121. The influence of frailty in spine surgery populations is highly variable: an analysis of 57,027 patients from the NSQIP database undergoing surgery for degenerative cervical myelopathy or lumbar spondylolisthesis

BACKGROUND CONTEXT Frailty is an emerging concept in risk prediction for spine surgery, but how the influence of frailty affects different populations of patients with different index pathologies has not been established. PURPOSE To compare the effect of frailty in predicting 30-day outcomes after surgery in two common patient populations: degenerative cervical myelopathy (DCM) and degenerative spondylolisthesis (DS). STUDY DESIGN/SETTING Statistical analysis of prospectively-collected registry data. PATIENT SAMPLE Patients included in the National Safety and Quality Improvement (NSQIP) database 2010-2018 undergoing a surgical procedure with a diagnosis of DCM or DS. OUTCOME MEASURES (1) Major complication (pneumonia, deep vein thrombosis (DVT), pulmonary embolism, myocardial infarction, cardiac arrest, wound infection or dehiscence, stroke, and sepsis) within 30 days, and (2) mortality, reoperation, readmission within 30 days, length of stay and discharge destination. METHODS Patients with a diagnosis of DCM or DS undergoing surgery were selected from an assimilated database of the NSQIP registry in the years 2010-2018 inclusive using the corresponding ICD-9/-10 codes. The 5-point modified frailty index (MFI-5) was then calculated for all patients using a score of 0 (Not-Frail), 1 (Pre-Frail) 2 (Frail) or 3 or more (Severely Frail). Simple demographics, including age, type of approach, number of levels, co-morbidities, fusion or nonfusion surgery and frailty distribution were compared using descriptive statistics for parametric and nonparametric data. Univariate analysis was used to define the odds ratio (OR) or beta coefficient (BC, continuous variables) of major complication, mortality, readmission, reoperation, length of hospital stay and discharge destination stratified by frailty, and then compared between DCM and DS cohorts. A multivariable model adjusting for age, gender and frailty was then used to test any significant effect seen. RESULTS A total of 41,369 patients underwent surgery for DCM compared to 15,658 patients who underwent surgery for DS. DCM patients (mean age 56.6[56.5-56.7]) were younger than DS patients (mean age 62.5[62.31-69.70]; p <0.001). In the DCM group, 80% underwent anterior approach, 17% posterior and 3% combined. In the DS group, 27.7% underwent posterior decompression alone compared to 72.3% with instrumented fusion. The DCM group had a higher proportion of non-frail patients (45%) compared to the DS group (36.8%), and the distribution of frailty was significantly different between cohorts (p <0.001). OR of a major complication was greater in the DCM patients compared to DS patients, for all strata of frailty (2.40[2.06-2.80] vs 1.75[1.38-2.24] for pre-frail, 3.80[3.22-4.48] vs 2.02[1.50-2.69] for frail, and 11.63[9.44-14.33] vs 2.68[1.51-4.75] for severely frail patients; p <0.001 for each). Frailty demonstrated a larger effect on the mortality in DCM patients compared to DS patients (OR for pre-frail 4.89[2.72-8.79] vs 5.34[1.21-23.69], OR for severely frail 27.70[14.29-53.69] vs 9.15[0.83-101.20]). This was also true for length of hospital stay and discharge to non-home destination. The effect of frailty on reoperation or readmission was equivalent. On multivariable analysis, DCM patients demonstrated significantly greater OR of major complication (1.35[1.20-1.52]; p <0.001), mortality (3.02[1.89-4.83] but a shorter length of hospital stay (BC -0.33[-0.42 - -0.24]; p <0.001). CONCLUSIONS This study shows the influence of frailty amongst patients undergoing surgery for DCM and DS populations is not equivalent. In DCM populations, frailty has a greater effect size for metrics such as major complication or mortality compared to DS patients, but the opposite is true for length of hospital stay. Further work is required to define the key elements of frailty assessment in these discrete populations. FDA DEVICE/DRUG STATUS This abstract does not discuss or include any applicable devices or drugs. Frailty is an emerging concept in risk prediction for spine surgery, but how the influence of frailty affects different populations of patients with different index pathologies has not been established. To compare the effect of frailty in predicting 30-day outcomes after surgery in two common patient populations: degenerative cervical myelopathy (DCM) and degenerative spondylolisthesis (DS). Statistical analysis of prospectively-collected registry data. Patients included in the National Safety and Quality Improvement (NSQIP) database 2010-2018 undergoing a surgical procedure with a diagnosis of DCM or DS. (1) Major complication (pneumonia, deep vein thrombosis (DVT), pulmonary embolism, myocardial infarction, cardiac arrest, wound infection or dehiscence, stroke, and sepsis) within 30 days, and (2) mortality, reoperation, readmission within 30 days, length of stay and discharge destination. Patients with a diagnosis of DCM or DS undergoing surgery were selected from an assimilated database of the NSQIP registry in the years 2010-2018 inclusive using the corresponding ICD-9/-10 codes. The 5-point modified frailty index (MFI-5) was then calculated for all patients using a score of 0 (Not-Frail), 1 (Pre-Frail) 2 (Frail) or 3 or more (Severely Frail). Simple demographics, including age, type of approach, number of levels, co-morbidities, fusion or nonfusion surgery and frailty distribution were compared using descriptive statistics for parametric and nonparametric data. Univariate analysis was used to define the odds ratio (OR) or beta coefficient (BC, continuous variables) of major complication, mortality, readmission, reoperation, length of hospital stay and discharge destination stratified by frailty, and then compared between DCM and DS cohorts. A multivariable model adjusting for age, gender and frailty was then used to test any significant effect seen. A total of 41,369 patients underwent surgery for DCM compared to 15,658 patients who underwent surgery for DS. DCM patients (mean age 56.6[56.5-56.7]) were younger than DS patients (mean age 62.5[62.31-69.70]; p <0.001). In the DCM group, 80% underwent anterior approach, 17% posterior and 3% combined. In the DS group, 27.7% underwent posterior decompression alone compared to 72.3% with instrumented fusion. The DCM group had a higher proportion of non-frail patients (45%) compared to the DS group (36.8%), and the distribution of frailty was significantly different between cohorts (p <0.001). OR of a major complication was greater in the DCM patients compared to DS patients, for all strata of frailty (2.40[2.06-2.80] vs 1.75[1.38-2.24] for pre-frail, 3.80[3.22-4.48] vs 2.02[1.50-2.69] for frail, and 11.63[9.44-14.33] vs 2.68[1.51-4.75] for severely frail patients; p <0.001 for each). Frailty demonstrated a larger effect on the mortality in DCM patients compared to DS patients (OR for pre-frail 4.89[2.72-8.79] vs 5.34[1.21-23.69], OR for severely frail 27.70[14.29-53.69] vs 9.15[0.83-101.20]). This was also true for length of hospital stay and discharge to non-home destination. The effect of frailty on reoperation or readmission was equivalent. On multivariable analysis, DCM patients demonstrated significantly greater OR of major complication (1.35[1.20-1.52]; p <0.001), mortality (3.02[1.89-4.83] but a shorter length of hospital stay (BC -0.33[-0.42 - -0.24]; p <0.001). This study shows the influence of frailty amongst patients undergoing surgery for DCM and DS populations is not equivalent. In DCM populations, frailty has a greater effect size for metrics such as major complication or mortality compared to DS patients, but the opposite is true for length of hospital stay. Further work is required to define the key elements of frailty assessment in these discrete populations.