Intensive management of type I diabetes mellitus may lead to a significant increase in hypoglycemia. This study evaluated the use of uncooked cornstarch to avert hypoglycemia in 13 patients, ages 3.0-17.5 years, with diabetes and a history of nighttime hypoglycemia. The usual bedtime snack (Standard Snack Period) was given for 14 days followed by 14 days in which 25%-50% of the carbohydrate content was given as uncooked cornstarch in milk (Test Snack Period). Blood glucose levels were obtained at 0200 and before breakfast during the 28-day study. No difference was found between the mean (+/- SD) before breakfast blood glucose levels during the two periods; however, the mean (+/- SD) number of hypoglycemic episodes (blood glucose < 60 mg/dL or 3.3 mmol/L) was reduced significantly during the Test Snack Period at both 0200 (p < 0.025) and before breakfast (p < 0.01) compared to Standard Snack Period (Standard Snack 2.00 +/- 2.12 versus Test Snack 0.61 +/- 0.87 at 0200, Standard Snack 2.61 +/- 2.25 versus Test Snack 0.69 +/- 1.03 before breakfast). In addition, in four children who were evaluated with every-2-h glucose levels through the night, stable glycemia was achieved after cornstarch ingestion. Our results suggest that uncooked cornstarch may be useful to decrease the frequency of nocturnal hypoglycemia in type I diabetes patients. This low-cost, simple intervention might be considered as an adjunctive therapy to diminish the risk of intensive diabetes management.