Evaluation of a point‐of‐care blood glucose monitor in healthy goats

OBJECTIVE: To assess agreement between a point-of-care glucometer (POCG) and a laboratory chemistry analyzer for blood glucose measurements in goats. DESIGN: Prospective study. SETTING: University teaching hospital. ANIMALS: Eighteen healthy adult goats. INVESTIGATIONS: Whole blood samples were obtained via jugular venipuncture prior to premedication with xylazine and butorphanol (T0), following premedication (T20), and after 1 hour of inhalant anesthesia (T60). Each sample was tested with a POCG and a laboratory analyzer (HITA). Agreement was assessed using concordance correlation coefficients and calculation of bias and 95% limits of agreement. MEASUREMENTS AND MAIN RESULTS: Mean blood glucose concentration at T0 was 3.9 ± 0.6 mmol/L (70 ± 10 mg/dL; POCG) and 2.9 ± 0.4 mmol/dL (53 ± 8 mg/dL; HITA). Glucose concentrations at T20 were 6.7 ± 2.4 mmol/L (121 ± 43 mg/dL) and 5.4 ± 2.1 mmol/L (97 ± 37 mg/dL) and at T60 were 5.7 ± 1.7 mmol/L (102 ± 31 mg/dL) and 4.7 ± 1.3 mmol/L (85 ± 24 mg/dL) when measured with the POCG and HITA, respectively. The POCG overestimated blood glucose compared to the HITA. The bias ± SD was 1.08 ± 0.53 mmol/L (19.4 ± 9.5 mg/dL) (95% LOA 0.04 to 2.11 mmol/L [0.7 to 38.0 mg/dL]) and the concordance correlation coefficient was 0.82. After correcting the results of the POCG using a mixed-effects linear model, the bias was 0.0 ± 0.38 mmol/L (0.0 ± 6.8 mg/dL) (95% LOA ± 0.74 mmol/L [± 13.4 mg/dL]) and the concordance correlation coefficient was 0.98. CONCLUSIONS: The POCG overestimated blood glucose concentrations in goats, compared to the HITA, but when the POCG concentrations were corrected, the agreement was excellent.