Analysis Of Risk Factors And Establishment Of Clinical Prediction Model For Latent Blood Loss After Total Hip Replacement In Overweight And Obese Patients

Background and Purpose: Since its introduction by Philip Wilesti in 1938,total hip arthroplasty has become the primary option for many elderly patients with end-stage hip disease.A European multicenter observational study by Lasocki et al.showed that the average perioperative blood loss during initial unilateral total hip arthroplasty(THA)was 1944 ml.The high blood loss means that the probability of postoperative allograft transfusion increases significantly,which may lead to complications or even death,as well as increasing the length of hospitalization and the financial burden on the patient.duration of hospitalization as well as the financial burden on the patient.In fact,the total blood loss after hip surgery is much higher than the actual observed blood loss.This phenomenon was first described by Pattison et al.in 1973 and was formally defined as invisible blood loss(HBL)in 2000.Currently,there is evidence that blood infiltration into tissues,periarticular and hemolysis may contribute to the formation of HBL.Previous literature has found that overweight and obese patients tend to have higher postoperative HBL.The author reviewed the literature and found that there is a lack of predictive models to guide clinicians.Therefore,the aim of this study was to identify risk factors that may contribute to elevated HBL in overweight and obese patients after THA through a retrospective analysis and to develop a relevant line graph predictive model to provide guidance for reducing HBL in overweight and obese patients after THA.Patients and Methods: A total of 505 overweight and obese patients treated with THA from January 2018 to November 2021 were included and randomized into the modeling and validation sets in a 7:3 ratio.Patient demographic information and relevant clinical data were collected,including gender,age,body mass index(BMI),diagnosis,American Society of Anesthesiologists score(ASA),hypertension,diabetes,cardiovascular disease,estimated systemic blood volume(EBV),preoperative hemoglobin(HGB),preoperative hematocrit(HCT),preoperative platelets(PLT),preoperative red blood cells(RBC)preoperative activated prothrombin partial time(APTT),preoperative prothrombin time(PT),preoperative international normalized ratio(INR),preoperative fibrinogen(FIB),preoperative prothrombin time(TT),preoperative blood calcium,preoperative erythrocyte sedimentation rate(ESR),preoperative D-dimer,preoperative blood calcium,intraoperative blood loss,operative time,preoperative systolic blood pressure,3 days postoperative HCT and The independent risk factors affecting HBL after THA in overweight and obese patients were finally obtained by Pearson correlation analysis,independent sample t-test and multiple linear regression analysis using SPSS 22.0 statistical software.The obtained independent risk factors were also used as independent variables to establish the nomogram graph prediction model of postoperative HBL in overweight and obese patients using R software(4.1.1),and their ROC curves,calibration curves and decision curve analysis(DCA)were plotted based on the modeling set and validation set data,respectively.Results: The patient’s HBL was 911±438 m L and the total perioperative blood loss was 1104±468 m L.Univariate analysis showed statistically significant differences(P<0.05)between HBL and age,operative time,preoperative red blood cells,preoperative D-dimer,intraoperative blood loss,diagnosis,hypertension,preoperative blood calcium level,and ASA.Multiple linear regression analysis showed that HBL was associated with diagnosis(regression coefficient=-344.397,P<0.001),hypertension(regression coefficient=-257.271,P<0.001),operative time(regression coefficient=2.445,P=0.001),preoperative red blood cells(regression coefficient=104.005,P=0.012),and preoperative D-dimer(regression coefficient=11.947,P=0.048)were strongly correlated.The area under the ROC curve(AUC)was 0.758 and 0.736 for the modeling and validation sets,respectively,and the slope of the calibration curve was close to 1.The DCA curve showed that the prediction model could result in better net clinical benefit for patients.Conclusion: Based on five independent risk factors,including preoperative D-dimer,time to surgery,preoperative RBC,hypertension,and diagnosis,the nomogram graph can predict the risk of excess HBL after THA in overweight and obese patients.The nomogram graph prediction model has good discrimination and accuracy and may result in better net clinical benefit for patients.