Predictive model for preoperative deep vein thrombosis in distal femur fractures using inflammatory blood markers: a single-center retrospective cohort study

Background The incidence of distal femur fractures (DFF) has shown a continuous annual increase. Preoperative deep vein thrombosis (DVT) of the lower limbs in DFF patients is associated with unfavorable clinical outcomes. Inflammation has been recognized as an essential contributor to thrombus formation, drawing increasing research attention in recent years. Objectives This study aimed to construct a predictive model utilizing inflammatory blood indicators to estimate the preoperative risk of DVT in patients with DFF. Methods A retrospective analysis was performed on 493 eligible patients treated for DFF, admitted to a single-center, between January 2022 and September 2025. Clinical baseline characteristics and laboratory parameters were collected. Independent risk factors for preoperative lower limb DVT were identified through univariate and multivariate binary logistic regression analyses. The predictive model was established using R 4.2.0, with patients randomly allocated to training and validation cohorts in a 7:3 ratio. Receiver operating characteristic (ROC) curves and corresponding areas under the curve (AUC) were computed for internal validation. Model performance was further assessed using calibration curves (CC) and decision curve analysis (DCA) in both cohorts. Results A total of 493 eligible patients were analyzed, comprising 205 cases with DVT and 288 without DVT. Univariate and multivariate binary logistic regression identified neutrophil- to-lymphocyte (NLR) (OR = 0.86, 95% confidence interval (CI) [0.774–0.95], P = 0.004), lymphocyte-to-monocyte ratio (LMR) (OR = 1.545, 95% CI [1.3–1.851], P = 0.000), white blood cell count (OR = 1.245, 95% CI [1.152–1.353], P = 0.000), neutrophil count (OR = 1.563, 95% CI [1.398–1.766], P = 0.000), lymphocyte count (OR = 0.151, 95% CI [0.077–0.285], P = 0.000), and eosinophil count (OR = 2.343, 95% CI [0.657–8.447], P = 0.19) as independent variables associated with preoperative lower limb DVT in DFF patients. Incorporating age with these indicators, a predictive model was developed. The ROC analysis demonstrated AUC values of 0.926 and 0.939 in the training and validation cohorts, respectively. Conclusions Among the preoperative inflammatory indicators evaluated, NLR, LMR, white blood cell count, neutrophil count, lymphocyte count, and eosinophil count were identified as independent determinants of preoperative lower limb DVT in patients with DFF. The predictive model constructed based on independent risk factors has excellent predictive efficacy and clinical application value.