Continuous assessment of cerebral autoregulation in subarachnoid hemorrhage

Cerebral vasospasm remains a leading cause of morbidity and mortality after subarachnoid hemorrhage (SAH). Cerebral ischemia may ensue when autoregulation fails to compensate for spasm. We examined how autoregulation is affected by vasospasm by using transcranial Doppler. The moving correlation coefficient between slow changes of arterial blood pressure and mean or systolic flow velocity (FV), termed "Mx" and "Sx," respectively, was used to characterize cerebral autoregulation. Vasospasm was declared when the mean FV increased to more than 120 cm/s and the Lindegaard ratio was more than 3. This occurred in 15 of 32 SAH patients. On the basis of the bilateral transcranial Doppler recordings of the middle cerebral artery in vasospastic patients, Mx and Sx were calculated for baseline and vasospasm. Mx increased during vasospasm (0.46 +/- 0.32; mean +/- SD) and was significantly higher (P = 0.021) than at baseline (0.21 +/- 0.24). Sx was also increased (0.22 +/- 0.26 vs 0.05 +/- 0.21 at baseline; P = 0.03). Mx correlated with mean FV (r = 0.577; P = 0.025) and the Lindegaard ratio (r = 0.672; P < 0.006). Mx (P = 0.006) and Sx (P = 0.044) were higher on the vasospastic side (Mx, 0.44 +/- 0.27; Sx, 0.24 +/- 0.23) when compared with the contralateral side (Mx, 0.34 +/- 0.29; Sx, 0.16 +/- 0.25). The increased Mx and Sx during cerebral vasospasm demonstrate impaired cerebral autoregulation. Mx and Sx provide additional information on changes in autoregulation in SAH patients.

The moving correlation coefficients between slow changes of arterial blood pressure and mean or systolic flow velocity, termed "Mx" and "Sx," respectively, characterize cerebral autoregulation but have not been applied to subarachnoid hemorrhage. A study in 15 patients revealed that Mx and Sx were significantly increased, indicating impaired autoregulation during vasospasm as compared with baseline, as well as on the side of vasospasm in comparison with the contralateral side.