Abstract
To compare the accuracy of clinical observation and detailed respiratory recordings in identifying infants at high risk of developing pneumothoraces 10 infants, with idiopathic respiratory distress syndrome, were studied at three different ventilator rates. All infants with synchronous respiration at fast rates were correctly identified by clinical observation. The clinical signs used to identify 'high risk' interactions--that is, active expiration and asynchronous breathing--were obvious respiratory efforts and a failure of improvement in oxygenation at increased rates. These criteria enabled correct identification of 'high risk' respiratory patterns on 15 (88%) of the 17 study occasions. These clinical criteria were then used as criteria for selective paralysis; no infant developed a pneumothorax during ventilation.