Risk Factors Analysis And Prediction Model Construction Of Dislocation After Total Hip Arthroplasty

BackgroundTotal hip arthroplasty（THA）has been widely used in clinical practice and is a significant and effective choice for patients with end-stage hip joints.However,postoperative prosthesis dislocation is undoubtedly a devastating complication for both the patients and surgeons.It is one of the leading causes of revision after THA and a complication that can jeopardize the quality of life of patients.Unfortunately,there is no consensus on the influencing causes of prosthesis dislocation,and there is controversy about the main factors that induce dislocation of the prosthesis.ObjectiveTo explore and analyze the related risk factors of prosthetic dislocation after THA,a retrospective study was conducted and a risk prediction model was constructed,to provide a reliable tool for clinicians to predict the dislocation risk.Meanwhile,it provides a theoretical basis for reducing the chance of postoperative dislocation or developing into repeated dislocation and improving the quality of life of patients.MethodsThis study retrospectively collected 41 patients with prosthetic dislocation after the first THA in our hospital from Jan 2013 to Dec 2021,including 32 cases with only one dislocation and 9 cases with multiple dislocations during the follow-up period.1169 patients without dislocation after THA were enrolled in the control group,and the follow-up time ended in April 2022.The baseline data（including the hospital reason,gender,age,BMI,history of basic diseases,neurological disease history,and lumbosacral disease history）,intraoperative indicators（including anesthesia mode,operation duration,acetabular cup size,and prosthesis head diameter）,and postoperative indicators（including femoral eccentricity,acetabular abduction angle,femoral anteversion,the vertical distance from small rotation to teardrop,combined anteversion,acetabular-head ratio,and changes of femoral eccentricity and limb length before and after surgery）were compared between the two groups.All the related factors that can cause postoperative dislocation of the prosthesis were included in univariate data analysis and multivariate unconditional logistic regression analysis,to determine the related risk factors.R software is used to draw a risk prediction model to predict the risk of dislocation after THA.ResultsThe results showed that there was no significant difference between the two groups in the hospital reason,gender,chronic medical history,anesthesia method,BMI,operation time,acetabular cup size,lesser trochanter to teardrop（P>0.05）.However,there was a significant difference between the two groups in the age,history of neurological disease,history of lumbosacral disease,normal acetabular abduction angle,normal femoral anteversion angle,normal combined anteversion angle,prosthetic head diameter,acetabular-to-head ratio,postoperative femoral eccentricity（P<0.05）.Multivariate logistic regression analysis indicated that age,neurological disease history,lumbosacral disease history,abnormal acetabular abduction angle,abnormal femoral anteversion,abnormal combined anteversion,and femoral prosthesis head size had an obvious correlation with prosthesis dislocation after THA（P<0.05）.Based on the results of logistic regression analysis,independent risk factors with statistical significance are incorporated into R software,and the risk prediction model of the nomogram is established by the RMS package.A calibration curve is used to evaluate the calibration degree of the nomogram model and found that the fitting degree between the calibration curve and the ideal measured curve is ideal,indicating that the predicted value is consistent with the ideal measured value.The validity of the calibration curve is evaluated by C-index and shows that the C-index is 0.943（95%CI:0.909-0.976）and the corrected C-index is 0.933.ROC analysis is used to verify the prediction ability of the nomogram model for postoperative prosthesis dislocation risk.The area under the ROC curve is 0.943,and the optimal critical values in the ROC curve are 0.054（0.908,0.854）,respectively.These results show that this model has good prediction performance.ConclusionTherefore,the patient’s age,neurological disease history,lumbosacral disease history,abnormal acetabular abduction angle,abnormal femoral anteversion angle,abnormal combined anteversion angle,and size of femoral prosthesis head may be independent risk factors for dislocation of the prosthesis after THA.The nomogram model established in this paper can predict the dislocation risk after THA.However,a large sample and multi-center study are needed to support these conclusions.