Frailty Is a Better Predictor than Age of Mortality and Perioperative Complications after Surgery for Degenerative Cervical Myelopathy: An Analysis of 41,369 Patients from the NSQIP Database 2010-2018

Abstract

Background: The ability of frailty compared to age alone to predict adverse events in the surgical management of Degenerative Cervical Myelopathy (DCM) has not been defined in the literature.

Methods: 41,369 patients with a diagnosis of DCM undergoing surgery were collected from the National Surgical Quality Improvement Program (NSQIP) Database 2010-2018. Univariate analysis for each measure of frailty (modified frailty index 11- and 5-point; MFI-11, MFI-5), modified Charlson Co-morbidity index and ASA grade) were calculated for the following outcomes: mortality, major complication, unplanned reoperation, unplanned readmission, length of hospital stay, and discharge to a non-home destination. Multivariable modeling of age and frailty with a base model was performed to define the discriminative ability of each measure.

Results: Age and frailty have a significant effect on all outcomes, but the MFI-5 has the largest effect size. Increasing frailty correlated significantly with the risk of perioperative adverse events, longer hospital stay, and risk of a non-home discharge destination. Multivariable modeling incorporating MFI-5 with age and the base model had a robust predictive value (0.85). MFI-5 had a high categorical assessment correlation with a MFI-11 of 0.988 (p < 0.001).

Conclusions and relevance: Measures of frailty have a greater effect size and a higher discriminative value to predict adverse events than age alone. MFI-5 categorical assessment is essentially equivalent to the MFI-11 score for DCM patients. A multivariable model using MFI-5 provides an accurate predictive tool that has important clinical applications.

Keywords: age; complications; degenerative cervical myelopathy; frailty; mortality.

Figure 1

Predictive margins of the final multivariable model for each frailty level stratified per decade of age, for each outcome studied. Decades of age were removed if co-linearity was present. 95% CI, 95% Confidence Interval.