Abstract
Anaemia is a common finding in critically ill dogs, and understanding its potential impact on the accuracy of point-of-care glucometers is essential. In human medicine, it is well reported that haematocrit can affect blood glucose measurements and influence clinical outcomes when measured using certain hand-held glucometers, whereas evidence in veterinary medicine is still scarce. This prospective case-control study included 72 client-owned dogs, 39 of which were anaemic. Haematocrit values were used to classify dogs into anaemic and non-anaemic categories. Whole blood glucose concentrations were measured using two new models of veterinary hand-held glucometers (G-PET PLUS(®) and AlphaTrak3(®)) and compared with the serum blood glucose measurements obtained with an automated wet biochemistry analyser. Agreement between the point-of-care blood glucose analysers and the reference method, as well as the effects of haematocrit on measurement accuracy, was assessed. Both hand-held glucometers had a significant positive correlation with the wet biochemistry analyser (AlphaTrak3(®): r = 0.664; 95% CI 0.502, 0.78; G-PET PLUS(®): r = 0.769; 95% CI 0.643, 0.855). However, when comparing the overall blood glucose results, there was a significant negative percentage bias (p < 0.001) for both devices. Haematocrit reductions did not affect the degree of bias for AlphaTrak3(®), and for the G-PET PLUS(®), bias was also present in control patients and in patients with a mild haematocrit reduction. Despite these statistical differences, clinical impact assessments of the results via specialist clinician questionnaires and Parkes error grid analysis showed no change in clinical decision-making or patient outcome, supporting the cautious use of these devices in anaemic but normoglycaemic dogs when testing whole blood EDTA samples.