Association between serum fibroblast growth factor‐23 concentration and development of hyperphosphatemia in normophosphatemic dogs with chronic kidney disease

Abstract Background Fibroblast growth factor (FGF)‐23 is increased first in the sequence of changes associated with chronic kidney disease (CKD)‐mineral and bone disorder. Thus, its measurement may serve as a predictive indicator of incident hyperphosphatemia. Objectives To investigate whether serum FGF‐23 concentration in normophosphatemic dogs with CKD is associated with the risk of the subsequent development of hyperphosphatemia and CKD progression. Animals Forty‐two normophosphatemic dogs with CKD. Methods Blood samples and medical records were retrospectively investigated. Hyperphosphatemia was defined as a serum phosphorous concentration >5.0 mg/dL. Progression was defined as a >1.5‐fold increase in serum creatinine concentration. The time periods and hazard ratios for these outcomes were assessed using Kaplan‐Meier analysis, log‐rank test, and univariate Cox regression analysis. The variables associated with the outcomes in the univariate analysis were included in the multivariate Cox regression model with backward selection. Results Serum FGF‐23 concentration >528 pg/mL was associated with a shorter time to development of hyperphosphatemia (P < .001) and CKD progression (P < .001). In multiple Cox regression analysis, increased FGF‐23 concentration remained a significant variable associated with these outcomes (P < .05). Conclusions and Clinical Importance Increased FGF‐23 concentration in normophosphatemic dogs with CKD was associated with significant risk of development of hyperphosphatemia, independent of CKD stage, and of the progression of CKD. Future research focusing on whether interventions that decrease FGF‐23 secretion will slow the development of hyperphosphatemia and CKD progression is needed.


| INTRODUCTION
Chronic kidney disease (CKD) is a common, irreversible, and progressive disease in dogs. 1 In patients with CKD, mineral metabolism disorders such as hyperphosphatemia, renal hyperparathyroidism, and decreased calcitriol synthesis occur because of impaired renal function and are associated with a poor prognosis. [2][3][4] Therefore, these disorders are considered important complications and are termed CKD-mineral and bone disorder (MBD). [2][3][4][5] Hyperphosphatemia is known to be a prognostic factor for shorter survival time in dogs with CKD. [6][7][8][9][10] Therefore, prevention and management of hyperphosphatemia may improve the prognosis of dogs with CKD.
Fibroblast growth factor (FGF)-23 is a phosphaturic hormone associated with CKD-MBD. 2,3,5 Fibroblast growth factor-23 is released from osteocytes in response to increased serum phosphorus and calcitriol concentrations and promotes phosphate excretion into the urine by downregulation of the sodium-phosphate co-transporter in renal proximal tubular cells and inhibition of calcitriol synthesis. 11,12 Fibroblast growth factor-23 acts by binding to the FGF receptorα-klotho complex. 13 In humans, cats, and dogs with CKD, circulating FGF-23 concentrations have been shown to increase in the advanced CKD stages. [14][15][16][17][18][19] Increased FGF-23 concentration in patients with CKD is associated with various mechanisms, such as a decreased clearance of FGF-23 because of decreasing glomerular filtration rate (GFR), compensation for the accumulation of phosphate in the body, and compensation for decreased klotho protein concentrations. [20][21][22] Studies of humans and cats with CKD have reported that concentrations of FGF-23 increased earlier than did those of parathyroid hormone (PTH) and phosphorus, [15][16][17]23 and were associated with shorter survival time. [24][25][26] Therefore, FGF-23 has been noted as an early marker of CKD-MBD. 2 Similarly, blood FGF-23 concentrations in dogs with CKD have been shown to become increased earlier than serum phosphorous concentrations. 18,19 Therefore, increased blood FGF-23 concentrations plausibly can reflect phosphate accumulation in dogs with CKD but without overt hyperphosphatemia. Thus, we hypothesized that FGF-23 predicts development of hyperphosphatemia in normophosphatemic dogs with CKD. If correct, FGF-23 may be a useful marker of when to initiate a phosphate-restricted diet to prevent the development of hyperphosphatemia.
Our purpose was to investigate the relationship between FGF-23 concentration and subsequent development of hyperphosphatemia and progression of CKD and determine the clinical relevance of increased FGF-23 concentration in normophosphatemic dogs with CKD.

| Case selection
Ours was a retrospective study performed using medical records and stored serum samples that were collected between September 2014 and September 2020 from client-owned dogs diagnosed with CKD at the nephrology service of a veterinary medical teaching hospital in Japan. Dogs with CKD were selected from a database using commercially available software (Microsoft Excel, Microsoft Japan Co., Ltd., Tokyo, Japan) and the search terms "CKD" and "dog." Then, the medical records and presence or absence of serum samples for the dogs identified from the database were reviewed. If serum samples at the time of CKD diagnosis were unavailable, samples obtained at a subsequent hospital visit were selected for measurement of serum FGF-23 and intact PTH concentrations. A diagnosis of CKD was based on persistently (≥3 months) having at least 1 of the following criteria: renal azotemia, renal proteinuria, or renal abnormalities on abdominal ultrasonography. Renal azotemia was defined as a serum creatinine concentration >1.45 mg/dL (the upper limit of reference range at our institution) and urine specific gravity (USG) <1.030. Renal proteinuria was assessed based on a urine protein/creatinine ratio (UPC) ≥0.5 without an apparent prerenal or postrenal cause. Ultrasonographic renal abnormalities included small kidneys, decreased corticomedullary differentiation, irregular renal contours, or some combination of these findings. Dogs with CKD were classified according to the 2019 International Renal Interest Society (IRIS) CKD guidelines for staging based on serum creatinine concentrations as stage 1 (<1.4 mg/dL), stage 2 (1.4-2.8 mg/dL), stage 3 (2.9-5.0 mg/dL), and stage 4 (>5.0 mg/dL). 27 All dogs with stage 1 CKD had renal proteinuria, ultrasonographic renal abnormalities, or both.
Dogs suspected or diagnosed with neoplasia, hyperadrenocorticism, hypoadrenocorticism, acute kidney injury or acute CKD exacerbation, or having a history of hyperphosphatemia were excluded from the study.
Hyperphosphatemia was defined as a serum phosphorous concentration >5.0 mg/dL, based on the reference interval at our institution. In addition, dogs for which follow-up information could not be obtained were excluded.

| Medical records review and data collection
Information regarding signalment, serum biochemical analysis, blood gas analysis, urinalysis, noninvasively determined systolic blood pressure (SBP) and treatment were collected from the medical records.
Serum biochemical analyses were performed using an automated analyzer (7180 Biochemistry Automatic Analyzer, Hitachi High-Technologies Corp., Tokyo, Japan). Blood ionized calcium and bicarbonate concentrations were measured using a blood gas analyzer (GEM PREMIER 3500, Instruments Laboratory Co., Ltd, Tokyo, Japan). owners. Blood samples in tubes containing serum separators were centrifuged at 1181g for 5 minutes, and the obtained serum samples were stored at À30 C for submission to an external laboratory (FUJIFILM VET Systems Co., Ltd, Tokyo, Japan) for FGF-23 and intact PTH analysis. Serum FGF-23 concentrations were measured using a sandwich ELISA kit (MedFrontier FGF-23, Hitachi Chemical Diagnostics Systems Co., Ltd, Tokyo, Japan). Serum intact PTH concentrations were analyzed using a chemiluminescent enzyme immunoassay (Siemens Immulyze intact PTH III, Siemens Healthcare Diagnostics Co., Ltd, Tokyo, Japan). Validation of the FGF-23 and intact PTH assays was reported in a previous study. 19  Dogs that did not develop hyperphosphatemia did not experience progression at the end of the follow-up period and those that were lost to follow-up were censored.

| Statistical analysis
Kaplan-Meier curves were used to plot the time to development of hyperphosphatemia and CKD progression. The cutoffs for FGF-23 and intact PTH were based on the reference ranges (528 and 8.5 pg/mL, respectively) reported in a previous study. 19 The cutoffs for serum biochemical variables other than serum phosphorous concentration were based on the reference ranges at our institution. The cutoff for serum phosphorous concentration (4.6 mg/dL) was set according to the therapeutic aim proposed by the IRIS CKD guidelines. 30 The cutoffs for bicarbonate and ionized calcium concentrations were determined based on a previous report. 31 Figure 2A). International Renal Interest Society stage also was significantly associated with the interval to outcome. Median durations (95% CI) to the outcome in stages 1, 2, and 3-4 groups were 581 days (185-977 days), 385 days (0-871 days), and 35 days (0-240 days), respectively ( Figure 2B).

The univariate Cox regression analysis showed that FGF-23
>528 pg/mL, UN >29.2 mg/dL, and CKD stages 3-4 were significantly associated with risk for development of hyperphosphatemia (P < .001, P = .003, and P = .01, respectively; Table 2). In addition, time between re-examination visits correlated with significant risk for outcome (    Moreover, our results suggest that the period before phosphate accumulation becomes apparent as hyperphosphatemia was shorter in dogs with increased FGF-23 concentrations. Another mechanism that contributes to increased blood FGF-23 concentrations in CKD is decreased renal clearance itself because of decreased GFR. 20 In humans, impaired GFR was the most important factor contributing to increased FGF-23 concentrations in early stages of CKD. 35 However, multivariate Cox regression analysis in our study showed that increased FGF-23 remained a significant risk factor for development of hyperphosphatemia after adjusting for CKD stage (ie, magnitude of GFR). This finding suggests that phosphate accumulation in the body is an important factor involved in increased FGF-23 concentrations, independent of GFR in dogs with CKD.

| Relationship between FGF-23 and progression of CKD
In addition, in our study, serum FGF-23 concentration was significantly associated with the progression of CKD in dogs. In humans with CKD, increased FGF-23 concentrations are related to shorter survival time. 24,25 One of the reasons for this phenomenon is that increased blood FGF-23 concentrations increase the risk of cardiovascular comorbidities, such as myocardial infarction and heart failure in humans. 25,36,37 In dogs and cats with CKD, increased plasma FGF-23 concentrations also have been associated with shorter survival times. 9,26 However, myocardial infarction is a rare event in dogs and cats, and the relationship between FGF-23 in CKD and heart failure has not been determined. Thus, the association between FGF-23 and worse prognosis in dogs and cats with CKD cannot be explained by cardiovascular comorbidities. Another mechanism by which increased FGF-23 may shorten survival time is that FGF-23 is related to the progression of CKD in humans. [37][38][39] Similar results also were observed in cats with CKD. 26,40 Our results are consistent with those of previous studies in humans and cats, but it is unknown whether FGF-23 directly promotes CKD progression. In rats and beagles, phosphate loading was shown to lead to renal toxicity by calcification of the kidneys. [41][42][43] Given the results of these studies, it is thought that the accumulation of phosphate in the body contributes to the progression of CKD by renal calcification. Therefore, increased blood FGF-23 concentrations, which reflect phosphate accumulation, could be associated with progression of CKD and shorter survival time in dogs and cats with CKD. In previous studies, the prevalence of hyperphosphatemia in dogs with CKD was 0% to 18% in stage 1, 19% to 50% in stage 2, 50% to 90% in stage 3, and 100% in stage 4. 19,44,45 Thus, the prevalence of hyperphosphatemia is associated with CKD stage and GFR. Moreover, our study was retrospective and included dogs that were already being treated for proteinuria, systemic hypertension or both, most of which were treated with telmisartan. The efficacy of telmisartan on proteinuria and blood pressure recently was found to be superior to that of angiotensin converting enzyme inhibitors. 56,57 Therefore, telmisartan could have influenced the effects of proteinuria and systemic hypertension on the progression of CKD. 33,34,55 Our study had some limitations. First, the study was a retrospective investigation, and it may have been subject to selection bias.
Some data sets had missing values, and the frequency of reassessment and therapeutic regimens such as dietary treatment and renin-angiotensin system inhibitors were not standardized. Because time between re-examination visits was significantly associated with the risk for outcomes, dogs with the development of hyperphosphatemia or CKD progression could be rigorously followed up and managed. In addition, follow-up periods of some dogs with FGF-23 <528 pg/mL and censored were not adequate to completely overcome these limitations. Second, sample size was small. Third, we did not consider the underlying causes of CKD in the dogs. Our study included various underlying diseases, such as protein-losing nephropathy, renal dysplasia and other tubulointerstitial diseases. Because the progression rate of CKD and treatment strategies may differ based on the underlying disease, this factor could have affected the results of our study. Fourth, the upper limit of the reference range of FGF-23 used in our study was determined using a small sample of healthy dogs. 19 Therefore, a more precise cut-off for the FGF-23 concentration that is associated with the risk of the development of hyperphosphatemia and CKD progression is needed. Finally, our study did not consider the storage period of serum samples used for the measurement of FGF-23 and intact PTH. Although storage stability in dogs was investigated for up to 28 days, 19 the storage period of samples used in our study ranged from 3 months to 5 years. The storage period of the serum samples used may have affected the FGF-23 and intact PTH concentrations obtained.
In conclusion, increased serum FGF-23 concentration was a significant risk factor for the subsequent development of hyperphosphatemia in normophosphatemic dogs with CKD. In addition, serum FGF-23 concentration was associated with CKD progression in the affected dogs.
Future research that investigates whether a decrease in FGF-23 concentrations prevents or delays the onset of hyperphosphatemia and associated progression is needed.

ACKNOWLEDGMENT
No funding was received for this study. We thank FUJIFILM VET Systems Co., Ltd for the analysis of FGF-23 and intact PTH. In addition, the authors thank Enago (https://www.enago.jp/) for English language review.

CONFLICT OF INTEREST DECLARATION
Yuichi Miyagawa is in receipt of speaker honoraria from FUJIFILM VET Systems Co., Ltd. The other authors declare no potential conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION
Authors declare no off-label use of antimicrobials.