The polygraphic findings from 11 future victims of sudden infant death syndrome (SIDS) are reported and compared with those of matched pairs of control infants. The recordings had been done to alleviate parental anxiety about sleep apnea. Four infants had siblings who were victims of SIDS. Two infants were studied 3.5 to 9.5 weeks before their deaths because of an unexplained apparent life-threatening event that had occurred during sleep. For each victim of SIDS, two control infants were selected from the 2,000 infants who had been tested in the same hospitals. They were matched for sex, gestational age, postnatal age, and weight at birth with the SIDS victims. Their polygraphic recordings had been performed within similar conditions. Each record was allocated a random code number and was analyzed without knowledge of the patient's identity by two independent scorers. Most sleep and cardiorespiratory variables studied did not differentiate SIDS victims from control infants. Only four variables significantly characterized the future SIDS victims: the maximal duration of central apneas, the number of sighs followed by a central apnea, the presence of obstructive apneas, and the presence of mixed apneas. Central apneas were longer during all sleep states in the SIDS victims compared with their matched controls, but none exceeded 14 seconds. Sighs immediately followed by an apnea were significantly less frequent in the future SIDS group. Obstructive and mixed sleep apneas were seen in eight of 11 SIDS victims and in only three of 22 control infants.(ABSTRACT TRUNCATED AT 250 WORDS)