Before the era of therapeutic hypothermia, a practice parameter including absent or extensor motor response at day 3 and the presence of myoclonus status epilepticus within 24 hours was used to assist prognostication of poor neurologic outcomes. There are conflicting results concerning whether hypothermia influences the reliability of the predictors and the accurate predictors and optimal timing for assessing neurologic recovery are largely unknown. Several prognostic indicators other than guidelines are also applied to help determining prognosis, including electroencephalogram, cerebral computed tomographic scan, and cerebral perfusion scintigraphy single-photon emission computed tomographic scan. Here, we present a cardiac arrest survivor treated with therapeutic hypothermia waked up finally on the 13th day, although clinical and laboratory examinations after return of spontaneous circulation all indicated poor neurologic prognosis. However, life support was reported to be withdrawn within 3 to 5 days in 25% to 50% cardiac arrest survivors treated with hypothermia when grave prognosis was predicted. The clinical course of the patient raises some important questions concerning the accuracy of current predictors, the optimal observation period for neurologic recovery, and the appropriate timing to determine withdrawal of life support in cardiac arrest victims receiving therapeutic hypothermia.