Diabetics well controlled on chlorpropamide received, in randomized manner either 1200 mg of ibuprofen or 300 mg of phenylbutazone per day for rheumatic pains, for a period of 4 weeks. Fasting and postlunch, whole blood, true sugars (FBS and PLBS) were estimated at weekly intervals. Subjects taking phenylbutazone showed reduction in FBS values throughout the treatment; the reduction became statistically significant at the 3rd and 4th week. Clinical hypoglycemia, however, was not observed. The FBS values returned to pretreatment levels after stopping phenylbutazone. No significant reduction was seen in FBS in subjects taking ibuprofen. There was no significant change in PLBS values in either group.