The use of continuous glucose monitoring in clinical trials in diabetes yields comprehensive data on glycaemia far beyond glycated haemoglobin and SMBG. From CGM traces e.g. glycaemic variability and glycaemic excursions can be evaluated. From the data, single hypoglycaemic episodes can be characterized by their timing, depth and duration. For such complex data, consensus about classification and reporting is needed. Recently, an international consensus report was elaborated at the initiative of Advanced Technologies & Treatments for Diabetes (ATTD) proposing glucose level and duration of events to be reported in trials. The added value of using CGM to record hypoglycaemia in clinical trials include comprehensive recording of events that may provide an estimate of the real rates experienced by patients. Furthermore, CGM can record asymptomatic hypoglycaemia, which at least in type 1 diabetes constitute most events. The main limitation of using CGM to detect hypoglycaemia is the inaccuracy in the lower glucose range, which is, however, improved in newer systems. Whereas blinded CGM data are rather unbiased, open real-time CGM may promote avoiding behavior, resulting in reduced hypoglycaemic rates. CGM may contribute significantly to the assessment of glycaemic interventions in clinical trials.