Moderators, Mediators, and Prognostic Indicators of Treatment With Hip Arthroscopy or Physical Therapy for Femoroacetabular Impingement Syndrome: Secondary Analyses From the Australian FASHIoN Trial

Background: Although randomized controlled trials comparing hip arthroscopy with physical therapy for the treatment of femoroacetabular impingement (FAI) syndrome have emerged, no studies have investigated potential moderators or mediators of change in hip-related quality of life. Purpose: To explore potential moderators, mediators, and prognostic indicators of the effect of hip arthroscopy and physical therapy on change in 33-item international Hip Outcome Tool (iHOT-33) score for FAI syndrome. Study Design: Cohort study; Level of evidence, 2. Methods: Overall, 99 participants were recruited from the clinics of orthopaedic surgeons and randomly allocated to treatment with hip arthroscopy or physical therapy. Change in iHOT-33 score from baseline to 12 months was the dependent outcome for analyses of moderators, mediators, and prognostic indicators. Variables investigated as potential moderators/prognostic indicators were demographic variables, symptom duration, alpha angle, lateral center-edge angle (LCEA), Hip Osteoarthritis MRI Scoring System (HOAMS) for selected magnetic resonance imaging (MRI) features, and delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) score. Potential mediators investigated were change in chosen bony morphology measures, HOAMS, and dGEMRIC score from baseline to 12 months. For hip arthroscopy, intraoperative procedures performed (femoral ostectomy +/- acetabular ostectomy +/- labral repair +/- ligamentum teres debridement) and quality of surgery graded by a blinded surgical review panel were investigated for potential association with iHOT-33 change. For physical therapy, fidelity to the physical therapy program was investigated for potential association with iHOT-33 change. Results: A total of 81 participants were included in the final moderator/prognostic indicator analysis and 85 participants in the final mediator analysis after exclusion of those with missing data. No significant moderators or mediators of change in iHOT-33 score from baseline to 12 months were identified. Patients with smaller baseline LCEA (beta = -0.82; P = .034), access to private health care (beta = 12.91; P = .013), and worse baseline iHOT-33 score (beta = -0.48; P < .001) had greater iHOT-33 improvement from baseline to 12 months, irrespective of treatment allocation, and thus were prognostic indicators of treatment response. Unsatisfactory treatment fidelity was associated with worse treatment response (beta = -24.27; P = .013) for physical therapy. The quality of surgery and procedures performed were not associated with iHOT-33 change for hip arthroscopy (P = .460-.665 and P = .096-.824, respectively). Conclusion: No moderators or mediators of change in hip-related quality of life were identified for treatment of FAI syndrome with hip arthroscopy or physical therapy in these exploratory analyses. Patients who accessed the Australian private health care system, had smaller LCEAs, and had worse baseline iHOT-33 scores, experienced greater iHOT-33 improvement, irrespective of treatment allocation.