Serum adhesion molecules as predictors of bacteremia in adult severe sepsis patients at the emergency department.

BACKGROUND: Bacteremia is a severe bacterial infection with significant mortality. Clinical parameters that reliably predict it are less elucidated. We assessed the potential of serum adhesion molecules for predicting bacteremia and compare it with current available infection biomarkers to determine a more timely predictor of adult severe sepsis patients on admission to the emergency department (ED).
METHODS: Sixty-seven consecutive non-traumatic, non-surgical adult patients with severe sepsis admitted to the ED were evaluated. Serum samples were collected and assessed while serum adhesion molecules were analyzed.
RESULTS: Thirty-one (46.2%) study patients had bacteremia. There were significant differences in both sICAM-1 and sE-selectin on admission between bacteremic and non-bacteremic patients. By stepwise logistic regression model, only sE-selectin was independently associated with bacteremia and any 1 ng/ml increase in level increased bacteremia rate by 0.8%. The cut-off value of sE-selectin level for predicting bacteremia was 117 ng/ml (84% sensitivity and 69% specificity).
CONCLUSION: Although serum cell adhesion markers are not specific for predicting bacteremia in septic patients, higher mean serum cell adhesion molecules levels on admission may imply both more severe infection and presence of bacteremia. Assay of serum adhesion molecules may be added as an infectious marker among the panel of bacteremic parameters in clinical practice, especially since early diagnosis and prompt antimicrobial therapy are essentially for survival.