Vitamin D Testing – What About Reliability?

Following a series of questions raised by colleagues concerning the accuracy of vitamin D testing we asked Doctors Data, based in the USA to respond to some of the general concerns raised, as their lab has recently undertaken an extensive review of the methodologies utilised, to prepare for the next generation of vitamin D testing.

This is their reply:

Authored by: David Quig PhD, DACBN & Jack Maggiore PhD, MT(ASCP)

Concern has been raised about the reliability of test results for 25-OH vitamin D from both serum and dried blood spots samples.[i],[ii]  Recently a plethora of scientific research has increased knowledge regarding key roles of vitamin D in health and disease, and sparked a tremendous increase in demand for laboratory analysis of vitamin D status.  In effort to meet the increased demand several automated high throughput vitamin D assays have become available.  Unfortunately several commonly used immunoassays, one of which has been utilised extensively in the past to establish reference ranges, have been found to be inaccurate and associated with high inter-laboratory variability.

Doctor’s Data, Inc. (DDI) has developed very precise and accurate liquid chromatography-tandem mass spectrometry (LC-MS/MS) method for 25-OH vitamin D₂ and D₃ levels from both serum and dried blood spot samples. The assay performance data provided below provides assurance that clinicians can have full confidence in DDI’s vitamin D results for their patients using both biological matrices.

Clinicians rely on precise and accurate laboratory results for total 25-OH-vitamin D as well as 25-OH D₂ and D₃ , and the U.S. Food and Drug Administration (the USA equivalent of the MHRA in the UK) stipulates that vitamin D assays must detect both 25-OH D₂ and D₃.² The isotope dilution LC-MS/MS approach is currently considered to be the “gold standard” for measurement of vitamin D status. In effort to circumvent the significant intra- and inter-laboratory variability in vitamin D results, the U.S. National Institute of Standards and Technology (NIST) developed a candidate reference procedure that incorporates LC-MS/MS analysis. Further to that end NIST has made available a standardised reference material (SRM 968e, Fat Soluble Vitamins in human serum) for use in standardisation of LC-MS/MS methods and results within and across laboratories.

The latter is important because the results of one recent study (n = 50) suggest that results obtained using routine LC-MS/MS methods may be about 11% higher than those measured using the NIST candidate reference LC-MS/MS procedure.¹  It is noted that currently no certified reference materials (CRM) are available for 25- OH vitamin D, but the NIST SRM 968e is the best alternative available.

DDI utilises a proprietary method for the extraction of the 25-OH Vitamin Ds from serum and dried blood spots prior to analysis by LC-MS/MS. To correct for sample extraction and analytical variability, aliquots of deuterated 25-OH D₂ and D₃ internal standards are added to all samples, control specimens and calibration standards for every analytical run.  Intra-assay precision and accuracy of the DDI assay for serum 25-OH D₃ was determined by repeat analysis of three different NIST SRM samples over four consecutive days.

The values for the three NIST SRM standards were established using the NIST candidate reference LC-MS/MS method. The data presented in Table 1 indicate that the DDI method for serum is very precise and accurate maintaining 97-100% concordance, even at a very low level of 25-OH D₃. The data also indicate that there does not appear to be any bias inherent in the DDI assay.  We were not able to evaluate the performance of our assay for 25-OH D₂ in this study because the NIST reference materials do not contain appreciable levels of that form of vitamin D.

25-OH D₃

Table 1.  Concordance of serum 25-OH D₃ levels in NIST Standard Reference Materials with values obtained from the DDI LC-MS/MS method.

Participation in external proficiency testing programs is a means by which laboratories can evaluate the accuracy of analytical methods on a given day, and see how they compare with other labs using similar or different methodologies.  The College of American Pathologists (CAP) sends serum specimens to laboratories for analysis and the lab reports its values obtained to CAP prior to notification of the expected results.  Currently CAP is evaluating labs on a voluntary basis only for total serum 25-OH-Vitamin D (Accuracy Based Vitamin D, ABVD).  Shown in Table 2 are the results of the five proficiency tests that DDI participated in during 2012.  All DDI results for serum total 25-OH-D were within +/-  6.1 % of the reference values, and within +/- 7.2 % of values reported for other laboratories that use LC-MS/MS methods (peer group, n = 61).  No apparent bias was observed for the DDI results, or among the peer group, against the reference values.

Total 25-OH- Vitamin D

Table 2.  DDI  CAP  accuracy-based proficiency test results  for 2012.

It is very important to clinicians that laboratories are able to reproduce precise values from day-to-day and across extended periods of time. Table 3 confirms that DDI has excellent precision with respect to measurement of 25-OH D₃ in serum and dried blood spots; cumulative interassay CVs of under 3% and 4%, respectively.   Such low interassay variability provides confidence that clinicians can rely on the DDI vitamin D tests to monitor the efficacy of therapeutic intervention to affect patients’ 25-OH D₃ levels.

Table 3.  Interassay cumulative CVs for 25-OH D₃  in serum and blood spot specimens.

There is considerable demand for a precise and accurate method for assessing vitamin D status from dried blood samples (finger prick) collected by patients or clinicians. Dried blood spots have been validated to be an appropriate means of sampling for accurate analysis of several (e.g. insulin), but not all circulating metabolites. In a small study blood spot cards were meticulously prepared under highly controlled conditions in a laboratory setting prior to shipment to a second laboratory.[iii]  The levels of 25-OH-D from the blood spots and serum from each subject (n = 20) were analyzed by an LC-MS/MS method and statistically similar levels of total 25-OH D were reported for the paired matrices.

There are several inherent problems associated with collection of blood spots that can significantly affect test results.  Therefore DDI developed a robust blood spot test for vitamin D that provides results deemed equivalent to those from serum.

In the clinical setting blood spots were collected (finger prick) by 31 subjects and serum was obtained after blood was drawn by a phlebotomist.  All samples were sent to DDI for analysis of Total 25-OH D.  Figure 1 indicates that there was excellent agreement for the results from the blood spots and serum, and there was no apparent systematic bias.  The data provide evidence that substantially equivalent results for 25-OH D status can be obtained in the real life clinical setting for patients from self-collected blood spots or serum.

Figure 1. Blood spot verses serum total 25-OH vitamin D (LC-MS/MS).

To further evaluate the concordance of DDI results for the two collection methods similar testing was performed for an additional 46 subjects. In the second study 26 subjects self-collected blood spots and 20 subjects had the finger stick and blood spotting performed by a clinician; all serum samples were prepared by a professional phlebotomist.  The self-collection arm of the study was performed under observance of a silent investigator in order to note potential common errors associated with self-collection of blood spots.

The observations have been utilised by DDI in preparation of detailed collection instructions that facilitate proper sample collection.  Figures 2 and 3 depict and confirm the results of the previous DDI study and indicate excellent agreement between a patients levels of Total 25-OH D and 25-OH D₃ ,as determined from dried blood spots and serum.   We were unable to perform meaningful correlation analysis for 25-OH D₂ because, as expected, too few subjects had reportable levels of D₂. As a result of explicit collection instructions, to date DDI has had to reject only about 1% of blood spot card samples.