Establishment and Validation of A Nomogram Model of the Risk of Femoral Head Necrosis Complicated by Cannulated Screw Fixation Surgery After Femoral Neck Fracture

ZHANG Fanghui; YANG Yunjian; LI Yibo; QIN Jiajun; ZHANG Yaoyu; OUYANG Fucheng; CAI Xianhua; WEI Shijun; XU Feng; LI Zhi

doi:10.13730/j.issn.2097-2148.2024.05.009

Abstract：Objective To explore the risk factors of femoral head necrosis complicated by cannulated screw fixation(CSF)surgery after femoral neck fracture(FNF),and to construct an individualized nomogram model for predicting the risk of femoral head necrosis.Methods A retrospective analysis was performed for 252 eligible FNF patients in the orthopedics depart-ment of the author's hospital,and clinical data were collected.Univariate analysis was used to screen the factors associated with femoral head necrosis after CSF surgery,and then they were included in the multivariate Logistic regression model to analyze the risk factors.A nomogram model was established by R software,which was internally verified by the Bootstrap method and the calibration curve was drawn;the discrimination of the predictive model was evaluated by receiver operating characteristic(ROC)curve and area under the curve(AUC),the decision curve analysis(DCA)curve was drawn to evaluate the applicative value of the predictive model.Results After the 252 FNF patients with CSF surgery,61 had femoral head necrosis,with an incidence of 24.21％.The Logistic regression analysis results showed that age ≥60,history of diabetes mellitus,subcephalic fractures,Ⅲ-Ⅳ Garden classification,Ⅱ Pauwels classification and Ⅱ-Ⅲ VN classification were risk factors for femoral head necrosis after CSF surgery in FNF patients(P＜0.05).The consistency predicted good agreement between the risk of the femoral head necrosis after CSF surgery and the actual occurrence of risk in FNF patients,the AUC of ROC curve for internal validation of the nomogram model predicting the risk of femoral head necrosis after CSF surgery in FNF patients was 0.968(95％CI:0.949-0.987),the DCA showed that the model was able to bring clinical benefit when the threshold probability of the patient was between 0 and 1.Conclusion In this study,a high-precision nomogram predictive model is successfully constructed,which included predictors such as age,history of diabetes mellitus,fracture line site,Garden classification,Pauwels classification,VN classification,etc.,which is helpful to improve the early identification and management efficiency of FNF patients at high risk of femoral head necrosis after CSF surgery.