Drivers of Readmission and Reoperation After Surgery for Vertebral Column Metastases.

To determine those clinical, demographic, and operative factors that predict 30-day unplanned reoperation and readmission within a population of adults who underwent spinal metastasis surgery at a comprehensive cancer center. Adults who underwent spinal metastasis surgery at a comprehensive cancer center were analyzed. Data included baseline laboratory values, cancer history, demographics, operative characteristics and medical comorbidities. Medical comorbidities were quantified using the modified Charlson Comorbidity Index (CCI). Values associated with the outcomes of interest were then subjected to multivariable logistic regression to identify independent predictors of readmission and reoperation. A total of 345 cases were identified. Mean age was 59.4 ± 11.7 years, 56% were male, and the racial makeup was 64% white, 29% black, and 7.3% other. Forty-two patients (12.2%) had unplanned readmissions, most commonly for wound infection with dehiscence (14.2%), venous thromboembolism (14.2%), and bowel obstruction/complication (11.9%). Thirteen patients required reoperation (4%), most commonly for wound infection with dehiscence (39%) or local recurrence (23%). Multivariable analysis showed that the modified CCI (odds ratio [OR], 1.25; 95% confidence interval [CI] 1.03–1.52; P = 0.03) was an independent predictor of 30-day readmission. Independent predictors of 30-day unplanned reoperation were: black (vs. white) race (OR, 0.08; 95% CI, 0.01–0.41; P < 0.01), length of stay (OR, 1.05 per day; 95% CI, 1.00–1.09; P = 0.04), and CCI (OR, 1.72 per point; 95% CI, 1.29–2.28; P < 0.01). Increasing medical comorbidities is independently predictive of both 30-day unplanned readmission and reoperation after spinal metastasis surgery. Unplanned reoperation is also positively predicted by a longer index admission. Neither tumor pathology nor age predicted outcome, suggesting that poor wound-healing factors and increased surgical morbidity may best predict these adverse outcomes. [ABSTRACT FROM AUTHOR]