Infection risk for primary and revision instrumented lumbar spine fusion in the Medicare population

Abstract

Object: This retrospective analysis of Medicare administrative data was performed to evaluate the risk of infection following instrumented lumbar fusion over a 10-year follow-up period in the Medicare population. Although infection can be a devastating complication, due to its rarity it is difficult to characterize infection risk except in large patient populations.

Methods: Using ICD-9-CM and CPT4 procedure codes, the Medicare 5% analytical research files for inpatient, outpatient, and physician carrier claims were checked to identify patients who were treated between 1997 and 2009 with lumbar spine fusion in which cages or posterior instrumentation were used. Patients younger than 65 years old were excluded. Patients were followed continuously by using the matching denominator file until they withdrew from Medicare or died. The authors identified 15,069 patients with primary fusion procedures and 605 with revision of instrumented lumbar fusion. Infections were identified by the related ICD-9 codes (998.59 or 996.67) after fusion. Kaplan-Meier survival analysis and Cox regression were performed to determine adjusted infection risk for each type of spine procedure (primary vs revision) and surgical approach (anterior, posterior, combined anteroposterior), accounting for patient (for example, age, sex, comorbidities/Charlson Comorbidity Index [CCI], and state buy-in) and hospital (census region) characteristics.

Results: At 10 years, the overall infection incidence, including superficial and deep infections, was 8.5% in primary procedures and 12.2% in revisions. Among the factors considered, infection risk within 10 years was most influenced by comorbidities: for a CCI of 5 versus 0, the adjusted hazard ratio (AHR) was 2.48 (95% CI 1.93-3.19, p < 0.001); for ≥ 9 versus 2-3 fused vertebrae, the AHR was 2.39 (95% CI 1.20-4.76, p < 0.001); for revision versus primary fusion procedures, the AHR was 1.66 (95% CI 1.28-2.15, p < 0.001). Other significant predictors of 10-year infection risk included diagnosis of obesity (p < 0.001); state buy-in--a proxy for socioeconomic status (p = 0.02); age (p = 0.003); surgical approach (p = 0.03); census region (p = 0.02); and the year of the index procedure (p = 0.03).

Conclusions: Patient comorbidities were the greatest predictor of infection risk for the Medicare population. The high incidence of infection following instrumented fusion warrants increased focus on infection risk mitigation, especially for patients with comorbid conditions.