Variability in the analysis of 25-hydroxyvitamin D by liquid chromatography–tandem mass spectrometry: The devil is in the detail

Abstract

Background

Liquid chromatography–tandem mass spectrometry (LC–MS/MS) is increasingly used in clinical laboratories for the analysis of 25-hydroxyvitamin D (25OHD), but measurement is not straightforward. Importantly, LC–MS/MS is not a single technique: variables in sample preparation, chromatography and ionisation/fragmentation should each be considered.

Methods

We analysed results from a survey organised by the international Vitamin D External Quality Assessment Scheme (DEQAS), to determine the influence of such variables on the results for two DEQAS distributions.

Results

65 laboratories returned questionnaires. 346 (57%) individual results were from laboratories using electrospray ionisation (ESI), and 259 (43%) from laboratories using atmospheric pressure chemical ionisation (APCI). Although the mean ratio of results was not significantly different between ESI and APCI (P = 0.5828), there was greater variation (P < 0.0001) in results obtained by laboratories using ESI. Greater variation (P < 0.05) was also observed between results from laboratories monitoring non-specific water-loss transitions. Only 3 laboratories (5%) could resolve the isobaric metabolite 3-epi-25OHD₃ from 25OHD₃.

Conclusions

There are many variables to consider when using LC–MS/MS, including assay standardisation/calibration, chromatography and MS conditions. MS/MS alone cannot distinguish isobaric metabolites such as 3-epi-25OHD₃. Interference can also occur if non-specific transitions are used. Laboratories should always subscribe to an EQA scheme for 25OHD analysis.

Introduction

‘Vitamin D’ is a term used to describe a group of seco-steroid compounds, the most important of which are cholecalciferol (vitamin D₃) and ergocalciferol (vitamin D₂). Often, the term is incorrectly used to incorporate a number of related vitamin D metabolites, including 25-hydroxyvitamin D (25OHD) and the active hormone 1α,25-dihydroxyvitamin D amongst others. Measurement of total serum 25OHD (i.e. 25-hydroxyvitamin D₂ and 25-hydroxyvitamin D₃ — 25OHD₂, 25OHD₃) is universally considered a reliable and robust marker of vitamin D status and for monitoring supplementation in vitamin D deficient subjects, since this concentration reflects both dietary and/or supplementary intake and dermal production [1].

It is known that severe deficiency (total serum 25OHD < 10 ng/mL) leads to rickets in children and osteomalacia in adults [2]. The concentrations of total 25OHD that relate to adequacy and sufficiency are less well defined, and are a subject of ongoing debate, although current opinion is that the optimum concentration of 25OHD₃ should be greater than 30 ng/mL [3], [4]. Vitamin D status has been the focus of much recent literature attention and has now been linked to a range of pathologies, including heart disease, hypertension, diabetes, cancer and autoimmune diseases [5]. As a consequence, there has been a dramatic increase in the number of requests for the measurement of serum 25OHD [6]. In order to meet demands, many laboratories considered high-throughput, automated immunoassays to replace more laborious solvent extraction methods, such as the initial competitive protein binding assay reported by Haddad and Chyu [7]. However, method differences between laboratories soon became apparent [8], [9], [10], [11], [12], as was shown by the results of the Vitamin D External Quality Assessment Scheme (DEQAS), which was established in 1989. Immunoassays for 25OHD have an inherent problem of inability to differentiate between a myriad of polar metabolites and vitamin D-like seco-steroids which contribute to the total 25OHD assay. In addition, analytical methodology details, including assay calibration, are not always disclosed because of commercial sensitivities.

There has also been a move in the last decade towards clinical laboratories using liquid chromatography–tandem mass spectrometry (LC–MS/MS) for a number of analyses which were previously performed using immunometric techniques [13]. LC–MS/MS can distinguish and independently quantify, for example, 25OHD₂ from 25OHD₃ by mass-to-charge ratio (m/z) alone. Since the first reported use of LC–MS/MS for the analysis of 25OHD by Watson et al. [14], improvements in the automation of sample preparation, the speed of chromatographic steps, and the sensitivity of MS instrumentation have meant that LC–MS/MS assays with higher sample throughput are increasingly being adopted, and the technique now accounts for over 10% of all DEQAS returns (October 2010). LC–MS/MS is considered by some the future ‘gold standard’ for analysis of 25OHD [15], [16], [17], [18].

However, LC–MS/MS for the analysis of serum 25OHD is not straightforward. Initially, inter-laboratory (% CV) agreement between LC–MS/MS users was poor. Some of the variation could be explained by the use of ‘in-house’ calibration standards, the preparation of which varied between laboratories. Indeed, in a DEQAS study reported by Carter and Jones [19], it was found that use of a common standard (from Chromsystems® — now traceable to the National Institute of Standards and Technology reference material, NIST SRM 972) improved the mean inter-laboratory imprecision for total 25OHD from 16.4% to 10.4%. Better insight into the importance of assay standardisation and other factors affecting LC–MS/MS methods, such as the tubes used for sample collection and preparation [6], [20] and interference from other vitamin D metabolites has further improved inter-laboratory agreement, but variability between laboratories does still exist and requires investigation. Importantly, LC–MS/MS is not a single, ‘off-the-shelf’ technique: variability in sample preparation, chromatographic separation and finally ionisation/fragmentation should each be considered. That said, the flexibility of LC–MS/MS systems may be beneficial, since there is the opportunity to adapt, individualise and standardise methods.

We analysed the results from a survey of LC–MS/MS users organised by the international Vitamin D External Quality Assessment Scheme, to determine the influence of such variables on the reported results for two DEQAS distributions (10 samples, July and October 2010).