INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the secondary prevention trials of patients with a prior myocardial infarction, the use of -blockers is associated with a reduced risk of recurrent coronary events and all cause mortality (1, 2). Chronic use of -blockers leads to an upregulation of the -receptors (3, 4). Cessation of -blockers, which produces a withdrawal syndrome in susceptible individuals, is associated with a transitory fourfold increase in the risk of myocardial infarction in patients with hypertension (5). -agonists, which exert the opposite pharmacologic effects of -blockers, have been associated with an increased risk of asthma mortality (6). In addition, oral and nebulized forms of -agonists have been associated with an increased risk of acute cardiovascular mortality (10). Whether the use of -agonist metered dose inhaler (MDI) increases the risk of myocardial infarction remains unknown. Using data from an existing case-control study, we assessed the association between the use of MDI -agonists and the risk of myocardial infarction.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Setting

The setting was the Group Health Cooperative of Puget Sound (GHC), a staff-model nonprofit health maintenance organization (HMO) in western Washington State. The study subjects had been identified for two previous studies, one designed to assess the association of hormone replacement therapy with the risk of myocardial infarction among postmenopausal women (11, 12), and the other designed to assess the association of antihypertensive medications with the risk of myocardial infarction (13). The current analysis includes data on men and women with treated hypertension and women without hypertension. Data on men without hypertension were not available. Previously described methods are summarized briefly (11).

Study Subjects

Cases were postmenopausal female GHC enrollees, age 30 to 79 yr, who suffered an incident nonfatal or fatal myocardial infarction between July 1986 and December 1994 (11, 12), and hypertensive male GHC enrollees, age 30 to 79 yr, who suffered an incident nonfatal or fatal myocardial infarction between July 1989 and December 1994 (13). Control subjects were a stratified random sample of GHC enrollees frequency-matched to the cases by sex, age (within a decade), hypertension, and calendar year of the index date at a case-to-control ratio of 1:2 to 1:3.

Index Dates and Eligibility

All subjects had an index date. For the hospitalized cases, the index date was the date of admission for the first myocardial infarction; for the out-of-hospital fatal cases, the index date was the date of death; and for the control subjects, the index date was a computer-generated random date within the calendar year for which they had been sampled. For all subjects, we collected information about eligibility and risk-factor data available only before the index date. We excluded subjects who had (1) been enrolled less than 2 yr or fewer than four visits before their index dates; (2) a prior myocardial infarction; and (3) a myocardial infarction as a complication of an operation.

Data Collection

Data collection included a review of the outpatient medical record. Based on the medical record, trained research assistants determined eligibility and collected information about traditional risk factors for coronary heart disease, such as smoking, angina, and diabetes. The GHC computerized pharmacy database, which contains records of all prescriptions dispensed to GHC enrollees, was used to assess prescriptions for -agonists that were dispensed in MDIs. A telephone survey conducted in conjunction with the study of postmenopausal women determined that 95% of subjects filled all or almost all of their prescriptions at a GHC pharmacy (12). Tobacco use was assessed by review of outpatient clinical records for all subjects and complemented with telephone interviews for 53% of subjects. In total, we assessed tobacco use history for greater than 98% of subjects. Among those subjects for whom both medical record review and telephone interviews were available, we obtained an 84% agreement (kappa 0.74).

Definition of Variables and Statistical Analysis

We defined never users, one-time users, and regular users of MDI -agonists as subjects who filled zero, one, and two or more prescriptions respectively in the 2 yr before index date. We defined new users of -agonists as those one-time users who received their only -agonist MDI canister in the 3 mo prior to their index date. The exposure of interest was the MDI -agonists filled at GHC during the study period and included albuterol, metaproterenol, and terbutaline. We defined cardiovascular disease as a diagnosis of angina, transient ischemic attack, stroke, congestive heart failure, arrhythmia, or claudication, or a history of coronary angioplasty, coronary bypass surgery, peripheral vascular surgery, or carotid endarterectomy (11). We used stratification and logistic regression to control for potential confounding factors and to estimate risk ratios. All statistical tests were two-tailed and a p value of 0.05 was used to define statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

During the study period, we identified 1,444 cases and 4,094 control subjects. Table 1 shows the characteristics of the cases and control subjects. As expected, cases differed from control subjects for a number of cardiovascular risk factors including diabetes mellitus, hyperlipidemia, and tobacco use. In addition, cases had a higher prevalence of cardiovascular diseases and on average, had more clinic visits in the year before the index date.

Among all subjects, the adjusted odds ratio (OR) of myocardial infarction was 1.67 (95% confidence interval [CI], 1.07 to 2.60) for subjects who received one MDI canister in the 3-mo period before the index date compared with subjects who did not receive a -agonist MDI canister (Table 2). The adjusted OR was not significantly elevated among subjects who received 2 to 3 or 4 or more canisters in the 3 mo prior to the index date.

Table 2 also shows the association between myocardial infarction and inhaled -agonist use among subjects stratified by the presence or absence of a history of cardiovascular disease. For subjects who had a history of cardiovascular disease and filled one MDI canister in the 3-mo period prior to the index date, the adjusted OR of myocardial infarction was 3.22 (95% CI, 1.63 to 6.35). In contrast, subjects without a history of cardiovascular disease had no increase in the risk of myocardial infarction. Regardless of the presence or absence of cardiovascular disease, the adjusted OR of myocardial infarction was not significantly elevated among subjects who had filled two or more -agonist canisters. Controlling for a history of cigarette smoking in the multivariate analysis had little effect on the results. Adjustment for -blocker use and restricting the analysis to subjects with hypertension had trivial effects on the results of this study.

To determine if the association between a single MDI canister dispensed and myocardial infarction differed based on the recency of the -agonist prescription, we classified subjects in categories of frequency and recency of use. For this analysis, we focused on the 678 cases and 1,140 control subjects with a history of cardiovascular disease and examined the recency of the -agonist prescription among one-time users. Compared with subjects who had never received a -agonist prescription in the 2 yr prior to the index date, the estimated risk of myocardial infarction among new users was 7.32 (95% CI, 2.34 to 22.8, Table 3). Compared with one-time users who had not received a -agonist MDI in the 3-mo period prior to the index date, the estimated risk of myocardial infarction among new users was 6.02 (95% CI, 1.78 to 20.3). Among subjects who had filled more than one -agonist prescription in the 2 yr prior to the index date, the estimated risk of myocardial infarction also was not significantly elevated.

We abstracted charts for 19 of the 21 subjects classified as new users to examine indications for -agonists. All subjects had symptoms that were consistent with asthma or chronic obstructive pulmonary disease (COPD). Five subjects had chest discomfort documented: two with chest pressure, one with atypical chest pain and pressure, and two with pleuritic chest pain. All 19 subjects were given a clinical diagnosis of asthma or COPD, but three charts also had documented the possibility of angina or congestive heart failure in combination with pulmonary disease. No additional information regarding the possibility of undiagnosed angina was found.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The primary goal of this study was to examine the relationship between inhaled -agonist use and the risk of myocardial infarction. Patients with a myocardial infarction were significantly more likely than matched control subjects to have been dispensed their first -agonist in the 3 mo immediately prior to their myocardial infarction. The increase in risk was present only among patients with a history of cardiovascular disease whose -agonist prescription was the first MDI in the preceding 2 yr. Furthermore, the increased risk of myocardial infarction was independent of other factors including history of tobacco use.

The strengths of this observational study include the use of population-based cases and control subjects, the validation and completeness of case diagnosis and identification as well as the use of a complete pharmacy data base to assess -agonist use among cases and control subjects in a comparable and unbiased fashion. In addition, because all subjects were enrolled in a common health care delivery system, in principle, they had equal access to health care.

This study design has several important limitations. Because the use of inhaled -agonists was assessed through pharmacy data for cases and control subjects, we could not assess when and if the subjects actually used their -agonist MDI. The increased risk of myocardial infarction was essentially limited to new users, a stratum with only four control subjects and 17 cases in this study. Although adjustment for covariates had little effect on the OR, these adjustments may not have been effective with small numbers and the possibility of confounding the association of myocardial infarction and MDI use by any of the covariates used in the analyses remains. Adjustments for unmeasured potential confounding variables, such as the severity or type of pulmonary disease, could not be made.

We pooled samples from two separate studies, one in hypertensive men and women and one in postmenopausal women. In each study, the case and control definitions were identical. Because data were not available on nonhypertensive men, our findings cannot be generalized to nonhypertensive men. Restricting the population to hypertensive subjects, however, had trivial effects on the OR.

Inhaled -agonists are absorbed systemically through the lungs and do not undergo first-pass metabolism before reaching the heart (14). Selective ₂-agonists cause a positive chronotropic (15) and inotropic response (16) and are associated with ventricular and atrial ectopy (17). Based on these known physiologic effects, our a priori hypothesis was that there might be a dose- response relationship between the number of -agonist prescriptions filled and the risk of myocardial infarction. To attempt to control for confounding by indication, we initially selected subjects who had filled at least two -agonist prescriptions in the preceding 2 yr. We found no dose-response relationship between -agonists and myocardial infarction, and since we were aware of the downregulation that occurs soon after initiation of -agonists (18), we sought to assess the risk of myocardial infarction among new users of -agonists. This second hypothesis is supported by in vitro data demonstrating that desensitization of the -adrenoceptor to -agonist stimulation occurs within minutes to hours (18) and in vivo data demonstrating that tolerance to the systemic effects of -agonists can occur within 5 d (21, 22). Given that the analyses presented here were in response to a negative initial analysis, we consider these findings to be preliminary and hypothesis-generating.

This study suggests that new users of inhaled -agonists who have cardiovascular disease may be at increased risk for myocardial infarction. While a causal relationship is biologically plausible given the physiologic side effects of -agonists and the downregulation of the -adrenoceptor, it is also plausible that this relationship may exist for noncausal reasons. The possible explanations for our findings are not necessarily mutually exclusive. The causal possibilities include the fact that -agonists prescribed for obstructive lung disease may precipitate myocardial ischemia and infarction as a direct adverse effect. Also, -agonists prescribed for airflow obstruction may cause hypoxemia by increasing ventilation/perfusion heterogeneity and thereby precipitate myocardial infarction. Finally, -agonists given for airflow limitation or nonspecific chest symptoms may have precipitated a dysrhythmia that led to myocardial infarction. The noncausal associations include -agonists prescribed for nonspecific respiratory symptoms or chest discomfort that represents undiagnosed angina, which is a major risk factor for myocardial infarction. In addition, perhaps -agonists were prescribed for airflow obstruction and, despite treatment, airflow obstruction caused hypoxemia and precipitated myocardial infarction. Finally, -agonists may be prescribed for a respiratory illness that is associated with an increase in risk of myocardial infarction (23).

This study demonstrates that new -agonist prescriptions among patients with cardiovascular disease are associated with an increased risk of myocardial infarction. Regardless of the explanation, the implication is that clinicians need to exercise caution before prescribing a -agonist in this setting and they should be cautious to exclude angina in patients presenting with dyspnea or nonspecific chest complaints. If these findings are confirmed in other studies, the guidelines for the diagnosis and treatment of asthma, which state that -agonists may aggravate existing cardiovascular disease (24), should perhaps be strengthened to include a caution when prescribing a first-time -agonist to patients with cardiovascular disease. Additional research is needed to confirm this relationship between first-time -agonist use and myocardial infarction, elucidate mechanisms, and test intervention strategies.