Abstract
Study participants with rheumatoid arthritis (RA) have an elevated risk for nontuberculous mycobacterial pulmonary disease (NTM-PD), which limits the treatments for RA. Biomarkers for NTM-PD in study participants with RA are required. Patients with NTM-PD have been studied for small-molecule metabolites, although few have been performed for NTM-PD associated with RA. Therefore, we performed lipidomic profiling of NTM-PD in the urine specimens of study participants with RA to discover useful biomarkers. Urine specimens provided by 90 study participants with RA, with or without NTM-PD were subjected to lipidomic analysis. Univariate analysis found that the urinary concentrations of lysophosphatidic acid (LPA) 22:5 and phosphatidic acid (PA) 36:1 were altered in study participants with RA and NTM-PD (respective areas under the curves of receiver operating characteristic (AUROCs) were 0.977 and 0.811; P = 3.83 × 10(-20) and 1.37 × 10(-5)), when compared with the levels in urine specimens of study participants with RA without NTM-PD. The partial least squares-discriminant analysis model created from these two phospholipids validated their ability to discriminate between the study participants with or without NTM-PD (AUROC: 0.988 [95% confidence interval 0.958-1.000]). Differences between the urinary levels of the phospholipids LPA and PA in study participants with RA and with or without NTM-PD were significant. Lipidomic profiling of urine samples should be effective in the process of evaluating biomarkers for NTM-PD associated with RA.