Objective - To evaluate the applicability and accuracy of a continuous glucose monitoring system (CGMS) in critically ill foals by comparing the performance of the CGMS, a point-of-care (POC) glucometer, and an in-house standard laboratory assay (SLA). Design - Prospective study. Setting - University teaching hospital. Animals - Seven critically ill neonatal foals requiring intensive care. Interventions - Foals were instrumented with a CGMS that measured interstitial glucose concentration every 5 minutes for the duration of 15-79 hours. Capillary and venous blood samples were taken every 4-6 hours for POC and SLA measurements, respectively. Measurements and Main Results - Bland-Altman analysis showed a mean bias (95% limits of agreement) of -0.1 (-3.9 to 3.5) mmol/L for comparison of CGMS versus SLA, 0.06 (-3.9 to 4.0) mmol/L for comparison of CGMS versus POC glucometer, and -0.16 (-1.8 to 1.5) mmol/L for comparison of POC glucometer versus SLA. Percent agreement and weighted kappa for classification in hypoglycemia, normoglycemia, and hyperglycemia were 68.4% and 0.296 for CGMS versus SLA, 72.4% and 0.442 for CGMS versus POC glucometer, and 80.7% and 0.568 for POC glucometer versus SLA. Conclusions - The CGMS may be helpful for monitoring a trend in interstitial glucose concentration in critically ill neonatal foals. However, considering the wide limits of agreement between methods, the CGMS should only be used as an adjunctive device to other, more accurate and readily available methods that are able to detect acute changes in glucose concentration. Its use is further limited by the relatively high costs of the sensors, the mandatory 2-hour initialization period, and the difficulties of keeping the transducer in place in an active foal. The POC glucometer used in this study is easy to use and proved to be sufficiently accurate for repeated, stall-sided glucose monitoring in neonatal foals.
