Progress of QRS complex width
In the 77 successful cases (29 with narrow baseline QRS and 48 with wide baseline QRS), the mean QRS width was significantly reducedversus baseline (141.3 ± 41 vs. 112.7 ± 15.3 ms;P  < .001). In the successful cases with a baseline QRS of <120 ms (14 in HBP and 15 in LBBAP), the paced QRS was significantly increased (97.3 ± 7.1 vs. 105.5 ± 10.3 ms;P  < .001) and the mean QCI was 8.7 ± 10.5% (range, 0-43.4%), with a slightly lower increase in the HBP group (5.2 ± 10.2% vs. 12 ± 10%; P  = .08). Paced QRS width was prolonged in 19 patients, 15 in the LBBAP group and 4 in the HBP group; a QCI of 0 was obtained in the other 10 patients in the HBP group (7 cases of selective and 3 of non-selective HBP). Paced QRS remained at ≥ 120 ms and <130 m in only three patients (1 LBBAP and 2 HBP). Selective HBP obtained better results than non-selective HBP or LBBAP (Figure 4). In the 48 successful cases with baseline QRS of ≥ 130 ms (25 in LBBAP and 23 in HBP), 23 had right bundle branch block (RBBB), 19 had left bundle branch block (LBBB), and 6 had QRS paced by a previously implanted device. The QCI was related to age, baseline QRS width, baseline LVEF, presence of LBBB or RBBB, and the technique (HBP or LBBAP). In the multiple linear regression, only baseline QRS width (OR –3.4, 95% CI –0.32 to –0.08; P  = .001) and LBBAP technique (OR 3.8, 95% CI 4.4 to 14.5; P  < .001) were retained as independent predictors. The reduction in QCI was greater with LBBAP than with HBP and even greater than with conventional CRT in failed cases (Figure 5).