Outcome With Calcium Channel Antagonists After Myocardial Infarction: A Community-Based Study

1.1 Study Protocol

The Newcastle collaborating center of the WHO MONICA Project monitored all suspected cases of sudden coronary death and acute myocardial infarction among residents of the Lower Hunter region of New South Wales who were aged 25 to 69 years old. Study nurses monitored all hospitals in the area, registering every patient likely to meet the study criteria. Patients were interviewed while in the hospital to obtain information on symptoms, medical history and smoking status. Cardiac enzyme results were extracted from hospital records, and electrocardiograms (ECGs) were copied and subsequently coded according to the Minnesota code [23]. Enzyme levels were reported as a percentage of the upper limit of the normal range for the relevant laboratory. For this study, ECG results were classified according to categories used in the definition of the Norris index [24]. Details of medication use during the event and at discharge were obtained from the medical record. Medication dosage and formulation were not recorded. All death certificates for the cohort were also scrutinized. Details of fatal cases were obtained from the death certificates and postmortem records and from questionnaires sent to doctors, relatives and other informants. The study had ethics approval from the public hospitals in the region and from the University of Newcastle Research Ethics Committee.

1.2 Event Classification

The cardiac events reported here were those satisfying the criteria for definite or possible myocardial infarction and cardiac death used in the WHO MONICA Project [23, 25]. These definitions are summarized as follows: Definite myocardial infarctionwas defined by either unequivocal serial ECG progression (defined by Minnesota codes) during the attack or by cardiac enzyme levels twice the level of normal and specific combinations of symptoms and ECG changes. Possible myocardial infarctionwas defined as typical prolonged (20 min) chest pain with ECG and enzyme changes not fulfilling the criteria for definite myocardial infarction and with no other cause for the symptoms established. Therefore, in some nonfatal cases, this definition was consistent with a clinical diagnosis of unstable angina; for others, a clinical diagnosis of myocardial infarction was probable [26].

1.3 Classification of Fatal Events

Fatal eventswere classified as definite myocardial infarction if they satisfied the criteria for nonfatal definite myocardial infarction or if autopsy data confirmed acute myocardial infarction. Fatal cases with suggestive terminal symptoms, a history of coronary artery disease or autopsy evidence of chronic occlusive coronary disease without another cause of death established were classified as possible myocardial infarction. Patients who died shortly after the onset of symptoms of myocardial infarction were usually classified as having a possible myocardial infarction because serial measurements of ECG and enzymes were not possible. Some cardiac deaths were unclassifiable if the cause of death on the death certificate was coded 410–414 according to the International Classification of Disease, 9th Revision, but information required for the other MONICA criteria was lacking. For this study, events categorized as fatal, possible or definite myocardial infarction or unclassifiable cardiac death were grouped together as cardiac death.

1.4 Patients and Outcomes

This study involved patients who survived at least 28 days from the onset of symptoms of a definite or possible myocardial infarction between September 1988 and December 1993. The main outcome of interest was the time to the next major coronary event, defined as possible or definite myocardial infarction or cardiac death. We also examined as separate outcomes the times to definite myocardial infarction or cardiac death, cardiac death and all-cause mortality.

Subjects were deemed to have had another major coronary event if they were registered by the MONICA Project on a second occasion before March 31, 1994 and satisfied the diagnostic criteria outlined above. The MONICA register only collected information for people between 25 and 69 years old; therefore, as subjects turned 70, they were censored.

1.5 Statistical Analysis

Chi-square tests were used to compare patient groups defined by the use of beta-blocker and calcium antagonist therapy at discharge. Relative risk estimates were calculated using survival analysis techniques [27]. Accelerated failure time models that assumed the probability of having another major coronary event and followed a Weibull distribution were used. These models are equivalent to parametric proportional hazards models [28]. Adjustment was made for the possible confounding effects of age (5-year age groups); gender; history of hypertension; hypercholesterolemia; diabetes; previous myocardial infarction; angina and stroke; smoking status (current, previous or never); peak creatine kinase level; ECG classification; use of frusemide or digoxin, or both, in the hospital; use of aspirin; angiotensin-converting enzyme inhibitors; diuretic drugs; digoxin; antiarrhythmic and bronchodilator medication at discharge from hospital; and year of enrollment. There were 12 patients with missing ECG information and 14 with missing enzyme data for the initial event: For the multivariate analysis, these patients were included by creating separate categories for missing data. The analysis was repeated after exclusion of patients with missing data, with only trivial effects on the results.

The initial analysis was performed with all patients who had a definite or possible myocardial infarction. A separate analysis was performed restricting the cohort to patients with definite myocardial infarction as the first registered event. In these analyses, the reference group was patients discharged with beta-blockers but not calcium antagonists. To evaluate the risks associated with the use of individual calcium antagonists, analyses were performed with patients given neither calcium antagonists nor beta-blockers as the reference group. For each of these analyses, the outcome variables were times to another major coronary event; definite myocardial infarction or cardiac death, or both; cardiac death; and all-cause mortality. When a patient satisfied the criteria for more than one outcome, the time to the first relevant outcome was used. For example, for a patient with recurrent myocardial infarction who later died during a further myocardial infarction, the time to the first recurrent myocardial infarction would be used in the analysis of major coronary events and the time to death in the analyses of cardiac and all-cause mortality.

2.1 Patients

There were 1,001 patients treated with beta-blockers, 923 with calcium antagonists, 711 with both beta-blockers and calcium antagonists and 1,346 with neither drug at the time of hospital discharge after a definite or possible myocardial infarction during the study period. The median observation time was 664 days (interquartile range 226 to 1,227), with a total observation period of 8,430 person-years.

2.2 Baseline Characteristics

There were substantial differences in baseline characteristics of patients according to the use of beta-blockers and calcium antagonists at the time of hospital discharge (Table 1). Compared with patients treated with beta-blockers, patients treated with calcium antagonists were older, were more likely to have angina and were treated more frequently with diuretic drugs, digoxin, angiotensin-converting enzyme inhibitors and bronchodilators. The differences between patients treated with beta-blockers and those treated with neither beta-blockers nor calcium antagonists were smaller, except for the prevalence of hypertension and the use of aspirin (which were more common in patients treated with beta-blockers) and the use of diuretic drugs, digoxin and bronchodilators (which were more prevalent in patients given neither beta-blockers nor calcium antagonists).

2.3 Risk of Major Coronary Events

The results from the survival analysis are shown in Table 2. In total, there were 493 nonfatal possible myocardial infarctions, 247 nonfatal definite myocardial infarctions, 211 cardiac deaths and 406 deaths from all causes (including cardiac deaths) in the study participants. When adjusted for all covariates, there was an increased risk of another major coronary event for subjects treated with calcium antagonists compared with those treated with beta-blockers (adjusted relative risk [RR] 1.5, 95% confidence interval [CI] 1.2 to 1.9). Relative risks for the outcomes of definite myocardial infarction or cardiac death (1.4, 95% CI 0.98 to 1.9), cardiac death (1.6, 95% CI 0.97 to 2.7) and all-cause mortality (1.7, 95% CI 1.1 to 2.6) were also higher for this patient group.

Patients treated with neither beta-blockers nor calcium antagonists were also at increased risk of recurrent definite myocardial infarction or cardiac death (RR 1.4, 95% CI 1.0 to 1.9), cardiac death (RR 2.0, 95% CI 1.2 to 3.3) and all-cause mortality (RR 1.8, 95% CI 1.2 to 2.8) compared with patients treated with beta-blockers.

2.4 Results in Patients With a Definite Myocardial Infarction

Relative risks were slightly but consistently higher when the analysis was restricted to the 1,855 patients with a definite myocardial infarction as the initial registered event. In this cohort, the adjusted relative risk of recurrent myocardial infarction or cardiac death was 1.5 (95% CI 1.0 to 2.3), cardiac death 2.0 (95% CI 1.1 to 3.95) and all-cause mortality 1.9 (95% CI 1.1 to 3.4) for patients taking calcium antagonists compared with patients taking beta-blockers.

2.5 Patients Treated With Both Beta-Blockers and Calcium Antagonists

Patients treated with both beta-blockers and calcium antagonists were at increased risk of a major coronary event (RR 1.6, 95% CI 1.3 to 2.0) compared with patients treated with beta-blockers only. For the outcomes of cardiac death and death by all causes, patients taking both beta-blockers and calcium antagonists had risks that were intermediate between those treated with calcium antagonists and those treated with beta-blockers, and the confidence intervals included unity (Table 2).

2.6 Verapamil, Diltiazem and Nifedipine

The major calcium antagonists prescribed were verapamil (185 patients), diltiazem (596 patients) and nifedipine (100 patients). Compared with patients given neither beta-blockers nor calcium antagonists, patients given diltiazem (RR 1.4, 95% CI 1.2 to 1.8) and nifedipine (RR 1.7, 95% CI 1.1 to 2.5) were at increased risk of another coronary event (Table 3). Most of the increased risk of recurrent events in these patients was attributable to an increased risk of nonfatal possible myocardial infarction. For the other outcomes of definite myocardial infarction or cardiac death, cardiac death and all-cause mortality, results were similar among the patients treated with different calcium antagonists and were not significantly different from patients treated with neither agent.

3.1 Adverse Effects of Calcium Antagonists

The role of calcium antagonists is currently unsettled because of lack of long-term data about the effects of these drugs on mortality and recent observational studies demonstrating an increased risk of adverse events, including recurrent myocardial infarction [31], gastrointestinal hemorrhage [32]and cancer [33]. The results of the present study do not support a major increase in risk of adverse events from the use of calcium antagonists after myocardial infarction. Increased mortality among patients given calcium antagonists was observed only in comparison to a reference group given beta-blockers, indicating that the difference between the groups was due to the beneficial effect of beta-blockers rather than an adverse effect of calcium antagonists. The results of the present study are consistent with another community-based study of beta-blocker and calcium antagonist use after myocardial infarction [34]. In that study, the use of a calcium antagonist instead of a beta-blocker after myocardial infarction was associated with an adjusted relative risk of death of 1.98 (95% CI 1.44 to 2.72). In the present study, the risk estimate for the same comparison, in patients with definite myocardial infarction, was 1.90 (95% CI 1.05 to 3.41).

3.2 Verapamil, Diltiazem and Nifedipine

The risks of definite myocardial infarction and death were not significantly different among patients given verapamil, diltiazem and nifedipine (Table 3), with risk estimates all close to 1 compared with patients taking neither beta-blockers nor calcium antagonists. However, in patients given nifedipine or diltiazem, there was an increased risk of readmission to the hospital with chest pain without unequivocal evidence of myocardial infarction (possible myocardial infarction in the MONICA classification). Some of these events would be classified clinically as unstable angina [26]. Short-acting calcium antagonists of the dihydropyridine group have been associated with exacerbation of angina as a result of reflex catecholamine stimulation [14–17]. In the present study, the estimates of risk for readmission with chest pain, as well as for other adverse outcomes, were highest with nifedipine, but relatively few patients were given nifedipine, and confidence intervals were wide.

3.3 Limitations of the Study

The strengths of the present study include its community-based cohort design, the prospective definition of end points and the intensive efforts to ascertain all cases of suspected myocardial infarction and cardiac death and to document medication use and other important factors associated with outcome after myocardial infarction. A major limitation of this observational study relates to the nonrandom assignment of drug therapy and, consequently, differences in baseline characteristics among the patients treated with calcium antagonists and beta-blockers (Table 1). Residual confounding by indication cannot be excluded as the cause for the excess risks associated with calcium antagonist use, particularly when these relative risks were in the range 1 to 2.

Outcome events were identified by repeat registrations of coronary events and surveillance of death certificates. This study design may have underestimated the relative risks because nonfatal cardiac events that occurred in hospitals outside the study area were not detected. Exposure to drugs after discharge from the hospital was not assessed and may also have introduced misclassification. In addition, drug formulation and dose were not recorded in the MONICA data. Therefore, we could not evaluate the relative effects of short- and long-acting calcium antagonists [35]. Very few patients in this study were treated with the newer long-acting calcium antagonists, such as amlodipine [36]and felodipine.

3.4 Underuse of Beta-Blockers

A noteworthy finding was the number of patients discharged from the hospital with calcium antagonists after a suspected myocardial infarction. Almost as many patients were discharged with calcium antagonists (41%) as were discharged with beta-blockers (43%). This pattern of drug use is difficult to understand in the context of trial data demonstrating unequivocal benefit for beta-blockers after myocardial infarction [30]and much less benefit for calcium antagonists [8, 11, 12]. In some cases, drug selection would have been influenced by contraindications to beta-blockade, but such contraindications have been estimated to affect no more than 30% of infarct survivors [34], yet <50% of patients in this cohort left the hospital with beta-blockers. Underuse of beta-blockers has been reported in several other studies [34, 37, 38], but the reasons for these apparently suboptimal prescribing patterns are poorly understood [37, 39].

3.5 Conclusions

The present results are consistent with randomized trial data showing benefit from beta-blockers after myocardial infarction and no effect on the risk of myocardial infarction and death with the use of calcium antagonists. These results confirm that whenever possible, beta-blockers should be used in preference to calcium antagonists in patients with a myocardial infarction.