Abstract / Summary
Abstract Cardiometabolic risk is increased in rheumatoid arthritis (RA), but early risk may remain unrecognized before overt metabolic abnormalities emerge. Whether neutrophil-associated inflammation can help identify this latent risk and its potential molecular link remains unclear. We analyzed 1,028 patients with RA to examine factors associated with elevated triglyceride-glucose body mass index (TyG-BMI) and developed machine learning models using routinely available clinical variables to distinguish patients in the highest (Q4) and lowest (Q1) TyG-BMI quartiles. An exploratory 12-month follow-up assessment was also conducted. Transcriptomic analyses identified candidate neutrophil-related markers, which were subsequently evaluated in an independent prospective multicenter cohort of 502 patients. Neutrophil count remained independently associated with elevated TyG-BMI after full multivariable adjustment and additional adjustment for antirheumatic medication use (OR = 1.17, 95% CI 1.08–1.27; P < 0.001). The selected logistic regression model achieved an AUC of 0.829 in the internal validation set for distinguishing the Q4 and Q1 TyG-BMI groups. Among patients without overt cardiometabolic disease and with normal baseline metabolic indicators, newly reported physician-diagnosed cardiometabolic conditions during the 12-month follow-up were reported by 21.3% of Q4 participants and 10.7% of Q1 participants (RR = 1.98, 95% CI 1.02–3.87; P = 0.044). Transcriptomic analyses identified lipocalin-2 (LCN2) as a candidate molecular link. In the prospective cohort, adding LCN2 modestly increased the AUC from 0.799 to 0.812 (DeLong P = 0.286); net reclassification improvement (NRI) was not significant, whereas integrated discrimination improvement (IDI) improved significantly ( P = 0.008). Exploratory mediation analysis yielded an estimated mediated proportion of 25.8% ( P < 0.001). Neutrophil-associated inflammation may help identify latent cardiometabolic risk in RA, with LCN2 emerging as a potential molecular link and complementary biomarker.