Development of a Nomogram Model to Predict Mortality in ANCA-Associated Vasculitis Patients With Pulmonary Involvement

Abstract

Objective

Risk assessment and prognosis prediction are crucial for patients with pulmonary involvement in antineutrophil cytoplasimc antibody associated vasculitis (AAV). This study was conducted to create and internally validate a prognostic model for mortality of pulmonary involvement in patients with AAV that provides individualized risk assessments.

Methods

A cohort of 150 patients diagnosed with AAV at the Second Affiliated Hospital of Nanchang University Hospital between January 2013 and July 2022 was included, using data obtained from the Chinese Rheumatism Data Center (CRDC). The model was developed using Cox proportional hazards regression and the least absolute shrinkage and selection operator. To validate the model, assessments were conducted for discrimination, calibration, and through decision curve analysis.

Results

The mean survival time of lung involvement AAV patients was 57.0 ± 4.1 months. In the final predictive model for death, four clinical variables were included: age at baseline, history of tumors, baseline hemoglobin level, and the level of the percentage of forced vital capacity to the normal predicted value. One-, two-, and three-year AAV patients with pulmonary involvement mortality probability-predictive nomogram were established. Internal validation of the model was conducted, yielding Harrell's concordance index (0.884), a Brier score of 0.088, and a calibration curve indicating satisfactory performance.

Conclusion

We constructed a risk model utilizing easily accessible clinical risk factors, which could accurately forecast the future mortality risk associated with pulmonary involvement in AAV patients.