OBJECTIVE
Perioperative and/or postoperative cerebrovascular accidents (PCVAs) after intracranial tumor resection (ITR) are serious complications with devastating effects on quality of life and survival. Here, the authors retrospectively analyzed a prospectively maintained, multicenter surgical registry to design a risk model for PCVA after ITR to support efforts in neurosurgical personalized medicine to risk stratify patients and potentially mitigate poor outcomes.
METHODS
The National Surgical Quality Improvement Program database was queried for ITR cases (2015–2019, n = 30,951). Patients with and without PCVAs were compared on baseline demographics, preoperative clinical characteristics, and outcomes. Frailty (physiological reserve for surgery) was measured by the Revised Risk Analysis Index (RAI-rev). Logistic regression analysis was performed to identify independent associations between preoperative covariates and PCVA occurrence. The ITR-PCVA risk model was generated based on logit effect sizes and assessed in area under the receiver operating characteristic curve (AUROC) analysis.
RESULTS
The rate of PCVA was 1.7% (n = 532). Patients with PCVAs, on average, were older and frailer, and had increased rates of nonelective surgery, interhospital transfer status, diabetes, hypertension, unintentional weight loss, and elevated BUN. PCVA was associated with higher rates of postoperative reintubation, infection, thromboembolic events, prolonged length of stay, readmission, reoperation, nonhome discharge destination, and 30-day mortality (all p < 0.001). In multivariable analysis, predictors of PCVAs included RAI “frail” category (OR 1.7, 95% CI 1.2–2.4; p = 0.006), Black (vs White) race (OR 1.5, 95% CI 1.1–2.1; p = 0.009), nonelective surgery (OR 1.4, 95% CI 1.1–1.7; p = 0.003), diabetes mellitus (OR 1.5, 95% CI 1.1–1.9; p = 0.002), hypertension (OR 1.4, 95% CI 1.1–1.7; p = 0.006), and preoperative elevated blood urea nitrogen (OR 1.4, 95% CI 1.1–1.8; p = 0.014). The ITR-PCVA predictive model was proposed from the resultant multivariable analysis and performed with a modest C-statistic in AUROC analysis of 0.64 (95% CI 0.61–0.66). Multicollinearity diagnostics did not detect any correlation between RAI-rev parameters and other covariates (variance inflation factor = 1).
CONCLUSIONS
The current study proposes a novel preoperative risk model for PCVA in patients undergoing ITR. Patients with poor physiological reserve (measured by frailty), multiple comorbidities, abnormal preoperative laboratory values, and those admitted under high acuity were at highest risk. The ITR-PCVA risk model may support patient-centered counseling striving to respect goals of care and maximize quality of life. Future prospective studies are warranted to validate the ITR-PCVA risk model and evaluate its utility as a bedside clinical tool.