Continued Search for Therapies to Favorably Modify Phosphate and FGF23 Levels in CKD

Over the last two decades, scientific developments have enhanced our understanding of the pathogenesis of cardiovascular disease in patients with CKD, and specific nontraditional risk factors have emerged. Elevated serum phosphate has been implicated in the development of vascular and valvular calcifications, microvascular disease, and endothelial dysfunction (¹). Elevated levels of fibroblast growth factor 23 (FGF23) are associated with development of left ventricular hypertrophy, atrial fibrillation, and heart failure, and preclinical data show direct effects of elevated FGF23 on cardiac myocyte hypertrophy, impaired leukocyte function, and chronic inflammation (¹). Taken together, multiple lines of evidence link phosphate and FGF23 excess to increased cardiovascular disease risk in CKD.

The excitement associated with identification of novel pathogenic mechanisms for cardiovascular disease fueled the development of phosphate- and FGF23-lowering therapies as potential therapeutic approaches to improve clinical outcomes in patients with CKD. Pilot studies testing the efficacy of phosphate and FGF23 reduction on biochemical and intermediate cardiovascular end points proliferated (²), and the 2009 Kidney Disease Improving Global Outcomes (KDIGO) clinical practice guideline recommended maintenance of serum phosphate in the normal range in individuals with CKD stages 3 and 4, suggesting that prevention of hyperphosphatemia would be beneficial. To date, completed studies have shown inconsistent results, and there are no large randomized clinical trials of phosphate binder therapy and hard clinical outcomes in CKD. Absence of a clear signal for benefit and potential for harm that emerged in one of the studies (³) led to revisions in the 2017 KDIGO guideline update, which recommended that treatment of CKD-mineral and bone disorder with phosphate binders should be reserved for individuals with sustained hyperphosphatemia (⁴).

In this issue of the Clinical Journal of the American Society of Nephrology, Liabeuf et al. (⁵) report the findings of the FGF23 Reduction Efficacy of a New Phosphate Binder in CKD (FRENCH) Trial. The FRENCH Trial was a multicenter, 12-week, double-blinded, randomized, placebo-controlled trial that tested the effects of sevelamer carbonate on FGF23 levels in patients with CKD stages 3B and 4. The authors attempted to answer an important question: can a noncalcium-containing phosphate binder, such as sevelamer carbonate, taken without dietary intervention lower FGF23 levels in patients with CKD who are normophosphatemic?

The FRENCH Trial investigators randomized 78 individuals with CKD stages 3B and 4, baseline FGF23 levels >80 RU/ml, and normophosphatemia to either 4.8 g of sevelamer carbonate (n=39) or placebo (n=39) three times daily for 12 weeks. All participants, regardless of baseline 25-hydroxyvitamin D levels, were treated with 100,000 IU of cholecalciferol at randomization. The primary outcome was change in C-terminal fibroblast growth factor 23 (cFGF23) levels. The mean adherence to sevelamer carbonate, as assessed by pill counts, was 86% in the intention to treat population. Adequate adherence is also suggested by significant decreases in urinary phosphate-to-creatinine ratio and cholesterol levels between baseline and week 12 in the sevelamer carbonate group. The authors found no difference between the sevelamer carbonate and placebo groups in changes in cFGF23 levels or their secondary outcomes: intact FGF23 levels, serum phosphate levels, parathyroid hormone levels, 1,25-dihydroxyvitamin D levels, or α-Klotho levels. Although there were no significant within-group changes in cFGF23 at the end of the study period, examination of median values over time is informative. Beginning at the second visit, cFGF23 levels rose by about 30% in both groups. Judging from the overlapping interquartile ranges, this elevation may not have reached significance, but it was sustained in both groups for the rest of the study period. There are important considerations to discuss from this study.

It is critical to understand the possible effects of vitamin D supplementation on the results of the FRENCH Trial. FGF23 and vitamin D are involved in an intricate feedback loop: FGF23 inhibits synthesis of 1,25-dihydroxyvitamin D by inhibiting expression of CYP27B1 and increases the degradation of 1,25-dihydroxyvitamin D by stimulating expression of CYP24A1. Vitamin D administration increases FGF23 transcription, and several clinical studies have reported that active vitamin D analogs and calciferols raise FGF23 levels (⁶). Given the observed rise in cFGF23 levels in both groups between randomization and the second follow-up visit, it is possible that the vitamin D supplementation that was administered to both groups at randomization is what drove the observed changes in FGF23 levels. Although data on the relationship of serial vitamin D and FGF23 levels might have further aided the interpretation of the findings, it is reasonable to conclude that vitamin D supplementation may have prevented the FRENCH Trial investigators from detecting an effect of sevelamer carbonate. It is also possible to speculate that the effect of vitamin D on the FGF23 axis is more powerful than the effect of sevelamer carbonate. Additionally, the results of the FRENCH Trial remind us that the FGF23-vitamin D-phosphate-parathyroid axis is complex, and alterations in one parameter change other parameters as well.

The FRENCH Trial investigators initially mandated that screening FGF23 levels be >120 RU/ml for inclusion in the study. This unique feature of the design had to be amended due to low recruitment in the first year, and the requirement was changed to FGF23 levels >80 RU/ml. Although the median value of FGF23 for all participants was 157 RU/ml (interquartile range, 120–241), 25% of participants had FGF23 levels <120 RU/ml. Given the severity of CKD in the patient population recruited into the FRENCH Trial as evidenced by their mean eGFR of 27±9 ml/min per 1.73 m², the median FGF23 value seems to be modestly lower than expected on the basis of existing literature. In the Chronic Renal Insufficiency Cohort (CRIC) Study, the mean eGFR was 43±14 ml/min per 1.73 m², and 72% of the population had FGF23 levels >100 RU/ml (⁷). In the subgroup of CRIC Study participants with mean eGFR between 20 and 29 ml/min per 1.73 m², median FGF23 levels were 247 (interquartile range, 169–378) (⁷). Therefore, it might be possible that the study population recruited for the FRENCH Trial had low cFGF23 levels at baseline, and perhaps, this feature reflected their reduced responsiveness to sevelamer carbonate, especially in face of exposure to vitamin D. It remains to be determined whether subpopulations of patients with CKD who have higher FGF23 levels than those of participants in the FRENCH Trial might be more amenable to FGF23 reduction with phosphate binders.

The search to identify strategies to lower phosphate and FGF23 levels in patients with CKD has led to important conclusions and new directions of investigation. Approaches focused on dietary phosphate reduction in patients with CKD who are normophosphatemic are not likely to reduce serum phosphate levels but may lower urinary phosphate excretion and FGF23 levels. We have also learned that regulation of FGF23 is dependent on multiple stimuli outside of the sphere of mineral metabolism. Recent clinical and preclinical studies have identified relationships between FGF23 and hematologic parameters of anemia, iron deficiency, and erythropoeitin (⁸). Positive results from clinical studies treating iron deficiency to lower FGF23 levels in patients with ESRD led to investigations that expanded into nondialysis CKD populations. The introduction of ferric citrate, a new iron-based phosphate binder, to our armamentarium of phosphate-lowering therapies now combines a noncalcium-based phosphate binder with an oral agent to replete iron deficiency. Encouraging data show the ability of ferric citrate to lower FGF23 levels in patients with CKD not on dialysis who are hyperphosphatemic (⁹). Although the effects may have been due to iron deficiency treatment, phosphate lowering, or a combination, these results suggest that further research is needed to understand the interplay between FGF23 regulation and the hematologic system. Inflammation and hypoxia are other recently recognized stimuli of FGF23 elevation (¹⁰,¹¹). Laboratory investigations show direct relationships between acute and chronic inflammation and increased FGF23 mRNA expression (¹⁰). Similarly, in vivo and in vitro studies support the role of hypoxia in FGF23 regulation (¹¹). Future lines of investigation hope to harness these laboratory discoveries into clinical research, with the goal of identifying novel FGF23-lowering therapies.

Although the results of the FRENCH Trial were unfavorable for FGF23 lowering by phosphate binder therapy alone, the investigators should be commended for attempting to answer critically important questions. As we continue to uncover factors that affect phosphate and FGF23 regulation, additional studies will be needed to test novel potential therapeutic approaches to lower phosphate and FGF23 levels and investigate the effect of phosphate- and FGF23-lowering therapies on intermediate end points. Ongoing clinical trials will hopefully provide new evidence that will inform our therapeutic approach to clinical management of patients with CKD who are normophosphatemic.

Disclosures

T.I. has received research support from Shire.