Assessing the utility of a clinical prediction score regarding 30-day morbidity and mortality following metastatic spinal surgery: the New England Spinal Metastasis Score (NESMS)

Background context: The New England Spinal Metastasis Score (NESMS) was recently proposed to help predict 1-year survival following surgery for spinal metastases. Its ability to predict short-term outcomes, including 30-day morbidity, mortality, and hospital length of stay, has not been evaluated.

Purpose: Assess the capacity of NESMS to predict 30-day morbidity and mortality, as well as hospital length of stay, following surgery for spinal metastases.

Study design: Validation study.

Patient sample: All patients who had undergone spinal surgery with a history of metastatic spinal disease within the National Surgical Quality Improvement Program (NSQIP; 2007-2013).

Outcome measure: Mortality, complications, failure to rescue, and length of stay.

Methods: Demographic, oncologic, laboratory, and surgical data were obtained from the NSQIP. All patients were assigned an NESMS score (0-3). The effect of the NESMS score on the outcomes of interest was assessed using multivariable logistic regression and negative binomial regression that controlled for confounders. Final model discrimination and calibration were assessed using the c-statistic and Hosmer-Lemeshow test, respectively. Internal validation was performed using a bootstrapping procedure.

Results: NSQIP data on 776 patients were included in this analysis. The 30-day mortality rate was 11% (N=87), and 51% of patients (N=395) sustained one or more complications. The final adjusted model demonstrated that the NESMS was a statistically significant predictor of 30-day mortality (p<.001), major systemic complications (p<.001), and failure to rescue (p=.03) following metastatic spinal surgery. Patients with an NESMS score of 3 had an 89% reduction in mortality (95% confidence interval [CI]: 0.04, 0.31), a 74% reduction in major systemic complications (95% CI: 0.11, 0.62), and an 88% reduction in failure to rescue (95% CI: 0.03, 0.47) as compared with those with a score of 0. The final model explained 71% of the variation in 30-day mortality. Findings were unchanged in the bootstrap analysis performed among 77,600 patient replicates.

Conclusion: This study demonstrates the clinical accuracy of the NESMS score for predicting short-term major morbidity and mortality following metastatic spinal surgery. The success of this score in an independent cohort of patients collected from centers across the United States indicates its potential for translation to clinical practice.