Risk Factors for Lumbar Spinal Fusion Reoperation: An Unusual Suspect

Commentary Severe, symptomatic degenerative disorders of the hip and spine are prevalent in the elderly, and disorders at both sites will occur in approximately 11% to 23% of patients1. The biomechanical relationships between the hip and spine are contingent upon the alignment and mobility of the pelvis, and current evidence recognizes the impact of sagittal spinopelvic balance on both hip and spinal degenerative disorders2. In patients with combined, advanced degenerative diseases of the hip and spine, preoperative assessment of the standing and sitting sagittal spinopelvic alignment parameters, particularly the pelvic tilt, is considered essential in achieving successful mobility and stability following total hip arthroplasty (THA). Although the impact of hip osteoarthritis (OA) progression on existing spinal degenerative disease has been reported3, the effects of hip OA on lumbar spinal surgery outcomes have not been clearly established. Lumbar spinal fusion for degenerative disease has become more prevalent. In 2012, a 10-year review of the Healthcare Cost and Utilization Project Nationwide Inpatient Sample found that the prevalence of lumbar spinal fusion in the U.S. had increased 2.4-fold (by 137%). This increase was higher than that for hip arthroplasty (49.1%) and that for knee arthroplasty (126.8%) in the same time period4. A 2022 study by Mills et al. demonstrated that, among patients with advanced hip OA who underwent lumbar spinal fusion, those who had undergone THA prior to their lumbar procedure demonstrated superior spine-related outcomes when compared with those who had not undergone THA. However, the extent to which a delay in treating hip OA could culminate in a lumbar spinal fusion reoperation was not fully explored5. This issue was examined by Mills et al. in the present study by reviewing deidentified 5% Medicare Part B claims data for patients who underwent primary elective lumbar fusion within a 15-year period. Large database studies have certainly become more prevalent in the orthopaedic literature; however, many of these investigations constitute large retrospective reviews that are incapable of providing meaningful data as a result of either misalignment with the question(s) being asked or the limitations of the database being utilized6. The question posed by Mills et al. regarding the potential association between hip OA and the risk of spinal reoperation following lumbar spinal fusion was appropriate for the data elements queried, and their detailed analysis generated strong and interesting conclusions. Among the patients undergoing lumbar spinal fusion who were included in the study, the severe hip OA group was skillfully determined by identifying those who underwent elective THA within 1 year after their index spinal procedure. This group was then compared with a control group of patients who did not undergo THA during the study period. Many of the major risk factors for lumbar spinal fusion reoperation (e.g., age, diabetes, obesity, sex, smoking, osteoporosis, the number of levels fused, surgical technique, and surgical approach) were documented and were examined with use of a multivariable regression analysis. The literature indicates that the incidence of reoperation after lumbar spinal fusion is small within the first several years (6.2% at 1 year and 8.2% at 3 years) but will progress over time7. Accordingly, Mills et al. assessed the outcomes following lumbar spinal fusion in the hip OA and control groups longitudinally at 1, 3, and 5 years. The results of the study by Mills et al. are interesting and quite revealing. At both 3 and 5 years of follow-up, the incidence of lumbar spinal fusion reoperation in the hip OA group (22.1% and 27.8%, respectively) was greater than that in the control group (13.8% and 15.7%, respectively), which indicates an association between hip OA and lumbar spinal reoperation. As previously noted, hip surgeons are acutely aware of the impact that spinal pathology has on THA; however, prior to this study, the reciprocal relationship has not been fully appreciated. The study is not without inherent limitations. Like all medical payer database studies, its validity is dependent on specific patient characteristics and coding accuracy. The authors could certainly infer the presence of advanced hip and spine degenerative diseases, but the precise extent and/or nature of these coexisting conditions could not be documented. Likewise, the preoperative and postoperative spinopelvic parameters that are required to understand the relationship between pathologies of the spine and hip could not be assessed. Finally, although failure of the lumbar spine to fuse is a common cause for reoperation, many of the relevant procedure-specific variables (e.g., the type of interbody surgical approach, implant, or bone graft material) could not accurately be evaluated. Properly performed large database studies can excel at detecting unsuspected adverse outcomes and their risk factors. However, these investigations should not be considered conclusive, but rather serve as an impetus for better controlled, prospective, randomized studies that can elucidate causative relationships between pathologies. Although I am quite fascinated by the findings of Mills et al., I am also cautiously reminded of Sir Winston Churchill, who so astutely said "Now this is not the end. It is not even the beginning of the end. But it is, perhaps, the end of the beginning."8 This sentiment highlights the importance of further research on this topic, in recognition of both its complexity and the need for a substantial time-effort commitment.