In mechanically ventilated patients with acute respiratory failure, the static compliance of the total respiratory system is conventionally obtained by dividing the tidal volume by the difference between the "plateau" pressure measured at the airway opening (PaO) during an occlusion at end-inspiration and positive end-expiratory pressure (PEEP) set by the ventilator. This analysis is valid only if the elastic recoil pressure of the respiratory system is zero at the end of expiration, indicating that the system has reached its elastic equilibrium point. To test if this is always the case, in 14 mechanically ventilated patients with acute respiratory failure, measurements were made of PaO and of flow and volume changes. In only 4 of the patients did expiratory flow become nil before end-expiration and inspiratory flow started synchronously with the onset of the positive-pressure swing delivered by the ventilator, indicating that in these 4 patients the end-expiratory elastic recoil pressure was indeed zero. By contrast, in the remaining 10 subjects, expiratory flow was still present when the ventilator had already begun to increase PaO, indicating that the end-expiratory elastic recoil pressure was not zero. Indeed, in all these 10 patients, a positive delta PaO (as much as 7.5 cm H2O) had to be applied by the ventilator before the actual onset of inspiratory flow. This delta PaO represents the pressure required to counterbalance the end-expiratory elastic recoil before inspiratory flow will begin, and can be termed intrinsic PEEP.(ABSTRACT TRUNCATED AT 250 WORDS)