ALL Metrics
-
Views
-
Downloads
Get PDF
Get XML
Cite
Export
Track
Brief Report
Revised

Prospective cohort study of the predictive value of inflammatory biomarkers over clinical variables in children and young people with cancer presenting with fever and neutropenia

[version 2; peer review: 2 approved]
PUBLISHED 03 Feb 2022
Author details Author details
OPEN PEER REVIEW
REVIEWER STATUS

This article is included in the Oncology gateway.

This article is included in the Cancer Early Detection and Screening collection.

Abstract

Introduction
Fever during chemotherapy induced neutropenia is a common and potentially life-threatening complication of the treatment of childhood cancer. Predictions of poor outcome could be enhanced by incorporating serum biomarkers of inflammation at presentation and reassessment.
Methods
A prospective cohort study was conducted of children under 18 years old, being treated for cancer or a cancer-like condition, who presented with fever (≥ 38.0°C) and neutropenia (neutrophil count < 0.5*109/L). Clinical features were recorded, along with three experimental inflammatory biomarkers: procalcitonin (PCT), interleukin-6 (IL-6) and interleukin-8 (IL-8). Outcomes included serious medical complications (SMC): any infection related mortality, critical care and organ support, severe sepsis, septic shock, significant microbiologically defined infection, or radiologically confirmed pneumonia.
Results
Biomarker assessments were undertaken in 43 episodes of fever and neutropenia, from 31 patients aged between four months and 17 years old (median six years): 20 were female and 22 had acute leukaemia. Five episodes of SMC were noted. PCT, IL-6 and IL-8 had poor individual discriminatory ability (C-statistic 0.48 to 0.60) and did not add to the value of clinical risk stratification tools. Insufficient data were collected to formally assess the value of repeated assessments.
Conclusions
Incorporating serum biomarkers of inflammation at presentation of episodes of fever with neutropenia in childhood does not clearly improve risk stratification. The value of serial assessments requires further investigation.

Keywords

febrile neutropenia, neutropenic sepsis, childhood cancer, inflammatory biomarkers, IL6, IL8, procalcitonin

Revised Amendments from Version 1

This version is in response to two peer reviewers' comments. The abstract final line has been modified to say "The value of serial assessments requires further investigation" which softens the tone. An extra paragraph is added to the discussion, exploring the reasons for the small numbers of patient episodes captured, and the range of the biomarkers as well as mean values have been added to Table 3.

See the author's detailed response to the review by Rejin Kebudi
See the author's detailed response to the review by Paul J. Gibson

Introduction

Infection, frequently presenting as fever during chemotherapy induced neutropenia, is a common and potentially life-threatening complication of the treatment of childhood cancer. Modern management approaches have promoted the use of clinical decision rules to stratify patients at low risk of serious medical complications during febrile neutropenia (FN), enabling them to be safely managed with less intensive therapy as an outpatient.1

In addition to the role of clinical risk stratification, there is an increased desire to use modern biochemical markers of inflammation in predicting the risk of severe sepsis. There is increasing interest in inflammatory biomarkers such as procalcitonin (PCT), an inflammatory marker that has been shown to rise in response to bacterial infection in non-immunosuppressed children,2 and various cytokines including interleukin (IL)-6 and IL-8 to improve the diagnostic accuracy of a prediction rule in children with FN.3

A systematic review and meta-analysis of the discriminatory ability of biomarkers in children with FN4,5 concluded that while several small studies suggest PCT, IL-6 and IL-8 may be valuable for predicting severe infection in children with FN, the true impact remains unknown. The lack of value of older biomarkers, such as C-reactive protein (CRP), is confirmed in these reviews. A smaller number of studies explored the role of serial (i.e., 0 h, 12–24 h, 48 h) biomarkers to detect documented infection or sepsis. In one study, the difference between CRP, PCT and IL-8 at 24 hours in children with and without sepsis was more pronounced than at presentation.6 These data suggest the optimal value for prediction may be made by incorporating biomarkers at early reassessment (i.e., within 12–24 h or 24–48 h), rather than at presentation.

This study aimed to undertake a focused analysis of three promising inflammatory biomarkers: PCT, IL-6 and IL-8 in both initial and value of serial testing and their additional discriminatory value above routine clinical features using the PICNICC model7 and AUS-score.8

Methods

This study took place in Leeds Children’s Hospital between March 2016 and March 2018. It recruited patients on presentation of FN or pre-enrolled them during routine appointments where they, or their parents or guardians, affirmed their consent in written form. In addition to formal information sheets, a link to an animated video summary of the research was provided https://www.youtube.com/watch?v=Z1AXzJqatds.9 Ethical approval was given by the Leeds West NHS Research Ethics Committee [15/YH/0357].

Children could be included who were: younger than 18 years old, who had cancer, or received a stem cell transplant, or who had Langerhans cell histiocytosis in need of cytotoxic chemotherapy, or haemophagocytic lymphohistiocytosis undergoing active treatment, or severe aplastic anaemia, and attended with febrile neutropenia. Fever was defined as temperature ≥ 38.0°C and neutropenia, an absolute neutrophil count ≤ 0.5 g/L.

All patients were admitted to hospital and managed according to institutional FN pathways; these consisted of full evaluation and empiric piperacillin/tazobactam (or suitable alternatives) until afebrile 48 hours and with no other reason for antimicrobials, regardless of neutrophil count. Inflammatory biomarkers, with the exception of CRP, were analysed in batches and their results masked from the treating team. Using this unselective approach, with treatment unaffected by the results of the biomarker analyses, we minimized the biases which can arise through cherry-picking of patients and altering medical treatment on the basis of the test under evaluation (selection and incorporation biases).

Demographic data, variables and outcomes included core items as devised by the International Paediatric Fever in Neutropenia Working Group (see Extended Data, Table 1)9. Clinical assessments and blood samples for biomarkers were taken at presentation (within 12 h of fever or admission) and daily until the end of the FN episode or discharge as part of usual clinical care.

Table 1. Distribution of diagnosis in 31 patients with biomarker data.

DiagnosisNumber of patients
Acute leukaemias21
Other haematological malignancies2
Solid tumours5
Brain tumours3

The primary outcome measure was ‘serious medical complication (SMC)’10 (defined by any of (i) infection related mortality (ii) admission to ICU/HDU/other ward/unit for organ support (iii) severe sepsis or (iv) septic shock) or significant microbiologically defined infection, or radiologically confirmed pneumonia. Secondary outcomes measures were the component parts of the primary outcome, death within 30 days, infection related mortality, clinically defined infection, infection with antibiotic-resistant bacteria and relapse of primary infection. Initial power calculations estimated ~400 episodes would detect an additional benefit C-statistic discriminatory of +0.10.

Results were analysed descriptively, and assessment made of the individual discriminatory value of the inflammatory biomarkers using the C-statistic for SMC, and their additional value to clinical prediction rules (PICNICC prediction and AUS-score). Analyses were undertaken using base R version 3.6.0 (R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria) (RRID:SCR_001905) and the package pROC (pROC: an open-source package for R and S+ to analyze and compare ROC curves) Further analysis of the results was planned to be by hierarchical logistic regression modelling of episodes within patients, assessing the predictive value of clinical and then biomarkers of inflammation at admission, day one and subsequent timepoints. Additional assessments of the sensitivity and specificity of the markers, alone and in combination, were also proposed.

Results

Data was collected from 65 patients with an episode of fever and neutropenia. Of those, 43 episodes in 31 patients had biomarker samples taken, with 31 episodes with ‘day one’ data. The distribution of diagnoses is shown in Table 1. The patients ranged from four months to 17 years old (median six years), and 20 were female.

Of these 31 episodes, five were assigned an SMC: two significant viral infections with oxygen requirement, and one each of: Fusobacterium blood stream infection, Escherichia coli urinary tract infection, and culture-negative severe sepsis requiring fluid boluses. No patient received intensive care support or died. There were eight non-serious viral infections and one non-serious central line colonization.

All biomarkers performed poorly in distinguishing those who developed an SMC (see Figure 1), with C-statistic estimates ranging from 0.48 to 0.60 (see Table 2). The biomarkers were ineffective in improving the discriminatory value of the PICNICC risk prediction (change in C-statistic: −0.10 to +0.04), and only marginally useful in improving the simpler AUS-score (+0.04 to +0.18) (See Extended Data, Table 2)9 Given the paucity of data and lack of diagnostic value, no complex analyses were undertaken.

8c34d9f2-3d8d-44ae-9128-7c17c683e6f9_figure1.gif

Figure 1. Plots of discriminatory values of admission biomarkers in distinguishing patients with serious medical complication (SMC).

CRP = C-reactive protein; PCT = procalcitonin; IL-6 = interleukin-6; IL-8 = interleukin-8.

Table 2. C-statistic values for biomarkers to distinguish patients with serious medical complication (SMC).

CRP = C-reactive protein; PCT = procalcitonin; IL-6 = interleukin-6; IL-8 = interleukin-8.

BiomarkerWith PICNICC predictionWith AUS-score
CRP0.600.690.71
PCT0.540.550.59
IL-60.570.550.57
IL-80.480.640.58
Rule alone0.650.53

The biomarker levels concurred better with the physician assessment of “severe” clinical illness, though in no case was this statistically significantly different (see Table 3, p > 0.10).

Table 3. Mean biomarker levels in patients who did, and did not, appear clinically unwell at presentation.

CRP = C-reactive protein; PCT = procalcitonin; IL-6 = interleukin-6; IL-8 = interleukin-8.

BiomarkerMean (range) ‘severe’ clinical appearance (n = 5)Mean (range) ‘non-severe’ clinical appearance (n = -26)
CRP111 mg/dL (5 − 213)42 mg/dL (3 − 155)
PCT5.8 ng/ml (<0.05 − 29.2)0.03 ng/ml (<0.05 − 3.6)
IL-6105 pg/ml (11.8 − 216)67 pg/ml (1 − 296)
IL-8210 pg/ml (40.9 − 514)176 pg/ml (5 − 718)

Repeated measures of the biomarkers were available in 25 episodes. Development of SMC did not occur in patients who presented with non-severe symptoms and consistently low inflammatory biomarkers despite ongoing fevers. Of those with SMC, 4/5 had reductions in biomarker levels as their infection resolved; they stayed high in the one case of culture negative severe sepsis.

Discussion

This prospective study of inflammatory biomarkers in paediatric febrile neutropenia found little support for their use as indicators of covert infection. Technical and administrative challenges limited the number of samples collected during the study, despite enthusiasm from the patients and their families. This adds to the body of evidence describing a relatively limited role in initial stratification.

Serial use of these markers, where they can be used to suggest an infection is under control, or controlled, is an area of active testing. PCT has been studied in several patient groups to diagnose sepsis and monitor response to treatment. A systematic review11 and subsequent large (>1500 patients) RCT with pragmatic study design examining PCT-guided decision-making in neonates (NeoPIns) with suspected infection and critically ill adults on ICU (SAPS) found significant reductions in antibiotic duration in the PCT arm.12,13 Two large UK studies looking at the same question in adults and children with suspected or proven serious bacterial infections are ongoing.14,15 These studies specifically exclude immunocompromised patients such as children with cancer, and work exploring this is required to enhance patient experience and reduce antimicrobial resistance.1618

The challenges in taking samples consistently over time during acute admissions, often after the ‘working hours’ of the research team, were considerable. In-patients, particularly in the stem cell transplant ward, had blood samples drawn consistently. In other areas the initial sample was often missed, and subsequently no more taken during that episode. The ward teams remembering to request daily additional samples was an unsurmounted barrier. Electronic order sets, now widespread in healthcare, were not present during this study and a subsequent pilot study using PCT had greater sample collection success. Extending the time for the study to run was not possible.

The results of this study should dissuade clinicians from routinely using inflammatory biomarkers in making initial stratification of children with febrile neutropenia. The onward investigation of their serial use in antimicrobial stewardship should be pursued within carefully monitored studies.

Data availability

Underlying data

The data for this study contains the following elements, which when combined would make the patients identifiable given the rarity of childhood cancer in the identified geographical area and time of this study:

Demographic data: (ii) episode date; (iii) age; (iv) cancer diagnosis & treatment.

FN episode data: (i) inpatient or outpatient onset; (ii) time of start of temperature; (iii) time of presentation.

A reduced dataset taking out all items to produce sufficient anonymity (for sensitive data; childhood cancer) would severely limit their utility for researchers.

The patients and families who took part in this study, and the Research Ethics Committee who granted permission for it, agreed the data should be available for sharing in ethically approved secondary use projects. Such studies are typically individual participant data meta-analysis collaboratives. Anyone who has such an approved project, investigating aspects of paediatric febrile neutropenia and biomarker profiles, is encouraged to approach the author at bob.phillips@york.ac.uk for access to the dataset.

Extended data

Open Science Framework (OSF): Extended data for ‘Prospective cohort study of the predictive value of inflammatory biomarkers over clinical variables in children and young people with cancer presenting with fever and neutropenia’, https://www.doi.org/10.17605/OSF.IO/CVFZB.9

This project contains the following extended data:

  • Extended Data Tables.docx (Extended Table 1: Data items collected and Extended Table 2: AUC-ROC (C-statistic) values for biomarkers to distinguish patients with SMC)

  • Consent Video (mp4 format)

Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication).

Consent

Written informed consent for publication of the patients’ details was obtained from the patients/parents/guardian of the patient.

Comments on this article Comments (0)

Version 2
VERSION 2 PUBLISHED 22 Oct 2021
Comment
Author details Author details
Competing interests
Grant information
Copyright
Download
 
Export To
metrics
Views Downloads
F1000Research - -
PubMed Central
Data from PMC are received and updated monthly.
- -
Citations
CITE
how to cite this article
Phillips B. Prospective cohort study of the predictive value of inflammatory biomarkers over clinical variables in children and young people with cancer presenting with fever and neutropenia [version 2; peer review: 2 approved]. F1000Research 2022, 10:1070 (https://doi.org/10.12688/f1000research.73075.2)
NOTE: If applicable, it is important to ensure the information in square brackets after the title is included in all citations of this article.
track
receive updates on this article
Track an article to receive email alerts on any updates to this article.

Open Peer Review

Current Reviewer Status: ?
Key to Reviewer Statuses VIEW
ApprovedThe paper is scientifically sound in its current form and only minor, if any, improvements are suggested
Approved with reservations A number of small changes, sometimes more significant revisions are required to address specific details and improve the papers academic merit.
Not approvedFundamental flaws in the paper seriously undermine the findings and conclusions
Version 2
VERSION 2
PUBLISHED 03 Feb 2022
Revised
Views
9
Cite
Reviewer Report 10 Feb 2022
Paul J. Gibson, Department of Pediatrics, McMaster University, Hamilton, Ontario, Canada 
Approved
VIEWS 9
Thanks to the author for his work and his ... Continue reading
CITE
CITE
HOW TO CITE THIS REPORT
Gibson PJ. Reviewer Report For: Prospective cohort study of the predictive value of inflammatory biomarkers over clinical variables in children and young people with cancer presenting with fever and neutropenia [version 2; peer review: 2 approved]. F1000Research 2022, 10:1070 (https://doi.org/10.5256/f1000research.120929.r122269)
NOTE: it is important to ensure the information in square brackets after the title is included in all citations of this article.
Views
5
Cite
Reviewer Report 07 Feb 2022
Rejin Kebudi, Division of Pediatric Hematology-Oncology, Oncology Institute, Istanbul University, Istanbul, Turkey 
Approved
VIEWS 5
The paper is very valuable ... Continue reading
CITE
CITE
HOW TO CITE THIS REPORT
Kebudi R. Reviewer Report For: Prospective cohort study of the predictive value of inflammatory biomarkers over clinical variables in children and young people with cancer presenting with fever and neutropenia [version 2; peer review: 2 approved]. F1000Research 2022, 10:1070 (https://doi.org/10.5256/f1000research.120929.r122270)
NOTE: it is important to ensure the information in square brackets after the title is included in all citations of this article.
Version 1
VERSION 1
PUBLISHED 22 Oct 2021
Views
19
Cite
Reviewer Report 21 Jan 2022
Rejin Kebudi, Division of Pediatric Hematology-Oncology, Oncology Institute, Istanbul University, Istanbul, Turkey 
Approved
VIEWS 19
The manuscript Prospective cohort study of the predictive value of inflammatory biomarkers over clinical variables in children and young people with cancer presenting with fever and neutropenia investigates the role of PCT, IL6, IL8 at first day and serial monitoring ... Continue reading
CITE
CITE
HOW TO CITE THIS REPORT
Kebudi R. Reviewer Report For: Prospective cohort study of the predictive value of inflammatory biomarkers over clinical variables in children and young people with cancer presenting with fever and neutropenia [version 2; peer review: 2 approved]. F1000Research 2022, 10:1070 (https://doi.org/10.5256/f1000research.76698.r100188)
NOTE: it is important to ensure the information in square brackets after the title is included in all citations of this article.
  • Author Response 03 Feb 2022
    Bob Phillips, Centre for Reviews and Dissemination, University of York, UK, York, YO10 5DD, UK
    03 Feb 2022
    Author Response
    There are a number of interesting questions raised in the review:
    1. Were there any differences in patients who had acute leukemia vs solid tumors?
    This is a ... Continue reading
COMMENTS ON THIS REPORT
  • Author Response 03 Feb 2022
    Bob Phillips, Centre for Reviews and Dissemination, University of York, UK, York, YO10 5DD, UK
    03 Feb 2022
    Author Response
    There are a number of interesting questions raised in the review:
    1. Were there any differences in patients who had acute leukemia vs solid tumors?
    This is a ... Continue reading
Views
20
Cite
Reviewer Report 10 Nov 2021
Paul J. Gibson, Department of Pediatrics, McMaster University, Hamilton, Ontario, Canada 
Approved with Reservations
VIEWS 20
General Comments:

This is a nice small study that tackles an interesting and still unanswered question with regards to the utility of biomarkers in the risk stratification of Pediatric Febrile Neutropenia. However, it appears there were challenges ... Continue reading
CITE
CITE
HOW TO CITE THIS REPORT
Gibson PJ. Reviewer Report For: Prospective cohort study of the predictive value of inflammatory biomarkers over clinical variables in children and young people with cancer presenting with fever and neutropenia [version 2; peer review: 2 approved]. F1000Research 2022, 10:1070 (https://doi.org/10.5256/f1000research.76698.r97615)
NOTE: it is important to ensure the information in square brackets after the title is included in all citations of this article.
  • Author Response 03 Feb 2022
    Bob Phillips, Centre for Reviews and Dissemination, University of York, UK, York, YO10 5DD, UK
    03 Feb 2022
    Author Response
    "General Comments:
    1. This is a nice small study that tackles an interesting and still unanswered question with regards to the utility of biomarkers in the risk stratification of
    ... Continue reading
COMMENTS ON THIS REPORT
  • Author Response 03 Feb 2022
    Bob Phillips, Centre for Reviews and Dissemination, University of York, UK, York, YO10 5DD, UK
    03 Feb 2022
    Author Response
    "General Comments:
    1. This is a nice small study that tackles an interesting and still unanswered question with regards to the utility of biomarkers in the risk stratification of
    ... Continue reading

Comments on this article Comments (0)

Version 2
VERSION 2 PUBLISHED 22 Oct 2021
Comment
Alongside their report, reviewers assign a status to the article:
Approved - the paper is scientifically sound in its current form and only minor, if any, improvements are suggested
Approved with reservations - A number of small changes, sometimes more significant revisions are required to address specific details and improve the papers academic merit.
Not approved - fundamental flaws in the paper seriously undermine the findings and conclusions
Sign In
If you've forgotten your password, please enter your email address below and we'll send you instructions on how to reset your password.

The email address should be the one you originally registered with F1000.

Email address not valid, please try again

You registered with F1000 via Google, so we cannot reset your password.

To sign in, please click here.

If you still need help with your Google account password, please click here.

You registered with F1000 via Facebook, so we cannot reset your password.

To sign in, please click here.

If you still need help with your Facebook account password, please click here.

Code not correct, please try again
Email us for further assistance.
Server error, please try again.