To test the hypothesis that sighs are mechanistically important in triggering apnea, we studied 10 preterm infants, group 1: body weight 1.8 +/- 0.1 kg, gestational age 33 +/- 1 weeks, postnatal age 21 +/- 4 days, and 10 term infants, group 2: body weight 3.9 +/- 0.15 kg, gestational age 40 +/- 0.4 weeks, postnatal age 1.4 +/- 0.2 days. Instantaneous ventilatory changes associated with a sigh were studied in another 10 preterm infants, group 3: body weight 1.6 +/- 0.11 kg, gestational age 32 +/- 0.4 weeks, postnatal age 25 +/- 4 days. Ventilation was measured using a nosepiece and a flow-through system. Sleep states were recorded. Sighs were more frequent in preterm than in term infants (0.4 +/- 0.04 vs. 0.18 +/- 0.03 sighs/min; p = 0.03) and in rapid eye movement than in quiet sleep (0.5 +/- 0.05 vs. 0.3 +/- 0.05 sighs/min; p = 0.05). Of 722 apneas, 235 (33%) were associated with a sigh; of these, 113 (48%) preceded and 122 (52%) followed a sigh. Sighs induced with airway occlusion (groups 1 and 2) were more frequent after occlusion on 21 than on 35% O2, particularly when O2 saturation was low and negative airway pressure high. Instantaneous ventilation measured over 10 breaths preceding a sigh did not show any trend indicating the possible appearance of a sigh. Tidal volume increased from 7.5 +/- 0.7 before the sigh to 18.9 +/- 0.7 ml/kg (p < 0.01) during a sigh, with a significant increase in inspiratory drive. Ventilation increased from 0.327 +/- 0.041 to 0.660 +/- 0.073 l/min/kg.(ABSTRACT TRUNCATED AT 250 WORDS)