Anti-neutrophil cytoplasmic antibody-associated vasculitis (AAV) faces a higher risk of venous thromboembolism (VTE). We aimed to identify the predicators for VTE in AAV patients, develop and validate an AAV-VTE model. 138 AAV patients were retrospectively enrolled to identify the risk factors for VTE and develop AAV-VTE prediction mode. A temporal-validation cohort was composed of 45 AAV patients to record the incidence of VTE within 18 months. It showed that AAV patients with VTE exhibited higher disease activity, more organ involvement and immune imbalance including elevated natural killer (NK) cells and reduced regulatory T (Treg) cells. The high Birmingham Vasculitis Activity Score (BVAS), cardiac involvement, elevated eosinophil and NK cells were identified as the independent predictors for VTE. These four predicators were used to establish a nomogram model to provide a visualized and individualized risk of VTE in AAV. The model demonstrated good discriminative ability and calibration. Each 10% increase in predicted probability was associated with a 1.99-fold increase in VTE odds (OR = 1.990, 95% CI = 1.599-2.646, P < 0.001). The temporal-validation also showed that it had good discrimination. The prediction model showed a clinical application value for distinguishing VTE patients and making clinical decision. We identified the immunological and clinical risk factors for VTE in AAV, especially the elevated NK cells and cardiac involvement. The AAV-VTE model might help to facilitate individualized risk prediction and support clinical decision-making in patients with AAV.
Wu et al. (Wed,) studied this question.