Selection bias in electrophysiologically-guided therapy of sustained ventricular tachyarrhythmias

To analyze if patient selection bias could contribute to the improved prognosis reported for ventricular tachycardia (VT) and ventricular fibrillation (VF) when therapy is guided by electrophysiologic studies (EPS), we studied 90 consecutive patients admitted to a tertiary referral center with recent VT/VF who were candidates for EPS. Seventeen patients (19%) died during the initial hospital admission, and 30 (33%) died after discharge. Survival probability was 0.83 (95% confidence interval [CI], 0.74-0.90); 0.67 (95% CI, 0.56-0.75); and 0.53 (95% CI, 0.42-0.63) at 1 month, 1 year, and 3 years, respectively. Of the 56 patients (62%) who underwent EPS during their initial hospitalization, only 1 died during that admission. Patients in whom EPS could not be performed had characteristics associated with a poorer prognosis. NYHA functional class (p = 0.005), inability to perform baseline EPS (p = 0.003) and use of digoxin (p = 0.016) were independent predictors of death. Early in-hospital mortality in patients with VT/VF remains high. Thus, omission of these deaths in reports of EPS-guided therapy creates incomplete, biased cohorts. Furthermore, there may be a bias toward a healthier population among hospital survivors undergoing EPS. These findings may contribute to better outcomes in current series compared to historical controls.