Abstract
The recent paper by Roth et al. 1 presents a comprehensive comparison of currently employed 25-hydroxyvitamin D assays using a highly selected patient population. The study was performed for a German population where, as in much of Europe, ergocalciferol (vitamin D2) supplementation is not in widespread use and as highlighted by Roth, the effect of 25-hydroxyvitamin D2 on assay results was not investigated. Thus, there is no confirmation in this study that the German population did not have circulating 25-hydroxyvitamin D2 from sources other than pharmaceutical supplementation.
Given the dramatic increase in vitamin D requests in recent years, 2 the performance of fully automated assays in the Roth study, such as the Roche Vitamin D3 (25-OH) electrochemiluminescence immunoassay (ECLIA), will be of interest to many laboratories. Recently, we evaluated the same ECLIA on a Roche E170 analyser for 54 patient samples from a north London population and 12 samples from the international vitamin D external quality assessment scheme (DEQAS). ECLIA results were compared with high-performance liquid chromatography–mass spectrometry (LC-MS), which resolves both 25-hydroxyvitamin D2 and 25-hydroxyvitamin D3, and was our reference method.
We agree with Roth et al. in that the ECLIA results are highly reproducible, with inter- and intra-assay CVs (4.9–5.6% and 2.2–2.9%, respectively, determined using Roche PreciControl Bone Quality Control samples) agreeing well with manufacturer values and showing a low bias (mean bias: +4.4 nmol/L, limits of agreement: –52.4 to +86.7 nmol/L, as determined by Bland-Altman analysis) against the LC-MS 25-hydroxyvitamin D3 values. While less bias was observed against the DEQAS ALTM (constant bias of +17.03% and a proportional bias of +0.81% using Deming regression) than reported for other automated assays, 3 none of these samples had measurable 25-hydroxyvitamin D2 concentrations. Close to the lower limit of detection (10 nmol/L), ECLIA overestimated 25-hydroxyvitamin D3 in all analyses performed, reducing the number of acutely deficient (<12.5 nmol/L) 4 subjects detected.
The concentrations for our patient samples (mean ±standard deviation) of 25-hydroxyvitamin D2, 25-hydroxyvitamin D3 and total 25-hydroxyvitamin D, as determined by LC-MS, were 2.5 ± 51.6 nmol/L, 62.5 ± 41.4 nmol/L and 65.6 ± 62.1 nmol/L, respectively. Of the 54 patients examined, nine were found to have appreciable 25-hydroxyvitamin D2 concentrations (>20 nmol/L). Five of these nine subjects were 25-hydroxyvitamin D sufficient (>50 nmol/L) 4 yet were misdiagnosed as insufficient (<50 nmol/L) by the ECLIA. One patient, falsely diagnosed as insufficient had a total 25-hydroxyvitamin D concentration of 334 nmol/L, yet was measured as 14 nmol/L by ECLIA. Given the observation that approximately 17% of our study population had significant 25-hydroxyvitamin D2 concentrations, we suggest that inability to detect 25-hydroxyvitamin D2 must still be considered a major drawback of 25-hydroxyvitamin D3 assays for routine testing. 3 While the recent study by Kimball and Veith suggests that the effects of a high vitamin D concentration may vary between individuals, 5 it is possible that low 25-hydroxyvitamin D3 measurements might accompany sufficient total vitamin D concentrations or indeed hypercalcaemia due to 25-hydroxyvitamin D2 toxicity. This may result in inappropriate supplementation, inaccurate monitoring and potential misdiagnosis of clinical hypercalcaemia. Studies have suggested that the reporting of individual values for 25-hydroxyvitamin D2 and 25-hydroxyvitamin D3 can lead to confusion and misdiagnoses among health-care practitioners, 6 a situation which may be compounded if the form of vitamin D supplementation for a given patient is not known.
Further, since there is a lack of clarity as to the bioequivalence, or otherwise, of 25-hydroxyvitamin D2 and 25-hydroxyvitamin D3, 7–9 an assay that measures either metabolite alone would appear to be of limited use. It is suggested that any 25-hydroxyvitamin D assay in routine clinical use must accurately detect both 25-hydroxyvitamin D2 and 25-hydroxyvitamin D3 on an equimolar basis.