We measured the ventilatory pattern and additional work of breathing (WOBadd) at three different levels of inspiratory pressure support [IPS 5, 10, 15 mbar above positive end-expiratory pressure (PEEP)] and in a new ventilatory mode, automatic tube compensation (ATC), in nine operative patients without lung injury nine patients ventilated for several following acute respiratory insufficiency (ARI). In ATC, endotracheal tube resistance is compensated automatically by means of closed-loop control of the calculated tracheal pressure. Pressure support in this mode, i.e. airway pressure above PEEP, is equal to the actual flow-dependent pressure drop across the endotracheal tube (ETT). Airway pressure rises at the beginning of inspiration and falls towards the end. As the tube resistance of ETT seriously hinders expiration and can cause desynchronization between ventilator and patient, airway pressure is reduced below PEEP during expiration in the same way as it is increased during inspiration. The result is a near-constant tracheal pressure at PEEP both during inspiration and during expiration. This mode could be best termed as "electronic extubation". The most striking difference between the postoperative patients and the ARI patients was their minute ventilation (17.8 +/- 1.85 l/min in ARI patients vs 7.3 +/- 3.1 l/min in the postoperative patients). In the postoperative patients augmentation of IPS from 5 to 15 mbar induced a steady increase in tidal volume (VT) and a consecutive decrease in respiratory rate (rr) compared with ATC (VTATC,postop = 463 +/- 78 ml; rrATC,postop = 16 +/- 4 min-1; VTIPS5.postop = 505 +/- 79 ml; rrIPS5,postop = 15 +/- 4 min-1; VTIPS10,postop = 562 +/- 86 ml; rrIPS15,postop = 14 +/- 4 min-1; VTIPS15.postop = 660 +/- 151 ml; rrTPS15,postop = 12 +/- 4 min-1), whereas the augmentation of IPS of 5 and 10 mbar in the ARI patients could not compensate for the increase in rr and the decrease in VT, after switching from ATC to IPS (VTATC,ARI 724 +/- 308 ml, rrATC,ARI = 24 +/- 6 min-1; VTIPS5,ARI = 649 +/- 315 ml; rrIPS5,ARI = 27 +/- 8 min-1; VTIPS10,ARI = 653 +/- 353 ml; rrIPS10,ARI = 25 +/- 8 min-1: Even IPS 15 was not able to reestablish VT at the values observed during ATC (VTIPS15,ARI = 680 +/- 312 ml). During ATC WOBadd was small in both postoperative and ARI patients (WOBadd,ATC,postop = 93 +/- 36 mJ/l, WOBadd,ATC,ARI = 116 +/- 72 mJ/l). In the postoperative patients, an inspiratory pressure support of 5 mbar was not sufficient to compensate WOBadd compared with ATC. However, IPS 10 and 15 mbar were able to compensate for WOBadd (WOBadd,ATC5.postop WOBadd,IPS5,postop = 189 +/- 77 mJ/l; WOBadd,IPS10,postop = 55 +/- 30 mJ/l; WOBadd,IPS15,postop = 21 +/- 11 mJ/l). In the ARI patients an IPS 5, 10 or 15 mbar was not sufficient to compensate for WOBadd (WOBadd,IPS 5,ARI = 1126 +/- 262 mJ/l; WOBadd,IPS 10,ARI 863 +/- 253 mJ/l; WOBadd,IPS15,ARI 763 +/- 298 mJ/l). Under ATC, WOBadd was only 15% of WOBadd under IPS of 15 mbar. All but two patients were successfully extubated after the investigation. These two patients were not extubated because they were dependent on an FIO2 > 0.5. Our results strongly indicate that ventilatory dependence in ARI patients may be caused by the ETT rather than by mechanical dysfunction of the lung. ATC is a very helpful mode to use in distinguishing between ventilatory failure caused by ETT and real ventilatory dependence.