Identifying Patients at Elevated Risk of Stroke After.

BACKGROUND: Transcarotid artery revascularization (TCAR) has assumed a significant role in the surgical treatment of extracranial carotid occlusive disease in current practice. Yet, TCAR specific factors which predispose to periprocedural stroke remain poorly characterized. The purpose of this analysis was to identify factors which may inherently predispose to adverse outcomes in patients undergoing TCAR. METHODS: We used Vascular Quality Initiative data to examine factors associated with in-hospital stroke after TCAR for patients treated between 2016-2024. We used two complementary machine learning methods to identify factors associated with stroke after TCAR. First, we used classification and regression trees (CART) to identify factors associated with stroke and then grouped patients according to these factors. Second, we used a random forest machine learning technique to identify the top features associated with stroke. Finally, we took the CART-defined subgroups and used logistic regression to account for confounding by the top variables from the random forest. RESULTS: We studied 61,111 patients who underwent TCAR (mean age: 73.4±8.9 years, 37.4% women, 42.1% symptomatic). The overall in-hospital stroke risk was 1.33% after TCAR. The CART analysis found that most important factors associated with in-hospital stroke were symptom status, non-compliance with taking a p2y12 inhibitor, and the use of two or more stents. This created four patient groups: (G1) asymptomatic and p2y12 inhibitor compliance, (G2) asymptomatic and p2y12 inhibitor not taking a preoperative P2Y12 inhibitor, (G3) symptomatic and with one stent, and (G4) symptomatic and with 2 or more stents. These groups had in-hospital stroke risk of 0.74%, 1.60%, 1.85%, and 3.96%, respectively. After adjustment for confounding, the groups had a stepwise increase in the likelihood of stroke, G1: reference, G2: adjusted odds ratio (aOR): 1.77 (95% CI: 1.27-2.47), G3: aOR: 2.36 (95% CI: 2.00-2.77), and G4: aOR: 5.29 (95% CI: 4.07-6.87). CONCLUSIONS: There appear to be important modifiable factors that are associated with increased periprocedural stroke risk among patients undergoing TCAR. Specifically, unreliable p2y12 inhibition and multi-stent procedures were strong predictors of in hospital stroke. These findings suggest that asymptomatic patients who cannot satisfactorily comply with preoperative medical therapy or symptomatic patients that cannot be treated with a single stent may wish to consider an alternative revascularization strategy.