The presence or absence of fetal breathing movements may be helpful in differentiating between true and false preterm labor. We attempted to demonstrate the clinical utility of this simple ultrasonic observation in predicting short-term delivery outcome during suspected preterm labor. A total of 50 pregnancies between 26 and 34 weeks' gestation with presumed preterm labor were observed in a prospective manner. During 20 minutes of observation with real-time ultrasound at the time of admission, fetal breathing movements were observed in 33 patients and considered absent in the remaining 17 patients. In those pregnancies with absent fetal breathing movements, true labor with subsequent delivery occurred in 16 patients. Of the 33 pregnancies with fetal breathing movements present, 29 continued for greater than 48 hours. It appears that the absence of fetal breathing movement is a reliable indicator of imminent preterm delivery, irrespective of fetal membrane status (p less than 0.0001). The observed mean sensitivity and specificity of this phenomenon in predicting short-term delivery outcome are 96.6% +/- 3.3% (mean +/- SD) and 80.0% +/- 8.9%, respectively. A multivariant statistical model based on the frequency of contractions, white blood cell counts, initial cervical examination results, and premature rupture of membranes could successfully predict delivery outcome in 40% to 75% of cases. The addition of fetal breathing movement analysis to the model allowed for the correct prediction of outcome in 90% of the cases.