Febrile neutropenia is an important cause of morbidity and mortality in children receiving treatment for cancer. Although cases of severe infection need inpatient treatment with broad-spectrum antibiotics there are a number of children with non-severe infection who are over treated. Previous meta-analysis showed further research is needed into the use of biomarkers, such as procalcitonin, as part of risk prediction in febrile neutropenia. No previous studies have looked at the use of biomarkers as an addition to existing clinical decision rules. Procalcitonin is currently being reviewed by NICE as part of it’s diagnostics assessment programme.

Aims To determine if procalcitonin can be used to diagnose or exclude severe infection on presentation with febrile neutropenia and if it has additional benefit when used with existing clinical decision rules.

Method This is a prospective cohort study of a diagnostic test. Patients between birth and 18 years old were who were admitted to the paediatric oncology and haematology wards with febrile neutropenia were included. Blood was taken for a procalcitonin level at admission as well as routine investigations.

Results Forty-one episodes were included from 26 patients. Procalcitonin level of >2 ng/dL has a likelihood ratio of 14.6 [95% CI 1.8, 120.4] and a sensitivity of 43% and specificity of 97%. For none of the clinical decision rules did the procalcitonin odds ratio reach significance although all of the odds ratios were over one.

Conclusion This study does not show a benefit in using procalcitonin in febrile neutropenia. The cut off of 2 ng/dL has a high likelihood ratio for severe infection but poor sensitivity. There is no significant additional benefit of procalcitonin when used with existing clinical and laboratory features in clinical decision rules.