Clinically predictive baseline labs for post-operative outcomes of brain tumors using NSQIP database

Abstract

PurposeThis study was performed to assess the effect of baseline Preoperative Laboratory Values (PLV) on post-operative Brain Tumor Resection (BTR) outcomes in a large national registry. MethodsWe extracted data from the National Surgical Quality Improvement Program (NSQIP) database for BTR patients 2015–2019 (n = 3 0,951). Uni- and multivariate analyses were performed for PLV and key surgical outcomes. ResultsThe most significant PLV predictors of 30-day mortality after BTR included hypernatremia (odds ratio, OR 4.184, 95% CI, 2.384–7.343, p < 0.001), high serum creatinine (OR 2.244, 95% CI 1.502–3.352, p < 0.001), thrombocytopenia (OR 1.997, 95% CI 1.438, 2.772, p < 0.001), and leukocytosis (OR 1.635, 95% CI 1.264, 2.116, p < 0.001). The most significant predictors of Clavien IV complications were increased INR (OR 2.653, 95% CI 1.444, 4.875, p < 0.01), thrombocytopenia (OR 1.514, 95% CI 1.280, 1.792, p < 0.001), hypoalbuminemia (OR 1.480, 95% CI 1.274, 1.719, p < 0.001), and leukocytosis (OR 1.467, 95% CI 1.306, 1.647, p < 0.001). The most robust predictors of eLOS were increased INR (OR 1.941, 95% CI 1.231, 3.060, p < 0.01) and hypoalbuminemia (OR 1.993, 95% CI 1.823, 2.179, p < 0.001), and those for non-routine discharge included increased INR (OR 1.897, 95% CI 1.196, 3.008, p < 0.01) and hypernatremia (OR 1.565, 95% CI 1.217, 2.012, p < 0.001). ConclusionsSeveral PLV independently predicted worse outcomes in BTR patients. Baseline labs should be routinely used for the pre-operative risk stratification of these patients.