Diagnostic performance of glycated haemoglobin (HbA1c) for diabetes mellitus in dogs.

The objective of this study was to determine the diagnostic performance of glycated haemoglobin (HbA1c) for canine diabetes mellitus (DM) and compare it with that of serum fructosamine. Aliquots of blood samples collected for diagnostic purposes from adult dogs were used. HbA1c was measured using a previously validated capillary electrophoresis assay. The dogs were allocated into four groups: (1) DM; (2) hyperadrenocorticism (HAC); (3) long-term corticosteroid therapy (CST); and (4) various chronic diseases (VD). In total, 88 dogs were included as follows: DM (n = 11), HAC (n = 10), CST (n = 14), and VD (n = 53). Fructosamine was measured in all four groups as follows: DM (n = 6), HAC (n = 7), CST (n = 9), and VD (n = 42). Median (range) serum glucose concentration was higher (P < 0.001) in the DM group (22.8 mmol/L; range, 15.6-29.3 mmol/L) compared to HAC (5.9 mmol/L; range, 4.2-6.8 mmol/L), CST (5.6 mmol/L; range, 4.3-23.3 mmol/L), and VD (5.5 mmol/L; range, 4.1-9.4 mmol/L) groups. Mean (± standard deviation) HbA1c was higher (P < 0.001) in the DM group (6.3% ± 1.5%) compared to HAC (1.9% ± 0.5%), CST (1.7% ± 0.5%), and VD (1.9% ± 0.5%) groups. All diabetic dogs and none of the other dogs had HbA1c levels above the cut-off value for DM (3.3%), indicating an accuracy of 100% in diagnosing DM. Significant differences (P < 0.01) were observed in median fructosamine between the DM group (389 μmol/L; range, 348-865 μmol/L) and the HAC (306 μmol/L; range, 167-348 μmol/L) and the VD (316 μmol/L; range, 189-500 μmol/L) groups. Fructosamine had an accuracy of 84.4% for the diagnosis of DM. When used for the diagnosis of canine DM, HbA1c measured with this specific assay had excellent diagnostic accuracy and was superior to serum fructosamine.