Assessment of intracranial venous blood flow after subarachnoid hemorrhage: a new approach to diagnose vasospasm with transcranial color-coded duplex sonography

Abstract

OBJECTIVETranscranial color-coded duplex sonography (TCCS) is a reliable tool that is used to assess vasospasm in the M1 segment of the middle cerebral artery (MCA) after subarachnoid hemorrhage (SAH). A distinct increase in blood flow velocity (BFV) is the principal criterion for vasospasm. The MCA/internal carotid artery (ICA) index (Lindegaard Index) is also widely used to distinguish between vasospasm and cerebral hyperperfusion. However, extracranial ultrasonography assessment of the neck vessels might be difficult in an intensive care unit. Therefore, the authors evaluated whether the relationship of intracranial arterial to venous BFV might indicate vasospasm with similar or even better accuracy.METHODSPatients who presented between 2008 and 2015 with aneurysmal SAH were prospectively enrolled in the study. Digital subtraction angiography (DSA) and TCCS were performed within 24 hours of each other to assess vasospasm 8–10 days after SAH. The following different TCCS parameters were analyzed to assess vasospasm in the MCA and were compared with the gold-standard DSA parameters: 1) mean time-averaged maximum BFV (Vmean) of the MCA, 2) peak systolic velocity (PSV) of the MCA, 3) the Lindegaard Index using Vmean as well as PSV, and 4) a new arteriovenous index (AVI) between the MCA and the basal vein of Rosenthal using Vmean and PSV. The best cutoff values for these parameters to distinguish vasospasm from normal perfusion or hyperperfusion were calculated using receiver operating characteristic curve analysis. Sensitivity, specificity, positive predictive value, and negative predictive value as well as the overall accuracy for each cutoff value were analyzed.RESULTSA total of 102 patients (mean age 52 ± 12 years) were evaluated. Bilateral MCA assessment by TCCS was successful in all patients. In 6 cases (3%), the BFV of the basal vein of Rosenthal could not be analyzed. The AVI could not be calculated in 50 of 204 cases (25%) because the insonation quality was very low in one of the ICAs. An AVI > 10 for Vmean and an AVI > 12 for systolic velocity provided the highest accuracies of 87% and 86%, respectively. Regarding the Lindegaard Index, the accuracy was highest using a threshold of > 3 for the mean BFV (84%) as well as systolic BFV (80%). BFVs in the MCA of ≥ 120 cm/sec (Vmean) and ≥ 200 cm/sec (PSV) predicted vasospasm with accuracies of 84% and 83%, respectively. A combined analysis of the MCA BFV and the AVI led to a slight increase in specificity (Vmean, 94%; PSV, 93%) and positive predictive value (Vmean, 88%; PSV 86%) without further improvement in accuracy (Vmean, 88%; PSV, 84%).CONCLUSIONSThe intracranial AVI is a reliable parameter that can be used to assess vasospasm after SAH. Its reliability for differentiating vasospasm and hyperperfusion is slightly higher than that for the established Lindegaard Index, and this method has the additional advantage of a remarkably lower failure rate.