The principal finding in the present study was that FGF23 was independently associated with CAAC in CKD patients who were not on dialysis. The prevalence of CAAC was significantly increased as the kidney function declined. FGF23levels were significantly higher in CKD stage 3 patients than in CKD stages 1–2 patients. In contrast, serum phosphate levels were significantly higher only in CKD stage 5patients compared with CKD stages 1–2 patients, and were not related to CAAC. These results suggest that measurement of FGF23 is more important than measurement of serum phosphorus for the detection of risk factors associated with CAAC.
In ESRD patients, hyperphosphatemia is thought to be associated with vascular calcification
[24–26]. In contrast, the association of serum phosphorus levels with vascular calcification in predialysis CKD patients has remained uncertain. Some studies have shown no significant correlation between serum phosphorus and vascular calcification
[1, 27], while it has been demonstrated that higher serum phosphate levels, although still within the normal range, are associated with a greater prevalence of vascular and valvular calcification in subjects with moderate CKD
. Most patients develop hyperphosphatemia only at CKD stages 4 and 5
[29, 30], despite the progressive elevation of serum PTH and FGF23 levels as a defense mechanism to prevent an increase in serum phosphate levels. The present study also showed that serum phosphate levels were significantly high only in CKD stage 5 patients compared with CKD stages 1–2 patients, whereas significantly higher FGF23 levels were already observed in CKD stage 3 patients compared with CKD stages 1–2 patients. These findings are in agreement with previous studies
[8, 10] reporting that an increase in FGF23 levels precedes an increase in serum phosphate levels in CKD patients, and they may support the hypothesis that elevated FGF23 concentrations are an early abnormality of disordered phosphorus metabolism in CKD patients
Vascular calcification appears early in the course of CKD (i.e., well before ESRD is reached)
. One study reported that a high prevalence (64%) of coronary artery calcification was found in CKD patients with a median creatinine clearance of 37 mL/min/1.73m2 and a median serum phosphorus of 3.8 mg/dL
. In the present study, a high prevalence of CAAC was already present in CKD stage 3 patients, who also had a mean serum phosphate level of 3.4 mg/dL. Moreover, concerning the recent findings that high normal serum phosphate levels are associated with vascular calcification in predialysis CKD patients
, it is possible that even lower serum phosphate levels contribute to vascular calcification. In addition, given the present finding that FGF23 levels, but not serum phosphate levels, were identified as an independent factor for CAAC, FGF23 measurement appears to be more important for detecting CAAC-associated risk factors in CKD patients who are not undergoing dialysis.
Although there is the possibility that elevated FGF23 levels affect bone metabolism or vascular calcification directly and/or indirectly, the precise mechanism of such an effect remains unknown. The potential finding of FGF23 levels as a biomarker of CAAC in the present study cannot lead to the conclusion of a real role of FGF23 in the pathogenesis of vascular calcification.
The main physiological stimuli of FGF23 secretion are increased dietary phosphorus intake and increased 1,25(OH)2D levels
. Accordingly, in the present study, we studied patients who had not been administered active vitamin D3 or phosphate binders, to more precisely determine whether FGF23 levels are associated with vascular calcification. Burnett et al. demonstrated that dietary phosphate loading increases the fractional excretion of phosphate and circulating FGF23 in 66 healthy males and females, suggesting that dietary phosphate is an important regulator of circulating FGF23 levels in humans
. However, an association of phosphate loading with changes in FGF23 levels has not been fully documented in CKD patients. In a small, randomized controlled trial of CKD patients, Isakova et al. recently reported that CKD patients receiving a higher phosphorus diet plus placebo for 2 weeks demonstrated a significant increase in FGF23 levels, with no significant changes over time in serum phosphate levels, suggesting an important effect of dietary phosphorus intake on FGF23 levels rather than serum phosphate levels
. However, it remains unknown whether high dietary phosphate loading contributes to the development of vascular calcification in CKD patients.
Moderately uremic mice fed a high phosphate diet are not hyperphosphatemic but have a significant rise in serum FGF23 levels that are significantly correlated with arterial medial calcification
. In the present study, phosphate intake or loading was not assessed in a quantitative manner over prolonged time periods before the patients’ admission. Accordingly, it remains unknown whether elevated FGF23 levels in some of our study population might reflect a reaction to phosphate loading, and the effect of phosphate loading on vascular calcification could not be assessed.
A previous study demonstrated that in normophosphatemic CKD patients there was a significant decrease in FGF23 levels, but not serum phosphorus levels, in patients treated with the non-calcium containing phosphate binder, sevelamer, for 6 weeks, whereas there was no decrease in FGF23 in calcium acetate-treated patients
. However, it is unclear whether a decrease in FGF23 by non-calcium containing phosphate binders is related to a simultaneous decrease in vascular calcification in CKD patients not on dialysis. Therefore, long, randomized-intervention studies are required to clarify the effect of non-calcium containing phosphate binders on changes in both FGF23 and vascular calcification in CKD.
The current study has some limitations. First, the patient cohort was relatively small. Second, the multidetector CT assessment depended on the presence or absence of CAAC, but calcium content was not scored. If the CAAC score is quantified, then the associated risk factors may be more precisely documented, and the relationship between CAAC scores and FGF23 levels should be defined. Third, serum klotho levels were not examined in the present study. It has been reported that klotho deficiency contributes to soft-tissue calcification in CKD
. FGF23 generally requires its cofactor, klotho, for its activity. High levels of FGF23, as well as deficient renal klotho expression and function, are found in CKD patients. It is possible that, in the absence of klotho, increased FGF23 levels exert pro-calcific effects via non-specific low affinity binding to its receptors
[39–41]. A recent study showed that in human arteries, restoration of klotho expression by vitamin D receptor activators could unmask anticalcific effect of FGF23
. Further investigations are warranted to determine whether high FGF23 levels, independent of klotho deficiency, contribute to the development of vascular calcification in CKD patients.