Article révisé par les pairs
Résumé : Patient response to acute bacterial infection is highly variable. Differing outcomes in this setting may be related to variations in the immune response to an infectious insult. Using quantitative real-time polymerase chain reaction, we quantified gene expression of the tumor necrosis factor alpha(TNFalpha), interferon gamma (IFNgamma), and interleukin 10 (IL10), IL12p35, and IL4 genes in 3 patient groups. These groups consisted of an intensive care unit (ICU) cohort who presented with severe sepsis or septic shock, a group of noncritically ill ward patients with documented Gram-negative bacteremia, and a group of healthy controls. Greater interleukin 10 messenger RNA (mRNA) levels were detected in the ICU group in comparison with both the bacteremic and control groups (P < 0.0001). More TNF-alpha mRNA was detected in the ICU group when compared with the control group (P < 0.0001). However, TNF-alpha mRNA was most abundant in the bacteremic group (P = 0.0007). Lesser IFN-gamma mRNA levels were detected in the ICU group when compared with both the bacteremic and control groups (P < 0.0003). Cytokine mRNA levels were not associated with the occurrence of shock upon admission to ICU. On the seventh day of ICU stay, the presence of shock was associated with lesser IFN-gamma mRNA (P = 0.0004) and lesser TNF-alpha mRNA (P = 0.001). Survivors had greater TNF-alpha mRNA copy numbers on day 7 of ICU stay than nonsurvivors (P = 0.002). We conclude that a proinflammatory response is the appropriate response in the setting of infection and is associated with lesser requirements for inotropes and lesser mortality. Quantitative real-time polymerase chain reaction can be used to predict infection outcome in clinically relevant situations where enzyme-linked immunosorbent assay testing has proved disappointing.

Documents en relation

DI-fusion