BACKGROUND: Efforts to mimic euglycemia depend upon targets from epidemiologic studies that rely on episodic measurements reduced to statistical summaries, leaving open the question, "What is normal glycemia?" We postulated that portrayal of euglycemia was possible through application of continuous glucose monitoring (CGM) and a novel analytical tool, the ambulatory glucose profile (AGP). METHODS: Individuals with normal glucose tolerance (NGT) and with diabetes used CGM for 30 days. AGP analysis, which graphs CGM data by time without regard to date, was used to characterize glucose exposure, variability, and stability. RESULTS: Sixty-two subjects completed the study, employing CGM for 28 +/- 4 days averaging 99 +/- 18 (range, 33-125) readings per day. NGT subjects (n = 32) had a mean CGM of 102 +/- 7 mg/dL, ranging between 94 and 117 mg/dL and averaging 105 +/- 8 mg/dL daytime and 97 +/- 6 mg/dL overnight. Glucose variability, as expressed by the interquartile range, was 21 +/- 4 mg/dL (range, 14-29 mg/dL). Stability in glycemic control (hourly change in the median) for NGT subjects averaged 3 +/- 1 mg/dL/h. Subjects with diabetes (n = 30) were significantly higher on all glycemic characteristics with the exception of the percentage of hypoglycemic (CGM <70 mg/dL) episodes for type 2 diabetes (2.9%), compared to 2.7% for subjects with NGT. CONCLUSIONS: CGM technologies enabled collection of verified data under normal living conditions, providing an exceptional vantage point from which to obtain important clinical information. This will facilitate an understanding of the range of euglycemic patterns, provide a sensitive means of detecting impaired glucose tolerance, and help set realistic treatment goals for individuals with diabetes.