Defining a Continuous Glucose Baseline to assess the impact of nutritional interventions

Accurate and robust estimation of individuals’ basal glucose level is a crucial measure in nutrition research but is typically estimated from one or more morning fasting samples. The use of Continuous Glucose Monitoring (CGM) devices presents an opportunity to define more robust basal glucose levels, which estimates can be generalized to any time of the day. However, to date, no standardized method has been delineated. The current paper seeks to define a reliable algorithm to characterize the individual’s basal glucose level over 24 h from CGM measurements. Data drawn from four nutritional intervention studies performed on adults free from chronic diseases were used to define that basal glucose levels were optimally estimated using the 40th percentile of the previous 24 h CGM data. This simple algorithm provides a Continuous Glucose Baseline over 24 h (24 h-CGB) that is an unbiased and highly correlated estimator (r = 0.86, p-value