| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
1 Clinical Toxicology Laboratory, University of Wisconsin, Hospital and Clinics, Madison, WI
Departments of2
Pathology, and 3
Medicine, Osteoporosis Clinical Center, and Research Program, University of Wisconsin-Madison, Madison, WI
4 Geriatric Research, Education and Clinical Center, William S. Middleton Veterans, Affairs Medical Center, Madison, WI
aAddress correspondence to this author at: Toxicology Laboratory, Rm D4/245, 600 Highland Ave., Madison, WI 53792.
To the Editor:
Dr. Schmidt’s comments give the false impression that our recently published HPLC method (1) is unable to identify the presence of the C-3 epimer apart from 25(OH)D3 or 25(OH)D2. We routinely achieve levels of chromatographic resolution illustrated in Fig. 1
. Chromatogram A displays elution of the C-3 epimer alone. Chromatogram B shows separation of 25(OH)D3 and its C-3 epimer, a scenario that would be observed for samples from infants <1 year old. Chromatogram C is from a patient receiving vitamin D2 as a supplement. Here the compounds are not baseline resolved, but each has a different retention time, and clearly they do not coelute. Most importantly, the C-3 epimer is not mistaken for 25(OH)D2. Our method has detected the C-3 epimer in nearly 25% of the average 650 samples we test each month. Concentrations of the C-3 epimer are usually <10 µg/L. Thus, we can analyze samples from infants <1 year of age (as well as older patients) without modifying the method. In contrast, Singh et al.(2) require 2 liquid chromatography tandem mass spectrometry (LC-MS/MS) procedures, one for samples from patients <1 year of age and another for patients >1 year. Singh et al.(3) have also reported that the C-3 epimer is "indistinguishable from 25(OH)D3 by most LC-MS/MS assays". In contrast, our HPLC method is one of a small number that can effectively resolve the C-3 epimer from 25(OH)D3 and 25(OH)D2.
|
The integrity of our comparison data (HPLC vs Diasorin RIA) has also been brought into question by Dr Schmidt. Interestingly, Diasorin, with whom Dr Schmidt is associated, generated the RIA data for our study. Binkley et al. (4) report that the reproducibility of the Diasorin RIA measurements is highly dependent on the laboratory in which the assay was performed. Furthermore, other investigators have compared Diasorin RIA vs HPLC(5) or LC-MS/MS(6)(7)(8). On close examination of these reports, inconsistencies appear in the interpretation of the statistical data, which often lack a regression plot of the data(5)(7)(8) or a statistic Sy/x, often called standard error of estimate, to describe the scatter about the regression line(5)(7)(8). Regression formulas and correlation coefficients (r2) appear to be the sole means by which data were interpreted with little attention to scatter. An incomplete picture of the comparative data can result from these omissions. Moreover, Turpeinen et al.(5) report that "several samples, however, displayed very large differences" between Diasorin RIA and their HPLC method. Mausell et al.(6) report significant scatter in the Bland–Altman plot encompassing a range of +25 to –40 nmol/L (+10 to –16 µg/L) for Diasorin RIA vs LC-MS/MS. These results are much in line with our report of an Sy/x 
18 nmol/L (7.3 µg/L) for HPLC vs Diasorin RIA. Furthermore, the Binkley study(4) that Dr. Schmidt refers to describes the discordant results among laboratories and methods for the determination of 25(OH)D. Their data do not allow us to draw any conclusions about the accuracy of the Diasorin method. Overall, our results are not significantly different from the data reported by other investigators.
The International Vitamin D External Quality Assessment Scheme (DEQAS) proficiency survey (9) is a valuable resource for assessing peer performance in measuring 25(OH)D and is based on how close a laboratory’s test result is to the all-laboratory trimmed mean. Unfortunately, the receipt of a passing grade does not scientifically support a method as being accurate and could give a false sense of security. The DEQAS program describes the relative performance among participants and, unless specifically controlled with traceable accurate reference standard materials for 25(OH)D3 or 25(OH)D2, such assessments present only a relative indication of performance. Accordingly, Dr. Schmidt’s judgment that the Diasorin RIA is accurate based on favorable DEQAS results is unwarranted and an overinterpretation of the proficiency testing data. We, too, have received passing grades. Acceptable performance in proficiency testing programs is not necessarily a measure of accuracy. A global effort to produce and use serum-based standard materials for 25(OH)D3 and 25(OH)D2 is essential to properly challenge all methods for accuracy. Currently, NIST is in the process of producing these standard materials.
Finally, we applaud Diasorin for taking steps to produce controls that should better describe the comparative performance of their 25(OH)D RIA and LIAISONTM immunoassays, and we look forward to the published results.
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  G. D. Carter 25-Hydroxyvitamin D Assays: The Quest For Accuracy Clin. Chem., July 1, 2009; 55(7): 1300 - 1302. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
