Sigma Metric Analysis

With Euromedlab back, the posters and abstracts are flowing in 2022. What do the latest methods and instruments look like for HbA1c?

Sigma metric analysis of HbA1c methods at the EuroMedLab Poster Abstracts

May 2022
Sten Westgard, MS

The key reference works that provide the foundation for this analysis are Basic Method Validation, 4th Edition and Six Sigma QC Design and Control, 2nd Edition.

HbA1c is described as one of the world's best examples for standardization and performance improvement. In the Milan Hierarchy, the top model uses HbA1c as the primary example of how to best develop performance specifications.

The IFCC standardization committee has suggested that for IFCC units (mmol/mmol), the TEa is essentially 10%, with the goal of getting the analytical Sigma-metric above 2 for routine methods, and Sigma-metric above 4 for clinical studies. CAP/NGSP - in NGSP units (% HbA1c) - has set their TEa at 6.0%, but without a specific recommendation of the level of performance achievement.

So, in 2022, are today's new HbA1c methods living up to the goals that were set?

A number of posters from EuroMedLab provide a stark answer. 

Poster Abstract M286: Evaluation of the new enzymatic Vitros A1c chemistry products slides on the Ortho Vitros 5600/XT 7600 Integrated and 4600/XT 3400 Chemistry Systems

G. Snodgrass, M. Barbro, T. Dimagno, J. Fyles, C. Hoffman, B. Johnson, A. Zeberi, Ortho Clinical Diagnostics, Rochester NY 14626

The imprecision data "was calculated using ANOVA across three analyzers each from the VITROS 4600/XT 3400 Chemistry and 5600/XT 7600 Integrated Systems with whole blood patient pools using two replicates per day, twice per day over 20 days (n-80). Accuracy was determined using EDTA venous whole blood patient samples following CLSI EP09c guidelines... against the NGSP secondary reference laboratory Tosoh G8 HPLC."

Poster Abstract M274: Performance Evaluation of the new Siemens Analyzer, Atellica DCA for Point of Care HbA1c measurement

Over the course of a month, 62 whole blood EDTA samples, were measured on both the Siemens Atellica DCA Analyzer and the  A.Menarini Diagnostics Premier Hb9210 analyser.  Imprecision was assess through 3 patient pools at 3 levels run in duplicate twice per day for 21 non-consecutive days.

Poster Abstract M261: Evaluation of Sekisui RC-W HPLC Analyzer for the measurement of HbA1c

Method comparison was performend according to CLSI EP-9, 105 samples run in parallel with ADAMS A1c HA-8180V, Passing-Bablock regression applied to results. Imprecision was evaluated according to CLSI EP-5.

Poster Abstract M027: Assessment of INCLIX HbA1c Point-of-Care Method

44 patient samples were analyzed in parallel on INCLIX and the Olympus AU 480. Dayto-day imprecision was determined with two baseline HBA1c levels (at 6.6% and 7.5%) and run for five days in triplicate

Is this HbA1c performance good or bad?

At the low end, most of these new methods do not meet even the rather generous goals that have been set. Achieving >2 Sigma in the view of IFCC might be acceptable, but in the normal Six Sigma world, 3 is the minimum performance, not the goal.

Since the development of QC Design and Frequency tools, we can also determine that these 3 Sigma methods will require all the Westgard Rules, large numbers of control measurements (2, 4 even 6 control measurements per run), and those controls will need to be made every 45 patient samples or even more frequently for 2 Sigma methods. What the IFCC goal means is that poor methods are "acceptable", but the QC required to adequately monitor those methods is essentially impossible.

HbA1c methods have indeed made great progress in the last 20 years, and continue to make progress now, but either the TEa goal needs to be adjusted, or the engineering and performance of these methods needs another significant improvement.