Hypoglycemia Prevention Behaviors Prevent Moderate Hypoglycemic Events: The Replace-Bg Trial

Background: Continuous glucose monitoring (CGM) provides real-time glucose levels and alarms for hypoglycemia. We analyze CGM data from the REPLACE-BG trial, which had 6 months of CGM data on people with well controlled type 1 diabetes (T1D), to characterize hypoglycemic events and evaluate their association with hypoglycemia-related behaviors. Methods: We conducted a psychometric analysis of the Hypoglycemia Fear Survey- Behavior scale to identify unique hypoglycemia behavior constructs. We identified and categorized hypoglycemic events in CGM data as mild (min glucose <70 mg/dL) and moderate (min glucose <50 mg/dL) and used ANOVA to evaluate differences in the frequency and duration of events by high and low score categories of hypoglycemia behavior constructs. We used multivariate linear regression to evaluate the association between hypoglycemia behavior constructs and 1) duration of moderate hypoglycemic events and 2) proportion of moderate hypoglycemic events. Results: We separated the HFS-B into three scales: hypoglycemia avoidance, reaction, and prevention. In low vs. high hypoglycemia prevention score categories, the duration of moderate hypoglycemic events was significantly longer (73 vs. 66 minutes, p<0.05) and proportion of moderate hypoglycemic events was higher (22.4 vs. 20.1, p<0.1). In multivariate models that included all three hypoglycemia behavior scales, hypoglycemia prevention behavior was significantly, inversely associated with 1) duration of moderate hypoglycemic events (β= -6.9, SE= 2.2, p< 0.01) and 2) proportion of moderate hypoglycemic events (β= -2.2, SE= 0.7, p-value < 0.01). Conclusion: Hypoglycemia is common in T1Ds. The increased likelihood of engaging in hypoglycemia prevention behaviors reduces the frequency and duration of dangerous moderate hypoglycemic events. Analyzing CGM data in its continuous form reveals the frequency and characteristics of individual hypoglycemic events. Disclosure M.A. Crawford: Employee; Self; Dexcom, Inc. M.M. White: None. D.R. Strong: None. J.P. Pierce: None.