Background: Patients with rheumatoid arthritis (RA) are at increased risk of deep vein thrombosis (DVT) after joint replacement, but the predictive value of iron metabolism markers remains unclear. This study aimed to investigate the predictive value of serum iron metabolism-related indicators for DVT after joint replacement in patients with RA, and to construct and internally validate a combined prediction model. Methods: A total of 180 RA patients who underwent joint replacement in our hospital from January 2019 to December 2023 were retrospectively enrolled. According to postoperative confirmed DVT, the patients were divided into a DVT group (n = 45) and a non-DVT group (n = 135). The general data, preoperative iron metabolism indices (serum iron, ferritin, total iron binding capacity), D-dimer, and DAS28 score were collected. Univariate and multivariate Logistic regression identified independent risk factors for postoperative DVT. The Box-Tidwell test was used to assess linearity; variables that violated the linearity assumption were categorized into quartiles, while variables that satisfied the linearity assumption were entered as continuous variables. Model discrimination was evaluated using Receiver Operating Characteristic (ROC) curve analysis, internal validation was performed using bootstrap resampling, calibration was performed using a calibration curve, and clinical utility was assessed using decision curve analysis (DCA). Results: The Box-Tidwell test showed that ferritin (p = 0.03) and D-dimer (p 0.05). DCA demonstrated that the prediction model simultaneously outperformed both the ‘treat all’ and ‘treat none’ strategies within the clinically relevant threshold range of 15% to 60%. Conclusions: Preoperative ferritin, D-dimer, and DAS28 score are independent risk factors for DVT after joint arthroplasty in RA patients. The model based on the above indicators has good predictive performance and clinical utility, which can provide a reference for the formulation of clinical individualized thrombosis prevention strategies.
Dai et al. (Thu,) studied this question.