Interaction of baseline characteristics with the hazard of encainide, flecainide, and moricizine therapy in patients with myocardial infarction. A possible explanation for increased mortality in the Cardiac Arrhythmia Suppression Trial (CAST).

Abstract

BACKGROUND The Cardiac Arrhythmia Suppression Trial (CAST) was designed to test the hypothesis that suppression of ventricular ectopy with antiarrhythmic drugs after a myocardial infarction reduces the incidence of sudden arrhythmic death. Patients in whom ventricular ectopy could be suppressed with encainide, flecainide, or moricizine were randomly assigned to receive either active drug or placebo. The encainide and flecainide arms of the study were discontinued in 1989 (CAST-I) and the moricizine arm in 1991 (CAST-II) because of excess mortality. To explore the mechanisms of these adverse outcomes, we examined the interaction of baseline characteristics with the hazard of therapy with encainide, flecainide, or moricizine compared with their respective placebos. METHODS AND RESULTS CAST-I comprised 755 patients assigned to flecainide or encainide and 743 patients assigned to placebo, whereas in CAST-II, 502 patients received moricizine and 491 patients received placebo. Clinical and laboratory baseline variables of patients receiving active drug and those receiving placebo were similar. In CAST-I patients, there was a significant interaction of active therapy with both all-cause death/cardiac arrest and arrhythmic death/cardiac arrest for non-Q-wave myocardial infarction (total mortality hazard ratios, 1.8 versus 7.9 for Q-wave versus non-Q-wave infarction, P = .03). Ventricular premature depolarization (VPD) frequency > or = 50/h and heart rate > or = 74 beats per minute each interacted significantly with total mortality/cardiac arrest only. In the sicker CAST-II patients (ejection fraction < or = 40%), only diuretic use at baseline interacted significantly with moricizine use for both all-cause death/cardiac arrest and arrhythmic death/cardiac arrest (total mortality hazard ratios, 1.9 versus 0.7 for diuretic use versus no use, P = .01). CONCLUSIONS Although active treatment in CAST-I was associated with greater mortality than placebo with respect to almost all baseline variables, the therapeutic hazard was more than expected in patients with non-Q-wave myocardial infarction and (for total mortality) frequent premature VPDs and higher heart rates, suggesting that the adverse effect of encainide or flecainide therapy is greater when ischemic and electrical instability are present. The relative hazard of therapy with moricizine in the sicker CAST-II population was greater in those using diuretics. Thus, although these drugs have the common ability to suppress ventricular ectopy after myocardial infarction, their detrimental effects on survival may be mediated by different mechanisms in different populations, emphasizing the complex, poorly understood hazards associated with antiarrhythmic drug treatment.