Vitamin D is a fat-soluble secosteroid that is extensively metabolized in the human body. Over the last 40 years, its synthesis and metabolism have been elucidated and more than 50 metabolites of vitamin D have been discovered.[6-8] However, to date, researchers have been able to develop measurement procedures for only a few of them (Table 1).

Vitamin D exists in two major forms, vitamin D2 (or ergocalciferol) and vitamin D3 (or cholecalciferol), which exhibit only minor differences in their structure. (Fig. 1). As a consequence, vitamin D2 and D3 have different molecular weights of 396.65 g/mol and 384.64 g/mol, respectively.[9] These differences in the chemical structure of vitamin D2 contribute to its lower affinity for vitamin D binding protein (VDBP), thus resulting in faster clearance from blood, a limited conversion to 25 hydroxyvitamin D [25(OH)D], and an altered catabolism by 24-hydroxyase (CYP24A1).[10-12] A recent meta-analysis found that vitamin D3 is more potent at raising serum 25(OH)D concentrations than is vitamin D2. Hence, vitamin D3 could potentially become the preferred choice for supplementation.[13]

Vitamin D3 is synthesized from 7-dehydrocholesterol (7-DHC) in the skin by UVB radiation while vitamin D2 is derived from plant/yeast by irradiation of ergosterol (Figs. 2 and 3).[14,15] In humans the main sources of vitamin D (e.g., D2 and/or D3), are sunlight, diet, and supplements. However, most foods (except for fatty fish) contain low levels of vitamin D unless fortified (Table 2). Exposure of human skin to solar UVB radiation (wavelengths 290–315 nm) leads to the conversion of 7-DHC to pre-vitamin D (pre-D) in the skin, which isomerizes to D3 in a non-catalytic, thermo-sensitive process.[16] Vitamin D3 production depends on the intensity of UV irradiation, which varies with season, latitude and altitude.[17] Skin pigmentation, sunscreen use, and clothing have been reported to affect the conversion of 7-DHC to vitamin D3.[18-20] Melanin in the skin blocks UVB from converting 7-DHC, thus limiting D3 production, as does extensive covering of the body with clothes and the use of sun-screen. A recent meta-analysis concluded that pigmented skin has less effective photoproduction of vitamin D and 25(OH)D. The quantity of sun exposure needed for dark-skinned, compared with light-skinned, people to achieve vitamin D sufficiency however remains uncertain.[21] However this view has been debated lately; Bogh et al., in an elegant study show that baseline vitamin D levels and total cholesterol levels are more important factors than skin pigmentation.[22,23]

Vitamin D synthesized in the skin diffuses into the bloodstream where it is transported by vitamin D binding protein (VDBP) to the liver. Vitamin D from the diet is absorbed in the small intestine, incorporated into chylomicrons, which are released into the lymphatic system, and enters the venous blood where it binds to VDBP and lipoproteins before being transported to the liver. Vitamin D is essentially biologically inactive and must be converted to hydroxylated metabolites to gain hormonal activity. Its activation involves two hydroxylation steps (Fig. 4).[24]

The first step occurs predominantly in the liver where vitamin D is hydroxylated at the C25 position by the cytochrome p450 enzyme CYP2R1 (also called 25-hydroxylase) to 25-hydroxyvitamin D [25(OH) D]. There are several other CYP enzymes that are capable of 25-hydroxylation, namely CYP27A1, CYP3A4, CYP2D25, and perhaps others, but the CYP2R1 is emerging as the most critical enzyme for 25-hydroxylation.[25-27] This step is poorly regulated by any feedback mechanism in the context of the vitamin D endocrine system and it seems to be dependent primarily on the concentration of vitamin D [28]. Hence, 25(OH)D levels increase in proportion to vitamin D intake and, plasma 25(OH)D levels are a good indicator of vitamin D status. The 25(OH)D produced in the liver is returned to circulation. Severe liver failure affects the function of the CYP2R1 enzyme. Moreover, loss-of-function mutations for the same enzyme are responsible for vitamin D-dependent rickets (VDDR), type 1B (VDDR-1B) [29] as shown in Table 3.

The second step in vitamin D activation is the formation of 1α,25-dihydroxy vitamin D [or calcitriol, 1α,25(OH)₂D]. It occurs, under physiological conditions, mainly in the kidney by another CYP450 enzyme, CYP27B1 or 1α-hydroxylase. This second hydroxylation takes place at the carbon in the C1 position. Calcitriol binds again to VBDP and re-enters the systemic circulation. It is now recognized that 1α-hydroxylase is expressed in many other extrarenal tissues, including skin, brain, and colon and serves as an autocrine/paracrine factor with cell specific functions (Table 4 and Fig. 5).[9,30-34].

The activity of renal 1α-hydroxylase is tightly controlled by 1α,25(OH)₂D itself, parathyroid hormone (PTH), fibroblast growth factor 23 (FGF23) and serum concentrations of calcium (Ca) and phosphate (PO4-3). Under normal conditions, FGF23 acts on kidney proximal tubular cells regulating phosphate excretion in order to maintain systemic phosphate homeostasis. However, FGF23 can also influence the synthesis of calcitriol in the proximal tubular cells by suppressing the expression of 1α-hydroxylase and increasing the expression of 24-hydroxylase.[30,35] The activity of extrarenal 1α-hydroxylase is not regulated by the same factors controlling its renal synthesis.[30] The kidney is the major source for circulating 1α,25(OH)₂D. Only in certain granulomatous diseases such as sarcoidosis does the extrarenal tissue produces sufficient 1,25(OH)₂D to contribute to the circulating levels, which is generally associated with hypercalcemia.[36] Inactivating mutations of this enzyme are responsible for vitamin D-dependent rickets (VDDR) type 1A [VDDR-1A] [28,32,33,37] as shown in Table 3.

To retain calcitriol levels within the strict boundaries required for appropriate calcium homeostasis and bone metabolism, both 1α,25(OH)₂D and 25(OH)D may undergo further hydroxylation by renal CYP24A1 (24-hydroxylase), leading to 1,24,25-trihydroxyvitamin D [1,24,25(OH)₃D] and 24R,25-dihydroxyvitamin D [24,25(OH)₂D], respectively (Fig. 6). Thus the main function of 24-hydroxylase is vitamin D inactivation, since [1] it limits the amount of 1α,25(OH)₂D3 in target tissues both by accelerating its catabolism to 1,24,25(OH)3D3 and ultimately in calcitroic acid or [2] by producing 24,25(OH)2D3 and thus decreasing the pool of 25(OH)D3 available for 1 hydroxylation.[38]

CYP24A1 has been found in many tissues that express the vitamin D receptor. In the kidney, it is found in the proximal and distal tubules. [39,40] The CYP24A1 gene is highly inducible by 1α,25(OH)₂D in all tissues in which it is found and it acts as a control mechanism to prevent intoxication from 1α,25(OH)₂D. [41] The importance of this feedback mechanism was demonstrated when inactivating mutations of CYP24A1 reported in children and adults with hypercalcemia.[29,42]

Another enzyme, CYP3A4, also plays a role in vitamin D catabolism. [43] This enzyme is involved in drug metabolism, and is located in the liver and the intestine. Recently, a gain-of-function mutation in CYP3A4 was described that leads to rickets with decreased serum calcium and phosphate and elevated PTH and alkaline phosphatase (Table 3).[44] This is a distinct form of vitamin D dependent rickets (named type 3 vitamin D-dependent rickets or VDDR3) since it does not involve a defect in synthesis of vitamin D metabolites but rather is due to accelerated inactivation of vitamin D metabolites as CYP3A4 was found to inactivate both 25(OH)D3 and 1,25(OH)₂D, leading to vitamin D deficiency through accelerated vitamin D metabolite inactivation (Table 3). [24,45] It is well known that CYP3A4 is induced by certain drugs, such as rifampicin.[46,47] Thus, the induction of CYP3A4 gene expression by certain drugs may enhance 25OHD and 1α,25(OH)2D3 catabolism.[43] and hence modulate vitamin D effects in the body and could present as an alternative therapeutic strategy to reduce serum levels of vitamin D metabolites in cases of patients with inactivating mutations of CYP24A1. [48]

Five types of pre-analytical variability will be examined in detail in this section:

sample collection and handling

factors that relate to the individual (i.e., age, sex, ethnicity, and lifestyle)

environmental factors

disease factors and pregnancy

genetic factors

Age, sex, body fat, and lifestyle do have a, often small, effect on 25(OH)D levels.[78]

Gene-environment interactions that could have an impact on various vitamin D-related disorders have recently drawn the attention of several researchers.[171,172] For example, it has been suggested that hypovitaminosis D occurs in the presence of specific gene variations related to vitamin D metabolism. Therefore, individuals with specific vitamin D-related genotypes may require specific personalized advice to optimize their vitamin D status. Data from twin and family-based studies have demonstrated that circulating vitamin D concentrations can be partially determined by genetic factors.[173,174] Moreover, it has been shown that genetic variants (e.g., mutation) and alterations (e.g., deletion, amplification, and inversion) in genes involved in the metabolism, catabolism, transport, or even binding of vitamin D to its receptor might have an effect on vitamin D levels.[175] However, the underlying genetic determinants of 25(OH)D plasma levels have not been fully elucidated. Furthermore, the association between epigenetic modifications such as DNA methylation and vitamin D levels have now been reported in several studies.[175]

Linkage studies, studies involving candidate genes in the vitamin D metabolism pathway, as well as genome wide association studies (GWAS) have shown human genetic variants to be related to vitamin D status.

The measurement of 25(OH)D is performed mainly for two reasons: [1] to determine the nutritional status of vitamin D, and [2] to monitor the efficacy of supplementation. As previously mentioned, vitamin D exists in two different forms and in order to adequately monitor therapy, both types of vitamin D need to be accurately quantitated. In fact, the accurate measurement of 25(OH)D for the assessment of vitamin D status has always been a major goal of all clinical laboratories involved in the measurement of vitamin D metabolites. 25(OH)D is the metabolite of choice to determine vitamin D status for several reasons:

25(OH)D levels in the blood are higher than those of any other vitamin D metabolite. The serum concentration of 25(OH)D is in the range of 25–200 nmol/L, which is 1000 times higher than that of 1α,25(OH)₂D, whose concentration is in the range of 50–150 pmol/L. Majority of 25(OH)D is found in the systemic circulation, with limited distribution in less accessible tissues (e.g., fat) [179].

It is well accepted that adequate levels of vitamin D are required to prevent nutritional rickets and osteomalacia, both of which are characterized by low levels of 25(OH)D as shown in Table 3. [9]

Several clinical studies have demonstrated that there is an association between serum levels of 25(OH)D and several clinical outcomes such as bone mineralization, fracture risk, fall risk, cancer, diabetes, and cardiovascular events. Meta-analyses and randomized control trials demonstrated a positive dose–response relationship between vitamin D supplementation and fracture prevention, which could partly be attributed to fall reduction.[180,181]

25(OH)D has a relatively long half-life (2–3 weeks) compared to that of 1α,25(OH)₂D (approximately 4–6 h), and, therefore, serum levels vary little within short periods of time [41,182].

The hydroxylase enzymes that metabolize vitamin D to 25(OH)D in vivo behave according to first-order reaction kinetics. This means that its rate of production is dependent on vitamin D levels and, therefore, its level in systemic circulation is the best indicator of vitamin D nutritional status [183].

Furthermore, 25(OH)D represents the sum of vitamin D intake and dermal production [184].

Serum levels of 25(OH)D are relatively stable and not affected by diet (i.e., calcium intake) and lifestyle (i.e., sedative life or regular physical exercise), whereas 1α,25(OH)₂D levels are affected by all the latter [179,182].

Serum levels of 25(OH)D can determine if there is enough 25(OH)D for the extrarenal tissues to produce 1α,25(OH)₂D for autocrine or paracrine action. Recent data have revealed that many of these tissues also contain CYP27B1, which is responsible for converting 25(OH)D to 1,25(OH)₂D. Regulation of CYP27B1 in these non-renal tissues is generally different from that in the kidney and may be more substrate-dependent. This finding has led to the concept that the maintenance of adequate 25OHD levels in the blood is required for vitamin D regulation of a large number of physiologic functions beyond those of the classic actions involved in bone mineral metabolism. Measurement of 1α,25(OH)₂D does not provide this information since its extrarenal production does not contribute much to the systemic load [52,185].

25(OH)D can be measured using various methods including immunoassays, which are mainly used, protein-binding assays, HPLC-UV, or LC-MS/MS.[59,183,186-188]

Serum total 25(OH)D, [the sum of 25(OH)D2 and 25(OH)D3], is considered to be the best biological marker of an individual’s vitamin D status as described above. However, in clinical guidelines, differences exist in the definition deficiency, insufficiency, and sufficiency, creating a great deal of controversy [189,190]. The most critical factor that confounds efforts to develop consensus clinical and nutritional public health guidelines for interpreting serum 25(OH)D concentrations is the substantial variability that existed (and still exists) in many assays that have been used over the years to measure 25(OH)D in clinical research studies [191]. The lack of assay standardization is the main source of bias, making it impossible to pool research results in order to develop consensus cut-off points [192].

To overcome these problems, the Vitamin D Standardization Program (VDSP) was established in 2010 by the Office of Dietary Supplements, National Institutes of Health, as an international collaborative effort with the National Institute of Standards and Technology (NIST), the Centers for Disease Control and Prevention (CDC), Ghent University in Belgium, the American Association for Clinical Chemistry (AACC), the IFCC, and national health and nutrition surveys from Australia, Canada, Germany, Ireland, Mexico, South Korea, United Kingdom, and the USA [193]

A reference measurement procedure (RMP) is now available. [194,195] NIST first developed a RMP based on ID-LC-MS/MS for the determination of 25(OH)D2 and 25(OH)D3 in human serum [195]. This method is now recognized by the Joint Committee for Traceability in Laboratory Medicine (JCTLM) as a RMP. Later, Stepman et al. (at Ghent University, Belgium) also described an ID-LC-MS/MS method for the determination of 25(OH)D2 and 25(OH)D3, which was also recognized as an RMP for 25(OH)D by the JCTLM [196]. Finally, the CDC developed an ID-LC-MS/MS method that was subsequently recognized by the JCTLM as an RMP for 25(OH)D [197]. These three methods are currently the only JCTLM-recognized RMPs for the determination of 25(OH)D2 and 25(OH)D3and make it possible to standardize 25(OH)D measurements. As standardization is crucial as cut-off values for deficiency, insufficiency, and sufficiency are used internationally, the Centers for Disease Control (CDC) started an international Vitamin D standardization certification program (VDSCP).[193] This led to an impressive improvement in the number of standardized 25(OH)D assays.

The achievements of VDSP and of the institutions that supported the objectives of the VDSP were significant: it resulted in a reference measurement system that is the backbone for standardizing 25(OH)D measurements in current and future assay systems [198,199]. The components of this reference measurement system include: [1] the gold standard RMPs, [2] NIST Standard Reference Materials (SRMs), [3] the VDSCP (developed, implemented and executed by the CDC), [4] the accuracy-based performance testing or external quality assessment schemes (PT/EQA) conducted by the College of American Pathologists (CAP) and the Vitamin D External Quality Assessment Scheme (DEQAS) [195,200,5] methods for retrospective standardization of studies completed in the past, and [6] a set of laboratory performance guidelines for both reference laboratories (those that govern the RMPs) and for routine laboratories developed by the VDSP [200].

The VDSP adopted the assay performance criteria for 25(OH)D that were developed by Stockl et al. for different types of labs and assays [200]. According to these criteria, for RMPs, the limits for total CV and mean bias should be less than or equal to 5% and less than or equal to 1.7%, respectively. For routine clinical laboratories, the limits for total CV and mean bias are not so strict and should be less than or equal to 10% and 5%, respectively.

However even today, several immunoassays suffer from other analytical issues that lead to continuing problems with the quality of 25(OH)D measurements. These immunoassays show patient or matrix dependent deviations, for instance in pregnant women, patients on intensive care, hemodialysis patients, osteoporotic patients, and patients with liver failure. Sera from these patient groups behave differently in these immunoassays than in LC-MS/MS assays. [168,201-205] One of the causes is a vitamin D binding protein (DBP) dependency in the immunoassay, but other causes are still unknown. Furthermore, immunoassays often demonstrate difficulties with the measurement of 25(OH)D₃ and 25(OH)D₂.[204,206,207] Where LC-MS/MS methods can separate 25(OH)D₃ and 25(OH)D₂, immunoassays make use of antibodies with a different affinity for 25(OH)D₃ and 25(OH)D₂ and therefore often over- or underestimate measured 25(OH)D depending on whether the subject uses 25(OH)D₂. In addition, some immunoassays suffer from cross reactivity with other vitamin D metabolites such as 24,25(OH)₂D. [208,209] On the other hand, LC-MS/MS assays are not always designed to separate the epimer of 25(OH)D. This can lead to falsely high 25(OH)D concentrations, especially in young children.[209-211] Most immunoassays do not show cross reactivity with the epimer.

The above-mentioned issues, including the different affinities for 25(OH)D₃ and 25(OH)D₂, cross reactivity with 24,25(OH)₂D, and matrix and/or patient-dependent biological variations make these currently available immunoassays difficult for standardization and accurate measurements.

Although 1α,25(OH)₂D is the active form of vitamin D, its measurement does not provide any additional value in determining an individual’s vitamin D status. Its measurement thus, should be limited in serious clinical conditions such as hypo- or hypercalcemia.[179,212]

Three types of conditions can influence 1α,25(OH)₂D disorders such as [1] disorders regarding CYP27B1, [2] disorders in the VDR, and [3] disorders in the extrarenal production of 1α,25(OH)₂D. For example, vitamin D dependent rickets type 1 or pseudo-vitamin D deficiency rickets is a genetic disorder leading to a CYP27B1 deficiency, which causes hypocalcemia and early onset rickets.[213]

Also X-linked hypophosphatemia, autosomal dominant hypophosphatemic rickets, autosomal recessive hypophosphatemic rickets 1 – 3, tumor induced osteomalacia, and other rare disorders leading to FGF23-mediated hypophosphatemia all led to inhibition of the CYP27B1 gene, abnormally low 1α,25(OH)₂D concentrations, and eventually osteomalacia or rickets [see Table 5 and for detailed review see references [214,215]]. Additionally, there are some rare disorders that may manifest as FGF23-mediated hypophosphatemia. These include, osteoglophonic dysplasia, McCune–Albright syndrome, epidermal nevus syndrome, neurofibromatosis, hypophosphatemic rickets with hyperparathyroidism, and Jansen metaphyseal chondrodysplasia. High FGF23 levels result in inhibition of CYP27B1 and therefore, abnormally low 1α,25(OH)₂D concentrations (Table 5).

High 1α,25(OH)₂D concentrations can be seen in disorders associated with the VDR such as hereditary vitamin D resistant rickets (or vitamin D dependent rickets type 2). Despite high 1α,25(OH)₂D concentrations, hypocalcemia and rickets appear as the VDR does not detect vitamin D (see Table 3). [216] Also, in diseases with excessive and uncontrolled extrarenal production of active vitamin D, high concentrations of 1α,25(OH)₂D are detected. Examples of these disorder types include sarcoidosis, tuberculosis, rheumatoid arthritis, inflammatory bowel disease, and lymphoproliferative disorders (Table 6). These disorders often result in a low bone mineral density/osteomalacia.[36,217-219] This extrarenal 1α-hydroxylation by local CYP27B1 is not controlled by PTH, FGFG23, phosphate, or 1α,25(OH)₂D itself, but is rather regulated by local factors such as IFN-γ and IL15, and is dependent on the availability of substrate.[36,220] When the locally produced 1α,25(OH)₂D concentration is too high, it may escape the confinements of the intracellular space, spill over to the systemic circulation, and raise blood concentrations to abnormally high levels.[219] In the case of hypo- or hypercalcemia or hypophosphatemia, the measurement of 1α,25(OH)₂D plays an important role in the verification of absence or onset of these conditions.

As serum 1,25(OH)₂D concentrations are very low (pmol/L), it is very difficult to measure this analyte. Over the few past decades 1,25(OH)₂D was often measured using manual competitive protein binding assays or radioimmunoassays. Many of these radioimmunoassays encountered problems with specificity, as cross-reactivity with other vitamin D metabolites influenced the results.[221] Recently automated immunoassays became commercially available, which appear to perform better regarding cross reactivity. However, these methods cannot separate 1,25(OH)₂D₂ from 1,25(OH)₂D₃, which might be problematic in countries with D2 supplementation.[222,223] In addition, over the last few years, several laboratories have developed LC-MS/MS methods to measure 1,25(OH)₂D. The sensitivity issue with these LC-MS/MS methods can be resolved by 2D chromatography, derivatization of the vitamin D molecule, and/or immunopurification. [212] This last option is not only advantageous for the sensitivity, but also regarding specificity. LC-MS/MS methods without immunopurification might suffer from isobaric interferences as it is difficult to separate 1β-25-dihydroxyvitamin D3 from its epimer.[224] Therefore, not all LC-MS/MS methods fully agree in method comparison studies. [225]

Unfortunately, no reference methods exist for measuring 1,25(OH)₂D. Without reference method standardization, an evaluation of the quality of currently available methods is not possible. Absent of standardization, the determination of reference values is subjective to the method used.

24,25(OH)₂D circulates in a concentration range of nmol/L and has a half-life of about 7 days. These characteristics make this metabolite attractive for quantitation and a candidate biomarker of vitamin D catabolism.[158,227] The amount of circulating 24,25(OH)₂D depends on the amount of its predecessor 25(OH)D and the activity of CYP24A1. The expression of CYP24A1 is upregulated by 1α,25(OH)₂D, and FGF23 and is downregulated by PTH. Moreover, it is partly regulated by VDR activity.[228,229]

Consequently, if there are sufficient levels of biologically active vitamin D and the expression of CYP24A1 is adequate, then calculating the ratio 25(OH)D/24,25(OH)₂D (also called vitamin D metabolite ratio or VMR) is a good indication of the catabolic clearance by CYP24A. If we also take into account that the production of 24,25(OH)₂D is dependent on 25(OH)D and on the expression of CYP24A1, then the absolute concentration of 24,25(OH)₂D or the VMR may be a better indicator of vitamin D sufficiency than 25(OH)D alone since it is not affected by race. [230] In addition, several studies have suggested that 24,25(OH)₂D has effects of its own [231-235], and that human bone cells and human mesenchymal stem cells (hMSCs) metabolize 25(OH)D3 into both 1α,25(OH)2D3 and 24R,25(OH)2D3.[236-238] These results demonstrate the ability of bone cells to convert 25(OH)D3 in vitro, indicate the importance of systemic and tissue-specific 24,25(OH)₂D3 actions, suggest a role in osteoblastic differentiation, and enhance the concept that the hydroxylation of 25(OH)D3 leads to 2 bioactive forms of vitamin D3, 24,25(OH)₂D3 and 1α,25(OH)₂D3, each with its own unique functions. [239] Moreover these studies demonstrated that 24,25(OH)₂D3 is an active form of vitamin D3 with an essential role in osteoblast maturation, Ca²⁺ mineralization, gene expression, and the regulation of cytochrome P450 expression, resulting in decreased 1α,25(OH)₂D3 biosynthesis.[239] These data suggest a direct role in bone cells—in particular, in osteoblasts. It should also be noted that 24-hydroxylation is the first step of a degradation cascade. Therefore, the biologically-active levels of 24,25D3 or 1,24,25D3 fully depend on the velocity of the subsequent steps in the degradation pathway.[240] It is of no surprise the biological significance of 24-hydroxylase has been continuously discussed because of its dual role first as a catalytic enzyme initiating the side chain catabolism of both 25(OH)D3 and more importantly 1α,25(OH)₂D3 in target tissues and second as an enzyme with a synthetic capacity since, in some situations, it is activated to produce 24,25(OH)₂D3.[241]

Chronic kidney disease (CKD) is characterized by a state of active vitamin D deficiency. In contrast to the focus placed on the decreased renal production of 1α,25(OH)₂D3, relatively little attention has been paid to the potential role of altered vitamin D catabolism in CKD. In healthy people, the concentrations of vitamin D metabolites in blood and target tissues represent a balance of production and catabolism. CYP24A1 is the primary enzyme responsible for the multistep catabolism of both 25(OH)D and 1α,25(OH)₂D3. CYP24A1 is present in most tissues in the body and is rapidly induced by 1α,25(OH)₂D3. In the kidney, CYP24A1 is also induced by FGF23 and suppressed by PTH. In CKD, the net effects of declining kidney function with increasing FGF23 and PTH concentrations on vitamin D catabolism are not clear. [40,158,229]

The measurement of 24,25(OH)₂D3 also seems to be useful in the detection of loss-of-activity mutations of the gene that encodes for CYP24A1.[242,243] Recent studies have shown that loss-of-function mutations for 24-hydroxylase are associated with a clinical phenotype/condition that is characterized by low PTH levels, increased 1α,25(OH)₂D, hypercalcemia, hypercalciuria, and/or the presence of kidney stones (Tables 6 and 7).[242,244-246]

For the evaluation and the diagnosis of patients with inactivating mutations of CYP24A1, it has been proposed that if a 25(OH)D/24,25(OH)₂D ratio is greater than 80, the presence of a homozygous mutation is very probable. If the ratio is between 25 and 80, the presence of a mutation is probable. However, confirmation with molecular testing is always necessary.[242,247] We must note here that the use of ratios for measured analytes is open to criticism. Indeed, there is not yet standardization of the expression of the ratio, which can lead to added criticism. In the literature, the VMR is expressed as the ratio of 24,25(OH)₂D to 25(OH)D, the ratio of 25(OH)D to (R)-24,25(OH)₂D, or even as a percentage (24,25(OH)₂D/(25(OH)D) × 100). Finally, in patients with a CYP24A1 mutation, 24,25(OH)₂D may not be quantifiable, making the calculation of the VMR impossible as the concentration of 24,25D approaches zero.[230,248-250] Cashman et al. described the presence of 24,25(OH)₂D in human serum as a “double-edged sword—an interferent for some immunoassays, yet potentially informative of nutritional status”.[208]

In summary, the molar ratio of 25(OH)D to 24,25(OH)₂D (or the VMR) has been recently used as an index of vitamin D deficiency and catabolism in healthy individuals, as well as in individuals with genetic mutations in the CYP24A1 gene such as individuals with idiopathic infantile hypercalcemia (IIH) and to evaluate the efficacy of supplementation. [208,251]

The measurement of 24,25(OH)₂D presents difficulties since the molecule is encountered in relatively low concentration in serum and is challenging to ionize in LC-MS/MS methods. The LC-MS/MS methods developed to measure 24,25(OH)₂D ensure a high sensitivity and specificity, but unfortunately, are not available in all clinical labs. [208,242,252,253] LC-MS/MS methods present several advantages. The pretreatment of the samples prior to LC-MS/MS analysis ensures the removal of binding proteins and the excellent chromatographic separation of various metabolites. As a result, LC-MS/MS methods often allow for the simultaneous quantitation of several vitamin D metabolites.[247,252-254]

Concerning reference intervals, one study showed that 24,25(OH)₂D levels in the general population are between 1.1 and 13.5 nmol/L.[252] Moreover, in this study, the investigators showed that levels greater than 4.2 nmol/L were indicative of 25(OH)D sufficiency. A more recent study calculated a slightly lower reference interval (0.4–8.9 nmol/L) using a method that was standardized using the recently developed NIST reference material.[255] In these two studies, the authors also calculated the ratio of 25(OH)D/24,25(OH)₂D with similar results.

VDSP has expanded its interest beyond 25(OH)D and to other vitamin D metabolites such as dihydroxyvitamin D metabolites and mainly 24,25(OH)₂D. Recently, NIST developed a candidate RMP based on ID-LC-MS/MS for the determination 24,25(OH)₂D. This method was published recently and recognized as a RMP by JCTLM in 2017.[256-258] Although several researchers have published methods based on ID-LC-MS/MS for the determination of 24,25(OH)₂D3 in human serum, the NIST method is the only RMP for this metabolite, and as such, it represents a key component in VDSP efforts to move toward the standardization of measurements for this metabolite. This method was used recently to assign values for 24,25(OH)₂D3 in two SRMs (SRM972a SRM2971) and in critical study samples.[256-259] DEQAS is offering an accuracy based external quality assessment scheme for 24,25(OH)₂D.

In addition to the primary pathway of vitamin D metabolism, there are also a number of minor metabolic pathways. It was recently discovered that vitamin D can alternatively be metabolized through the a C3-epimerization pathway that parallels the standard metabolic pathway.[260] This pathway creates the vitamin D epimers - a certain group of metabolites that has attracted much attention (Fig. 7). The C3 epimerization pathway leads to the conversion of the configuration of the hydroxyl group at C3 of the A-ring. In the C3 epimerization pathway, the hydroxyl group at position C3 of the A-ring is inverted from the β position to its diastereomer (α) while the other chiral centers remain unchanged. Therefore, epimers are molecules with an identical structure, but different a stereochemical configuration (diastereoisomers) at one chiral center.[261,262] Both vitamin D2 and D3 can be epimerized. The C3-epimers of 25(OH)D are produced by 25(OH)D3-C3-epimerase. This enzyme is present in the endoplasmic reticulum of a range of cells/tissues including liver, bone and skin, but not kidney. The gene responsible for encoding this enzyme has not yet been identified.[8] The process of epimerization is irreversible. Epimerase enzymes can also carry out the epimerization process of 1α,25(OH)₂D3 and 24,25(OH)₂D although it is not performed at the same rate as 25(OH)D.[263] Microsomes containing the epimerase have been reported to act on 1α,25(OH)₂D3 and 24R,25(OH)₂D3, producing 3α-epimers.[264]

The C3-epimer of 25(OH)D [C3-epi-25(OH)D] is the most abundant epimer found in systemic circulation.[265] The C3-epi-25(OH)D can also undergo 1α hydroxylation to give C3-epi-1α,25(OH)₂D, and 24 hydroxylation to give C3-epi-24,25(OH)2D.[262] Subsequently, C3-epi-1α,25(OH)2D3 is metabolized to three polar compounds, C3-epi-1α,24 (R),25(OH)3D3, C3-epi-24-oxo-1α,25(OH)2D3, and C3-epi-24-oxo-1α,23(S),25-trihydroxyvitamin D3. Therefore, CYP27B1 and CYP24 are able to perform C-1α and C-24 hydroxylation irrespective of the stereochemistry of the C3 hydroxyl group.[262-265]

At present, the source of the C3-epimer from diet, supplements, or endogenous metabolism and factors underlying the epimerization of vitamin D metabolites are still unclear. Also, their physiological importance is not known very well and is under debate.[265] From in vitro studies and experiments with rodent models, we have learned that C3-epi-25(OH)D3 and the epimeric form of calcitriol (C3-epi-1α,25(OH)₂D3) can bind to VDBP at 36–46% and to vitamin D receptor (VDR) at 2–3% as compared to their non-epimeric forms.[262] Nevertheless, C3-epi-1α,25(OH)2D3 appears nearly as potent as calcitriol in suppressing PTH, although it has significantly reduced calcemic effects and a reduced ability to stimulate rat Cyp24a1 gene expression.[8,266] Detailed reviews of the metabolic pathway and physiological functions of the C3-epimer forms of vitamin D can be found elsewhere. [263,265,267]

The C3-epi-25(OH)D was initially found in neonates. In a 2006 publication by Singh et al., it was reported that the C3-epimer was detectable at a significantly high percentage (up to 60%) in neonates and children up to 1 years of age.[268] In another publication, C3-epi-25(OH)D3 was found to be present in all neonatal samples, but contributed less than 10% of the total 25(OH)D concentration, which was considered by the authors as unlikely to be clinically significant.[269]

Although recently significant concentrations (ranging from 0.1 to 24 ng/mL) have also been reported in adults, highly variable concentrations have been reported by several groups, which makes it difficult to evaluate the real importance of this metabolite in patient samples. [260,270-272] Detectable levels of the epimer range from 0 to 100% in the adults tested and it is estimated that on average, adults have a median concentration range for the epimer of 1.72 (0–9.01) ng/mL and that the epimer makes up to 6.1% (0–47.0%) of total 25(OH)D.[265]

Currently there are several assays that can quantitate the C3 epimer. These assays all implement LC-MS/MS and usually are applications that measure multiple metabolites of vitamin D.[49,265,273,274]

We must note here that the C3-epimer can also interfere with the measurement of 25(OH)D. Although the antibodies used in immunoassays are (supposedly) selective and do not cross-react with the epimer, commercial assays are reported to have variable degrees of cross reactivity (ranging from 0.07% to 91%) with the epimer.[57] However, one study showed that the C3-epimer cross-reactivity as calculated from exogenous 3-epi-25(OH)D3 recovery is different from the cross-reactivity determined from real neonatal samples with endogenous 3-epi-25(OH)D3. This study concluded that the cross recovery based on spiking experiments should be regarded as an in vitro anomaly. [275,276] We might therefore conclude that immunoassays seem to be free from C3-epimer interferences.[57]

Interference with the epimers is also a problem for HPLC and some LC-MS/MS applications. The differentiation of the epimers with LC-MS/MS methods is not routine in clinical laboratories, as many laboratory specialists choose not to distinguish these compounds. This is primarily because the epimers display similar MS/MS spectra and their separation requires extended chromatography times or specialty columns. Although numerous LC methods capable of chromatographic resolution have been published in the last 5 years, time of analysis generally exceeds 10–20 min per sample. A more rapid method based on ion mobility spectrometry (IMS) has also been recently described.[277] This interference with the epimers might be of clinical importance when we measure pediatric samples. However, in adult populations the epimers do not seem to have a significant impact.[278,279]

Currently, there is no reference measurement procedure (RMP) for the determination of 3-epi-25(OH)D3; however, NIST measures the epimer using a method similar to the RMPs, i.e., an ID-LC-MS/MS method that separates 25(OH)D3 and 3-epi-25(OH)D3 using isotopically-labeled 3-epi-25(OH)D3 as an internal standard.

Steroid hormones are highly lipophilic and need a serum carrier protein to ensure effective delivery to target cells. Given the abundance of proteins in serum, some of this transport will be non-specific. However most of their trafficking is accomplished with ligand-specific serum carriers.[280] The majority of circulating 25(OH)D and 1α25(OH)2D is tightly bound to VDBP and less tightly bound to albumin, with less than 1% circulating in an unbound form.[132] Consequently factors affecting VDBP levels alter the interpretation of 25(OH)D levels. Moreover, the free hormone hypothesis postulates that protein-bound hormones are biologically inactive and only unbound hormones (or free) are able to exert their physiologic activity.[281] This hypothesis has been proposed as a universal mechanism for the cellular uptake of steroid hormones mainly because these are highly lipophilic molecules such as 25(OH)D, which have the potential to passively diffuse across cell membranes.[282]

Vitamin D binding protein (VDBP) was discovered with Immunoelectrophoresis in 1959, but it was not until 1975 that its function as a carrier protein of vitamin D metabolites was established.[283,284] VDBP is a protein (alpha globulin) that was initially known as Gc-globulin (group-specific component of serum) and subsequently named VDBP because of its ability to bind the majority of (>85%) circulating 25(OH)D. VDBP is the ligand-specific carrier protein for 25(OH)D and 1α,25(OH)2D, and is a member of the same family of proteins that include albumin.[285,286] VDBP has actions beyond vitamin D binding.[286-288] As only 1–2% of its sterol binding sites are utilized, multiple metabolic roles beyond the transportation of vitamin D metabolites have been described for VDBP. Actin scavenging, modulation of inflammatory processes and innate immunity, binding of fatty acids, and influencing of bone metabolism are some.[285,288,289]

In 1985, the VDBP cDNA was cloned from an adult liver library. The gene that encodes VDBP in humans is located on chromosome 4 at the 4q12-q13 position. It is 35 kb in length and comprised of 13 exons encoding a 474 amino acids peptide, including a 16 amino acid leader sequence that is cleaved before release.[281] The mature VDBP molecule in humans is approximately 58 kDa in size, although differences in glycosylation of the protein for different alleles may alter the actual size, it has 458 amino acids, and possess a highly conserved number of cysteines and disulfide bridges.[281] It shares 25% and 19% sequence identity with albumin and alpha-fetoprotein, respectively. The VDBP molecule is comprised of 3 structurally similar domains. The disulfide bridges are responsible for the formation of three double loops in the three domains. The first domain (between aa 35 and aa 45) is the binding site of all vitamin D metabolites.[281] Differences in glycosylation of the protein for different alleles may alter the actual size. The half-life of VDBP in human plasma is about 1.7 days (markedly shorter than that of 25(OH)D, which is 215 days) and the estimated daily production in an adult person is about 700 – 900 mg/day.[132] Compared with albumin, its concentration is 300 times lower. About 40% of VDBP is intravascular and the remaining 60% is distributed in the interstitial space of various organs, similar to albumin.[132]

Several tissues can express VDBP, but the main site of its production is the liver.[290] The clearing site of VDBP is not fully understood. It is known that VDBP is filtered by the glomerulus and reabsorbed in the epithelial cells of the proximal tubules by a carrier receptor mechanism (megalin) where it is degraded intracellularly.[132] VDBP is produced at stable levels through life. However, its expression is increased by estrogen, as it was found to be elevated in pregnancy and upon oral contraceptive administration.[291-294] The exact mechanism for this induction is not clear, as a response element for the estrogen receptor in the VDBP promoter has not been identified.[281]

Androgens, on the other hand, do not appear to affect VDBP levels.[295] Glucocorticoids and certain cytokines increase VDBP production. Alternatively, primary hyperparathyroidism is associated with a reduction in VDBP levels, contributing to the lower total 25(OH)D levels seen in these patients as the free 25(OH)D levels are not reduced.[296] Vitamin D itself or any of its metabolites do not regulate VDBP production.[297] A more detailed presentation of the factors and conditions that affect the VDBP levels can be found in the article by Jassil et al. [295]

VDBP is a highly polymorphic protein. Gene sequencing has revealed many variations in the VDBP gene. However, the genetic effects of VDBP on25(OH)D are complicated and have not been elucidated completely. VDBP was originally characterized in humans by serum electrophoresis as the product of two autosomal, codominant alleles GC1 and GC2. Isoelectric focusing (IEF) was subsequently used and allowed for the further characterization of two subtypes of the GC1, the “fast” GC1F and the “slow” GC1S, resulting in six common phenotypes.

In addition to the three common alleles, more than 120 variants have been described in humans. However, three major genotypes account for the majority of VDBP variants. The geographic distribution of these variants often correspond to patterns of human migrations and thus are of anthropological interest. GC1F is most abundant among those of African ancestry, whereas GC1S is most abundant in the European populations, with Asians exhibiting intermediate frequencies of both GC1 forms. The GC2 form is exceedingly rare in the black ethnic groups and found in similar frequencies in people of Asian and European ancestry. These VDBP variants exhibit differences in their affinity to 25(OH)D and 1α,25(OH)2D, as well as in their serum concentration with the hierarchy in both cases being GC1F > GC1S > GC2.[298,299] The existence of the different forms of VDBP with different affinities to vitamin D metabolites could be explained by the differences in skin pigmentation of populations living in geographic locations with different UV exposures, which results in different cutaneous vitamin D synthesis. As such, people groups that migrated to northern latitudes were exposed to less UV, resulting in “selective pressure” for lighter skin tones and lower affinity GC1S and GC2 forms, which allowed for greater free vitamin D levels. Conversely, in darker-skinned individuals, the combination of dark skin and the VDBP variant was ideal in geographical areas with latitudes that were closer to equator. However, upon immigration to northern latitudes, a disadvantage may have been present because of a diminished cutaneous vitamin D synthesis imposed by darker skin and reduced free vitamin D levels due to higher affinity GC1F VDBP. Thus, these individuals may require greater vitamin D supplementation to compensate for these genetic and environmental factors.[288]

The mechanism by which the ligand is released from its binding protein and acquired by the target cell is crucial to hormone signaling pathways. This is very important for vitamin D since there is evidence for an extrarenal, intracrine, conversion of 25(OH)D to 1α,25(OH)2D. Factors affecting this conversion include the tissue specific expression of 1α-hydroxylase, the nuclear receptor VDR, and the availability of 25(OH)D.

Although greater than 99% of 25(OH)D circulates in the serum bound to VDBP or other carrier proteins, the general assumption is that biological activity involves the unbound or free fractions, even though this component is very small.[141,300] The free hormone hypothesis has been proposed as a universal mechanism for the cellular uptake of steroid hormones mainly because these molecules are highly lipophilic and have the potential to diffuse passively and rapidly across cell membranes.[282] However, it is still unknown whether the free hormone hypothesis is applicable to 25(OH)D as other free steroid hormones are feedback regulated whereas in case of vitamin D, the active vitamin D 1α,25(OH)2D is feedback regulated and not 25(OH)D.[301] Moreover, this hypothesis has been under “debate” because of the discrepancy between the likely available amounts of free hormone for passive diffusion and the levels required to efficiently occupy the intracellular receptors.[141,280,302,303] A second reservation concerning the free hormone hypothesis is the megalin-dependent uptake of the VDBP-25(OH)D complex into cells. (the role of megalin in proximal tubule protein reabsorption has been reviewed in detail elsewhere: [304,305] Although this type of uptake has a clear role in the renal proximal tubular cell uptake of 25(OH)D, it is not clear whether this mechanism is utilized by other target tissues of vitamin D. Outside the kidney, megalin is expressed by several tissues, including the placenta, mammary gland, parathyroid, and thyroid glands, which also exhibit 1α-hydroxylase activity so that a VDBP-megalin interaction can be used for vitamin D internalization similar to that described for the kidney. Therefore, it is possible that some extrarenal tissues employ the megalin-mediated receptor uptake of 25(OH)D bound to VDBP similar to that found in the proximal tubular cells of the kidney. Most of the extrarenal tissues that do not appear to express megalin or its associated coreceptors are more likely to acquire free or bioavailable 25(OH)D.[303]

In serum, vitamin D metabolites mainly circulate bound to VDBP (85–90%), but they are also known to associate with serum albumin (10–15%). The affinities of 25OHD and 1,25(OH)2D for VDBP are K_a = 6 × 10⁵ M⁻¹ and K_a = 5.4 × 10⁴ M⁻¹, respectively. For albumin, the affinities are substantially lower (K_a = 7 × 10⁸ M⁻¹ and K_a = 4 × 10⁷ M⁻¹, respectively). However, because of the relative abundance of albumin in serum (650 μM) compared to VDBP (5 μM), it seems logical that some vitamin D metabolites are transported in circulation by albumin. Moreover, the vast majority of VDBP in serum circulates unbound because its concentration is much higher than the concentrations of vitamin D metabolites found in circulation. Therefore, it seems likely that most circulating vitamin D metabolites are bound to a carrier protein of some sort. Moreover, at any given time, a small proportion of vitamin D metabolites will not be bound to VDBP or albumin, but will instead be unbound or ‘free’.[280] In the case of 25OHD, it is estimated that less than 0.1% of the total circulating levels of this metabolite are ‘free’.[306] An alternative term, ‘bioavailable’ 25OHD, has also been proposed.[307] Bioavailable 25OHD refers to all the circulating 25OHD that is not bound to VDBP, in other words the sum of free plus that bound to albumin. Calculation of bioavailable 25OHD has been used as an alternative to free 25OHD in some clinical studies.

It is obvious that the variation observed in serum VDBP concentrations under certain conditions results in higher or lower free 25(OH)D levels. These observations have provoked an interest in the measurement of free 25(OH)D since it was hypothesized that it may be a better marker of vitamin D activity than the classical measurement of total 25(OH)D. Free 25(OH)D can be measured either directly or estimated through calculations, based on the measurements of total 25(OH)D, VDBP, and albumin serum levels. In principle, both methods should give similar results, but this not always the case. The direct measurement of free 25(OH)D can be performed either by centrifugal ultrafiltration or by a recently-developed commercially-available ELISA.[170,303] Centrifugal ultrafiltration was a method introduced in the 1980s that did not become commonly used due to its high costs and technical difficulties in application although it was proved to be quite accurate.[141,308,309]

In the early 2010s, a two-step ELISA was developed using monoclonal antibodies that is now commercially available. In this assay, an anti-vitamin D antibody is coated on a microtiter plate. Serum samples and calibrators are pipetted into the wells of the microtiter plate. Free 25(OH)D is captured by the antibody during a first incubation. After washing, a biotin-labeled 25(OH)D analog is allowed to react with the unoccupied antibody binding sites in a second incubation. After a second washing step and incubation with a streptavidin-peroxidase conjugate, bound enzyme is quantitated using a colorimetric reaction. Intensity of the signal is inversely proportional to the level of free 25(OH)D in the sample. The limit of detection of this assay is 2.8 pg/mL. However, the antibody used in the current assay does not equally recognize 25(OH)D2 and 25(OH)D3 (77% of the 25(OH)D3 value). Therefore, it underestimates free 25(OH)D2. However, under most situations where the predominant vitamin D metabolite is 25(OH)D3, this issue is not a major concern. Data from both normal subjects and from subjects with variations in VDBP levels (e.g., cirrhotic patients and pregnant women) correlate well to that obtained from similar populations using the centrifugal ultrafiltration assay.[281,310] However, the accuracy of this assay cannot be determined as no reference method is currently available.

The calculation of the free portion of 25(OH)D requires the measurements of total 25(OH)D, albumin, and VDBP as well as a mathematical equation that will take into account the affinity of both binding proteins.[308,311] According to the theory of protein–ligand binding kinetics, total 25(OH) vitamin D can be defined as follows:[312,313] total 25(OH)D=free25(OH)D + albumin−25(OH)D + VDBP−25(OH)D

Free 25(OH)D can be assessed by the following formula, first suggested by Bikle et al. in 1986 [311] and subsequently used in several studies: Free25(OH)D=(Total25(OH)D)∕(Kalb∗Alb+Kvdbp∗VDBP) where, Free25(OH)D is the calculated concentration of free 25(OH)D in mol/L, K_alb is the affinity constant between albumin and 25(OH)D in mol⁻¹, K_vdbp is the affinity constant between VDBP and 25(OH)D in mol⁻¹, Alb is the concentration of albumin in serum in mol/L, and VDBP is the concentration of total VDBP in serum in mol/L. After the calculation, the value of Free25(OH)D is converted to pmol/L.

After determining the concentration of free 25(OH)D, the bioavailable amount can be determined using the following formula: bioavailable25(OH)D=(Kalb∗alb+1)∗Free25(OH)D where Free25(OH)D is the concentration of free 25(OH)D in mol/L, K_alb is the affinity constant between albumin and 25(OH)D in mol⁻¹, and Alb is the concentration of albumin in serum in mol/L. After this calculation, the value of bioavailable 25(OH)D is converted to nmol/L.

The direct measurement of free vitamin D shows various correlation coefficients for the calculated free vitamin D concentration (e.g., R² = 0.69 in a transgender population; R² = 0.13 in a combined group of pregnant women, patients with liver failure, and control group). [314,315] These correlations are difficult to interpret, as many variables are present (i.e., CV% of the direct free vitamin D, 25(OH)D, VDBP, and albumin measurement) and the quality/accuracy of all these measurements should be taken into account.

It is obvious that in order to use these formulas, one requires an accurate measurement of VDBP, albumin, and the vitamin D metabolite of interest. Moreover, the calculation depends on an assumption that the affinity constants are invariant.

25(OH)D immunoassays, especially automated immunoassays, are limited by their possibilities in displacement of vitamin D from the VDBP [168]. Therefore, the VDBP concentration can influence the vitamin D levels reported using these assays. If immunoassays are used to calculate free 25(OH)D, different results will be generated compared to calculated free 25(OH)D levels determined using LC-MS/MS methods for 25(OH)D.

In initial studies using normal serum, in which VDBP levels were measured with polyclonal assays, the calculated values correlated well with the directly measured values using the centrifugal ultrafiltration method.[141,311] However, commercially-developed VDBP assays that use monoclonal antibodies appear to differ in their ability to detect the different VDBP alleles compared with the polyclonal antibody assays [316,317], to determine the apparent change in the VDBP affinities for the vitamin D metabolites under different physiologic/pathologic conditions [141,309,315], and to calculate values diverged substantially from the directly measured levels.[315]

For the measurement of VDBP, several immunological methods have been used, which included assays using polyclonal antibodies (e.g., radial immunodiffusion, nephelometry, turbidimetry, rocket immunodiffusion, and radioimmunoassay) or commercial ELISA assays that employed either polyclonal or monoclonal antibodies. Given the polymorphic nature of human VDBP and the isotype-specific post-translational modifications (mainly glycosylations), the assays should be able to equally detect all isoforms of VDBP. However, this is not always the case.

In 2013, Powe et al.[128] reported that black Americans compared with white Americans had similar levels of calculated bioavailable 25OHD despite lower levels of total 25OHD. Similar results came from other studies that used the same assay and all seemed to conclude that the best biomarker to test vitamin D status was free vitamin D rather than total 25(OH)D. However, the VDBP assay they used was a monoclonal antibody assay which resulted in lower VDBP levels and therefore an increased bioavailable fraction in subjects of African-American descent (with the GC1F allele) Previous studies with assays using polyclonal antibodies did not find racial differences in serum levels of VDBP. [318,319] In studies that followed using assays that employed polyclonal antibodies or LC-MS/MS methods demonstrated that the monoclonal antibody-based assay discriminated against the GC1F variant and that the differences in VDBP levels between black and white Americans were minimal when VDBP was measured with polyclonal assays. [316,317,320,321] The results of the polyclonal assays were confirmed by direct measurement of free 25(OH)D using a commercial ELISA assay and by an LC-MS/MS method. Thus, although it is not clear whether the different VDBP alleles have different affinities for the vitamin D metabolites, the alleles do affect the results of immunoassays when a monoclonal antibody is used.[320-322] Polyclonal antibody techniques do not exhibit this kind of bias, but the absolute concentration of VDBP differs according to the assay technique.[301,322]

LC-MS/MS methods may avoid the potential bias observed in antibody-based assays because such methods allow for the measuring of peptides from a region of VDBP that is common to all genetic variants and separate from genotype-specific sequences.[320] However, there are other sources of error in this type of assay that may also lead to poor results if not carefully controlled.[323] Aside from the investigation of optimal digestion conditions, the quantitation of different isoforms should take into account the various degrees and sites of glycosylation. Therefore, assay standardization is needed, and it is among the priorities of the IFCC Committee of Bone metabolism, which is in close collaboration with NIST, CDC VDSCP, VDSP, and research laboratories. NIST recently developed a candidate reference method that is able to measure all three major isoforms of VDBP and this method has been used to assign values to reference materials.[323]

In order to have an accurate calculation of free vitamin D metabolites, we need valid estimations of the affinity constants of these metabolites with VDBP at 37 °C. It is understood that 25(OH)D has a high affinity for VDBP (i.e., the K_a is in the nmol/L range, approx. 1.5 × 10⁸ M⁻¹) and that the affinity of calcitriol is somewhat 10 to 100 times lower.[132] Whether the different isoforms of VDBP have different affinities is still a matter of debate. This also needs to be further researched since the affinity constant is incorporated into the equations that calculate free-25(OH)D and affect the accuracy of the calculations.