Events in the cardiac arrhythmia suppression trial (CAST): Mortality in patients surviving open label titration but not randomized to double-blind therapy

The patient characteristics and outcomes were studied in the 318 patients who survived open label drug titration in the Cardiac Arrhythmia Suppression Trial (CAST) and who were not randomized to double-blind therapy and in 942 patients, who were randomized to double-blind placebo therapy. The patients randomized to placebo therapy had a lower total mortality or resuscitated cardiac arrest rate (4% vs. 8.5%). However, at baseline, nonrandomized patients were dissimilar from patients randomized to placebo in the following ways: older; lower left ventricular ejection fraction; greater use of digitalis, diuretic drugs and antihypertensive agents; lesser use of beta-adrenoceptor blocking agents and more frequent prior cardiac problems, including runs of ventricular tachycardia and left bundle branch block. A matched comparison that took these inequities into account showed no significant differences in mortality or rate of resuscitation from cardiac arrest between nonrandomized patients and clinically equivalent patients randomized to placebo. Cox regression analysis indicated that two factors significantly increased the hazard ratio for arrhythmic death or resuscitated cardiac arrest in the nonrandomized patients: female gender (4.7, p < 0.05) and electrocardiographic events (ventricular tachycardia, proarrhythmia, widened QRS complex, heart block, bradycardia) during open label titration (7.0, p < 0.005). However, some potentially important differences between men and women were not included in the Cox regression model. Of the nonrandomized patients, approximately 70% were not randomized because of lack of suppression of ventricular premature depolarizations or adverse events, or both, and the remaining 30% because of patient or private physician request with no indication of another reason. In conclusion, nonrandomized patients had more extensive coronary heart disease and experienced higher mortality and resuscitated cardiac arrest rates than did patients randomized to placebo. This finding at least partly explains the low mortality rate in the CAST patients randomized to placebo. Events such as proarrhythmia during drug titration portend a high risk of arrhythmic death. Many of the patients who most need antiarrhythmic benefit are unable to tolerate these drugs. In CAST, virtually all patients who could and would be treated with an antiarrhythmic drug were randomized.