Sighs, the interspersed breaths of larger tidal volume (V-T) than those preceding or following them, occur more frequently in the newborn than in adults. It has been postulated that the sigh augments functional residual volume (FRC), increases lung compliance (C-L), and recruits atelectatic alveoli. Since the premature infant has relatively low FRC, we hypothesized that sighs might be more frequent in these babies and that C-L could be shown to change before and/or after a sigh. We studied 14 premature infants (gestational ape, 30 +/- 2.2 weeks; mean +/- SD) without lung disease during quiet sleep. Minute ventilation (V-E) was measured by pneumotachography and pleural pressure changes via a water-filled catheter in the esophagus. In the 78 sighs examined, the incidence was 0.71 sighs/min, similar to previous reports for full-term newborns. There was no change in respiratory frequency, V-T, V-E, nor mean inspiratory flow (V-T/T-I, where T-I is the inspiratory time) before or after the event. The volume of the sigh increased from 7.8 +/- 0.2 mL/kg to 21.5 +/- 0.5 mL/kg, in concert with an increase in breath duration (T-TOT) from 0.98 +/- 0.02 s to 2.56 +/- 0.12 s, resulting in an unchanged V-E, V-T/T-I, and respiratory duty cycle (T-I/T-TOT). This suggests that the increased V-T was matched by a prolongation in both T-I and T-TOT. Following the sigh, C-L increased from 1.43 +/- 0.13 mL/cmH(2)O/kg to 1.66 +/- 0.27 mL/cmH(2)O/kg (paired t-test, P < 0.001) whilst the lung resistance R(L) fell from 62 +/- 8 cmH(2)O/L/s/kg to 55 +/- 3.8 cmH(2)O/L/s/kg (P < 0.05). As reflected by the end-expiratory volume (EEV), FRC did not change. This stability of FRC, coupled with the increase in C-L and fall in R(L), suggests a recruitment of atelectatic acini with a redistribution of lung volume, such that the previously atelectatic and overdistended acini achieve a more optimal pressure/volume relationship following a sigh. (C) 1994 Wiley-Liss, Inc.