Diagnostic accuracy of G-CSF, IL-8, and IL-1ra in critically ill children with suspected infection

Joachim E Fischer,Anne Benn,Stephan Harbarth,David Nadal,Sergio Fanconi
DOI: https://doi.org/10.1007/s00134-002-1423-2
Abstract:Objective: To elucidate the diagnostic accuracy of granulocyte colony-stimulating factor (G-CSF), interleukin-8 (IL-8), and interleukin-1 receptor antagonist (IL-1ra) in identifying patients with sepsis among critically ill pediatric patients with suspected infection. Design and setting: Nested case-control study in a multidisciplinary neonatal and pediatric intensive care unit (PICU) PATIENTS: PICU patients during a 12-month period with suspected infection, and plasma available from the time of clinical suspicion (254 episodes, 190 patients). Measurements and results: Plasma levels of G-CSF, IL-8, and IL-1ra. Episodes classified on the basis of clinical and bacteriological findings into: culture-confirmed sepsis, probable sepsis, localized infection, viral infection, and no infection. Plasma levels were significantly higher in episodes of culture-confirmed sepsis than in episodes with ruled-out infection. The area under the receiver operating characteristic curve was higher for IL-8 and G-CSF than for IL-1ra. Combining IL-8 and G-CSF improved the diagnostic performance, particularly as to the detection of Gram-negative sepsis. Sensitivity was low (<50%) in detecting Staphylococcus epidermidis bacteremia or localized infections. Conclusions: In this heterogeneous population of critically ill children with suspected infection, a model combining plasma levels of IL-8 and G-CSF identified patients with sepsis. Negative results do not rule out S. epidermidis bacteremia or locally confined infectious processes. The model requires validation in an independent data-set.
What problem does this paper attempt to address?