To determine the influence of baseline paCO2 on the results of apnea testing in the diagnosis of brain death, we performed an open prospective study on 36 patients fulfilling all other criteria for the diagnosis of brain death according to the criteria proposed by the Advisory Board of the German Federal Chamber of Physicians. For testing of apnea, patients underwent hypoventilation with 100% oxygen supply until a baseline paCO2 of 40 torr (5.3 kPa, n = 24, group 1) or 60 torr (8.0 kPa, n = 12, group 2) was reached. Then, patients were disconnected from the ventilator and apneic oxygenation with insufflation of 61 O2/min into the tracheal cannula was performed for five minutes. Arterial blood gas samples were obtained every minute during the testing period. In parallel, patients were observed for signs of spontaneous breathing. All patients remained apneic during the five minute test period. No relevant hypoxia (paO2 < 80 torr [10.6 kPa]) was observed in either group. In group 1, a mean baseline paCO2 of 45 torr (6.0 kPa) was registered, mean end-paCO2 was 75 torr (10.0 kPa). In group 2, paCO2 values were 66 torr (8.8 kPa) and 90 torr (12 kPa), respectively. Baseline pH in group 1 (7.32) decreased to 7.18 at the end of testing and from 7.23 to 7.13 in group 2. Patients in group 2 were in possible danger of developing a CO2-induced narcosis mimicking apnea. Secondary organ damage due to severe respiratory acidosis could not be excluded in the patients of group 2. As no complications were observed in group 1 and apnea was evident in all these patients, we consider a baseline paCO2 of 40 torr (5.3 kPa) sufficient to establish apnea after five minutes of apneic oxygenation if an increase of baseline paCO2 of at least 20 mmHg is documented by arterial blood gas sampling. A higher baseline paCO2 may endanger patients without yielding more specific testing results.