Predictors of stroke after carotid endarterectomy

Abstract

To investigate factors leading to the occurrence of stroke after carotid enderaterectomy. N=206 patients were subject to a longitudinal study (median observation time 108 months, max 141 month, min 36 months). During the observation 33 patients had post-surgery stroke, and 37 died, including 21 having stroke before death. The occurrence of post-surgery stroke was modeled using Cox survival model, both in the traditional form (W. N. Venables and B. D. Ripley. Modern Applied Statistics with S. Springer, 2010) and the Bayesian one (Zaixiang Tanget al. BMC Bioinformatics, 20(1):94, 2019). The model used the following parameters: age, gender, employment, marital status, education, alcohol use, smoking, presence of angina and diabetes, initial diagnosis, taking of statins and aspirin, patient's compliance, regular visits to neurologists and cardiologists, Barthel, FIM and MMSE scores, anxiety and depression, and cholesterol ratio. The most important predictors are shown in Table 1. Both models chose age and Barthel score as the most predictive parameters. Traditional model added MMSE and retirement as predictors. Presumably the work-related stress makes the prognosis better for retired patients. This finding underscores the importance of non-medical factors in post-surgery prognosis. The study confirms the value of Barthel index for the prognosis: the more independent is the patient in the daily life, the better is the expected outcome.