Biventricular pacing results in left ventricular (LV) reverse remodeling in heart failure patients with wide QRS complexes. This study examines potential predictors of reverse remodeling. Echocardiography with tissue Doppler imaging was performed at baseline and 3 months after biventricular pacing in 30 patients (21 men and 9 women, mean age 62 +/- 14 years). There were 17 responders to reverse remodeling (defined as a reduction in LV end-systolic volume by >15%) and 13 nonresponders. Responders had significant improvement in 6-minute hall-walking distance (p = 0.006), metabolic equivalents (p = 0.02), peak oxygen uptake (p = 0.02), New York Heart Association functional class (p <0.001), and quality of life (p <0.001); an increase in the sphericity index (p = 0.007), ejection fraction (p <0.001), and diastolic filling time (p = 0.03); a decrease in myocardial performance index (p = 0.02), isovolumic relaxation time (p = 0.004), and mitral regurgitation (p = 0.007); and an improvement in systolic dyssynchrony (SD of the time to peak myocardial systolic contraction of the 12 LV segments as dyssynchrony index) (45.0 +/- 8.3 vs 32.5 +/- 14.5 ms, p = 0.003). In contrast, nonresponders only had a small degree of clinical improvement in New York Heart Association class (p = 0.03) and quality-of-life scores (p = 0.03), without any change in cardiac function, and worsening of systolic dyssynchrony (24.8 +/- 4.5 vs 34.1 +/- 13.5 ms, p = 0.02). When all the above factors were put into univariate and multivariate analyses models, systolic dyssynchrony was the only independent predictor of reverse remodeling (r = -0.76, p <0.001) (beta = -1.54, p = 0.007). A preimplant dyssynchrony index of 32.6 ms (+2 SDs from mean of 88 normal controls) was able to totally segregate responders from nonresponders of biventricular pacing. Thus, responders of LV reverse remodeling were associated with improvement in clinical status, cardiac function, and systolic synchronicity. Direct assessment of systolic synchronicity by tissue Doppler imaging is highly accurate in predicting responders to therapy.