The purpose of this study was to determine whether high sensitivity C-reactive protein (hsCRP) before cardiac re-synchronization therapy (CRT) implantation was able to predict the response to CRT and cardiac deaths in severe heart failure patients. The study population consisted of 65 heart failure patients (46 males, mean age 65.0 ± 11.8 years, NYHA class III/IV) with CRT implantation. Levels of hsCRP and B-type natriuretic peptide (BNP) were measured before CRT implantation. Left ventricular end-diastolic volume index (LVEDVI), left ventricular end-systolic volume index (LVESVI), and left ventricular ejection fraction (LVEF) were assessed by echocardiography at the same time. At 6 months after device implantation follow-up, echocardiography was performed and reverse remodeling was defined as > 15% reduction in LVESV. Of the 61 patients (4 patients died within 6 months), 41 patients (67%) and 20 patients (33%) were classified as responders (group-R) and nonresponders (group-NR), respectively. Cardiac deaths occurred more frequently in group-NR than in group-R (29% versus 5%, P < 0.05). Hs-CRP level was significantly higher in group-NR than in group-R (P < 0.01). Multivariate logistic regression analysis showed an independent relationship between hsCRP and the incidence of nonresponders (odds ratio: 1.499, P = 0.011). Stepwise multivariate Cox proportional hazard analysis identified the hsCRP level as the strongest predictive factor for cardiac death (hazard ratio: 1.337, P = 0.001). Receiver operating characteristic (ROC) analysis revealed hsCRP levels of 3.0 mg/L as the cut-off value for cardiac mortality. The hsCRP level may provide a new insight into CRT implantation for severe heart failure by predicting responses to CRT and cardiac death.