To evaluate the relationship between blood glucose and neurological recovery after cardiac arrest, we retrospectively reviewed our experience with 430 consecutive patients resuscitated after out-of-hospital cardiac arrest. All these patients had received variable amounts of intravenous 5% glucose solutions before admission. Awakening patients, those following commands or having comprehensible speech, were classified as to whether they suffered persistent neurological deficits. The mean blood glucose level on hospital admission was higher in 154 patients who never awakened than in 276 who did awaken (mean, 341 versus 262 mg per 100 milliliters; p less than 0.0005). Among the 276 who awakened, 90 patients with persistent neurological deficits had higher mean glucose levels on admission than did 186 without deficits (286 versus 251 mg per 100 milliliters; p less than 0.02). These significant differences persisted after excluding all patients whose glucose levels were higher than 500 mg per 100 milliliters and after controlling for potentially confounding variables using multiple regression analysis.