Measurement uncertainty for the analysis of serum 25-hydroxyvitamin D: response to Stepman and Thienpont

Abstract

Dear Editor, We thank Drs. Stepman and Thienpont [1] for their comment on our recent article in which we reported “a relatively high measurement uncertainty for the measurement of serum 25(OH)D levels” [2]. Indeed, we found that to be sure that the 25(OH)D serum concentration of a given patient is >80 nmol/L, a value >100 nmol/L should be targeted. Drs. Stepman and Thienpont correctly underlined that this translates into an analytical coefficient of variation (CV) of approximately 8% which compares well with the analytical CV of the measurement of most steroid compounds [3]. Thus, what we called “a relatively high measurement uncertainty” is in fact very common when measuring steroid hormones. We agree that one must aim to achieve the best analytical performances for the measurement of 25(OH)D. However, this measurement is very frequently prescribed and, from a pragmatic point of view, we must deal with the assays that are currently available and take the measurement uncertainty into account for an optimal interpretation of the measured values. According to our data, this means that the “true” concentration of a patient whose measured value is 100 nmol/L for example, could be (grossly) anywhere between 80 and 120 nmol/L. On the one hand, it will be above the desired value of 80 nmol/L, but on the other hand, it may be as high as approximately 120 nmol/L. This seems to alarm Drs. Stepman and Thienpont. However, in our opinion, it cannot be considered as a potential “burden” for the patients, as 25(OH)D serum levels up to 250 nmol/L (and probably largely more) are safe [4]. Moreover, many experts argue for increasing the minimum 25(OH)D levels that defines optimal vitamin D status to at least 100 nmol/L [5–7]. We acknowledge that the 25(OH)D threshold level below which one can consider that the vitamin D status is insufficient is not consensual, as some other experts consider a value of 50 nmol/L as sufficient [8]. However, measurement uncertainty must be applied with any cut-off and, according to our experience, if one wants to ensure that a measured 25(OH)D concentration is really >50 nmol/L, a value of 70–75 nmol/L at least should be targeted. We also agree with Drs. Stepman and Thienpont that averaging 25(OH)D results from repeated sampling when monitoring vitamin D supplementation would significantly (but not completely) reduce measurement uncertainty of 25(OH)D results. However, this would necessitate several blood samples which, in our opinion, is clearly a kind of burden for the patients. Furthermore, 25(OH)D E. Cavalier (*) Department of Clinical Chemistry, University of Liege, University Hospital of Liege, Domaine du Sart-Tilman, 4000 Liege, Belgium e-mail: Etienne.cavalier@chu.ulg.ac.be