Abstract
Background
Fibroblast growth factor 23 (FGF-23) is a recently discovered hormone, which plays a key role in phosphate regulation. To investigate whether FGF-23 can be determined reliably, we validated the three available FGF-23 assays.
Methods
Currently, two intact FGF-23 assays (Kainos; Immutopics) and one C-terminal FGF-23 assay (Immutopics) are available. We determined intra- and inter-assay variation, linearity and matrix interference in these assays. Moreover, we compared assay results from healthy subjects with those of patient groups with expected high FGF-23 concentrations.
Results
Intra-assay variation was reasonably good in all three assays. Inter-assay variation and linearity were poor for the intact Immutopics assay but reasonable for both other assays. Immutopics assays gave best results in ethylenediaminetetraacetic acid (EDTA) plasma, while the Kainos assay showed comparable reproducibility in EDTA plasma and serum. Although the manual of the Kainos assay states that an automatic washing machine can be used, acceptable results were only found by manual washing. Patients with kidney disease and patients with hypophosphatemic osteomalacia had increased C-terminal FGF-23 concentrations compared with healthy controls.
Conclusion
Two assays of reasonable quality are available for FGF-23, the intact FGF-23 assay (Kainos) provided proper attention is paid to the washing procedure, and the C-terminal assay (Immutopics).
Introduction
Fibroblast growth factor 23 (FGF-23) is a recently discovered growth factor produced by osteocytes and behaving as a hormone in phosphate homeostasis. 1 FGF-23 decreases phosphate reabsorption in the kidney by down-regulating the expression of sodium-phosphate co-transporters in the proximal tubule. Moreover, FGF-23 inhibits 1α-hydroxylase expression, leading to decreased hydroxylation of 25-hydroxyvitamin D to 1,25-dihydroxyvitamin D.
Increased FGF-23 concentrations, due to genetic mutations, cause excessive renal phosphate excretion and inappropriately low plasma 1,25-dihydroxyvitamin D concentrations in several types of hereditary hypophosphatemic rickets. 2,3 Tumour-induced osteomalacia is a syndrome consisting of hypophosphataemia, due to decreased renal tubular phosphate reabsorption, inappropriately low 1,25-dihydroxyvitamin D, myopathy and osteomalacia, caused by tumour production of FGF-23. 4
Patients undergoing haemodialysis often have both increased serum phosphate and FGF-23 concentrations. FGF-23 concentrations in these patients appear to be independently associated with mortality. 5
Since its discovery, it has become clear that FGF-23 plays a pivotal role in numerous processes associated with phosphate homeostasis. However, reliability of the available FGF-23 assays is unknown. In the present study, we evaluated the performance of all three commercially available assays.
Methods
Assays
Three FGF-23 assays were evaluated. One assay was designed to measure C-terminal FGF-23 (cFGF-23) (Immutopics, San Clemente, CA, USA), whereas two other assays were designed to measure intact FGF-23 (iFGF-23) (Immutopics, Kainos, Tokyo, Japan). All three assays are sandwich enzyme-linked immunosorbant assays. According to the Immutopics package inserts, cFGF-23 and iFGF-23 (Immutopics) should be measured in ethylenediaminetetraacetic acid (EDTA) plasma while the Kainos assay recommends measuring iFGF-23 in serum. Unfortunately, synthetic or purified natural full-length FGF-23 for use as a standard is not available. The C-terminal kit employs a cell culture supernatant into which FGF-23 has been secreted to this end. The composition of the calibrators for the other assays was not disclosed. By consequence, all reported FGF-23 levels are relative. The BioTek ELx50 (BioTek, Bad Friedrichshall, Germany) was used as an automatic washing machine for all three assays. All analyses were performed according to the manufacturer's protocol, unless otherwise described.
Intra- and inter-assay variation
Intra- and inter-assay variations were determined using human samples with low, normal and high concentrations of FGF-23.
Linearity
To determine the linearity, samples were diluted in phosphate-buffered saline and FGF-23 was measured in all three assays. The recovery was calculated by dividing the result obtained by the expected value. A recovery of 100% represents a good linearity.
Matrix interference
iFGF-23, cFGF-23 (Immutopics) and iFGF-23 (Kainos) were measured concurrently in paired drawn specimens in EDTA plasma and serum.
Patients
cFGF-23 and iFGF-23 were determined in 78 and 52 healthy subjects, respectively, 20 and 19 predialysis patients, respectively (glomerular filtration rates 35 ± 11 and 35 ± 11 mL/min/1.73 m2, respectively), 47 and 21 patients undergoing dialysis, respectively (samples taken before dialysis) and four patients with hypophosphatemic osteomalacia, proven by non-decalcified iliac crest biopsy. All patients gave informed consent to the study.
Statistics
Differences between groups were determined by Mann-Whitney U test. A P value below 0.05 was considered significant.
Results
Results of intra- and inter-assay variation are shown in Table 1. For the Kainos intact assay, the results are shown for automatic washing (as recommended by Kainos) as well as for manual washing. The high inter-assay variation in the intact Immutopics assay was mainly caused by lot-to-lot variability.
Intra- and inter-assay variation
*N is based on duplicate measurements
Nd, not determined; CV, coefficient of variation
For Immutopics, a recovery of around 100% was found for dilutions up to 10% for the C-terminal assay. In contrast, recovery for the intact assay was above 100% and increased with dilutions of 60% and more. The Kainos assay showed a recovery of around 100% for the entire range.
cFGF-23 and iFGF-23 (Immutopics) could only be determined in EDTA plasma, as measurements in serum gave lower or even undetectable results. iFGF-23 by Kainos could be determined in both serum and EDTA plasma.
cFGF-23 and iFGF-23 concentrations were significantly increased in predialysis patients (397 ± 644 RU/mL; 312 ± 786 ng/L; both P < 0.01) and patients undergoing dialysis (13915 ± 17782 RU/mL; 5091 ± 7103 ng/L; both P < 0.01) compared with healthy subjects (39 ± 24 RU/mL; 34 ± 8 ng/L). A significant difference in concentrations of cFGF-23 and iFGF-23 between predialysis and dialysis patients was also seen (P < 0.01). Patients with hypophosphatemic osteomalacia showed elevated cFGF-23 concentrations from 162 to 558 RU/mL compared with healthy subjects (P < 0.01).
Discussion
The three assays evaluated in this study show reasonably good intra-assay variation. Inter-assay variation and linearity, however, were unacceptable in the intact Immutopics assay, but reasonable in the other two assays. The iFGF-23 assay by Immutopics was therefore discarded in further analysis.
The manual of the iFGF-23 assay by Kainos states that an automatic washing machine can be used when washing the plates. We started out by using an automatic washing machine, but were confronted with poor reproducibility of the assay. The effect of washing the plates manually was striking, and improved the coefficient of variation (CV) from 43% to 8% (Table 1). We therefore recommend that manual washing is absolutely necessary to obtain acceptable results for the iFGF-23 Kainos assay.
The obtained intra- and inter-assay CVs show that both the C-terminal assay by Immutopics and the intact assay by Kainos give reproducible results. However, the question remains how accurate these assays are. The antibodies in the C-terminal assay are directed against the C-terminal region of FGF-23, and therefore detect both putative C-terminal fragments and intact FGF-23. The antibodies in the intact assay are described as being directed against the C-terminal and the N-terminal region of FGF-23 and may therefore detect only iFGF-23. 6 However, as long as there are no well-defined standards available, we do not know the exact accuracy of these assays. In diseases with known pathological high concentrations of FGF-23 such as hypophophatemic rickets and tumour-induced osteomalacia, FGF-23 was increased when analysed with both assays. 6 We observed grossly elevated concentrations of FGF-23 in four patients with hypophosphatemic osteomalcia. Moreover, predialysis and dialysis patients showed increased FGF-23 concentrations. These results indicate that the assays probably detect FGF-23.
In analogy with parathyroid hormone assays, it is important to know whether FGF-23 fragments are measured in these assays, as measurement of different fragments in patients with renal failure may cause problems. 7,8 More research has to be performed with regard to clearance of FGF-23 fragments to clarify this problem. Also the possibility that potential FGF-23 fragments might have unexpected biological activity should be addressed.
Conclusion
On the basis of this study, there are two assays for FGF-23 of reasonable quality on the market, the iFGF-23 assay by Kainos, provided proper attention is given to the washing procedure, and the cFGF-23 assay by Immutopics.
DECLARATIONS