Intact and C-Terminal FGF23 Assays-Do Kidney Function, Inflammation, and Low Iron Influence Relationships With Outcomes?

Abstract

Higher fibroblast growth factor-23 (FGF23) concentrations are associated with heart failure and mortality in diverse populations, but the strengths of associations differ markedly depending up on which assay is used. We sought to evaluate whether iron deficiency, inflammation, or kidney function account for differences in the strengths of associations between these 2 FGF23 assays with clinical outcomes. Case cohort study from the Cardiovascular Health Study. A total of 844 community-dwelling individuals aged 65 years or older with and without chronic kidney disease were followed for 10 years. Outcomes included death, incident heart failure (HF), and incident myocardial infarction (MI). Exposure was baseline intact and C-terminal FGF23. Using modified Cox models, adjusting sequentially we tested whether observed associations of each assay with outcomes were attenuated by iron status, inflammation, kidney function, or their combinations. FGF23 measured by either assay was associated with mortality in unadjusted analysis (intact FGF23 hazard ratio [HR] per 2-fold higher 1.45; 95% CI, 1.25-1.68; C-terminal FGF23 HR 1.38; 95% CI, 1.26-1.50). Adjustment for kidney function completely attenuated associations of intact FGF23 with mortality (HR 1.00; 95% CI, 0.85-1.17), but had much less influence on the association of C-terminal FGF23, for which results remained significant after adjustment (HR 1.15; 95% CI, 1.04-1.28). Attenuation was much less with adjustment for iron status or inflammation. Results were similar for the HF end point. Neither C-terminal or intact FGF23 was associated with MI risk. The relationship of FGF23 with clinical end points is markedly different depending on the type of FGF23 assay used. The associations of biologically active FGF23 with clinical end points may be confounded by kidney disease, and thus much weaker than previously thought.