Intra-arterial (i.a.) and transcutaneous (t.c.) blood gas monitoring were compared with in vitro blood gas analysis (abg) during apnoea testing for the determination of brain death in a prospective observational study. All three methods were used simultaneously in 19 patients in whom brain death was suspected. Brain death was confirmed in each case adhering to the recommendations of the Scientific Advisory Board of the German Federal Chamber of Physicians which demand a PCO2 of at least 60 mmHg. In vitro parameters ranged from 23.2 to 80.4 mmHg (PCO2), 52.7 to 509.9 mmHg (PO2), and 7.072 to 7.591 (pH). The intra-individual correlations between both monitoring methods (rPCO2=0.958, rPO2=0.859) and between each of them and abg (r>0.960) were high. Absolute deviations from abg for the corrected as well as uncorrected measurements were similar for both methods, except with regard to group bias where an advantage for the i.a. values emerged. Since many of the i.a. measurements failed and the disposable i.a. probes cost much more than the t.c. electrodes, the i.a. technique at present holds no advantage over t.c. measurements in testing for apnoea in suspected brain death except where simultaneous monitoring of pH and temperature are desired.