Impaired ex vivo lipopolysaccharide-stimulated whole blood tumor necrosis factor production may identify "septic" intensive care unit patients.

Currently, there is no reliable diagnostic test to identify septic intensive care unit (ICU) patients. We initiated studies to test the hypothesis that in sepsis, the in vivo exposure to endotoxin is detectable by the ex vivo analysis of lipopolysaccharide (LPS)-stimulated tumor necrosis factor (TNF) production. We obtained heparinized whole blood (WB) from 58 ICU patients and 14 healthy controls. The samples were incubated +/-10 ng/mL of LPS at 37 degrees C for 3 h. Plasma TNF levels were measured using enzyme-linked immunoassay (mean +/- standard error of the mean). Clinical data, including ICU length of stay (LOS), ventilator days (VentD), WBC, and positive cultures (Clt+), were obtained retrospectively. A wide range of LPS-stimulated WB TNF production (pg/mL) was observed in ICU patients (4481+/-469) and controls (6706+/-715). Patients were stratified into quartiles (I-IV) on the basis of the distribution of plotted LPS-stimulated TNF values (pg/mL). Patients in quartile I (N = 14) had significantly lower TNF production (< 2000 pg/mL, P < 0.05) and required increased VentD (16 vs. 10 days, P < 0.05) compared to quartiles II-IV (N = 44). Patients in quartile I also had a higher incidence of infection (79 vs. 50%) and longer LOS (18 vs. 13 d) compared to quartiles II-IV. Impaired TNF release may be a manifestation of monocyte endotoxin tolerance and may be useful to diagnose sepsis.