Trans-cranial Doppler predicts early neurologic deterioration in anterior circulation ischemic stroke after successful endovascular treatment

Abstract

Abstract. Background. Early neurologic deterioration (END) may occur in patients with anterior circulation ischemic stroke (ACIS) after receiving endovascular treatment (EVT). Hemodynamic insufficiency, re-occlusion, and post-re-canalization hyper-perfusion are likely to play a critical role in END. We hypothesized that hemodynamic changes can predict END in patients with ACIS post-successful EVT using trans-cranial Doppler (TCD). Methods. We utilized a prospectively maintained database of ACIS patients treated with EVT between September 2016 and June 2018 in the Xuanwu Hospital, Capital Medical University. TCD parameters including peak systolic velocity (PSV), bilateral mean flow velocity (MFV), and pulse index (PI) were determined via the middle cerebral arteries within 72 h post-EVT. A logistic regression model was applied to detect independent predictors for END. Results. Totally, 112 EVT patients were included in this study and 80/112 patients experienced successful re-canalization with <50% residual stenosis, while 17/80 (21.3%) patients suffered END, for which vasogenic cerebral edema (11/17) was considered as a leading role and followed by symptomatic intra-cranial hemorrhage (4/17) and ischemia progression (2/17). For the 80 patients, the PSV (median: 127 cm/s vs. 116 cm/s, P = 0.039), the ratio of ipsilateral-MFV/contra-lateral-MFV (iMFV/cMFV) (median: 1.29 vs. 1.02, P = 0.036) and iMFV/mean blood pressure (MBP) (median: 0.97 vs. 0.79, P = 0.008) in END patients were higher than those of non-END. Using the receiver-operating characteristic curve to obtain cut-off values for PSV, PI, iMFV/cMFV, and iMFV/MBP for END, we found that PI ≥0.85 (odds ratio: 11.03, 95% confidence interval: 1.92–63.46, P = 0.007) and iMFV/MBP ≥0.84 (odds ratio: 9.20, 95% confidence interval: 2.07–40.84, P = 0.004) were independent predictors of END in a multivariate logistic regression model, with a sensitivity of 82.4% and 76.5% and a specificity of 42.9% and 66.7%, respectively, and had the positive predictive values of 29.0% and 38.2%, and negative predictive values of 90.0% and 91.3%, with an area under the receiver-operating characteristic curve of 0.57 and 0.71, respectively. Conclusion. TCD examination of EVT patients may be used as a real-time tool to detect END predictors, such as the higher PI and iMFV/MBP, allowing for better post-thrombectomy management in ACIS patients.