Aim To compare diagnostic significance of different criteria for complete left bundle branch block (cLBBB) in prediction of reverse left ventricular (LV) remodeling associated with cardiac resynchronization therapy (CRT).Materials and methods This study included 93 patients (men, 81.7 %; mean age at the time of implantation, 56.6±9.3 years). Achievement of a maximum decrease in LV end-systolic volume (ESV) was recorded during the entire follow-up period for evaluation of LV reversibility by CRT. Based on the dynamics of LV ESV, patients were divided into two groups, non-responders (n=27) and responders (n=66). cLBBB was determined by 9 criteria (ESC 2006 and 2013, AHA 2009, Strauss, and MIRACLE, CARE-HF, MADIT-CRT, REVERSE, and RAFT used in large multicenter studies).Results Incidence of cLBBB was significantly higher in the group of responders as demonstrated by the AHA (p=0.001), ESC 2013 (p=0.014), Strauss (p=0.002), MADIT-CRT (p=0.014), REVERSE (p=0.013), and RAFT (p<0.001) criteria. The highest specificity was shown for the AHA and RAFT (92.6 %) criteria, and the highest sensitivity and overall accuracy were shown for the Strauss (80.3 % and 72.04 %, respectively) criterium. The criteria proposed in actual clinical guidelines (AHA and ESC 2013) demonstrated a strong consistency in detecting cLBBB (κ=0.818, 95 % CI, 0.7–0.936; p<0.001). However, the Strauss and ESC 2006 / AHA / ESC 2013 showed the least consistency in identifying cLBBB. For the criteria described in large multicenter studies, consistency in detecting cLBBB was minimal in most cases. However, criteria with moderate or strong consistency were used in the studies, which results have substantiated the use of cLBBB as a selection criterium (MADIT-CRT, REVERSE, and RAFT).Conclusion The reversibility of LV remodeling associated with CRT was different in patients with cLBBB determined by different criteria. All actual cLBBB criteria (AHA, ESC 2013, and Strauss) were significantly more frequently observed in the responder group. Nevertheless, these criteria differed in their sensitivity and specificity. A number of large multicenter studies have used criteria with minimal consistency in detecting cLBBB, which should be taken into account in interpreting results of these studies.