Issues on hypoglycemia prediction and detection

Current guidelines in the treatment of diabetes stress the importance of tight glycemic control in order to avoid long term complications. At the same time, the intensive approach to maintaining euglycemia increases the incidence of hypoglycemia. Continuous glucose monitoring systems (CGMS) have led to the possibility of predicting and detecting hypoglycemic episodes, and alerting the user to take action. We use an optimal estimation theory approach to hypoglycemia detection and prediction. We demonstrate the effect of measurement sampling frequency, threshold level, and prediction horizon on the sensitivity and specificity of the prediction of hypoglycemia. For characteristic glucose changes we first show the inherent limitation in sensor performance at predicting future glucose values. We then discuss how optimal estimators can be tuned to trade-off the false alarm rate with the rate of missed predicted hypoglycemic episodes. We also suggest the use of different alarm levels as a function of current and future estimates of glucose and the hypoglycemic threshold and prediction horizon.