Accuracy of a continuous glucose monitoring system in dogs and cats with diabetic ketoacidosis

Objective – (1) To determine the ability of a continuous interstitial glucose monitoring system (CGMS) to accurately estimate blood glucose (BG) in dogs and cats with diabetic ketoacidosis. (2) To determine the effect of perfusion, hydration, body condition score, severity of ketosis, and frequency of calibration on the accuracy of the CGMS. Design – Prospective study. Setting – University Teaching Hospital. Animals – Thirteen dogs and 11 cats diagnosed with diabetic ketoacidosis were enrolled in the study within 24 hours of presentation. Interventions – Once BG dropped below 22.2 mmol/L (400 mg/dL), a sterile flexible glucose sensor was placed aseptically in the interstitial space and attached to the continuous glucose monitoring device for estimation of the interstitial glucose every 5 minutes. Measurements and Main Results – BG measurements were taken with a portable BG meter every 2–4 hours at the discretion of the primary clinician and compared with CGMS glucose measurements. The CGMS estimates of BG and BG measured on the glucometer were strongly associated regardless of calibration frequency (calibration every 8 h: r=0.86, P<0.001; calibration every 12 h: r=0.85, P<0.001). Evaluation of this data using both the Clarke and Consensus error grids showed that 96.7% and 99% of the CGMS readings, respectively, were deemed clinically acceptable (Zones A and B errors). Interpatient variability in the accuracy of the CGMS glucose measurements was found but was not associated with body condition, perfusion, or degree of ketosis. A weak association between hydration status of the patient as assessed with the visual analog scale and absolute percent error (Spearman's rank correlation, ρ=?0.079, 95% CI=?0.15 to ?0.01, P=0.03) was found, with the device being more accurate in the more hydrated patients. Conclusions – The CGMS provides clinically accurate estimates of BG in patients with diabetic ketoacidosis.