Interleukin-5, interleukin-6, interleukin-8 and tumour necrosis factor-alpha levels obtained within 24-h of admission do not predict high-risk infection in children with febrile neutropenia

Purpose: Biomarkers that can predict the severity of febrile neutropenia (FN) are potential tools for clinical practice. Objective: The objective of this study is to evaluate the reliability of plasma interleukin (IL) levels as indicators of high-risk FN. Materials and Methods: Children with haematological malignancies and FN were enrolled prospectively. A blood sample was obtained within 24-h of admission for estimation of IL-5, IL-6, IL-8 and tumour necrosis factor-alpha (TNF-alpha) level by the enzyme-linked immunosorbent assay. Patients were stratified into three groups. Group I (low-risk): No focus of infection; Group II: Clinical/radiological focus of infection; Group III: Microbiologically proven infection or FN related mortality. Groups II and III were analysed as high-risk. The cytokines were assessed at three different cut-off levels. Results: A total of 52 episodes of FN in 48 patients were evaluated. The mean age was 6 years (range: 2-13). Primary diagnosis included acute lymphoblastic leukaemia (82%), non-Hodgkin's lymphoma (13%) and acute myeloid leukaemia (5%). Absolute neutrophil count was < 200 cells/mu l in half and 200-500 in 23%. Majority were categorised as Group I (69%), followed by Group II (16%) and III (15%). The range of IL-5 was too narrow and similar in the two risk-groups to be of any relevance. The best sensitivity of TNF-alpha and IL-6 for high-risk group was 78% and 70%, respectively. The highest specificity observed was 35%. The negative predictive value of IL-6, IL-8 and TNF-alpha exceeded 80%. Conclusion: IL-5, IL-6, IL-8 and TNF-alpha failed as predictors of clinically localised or microbiologically documented infection in children with chemotherapy induced FN. However, IL-6, IL-8 and TNF-alpha could be useful in excluding the possibility of high-risk infection.