Complete blood count-derived myeloid-lymphoid inflammatory phenotype in rheumatoid arthritis complicated by acute myocardial infarction: an age- and sex-balanced retrospective study

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Background Rheumatoid arthritis (RA) is associated with excess cardiovascular morbidity, including acute myocardial infarction (AMI). Routine complete blood count (CBC)-derived inflammatory indices are clinically accessible, but their value for characterizing cardiovascular comorbidity in RA remains incompletely defined. We examined whether CBC-derived markers form a coordinated myeloid-lymphoid inflammatory phenotype associated with AMI in patients with RA. Methods We retrospectively analyzed 110 healthy controls, 246 consecutive patients with RA without AMI, and 90 consecutive patients with RA and AMI treated at Weifang People's Hospital between January 2020 and December 2025. The primary phenotyping analysis used an age- and sex-balanced sample of 57 participants per group. CBC-derived indices were log2-transformed and standardized. Principal component analysis (PCA), permutational multivariate analysis of variance (PERMANOVA), ordinal logistic regression, RA-only multivariable logistic regression, bootstrap elastic-net stability analysis, and correlation-network mapping characterized CBC-derived inflammatory patterns. Results In the full cohort, RA-AMI participants were older and less often female than RA-noAMI and healthy controls. Within RA, AMI was accompanied by longer disease duration, greater glucocorticoid exposure, higher CRP/ESR/DAS28-CRP, higher triglycerides, lower HDL-C, and more frequent cardiometabolic risk. In the matched sample, CBC-derived inflammatory indices increased stepwise from healthy controls to RA without AMI and RA with AMI. The first CBC phenotype principal component (PC1) explained 61.1% of multivariable variance and showed positive loadings for AISI, SIRI, SII, NLR, MLR, neutrophils, monocytes, and platelets, with an opposite loading for lymphocytes. Group-level multivariable separation was confirmed by PERMANOVA (R2 = 0.20, pseudo-F = 20.85, P = 0.0002). Each 1-SD increase in PC1 was associated with higher clinical tier (OR 3.64, 95% CI 2.51–5.28). Within RA, PC1 remained associated with AMI status after adjustment for age, sex, glucocorticoid dose, CRP, HDL-C, LDL-C, and cardiometabolic risk (adjusted OR 1.99, 95% CI 1.32–3.02). Bootstrap elastic-net analysis most consistently retained NLR and neutrophil burden. Conclusion Routine CBC-derived indices characterize a coordinated myeloid-lymphoid inflammatory phenotype associated with concurrent AMI in RA. These findings support CBC-based inflammatory phenotyping as a low-cost, hypothesis-generating research approach, while direct mechanistic, longitudinal, and prospective validation remains necessary.