Abstract
OBJECTIVE: Periprocedural thromboembolic events are a serious complication associated with coil embolization of unruptured intracranial aneurysms. However, no established clinical rule for predicting thromboembolic events exists. This study aimed to clarify the significance of adding preoperative clopidogrel response value to clinical factors when predicting the occurrence of thromboembolic events during/after coil embolization and to develop a nomogram for thromboembolic event prediction. METHODS: In this prospective, single-center, cohort study, we included 345 patients undergoing elective coil embolization for unruptured intracranial aneurysm. Thromboembolic event was defined as the occurrence of intra-procedural thrombus formation and postprocedural symptomatic cerebral infarction within 7 days. We evaluated preoperative clopidogrel response and patients' clinical information. We developed a patient-clinical-information model for thromboembolic event using multivariate analysis and compared its efficiency with that of patient-clinical-information plus preoperative clopidogrel response model. The predictive performances of the two models were assessed using area under the receiver-operating characteristic curve (AUC-ROC) with bootstrap method and compared using net reclassification improvement (NRI) and integrated discrimination improvement (IDI). RESULTS: Twenty-eight patients experienced thromboembolic events. The clinical model included age, aneurysm location, aneurysm dome and neck size, and treatment technique. AUC-ROC for the clinical model improved from 0.707 to 0.779 after adding the clopidogrel response value. Significant intergroup differences were noted in NRI (0.617, 95% CI: 0.247-0.987, p < .001) and IDI (0.068, 95% CI: 0.021-0.116, p = .005). CONCLUSIONS: Evaluation of preoperative clopidogrel response in addition to clinical variables improves the prediction accuracy of thromboembolic event occurrence during/after coil embolization of unruptured intracranial aneurysm.