It has been proposed that high-carbohydrate, high-fiber (HCF) diets might serve as useful therapeutic modality in non-insulin-dependent diabetes mellitus (NIDDM). One problem in evaluating clinical trials of this therapy is that, by their very nature, the trials cannot be double blinded. We have developed HCF and placebo granola-type bars using complex absorbable carbohydrate and guar gum fiber to circumvent this methodologic problem. The HCF bars, when consumed with an ad lib. diet, assure an HCF intake without imposing other dietary restrictions. To test the short-term efficacy of the bars, 9 normal adult volunteers, 2 women with impaired glucose tolerance, and 20 patients with NIDDM consumed the bars alone or with meals. Blood glucose responses when HCF bars were consumed alone were blunted when compared with the placebo response (P < 0.0005 to P < 0.002), with the most marked suppression occurring in the early postprandial period. In contrast, when the bars were consumed along with breakfast, HCF and placebo responses were virtually identical in the early postprandial period, but showed a progressively greater difference from 90 to 240 min (P < 0.02 to P < 0.0005). When consumed with lunch as well as breakfast, the HCF bars caused flattening of blood glucose responses during the late postprandial period after breakfast and maintained flattened responses during the early and late postprandial periods after lunch (P < 0.05 to P < 0.005). It is concluded that these HCF bars can be used to blunt postprandial blood glucose responses, in subjects with either normal or abnormal carbohydrate metabolism.