Gabrielle M Haeusler1, Fabianne Carlesse, Robert S Phillips. 1. From the *Children's Cancer Centre, The Royal Children's Hospital Melbourne, Parkville; †Department of Infectious Diseases and Infection Control, Peter MacCallum Cancer Centre, East Melbourne, Australia; ‡Department of Pediatrics, Pediatric Oncology Institute (IOP-GRAACC), Federal University of Sao Paulo, São Paulo, Brazil; §Centre for Reviews and Dissemination, Alcuin College, University of York, York; and ¶Regional Department of Paediatric Haematology/Oncology, Leeds Teaching Hospitals Trust, Leeds, United Kingdom.
Abstract
BACKGROUND: Fever during neutropenia (FN) is a frequent and potentially life-threatening complication of the treatment of childhood cancer. The role of biomarkers in predicting morbidity and mortality associated with FN in children has been explored with varying results. This systematic review identified, critically appraised and synthesized information on the use of biomarkers for the prediction of outcome of FN in children/young adults, updating a review of initial assessment and adding further analysis of their value at reassessment. METHODS: This review was conducted in accordance with the Centre for Reviews and Dissemination Methods, using 3 different random effects meta-analysis models. RESULTS: Thirty-seven studies involving over 4689 episodes of FN in children were assessed, including an additional 13 studies investigating 18 biomarkers in 1670 FN episodes since the original review. Meta-analysis was possible for admission C-reactive protein (CRP), procalcitonin (PCT), interleukin-6 and interleukin-8 in their ability to detect significant infection. Marked heterogeneity exists, precluding clear clinical interpretation of the results. Qualitative synthesis of the role of serial biomarkers suggests their predictive ability may be more pronounced at 24 to 48 hours compared with admission. Direct comparisons of the discriminatory power of admission values of PCT and CRP showed PCT generally had a better discriminatory estimate of serious infection than CRP. CONCLUSIONS: There remains a paucity of robust and reproducible data on the use of biomarkers in prediction of serious infection in children with FN. Available evidence suggests PCT has better discriminatory ability than CRP and that the role of serial biomarkers warrants further study.
BACKGROUND:Fever during neutropenia (FN) is a frequent and potentially life-threatening complication of the treatment of childhood cancer. The role of biomarkers in predicting morbidity and mortality associated with FN in children has been explored with varying results. This systematic review identified, critically appraised and synthesized information on the use of biomarkers for the prediction of outcome of FN in children/young adults, updating a review of initial assessment and adding further analysis of their value at reassessment. METHODS: This review was conducted in accordance with the Centre for Reviews and Dissemination Methods, using 3 different random effects meta-analysis models. RESULTS: Thirty-seven studies involving over 4689 episodes of FN in children were assessed, including an additional 13 studies investigating 18 biomarkers in 1670 FN episodes since the original review. Meta-analysis was possible for admission C-reactive protein (CRP), procalcitonin (PCT), interleukin-6 and interleukin-8 in their ability to detect significant infection. Marked heterogeneity exists, precluding clear clinical interpretation of the results. Qualitative synthesis of the role of serial biomarkers suggests their predictive ability may be more pronounced at 24 to 48 hours compared with admission. Direct comparisons of the discriminatory power of admission values of PCT and CRP showed PCT generally had a better discriminatory estimate of serious infection than CRP. CONCLUSIONS: There remains a paucity of robust and reproducible data on the use of biomarkers in prediction of serious infection in children with FN. Available evidence suggests PCT has better discriminatory ability than CRP and that the role of serial biomarkers warrants further study.
Authors: Robert S Phillips; Lillian Sung; Roland A Ammann; Richard D Riley; Elio Castagnola; Gabrielle M Haeusler; Robert Klaassen; Wim J E Tissing; Thomas Lehrnbecher; Julia Chisholm; Hana Hakim; Neil Ranasinghe; Marianne Paesmans; Ian M Hann; Lesley A Stewart Journal: Br J Cancer Date: 2016-03-08 Impact factor: 7.640