Thirty patients received one of the lidocaine analogues--mexiletine or tocainide--orally for treatment of symptomatic ventricular arrhythmias. Crossover to the other analogue was allowed if initial drug treatment was unsuccessful, and the controlled use of other marketed oral antiarrhythmic agents was permitted. After follow-up of 7 +/- 3 months (SD), mexiletine was successful in 5 of 13 patients initially and in 5 of 14 patients who failed to respond to tocainide. Tocainide was successful in 1 of 17 patients initially and in 2 of 7 who did not respond to mexiletine. Combination therapy was used in nearly half of all ultimately successful drug trials. A common cause of drug trial failure for both drugs was the occurrence of adverse effects that frequently appeared well after hospital discharge. Response to lidocaine was a sensitive but nonspecific predictor of clinical outcome with mexiletine or tocainide that helped to identify drug-resistant patients. Finally, although mexiletine provided effective antiarrhythmic therapy more often than tocainide, response to one lidocaine analogue did not predict response to the other.