The use of phosphate therapy in the management of diabetic ketoacidosis (DKA) has been controversial, particularly with respect to the effect of phosphate intermediates on tissue oxygenation. In a prospective randomized study we evaluated the effect of phosphate (8.5 mmol/h or approximately 6 g phosphate/24 h) (experimental group) vs. no phosphate therapy (control group) in 30 DKA patients, 15 in each group. Various determinations including erythrocyte 2,3-diphosphoglycerate (2,3-DPG), oxyhemoglobin dissociation (p50), serum phosphate, calcium, lactate, pyruvate, electrolytes, and response time to reach predetermined values for glucose, bicarbonate, and pH were measured at frequent intervals during the first 24 h of therapy and daily for 5 days after metabolic control. Initial electrolytes, glucose, pH, erythrocyte 2,3-DPG, lactate, and p50 were not different in either group. Whereas the experimental group had a greater level of 2,3-DPG than the control group by 48 h, the difference was not statistically significant. Recovery indices, including hours to reach glucose of 250 mg/dl, bicarbonate greater than 15 meq/liter, pH greater than 7.3, and mental alertness, were not different in the two groups nor were the p50 or lactate measurements. The experimental group exhibited significantly lower plasma ionized calcium values during therapy. We conclude that phosphate therapy may accelerate regeneration of erythrocyte 2,3-DPG but in the relatively small number of patients studied it had no demonstrable influence on tissue oxygenation or clinical response to low dose insulin therapy of DKA. Furthermore, the exaggeration of hypocalcemia seen in phosphate-treated patients may be reason for caution in the use of such therapy.