Quality specifications and their daily application to evaluate the accuracy of reference measurements for serum concentrations of 25-hydroxyvitamin D3 and 25-hydroxyvitamin D2

Abstract Background Reference measurement procedures (RMP) have rigorous accuracy specifications. For total 25-hydroxyvitamin D, 25(OH)D, bias ≤1.7% and CV ≤5% are recommended. These quality specifications are impractical for minor analytes, such as 25(OH)D 2 . Furthermore, documentation on RMP quality performance specifications for the individual 25(OH)D metabolites and their daily application are missing. Methods To assess accuracy, we used zeta-scores. Daily, 5–10 specimens (duplicate) and 3 reference materials (singleton or duplicate) were measured for 25(OH)D 3 and 25(OH)D 2 using JCTLM-accepted LC-MS/MS RMPs. Protocols were repeated on 3–4 occasions to generate campaign results. We used separate zeta-score acceptability criteria for daily (≤|2|) and campaign (≤|1|) evaluations. Allowable imprecision was determined experimentally. Results Across 7 campaigns, unacceptable daily zeta-scores required repeating 2 runs for 25(OH)D 3 and 5 runs for 25(OH)D 2 . Hence, the zeta-scores of acceptable reference material results indicated high accuracy. The allowable imprecision for the RMPs was ≤5% (daily) and ≤ 3% (campaign) for 25(OH)D 3 and ≤ 7% (daily) and ≤ 4% (campaign) for 25(OH)D 2 , respectively. Conclusions Using zeta-scores and experimentally derived imprecision, we developed a straightforward approach to assess the acceptability of individual 25(OH)D reference measurements, providing also much-needed practical accuracy specifications for 25(OH)D 2 .