Analysis Of Risk Factors And Construction Of Prediction Model Of PLOS In Patients With Femoral Head Necrosis After Primary Total Hip Replacement

Objective: To investigate the risk factors of Prolonged hospital Stay(p LOS)after primary Total Hip Arthroplasty(THA)in patients with femoral head necrosis and construct clinical prediction modelsMethod : The clinical data of 159 patients aged over 55 years who underwent initial total hip replacement due to femoral head necrosis in A hospital Autonomous Region from December 2017 to December 2022 were retrospectively analyzed.The main outcome of concern was the larger-than-expected duration of postoperative LOS,which were divided into normal((LOS<75th)and prolonged(LOS ≥75th)groups based on postoperative LOS.Gender,age,body mass index(BMI),preoperative anemia and hypoproteinemia,number of medical diseases,method of anesthesia,American Society of Anesthesiologists were collected Anesthesiologists(ASA)grading,duration of surgery,amount of intraoperative blood loss,postoperative anemia and hypoproteinemia,and postoperative adverse events.The risk factors that may affect the length of postoperative hospital stay were screened out by binary Logistic regression analysis.All variables were included in Lasso regression,the optimal penalty coefficient λ was determined according to internal cross-validation,and then the Logistic regression was re-fitted to establish the prognosis prediction model and draw the column graph.Bootstrap was used for internal verification,and the differentiation and calibration degree of C index and Brier score were calculatedResult: A total of 159 patients were included in the clinical data,of which93 were male,accounting for 58.5%,and 66 were female,accounting for 41.5%.Patients ranged in age from 55 to 88 years,with an average age of(65.6±6.8)years.Postoperative LOS5-35 days,mean LOS(11.8±4.2)days,postoperative LOS75th=14 days,were divided into normal length of stay group(LOS<14 days108 cases about 67.9%)and extended hospitalization group(LOS≥14 days 51 cases about 32.1%).Binary Logistic regression analysis showed that there were statistically significant indicators:BMI<18.5kg/m2[OR=6.111;CI 95%CI,1.481-25.221;P=0.012],ASA3 and Grade 4 [OR=4.238;CI 95%CI,1.745-10.292;P=0.001],operation time >120min[OR=3.213;CI 95%CI,1.345-7.676;P=0.009],blood loss >400ml[OR=4.092;CI 95%CI,1.592-10.516;P=0.003],postoperative adverse events [OR=5.529;confidence interval 95%CI,2.093-14.606;P=0.001]were independent risk factors for p LOS in patients with femoral head necrosis after THA.In R language,five predictors of postoperative p LOS were selected according to Lasso regression,including ASA III or IV grade,operation time >120min,blood loss >400ml,postoperative anemia and postoperative adverse events.The above variables were used to establish a prediction model for the length of hospital stay after THA for femoral head necrosis and converted into a column graph.Model verification: the C-index and Brier of the prediction model were 0.769 and 0.174,respectively.In Bootstrap internal verification,the C-index and Brier of the prediction model were 0.702 and 0.204,respectively.Conclusion : BMI<18.5kg/m2,ASA Grade III and IV,operation time >120min,blood loss >400ml and postoperative adverse events were independent risk factors for prolonged hospital stay after THA.The established clinical prediction model can better distinguish the patients who need prolonged hospitalization after THA for femoral head necrosis.