BACES Score for Predicting Mortality in Nontuberculous Mycobacterial Pulmonary Disease

Abstract

Rationale: Because the prognosis of nontuberculous mycobacterial pulmonary disease varies, a scoring system predicting mortality is needed. Objectives: We aimed to develop a novel scoring system to predict mortality among patients with nontuberculous mycobacterial pulmonary disease. Methods: We included patients age >= 20 years with newly diagnosed nontuberculous mycobacterial pulmonary disease, with Mycobacterium avium, M. intracellulare, M. abscessus subsp. abscessus, or M. abscessus subsp. massiliense. Cox proportional hazards models were used to identify predictors of mortality in a derivation cohort, and a scoring system was developed. It was validated in an independent prospective cohort. Measurements and Main Results: A total 1,181 and 377 patients were included in the derivation and validation cohorts, respectively. In the final model, body mass index <183 kg/m(2) (1 point), age >= 65 years (1 point), presence of cavity (1 point), elevated erythrocyte sedimentation rate (1 point), and male sex (1 point) were selected as predictors for mortality. We named this novel scoring system BACES (body mass index, age, cavity, erythrocyte sedimentation rate, and sex). Harrell's C-index for the BACES score was 0.812 (95% confidence interval, 0.786-0.837) in the derivation cohort and 0.854 (95% confidence interval, 0.797-0.911) in the validation cohort, indicating excellent discrimination performance. The estimated 5-year risk of mortality was 1.2% with BACES score 0 and 82.9% with BACES score 5. Conclusions: We developed the BACES score, which could accurately predict mortality among patients with nontuberculous mycobacterial pulmonary disease caused by M. avium, M. intracellulare, M. abscessus subsp. abscessus, or M. abscessus subsp. massiliense.