Risk factors for in-hospital mortality after spine surgery: a matched case-control study using a multicenter database

Abstract

BACKGROUND/CONTEXTIt is yet unclear what preoperative and intraoperative factors affect mortality after spine surgery. PURPOSETo identify the preoperative and intraoperative risk factors for in-hospital mortality after spine surgery using a matched case-control study based on a multicenter database. STUDY DESIGN/SETTINGA retrospective matched case-control study based on a registry of prospectively collected multicenter data. PATIENT SAMPLEWe identified 25 patients who died in the hospital (the mortality group) from the 26,604 patients in the database who underwent spine surgery at our 27 affiliated institutions between 2012 and 2018. An age-, sex-, spinal disease-, and surgical procedure-matched control group of patients (n=100, 4:1 ratio with the mortality group) was selected from the same database. OUTCOME MEASURESData relating to comorbidities, preoperative blood tests, operative factors, and perioperative complications. METHODSWe retrospectively reviewed all the medical records of each patient in the two groups to nullify the effects of overt risk factors such as age, sex, diseases, and surgical procedures. Risk factors for in-hospital mortality were initially evaluated by univariate analysis. Then, multivariate logistic regression models were generated to analyze independent risk factors for in-hospital mortality. RESULTSThe overall in-hospital mortality rate was 0.09% (25/26,604). Mortality was lowest in patients with degenerative cervical (0.04%, 2/5,027) or lumbar disease (0.03%, 5/15,630). In contrast, mortality was highest in patients with dialysis-related spondyloarthropathy (3.0%, 3/99), patients with infectious spondylodiscitis (1.5%, 6/401), and patients with metastatic spinal tumors (0.9%, 3/334). Multivariate logistic regression analysis revealed that massive intraoperative hemorrhage (>2 L) (odds ratio [OR], 28.2; 95% confidence interval [CI], 2.27–349), preoperative renal comorbidity (OR, 4.33; 95% CI, 1.38–13.6), and elevated preoperative aspartate aminotransferase levels (OR, 1.51 per 10 units; 95% CI, 1.04–2.20) were risk factors. CONCLUSIONSSpine surgery for patients with dialysis-dependency, infectious diseases or metastatic tumors had much more potential of in-hospital mortality compared with those for patients with degenerative diseases. Massive intraoperative hemorrhage and preoperative renal and liver comorbidities were identified as risk factors for in-hospital mortality in patients who underwent spine surgery.