OBJECTIVE Out-of-hospital cardiac arrest (OHCA) is common and carries a bleak prognosis. Early prediction of unfavourable outcomes is difficult but crucial to improve resource allocation. The aim of this study was to develop a simple tool for predicting survival with good neurological function in the overall population of patients with successfully resuscitated cardiac arrest. RESULTS We used logistic regression analysis to identify clinical and laboratory variables that were both readily available at admission and predictive of poor outcomes (death or severe neurological impairment) in a development cohort of 130 consecutive OHCA patients admitted to a French intensive care unit (ICU) between 1999 and 2003. To test the prediction score built from these variables, we used a validation cohort of 210 patients recruited in four French ICUs between 2003 and 2005. Initial rhythm, estimated no-flow and low-flow intervals, blood lactate, and creatinine levels determined using whole blood analyzers were independently associated with poor outcomes and were used to build a continuous severity score. Goodness-of-fit tests indicated good performance (P=0.79 in the development cohort and P=0.13 in the validation cohort). The area under the receiver-operating characteristics curve was 0.82 in the development cohort and 0.88 in the validation cohort. CONCLUSIONS The outcome can be accurately predicted after OHCA using variables that are readily available at ICU admission.