Left ventricular activation delay (LVAT) >100 ms may determine response to cardiac resynchronization therapy, but its prevalence and relation to QRS configuration are unknown. QRS duration and LVAT in control subjects (n = 30) were compared with those in patients with heart failure (HF; LV ejection fraction 23 +/- 8%, n = 120) with a QRS duration <120 ms (NQRS(HF), n = 35) or > or = 120 ms (left bundle branch block [LBBB(HF)], n = 54; right bundle branch block [RBBB(HF)], n = 31). LVAT was estimated by interval from QRS onset to basal inferolateral LV depolarization. In controls, QRS duration was 82 +/- 13 ms and LVAT was 55 +/- 18 ms. LVAT was always <100 ms. In patients with NQRS(HF), QRS duration (104 +/- 10 ms) and LVAT (82 +/- 22 ms) were prolonged versus controls (p <0.001). LVAT exceeded 100 ms in 8 of 35 patients. In patients with LBBB(HF), QRS duration (161 +/- 29 ms) and LVAT (136 +/- 33 ms) were prolonged compared with controls and patients with NQRS(HF) (p <0.001). LVAT exceeded 100 ms in 47 of 54 patients. In patients with RBBB(HF), QRS duration did not differ from that in patients with LBBB(HF), but LVAT (100 +/- 24 ms) was shorter (p <0.001). In 17 of 31 patients with RBBB(HF) LVAT was <100 ms (82 +/- 12), similar to those with NQRS(HF) (p = NS), indicating no LV conduction delay. However, in 7 of 31, LVAT (135 +/- 13 ms) was similar to that in patients with LBBB(HF) (p = NS). LVAT correlation with QRS duration varied (control p = 0.004, NQRS(HF) p = 0.15, RBBB(HF) p = 0.01, LBBB(HF) p <0.001). In conclusion, LV conduction delays in patients with HF varied with QRS configuration and duration, exceeding 100 ms in only 23% of patients with narrow QRS configuration and 45% with RBBB(HF) compared with 87% with LBBB(HF). Fewer than 25% of patients with RBBB(HF) demonstrated delays equivalent to those in patients with LBBB(HF.) These variations may affect efficacy to cardiac resynchronization therapy.