Loss of walking independence one year after primary total hip arthroplasty for osteonecrosis of the femoral head: incidence and risk prediction model.

BACKGROUND

Assessment of postoperative ambulation in osteonecrosis of the femoral head (ONFH) patients treated with total hip arthroplasty (THA) is limited. This study aimed to define the incidence and risk factors for losing walking independence (LWI) at one-year postoperatively in patients with ONFH undergoing primary THA, and to establish and validate a predictive nomogram.

METHODS

This was a retrospective analysis of prospective collected data from patients admitted to a tertiary referral hospital with ONFH who underwent primary unilateral THA from October 2014 to March 2018. The Functional Independence Measure-Locomotion scale was used to quantify walking independence and was documented at a one-year continuous postoperative follow-up, which classified patients with a final score below 6 as LWI. Multivariate logistic regression identified independent risk factors for LWI, and a predictive nomogram was constructed based on the analysis results. The stability of the model was assessed using patients from April 2018 to April 2019 as an external validation set.

RESULTS

1152 patients were enrolled in the study, of which 810 were used in the training cohort and the other 342 for the validation cohort. The incidence of LWI was 5.93%. Multivariate analysis revealed that age 62 years or older (odd ratio (OR) = 2.37, 95% confidence interval (CI) 1.07-5.24), Charlson's comorbidity index 3 or higher (OR = 3.64, 95% CI 1.09-12.14), Association Research Circulation Osseous stage IV (OR = 2.16, 95% CI 1.03-4.54), reduced femoral offset (OR = 2.41, 95% CI 1.16-5.03), and a higher controlling nutritional status score (OR = 1.14, 95% CI 1.01-1.30) were independent risk factors of LWI. The nomogram had a concordance index of 0.773 and a Brier score of 0.049 in the training set, with corrected values of 0.747 and 0.051 after internal validation. The receiver-operating characteristic curve, calibration curve, Hosmer-Lemeshow test, and decision curve analysis all performed well in both the training and validation cohorts.

CONCLUSIONS

This study reported a 5.93% incidence of LWI and established a risk prediction model in patients undergoing THA for ONFH, supporting targeted screening and intervention to assist surgeons in assessing ambulation capacity and managing rehabilitation.