BACKGROUND: Cerebral vasoreactivity is an index of the autoregulation of cerebral perfusion, and can be measured by functional imaging such as Xe CT, SPECT, PET in reponse to a hypercapneic stimulus. Recently, a transcranial Doppler (TCD) study helped in the evaluation of cerebral CO2 vasoreactivity by a breath holding method and CO2 inhalation. However, the breath holding method had some constrictions in reliability and the CO2 inhalation method is not as feasible to apply to routine study. In order to measure cerebral CO2 vasoreactivity in a routine TCD study conveniently and reliably, we devised a method of rebreathing into a closed volume of a reservoir bag as a hypercapneic stimulus, and applied it to normal volunteers and patients with ischemic stroke. METHODS: We performed a cerebral CO2 vasoreactivity study by the rebreathing method in 22 normal volunteers and 61 ischemic stroke patients. As a hypercapneic stimulus, we applied a fitted mask connected with a closed reservoir bag for about 90 seconds, and the mean flow velocity (MFV) and pulsatility index (PI) were evaluated at the proximal middle cerebral arteries of 50-55 mm depth, before and after the hypercapneic stimulus. Cerebral vasoreactivities as percent increments of MFV in MCA were compared between the two groups. RESULTS: MFV of both the normal control and patient groups increased continuously during the hypercapneic stimulus, but the rate of the increments of MFV markedly diminished after about 60 seconds of rebreathing. The mean cerebral vasoreactivity of 61 MCA's in the ischemic stroke patient group (47.0 +/- 22.0%) was significantly lower than that of 22 MCA's in the normal control group (60.2 +/- 16.0%). But, the percent decrement of PI showed no significant difference between both groups. CONCLUSIONS: We suppose that the rebreathing method is a relatively reliable and convenient technique as a hypercapneic stimulus in determining cerebral CO2 vasoreactivity by TCD. Cerebral vasoreactivity measured by the rebreathing method is significantly lower in the ischemic stroke patient group than in the normal control group, which may reflect that the atherosclerosis itself could cause reduced distensibility of the small arteriole as well as the structural narrowing of the cerebral arteries.