To examine the role of systemic plasma tumor necrosis factor (TNF) in the septic response following trauma, an endotoxin (lipopolysaccharide (LPS)) challenge was administered to anesthetized mongrel pigs 72 h following either hemorrhagic shock/resuscitation or sham shock. For TNF to be considered a mediator, at least two conditions should be satisfied: a TNF increase should precede other manifestations of the septic response and the magnitude of that increase should correlate with the symptoms. Immediately following resuscitation from shock, hemodynamics were stable, but heart rate, cardiac index (CI), and systemic oxygen delivery (DO2) were elevated 20-60%, and systemic vascular resistance (SVR) was decreased 40%, relative to the preshock baseline. After 72 h, the animals were reanesthetized, reinstrumented, and all hemodynamic values were near normal in both groups. At this point, either 1.5 (shock, n = 2; sham, n = 2), 15 (shock, n = 7; sham, n = 6) or 150 (shock, n = 11; sham, n = 4) micrograms/kg of Escherichia coli LPS was administered intravenously over 30 min. Serial hemodynamic data, complete blood counts, and TNF were recorded for 3 h post-LPS. LPS evoked profound leukopenia and pulmonary hypertension within 15 min that was followed by a hyperdynamic septic response (i.e., progressive arterial desaturation, tachypnea, tachycardia, increased CI, and decreased SVR) and rise in plasma TNF at 60-90 min. In the shock group, LPS-evoked TNF changes were less than or equal to those in the sham group, even though mortality was higher after shock.(ABSTRACT TRUNCATED AT 250 WORDS)