Discordance Between Fasting Plasma Glucose- and HbA1c-Based Glycemic Classifications and Associated Metabolic and Hemodynamic Profiles in Thai Older Adults

Background: Fasting plasma glucose (FPG) and glycated hemoglobin (HbA1c) reflect different aspects of glycemic regulation and may yield discordant laboratory-defined classifications, particularly among older adults. Methods: This cross-sectional analytical study included 203 community-dwelling Thai adults aged ≥60 years. Single FPG and HbA1c measurements were classified as normal-range, prediabetes-range, or diabetes-range values. Agreement was evaluated using exact agreement, Cohen’s kappa (κ), linear-weighted κ, and symmetry tests. Four binary laboratory-defined phenotypes were compared using age- and sex-adjusted HC3 robust regression. The Benjamini–Hochberg procedure was applied separately across 13 omnibus tests and 39 phenotype-specific contrasts. Results: Exact three-category agreement was 70.9% (unweighted κ = 0.376; 95% CI, 0.247–0.498; weighted κ = 0.486; 95% CI, 0.355–0.601). Binary classifications were discordant in 25.6% of participants; HbA1c-only dysglycemia was numerically more frequent than FPG-only dysglycemia (15.8% vs. 9.9%), but directional asymmetry was not significant (p = 0.127). FDR-supported overall phenotype differences involved adiposity, fasting insulin, and lipid measures, whereas hemodynamic measures did not differ. Conclusions: FPG and HbA1c identified overlapping but non-equivalent laboratory-defined groups. Discordant single measurements warrant confirmation and evaluation of biological, analytical, and treatment-related influences; this study does not establish that routine dual testing improves diagnosis or outcomes.