Conventional assessment of glycemic control in diabetes mellitus (DM) includes blood glucose attention to glycemia and glycated hemoglobin levels. Recently, we introduced the continuous glucose-monitoring test (CGM) (Medtronic Minimed-CGMS System Gold). Here we describe our experience with this methodology over the year 2004. A total of 141 CGM tests were performed over this period of time. Overall, 88% (n= 124) patients were diabetics (DM), 99 of them were insulin users. We found a strong correlation between glucose values obtained by CGM and capillary glucose measures (r= 0.926; p< 0.005). In diabetic patients, nocturnal hypoglycemia (< 50 mg/dL) was identified in approximately 35% (n= 44), hyperglycemic patterns (> 220 mg/dL) at specific times of day in approximately 44% and sustained hyperglycemia throughout the whole monitoring period in thirteen cases (10%). Twelve tests were performed to investigate the occurrence of hypoglycemia in non-diabetic subjects. Two tests came out very suggestive of "dumping", and in one case the CGMS supported the hypothesis of insulinoma. Partial monitoring interruptions have occurred in 15% of all tests. We concluded that CGMS is a useful methodology to investigate glycemic fluctuations, and it is also an important tool to adjust therapy in diabetic patients.