.5 times normal and LD1 proportion greater than that of LD2, or 2 of the following 3 criteria: chest pain, doubling of the creatine kinase level above each particular hospital's level for the upper limit of normal, or a new AMI on the official electrocardiogram report. We included only the patient's first admission during the sample period to avoid double counting. Patients who subsequently were transferred to another acute care institution were excluded because we were unable to determine their discharge medications. We also excluded patients considered to have a terminal illness (chart documented as unlikely to live past 6 months) or metastatic cancer, since the focus of their treatment may not have been targeted toward a survival benefit.
We restricted the study sample to patients without strong contraindications to β-blocker therapy based on the clinical guidelines established by the ACC-AHA for the treatment of AMI. Contraindications included bradycardia (last recorded heart rate before discharge, <50 beats/min), low blood pressure (last recorded systolic blood pressure before discharge, <100 mm Hg), high-grade (second- or third-degree) atrioventricular block, chronic obstructive pulmonary disease, asthma, or chart-documented intolerance to β-blockers. Of the 234 769 patients in the initial CCP sample, 84 457 remained eligible for this analysis (Table 1).
The primary study variables included prescription of β-blocker therapy at discharge, mortality 1 year after discharge, and coronary revascularization in the hospital. Dates of death were obtained from the Medicare Enrollment Database. This database was derived from discharge dates of billing records indicating in-hospital death, and from the Social Security Administration Master Beneficiary Record.9 We excluded 3 patients in whom 1-year mortality could not be confirmed or dated. We identified all patients receiving an oral β-blocker as a discharge medication, excluding topical β-blockers. We categorized coronary revascularization during the index hospitalization as coronary artery bypass surgery (CABG) or percutaneous transluminal coronary angioplasty (PTCA) based on medical chart review. There were 1047 patients who received both procedures. Because it was likely that most of these patients underwent bypass surgery as a result of failed PTCA, they were included in the CABG category.
Additional abstracted variables included demographics (aged 65-74, 75-84, and ≥85 years; sex; and race), comorbidities (hypertension, current smoker, previous MI, congestive heart failure, CABG, PTCA, peripheral vascular disease, stroke, dementia, and diabetes), clinical characteristics (anterior infarction, left bundle-branch block, atrial fibrillation or flutter, respiratory rate >25 breaths/min on admission, serum urea nitrogen level >14.3 mmol/L [>40 mg/dL] or creatinine level >221.0 µmol/L [>2.5 mg/dL], albumin level <0.03 g/L, prothrombin time >16 seconds, hematocrit <0.3, and stroke on admission), hospital treatment and course (thrombolytic therapy, recurrent chest pain, and creatine kinase level >4 times normal), functional status (urinary incontinence and inability to ambulate), discharge disposition (home, skilled nursing facility, or other), discharge medications (aspirin, calcium channel blockers, angiotensin-converting enzyme inhibitors, loop diuretics, or digoxin), physician specialty, census region, and length of stay. We adjusted for severity of heart failure during the hospitalization using several variables, including clinical diagnosis of heart failure, radiological signs of heart failure, rales, gallop, shock, and estimated left ventricular ejection fraction (LVEF, as determined by radionuclide ventriculography studies, cardiac catheterization, and echocardiogram, prioritized in that order). Specialty of the admitting physician was classified as cardiology, internal medicine, family or general practice, or other by linking the attending physician's Unique Physician Identification Number from Medicare Part A claims to a directory of physician-reported specialties maintained by the Health Care Financing Administration, as in a previous report.10 We also classified patients according to the 9 major US census regions, based on the site of hospitalization.11 Length of stay was coded as greater than 12 days (yes or no), approximately the 75th percentile. For variables with more than 3% missing values (eg, albumin level), a dummy variable was created and included in the multivariate analyses.
We evaluated differences in baseline clinical characteristics of patients undergoing or not undergoing revascularization using the χ2 test for categorical variables and analysis of variance for continuous variables. In the first part of the study, we compared the use of β-blocker therapy in patients who underwent PTCA or CABG with those of the nonrevascularized group. We used bivariate and multivariate logistic regression analyses to determine whether prescribed use of β-blockers differed between revascularized and nonrevascularized groups. Variables were included in the multivariate models on the basis of backward stepwise selection with prescription of β-blockers as the dependent variable, as described previously.12 The analysis was repeated among patients with no documentation of receiving β-blockers before admission.
In the second part of the study, we examined the relationship between β-blocker therapy and 1-year mortality. We first evaluated the unadjusted relationship between β-blocker therapy and 1-year mortality for each of the 3 patient categories (nonrevascularization, PTCA, and CABG). Because patients receiving β-blocker therapy were more likely to have fewer comorbid conditions and lower severity of illness than those not receiving therapy, we conducted analyses that adjusted for potential confounders using Cox proportional hazards models. Covariates used in the model were selected on the basis of clinical relevance and previous work.12 We evaluated the assumption of proportionality graphically and found the results satisfactory. We also tested the hypothesis that β-blocker effectiveness differed between revascularized and nonrevascularized groups using a combined interaction model. In secondary analyses, we examined the association between prescription of β-blockers and 1-year mortality stratified by several subgroups (presence of heart failure during the hospitalization, ejection fraction, age, and history of MI). Although underpowered to detect whether β-blocker effectiveness differed within some of these subgroups, we sought to confirm consistency of the overall association between β-blocker therapy and mortality and to examine whether such therapy was associated with any overt harm in particular groups of patients.
All statistical calculations were performed using Stata 5.0 statistical software (StataCorp, College Station, Tex).
Of the 84 457 patients without contraindications to β-blocker therapy, 22 479 (26.6%) underwent coronary revascularization during their hospitalization. Of those patients, 8482 (10.0%) received CABG, and 13 997 (16.6%) received PTCA. The remaining 61 978 patients (73.4%) did not undergo revascularization. There were significant differences in baseline characteristics among revascularized and nonrevascularized groups (Table 2). On average, patients in the revascularized group were younger and had fewer comorbid conditions (previous AMI, heart failure, or stroke; diabetes mellitus; dementia; inability to ambulate; urinary incontinence; serum urea nitrogen level >14.3 mmol/L [>40 mg/dL]; or creatinine level >221.0 µmol/L [>2.5 mg/dL]) than in the nonrevascularized group. However, patients in the revascularized group were more likely to have anterior infarction, recurrent chest pain, prescribed use of aspirin at discharge, or a cardiologist as the admitting physician or to be discharged to home.
We found that 33.1% of the patients undergoing CABG, 51.4% of the patients undergoing PTCA, and 42.0% of the nonrevascularized group were prescribed β-blockers at discharge (P<.001). In the adjusted analysis, patients who had undergone CABG or PTCA were less likely to receive β-blockers at discharge (odds ratio [OR], 0.44; 95% confidence interval [CI], 0.41-0.47; and OR, 0.89; 95% CI, 0.85-0.93, respectively). The factors most strongly associated with β-blocker therapy at discharge were use of aspirin at discharge, history of hypertension, recurrent chest pain, and use of thrombolytic therapy. The strongest predictors of not receiving β-blockers were discharge prescriptions for calcium channel blockers, angiotensin-converting enzyme inhibitors, loop diuretics, or digoxin; discharge other than to home; increasing age; measured LVEF of less than 20%; and heart failure during the hospitalization. The area under the receiver operating characteristic curve was 0.74, indicating good discriminant ability of the model.13
Of the 68 340 patients not receiving β-blocker therapy before admission, 24 510 (35.9%) were prescribed β-blocker therapy at discharge. Overall, 28.0% of patients who had CABG and 45.2% of patients who had PTCA were discharged with prescriptions for β-blockers compared with 34.8% of the nonrevascularized group (P<.001). After adjusting for potential confounders, patients undergoing revascularization were less likely to receive β-blockers at discharge after CABG (OR, 0.41; 95% CI, 0.38-0.44) and PTCA (OR, 0.90; 95% CI, 0.86-0.95) compared with the nonrevascularized group.
The 1-year mortality rate for all patients prescribed β-blocker therapy at discharge was 12.3% compared with 23.6% for patients who were not prescribed therapy (P<.001). In the unadjusted analysis, β-blockers were associated with lower mortality among patients undergoing CABG (3.4% vs 5.9% [P<.001]), PTCA (4.6% vs 7.5% [P<.001]), and no revascularization (15.4% vs 29.4% [P<.001]). In the adjusted analyses, β-blocker therapy was significantly associated with lower mortality in patients who underwent CABG (hazard ratio [HR], 0.70; 95% CI, 0.55-0.89) or PTCA (HR, 0.86; 95% CI, 0.74-1.00) as well as in the nonrevascularized group (HR, 0.83; 95% CI, 0.80-0.87). In models that examined the interaction between β-blocker therapy and revascularization, there were no significant differences in β-blocker efficacy in patients who underwent CABG or PTCA compared with patients in the nonrevascularized group.
In the secondary analyses, we found that in the nonrevascularized group, β-blocker therapy was significantly associated with lower 1-year mortality when stratifying patients by presence of heart failure during the hospitalization (yes or no), LVEF (<35% or ≥35%), age (65-74 or ≥75 years), or history of myocardial infarction (yes or no). Among patients who underwent PTCA or CABG, β-blockers were significantly associated with lower mortality in those with heart failure, aged 65 to 74 years, or without previous MI. Although not reaching statistical significance in the remaining categories for the revascularized group, all of the risk ratio estimates were at or below 1.00, suggesting that there was no evidence that β-blockers were associated with harm for patients in these subgroups.
We compared the use and effectiveness of β-blocker therapy among elderly patients who underwent revascularization after AMI compared with a nonrevascularized group. We found that the prescribed use of β-blocker therapy was associated with reductions in 1-year mortality among patients who underwent CABG or PTCA to an extent similar to that in patients who did not undergo revascularization. The rate of β-blocker use was low among revascularized and nonrevascularized groups, as previously reported.12
Is it possible that β-blocker effectiveness differs in patients treated vs those not treated with revascularization? The risk reduction for CABG-treated patients prescribed β-blockers appeared to be slightly greater than that for the nonrevascularized group, although our study was underpowered to detect whether this difference in the magnitude of the benefit was statistically significant. Many studies have demonstrated that β-blockers reduce the incidence of atrial fibrillation and other supraventricular arrhythmias after CABG, a common complication of cardiac surgery,14,15 presumably by blunting the hyperadrenergic state associated with the stress response from CABG surgery.16 This protective mechanism may explain additional benefits from β-blocker therapy in CABG-treated patients by reducing the risk for cardiac arrhythmia or its complications in the postoperative period. In contrast, there is little evidence that β-blockers provide a benefit to patients who undergo PTCA beyond that expected for patients who do not undergo revascularization. In a study of 455 patients who received β-blockers and 483 patients who did not receive therapy after PTCA, no association was found between β-blockers and increased success rate during the procedure or the subsequent rate of complications or restenosis.17
Regardless of relative differences in β-blocker effectiveness, physicians would still need to treat more patients undergoing revascularization with β-blockers to save 1 life because of a smaller absolute reduction in mortality. Using CCP mortality rates as examples and assuming the relative reduction in 1-year mortality from β-blockers is 20%, 77 patients undergoing revascularization would need to be treated with β-blockers to save 1 life by the end of 1 year (assuming a reduction in risk from 6.5% to 5.2%, for an absolute difference of 1.3%), compared with 17 patients not undergoing revascularization (risk reduction from 30.0% to 24.0%, an absolute difference of 6.0%). Although the number needed to treat to save 1 life is higher for patients undergoing revascularization, many physicians may still find that this ratio favors treatment.
Few randomized controlled trials have examined the long-term outcome of patients receiving β-blocker therapy after coronary revascularization. The Metoprolol After Coronary Bypass study18 examined 967 Swedish survivors of CABG who were randomized to metoprolol tartrate or placebo for 2 years, with a composite end point of mortality, reinfarction, development of unstable angina, or need for repeated revascularization. There were no significant differences between treatment and placebo groups for the composite end point (8.8% vs 8.0%, respectively [P=.73]) or for mortality alone (3.3% vs 1.8%, respectively [P=.16]). The authors caution, however, that their negative findings may have resulted from the study's low event rate or the fact that a large proportion of patients withdrew from blind treatment.
Among observational studies, our findings are consistent with those of a study that used the CCP data to examine the association between β-blockers and mortality in high- and low-risk patients,19 although our estimates of the β-blocker effectiveness differ. Gottlieb et al19 found β-blockers associated with a relative risk of 0.60 (95% CI, 0.57-0.63) for patients who underwent PTCA or CABG, a benefit greater than that found in our study. One potential explanation for this difference is the presence of confounding. The earlier study included patients with relative contraindications to therapy, without detailed adjustment for severity of these conditions, such as chronic obstructive pulmonary disease or heart failure. Our study took several steps to reduce potential confounding. First, we restricted our sample to patients without contraindications to β-blockers to minimize confounding due to differential severity of contraindications. Second, we included several variables to control specifically for severity of heart failure. As a result, our estimate of β-blocker effectiveness is close to those found in the clinical trials.1,20
Our study has several limitations. Although the use of β-blockers was based on retrospective chart review, we do not have information regarding which patients continued therapy during the year after discharge. Patients who underwent CABG or PTCA may have been less likely to continue β-blocker therapy, perhaps believing that revascularization was a definitive cure. However, the resulting misclassification would have biased the association between β-blockers and survival toward the null, underestimating the observed associations. In addition, our ability to adjust for confounding risk factors was limited by the information documented in the medical charts. Although we used a comprehensive set of variables to examine patients at similar likelihood of receiving β-blockers and similar severity of clinical presentation, we cannot exclude the possibility that the use of β-blockers was associated with additional factors not found in the CCP that were associated with decreased mortality. Last, this study was unable to determine whether the optimal duration of β-blocker therapy differs between patients undergoing and those not undergoing revascularization.
In a nationwide community-based cohort, we demonstrated that elderly patients who undergo CABG or PTCA after AMI were significantly less likely to be prescribed β-blocker therapy at discharge. Our findings suggest that β-blocker therapy was as effective in reducing 1-year mortality for patients who undergo revascularization (CABG or PTCA) as for patients not undergoing revascularization. Because β-blocker use was low among both groups of elderly patients, this represents an opportunity to improve the care and outcomes for both groups. Our findings provide evidence supporting the ACC-AHA practice guidelines that β-blockers after AMI are as effective in patients undergoing revascularization as in those not undergoing such procedures. Although additional randomized trials would be needed to confirm our results, this study supports the ACC-AHA recommendation that β-blockers routinely be considered for patients with AMI who undergo CABG or PTCA.
Accepted for publication June 29, 1999.
The analyses on which this publication is based were performed under contract 500-96-P549, entitled "Utilization and Quality Control Peer Review Organization for the State of Connecticut," sponsored by the Health Care Financing Administration, Department of Health and Human Services, Washington, DC.
Mr Chen is a Merck/American Federation for Aging Research Student Scholar in Geriatric Pharmacology and an American Heart Association Student Scholar in Cardiovascular Disease and Stroke. Dr Krumholz is a Paul Beeson Faculty Scholar. This article is a direct result of the Health Care Quality Improvement Program initiated by the Health Care Financing Administration, Baltimore, Md, which has encouraged identification of quality improvement projects derived from analysis of patterns of care, and therefore required no special funding on the part of this contractor. Ideas and contributions to the authors concerning experience in engaging with issues presented are welcomed.
The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. The authors assume full responsibility for the accuracy and completeness of the ideas presented herein.
Reprints: Harlan M. Krumholz, MD, Yale University School of Medicine, 333 Cedar St, PO Box 208025, New Haven, CT 06520-8025 (e-mail: email@example.com).