It has been hypothesized that hypoxanthine concentrations in the blood of newborn infants are a marker of asphyxia. To test this hypothesis, we measured serum hypoxanthine levels in relationship to perinatal and neonatal asphyxia, and compared arterial hypoxanthine levels with arterial pH and base deficit. We also compared hypoxanthine levels of survivors with those of asphyxiated non-survivors. Forty-two newborns were classified as asphyxiated by either of two methods: 1) Infants from whom umbilical cord hypoxanthine levels were taken were classified as asphyxiated if they had an Apgar score of 6 or less at 1 or 5 minutes, fetal heart rate below 100 beats per minute, or meconium-stained amniotic fluid; and 2) infants from whom peripheral arterial hypoxanthine samples were taken were classified by clinical assessment, whereby one author, blinded to the infants' hypoxanthine levels, prospectively assessed each patient's condition for evidence of asphyxia. Hypoxanthine levels correlated with increased base deficit (P less than .001; r = 0.8) and with decreased pH (P less than .001; r = -0.5). By both of our asphyxia classification methods, hypoxanthine levels were significantly higher (P less than .002) in the asphyxiated groups. We also noted a higher hypoxanthine level in asphyxiated non-survivors as compared with all survivors (P less than .02). We propose that serum hypoxanthine levels may help define asphyxia. Because hypoxanthine, when metabolized by xanthine oxidase, generates oxygen radicals that are highly destructive to tissue, hypoxanthine levels may have important therapeutic implications for asphyxiated patients.