Fibroblast Growth Factor 23 (FGF23) Predicts Progression of Chronic Kidney Disease: The Mild to Moderate Kidney Disease (MMKD) Study
Danilo Fliser*,
Barbara Kollerits,
Ulrich Neyer,
Donna P. Ankerst,,
Karl Lhotta||,
Arno Lingenhel,
Eberhard Ritz¶,
Florian Kronenberg for the MMKD Study Group
* Department of Internal Medicine, Hannover Medical School, Hannover, Germany; Division of Genetic Epidemiology; Department of Medical Genetics, Molecular and Clinical Pharmacology, Innsbruck Medical University, Innsbruck, Austria; Department of Nephrology and Dialysis, Academic Teaching Hospital, Feldkirch, Austria; Department of Urology, University of Texas Health Science Center at San Antonio, San Antonio, Texas; || Department of Clinical Nephrology, Innsbruck University Hospital, Innsbruck, Austria; and ¶ Department of Internal Medicine, Division of Nephrology, Ruperto-Carola-University, Heidelberg, Germany
Correspondence: Dr. Danilo Fliser, Department of Internal Medicine, Hannover Medical School, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany. Phone: +49-511-532-6319; Fax: +49-511-552366; E-mail: fliser.danilo{at}mh-hannover.de
Received for publication August 29, 2006.
Accepted for publication May 22, 2007.
It has not been firmly established whether disturbed calcium-phosphatemetabolism affects progression of chronic kidney disease (CKD)in humans. In this cohort study of 227 nondiabetic patientswith CKD, we assessed fibroblast growth factor 23 (FGF23) plasmaconcentrations in addition to other variables involved in calcium-phosphatemetabolism, and we followed 177 of the patients prospectivelyfor a median of 53 months to assess progression of renal disease.In the baseline cohort, we found a significant inverse correlationbetween glomerular filtration rate and both c-terminal and intactFGF23 levels (both P < 0.001). The 65 patients who experienceda doubling of serum creatinine and/or terminal renal failurewere significantly older, had a significantly lower glomerularfiltration rate at baseline, and significantly higher levelsof intact parathormone, c-terminal and intact FGF23, and serumphosphate (all P < 0.001). Cox regression analysis revealedthat both c-terminal and intact FGF23 independently predictprogression of CKD after adjustment for age, gender, GFR, proteinuria,and serum levels of calcium, phosphate, and parathyroid hormone.The mean follow-up time to a progression end point was 46.9(95% CI 40.2 to 53.6) months versus 72.5 (95% CI 67.7 to 77.3)months for patients with c-terminal FGF23 levels above or belowthe optimal cut-off level of 104 rU/mL (derived by receiveroperator curve analysis), respectively. In conclusion, FGF23is a novel independent predictor of progression of renal diseasein patients with nondiabetic CKD. Its pathophysiological significanceremains to be elucidated.
Disturbed calcium-phosphate metabolism affects cardiovascularmorbidity and mortality in patients with chronic kidney disease(CKD), particularly in patients with ESRD.1,2 So far, it hasnot been firmly established whether it is also related to CKDprogression. Among factors that are related to calcium-phosphatemetabolism in patients with CKD, potential culprits for progressionare hyperphosphatemia, hyperparathyroidism, lack of active vitaminD, and possibly excess of the recently discovered phosphaturichormone fibroblast growth factor 23 (FGF23).3,4 Early experimentalwork suggested a parathyroid hormone (PTH)-independent beneficialrole of phosphate restriction on progression in rats,5 but ithas to be pointed out that these animals have physiologic hyperphosphatemia.Furthermore, there is also little direct experimental or clinicalevidence for a role of PTH in accelerating progression,3 althoughresults from recent experimental studies documented that progressionis significantly attenuated by administration of calcimimeticsor by parathyroidectomy.6 However, a confounding effect of lowerBP values in these experimental settings cannot be excluded.The most solid evidence from experimental studies for an effecton progression exists for active vitamin D (1,25-OH2D3). Althoughin the past it was assumed that vitamin D therapy is "nephrotoxic,"probably as a result of vitamin D–induced hypercalcemiain patients with CKD,7 recent experimental evidence clearlyrevealed that 1,25-OH2D3 and its analogues attenuate progressionin various CKD models.3,8,9 Finally, the role of FGF23 in CKDprogression is unknown.
We assessed various parameters of calcium-phosphate metabolismincluding c-terminal and biologically active intact FGF23 plasmaconcentrations in 227 patients who did not have diabetes andhad primary CKD, 177 of whom were prospectively followed fora median of 53 mo. We examined the working hypothesis that FGF23is a predictor of CKD progression.
Stages of CKD and Calcium-Phosphate Metabolism
Baseline clinical characteristics and laboratory data of 227patients with CKD are reported in Table 1. To elucidate therelationship between renal function and parameters of calcium-phosphatemetabolism, we stratified renal patients into four groups accordingto National Kidney Foundation (NKF) criteria for CKD: GFR 90ml/min per 1.73 m2, GFR 60 to 89 ml/min per 1.73 m2, GFR 30to 59 ml/min per 1.73 m2, and GFR <30 ml/min per 1.73 m2(Table 1). We found a continuous and significant increase ofCa x P, PTH, and both c-terminal and intact FGF23 concentrationacross the various NKF stages of renal dysfunction. In addition,serum phosphate was significantly higher in patients with moreadvanced renal failure. We found significant correlations betweenc-terminal FGF23 as well as intact FGF23 with GFR, phosphate,and PTH (Figure 1). When we analyzed the data of patients withonly CKD stages 2 through 5, we found similar results as forthe entire population (data not shown).
Figure 1. Correlation analysis of c-terminal (A through C) as well as intact fibroblast growth factor 23 (FGF23; D through F) with GFR, serum phosphate, and parathormone. All parameters have been analyzed based on the log scale.
Calcium-Phosphate Metabolism and Progression of CKD
Clinical characteristics and laboratory data of patients withfollow-up are reported in Table 2. The median follow-up aftercompletion of the baseline investigation was 53 mo (range 3to 84 mo), and during this follow-up, 65 patients had progressedto a renal end point (33 patients had doubled their serum creatinine,and 29 had reached terminal renal failure necessitating renalreplacement therapy; the two groups differed with respect tobaseline serum creatinine, GFR, c-terminal FGF23, Ca x P, andthe use of phosphate binders). Patients who had reached a progressionend point were significantly older and had higher protein excretionrates and lower GFR. In addition, they had significantly higherphosphate, PTH, and FGF23 levels and Ca x P. There were no differencesfor surrogate parameters of nutritional (body mass index, albumin)and inflammatory status (high-sensitivity C-reactive protein[hsCRP]).
Table 2. Baseline clinical and laboratory data of 177 patients with completed follow-up with further stratification of those with and without progression during the follow-up perioda
Age- and gender-adjusted Cox regression analysis revealed thatGFR, c-terminal and intact FGF23, phosphate, Ca x P, and PTHall showed a strong association with progression-free survival(all P < 0.001; Table 3, model 1), and proteinuria showeda weaker association (P = 0.03). In a multivariable analysisadjusting for all variables in addition to age and gender, onlybaseline GFR (P < 0.001), c-terminal FGF23 (P = 0.002), andintact FGF23 (P = 0.005) remained significant predictors forprogression, and serum calcium, phosphate, and PTH concentrationswere not independently associated with disease progression (Table 3,model 2). When we used only parameters of the calcium-phosphatemetabolism without FGF23 levels in a third model, we observedPTH and Ca x P besides baseline GFR to predict CKD progression.Phosphate concentrations were of borderline significance (P= 0.063), and calcium was not associated with CKD progression(Table 3, model 3). To test the goodness of fit of the modelsto the observed data, we also analyzed models with c-terminaland intact FGF23 concentrations that were transformed usingthe log base 2 logarithm (log2 c-terminal FGF23 or log2 intactFGF23). We found that both log-transformed parameters remainedsignificant, and results were comparable to models interpretinghazard ratios for an increment of 10 units in FGF23 concentration(data not shown).
Table 3. Association of baseline variables with progression of kidney disease during the observation period using multiple Cox proportional hazards regression modelsa
To evaluate the operating characteristics of FGF23 as a prognostictool for the progression of kidney disease, we performed a receiveroperating characteristic (ROC) analysis for c-terminal and intactFGF23 in comparison with GFR (Figure 2). GFR obtained the highestarea under the curve (AUC; AUC = 0.84; 95% confidence interval[CI] 0.78 to 0.90; P < 0.001), followed by c-terminal FGF23(AUC = 0.81; 95% CI 0.74 to 0.88; P < 0.001) and intact FGF23(AUC = 0.72; 95% CI 0.64 to 0.80; P < 0.001). We furtherconstructed Kaplan-Meier curves of progression-free survivalcomparing patients with c-terminal FGF23 values above and belowthe optimal cutoff level of 104 rU/ml and for intact FGF23 valuesabove and below the median of 35 pg/ml (a value very close tothe optimal cutoff derived by ROC analysis). Patients who hadc-terminal and intact FGF23 levels above these threshold valueshad a worse prognosis and statistically significant shorterprogression times compared with patients with values below thethreshold (Figure 3): Mean time to progression 46.9 mo (95%CI 40.2 to 53.6) compared with 72.5 mo (95% CI 67.7 to 77.3;P < 0.0001), respectively, for c-terminal FGF23 and 54.6mo (95% CI 48.2 to 61.0) compared with 69.8 mo (95% CI 64.8to 74.8; P = 0.0004), respectively, for intact FGF23. We observedsimilar separation of survival curves when we used the medianof c-terminal FGF23 (85 rU/ml) instead of the ROC-derived cutoffof 104 rU/ml (data not shown).
Figure 2. Receiver operating characteristics (ROC) curve of GFR and plasma c-terminal and intact FGF23 concentrations with progression of kidney disease as status variable. The area under the curve (AUC) for GFR, c-terminal, and intact FGF23 are 0.84, 0.81, and 0.72, respectively. The inserted tables present sensitivity and specificity of the four cutoffs derived from the quintile thresholds for c-terminal and intact FGF23, respectively. For c-terminal FGF23, the threshold between the third and fourth quintiles was equal to the ROC-derived optimal cutoff. For intact FGF23, the optimal cutoff is presented additionally.
Figure 3. Kaplan-Meier curves of renal end points in patients with below and above optimal cutoff of plasma c-terminal (A) FGF23 concentrations and for intact (B) FGF23 concentrations below and above the median. In patients with c-terminal FGF23 levels above the optimal cutoff (>104 rU/ml), progression was significantly faster (log-rank test, P < 0.0001); the same was observed in patients with intact FGF23 levels above the median (35 pg/ml; log-rank test, P = 0.0004). Numbers near the survival curves represent the number of patients at risk with FGF23 levels below and above the optimal cutoff at the times 0, 12, 24, 36, 48, 60, and 72 mo.
The results of this prospective long-term study in a sizablecohort of white patients with nondiabetic CKD have identifiedFGF23 as a novel risk marker for the progression of CKD. Remarkably,apart from baseline GFR, FGF23 was the only independent predictorof progression among several parameters of the calcium-phosphatemetabolism assessed. Moreover, both the c-terminal fragmentand the biologically active intact FGF23 were independent predictorsof progression. This finding is of particular interest becauseit points to a role of FGF23 in CKD progression apart from theproblem of accumulation as a result of reduced renal clearance,which could be one potential explanation for elevated concentrationsof c-terminal FGF23.
FGF23 is a recently identified "phosphatonin" that is thoughtto be implicated in the systemic balance of phosphate maintainedby the interaction of intestine, bone, and kidneys.9–11In several clinical conditions, excessive activity of FGF23resulted in hypophosphatemia, low 1,25-OH2D3 levels, and osteomalacia.12,13Administration of recombinant FGF23 to experimental animalsor overexpression of the FGF23 gene in vivo produced similarderangements of calcium-phosphate metabolism, whereas inactivationof this gene caused hyperphosphatemia and high circulating 1,25-OH2D3levels.14–16 The physiologic stimulus for FGF23 secretionseems to be hyperphosphatemia caused by a dietary phosphateload.17 The increase in FGF23 levels in response to dietaryphosphate promotes phosphaturia and suppresses renal productionof active vitamin D. However, in the presence of progressiveCKD, serum FGF23 levels increase in parallel with the deteriorationof renal function and the increase of serum phosphate and PTHconcentrations.9,18–21 In predialysis patients and inpatients who were on maintenance hemodialysis, high FGF23 serumlevels were correlated with those of phosphate, pointing toa disrupted feedback loop resulting in very high levels of serumFGF23. We could clearly confirm these findings in our studypopulation with mild to moderate CKD, in which the decreaseof renal function across a wide range of GFR was paralleledby a significant and marked increase particularly of c-terminalbut also of intact FGF23 blood concentrations (Table 1, Figure 1).However, because we measured only baseline serum phosphate,calcium, PTH, and FGF23 levels in our cross-sectional study,we cannot distinguish whether the hormone levels are the causeof abnormal serum calcium and phosphate levels or their consequence.
In experimental studies, the increase of FGF23 levels precededthe decrease of serum 1,25-OH2D3 concentrations, suggestingan important role of FGF23 in the development of secondary hyperparathyroidismof patients with CKD. This point is further corroborated byresults of recent experimental22 and clinical studies.21 Forexample, Gutierrez et al.21 found an increase in FGF23 serumlevels in early stages of CKD, even before serum phosphate andcalcium concentrations had become abnormal. The authors concludedthat increased FGF23 levels are presumably a central factorin the early pathogenesis of secondary hyperparathyroidism.Collectively, these findings implicate that circulating FGF23is a physiologic regulator of phosphate balance and as suchalso a potential uremic toxin.10,11 An alternative interpretationof our findings is that FGF23 levels represent the overall burdenof phosphate loading above and beyond that represented by asingle measurement of fasting serum phosphate. At any rate,FGF23 is an excellent indicator of the complex derangementsof calcium-phosphate metabolism induced by CKD and probablyalso a valuable surrogate parameter to indicate more distalsequelae of the deranged mineral metabolism. This may be clinicallyimportant: It has been known for decades that in ESRD, hyperphosphatemiacauses soft tissue calcification including vascular calcification.23However, until recently, its impact on survival in patientswith CKD had not been appreciated. In 1998 Block et al.24 foundthat in dialysis patients, survival was significantly less whenthe predialysis serum phosphate concentration exceeded 6.5 mg/dl.This increase was related to death from coronary heart disease,possibly as a result of accelerated coronary plaque calcification.25Phosphate—more specific, intracellular phosphate—playsa major role in the genesis of vascular calcification, particularlyin the presence of ionized calcium.26 However, the adverse roleof high serum phosphate concentrations is not restricted topatients with CKD,27 and even in nonrenal patients, serum phosphateconcentrations were positively and significantly correlatedwith the severity of coronary artery disease and with the severityof coronary artery stenoses and presence of occlusions.28 Ina post hoc analysis of the Cholesterol and Recurrent Events(CARE) study in 4127 patients, Tonelli et al.29 found that serumphosphate levels (even within the upper normal range) were associatedwith more adverse cardiovascular outcomes. They also found adirect association between GFR and serum phosphate concentrations,but the relation between phosphate and outcome still persistedwhen individuals with a GFR <60 ml/min were excluded fromthe analysis. It is interesting that the calcium-phosphate productwas not independently associated with adverse outcome.29 Takentogether, these data suggest that even minor derangements ofcalcium-phosphate metabolism, particularly of serum phosphatelevels, may contribute to cardiovascular complications in nonrenalpatients as well as in patients with CKD.
This study points to an additional potential consequence, namelythat serum phosphate levels have an impact on progression ofrenal disease. The experimental work of Haut and Alfrey andcolleagues30,31 showed that renal damage was aggravated by phosphateloads. To our knowledge, our observation is the first largeprospective study to investigate the influence of changes incalcium-phosphate metabolism and the role of FGF23 on progressionof CKD. Given the physiologic role of FGF23 in phosphate metabolism,this "phosphatonin" may turn out to be an excellent indicatorof cardiovascular risk even in patients without CKD. Furtherstudies on this issue are warranted.
The cohort investigated in our study comprised relatively youngpatients with mostly mild to moderate impairment of renal functionor even normal GFR at the baseline examination. The remarkablepower of baseline GFR as a predictor of progression highlightsthe importance of impaired renal function as a key determinantof progressive renal damage. In addition, this finding may explainwhy more patients in the "progressor" subgroup received vitaminD and phosphate binders, because derangements of calcium-phosphatemetabolism become more evident with advanced CKD. In our patients,FGF23 levels increased with decreasing GFR, but it is importantto point out that the impressive prediction of progression byboth c-terminal and intact FGF23 is already adjusted for GFR.We have to admit that the measurement of GFR even with iohexolclearance may not be perfect and also does not reflect all aspectsof kidney dysfunction. Therefore, we cannot exclude the possibilitythat the prediction of progression by intact FGF23 might alsoresult from commutation of biologically inert degradation productsreflecting residual confounding by severity of kidney function.However, the estimates of CKD progression by FGF23 were stillhighly significant after adjustment for GFR. This finding suggeststhat FGF23 is not simply a surrogate marker for baseline GFRand that FGF23 is a marker on its own providing prediction beyondthe information obtained by the measurement of GFR.
In our cohort, systolic and diastolic BP at baseline were comparablein patients who progressed to a renal end point during follow-upand those who did not. The former needed more aggressive antihypertensivetreatment, however, to achieve the same level of BP control.We emphasize that the use of angiotensin-converting enzyme inhibitorswas comparable between groups. It is likely that identificationof further progression predictors and potentially progressionpromoters such as FGF23 was facilitated by the almost equalBP control in progressors and nonprogressors.
The exclusion criteria in this study yielded a selected groupof patients. Further studies must show whether the associationbetween FGF23 and progression can be found also in other typesof CKD, such as diabetic nephropathy, or nephrotic forms ofkidney disease. The study had a follow-up time of adequate duration,and the primary end points were reached by one third of theparticipants. Therefore, the data are sufficiently solid todraw firm conclusions for the entire cohort, despite a considerableloss during follow-up of participants who had superior kidneyfunction at baseline when compared with the followed patients.An additional limitation is the lack of data on phosphate intakein our patients.
In this prospective study comprising a cohort of patients withprimary CKD, we identified intact FGF23 and, additionally, the-easier-to-measurec-terminal FGF23 fragment as novel predictors of CKD progression.Given the physiologic role of FGF23 in calcium-phosphate metabolism,it may also serve as a clinically useful indicator of earlyalterations of calcium-phosphate metabolism, correction of whichmight modify progression of CKD31 and possibly also cardiovascularrisk.29
Patients and Baseline Investigations
We examined at baseline 227 white patients who were between18 and 65 yr of age and had nondiabetic CKD and various degreesof renal impairment. These patients were recruited from eightnephrology departments in Germany, Austria, and South Tyrolas described previously.32 The study was approved by the institutionalethic committees, and all participants gave written informedconsent. They had stable renal function for at least 3 mo beforeentry into the study. Exclusion criteria were treatment withimmunosuppressive agents, fish oil, or erythropoietin; serumcreatinine >6 mg/dl; diabetes of any type; malignancy; liver,thyroid, or infectious disease; nephrotic syndrome (definedas proteinuria >3.5 g/1.73 m2 per d); organ transplantation;allergy to ionic contrast media; and pregnancy. According tothe NKF classification of CKD, our study cohort showed the followingstages of CKD: GFR 90 ml/min per 1.73 m2 (stage 1) in 72 (31.7%)patients, GFR 60 to 89 ml/min per 1.73 m2 (stage 2) in 49 (21.6%)patients, GFR 30 to 59 ml/min per 1.73 m2 (stage 3) in 63 (27.8%)patients, and GFR <30 ml/min per 1.73 m2 (stages 4 and 5)in 43 (18.9%) patients. The primary cause of kidney diseasewas glomerulonephritis in 97 (biopsy-confirmed in 90) patients,adult polycystic kidney disease in 37 patients, interstitialnephritis in 24 patients, other types of kidney disease in 43patients, and unknown in 26 patients.
For avoiding interobserver differences, all patients were recruitedby one physician who visited all participating centers. Patienthistory, including smoking habits and antihypertensive treatmentat baseline, was recorded by interview and confirmed by checkingpatient records. This was complemented by clinical examination,including assessment of body mass index and BP. Hypertensionwas defined as BP >140/90 mmHg and/or the use of antihypertensivemedication. Antihypertensive medication was withheld on theday of the study to minimize interference with measurementsof the GFR. Antihypertensive drugs were taken by 179 (79%) patients:Diuretics (n = 83; 37%), angiotensin-converting enzyme inhibitors(n = 123; 54%), calcium channel blockers (n = 78; 34%), receptorblockers (n = 67; 30%), and -1 receptor blockers (n = 36; 16%).
Prospective Follow-Up
After the baseline investigation, patients were followed pro-spectivelyuntil the primary study end point or the end of the observationperiod was reached. The primary end point was defined as doublingof baseline serum creatinine and/or terminal renal failure necessitatingrenal replacement therapy. A total of 177 (78%) patients fromthe baseline cohort could be assessed during the follow-up.Patients who were lost to follow-up (n = 50) had at baselinea significantly better renal function than patients who werenot lost for follow-up (i.e., a higher mean GFR [91 ±44 versus 64 ± 39 ml/min per 1.73 m2; P < 0.01]).However, both groups did not differ significantly with respectto age and gender. Patients who were lost to follow-up had movedaway or were not referred by their physicians for follow-upvisits in the renal units.
Laboratory Measurements
Blood samples for measurement of routine chemistry, hsCRP, PTH,and c-terminal and intact FGF23 were taken after an overnightfast of at least 12 h. The samples were immediately centrifugedat 1500 x g and 4°C for 10 min, and the supernatants werestored in aliquots at –80°C until further use. PTHwas measured with an immunoradiometric assay, and FGF23 wasmeasured using both the human c-terminal and the intact ELISA(Immutopics, San Clemente, CA).10 The interassay coefficientsof variability of c-terminal FGF23 are 6.5% at 40 rU/ml and7.5% at 175 rU/ml, respectively, with a lower detection limitof 3.0 rU/ml. For intact FGF23, the interassay coefficientsof variability are 6.1% at 15.6 pg/ml and 6.5% at 166 pg/ml,respectively, with a lower detection limit of 1.0 pg/ml. Measurementsof routine chemistry including hsCRP were performed with routinelaboratory tests. In addition, GFR was assessed in all patientsusing the iohexol clearance technique as described in detailelsewhere.33
Statistical Analyses
Statistical analysis was performed with the SPSS for Windows12.01 (SPSS, Chicago, IL). Univariate comparisons of continuousvariables between various groups were performed using an unpairedt test or the nonparametric Wilcoxon rank sum test in case ofnon-normally distributed variables. Dichotomized variables werecompared using Pearson 2 test. All statistical tests were performedat the two-sided 0.05 level of significance. Data are presentedas means ± SD and as median and 25th and 75th percentilesfor skewed variables where appropriate. Univariate correlationanalysis was performed by Spearman correlation analysis. Forcalculation of ROC curves for the predictor variables c-terminaland intact FGF23, patients were stratified into two groups (thosewho progressed and those who did not over the study period),and for each cutoff value c, sensitivity was defined as theproportion of progressors with predictor variables above c andspecificity as the proportion of nonprogressors with valuesbelow c. The optimal cutoff value for the predictor variableswas defined as the cutoff obtaining the highest sum of sensitivityand specificity. Calculation of the AUC and a test of the nullhypothesis that it equals 50% (equivalent to a random 50:50prognosis) were performed using the Mann-Whitney test. Kaplan-Meiertime-to-event curves were generated for patients with c-terminaland intact FGF23 concentrations above and below optimal cutofflevels. Adjusted risk estimates for progression end points werecalculated using multivariable Cox proportional hazards regressionanalysis. Because Ca x P is the product term of the variablecalcium and phosphate, estimates of Ca x P were based on a modelthat did not include calcium and phosphate at the same time,to avoid multicollinearity. Similarly, intact and c-terminalFGF23 were not offered at the same time in a particular model.Exploratory graphical analysis and tests of specific violationsindicated no departure from the assumption of proportional hazards.
Parts of this work were supported by the "Genomics of Lipid-associatedDisorders—GOLD" of the Austrian Genome Research ProgrammeGEN-AU and by grants from the Austrian Nationalbank (Project9331) and the Austrian Heart Fund to F.K.
The following members of the Mild and Moderate Kidney Disease(MMKD) Study Group collaborated with the authors of this project:Erich Kuen, Division of Genetic Epidemiology, Innsbruck MedicalUniversity (Innsbruck, Austria); Paul König, InnsbruckUniversity Hospital (Innsbruck, Austria); Günter Kraatz,Ernst Moritz Arndt University (Greifswald, Germany); JohannesF.E. Mann, München Schwabing Hospital (Munich, Germany);Gerhard A. Müller, Georg August University (Göttingen,Germany); Ulrich Neyer, Feldkirch Hospital (Feldkirch, Austria);Hans Köhler, Medizinische Universitätskliniken desSaarlandes (Homburg/Saar, Germany); and Peter Riegler, BozenHospital (Bozen, Italy).
Footnotes
Published online ahead of print. Publication date availableat www.jasn.org.
Block GA, Klassen PS, Lazarus JM, Ofsthun N, Lowrie EG, Chertow GM: Mineral metabolism, mortality, and morbidity in maintenance hemodialysis.
J Am Soc Nephrol 15
: 2208
–2218, 2004[Abstract/Free Full Text]
Schmitt CP, Odenwald T, Ritz E: Calcium, calcium regulatory hormones, and calcimimetics: Impact on cardiovascular mortality.
J Am Soc Nephrol 17[Suppl 2]
: S78
–S80, 2006
Ritz E, Gross ML, Dikow R: Role of calcium-phosphorous disorders in the progression of renal failure.
Kidney Int Suppl 99
: S66
–S70, 2005[Medline]
Berndt TJ, Schiavi S, Kumar R: "Phosphatonins" and the regulation of phosphorus homeostasis.
Am J Physiol Renal Physiol 289
: F1170
–F1182, 2005[Abstract/Free Full Text]
Tomford RC, Karlinsky ML, Buddington B, Alfrey AC: Effect of thyroparathyroidectomy and parathyroidectomy on renal function and the nephrotic syndrome in rat nephrotoxic serum nephritis.
J Clin Invest 68
: 655
–664, 1981[Medline]
Ogata H, Ritz E, Odoni G, Amann K, Orth SR: Beneficial effects of calcimimetics on progression of renal failure and cardiovascular risk factors.
J Am Soc Nephrol 14
: 959
–967, 2003[Abstract/Free Full Text]
Christiansen C, Rodbro P, Christensen MS, Hartnack B, Transbol I: Deterioration of renal function during treatment of chronic renal failure with 1,25-dihydroxycholecalciferol.
Lancet 2
: 700
–703, 1978[CrossRef][Medline]
Schwarz U, Amann K, Orth SR, Simonaviciene A, Wessels S, Ritz E: Effect of 1,25(OH)2 vitamin D3 on glomerulosclerosis in subtotally nephrectomized rats.
Kidney Int 53
: 1696
–1705, 1998[CrossRef][Medline]
Panichi V, Migliori M, Taccola D, Filippi C, De Nisco L, Giovannini L, Palla R, Tetta C, Camussi G: Effects of 1,25(OH)2D3 in experimental mesangial proliferative nephritis in rats.
Kidney Int 60
: 87
–95, 2001[CrossRef][Medline]
Weber TJ, Liu S, Indridason OS, Quarles LD: Serum FGF23 levels in normal and disordered phosphorus homeostasis.
J Bone Miner Res 18
: 1227
–1234, 2003[CrossRef][Medline]
Fugakawa M, Kazama JJ: With or without the kidney: The role of FGF23 in CKD.
Nephrol Dial Transplant 20
: 1295
–1298, 2005[Free Full Text]
ADHR Consortium: Autosomal dominant hypophosphataemic rickets is associated with mutations in FGF23.
Nat Genet 26
: 345
–348, 2000[CrossRef][Medline]
Jonsson KB, Zahradnik R, Larsson T, White KE, Sugimoto T, Imanishi Y, Yamamoto T, Hampson G, Koshiyama H, Ljunggren O, Oba K, Yang IM, Miyauchi A, Econs MJ, Lavigne J, Juppner H: Fibroblast growth factor 23 in oncogenic osteomalacia and X-linked hypophosphatemia.
N Engl J Med 348
: 1656
–1663, 2003[Abstract/Free Full Text]
Shimada T, Mizutani S, Muto T, Yoneya T, Hino R, Takeda S, Takeuchi Y, Fujita T, Fukumoto S, Yamashita T: Cloning and characterization of FGF23 as a causative factor of tumor-induced osteomalacia.
Proc Natl Acad Sci U S A 98
: 6500
–6505, 2001[Abstract/Free Full Text]
Shimada T, Urakawa I, Yamazaki Y, Hasegawa H, Hino R, Yoneya T, Takeuchi Y, Fujita T, Fukumoto S, Yamashita T: FGF-23 transgenic mice demonstrate hypophosphatemic rickets with reduced expression of sodium phosphate cotransporter type IIa.
Biochem Biophys Res Commun 314
: 409
–414, 2004[CrossRef][Medline]
Shimada T, Kakitani M, Yamazaki Y, Hasegawa H, Takeuchi Y, Fujita T, Fukumoto S, Tomizuka K, Yamashita T: Targeted ablation of FGF23 demonstrates an essential physiological role of FGF23 in phosphate and vitamin D metabolism.
J Clin Invest 113
: 561
–568, 2004[CrossRef][Medline]
Ferrari SL, Bonjour JP, Rizzoli R: Fibroblast growth factor-23 relationship to dietary phosphate and renal phosphate handling in healthy young men.
J Clin Endocrinol Metab 90
: 1519
–1524, 2005[Abstract/Free Full Text]
Larsson T, Nisbeth U, Ljunggren O, Juppner H, Jonsson KB: Circulating concentration of FGF-23 increases as renal function declines in patients with chronic kidney disease, but does not change in response to variation in phosphate intake in healthy volunteers.
Kidney Int 64
: 2272
–2279, 2003[CrossRef][Medline]
Imanishi Y, Inaba M, Nakatsuka K, Nagasue K, Okuno S, Yoshihara A, Miura M, Miyauchi A, Kobayashi K, Miki T, Shoji T, Ishimura E, Nishizawa Y: FGF-23 in patients with end-stage renal disease on hemodialysis.
Kidney Int 65
: 1943
–1946, 2004[CrossRef][Medline]
Shigematsu T, Kazama JJ, Yamashita T, Fukumoto S, Hosoya T, Gejyo F, Fukagawa M: Possible involvement of circulating fibroblast growth factor-23 in the development of secondary hyperparathyroidism associated with renal insufficiency.
Am J Kidney Dis 44
: 250
–256, 2004[Medline]
Gutierrez O, Isakova T, Rhee E, Shah A, Holmes J, Collerone G, Juppner H, Wolf M: Fibroblast growth factor-23 mitigates hyperphosphatemia but accentuates calcitriol deficiency in chronic kidney disease.
J Am Soc Nephrol 16
: 2205
–2215, 2005[Abstract/Free Full Text]
Nagano N, Miyata S, Abe M, Kobayashi N, Wakita S, Yamashita T, Wada M: Effect of manipulating serum phosphorus with phosphate binder on circulating PTH and FGF23 in renal failure rats.
Kidney Int 69
: 531
–537, 2006[CrossRef][Medline]
Ibels LS, Alfrey AC, Huffer WE, Craswell PW, Anderson JT, Weil R: Arterial calcification and pathology in uremic patients undergoing dialysis.
Am J Med 66
: 790
–796, 1979[CrossRef][Medline]
Block GA, Hulbert-Shearon TE, Levin NW, Port FK: Association of serum phosphorus and calcium x phosphate product with mortality risk in chronic hemodialysis patients: A national study.
Am J Kidney Dis 31
: 607
–617, 1998[Medline]
Ganesh SK, Stack AG, Levin NW, Hulbert-Shearon T, Port FK: Association of elevated serum PO(4), Ca x PO(4) product, and parathyroid hormone with cardiac mortality risk in chronic hemodialysis patients.
J Am Soc Nephrol 12
: 2131
–2138, 2001[Abstract/Free Full Text]
Giachelli CM, Speer MY, Li X, Rajachar RM, Yang H: Regulation of vascular calcification: Roles of phosphate and osteopontin.
Circ Res 96
: 717
–722, 2005[Abstract/Free Full Text]
Kestenbaum B, Sampson JN, Rudser KD, Patterson DJ, Seliger SL, Young B, Sherrard DJ, Andress DL: Serum phosphate levels and mortality risk among people with chronic kidney disease.
J Am Soc Nephrol 16
: 520
–528, 2005[Abstract/Free Full Text]
Narang R, Ridout D, Nonis C, Kooner JS: Serum calcium, phosphorus and albumin levels in relation to the angiographic severity of coronary artery disease.
Int J Cardiol 60
: 73
–79, 1997[CrossRef][Medline]
Tonelli M, Sacks F, Pfeffer M, Gao Z, Curhan G; Cholesterol and Recurrent Events Trial Investigators: Relation between serum phosphate level and cardiovascular event rate in people with coronary disease.
Circulation 112
: 2627
–2633, 2005[Abstract/Free Full Text]
Haut LL, Alfrey AC, Guggenheim S, Buddington B, Schrier N: Renal toxicity of phosphate in rats.
Kidney Int 17
: 722
–731, 1980[Medline]
Alfrey AC: The role of abnormal phosphorus metabolism in the progression of chronic kidney disease and metastatic calcification.
Kidney Int Suppl 90
: S13
–S17, 2004[Medline]
Boes E, Fliser D, Ritz E, Konig P, Kraatz G, Lhotta K, Mann JF, Muller GA, Neyer U, Riegel W, Riegler P, Kronenberg F; for the MMKD Study Group: Apolipoprotein A-IV predicts progression of chronic kidney disease: The Mild to Moderate Kidney Disease (MMKD) Study.
J Am Soc Nephrol 17
: 528
–536, 2006[Abstract/Free Full Text]
Bostom AG, Kronenberg F, Ritz E: Predictive performance of renal function equations for patients with chronic kidney disease and normal serum creatinine levels.
J Am Soc Nephrol 13
: 2140
–2144, 2002[Abstract/Free Full Text]
Related Article
This Month's Highlights
J. Am. Soc. Nephrol. 2007 18: A12.
[Full Text][PDF]
This article has been cited by other articles:
J. Danziger The bone-renal axis in early chronic kidney disease: an emerging paradigm
Nephrol. Dial. Transplant.,
September 1, 2008;
23(9):
2733 - 2737.
[Full Text][PDF]
Top
Abstract
Introduction
90 ml/min per 1.73 m2, GFR 60 to 89 ml/min per 1.73 m2, GFR 30 to 59 ml/min per 1.73 m2, and GFR <30 ml/min per 1.73 m2 (Table 1). We found a continuous and significant increase of Ca x P, PTH, and both c-terminal and intact FGF23 concentration across the various NKF stages of renal dysfunction. In addition, serum phosphate was significantly higher in patients with more advanced renal failure. We found significant correlations between c-terminal FGF23 as well as intact FGF23 with GFR, phosphate, and PTH (Figure 1). When we analyzed the data of patients with only CKD stages 2 through 5, we found similar results as for the entire population (data not shown). 
receptor blockers (n = 67; 30%), and 
-1 receptor blockers (n = 36; 16%). 
2 test. All statistical tests were performed at the two-sided 0.05 level of significance. Data are presented as means ± SD and as median and 25th and 75th percentiles for skewed variables where appropriate. Univariate correlation analysis was performed by Spearman correlation analysis. For calculation of ROC curves for the predictor variables c-terminal and intact FGF23, patients were stratified into two groups (those who progressed and those who did not over the study period), and for each cutoff value c, sensitivity was defined as the proportion of progressors with predictor variables above c and specificity as the proportion of nonprogressors with values below c. The optimal cutoff value for the predictor variables was defined as the cutoff obtaining the highest sum of sensitivity and specificity. Calculation of the AUC and a test of the null hypothesis that it equals 50% (equivalent to a random 50:50 prognosis) were performed using the Mann-Whitney test. Kaplan-Meier time-to-event curves were generated for patients with c-terminal and intact FGF23 concentrations above and below optimal cutoff levels. Adjusted risk estimates for progression end points were calculated using multivariable Cox proportional hazards regression analysis. Because Ca x P is the product term of the variable calcium and phosphate, estimates of Ca x P were based on a model that did not include calcium and phosphate at the same time, to avoid multicollinearity. Similarly, intact and c-terminal FGF23 were not offered at the same time in a particular model. Exploratory graphical analysis and tests of specific violations indicated no departure from the assumption of proportional hazards. 
