A randomized crossover study of 5-g guar minitablets against placebo, given three times per day with main meals for 8 wk, was done in 29 non-insulin-dependent diabetes mellitus (NIDDM) patients who had near-normal fasting plasma glucose concentrations on treatment with diet alone, additional sulfonylurea, or ultralente insulin. Guar did not reduce the excessive postprandial glycemic excursion, glycosylated hemoglobin values, basal plasma glucose concentrations, basal or incremental plasma C-peptide values, or body weight. There were few side effects with either guar or placebo therapy. Mean low-density lipoprotein cholesterol levels were significantly reduced (P less than .001) by guar administration (116 +/- 23 vs. 104 +/- 19 mg/dl). Guar additives did not improve the excessive postprandial glycemia found in NIDDM patients in whom near-normal fasting plasma glucose levels had been obtained.