Abstract
BACKGROUND: The need for a transfusion is one of the adverse events following total knee arthroplasty (TKA), and accurately predicting this need remains challenging for arthroplasty surgeons. The purpose of the present research is to study the preoperative predictors of transfusion risk in patients following TKA and develop a nomogram. METHODS: The nomogram was developed based on a training set of 5402 patients who underwent TKA at the Affiliated Hospital of Qingdao University between September 2013 and November 2018. The independent predictors of transfusion were identified by univariate, LASSO, and binary logistic regression analyses. Then, a nomogram was established based on these independent predictors. The area under the curve (AUC), calibration curve, and decision curve analysis (DCA) were selected to evaluate the nomogram. The results were validated using an independent set of 1116 patients who underwent TKA between December 2018 and September 2019. In addition, we also carried out subgroup analyses in the training and testing sets based on the independent predictors. RESULTS: Five independent predictors were identified by multivariate analysis and were used to establish the nomogram. The AUCs of the nomogram were 0.884 (95% CI: 0.865-0.903) and 0.839 (95% CI, 0.773-0.905) in the training and testing sets, respectively. In both the training and testing sets, the calibration curve indicated that the prediction by the nomogram was highly consistent with the actual observation, and the DCA indicated that the nomogram had a favorable level of clinical usefulness. In addition, the AUC of the nomogram was significantly higher than the AUC of any independent predictor for predicting transfusion risk following TKA, and the subgroup analysis showed good performance in 20 subgroups. CONCLUSION: Lower preoperative Hb levels, simultaneous bilateral TKA, lower BMI, older age, and coronary heart disease were identified as independent predictors of postoperative transfusion in patients following TKA. A nomogram incorporating the above five predictors could accurately predict the transfusion risk.