Bedside measurement of cerebral blood flow (CBF) represents an important feature in monitoring of neurointensive care patients which is hard to establish. Therefore, we adopted a recently described thermo-dye-dilution-based approach for monitoring CBF in patients suffering from severe cerebral insults, that is, traumatic brain injury (TBI) or subarachnoid hemorrhage (SAH). Combined fiberoptic-thermistor catheters were placed in one jugular venous bulb and in the abdominal aorta of 16 patients. Following central venous injection of a 50-mL bolus of precooled indocyanine green (ICG) solution, CBF was determined as a function of the mean transit times of coldness and dye. In addition, measurements of CBF using stable xenon-enhanced computerized tomography (sXe-CT) were simultaneously performed in 10 patients. A total of 272 thermo-dye-dilution measurements yielded 196 valid results, with CBF ranging from 26.2 to 144.8 mL 100 g(-1) min(-1). Reproducibility was fairly good, with the standard deviation within sets of repeated measurements being 6.3 mL 100 g(-1) min(-1) and 9.4 as the mean coefficient of variation. Simultaneously obtained values with sXe-CT displayed a good correlation (r = 0.843, p < 0.01); however, the thermo-dye-dilution method consistently overestimated CBF. Data analysis using the Bland and Altman methodology revealed a large bias of 45.7 mL 100 g(-1) min(-1) with a +/- 2 SD range of 37 mL 100 g(-1) min(-1), indicating a rather poor agreement. The thermo-dye-dilution method proved a reasonably reproducible technique, enabling repeated long-term bedside measurements of CBF in neurointensive care patients with a minimum of time effort. However, a high failure rate was also noted, and consistent overestimation of perfusion was observed in comparison to sXe-CT measurements. Although the thermo-dye-dilution technique has been successfully validated in patients with normal neurovascular function, its applicability for bedside monitoring of CBF appears uncertain in patients suffering from severe TBI or SAH.