[en] We used the end-tidal concentration of xenon-133 (air curve) to estimate the profile of its arterial concentration in calculating cerebral blood flow. We examined the effects of pulmonary disease and artificial ventilation on the air curve and the calculated cerebral blood flow. We studied the relation between arterial and end-tidal xenon activities in 19 subjects, of whom 15 had pulmonary dysfunction. The t 1/2 of the declining phases of the arterial and air curves were used to express their shapes. The mean +/- SD reference t 1/2 from 15 normal volunteers was 26.8 +/- 8.4 seconds. The mean +/- SD t 1/2 s of the air and arterial curves from the 15 patients with pulmonary dysfunction were 10.4 +/- 2.9 and 33.8 +/- 10.9 seconds. The degree of pulmonary dysfunction (expressed as the pulmonary shunt percentage) correlated with distortion of the air curve. Substituting the arterial for the air curve, mean calculated cerebral blood flow (as the initial slope index) increased from 40 to 61 for the 12 patients with chronic obstructive pulmonary disease. The degree of underestimation of cerebral blood flow using the air curve correlated with the pulmonary shunt percentage. Our work confirms the problems of estimating cerebral blood flow in subjects with pulmonary dysfunction