Mid-to long-term survival analysis following allograft fibular support for Japanese investigation committee type C non-traumatic osteonecrosis of the femoral head: a single-center retrospective cohort study

Abstract

This study evaluated mid- to long-term femoral-head survival following allograft fibular support (AFS) for Japanese Investigation Committee (JIC) type C non-traumatic osteonecrosis of the femoral head (ONFH) at the pre-collapse or early-collapse stage and examined factors associated with prognosis. This single-center retrospective cohort study enrolled 68 patients (74 hips) with JIC type C1 or C2 non-traumatic ONFH at Association Research Circulation Osseous (ARCO) stage II or IIIA treated with AFS between 2015 and 2022. Each hip was the analysis unit, and patient-clustered robust variance was used to account for bilateral hips. Kaplan‑Meier survival estimates and univariable Cox proportional hazards models were used to evaluate B-grade collapse (BGC) and total hip arthroplasty (THA) as two separate endpoint events. Multivariable models were not fitted because only 16 BGC and seven THA events occurred. Sixteen hips (21.6%) developed BGC, and seven hips (9.5%) underwent THA. Median observation time was 30.39 months (IQR, 24.93–40.13; range, 16.87–62.47 months). Estimated 3-year survival was 74.6% (95% CI, 60.6%–84.3%) with BGC as the endpoint and 88.9% (95% CI, 75.7%–95.2%) with THA as the endpoint. The 5-year estimates were numerically unchanged, but only two hips remained at risk at 60 months. Compared with C1, C2 was associated with higher average hazards of BGC (HR, 3.72; 95% CI, 1.28–10.77; p = 0.015) and THA (HR, 8.22; 95% CI, 1.50–45.17; p = 0.015). AFS may be a promising structural hip-preserving option for selected patients with JIC type C ONFH at ARCO stage II or IIIA. Type C2 lesions had poorer observed survival than type C1 lesions. The 3-year estimates are supported by the available follow-up, whereas the numerically unchanged 5-year estimates are exploratory. Prospective comparative studies incorporating validated patient-reported outcomes are required before treatment efficacy can be established. Not applicable; this was a retrospective observational cohort and participants were not prospectively assigned to an intervention.