Severely burned patients [74] were resuscitated with i.v. crystalloid solutions that varied in their average Na+ concentration. No colloid solution was used. During the 1st 24 h, when requirements were greatest, 21 patients received fluids containing less than 150 meq Na+/O, nearly all of which was lactated Ringer''s solution, while for 31 patients the fluids contained more than 199 meq Na+/l; the remaining 22 patients were treated with fluids of intermediate Na+ content (150-199 meq/l). Detailed, computer-assisted analysis of the data was performed. There were no significant intergroup differences in mean hourly urine output, which was in the targeted range of 0.5-1 ml/kg, or in urinary Na+ excretion. Both the Na+ and water loads administered were significantly larger in patients who died than in survivors (P < 0.0025). This observation is new. The use of concentrated Na+ solutions did not increase the Na+ requirement, but water loads were significantly lower in patients who received the most concentrated Na+ solutions compared to those resuscitated with the least concentrated Na+ solutions (P < 0.014). Minimizing the water loads during resuscitation by increasing the Na+ content of the administered fluid might improve the chance for survival of severely burned patients.