Flexible fiberoptic bronchoscopy of intubated patients can be performed in general or local anesthesia (LA). Up to now, no results have been published on the effects of LA for bronchoscopy in ventilated patients. We studied the hemodynamic changes caused by bronchoscopy under LA in mechanically ventilated patients and the effect of LA on the endoscopic decline in arterial pO2. Differences between the widely used agents lidocaine and oxybuprocaine hydrochloride were also studied. We found a decline in paO2 after the administration of LA and further investigated the influence of bronchial lavage on paO2. METHOD. A total of 70 ventilator patients, excluding patients with atelectasis, massive mucous-plug retention, and those under muscle relaxants, were examined in a surgical intensive care unit. In 40 long-term ventilator patients bronchoscopy was performed with either oxybuprocaine hydrochloride 1% (Novesin) (group 1; n = 20) or Lidocaine 1% (Xylocaine) (group 2; n = 20) (2-3 ml LA in repeated doses into the trachea and main bronchi; total amount 10 ml). We looked for hemodynamic changes and effects of LA on the bronchoscopic decline in paO2. In 15 long-term ventilator patients (group 3), LA was applied without bronchoscopy to investigate the duration of the LA-caused decline in paO2. In 15 intubated patients (group 4), the influence of intratracheal administration of 10 ml normal saline was examined. Patient data are shown in Table 2. Measurements were performed in groups 1 and 2 before and after LA, immediately after bronchoscopy and 15, 30, and 60 min after bronchoscopy and in groups 3 and 4 before and 5, 15, 30, and 60 min after LA. RESULTS. There was no effect on cardiocirculatory function during bronchoscopy in LA, but we found a decrease in paO2 after administration of LA in all patients (median in group 1 from 100 to 78 mmHg in group 2 from 104 to 86 mmHg). The subsequent bronchoscopy caused only a small, nonsignificant further decline in paO2. The administration of LA without bronchoscopy (group 3) was followed by a fall in paO2 from 86 +/- 12.5 to 69 +/- 11.7 mmHg (mean +/- SD) with oxybuprocaine hydrochloride and from 87 +/- 12.4 to 72 +/- 8.7 mmHg with lidocaine. Even after 30 min the paO2 had not returned to the initial value. The intratracheal application of 10 ml 0.9% NaCl caused a decline in paO2 from 101 +/- 20 to 78 +/- 12 mmHg (mean +/- SD), which also persisted for more than 30 min. CONCLUSIONS. The study shows that in ventilator patients undergoing fiberoptic bronchoscopy in LA, the administration of the LA is an essential factor in the decline in paO2 associated with bronchoscopy. A similar fall in paO2 is observed by intratracheal