Communitywide Trends in the Use and Outcomes Associated With β-Blockers in Patients With Acute Myocardial Infarction: The Worcester Heart Attack Study

Background  Despite the benefits associated with β-blocker therapy in patients with acute myocardial infarction (AMI), limited recent data are available describing the extent of use of this therapy and the associated hospital and long-term outcomes, particularly from the perspective of a population-based study. Data are also limited about the characteristics of patients with AMI who do not receive β-blockers. This study examines more than 2 decades of trends in the use of β-blockers in hospitalized patients with AMI.

Methods  Communitywide study of 10 374 patients hospitalized with confirmed AMI in all metropolitan Worcester hospitals during 12 annual periods between 1975 and 1999.

Results  There was a marked increase in the use of β-blockers in hospitalized patients between 1975 (11%) and 1999 (82%). Older patients, women, and patients with comorbidities were significantly less likely to be treated with β-blockers. After controlling for other prognostic factors, patients treated with β-blockers were less likely to develop heart failure (adjusted odds ratio [OR], 0.58; 95% confidence interval [CI], 0.53-0.63), cardiogenic shock (OR, 0.46; 95% CI, 0.39-0.54), and primary ventricular fibrillation (OR, 0.84; 95% CI, 0.65-1.08) and were less likely to die (OR, 0.26; 95% CI, 0.22-0.29) during hospitalization than were patients who did not receive this therapy. Patients who used β-blockers during hospitalization had significantly lower death rates after hospital discharge.

Conclusions  The results of this observational study demonstrate encouraging trends in the use of β-blockers in hospitalized patients with AMI and document the benefits to be gained from this treatment.

DESPITE the magnitude and impact of cardiovascular disease,^1,2 encouraging declines in mortality attributed to coronary heart disease have occurred in the United States in the past several decades. An important contributor to the decline in coronary heart disease mortality observed in the United States is the increased use of effective cardiac treatments, including aspirin, thrombolysis, primary angioplasty, and β-blockers, in patients with acute myocardial infarction (AMI).^3,4

Several large-scale clinical trials^5-10 have demonstrated the beneficial effects of β-blocker therapy on morbidity and mortality rates in patients with AMI. Based on these findings, national guidelines¹¹ have recommended the use of β-blockers in all patients with AMI who do not have contraindications to this therapy. In revised guidelines¹² published by the American College of Cardiology/American Heart Association in 1999, indications for the use of β-blockers in patients with coronary heart disease have become even broader. Despite these widely disseminated guidelines and documented benefits associated with the use of β-blockers in patients with coronary heart disease, some data suggest that β-blockers, although increasingly used, remain underused in patients with AMI.^4,13,14

The Worcester Heart Attack Study is an ongoing observational study of greater Worcester residents hospitalized with confirmed AMI.^15-18 This population-based study offers a unique opportunity and perspective to examine recent and more than 2 decades (1975-1999) of trends in the use of β-blockers in metropolitan Worcester residents hospitalized with AMI. Additional objectives of the present study were to identify groups of patients in whom β-blockers are underused and the relation of β-blocker use to hospital and postdischarge outcomes after AMI.

This study is part of an ongoing population-based investigation of long-term trends in the incidence, hospital, and postdischarge case-fatality rates in residents of the Worcester metropolitan area hospitalized with a discharge diagnosis of AMI from all 16 greater Worcester hospitals during 1975, 1978, 1981, 1984, 1986, 1988, 1990, 1991, 1993, 1995, 1997, and 1999. Fewer hospitals were included in more recent study years owing to hospital closures, mergers, or conversions to chronic care or rehabilitation facilities. These periods were selected on the basis of funding availability and for the purpose of examining trends over time in study outcomes on an approximately alternating yearly basis. The details of this study have been described previously.^15-18 In brief, the medical records of residents of the Worcester metropolitan area (1990 census estimate, 437 000) hospitalized for possible AMI were individually reviewed, and the diagnosis of AMI was validated according to predefined criteria.^15-18 A total of 10 374 metropolitan Worcester residents with validated AMI composed the study population.

Sociodemographic, medical history, and clinical data were abstracted from the hospital medical records of geographically eligible patients with confirmed AMI by trained study physicians and nurses. Information was collected about the patient's age, sex, and comorbidities; AMI order (initial vs previous), type (Q wave vs non–Q wave), and location (anterior vs inferior or posterior); and hospital discharge status. Information was also collected about the occurrence of clinically significant hospital complications, including heart failure, cardiogenic shock, and primary ventricular fibrillation.^19-21 Nurse and physician notes and medication administration records were reviewed to ascertain use of β-blockers, other cardiac medications, and coronary interventional procedures.²² Patients were considered to have received β-blockers if they were prescribed this medication at any time during hospitalization, irrespective of whether they were previously treated with β-blockers. Owing to the manner in which the data were collected, we did not obtain information about duration, dosage, or timing of administration of β-blockers (or the specific agent used) in relation to hospital presentation. We also did not collect information about the use of β-blockers, or other cardiac therapies, after hospital discharge. Survival status after hospital discharge was determined through a review of hospital medical records of additional hospitalizations at area hospitals and a review of death certificates. Some form of additional follow-up was performed for most patients (approximately 99%) discharged after AMI.

Differences in the distribution of demographic, medical history, and clinical characteristics between patients with AMI treated and not treated with β-blockers during the acute hospitalization were examined through the use of χ² and t tests for discrete and continuous variables, respectively. The significance of changes during the nearly 25-year study in the use of β-blockers was examined through the use of χ² tests for trends. A logistic multivariate regression approach was used to examine the association between demographic, medical history, and clinical characteristics and the prescribing of β-blockers during the index hospitalization. A logistic regression analysis was also used to examine the association between β-blocker use and hospital case-fatality rates while controlling for potentially confounding demographic and clinical prognostic factors. The factors controlled for in the regression analyses included age, sex, previous angina, hypertension, diabetes mellitus, stroke and heart failure, and AMI order (initial vs previous), type (Q wave vs non–Q wave), and location (anterior vs inferior or posterior). Given the magnitude and adverse prognostic impact of several major complications of AMI, namely, heart failure, cardiogenic shock, and primary ventricular fibrillation,^19-21 we examined the relation between the use of β-blockers during hospitalization and the development of these clinical end points while controlling for potentially confounding factors. A life-table approach was used to examine the association between administration of β-blockers during hospitalization and long-term survival, including patients with varying lengths of follow-up after hospital discharge. A proportional hazards regression model was used to examine the association between β-blocker use during hospitalization for AMI and postdischarge survival while controlling for potentially confounding demographic, medical history, and clinical characteristics.

A total of 10 374 patients with validated AMI were included in the study. Of these, 5505 (53%) received β-blockers at some time during the index hospitalization during the study years.

Use of β-blockers in greater Worcester residents increased significantly during the nearly 25 years under study (Figure 1). In 1975, approximately 11% of patients with AMI were treated with β-blockers. The use of β-blockers increased sharply during subsequent years, particularly during the 1990s. Reflective of these trends, 82% of greater Worcester residents hospitalized with AMI in 1999 were treated with β-blockers during the acute hospitalization.

Because some studies have shown underuse of β-blockers in certain demographic or clinically defined subgroups, we examined differences in the characteristics of patients treated vs not treated with β-blockers during hospitalization.

Patients treated with β-blockers were significantly younger, were more likely to be men, and were significantly more likely to have a history of angina or hypertension (Table 1). On the other hand, patients with a history of stroke or heart failure were less likely to receive β-blockers than were respective comparison groups. Patients who experienced an initial, inferior or posterior, or non–Q wave AMI were more likely to be treated with β-blockers. Patients receiving β-blockers were significantly more likely to receive other effective cardiac medications, including angiotensin-converting enzyme inhibitors, aspirin, lipid-lowering agents, and thrombolytic agents. Patients treated with β-blockers were more likely to undergo cardiac catheterization and percutaneous revascularization but were less likely to undergo coronary artery bypass graft surgery than were patients who did not receive this medication (Table 1).

Because there were marked increases in the use of β-blockers in hospitalized patients over time, we examined whether the profile of patients treated with β-blockers had changed during the study years. For this analysis, we included patients who were hospitalized during the initial (1975/1978), midpoint (1986/1988), and most recent (1997/1999) study years for ease of data interpretation (Table 2). There were marked increases in the use of β-blockers over time in most of the patient subgroups examined.

β-Blockers have been increasingly prescribed to patients of all ages over time, with particularly marked increases in the use of this therapy in patients 65 years and older. Men were more likely than women to receive β-blockers during each study period, although these treatment differences narrowed during more recent study years. Although there were some differences in the early study years in the prescribing of β-blockers to patients with certain comorbidities or characteristics associated with AMI, most of these treatment differences narrowed over time. Patients treated with β-blockers were consistently more likely to receive other beneficial medications for AMI and were also more likely to undergo cardiac catheterization, percutaneous transluminal coronary angiography, and, in more recent years, coronary artery bypass graft surgery.

We carried out several multivariate regression analyses to identify high-risk patient groups that did not receive β-blockers during hospitalization (Table 3). In the first analysis, we examined factors associated with the nonuse of β-blockers in the entire cohort of patients with AMI. In the second analysis, to provide more recent insights into factors associated with the lack of receipt of β-blockers and to examine the relation between contemporary treatment modalities, including coronary interventional procedures, and prescribing of β-blockers, we restricted the sample to patients hospitalized between 1986 and 1999. The inclusion of hospital discharge status, length of hospital stay, or year of hospitalization in each of these regression analyses did not appreciably change the observed results.

Increasing age, female sex, history of heart failure or stroke, and a previous, anterior, or Q wave AMI were significantly associated with failure to be treated with β-blockers (Table 3). Patients without a history of angina or hypertension were significantly less likely to receive this therapy. Relatively similar, albeit fewer, factors were associated with failure to receive β-blockers when we examined the use of β-blockers in patients hospitalized between 1986 and 1999. Patients treated with calcium antagonists or thrombolytics, those who did not receive aspirin or lipid-lowering agents, and those who underwent coronary interventional procedures, were significantly less likely to be treated with β-blockers.

We also examined whether the characteristics of patients with AMI who did not receive β-blockers had changed in the 3 most recently hospitalized cohorts (1995, 1997, and 1999). Advanced age; absence of a history of angina, hypertension, and heart failure; lack of receipt of previously described cardiac medications; and undergoing cardiac catheterization and coronary artery bypass graft surgery remained significantly associated with failure to be treated with β-blockers.

During the combined periods, patients receiving β-blockers during hospitalization for AMI experienced lower rates of heart failure, primary ventricular fibrillation, cardiogenic shock, and hospital death than patients who were not treated with β-blockers (Table 4).

Because patients receiving vs not receiving β-blockers differed in a variety of characteristics that might affect these important hospital outcomes, we carried out a series of multiple regression analyses to more systematically examine the association of β-blocker use with these hospital end points. For these analyses, each of the aforementioned clinical complications and hospital death were considered to be the outcomes. The relation of use of β-blockers to each of these end points was examined while controlling for previously described potentially confounding demographic, medical history, and clinical characteristics. The results of these analyses confirm the beneficial relation between β-blocker use and each of these hospital end points. Patients treated with β-blockers were considerably less likely to develop heart failure (adjusted odds ratio [OR], 0.58; 95% confidence interval [CI], 0.53-0.63), cardiogenic shock (OR, 0.46; 95% CI, 0.39-0.54), and primary ventricular fibrillation (OR, 0.84; 95% CI, 0.65-1.08) and were also less likely to die (adjusted OR, 0.26; 95% CI, 0.22-0.29) during the acute hospitalization than were patients who were not treated with these agents. These associations remained even after we controlled for the effect of other cardiac medications and coronary interventional procedures. Similar findings were observed when we restricted the analysis to patients hospitalized between 1986 and 1999 and when we controlled for the use of other effective cardiac medications and coronary interventional procedures.

Because the beneficial effects of β-blocker therapy may not be immediate, we carried out an additional subgroup analysis excluding patients who died or who developed heart failure, cardiogenic shock, or primary ventricular fibrillation during the first 2 days of hospital admission. The previously observed beneficial effects of β-blocker therapy remained in this subgroup analysis. Hospital death rates were 4.9% and 16.8% for patients treated and not treated with β-blockers, respectively. The rates of heart failure (5.9% vs 8.2%), cardiogenic shock (1.2% vs 2.1%), and primary ventricular fibrillation (0.4% vs 0.6%) were also lower in patients who received vs did not receive β-blockers.

We also examined whether the beneficial effects associated with the use of β-blockers varied according to patient demographic and clinical characteristics (Table 5). After simultaneously controlling for each of the factors examined, patients of varying ages, both sexes, and with selected medical history and clinical characteristics all shared in the positive effects of β-blocker therapy on hospital survival after AMI.

We examined differences in the long-term survival of patients discharged from all greater Worcester hospitals after AMI according to the use of β-blockers during their index hospitalization. Patients who had been prescribed β-blockers while hospitalized for AMI had significantly lower long-term death rates than patients who had not been given β-blockers (Figure 2). These differences were present throughout the study years but were especially striking during more recent periods.

We also examined whether differences in long-term survival according to treatment with β-blockers remained after controlling for other prognostic factors. In the first analysis, we examined differences in the death rates 1 year after hospital discharge in hospital survivors according to hospital receipt of β-blockers. One year was selected because it has been shown to be a high-risk period after hospital discharge for AMI. After controlling for various demographic, medical history, and clinical characteristics, patients treated with β-blockers during hospitalization were significantly less likely to die during the first year after hospital discharge than patients who were not treated with β-blockers (adjusted OR, 0.53; 95% CI, 0.47-0.61). Similar long-term benefits of the use of β-blockers were observed when we restricted the analysis to patients discharged from the hospital in 1986 to the most recent cohorts and additionally controlled for the use of effective cardiac medications and coronary interventional procedures (OR, 0.60; 95% CI, 0.50-0.71).

Last, we examined the association between hospital treatment with β-blockers and overall long-term survival for greater than 25 years. After controlling for previously described factors, patients treated with β-blockers were significantly less likely to die during follow-up (adjusted hazard rate, 0.62; 95% CI, 0.58-0.66). Similar results were observed when we restricted the analysis to patients discharged from all area hospitals between 1986 and 1999 and after controlling for previously described demographic and clinical characteristics and interventional procedures and medications (adjusted hazard rate, 0.70; 95% CI, 0.64-0.76).

The results of this community-based study suggest significant increases in the use of β-blockers in hospitalized patients with AMI during the 25 years under study, with particularly marked increases in the use of this therapy in the 1990s. Although increases in the use of β-blockers were noted for most of the subgroups examined, there remains substantial room for improvement in certain high-risk groups, including the elderly and women. Although direct conclusions about the efficacy of this therapy cannot be drawn from data collected in the context of this nonrandomized observational study, administration of β-blockers was associated with better outcomes in the total study sample and in each of the high-risk patient subgroups examined.

Several population-based studies,^18,23-25 including the Worcester Heart Attack Study, have shown consistent declines during the past several decades in hospital case-fatality rates in patients with AMI. These improvements in hospital survival have largely been attributed to the development and increased use of effective coronary reperfusion strategies and adjunctive medical therapies.

β-Blockers have been a proven mainstay treatment for patients with AMI since the publication of several large-scale, multicenter, randomized controlled trials in the 1980s. In the First International Study of Infarct Survival,⁵ receipt of atenolol after AMI was associated with a 15% reduction in mortality from vascular causes. Two other large multicenter trials, the Norwegian Multicenter Study⁸ and the Beta-Blocker Heart Attack Trial,⁹ also showed improved survival rates with the use of timolol and propranolol, respectively, after AMI. In a meta-analysis²⁶ of 28 published clinical trials carried out before the widespread use of aspirin and thrombolytic therapy, treatment with β-blockers after AMI was associated with a 13% reduction in all-cause mortality.

Despite clear evidence of the benefits of β-blocker therapy on mortality after AMI, and current guidelines that advocate the early use of β-blockade in a broader range of patients with AMI,^11,12 data from several earlier descriptive studies suggest that this therapy is underused. In an observational study²⁷ of more than 5000 elderly patients who had survived for 30 days after AMI between 1987 and 1992, only one fifth of the potentially eligible patients were receiving β-blockers. Receipt of β-blockers was associated with a marked reduction in mortality. A more recent study²⁸ from the Cooperative Cardiovascular Project reviewed the medical records of more than 200 000 Medicare beneficiaries hospitalized with AMI in 1994 and 1995. Approximately one third of the patients studied received β-blockers during the acute hospitalization. Patients treated with β-blockers had a 40% reduction in mortality during 2-year follow-up.

Our findings are in agreement with data from other observational studies that have shown increases in the use of β-blockers in patients with AMI over time. A community-based study²⁹ in Canada examined trends in β-blocker use in patients with AMI between 1979 and 1992. β-Blocker use increased from 19% in 1979 to 48% in 1992.²⁹ Investigators from the Cooperative Cardiovascular Project recently published findings³⁰ from a quality improvement study. Between 1992 and 1995, the prescription of β-blockers to elderly patients with AMI at the time of hospital discharge increased from 47% to 68% in the hospitals that received continuous information feedback.³⁰ In the large National Registry of Myocardial Infarction, use of β-blockers during the first 24 hours of hospitalization for AMI increased from 48% in 1994 to 67% in 1999.³¹

The results of the present study, although confirming the underuse of β-blockers in the past, demonstrate marked improvements in the use of this effective cardiac medication, particularly during the past decade, from a community-based perspective. These changes followed the publication of major randomized controlled trials confirming the benefits of these agents after AMI, studies demonstrating the underuse of β-blockers from a broader population-based perspective, and revised guidelines from the American College of Cardiology/American Heart Association extending the indications for β-blocker use after AMI.^11,12

We identified a variety of patient demographic and clinical characteristics associated with failure to receive β-blockers. The elderly, women, patients with significant comorbidities, and those with an anterior or Q wave AMI were least likely to receive these agents.

Low utilization rates of β-blockers in elderly patients with AMI have been reported previously.²⁷ Undertreatment of this particularly high-risk group may have been due to these patients having a greater prevalence of comorbidities and contraindications to the receipt of β-blockers.³² Although several clinical variables that may be contraindications to the use of β-blockers were not evaluated in the present study, it is apparent that the treatment of elderly patients has increased significantly over time. Physicians may now be more likely to prescribe this treatment to elderly patients with relative contraindications.

Although the results of this study confirm those of previous studies that have shown lower utilization rates of β-blockers in women than in men,³² encouraging trends in the use of this therapy in women over time were noted, narrowing previously observed sex gaps. Although this study did not find differences in the use of β-blockers in patients with or without a history of diabetes mellitus, diabetic patients with AMI receive important benefits from this treatment in terms of survival and reinfarction rates.^33,34

Previous studies²⁸ have found that patients with various comorbidities, including heart failure, are less likely to receive β-blockers after AMI. These findings were also observed in the present study. This is an important finding because previous studies^28,35 have shown that sicker patients with AMI have a greater absolute risk reduction with β-blocker use. Also, patients with non–Q wave AMI were nearly as likely to receive β-blockers as patients with Q wave MI during the most recent study years. This finding is somewhat surprising since American College of Cardiology/American Heart Association guidelines only recently upgraded the recommendation of β-blocker use in patients with non–Q wave AMI from a class IIb recommendation to a class I recommendation in 1999.¹²

Unlike the analysis of other therapeutic agents, such as aspirin, used in the treatment of patients with AMI in which a plausible rationale for the underuse of therapy in high-risk subsets is difficult to discern, there are some grounds for caution in the use of β-blockade in certain patient groups. This therapy must be carefully used in patients with heart failure complicating their AMI and, by extension, in those at high risk for heart failure for fear of incurring progression of this clinical syndrome. These data notwithstanding, some information suggests that these patients may have much to gain from this therapy.^9,10,28

The use of β-blockers in this study was associated with a marked reduction in the likelihood of dying after AMI for the total study sample and for all subgroups analyzed, including the elderly, patients with diabetes mellitus, patients with previous heart failure, and those with AMI-associated characteristics. The magnitude of this reduced risk did not appreciably differ for each of the subgroups examined. These data lend indirect support to the premise that the increase in β-blocker use during the past 25 years has played at least a partial role in decreasing hospital mortality after AMI.

Since the convincing results of randomized trials carried out in the 1980s that showed increased survival with the use of β-blockers after AMI, there have been no further randomized trials of this therapy, to our knowledge. A quality improvement endeavor recorded data on β-blocker use (and other therapies) in patients with AMI from hospitals in 4 states.³⁰ Participating hospitals were provided data on their own utilization patterns and were encouraged to devise strategies to improve these rates. In 1 year, the use of β-blocker therapy increased from 32% to 50%.

The positive impact of β-blockade on outcomes after AMI in the present community-based study is further supported by the strong association between receipt of β-blockers and hospital and long-term survival. Even after controlling for multiple potentially confounding prognostic variables, patients receiving β-blockers were much less likely to die during the first year after hospital discharge or during 25-year follow-up. To our knowledge, this is the first population-based study demonstrating a positive association between hospital administration of β-blockers and improved long-term survival. This finding also likely reflects an increased use of β-blockers on an outpatient basis in patients administered this therapy during the acute hospitalization. Our data suggest that the in-hospital period represents an important opportunity for the initiation of efficacious therapies.

Caution must be exercised in any attempt to identify an association between prescribing of a specific therapy and hospital or postdischarge outcomes in an observational study. This is particularly problematic in the assessment of hospital events when the timing of therapy relative to a specific clinical complication cannot be accurately ascertained. Despite appropriate reservations, observational studies remain an important tool for assessing the impact of a given therapy in the community setting. In the present study, as in others, β-blockade was associated with significantly lower rates of heart failure, cardiogenic shock, and hospital death.

There are several strengths and limitations of this communitywide observational study. The strengths include the large population of residents from a representative northeast metropolitan area hospitalized at all medical centers with documented AMI, enhancing the generalizability of the present findings. The results of this study provide insights into the treatment practices and outcomes of more typical patients with AMI treated in the community by areawide cardiologists and internists. This is in contrast to patients treated in the context of randomized controlled trials, with their more narrowly defined inclusion criteria. Note, however, that previous studies^32,36 have shown differences in the use of β-blockers in patients with AMI from different regions of the United States, with patients from the northeast being more likely to be treated with this therapy. Use of multivariate analytic techniques to control for the effects of other potentially confounding factors in examining the characteristics of patients more likely to be treated with β-blockers, and associated outcomes, strengthens the interpretation of any associations observed.

Limitations of this study include the failure to adequately measure (or not measure at all) other factors, such as pulmonary or peripheral vascular disease, that may be associated with β-blocker use and hospital and postdischarge outcomes; the potential for direct or residual confounding due to medication indication; the lack of information about the timing of clinically significant hospital complications relative to administration of β-blockers; and the lack of information about long-term use rates of β-blockers after hospital discharge for AMI. Moreover, the biases inherent in observational studies must be kept clearly in mind in interpreting the present findings. These biases may have contributed to the larger beneficial effects associated with β-blocker therapy in the present study than in previous randomized trials.

In summary, the results of this descriptive, communitywide study demonstrate encouraging increases in the use of β-blockers in patients with AMI. Although increases in the use of this treatment approach have been observed in a variety of patient subgroups that have been historically undertreated with this therapy, there remains room for improvement in many high-risk groups.

Corresponding author and reprints: Robert J. Goldberg, PhD, Division of Cardiovascular Medicine, Department of Medicine, University of Massachusetts Medical School, 55 Lake Ave N, Worcester, MA 01655 (e-mail: goldberr@ummhc.org).

Accepted for publication November 22, 2002.

This study was supported by grant RO1 HL35434 from the National Heart, Lung, and Blood Institute, Bethesda, Md.

We thank the administration, medical records, and cardiology departments of the participating Worcester metropolitan area hospitals for their cooperation.