Xylitol has been suggested as a potentially useful sweetener in the diabetic diet. In 14 insulin-dependent diabetics a standard diabetic diet regimen was compared with diets in which starch was isocalorically exchanged in the breakfast meal by either 30 g xylitol or 30 g sucrose. Insulin requirement and blood glucose were measured using a glucose-controlled insulin infusion system. The results following breakfast with xylitol were similar to those after starch breakfasts. Sucrose, in contrast, induced a greater post-prandial rise in blood glucose levels despite counter-regulation by the glucose-controlled insulin infusion system. Insulin requirement after sucrose significantly exceeded (p less than 0.01) that after xylitol or starch during the first 60 min and 2 h respectively. No short-term side effects of xylitol were found.