Agreement Between Glycated Haemoglobin and Fasting Plasma Glucose in Screening for Diabetes

Abstract

Fasting plasma glucose and glycated haemoglobin (HbA1c) are the two blood tests most often used to screen for diabetes, and since 2009 both have carried an agreed diagnostic threshold. They are not interchangeable. Glucose reports a single moment after a night without food; HbA1c reflects glycation of red-cell haemoglobin over the preceding weeks and is therefore exposed to everything that changes red-cell biology. This narrative review asks how well the two tests agree, why they sometimes do not, and what disagreement means for screening. It draws on published diagnostic-criteria documents, population surveys, laboratory-standardisation work, studies of ethnic and haematological influences, and prospective studies of risk, and it reports no new measurements. Across settings the tests identify overlapping but different groups of people, and the direction and size of the difference depend on age, ancestry, anaemia and iron status, haemoglobin variants, assay quality and the threshold chosen. Agreement statistics are also easy to over-read, because kappa and correlation are sensitive to prevalence and to range. Evidence that screening itself improves outcomes at population level remains limited. The review concludes that HbA1c and fasting glucose are best treated as complementary, that a discordant result is information and not an error, and that studies of agreement should report methods, populations and sample handling in full.