Recommendations for the treatment of insulin reactions are based more on habit than data. We investigated the efficacy in correcting blood glucose levels and alleviating clinical symptoms of hypoglycemia of seven orally administered carbohydrates--glucose in solution, tablets, and gel; sucrose in solution and tablets; a hydrolized polysaccharide solution; and orange juice--each of which provided 15 g of carbohydrate. Forty-one type I diabetic patients recently treated with insulin agreed to submit to artificially induced hypoglycemia by an intravenous injection of insulin. Corrective therapy was given when patients experienced symptoms and asked for treatment. Mean blood glucose levels 10 minutes after ingestion were found to be similar whether correction was dispensed with the tablets and the solutions of glucose, those of sucrose, or the polysaccharide preparation. However, almost no increment was obtained at this time point with the gel or the fruit juice. Fifteen and 20 minutes after carbohydrate intake, blood glucose levels were higher with the tablet forms than with the solutions, although differences only became signifiant for sucrose. Glycemic responses were again consistently lower with the sucrose gel and the orange juice. Clinical symptoms were alleviated in 14.0 +/- 0.8 minutes (mean +/- SEM) with sucrose and glucose in solution or tablets. We conclude that in moderately severe hypoglycemia, ingestion of 15 g of carbohydrate in the form of glucose or sucrose tablets or as a solution provides an effective therapy; both sugars seem equivalent. Even if sucrose lumps are better recommended in terms of cost and availability, they may not be recommendable in terms of palatability. Glucose gel or orange juice cannot be recommended, at least in light of our experimental procedure and at the dosage used therein.