Combined measurement uncertainty of HbA1c and glucose using ISO 20914:2019 and outcome-based simulationEmine Feyza Yurton May 14, 2026 at 10:00 am

Clin Chem Lab Med. 2026 May 15. doi: 10.1515/cclm-2026-0388. Online ahead of print.

ABSTRACT

OBJECTIVES: This study aimed to calculate the measurement uncertainty (MU) of HbA_1c and glucose according to ISO/TS 20914:2019 guideline and to evaluate using indirect outcome-based analytical performance specifications (APS).

METHODS: MU was estimated using internal quality control data collected between October 2023 and August 2025. A total of 13,270 and 13,824 IQC results for HbA_1c and 1,329 and 1,208 results for glucose were analyzed for Level 1 and Level 2 controls, respectively. HbA_1c was measured by capillary electrophoresis using two analyzers (12 capillaries each), and glucose was measured using the hexokinase method on three platforms. Maximum allowable uncertainty (MAU) targets were determined using the APS Simulator based on 2,000 patient samples with simultaneous HbA_1c and fasting plasma glucose results. Clinical decision limits were defined as 5.7 % (39 mmol/mol) and 6.5 % (48 mmol/mol) for HbA_1c, and 5.6 and 7 mmol/L for glucose. Overall agreement was used as the primary performance metric.

RESULTS: Relative expanded MU (Urel) ranged from 4.56 to 4.59 % for HbA_1c expressed in %, 6.36-7.87 % for HbA_1c expressed in mmol/mol, and 3.68-4.36 % for glucose. Glucose Urel values remained below the minimum MAU limit (5.1 %), whereas HbA_1c values exceeded the corresponding limits (2.8 and 4.7 % for % and mmol/mol). Sensitivity and specificity ranged from 83.9 to 96 % and 93.4 to 99 % for glucose, and 85.4 to 93.4 % and 89 to 97.6 % for HbA_1c.

CONCLUSIONS: Glucose measurements met the minimum APS targets derived from the simulation model, whereas HbA_1c measurements exceeded these limits, indicating higher MU and a potential impact on patient classification around clinical decision limits.

PMID:42132101 | DOI:10.1515/cclm-2026-0388