Abstract
BACKGROUND: Continuous glucose monitoring systems have been validated for eu- and hyperglycemic cats. The FreeStyle Libre 2 (FSL2) is sufficiently accurate in people during hypoglycemia to guide critical treatment decisions without confirmation of blood glucose concentration (BG). OBJECTIVES: Assess FSL2 accuracy in cats with hypoglycemia. ANIMALS: Nine healthy, purpose-bred cats. METHODS: Hyperinsulinemic-hypoglycemic clamps were performed by IV infusion of regular insulin (constant rate) and glucose (variable rate). Interstitial glucose concentration (IG), measured by FSL2, was compared to BG measured by AlphaTrak2. Data were analyzed for all paired measurements (n = 364) and separately during stable BG (≤1 mg/dL/min change over 10 minutes). Pearson's r test, Bland-Altman test, and Parkes Error Grid analysis respectively were used to determine correlation, bias, and clinical accuracy (P < .05 considered significant). RESULTS: Overall, BG and IG correlated strongly (r = 0.83, P < .0001) in stable glycemia and moderately at all rates of change (r = 0.69, P < .0001). Interstitial glucose concentration underestimated BG in euglycemia, but the BG-IG difference was progressively smaller as BG decreased (12.9 ± 12.2, 8.8 ± 11.2, -3.2 ± 7.4, and -7.8 ± 5.2 mg/dL in the ranges of 80-120 [n = 64], 60-79 [n = 29], 50-59 [n = 71], and 29-49 mg/dL [n = 53], respectively). CONCLUSIONS: Although IG underestimates BG throughout most of the hypo-euglycemic range, IG generally overestimates BG in marked hypoglycemia (<60 mg/dL). It is therefore imperative to evaluate FSL2 results in this critical range with caution.