Background/Purpose: Anti-neutrophil cytoplasmic antibody (ANCA)-associated vasculitis (AAV), combining granulomatosis with polyangiitis (GPA) and microscopic polyangiitis (MPA), is associated with a substantial risk of relapse or death. Particularly, the first year after diagnosis is characterized by a marked excess of vasculitis-related deaths. The factors determining risk of death or relapse in the long term have been widely investigated, but those determining short-term outcome are not well known. Accurate prediction of short-term prognosis is crucial to identify patients at highest risk of an adverse event. We aimed to develop a risk score to predict the 1-year risk of death or relapse in newly diagnosed AAV.

Methods: We studied patients with incident AAV enrolled in 4 international, randomized, multicenter clinical trials focusing on various subgroups with AAV. An initial set of 22 candidate variables, including the main demographic, clinical and laboratory data, all assessed at diagnosis, and GPA or MPA diagnoses, were considered for predicting 1-year survival or relapse risk. A stepwise approach using univariate and multivariate logistic regression models in 2,000 bootstrap samples was used to identify variables that consistently predicted death or relapse at 1 year. The regression coefficients computed in the final multivariate model were used to derive weights for each of the identified predictors of the composite outcome. The discriminative ability of the final score was evaluated by analysis of the area under the receiver operating characteristic curve (AUC). A sensitivity analysis was performed with multiple imputation methods to account for missing data and calculate an AUC for the complete dataset.

Results: Among the 535 subjects, we had complete data for all 22 analyzed variables for 441 subjects (244 GPA, 197 MPA). At 1-year follow-up, we recorded 44 deaths (10.0%) and 12 relapses (3.3%). We retained 9 variables, assigned weights of 1 to 3, for the final risk score model: age > 60 yr (1 point); female sex (1); ear, nose and throat involvement (1); serum creatinine level > 300 µmol/l (3); ANCAs directed against proteinase 3 (PR3-ANCA) (1) and myeloperoxydase (MPO-ANCA) (3); peripheral neutrophil count > 7,000/mm³(2); hemoglobin level < 10 g/dl (1) and C-reactive protein level > 10 mg/l (2). Each patient was assigned a risk score between 0 and 15. Accordingly, the 1-year risk for death or relapse was low (<5%) for 32% of patients, medium (5–20%) for 50%, and high (>20%) for 18%. The AUC for the prediction model was 0.79 (95% confidence interval 0.73–0.85) and 0.78 for the whole dataset of 535 subjects after missing-data imputation. 

Conclusion: This integrative AAV-risk prediction score may be useful in predicting a patient’s risk of death or relapse on short-term follow-up and may contribute to better risk-stratified characterization and management of AAV. Further validation in other AAV populations is required.

---

« Back to 2012 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/a-risk-score-for-predicting-short-term-incidence-of-death-or-relapse-in-anti-neutrophil-cytoplasmic-antibody-associated-vasculitis/