CONTEXT AND OBJECTIVE: Neonatal hypoxic-ischemic encephalopathy is associated with high morbidity and mortality. Studies have shown that therapeutic hypothermia decreases neurological sequelae and death. Our aim was therefore to report on a three-year experience of therapeutic hypothermia among asphyxiated newborns. DESIGN AND SETTING: Retrospective study, conducted in a university hospital. METHODS: Thirty-five patients with perinatal asphyxia undergoing body cooling between May 2009 and November 2012 were evaluated. RESULTS: Thirty-nine infants fulfilled the hypothermia protocol criteria. Four newborns were removed from study due to refractory septic shock, non-maintenance of temperature and severe coagulopathy. The median Apgar scores at 1 and 5 minutes were 2 and 5. The main complication was infection, diagnosed in seven mothers (20%) and 14 newborns (40%). Convulsions occurred in 15 infants (43%). Thirty-one patients (88.6%) required mechanical ventilation and 14 of them (45%) were extubated within 24 hours. The duration of mechanical ventilation among the others was 7.7 days. The cooling protocol was started 1.8 hours after birth. All patients showed elevated levels of creatine phosphokinase, creatine phosphokinase- MB and lactate dehydrogenase. There was no severe arrhythmia; one newborn (2.9%) presented controlled coagulopathy. Four patients (11.4%) presented controlled hypotension. Twenty-nine patients (82.9%) underwent cerebral ultrasonography and 10 of them (34.5%) presented white matter hyper-echogenicity. Brain magnetic resonance imaging was performed on 33 infants (94.3%) and 11 of them (33.3%) presented hypoxic-ischemic changes. The hospital stay was 23 days. All newborns were discharged. Two patients (5.8%) needed gastrostomy. CONCLUSION: Hypothermia as therapy for asphyxiated newborns was shown to be safe.