To assess the effects of crystalloid and colloid resuscitation on hemodynamic response and on lung water following thermal injury, 79 patients were assigned randomly to receive lactated Ringer's solution or 2.5% albumin-lactated Ringer's solution. Crystalloid-treated patients required more fluid for successful resuscitation than did those receiving colloid solutions (3.81 vs. 2.98 ml/kg body weight/% body surface burn, p < 0.01). In study phase 1 (29 patients), cardiac index and myocardial contractility (ejection fraction 'and mean rate of internal fiber shortening, Vcf) were determined by echocardiography during the first 48 hours postburn. Cardiac index was lower in the 12-to 24-hour postburn interval in the crystalloid group, but this difference between treatment groups had disappeared by 48 hours postburn. Ejection fractions were normal throughout the entire study, while Vcf was supranormal (p < 0.01 vs. normals) and equal in the two resuscitation groups. In study phase 2 (SO patients), extravascular lung water and cardiac index were measured by a standard rebreathing technique at least daily for the first postburn week. Lung water remained unchanged in the crystalloid-treated patients (p > 0.10), but progressively increased in the colloid-treated patients over the seven day study (p < 0.0001). The measured lung water in each treatment group was significantly different from one another (p < 0.001). Cardiac index increased progressively and identically in both treatment groups over the study period (p < 0.01). These data refute the existence of myocardial depression during postburn resuscitation and document hypercontractile left ventricular performance. The addition of colloid to crystalloid resuscitation fluids produces no long lasting benefit on total body blood flow, and promotes accumulation of lung water when edema fluid is being reabsorbed from the burn wound.