AIM:This study assessed whether the poor correlation between HbA1c and oral glucose tolerance test (OGTT) for dysglycaemia diagnosis may be explained by haemoglobin glycation (HbG). METHODS:A total of 1033 consecutive overweight or obese patients with no known diabetes underwent OGTT and measurement of HbA1c to diagnose diabetes and dysglycaemia (American Diabetes Association criteria). For each OGTT result category, low, medium and high HbG was defined according to the mean HbA1c/fructosamine ratio and mean fructosamine. High HbG was defined as values greater than mean values in each OGTT category for both HbA1c/fructosamine ratio and fructosamine levels, and low HbG was defined as lower values of both. The remaining patients were considered medium HbG. RESULTS:Based on OGTT and HbA1c values, 267 (25.8%) and 443 (42.8%) patients had intermediate hyperglycaemia, and 66 (6.4%) and 95 (9.2%) patients had diabetes, respectively. The results were discordant for intermediate hyperglycaemia or diabetes diagnosis in 41.7% and for diabetes diagnosis in 10.0% of the patients. The proportion of patients with HbA1c≥6.5%, but without OGTT-diagnosed diabetes, was 0%, 3.8% and 32.8% in the low-HbG, medium-HbG and high-HbG groups, respectively. In contrast, the proportion of patients with HbA1c<5.7%, but with an abnormal OGTT, was 30.4%, 11.1% and 0%, respectively. The AUROC of HbA1c to detect OGTT-diagnosed diabetes was better in the medium-HbG group [0.874 (0.816-0.931)] than in those with low or high HbG [0.628 (0.489-0.768); P<0.01]. Only age was independently associated with high-HbG status [10-year OR: 1.3 (1.1-1.5); P<0.0001]. CONCLUSION:Haemoglobin glycation may explain many of the discordant results between HbA1c and OGTT when used for dysglycaemia diagnosis.