Beware of Noncommutability of External Quality Assessment Materials for Hemoglobin A1c

To the Editor: A recent paper reported the results of the EurA1c study, a European trial that investigated the performance of hemoglobin A1c (HbA1c) assays from 24 manufacturers across 17 countries (1,). In that study, fresh whole blood and lyophilized hemolysate specimens manufactured from the same pool were used by 17 external quality assessment (EQA) organizers to evaluate analytical performance of 2166 laboratories. Results showed that there were no substantial differences in performance between groups using fresh whole blood or lyophilized hemolysate samples. A +0.2 mmol/mol bias over 1517 laboratories using fresh whole blood material and a −0.5 mmol/mol bias across 649 laboratories using the lyophilized version of the same pool were reported. Although a slight difference of bias was observed, these results could give the impression that lyophilization did not alter commutability of quality-control materials for HbA1c. Although we agree that lyophilized hemolysates of adequate commutability can sometimes be obtained, the results of a multicenter study that we recently performed actually suggested that commutability of processed EQA materials was highly heterogeneous and often insufficient to rigorously assess trueness of HbA1c assays (2,). In our study, the difference in bias approach (3, 4) was used to assess commutability of 23 processed quality-control materials for 17 of the most frequently used HbA1c assays, including immunoassays, enzymatic assays, ion-exchange HPLC, boronate affinity HPLC, and capillary electrophoresis. Our results showed that only a few lyophilized EQA materials were commutable for most assays and that commutability of lyophilized EQA materials was globally better for some assays than for others. The first consequence of commutability not being the same for the different assays is the inability of the EQA scheme to evaluate the ability of the different assays to provide equivalent results. In our study, mean intermethod CV was 3.2 ± 1.1% with fresh whole blood samples and 8.0 ± 5.8% with processed EQA materials (2). Use of noncommutable materials thus depicted an overly pessimistic picture that incorrectly suggested that standardization of HbA1c assays had not been achieved, which was not actually the case. In addition, heterogeneity in commutability levels among the different EQA materials resulted in discrepant estimates of the agreement between HbA1c assays. The second consequence of noncommutability of EQA materials is to skew trueness assessment of HbA1c assays when results are compared with those of the IFCC reference measurement procedure. For a given assay or group of assays, noncommutability of materials randomly resulted in a positive, negative, or negligible shift in bias, meaning that bias could be overestimated, underestimated, or correctly estimated in an unpredictable manner. This observation confirmed that heterogeneity in commutability of EQA materials made it very risky to assess trueness of HbA1c assays with quality-control materials for which commutability had not been properly evaluated (5). In Fig. 1, the noncommutability bias—the difference in bias as defined in IFCC recommendations (4)—of 23 lyophilized EQA materials is represented for 9 groups of assays representing a total of 16 assays. Although the mean noncommutability bias seemed reasonable for most assays (−1.1%, −1.0%, −0.5%, +1.2, +2.1% for Sebia, Arkray, Tosoh, Trinity, and Roche assays, respectively), noncommutability consistently resulted in a marked negative shift for some assays (e.g., mean noncommutability bias of −14.2% for the Abbott enzymatic assay) and a marked positive shift for other assays (e.g., mean noncommutability bias of +12.5% for the Siemens DCA Vantage assay). Calculating the mean noncommutability bias of all EQA materials for all assays would provide a misleading mean bias, but a closer look at detailed results actually demonstrated that commutability of EQA materials for HbA1c was heterogeneous and skewed trueness assessment differently depending on the assay. Noncommutability bias of 23 lyophilized EQA materials for HbA1c. Noncommutability bias (difference in bias) of 23 lyophilized EQA materials is shown for 9 peer groups of HbA1c assays: Abbott enzymatic (Architect C-4000), Sebia (Capillarys 2, Capillarys 3, and Minicap), Arkray, (Adams A1c HA-8180T) Tosoh (G8, G11, Gx), Trinity Biotech (Premier Hb 9210), Roche (Cobas C513), Beckman Coulter (DxC700AU B00389), Bio-Rad (VII NU, VII Turbo, D10, D100), Siemens (DCA Vantage). Our data support that a comparison of results obtained from fresh whole blood and lyophilized hemolysates should not be performed by calculating the mean results from all HbA1c assays together; this could hide different susceptibility of HbA1c assays to noncommutability of EQA materials and erroneously lead to the conclusion that lyophilized hemolysates are equivalent to fresh whole blood. More generally, we recommend that commutability of EQA materials for HbA1c should not be assumed and should be evaluated systematically when the goal is to evaluate bias of assays and/or agreement between results provided by the different assays. When commutability has not been evaluated, using EQA materials consisting of fresh whole blood remains the best option. Hb A1c, hemoglobin A1c; EQA, external quality assessment. All authors confirmed they have contributed to the intellectual content of this paper and have met the following 4 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved. P. Kaiser, statistical analysis. No authors declared any potential conflicts of interest.