Steroid Use Associated With Increased Odds of 30-Day Mortality in Surgical Patients With Metastatic Spinal Tumors in the Setting of Disseminated Disease.

Abstract

Steroid administration is part of a standard treatment regimen in metastatic spinal cord compression, though the appropriate dose, duration, efficacy, and risks remain controversial. To analyze the risk of preoperative steroid use on 30-d mortality in surgical metastatic spinal tumors with dissemination disease using a large multicenter national database. Adult patients who underwent surgical treatment for metastatic spine tumors between 2005 and 2014 were identified in the American College of Surgeons National Surgical Quality Improvement Program database. Demographic, preoperative risk factors, operative information, and postoperative events were extracted. Multivariate logistical regression modeling was used to investigate the association with preoperative steroid use with the outcome of interest, 30-d mortality. Other independent risk factors associated with 30-d mortality were also identified. Five hundred fifty-two patients underwent surgical treatment of spinal metastases with disseminated cancer present at time of surgery. Independent risk factors of 30-d mortality included prolonged steroid use (odds ratio [OR] 2.48, 95% confidence interval [CI]: 1.22-5.04, P=.012), dependent functional status (OR 2.91, 95% CI: 1.68-5.04, P<.001), history of bleeding disorder (OR 2.80, 95% CI: 1.16-6.74, P=.021), history of smoking (OR 2.26, 95% CI: 1.11-4.61, P=.024), preoperative transfusions (OR 2.91, 95% CI: 1.02-8.29, P=.049), and preoperative infection/sepsis (OR 2.67, 95% CI: 1.18-6.08, P=.02). Our model demonstrates very strong predictive capabilities, with an area under the receiver operating characteristic curve of 0.7447. Steroid use is associated with a significant increased risk of 30-d mortality in surgical metastatic spine tumor patients with disseminated disease. These findings warrant further investigation in controlled experimental environments.