Effect of reproducibility of baseline arrhythmia induction on drug efficacy predictions and outcome in the Electrophysiologic Study Versus Electrocardiographic Monitoring (ESVEM) trial.

Spontaneous variability over time in the ease of induction of ventricular arrhythmias may mimic a drug effect and affect the predictive value of drug therapy guided by programmed stimulation. We assessed the effect of baseline reproducibility of arrhythmia induction on the incidence and accuracy of drug efficacy predictions in the Electrophysiologic Study Versus Electrocardiographic Monitoring (ESVEM) trial. Patients with sustained ventricular tachyarrhythmias induced twice during baseline electrophysiologic testing with the same stimulation technique, i.e., induced at the same pacing site with the same drive cycle length and number of extrastimuli, were identified from the ESVEM database. These patients with highly reproducible arrhythmia induction were compared to those with less reproducible arrhythmias. Of 473 randomized patients with reproducibility data, 313 (66%) had highly reproducible arrhythmias. In patients randomized to electrophysiologic testing, baseline arrhythmia reproducibility did not affect the incidence of drug efficacy predictions (70 of 157 [45%], drug efficacy predictions in patients with highly reproducible arrhythmias vs 34 of 79 [43%] with less reproducible arrhythmias, p = 0.890). Drug efficacy predictions obtained by electrophysiologic testing in patients with highly reproducible arrhythmias were not associated with decreases in arrhythmia recurrence (p = 0.202), all-cause mortality (p = 0.301), cardiac death (p = 0.358), or arrhythmic death (p = 0.307) compared to those with less reproducible arrhythmias. Analysis of patients with highly reproducible sustained monomorphic ventricular tachycardia led to similar results. In the ESVEM trial, most patients had highly reproducible arrhythmia induction during baseline electrophysiologic testing. Reproducibility of arrhythmia induction in the baseline state had no effect on the incidence or accuracy of drug efficacy predictions.