Association of thigh muscle status and longitudinal changes assessed by MRI with subsequent total hip arthroplasty: insights from the osteoarthritis initiative

Abstract Background Evidence regarding the association of longitudinal thigh muscle MRI markers with subsequent total hip arthroplasty (THA) is limited. Such markers may help characterize risk among individuals with or at risk of hip osteoarthritis. Methods A longitudinal year-2 landmark analysis was conducted using the Osteoarthritis Initiative cohort. Hips that underwent THA after the year-2 landmark were matched to eligible control hips at a 1:4 ratio. Thigh MRI cross-sectional areas (CSAs) were automatically quantified using a validated U-Net segmentation network. The prespecified primary exposure was the 2-year change in total thigh muscle CSA. Individual muscle regions, adipose-tissue measures, baseline measurements, and alternative processing methods were evaluated as secondary or exploratory exposures. Cox proportional hazards models reported hazard ratios (HRs) per 1-standard-deviation increase. Secondary p -values were adjusted using the Benjamini–Hochberg false discovery rate procedure. This study evaluated associations and did not develop or validate a clinical prediction model. Results A total of 495 hips (99 THA cases and 396 matched controls) were included. In the primary analysis, a greater 2-year change in total thigh muscle CSA was associated with a lower incidence of subsequent THA (HR per 1-SD increase, 0.75; 95% CI 0.63–0.90; p  = 0.002). In secondary analyses, the 2-year changes in quadriceps, hamstrings, and adductors CSA showed nominal inverse associations with subsequent THA; however, none remained statistically significant after FDR correction (all q  ≥ 0.05). No baseline MRI measure was significantly associated with subsequent THA. Conclusion A greater 2-year change in total thigh muscle CSA, corresponding to greater gain or less loss, was associated with a lower incidence of subsequent THA. Conversely, lower change values or greater CSA loss were associated with higher subsequent THA incidence. Findings for individual muscle regions and alternative processing methods did not remain significant after false discovery rate correction and should be interpreted as exploratory. Longitudinal total thigh muscle CSA may warrant further study as a potential risk-stratification marker, but predictive performance and clinical utility were not evaluated. Trial registration The parent OAI cohort is registered at ClinicalTrials.gov (NCT00080171). This secondary analysis was not separately registered.