Analysis of the natural collapse course of non-traumatic osteonecrosis of the femoral head based on the matrix model

TL;DR

This study evaluates a new method to predict femoral head collapse in osteonecrosis by combining three existing models.

Contribution

The novel contribution is combining three classical prediction methods into a matrix model to improve collapse prediction accuracy.

Findings

The matrix model showed better predictive ability than existing staging systems (P < 0.05).

The median survival time of the femoral head was 3 years according to Kaplan–Meier analysis.

The matrix model had an AUC of 0.771, indicating strong predictive value for femoral head collapse.

Abstract

There are many predictions about the progression of natural collapse course of osteonecrosis of the femoral head. Here, we aimed to combine the three classical prediction methods to explore the progression of the natural collapse course.

This retrospective study included 127 patients admitted to our hospital from October 2016 to October 2017, in whom the femoral head had not collapsed. Logistic regression analysis was performed to determine the collapse risk factors, and Kaplan–Meier survival curves were used for femoral head survival analysis. The collapse rate of the femoral head was recorded within 5 years based on the matrix model. The specificity of the matrix model was analyzed using the receiver operating characteristic curve.

A total of 127 patients with a total of 202 hips were included in this study, and 98 hips collapsed during the follow-up period. Multivariate logistics regression analysis showed that the predictive ability of the matrix model was stronger than Association Research Circulation Osseous staging, Japanese Investigation Committee classification, and area (P < 0.05). Kaplan–Meier survival curve showed that the median survival time of femoral head in patients was 3 years. The result of the receiver operating characteristic curve analysis showed that the area under the curve (AUC) of the matrix model had better predictive value (AUC = 0.771, log-rank test: P < 0.001).

We creatively combined the three classical prediction methods for evaluating the progression of the natural collapse course based on the matrix model and found that the higher the score of the matrix model, the higher the femoral head collapse rate. Specifically, the matrix model has a potential value in predicting femoral head collapse and guiding treatment selection.

Figures