Twenty patient with Type 1 diabetes, using insulin glargine as basal insulin, participated in a prospective, controlled crossover trial comparing blood glucose profiles over two 24 h periods with and without a late night snack (a slice of rye bread, 20 g carbohydrates, at 10 p.m.), in randomized order. The main endpoint was the number of hypoglycemic episodes with a confirmed laboratory blood glucose ≤ 50 mg/dl between 10 p.m. and 8 p.m. the following day. Secondary endpoint was the blood glucose profile during this period.

The order of studies (with and without a late night snack) was determined by randomization. Study-related procedures started at 6 p.m. (standard blood glucose profiles with determinations at 6, 8, and 10 p.m, at midnight, at 2, 4, 6:45, 8:45 a.m., and at noon and 2, 6, and 8 p.m. of the following day). At 10 p.m. a late night snack (one slice of whole meal rye bread containing approximately 20 g of carbohydrate) was eaten or no nutrients were consumed, as determined by the randomization protocol. After one day without study-related activities, the study was repeated with the complementary protocol (crossover design). Patients were instructed to use fast-acting insulin to correct hyperglycemia three times per day (before major meals) or to take in additional carbohydrate when symptomatic or chemical hypoglycemia was noted. They were asked not to perform strenuous exercise beyond their normal daily activities.

Inclusion Criteria:

• Type 1 diabetes
• Basal insulin substitution with insulin glargine at least 3 month

Exclusion Criteria:

• Severe disorders
• Severe and clinical significant gastroparesis
• History of pancreatitis or pancreatic cancer
• Severe Liver disorder (liver enzymes > three times upper normal limit)
• Pregnancy
• Drug- or alcohol abuse