We compared the efficacy and safety of a single polyelectrolyte solution, Dhaka solution (DS), containing 133 mmol/L sodium, 13 mmol/L potassium, 98 mmol/L chloride, and 48 mmol/L acetate with and without 139 mmol/L (25 gm/L) dextrose in the rapid (4 hours) rehydration of 67 patients with diarrhea and moderate or severe dehydration requiring parenteral fluid therapy. Of the 67 patient, 31 were randomly assigned to receive the dextrose-containing solution (DS + D) and 36 DS without dextrose. On admission to the hospital, the two groups of patients were similar with respect to enteric pathogens detected, proportion with hyponatremia, magnitude of dehydration as assessed by clinical criteria, serum protein or creatinine concentration, and plasma glucose levels. At the end of the 4-hour infusion, both groups of patients had similar decreases in serum creatinine and protein levels and similar volume of urine output, but patients receiving DS + D had a significantly higher plasma glucose level than patients receiving DS (7.8 mmol/L (140 mg/dl) vs 5.39 mmol/L (97 mg/dl), P less than 0.01). One patient in the DS group had hypoglycemia (plasma glucose 2.0 mmol/L (36 mg/dl) at 4 hours. No other complications were noted. Serum protein values 24 hours after admission were little changed from 4-hour values, suggesting that rehydration was complete at the end of 4 hours. We conclude that, in our patients, rehydration can be carried out safely and rapidly with the use of a single solution and that adding 139 mmol/L (25 gm/L) of dextrose to the solution can prevent hypoglycemia without producing an osmotic diuresis.