Objective: To determine the influence of presence of caffeine in umbilical cord blood on apnea occurrence. Methods:A prospective cohort study with preterm newborns with birth weight lower than 2,000 g was undertaken. Exclusion criteria were: mothers who received opioids; mechanical ventilation during the first 4 days of life; cerebral and major cardiac malformations; perinatal asphyxia; severe periintraventricular hemorrhage; exchange transfusion before the fourth day of life; and those who received methylxantine prior to extubation. Neonates were divided into detectable and undetectable caffeine in umbilical cord blood. Newborns were followed for the first 4 days for occurrence of apnea spells.Results: Eighty-seven newborns with and 40 without detectable caffeine in umbilical cord blood were studied. Median caffeine concentration of the 87 patients with detectable caffeine in umbilical blood was 2.3 µg/mL (0.2-9.4 µg/mL). There was no association between occurrence of apnea spells and presence of caffeine in umbilical cord blood. Neonates with detectable caffeine in umbilical blood had borderline later apnea (66.3±4.14 hours) than those with undetectable levels (54.2±6.26 hours). Conclusion:Detected levels of caffeine in umbilical cord blood did not decrease occurrence of apnea of prematurity, but it had a borderline effect delaying its occurrence, suggesting that even a low level of caffeine in umbilical cord blood might delay occurrence of apnea spells.J Pediatr (Rio J). 2010;86(2):137-142: Prematurity, apnea, caffeine, low-birth-weight infant.