AIMS: Cardiac resynchronization therapy (CRT) is an effective treatment for heart failure patients with prolongation of QRS duration. Despite careful patient selection, some do not respond to CRT based on QRS complex duration. We sought to evaluate the presence of left ventricular dyssynchrony using tissue Doppler imaging (TDI) according to QRS duration in heart failure patients. METHODS AND RESULTS: Ninety-nine patients (mean age 52.6 +/- 15.3 years) with severe heart failure [left ventricular (LV) ejection fraction, <35%] were prospectively evaluated. On the basis of QRS width, the patients were divided into two groups. Forty-eight patients (48.5%) had a normal QRS duration (<120 ms), Group I, and 51 (51.5%) had a prolonged QRS duration, Group II. All patients underwent echocardiography coupled with TDI. Spectral displays of six basal and six middle LV segments with pulsed-wave TDI were obtained to assess the time to peak systolic point from R-wave on electrocardiogram (Ts). The standard deviation of Ts (Ts-SD) and the maximal temporal difference of Ts (Ts-diff) were measured. Interventricular dyssynchrony [defined as the presence of an interventricular mechanical delay (IVMD) >40 ms] and intra-LV mechanical delays (defined as Ts-SD >33.4 ms and Ts-diff >100 ms) were correlated with the QRS width and morphology. We found a greater IVMD in Group II patients, compared with patients in Group I (42.5 +/- 22.3 vs. 26.8 +/- 21, respectively, P < 0.001). Intraventricular dyssynchrony defined as Ts-SD > or =33.4 ms was found in 45.1% of patients in Group II compared with 23% of patients in Group I (P = 0.03). Similarly, the Ts-diff was prolonged in Group II patients compared with Group I (P = 0.02). By linear regression analysis, a weak relation was found between Ts-SD and QRS duration (P = 0.055). A substantial portion of patients with prolonged QRS did not exhibit ventricular dyssynchrony defined either as total asynchrony index > or =33.4 ms or as IVMD >40 ms. CONCLUSION: A substantial proportion of patients with prolonged QRS (32.1%) did not exhibit inter- or intraventricular dyssynchrony, which may represent a limitation in identifying the ideal QRS interval for the selection of patients for CRT.