The airway occlusion pressure, P0.1, is the negative airway pressure generated during the first 100 msec of an occluded inspiration. P0.1 is a parameter for the neuro-muscular activation of the respiratory system, which is an important determinant for the work of breathing. It has been shown to be a good predictor for successful weaning from mechanical ventilation. Standard P0.1 measurement techniques are based on a total occlusion of the inspiration for more than 100 msec. These measurements are technically complex and therefore not useful for clinical purposes. Furthermore, a significant breath-by-breath variability has been shown for P0.1, which is neglected by any single point measurement technique. Therefore, we have developed a continuous on-line measurement for breath-by-breath determination of P0.1 using the Siemens Servo 900C respirator. In triggered mechanical ventilation the delay time between the onset of the patient's inspiration and flow delivery from the respiratory is more than 100 msec for this respirator. During that time the inspiration is occluded. Therefore, the trigger effort was proposed to be a good estimate of P0.1. Based on this, we calculated P0.1 as follows: airway pressure (Paw) was registered at the endotracheal tube site of the respiratory tubing, digitized and acquired by a personal computer at 100 Hz. The recorder output of the Servo 900C was connected to the same computer, delivering the electronical signal for the inspiratory valve to open when the inspiratory effort has exceeded the trigger threshold, which needs a minimal delay time of 80 msec. Around 20 msec after this signal flow is delivered from the respirator. The computer runs an algorithm, which recognizes this signal and calculates P0.1 (Servo P0.1) as the slope of the pressure drop during this 100 msec. Paw tracings and the calculated P0.1 values were displayed on the computer screen and stored on disk. This method was validated by comparing it to the standard technique, using a Hans-Rudolph valve for inspiratory occlusion and calculating P0.1 from Paw tracings during the occluded inspiration. For validation we used a mechanical lung model which generated P0.1 values ranging between 1.1-10.3 mbar. For a given adjustment of the lung model two standard measurements (standard P0.1) were made and compared to the Servo P0.1. In a total of 21 measurements the mean Servo P0.1 was 4.9 +/- 2.9 mbar; the mean standard P0.1 was 4.3 +/- 2.5 mbar. The mean difference between Servo P0.1 and standard P0.1 was 0.6 +/- 0.6 mbar (range: -0.3-1.8 mbar). The regression equation for linear regression analysis was: Servo P0.1 = 1.15* standard P0.1-0.05. This correlation was significant (r = 0.99, p < 0.01). From these data we conclude that the described method for continuous P0.1 measurement provides reliable values with the advantage of a maneuver-free, breath-by-breath measurement technique. It thereby opens the possibility for monitoring the neuro-muscular activation of the respiratory system at the bedside, which is shown as an example for a patient during weaning from mechanical ventilation.