Background: On the basis of human perfusion studies it has been speculated that bicarbonate ions in oral rehydration salt solutions (ORS) to treat diarrhoea are more efficiently absorbed from the small bowel. We evaluated the role of bicarbonate in ORS by using a reduced purging rate in cholera as a proxy indicator for absorption efficiency in cholera-like severe diarrhoea. Methods: In a double-blind randomized trial 60 patients received standard ORS containing bicarbonate or an identical solution except that sodium bicarbonate was replaced by an equimolar amount of sodium chloride (sodium, 90mmol/l; potassium, 20mmol/l; chloride, 80mmol/l; bicarbonate, 30mmol/l; glucose, 111 mmol/l; and osmolality, 331 mmol/l) after initial intravenous rehydration to correct initial dehydration and shock and until diarrhoea ceased. Results: Five patients receiving standard ORS and eight receiving bicarbonate-free ORS required unscheduled intravenous therapy for recurrence of severe dehydration in spite of receiving ORS solution. ORS intake and purging rate, in ml/kg body weight/day, both including and excluding stool output during unscheduled intravenous therapy are closely similar in the two treatment groups. Conclusion: The results indicate that bicarbonate-containing ORS solution does not have any clinically significant effect on the absorption efficiency of ORS, either beneficial or adverse, and its use is relevant only for correction of metabolic acidosis of diarrhoeal dehydration.