A new study published in the Diabetes Technology & Therapeutics has found that personal continuous glucose monitors (CGMs) worn by diabetic patients during hospitalization offer reliable glucose readings, closely aligning with laboratory and point-of-care (POC) measurements in noncritical inpatient settings.

Researchers enrolled adult patients with diabetes who were already wearing outpatient-inserted CGMs at the time of hospital admission. The study analyzed 3,316 CGM-POC reading pairs from 101 patients and 771 CGM-Lab reading pairs from 97 patients, collected within five-minute intervals.

For CGM-POC comparisons, the mean absolute relative difference (MARD) was 13.7% overall, with 14.4% for Dexcom and 11.8% for Libre sensors. MARD increased to 22.6% during hypoglycemia (glucose <70 mg/dL), indicating less precision in low glucose ranges. For CGM-Lab comparisons, MARD was 13.6% overall.

Over 98% of both CGM-POC and CGM-Lab pairs fell within zones A and B of the Clarke Error Grid, indicating clinically acceptable accuracy. No significant correlation was found between CGM accuracy and vital signs such as blood pressure, hemoglobin levels, glomerular filtration rate, or oxygen saturation.

The study also assessed in-hospital glycemic metrics. Time below range (TBR) averaged just 2.2%, time in range (TIR) was 58.7%, and time above range was 39.9%. Glucose variability was 31.2%. TIR levels met inpatient consensus recommendations, and overall CGM performance was comparable to hospital-inserted devices.