BACKGROUND Several studies have demonstrated the efficiency and safety of mild hypothermia (33 degrees C) used for treating moderate encephalopathy. In animal models, deep hypothermia proved to be neuroprotective. OBJECTIVE To determine the safety of whole-body deep hypothermia between 30 and 33 degrees C in moderate-severe hypoxic-ischemic encephalopathy in newborn term infants. METHODS Mortality rates, incidence of brain damage detected by magnetic resonance imaging (MRI) and neurological outcomes of 39 term asphyxiated infants were retrospectively compared. A first group of patients (control group C) was treated with routine standard methods, a second group (MH) was treated with mild whole-body hypothermia (32-34 degrees C) and a third group (DH) was treated with deep whole-body hypothermia (30-33 degrees C), for 72 h. Mean arterial pH, basic excess (BE) and lactic acid in the blood were measured. Laboratory and clinical side effects of hypothermia were investigated. A conventional brain MRI was performed after the second week of life. RESULTS 39 term asphyxiated newborns were enrolled in the study: 11 in group C, 10 in group MH, and 18 in group DH. During the first 72 h, disseminated intravascular coagulation was recorded in 2 cases (18%) in group C, pulmonary hypertension in 2 patients (20%) in group MH, and pneumonia in 3 cases (16%) in group DH. Severe cerebral lesions and poor neurological outcome were observed in 4 cases (36%) in group C, 1 case (10%) in group MH, and 1 case (5%) in group DH. A statistically significant difference in brain damage and major clinical neurological abnormalities was observed between group C and groups MH and DH, whereas no differences were demonstrated between asphyxiated infants treated with mild or deep hypothermia. CONCLUSIONS The results support the safety of deep hypothermia. Further studies are needed to confirm these results and the neuroprotective effect of this approach.