Administration of 100% O2 to preterm infants induces an apnea that is usually central. We hypothesized that this apnea may be "mixed" at times with an obstructive component appearing late during the respiratory pause. In addition, we reasoned that obstruction would depend on the duration of the apnea. Thus, we gave 100% O2 to 61 healthy preterm infants. Group 1 was > or = 1,500 g [birth wt 1.8 +/- 0.1 (SE) kg, gestational age 32 +/- 1 wk, postnatal age 19 +/- 2 days, n = 26] and group 2 was < 1,500 g [birth wt 1.2 +/- 0.1 kg, gestational age 29 +/- 1 wk, postnatal age 30 +/- 4 days, n = 35]. Ventilation was measured using a flow-through system. Respiratory efforts in the absence of flow were detected using chest and abdominal displacements or diaphragmatic electromyography. In group 1, 19% of the central apneas became obstructive at 17 +/- 3 s, whereas in group 2, 34% did so at 12 +/- 2 s. Mixed apneas were longer than those without obstruction (28 +/- 3 vs. 12 +/- 1 s; P = 0.0001). The incidence of mixed apneas was 0, 14, and 66% in group 1 and 0, 27, and 69% in group 2 in apneas of 3-10, 11-20, and > 20 s, respectively. These findings suggest that 1) a percentage of the central apneas induced by inhaling 100% O2 became obstructive, 2) the incidence of the obstructive component increased with the duration of apnea, and 3) smaller infants became obstructed sooner and had a higher incidence of obstruction than larger infants.(ABSTRACT TRUNCATED AT 250 WORDS)