380-P: Hypoglycemia among Nursing Home Residents with Diabetes Detected by Continuous Glucose Monitoring (CGM)

Few data exist on the frequency of hypoglycemia among nursing home residents with diabetes.

We recruited patients 65 years or older, with type 1 or type 2 diabetes, on standing insulin or sulfonylureas, who were long-term residents at 10 Connecticut nursing homes. After informed consent, Freestyle Libre CGM devices were placed for 14 days of blinded monitoring. After omitting the first 24 hours of data, we calculated the percentage (%) of 24-hr periods with at least 2 consecutive blood glucose (BG) levels <70, <60, <50 mg/dl for all patients with at least 7 days of data. We examined the association between selected variables and the % of 24-hr periods with hypoglycemia using Spearman rank correlation and Wilcoxon rank sum tests.

Among 35 patients who completed the study, 33 had at least 7 days of CGM data. Median age was 80 [IQR 71-86], 46% men, 36% nonwhite, HbA1c 7.3 [IQR 6.5-8.5], ADL disability score 26 [IQR 22-27, max possible value=35], 52% had dementia, 39% chronic kidney disease, and 97% were on insulin. More than 1 in 3 (38% (IQR 15-62%)) of the patient-24 hr CGM periods contained at least 2 consecutive BGs <70mg/dl. The corresponding values for <60 mg/dl and <50 mg/dl were nearly 1 in 4 (23% (8-38%)) and nearly 1 in 12 (8% (0%-25%)), respectively. Hypoglycemia by fingerstick (FS) was extremely rare, with a total of just 4 FS <70 mg/dl during all observation periods combined, comprising a median of 29 FS/patient (IQR 27-41). In bivariate analyses, age was positively correlated (Spearman 0.37, p=0.03), weight loss in past 3 months negatively correlated (-0.37, p=0.03), and hypertension negatively associated (median for hypertensives=0, non-hypertensives=38%, p=0.01) with the % of 24-hour periods with hypoglycemia. Other factors, including HbA1c, were not significantly related but power was limited by small sample size.

Hypoglycemia detected by flash CGM is common but otherwise under-recognized among older, long-term residents of nursing homes treated with insulin or sulfonylureas.