Preoperative and intraoperative risk factor assessment and model for predicting stroke risk in patients diagnosed...

Perioperative stroke is a devastating surgical complication associated with high mortality rates. The goals of our study were to identify the predictors of postoperative stroke and to explore the possible role of intraoperative hypotension (IOH).

Medical records of patients undergoing non-cardiovascular, non-neurological surgery at Ningbo No.2 Hospital (January 1st, 2012-December 31st, 2023) were retrospectively analyzed. Logistic regression was used to screen for independent risk factors. The nomogram is used for the visualization of the model.

This retrospective cohort study included 574,102 patients. Among 419,358 patients from 2012 to 2019, 848 (0.2 %) experienced postoperative stroke; cancer-related surgeries increased this incidence by 67 % (P < 0.001). In the cancer subgroup (multivariate analysis), significant predictors included age, history of stroke, atrial fibrillation, renal failure, elevated preoperative D-Dimer, and sustained IOH (>40 % decrease in mean arterial pressure [MAP] from baseline for >15 min) (all P < 0.05). Furthermore, the nomogram model developed based on the above variables demonstrated effective performance in predicting the occurrence of postoperative stroke in both the training set (January 1st, 2012-December 31st, 2019; n = 65,312) and the validation set (January 1st, 2020-December 31st, 2023; n = 28,326), with areas under the curve (AUCs) of 0.843 and 0.852, respectively.

This study confirms that cancer-related surgeries significantly increase postoperative stroke risk and demonstrates IOH's independent association in this subgroup. A validated nomogram provides an effective tool for prediction.