Strength of Fibroblast Growth Factor 23 as a Cardiovascular Risk Predictor in Chronic Kidney Disease Weaken by ProBNP Adjustment

Abstract

Background: Various epidemiological studies linked high fibroblast growth factor 23 (FGF23) levels with cardiovascular events in chronic kidney disease (CKD). It remains enigmatic whether high FGF23 exerts adverse cardiovascular effects, or whether it reflects detrimental effects of residual confounders. Earlier studies adjusted for CKD-mineral bone disease (CKD-MBD) regulators of FGF23 rather than for recently discovered non-CKD-MBD regulators, among which iron deficiency and heart failure are of particular importance. Moreover, they used c-terminal FGF23 (cFGF23) assays rather than more specific intact FGF23 (iFGF23) assays. Methods: The CARE FOR HOMe study analyzed plasma ferritin, iFGF23, cFGF23 and N-terminal proBNP (NT-proBNP) along with conventional risk factors, among 575 CKD G2-G4 patients to determine the interaction between FGF23, its non-CKD-MBD regulators, and incident cardiovascular events in CKD patients. The participants were followed up for 5.1 ± 2.1 years for the occurrence of atherosclerotic events and hospitalization for acute decompensated heart failure. Results: cFGF23 correlated strongly with high iFGF23 (r = 0.607), fairly with high NT-proBNP (r = 0.453) and weakly with low ferritin (r = –0.207); correlation coefficients of iFGF23 with NT-proBNP and ferritin were numerically lower. In Kaplan-Meier analyses, both endpoints were predicted by cFGF23 and iFGF23. In Cox regression models, cFGF23 remained an outcome predictor after adjustment for conventional risk factors and ferritin. This prediction was largely eliminated when further adjusting for NT-proBNP. iFGF23 was less consistently associated with adverse outcome in partly adjusted models, and failed to predict outcome in fully adjusted models. Conclusion: In summary, iron deficiency and heart failure affect plasma FGF23. As adjustment for NT-proBNP virtually eliminates the association between plasma FGF23 and predefined outcome, we speculate that high FGF23, rather than exerting detrimental cardiovascular effects, mirrors prevalent heart disease.

Limited role for fibroblast growth factor 23 in assessing prognosis in heart failure patients: data from the TIME‐CHF trial

“…Hence, the present data are in contrast to some of the previously published cohort studies which suggested a strong and independent association with all‐cause mortality, cardiovascular mortality, or need for hospitalization even after adjustment for (NT‐pro) BNP. Our data are, however, in line with the CARE FOR HOMe study that initially showed a strong association between FGF23 and the development of HF in patients with chronic kidney disease, but this association was lost when adjusting for NT‐proBNP …”

Fibroblast growth factor 23: are we ready to use it in clinical practice?

“…Besides, cFGF23 is more consistently associated with outcome as shown in the meta-analysis of Xiao et al in which c-term FGF23 was associated with mortality in HD patients whereas iFGF23 measurement did not correlate to mortality [40]. Furthermore cFGF23 is a better predictor for identifying patients with declining renal function [41], atherosclerosis associated cardiovascular disease and heart failure [42]. Therefore, the cFGF23 assay may outperform the iFGF23 assay for clinical use, especially for the purpose of patients individuals risk assessment.…”

“…However, other studies showed no improvement at all [42,54]. Interestingly the study by Emrich et al found that when NT proBNP was added to the model the predictive value of FGF23 was largely eliminated and NT proBNP had a much stronger discriminating ability than FGF23 [42]. Overall, FGF23 only marginally improved the prediction for outcome.…”

“…Although one study showed that FGF23 improved prediction of fatal and non-fatal cardiovascular events in predialysis patients [49], another study showed small improvement of prediction only, yet without improved (NRI) classification. However, other studies showed no improvement at all [42,54]. Interestingly the study by Emrich et al found that when NT proBNP was added to the model the predictive value of FGF23 was largely eliminated and NT proBNP had a much stronger discriminating ability than FGF23 [42].…”

FGF23: A Review of Its Role in Mineral Metabolism and Renal and Cardiovascular Disease

“…According to meta-analysis of Xiao et al, C-terminal but not intact FGF23 was significantly associated with all-cause mortality risk [ 39 ]. Moreover, cFGF23 is considered a better predictor of CKD progression and atherosclerosis associated with cardiovascular diseases (CVD) and heart failure (HF) [ 40 , 41 ]. Thus, the cFGF23 assay may be a better choice for clinical use, especially regarding individual risk assessment.…”