Accepted Manuscript

Comparison of the diagnostic power of cytokine patterns and procalcitonin for

predicting infection among pediatric hematology/oncology patients

Tian Xia, Xiaojun Xu, Ning Zhao, Zebin Luo, Yongmin Tang

PII: S1198-743X(16)30401-3
DOI: 10.1016/j.cmi.2016.09.013
Reference: CMI 725

To appear in: Clinical Microbiology and Infection

Received Date: 4 April 2016

Revised Date: 14 September 2016
Accepted Date: 18 September 2016

Please cite this article as: Xia T, Xu X, Zhao N, Luo Z, Tang Y, Comparison of the diagnostic power
of cytokine patterns and procalcitonin for predicting infection among pediatric hematology/oncology
patients, Clinical Microbiology and Infection (2016), doi: 10.1016/j.cmi.2016.09.013.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Comparison of the diagnostic power of cytokine

patterns and procalcitonin for predicting infection

among pediatric hematology/oncology patients

PT
Tian Xia 1, 2, Xiaojun Xu 1, 2, Ning Zhao 1, 2, Zebin Luo 2, Yongmin

RI
Tang 1, 2, *

SC
1
Division of Hematology-Oncology, Children’s Hospital of Zhejiang University

U
School of Medicine, Hangzhou 310003, PR China
AN
2
Key Laboratory of Reproductive Genetics (Zhejiang University), Ministry of
M

Education, Hangzhou 310003, PR China

D

Abbreviations: IL, interleukin; TNF, tumor necrosis factor; FCM, flow cytometry;
TE

IFN, interferon; ROC, receiver operating characteristics; AUC, area under

EP

curve.
C

* Corresponding author Yongmin Tang, Division of Hematology–Oncology,

AC

Children’s Hospital of Zhejiang University School of Medicine, Hangzhou

310003, PR China. Tel.: +86(571) 88873450; fax: +86 (571) 87033296.

E-mail address: Y_M_TANG@zju.edu.cn (Y. -M. Tang).

1
 ACCEPTED MANUSCRIPT

Abstract

Purpose: The aim of this study was to evaluate the diagnostic power of the

cytokine patterns and serum procalcitonin (PCT) level for predicting infection in

PT
pediatric hematology/oncology patients. Methods: Retrospective study

RI
including hospitalized hematology-oncology children was conducted and their

SC
serum Th1/Th2 cytokines measured by a flow cytometric method were

U
analyzed. According to clinical symptoms, imaging and microbiological
AN
findings, febrile episodes were divided into five diagnostic groups serving as
M

reference standard; and also grouped according to disease severity. A control

D

group consisted of afebrile children. Results: A total of 3023 samples (2819

TE

febrile episodes and 204 control samples) derived from 992 (including 164
EP

afebrile control patients) were obtained. Interleukin (IL)-6 and IL-10 levels as
C

well as their positivity rates were significantly higher among children with
AC

bacteremia than for the viral infection and control groups. Among bacteremic

children, 92.8% (297/320) and 82.2% (263/320) had increased IL-6 and IL-10

levels that exceeded the upper limit of the normal range, respectively. The

2
 ACCEPTED MANUSCRIPT

positivity rates of PCT and C-reactive protein (CRP) were only 33.8% (108/320)

and 73.1% (234/320), respectively, significantly lower than those of IL-6 and

IL-10 (P<0.01). Based on the receiver operating characteristic (ROC) curves,

PT
PCT exhibited poorer sensitivity in the diagnosis of severe infection, as

RI
compared with IL-6 and IL-10 (p<0.01). Specificity of IL-6 and IL-10 was

SC
significantly higher than PCT in the diagnosis of Gram-negative bacteremia.

U
Conclusion: Cytokine patterns of IL-6 and IL-10 showed higher diagnostic
AN
accuracy than PCT for bacteremia and severe infections among febrile
M

hematology-oncology children.
D

Key words
TE

Procalcitonin, Hematology/oncology, Cytokine, Infection, Children

EP

1. Introduction
C

Infections are a common and serious problem in hematology/oncology

AC

patients. The mortality of severe sepsis/septic shock in children with cancer is

about 16% and even higher in adults (1).During profound immunosuppression

by chemotherapy, early diagnosis of infection is difficult, and sometimes even

impossible due to minimal symptoms and signs. Among these patients, fever is

3
 ACCEPTED MANUSCRIPT

often the only sign of infection (2). These patients receive prompt

broad-spectrum antibiotic therapy when fever develops. Increasingly,

neutropenic patients are also given antifungal agents as empiric therapy for

PT
persistent fever. It is crucial to improve the correct diagnosis of infections and

RI
their complications as early as possible with timely implementation of

SC
appropriate treatment.

U
Inflammatory markers, including procalcitonin (PCT), C-reactive protein (CRP),
AN
and cytokines such as IL-10, IL-8, IL-6 and interferon have become useful
M

diagnostic tools for monitoring and predicting prognosis in infections (3-5). An

D

increase in plasma PCT level is associated with sepsis, accompanied by a

TE

pro-inflammatory cytokine response. PCT has been reported to be a superior

EP

marker to CRP in differentiating infections from non-infectious causes of

C

inflammation. Also, PCT is significantly higher in patients with bacterial

AC

infection than those with viral infection (2, 6).

During the past decade, we have successfully applied cytometric bead arrays

(CBA) to determine Gram-positive and Gram-negative bacterial infection in

4
 ACCEPTED MANUSCRIPT

febrile hematology-oncology patients (7, 8). Flow cytometry-based

inflammatory cytokine measurement of IL-6, IL10, and tumor necrosis factor

(TNF)-α were closely related to the pediatric index of mortality (PIM) 2 score

PT
and mortality, which might be useful for timely selection of appropriate

RI
antibiotics (7). However, the diagnostic and predictive power for infection

SC
between proinflammatory cytokines and the current widely used PCT has not

U
been systematically compared in children with neutropenic infection. In the
AN
present study, we further evaluated and compared the diagnostic performance
M

of cytokine profiles (IL-2, IL-4, IL-6, IL-10, TNF-α, and IFN-γ) and PCT for
D

predicting infection in hospitalized febrile hematology-oncology children.

TE

2. Patients and Methods

EP

2.1. Study population

C

Pediatric hematology/oncology patients from January 2011 through

AC

September 2014, who were hospitalized and experienced febrile episodes in

the Hematology/Oncology Department of the Children’s Hospital of Zhejiang

University School of Medicine, were retrospectively analyzed. Inclusion criteria

5
 ACCEPTED MANUSCRIPT

included a febrile episode defined as at least one axillary temperature reading

of >38.5°C or two measurements of 38.0–38.4°C taken within a 24-hour period

and serum cytokine samples were drawn at the beginning of the episode.

PT
Separate episodes of infection were defined as occurring at least two weeks

RI
apart. All consecutive febrile children whose serum cytokines were determined

SC
at the time of febrile episodes were included in this study. A control group

U
consisting of all hematology-oncology children without fever or signs of
AN
infection whose serum cytokine levels were determined during the study
M

period was set. This study was approved by the ethic committee of the hospital,
D

and written informed consent was obtained from patients’ parents or

TE

guardians.
EP

2.2. Reference standard

C

Final diagnosis was made by the hematology-oncology experts in our

AC

department. Patients were assigned to five groups: 1) blood culture positive

sepsis, based on a presence of signs of acute systemic inflammation and

positive isolation of microorganism(s) in cultures of blood; 2) clinical diagnosis

6
 ACCEPTED MANUSCRIPT

of sepsis were based on clinical symptom without isolation of microorganism(s)

in blood cultures; 3) nonsepsis infection was defined if there was a clinically

evident source of infection and/or roentogenographic signs of infection; 4) viral

PT
infection and 5) systemic fungal infection. Viral infection and systemic fungal

RI
infection were considered when there was evidence of Epstein-Barr virus

SC
(EBV)/cytomegalovirus (CMV)/fungal infection other than sepsis. Patients

U
were also grouped by disease severity. Patients who suffered from septic
AN
shock, multiple organ dysfunction syndrome (MODS) or death, were
M

considered to have severe infection.

D

Neutropenia was defined as an absolute neutrophil count in peripheral blood

TE

<0.5×109/L(9). Septic shock was defined as a systolic blood pressure of

EP

<90mmHg or organ dysfunction attributable to sepsis (10).

C

2.3. Data collection

AC

During each febrile episode, patients were examined, and their serum cytokine

levels IL-2, IL-4, IL-6, IL-10, TNF-α, IFN-γ, PCT and CRP) were

simultaneously determined when the patients were admitted or within 24 hours

7
 ACCEPTED MANUSCRIPT

of fever onset in hospital. Blood cultures were collected. Specimens from other

potentially infected areas such as urinary tract and respiratory tract were also

taken as indicated. In the control group, Th1/Th2 cytokine levels were

PT
randomly done as part of routine testing (Figure 1.).

RI
2.4. Sample measurements

SC
The concentrations of IL-2, IL-4, IL-6, IL-10, TNF-α, and IFN-γ in the sera were

quantitatively determined by
U
a CBA Human Th1/Th2 Cytokine Kit II (BD
AN
Biosciences, San Jose, CA) as previously described (11, 12). Six standard
M

curves were obtained from one set of calibrators and six results were obtained
D

on one test sample. The maximum and minimum limits of detection for all the
TE

six cytokines were 1.0 and 5,000 pg/ml, respectively. All results were reported
EP

within 5 hours after the blood samples were drawn.

C

2.5. Statistical analysis

AC

All statistical analyses were performed with SPSS software, version 22.0.

Continuous data are expressed as median and range when not normally

distributed. Between groups, the differences were tested for significance using

8
 ACCEPTED MANUSCRIPT

the Mann–Whitney U-test and Kruskal–Wallis H-test when appropriate.

Receiver operating characteristic (ROC) curves were applied to determine the

sensitivities and specificities for cytokine measurements in febrile patients. The

PT
area under the curve (AUC) and the optimal cutoff value were also calculated.

RI
A two-sided p-value of <0.05 was considered to be statistically significant.

SC
3. Results

3.1. Patient characteristics

U
AN
A total of 3023 samples (2819 febrile episodes samples, 204 control samples)
M

were collected for determination of IL-2, IL-4, IL-6, IL-10, TNF-α, and IFN-γ.
D

Among 2819 febrile episodes, 1585 (56.2%) occurred in patients with

TE

neutropenia. Based on microbiological and clinical evidence, 828 febrile

EP

patients who developed 2819 febrile episodes were assigned to the five
C

reference standard groups (Figure 1). Blood cultures were positive in 320/2819
AC

(11.4%) of febrile episodes of which 164/320 (51.3%), 154/320 (48.1%) and

2/320 (0.6%) were attributable to Gram-positive bacteria, Gram-negative

bacteria and fungi, respectively.

9
 ACCEPTED MANUSCRIPT

Febrile episodes of 2819 children were also grouped according to disease

severity, with 154 severe and 2665 mild infections.

Of 828 febrile patients, there were 530 cases of acute lymphoblastic leukemia

PT
and 162 cases of acute myeloid leukemia. Other underlying malignancies and

RI
baseline characteristics of patients are shown in Table 1.

SC
3.2. Cytokine, PCT and CRP levels in patients with infection

U
The median (ranges) levels of cytokines in the different groups are shown in
AN
Table 2 and Figure 2. In the nonsepsis infection group, IL-6 was elevated at
M

least 2-fold compared with the control group (P < 0.001). PCT and CRP levels
D

were 0.5 (0.02-45) ng/ml and 36.7 (1-200) mg/L, respectively. In the positive
TE

blood culture sepsis group, IL-6 and IL-10 (688.2 pg/mL and 208.6.0 pg/mL,
EP

respectively) were significantly increased compared with the control group

C

(6.8pg/mL and 5.5pg/mL, respectively). PCT and CRP levels were 2.2
AC

(0.04-75) ng/ml and 54.4 (1-180) mg/L, respectively and both levels were

significantly higher than those found in the nonsepsis infection group (P

<0.001). Patients in all groups showed similar IL-2 and IL-4 levels. TNF-α was

10
 ACCEPTED MANUSCRIPT

found to be elevated in 18.1% (58/320) of positive blood culture sepsis children.

In the sepsis group, the 6 cytokine levels were not significantly different

between patients with neutropenia and those with neutrophil count 500

PT
cells/mm3 (P >0.05). Among the viral infection group, both IL-6 and IL-10

RI
levels were significantly lower than those in other febrile episode groups (P

SC
<0.001).

U
3.3. Positivity rates of Cytokines, PCT and CRP in groups
AN
IL-6 was predominantly elevated in patients with positive blood culture sepsis,
M

with 92.8% (297/320) patients having an level exceeding the upper limit of the
D

normal range (16.6 pg/mL) and 53 of 320 (16.6%) cases exceeding 1000
TE

pg/mL. Similarly, 263 (82.2%), 58 (18.1%), 82 (25.6%) patients in this group

EP

exhibited increased IL-10, TNF-α and IFN-γ levels that exceeded the upper
C

limit of the normal range, respectively. However, CRP was found positive in
AC

73.1% (234/320) cases, whereas PCT was only detected in 33.8% (108/320)

patients, suggesting an unacceptable low positivity rate of PCT in such

patients compared with IL-6 and IL-10 (P <0.05). In the Gram-negative

11
 ACCEPTED MANUSCRIPT

bacteremia group, 96.8% (149/154) and 90.3% (139/154) cases had increased

IL-6 and IL-10 levels that exceeded the upper limit of the normal range, in

contrast with 45.5% (70/154) and 77.3% (119/154) of positivity rates for PCT

PT
and CRP. In the Gram-positive bacteremia group, the positivity rates of IL-6,

RI
IL-10, PCT and CRP were 89.0% (146/164), 74.4% (122/164), 22.6% (37/164)

SC
and 68.9% (113/164), respectively.

U
3.4. Receiver operating characteristic curve analysis
AN
The Th1/Th2 cytokine patterns differed among infections (Table 2 and Figure
M

2). ROC curves were built based on the IL-6, IL-10, PCT and CRP levels for all
D

febrile patients. In the Gram-negative bacteremia group, the areas under the
TE

curve for IL-6, IL-10, PCT, and CRP were 0.686 (95% CI 0.622-0.750), 0.747
EP

(95% CI 0.688-0.806), 0.657 (95% CI 0.591-0.723) and 0.596 (95% CI

C

0.527-0.666), respectively. In the severe infection group, the areas under the
AC

curve for IL-6, IL-10, PCT, and CRP were 0.875 (95% CI 0.847-0.905), 0.839

(95% CI 0.806-0.880), 0.726 (95% CI 0.679-0.775), and 0.648 (95% CI

0.598-0.697) respectively, indicating that IL-6 and IL-10 could be significant

12
 ACCEPTED MANUSCRIPT

discriminators for patients with severe infection. The diagnostic performance of

IL-6, IL-10, PCT and CRP based on their optimal cut-off values are shown in

Table 3. When setting cutoff points of IL-6 and IL-10 at 500 pg/mL, and 100

PT
pg/mL, the sensitivity rates were 54.4% and 47.1%, and specificity rates were

RI
92.5% and 94.9%, respectively for predicting severe infection. Regarding PCT,

SC
when the cutoff point was set at 2 ng/ml, the sensitivity and specificity were

U
29.4% and 95%, respectively, revealing a significantly lower sensitivity for
AN
predicting severe infection as compared with IL-6 and IL-10. Taken together,
M

IL-6 and IL-10 presented a much better performance than PCT in predicting
D

severe infection for febrile hematology-oncology children.

TE

4. Discussion
EP

In this study we evaluated the role of PCT in infection prediction together with
C

cytokine profiles in febrile children. IL-6 and IL-10 provided better predictive
AC

value for the diagnosis of bacteremia and severe infection than PCT in febrile

hematology-oncology children. Our study has shown that, in febrile children,

elevated levels of IL-6 and IL-10 are closely related to bacterial infection and

13
 ACCEPTED MANUSCRIPT

infection severity, which is consistent with our previous reports (7, 11). Septic

shock was associated with further elevation of IL-6 and IL-10. The sensitivity of

PCT was significantly lower than IL-6 and IL-10 for predicting infection severity

PT
and the specificity of PCT was poor in the diagnosis of Gram-negative

RI
bacteremia.

SC
Previous studies have evaluated diverse markers of inflammation as

U
predictors for different types of infection (5, 13-18). Kitanovski et al. found that
AN
on admission and 24 hours later, IL-6 and PCT are more accurate for
M

predicting bacteremia or clinical sepsis compared to

D

the lipopolysaccharide-binding protein (LBP) in 90 febrile neutropenia

TE

episodes were experienced by 47 children (19). Surbatovic et al. reported that

EP

in a limited patient population (165 adult patients) with abdominal sepsis,

C

Gram-negative infections resulted in 2-fold increase in TNF-α, 3.3-fold in IL-8

AC

and 13-fold in IFN-γ as compared with Gram-positive infection, thereby

discriminating Gram-negative from Gram-positive bacteremia, also predicting

survival (20). Similar to our results, they demonstrated that IL-10 was a fairly

14
 ACCEPTED MANUSCRIPT

good predictor of Gram-negative bacteremia. Vanska et al. found a high

negative predictive value for PCT and provided new evidence for IL-10 as an

early predictor for complicated course of febrile neutropenia in hematological

PT
patients (21). Combining IL-10 with PCT improved the early prediction for a

RI
complicated course of febrile neutropenia. It has been reported that PCT and

SC
interleukin-2 receptor (IL-2R) determination might be used as an additional

U
diagnostic tool for the detection of bacteremia/sepsis in childhood oncology
AN
patients with febrile neutropenia (22).
M

PCT is helpful for the diagnosis of bacterial infection, especially in septic shock
D

and severe sepsis. IL-6 exhibits better role for monitoring the effectiveness of
TE

antibiotic treatment (23). Among 177 consecutive patients with a systemic

EP

inflammatory response syndrome, PCT showed good diagnostic performance,

C

with 74.4% and 93.7% sensitivity and 86.7% and 75.2% specificity for sepsis
AC

and severe sepsis/septic shock, respectively (23). . However, the findings in

this study with a very large sample size showed that IL-6 and IL-10 are much

better diagnostic power than PCT.

15
 ACCEPTED MANUSCRIPT

Our study provides new insight on diagnostic performance of IL-6, IL-10 and

PCT on infection in febrile pediatric hematology-oncology patients with a large

sample size. Compared with the troublesome and time-consuming

PT
enzyme-linked immunosorbent assay (ELISA) method, which determines a

RI
single cytokine level (24), CBA is more rapid and more cost- and time-effective.

SC
Results can be obtained within a few hours (<5 hours), providing promising

U
clinical application in urgent infections, including guiding the selection of
AN
antibiotics. The differences between underlying diseases warrant further
M

investigation. The shortcomings of this study were its retrospective nature and
D

the fact that this was a single center experience. A multi-centered and
TE

prospective study is warranted in the near feature. We are now planning to

EP

launch such a valuable study.

C

In summary, the specific cytokine pattern of highly elevated IL-6 and IL-10
AC

shows higher diagnostic accuracy than PCT for bacterial and severe infection

among febrile patients with pediatric hematology-oncology diseases.

Acknowledgement: This study was supported in part by grants from the National Natural

Science Foundation of China (No: 81170502, 81470304), the Zhejiang Provincial Natural

16
 ACCEPTED MANUSCRIPT

Science Foundation of China (No: LZ12H08001). The authors would also like to thank all

other physicians in hematology/oncology division in the Children’s Hospital of Zhejiang

University School of Medicine for their support.

PT
Conflict of Interest: All the authors declare no conflict of interests.

RI
U SC
AN
M
D
TE
C EP
AC

17
 ACCEPTED MANUSCRIPT

References
1. Weiss SL, Fitzgerald JC, Pappachan J, Wheeler D,
Jaramillo-Bustamante JC, Salloo A, Singhi SC, Erickson S, Roy JA,
Bush JL, Nadkarni VM, Thomas NJ. Global epidemiology of pediatric
severe sepsis: the sepsis prevalence, outcomes, and therapies study.

PT
American journal of respiratory and critical care medicine
2015;191:1147-1157.

RI
2. Mian A, Becton D, Saylors R, James L, Tang X, Bhutta A, Prodhan P.
Biomarkers for risk stratification of febrile neutropenia among children

SC
with malignancy: a pilot study. Pediatric blood & cancer
2012;59:238-245.

U
3. Harbarth S, Holeckova K, Froidevaux C, Pittet D, Ricou B, Grau GE,
Vadas L, Pugin J, Geneva Sepsis N. Diagnostic value of procalcitonin,
AN
interleukin-6, and interleukin-8 in critically ill patients admitted with
suspected sepsis. American journal of respiratory and critical care
M

medicine 2001;164:396-402.
4. Karakurt DG, Demirsoy U, Corapcioglu F, Oncel S, Karadogan M,
D

Arisoy ES. Do proinflammatory cytokine levels predict serious

TE

complication risk of infection in pediatric cancer patients? Pediatric

hematology and oncology 2014;31:415-424.
5. Badurdeen S, Hodge G, Osborn M, Scott J, St John-Green C, Tapp H,
EP

Zola H, Revesz T. Elevated serum cytokine levels using cytometric

bead arrays predict culture-positive infections in childhood oncology
C

patients with febrile neutropenia. Journal of pediatric

AC

hematology/oncology 2012;34:e36-38.
6. Karanwal AB, Parikh BJ, Goswami P, Panchal HP, Parekh BB, Patel KB.
Review of clinical profile and bacterial spectrum and sensitivity patterns
of pathogens in febrile neutropenic patients in hematological
malignancies: A retrospective analysis from a single center. Indian
journal of medical and paediatric oncology 2013;34:85-88.
7. Xu XJ, Tang YM, Liao C, Song H, Yang SL, Xu WQ, Shi SW, Zhao N.
Inflammatory cytokine measurement quickly discriminates

18
 ACCEPTED MANUSCRIPT
gram-negative from gram-positive bacteremia in pediatric
hematology/oncology patients with septic shock. Intensive care
medicine 2013;39:319-326.
8. Tang Y, Liao C, Xu X, Song H, Shi S, Yang S. Th1/Th2 cytokine profiles
in G+/G- bacteremia in pediatric hematology/oncology patients.
Pediatric blood & cancer 2012;58:50-54.

PT
9. Soker M, Colpan L, Ece A, Devecioglu C, Haspolat K. Serum levels of
IL-1 beta, sIL-2R, IL-6, IL-8, and TNF-alpha in febrile children with

RI
cancer and neutropenia. Medical oncology 2001;18:51-57.
10. Sprung CL, Annane D, Keh D, Moreno R, Singer M, Freivogel K, Weiss

SC
YG, Benbenishty J, Kalenka A, Forst H, Laterre PF, Reinhart K,
Cuthbertson BH, Payen D, Briegel J, Group CS. Hydrocortisone

U
therapy for patients with septic shock. The New England journal of
medicine 2008;358:111-124.
AN
11. Tang Y, Liao C, Xu X, Song H, Shi S, Yang S, Zhao F, Xu W, Chen X,
Mao J, Zhang L, Pan B. Evaluation of Th1/Th2 cytokines as a rapid
M

diagnostic tool for severe infection in paediatric haematology/oncology

patients by the use of cytometric bead array technology. Clinical
D

microbiology and infection 2011;17:1666-1673.

TE

12. Tang Y, Xu X, Song H, Yang S, Shi S, Wei J, Pan B, Zhao F, Liao C, Luo
C. Early diagnostic and prognostic significance of a specific Th1/Th2
cytokine pattern in children with haemophagocytic syndrome. British
EP

journal of haematology 2008;143:84-91.

13. von Lilienfeld-Toal M, Dietrich MP, Glasmacher A, Lehmann L, Breig P,
C

Hahn C, Schmidt-Wolf IG, Marklein G, Schroeder S, Stuber F. Markers

AC

of bacteremia in febrile neutropenic patients with hematological

malignancies: procalcitonin and IL-6 are more reliable than C-reactive
protein. European journal of clinical microbiology & infectious diseases
2004;23:539-544.
14. Cost CR, Stegner MM, Leonard D, Leavey P. IL-8 predicts pediatric
oncology patients with febrile neutropenia at low risk for bacteremia.
Journal of pediatric hematology/oncology 2013;35:206-211.
15. Hodge G, Osborn M, Hodge S, Nairn J, Tapp H, Kirby M, Sepulveda H,

19
 ACCEPTED MANUSCRIPT
Morgan E, Revesz T, Zola H. Rapid simultaneous measurement of
multiple cytokines in childhood oncology patients with febrile
neutropenia: increased interleukin (IL)-8 or IL-5 correlates with
culture-positive infection. British journal of haematology
2006;132:247-248.
16. El-Maghraby SM, Moneer MM, Ismail MM, Shalaby LM, El-Mahallawy

PT
HA. The diagnostic value of C-reactive protein, interleukin-8, and
monocyte chemotactic protein in risk stratification of febrile neutropenic

RI
children with hematologic malignancies. Journal of pediatric
hematology/oncology 2007;29:131-136.

SC
17. Kiszewska N, Bien E, Irga-Jaworska N, Adamkiewicz-Drozynska E.
Selected inflammatory markers in the diagnosis and monitoring of

U
infections in children treated for hematological malignancies.
Biomarkers in medicine 2015;9:461-471.
AN
18. Tromp YH, Daenen SM, Sluiter WJ, Vellenga E. The predictive value of
interleukin-8 (IL-8) in hospitalised patients with fever and
M

chemotherapy-induced neutropenia. European journal of cancer

2009;45:596-600.
D

19. Kitanovski L, Jazbec J, Hojker S, Derganc M. Diagnostic accuracy of

TE

lipopolysaccharide-binding protein for predicting bacteremia/clinical

sepsis in children with febrile neutropenia: comparison with interleukin-6,
procalcitonin, and C-reactive protein. Supportive care in cancer
EP

2014;22:269-277.
20. Surbatovic M, Popovic N, Vojvodic D, Milosevic I, Acimovic G, Stojicic M,
C

Veljovic M, Jevdjic J, Djordjevic D, Radakovic S. Cytokine profile in

AC

severe Gram-positive and Gram-negative abdominal sepsis. Scientific

reports 2015;5:11355.
21. Vanska M, Koivula I, Jantunen E, Hamalainen S, Purhonen AK, Pulkki K,
Juutilainen A. IL-10 combined with procalcitonin improves early
prediction of complications of febrile neutropenia in hematological
patients. Cytokine 2012;60:787-792.
22. Urbonas V, Eidukaite A, Tamuliene I. The predictive value of soluble
biomarkers (CD14 subtype, interleukin-2 receptor, human leucocyte

20
 ACCEPTED MANUSCRIPT
antigen-G) and procalcitonin in the detection of bacteremia and sepsis
in pediatric oncology patients with chemotherapy-induced febrile
neutropenia. Cytokine 2013;62:34-37.
23. Jekarl DW, Lee SY, Lee J, Park YJ, Kim Y, Park JH, Wee JH, Choi SP.
Procalcitonin as a diagnostic marker and IL-6 as a prognostic marker
for sepsis. Diagnostic microbiology and infectious disease

PT
2013;75:342-347.
24. Leng SX, McElhaney JE, Walston JD, Xie D, Fedarko NS, Kuchel GA.

RI
ELISA and multiplex technologies for cytokine measurement in
inflammation and aging research. The journals of gerontology.

SC
2008;63:879-884.

U
AN
M
D
TE
C EP
AC

21
 ACCEPTED MANUSCRIPT

Figure captions

FIG.1.TIF An overview of the study workflow. CBC: complete blood count,

LRTI: lower respiratory tract infection, URTI: upper respiratory tract infection.

FIG.2.TIF Comparison of cytokine levels among patients with blood culture

positive sepsis, nonsepsis infection, versus control. (A) Interleukin (IL)-2; (B)

PT
IL4; (C) IL-6; (D) IL10; (E) tumour necrosis factor (TNF)- α; (F) interferon
(IFN)-γ.

RI
FIG.3. TIF ROC curves in patients with Gram-negative bacterial infection and

SC
severe infection. (A) Gram-negative bacterial infection; (B) Severe infection

U
AN
M
D
TE
C EP
AC

22
 ACCEPTED MANUSCRIPT

TABLE 1. Characteristics of patients with infection in

pediatric hematology-oncology patients
Positive
blood clinical Nonsepsis Viral Systemic
Control
culture sepsis infection infection fungal

sepsis

Total no. of febrile episode 204 320 1155 1324 16 4

PT
7.4 6.8 7.6 6.8 6.9 9.4
Age
(0.1-16.3) (0.7-16.1) (0.1-17.0) (0.1-17.5) (0.6-14.7) (2.6-14.6)
Sex, M/F 142:62 202:118 691:464 822:502 8:8 2:2

RI
Temperature( ) 39.0 / 201 844 1017 15 4
39.0-40.0 / 112 294 292 1 0

SC
40.0 / 7 17 15 0 0
Underlying diseases
Acute lymphoblastic
95 231 654 888 8 2

U
leukemia
Acute myeloid leukemia 49 65 355 252 3 0
AN
Acute leukemia 0 0 3 11 0 0
Chronic myeloid leukemia 16 2 5 12 1 0
Lymphoma 15 7 86 85 3 0
M

Langerhans cell
18 6 17 22 0 0
histiocytosis
Myelodysplastic syndrome 2 0 1 6 1 0
D

Others 9 9 34 48 0 2
TE
C EP
AC
 ACCEPTED MANUSCRIPT

TABLE 2. Median levels and ranges of Th1/Th2 cytokines, PCT and CRP by group

IL-2 IL-4 IL-6 IL-10 TNF-α IFN-γ PCT CRP

Groups No.
(pg/ml) (pg/ml) (pg/ml) (pg/ml) (pg/ml) (pg/ml) (ng/mL) (mg/L)

2.8 2.4 6.8 5.5 2.2 6.0 0.1 3.5

Control 204
(2.6-3.0) (0.6-4.6) (1.0-32.0) (1-26.7) (1.0-10.1) (1.4-17.6) (0.02-0.24) (1-14)

PT
2.2 2.1 20.9 6.2 2.7 7.2 0.2 16.4
Viral infection 16
(1.0-4.5) (0.7-3.8) (1.7-55.0) (2.4-16.9) (1.4-4.5) (2.3-35.4) (0.04-0.4) (1-78)

2.8 2.4 58.9 6.2 2.1 46.8 0.6 19.0

Systemic fungal infection 4

RI
(1.9-3.5) (0.9-3.6) (2.7-129.8) (2.4-11.2) (1.5-2.8) (2.2-170.8) (0.1-1.2) (1-40)

3.3 2.7 688.2 208.6 60.1 36.4 2.2 54.4

Positive blood culture sepsis 320
(1.0-66.4) (0.6-18.8) (3.0-＞5000) (1.5-＞5000) (1.0-＞5000) (1.0-2475.8) (0.04-75) (1-180)

SC
3.2 2.7 291.5 92.6 10.7 25.1 1.2 58.0
Clinically diagnosed sepsis 1155
(1.0-118.6) (0.4-103.5) (2.0-＞5000) (1.0-＞5000) (1.0-2347.9) (1.0-1817.4) (0.02-84) (1-184)

3.1 2.6 163.2 21.1 7.0 23.7 0.5 36.7

Nonsepsis infection 1324
(1.0-97.8) (0.4-30) (1.6-＞5000) (1.0-＞5000) (1.0-602.2) (1.2-1175.7) (0.02-45) (1-200)

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

TABLE 3. Diagnostic accuracy of Th1/Th2 cytokines and PCT for patients

Prediction of Gram-negative
Prediction of severe infection (%)
bacteremia (%)

Cutoffs Sensitivity Specificity Sensitivity Specificity

IL-6 (pg/ml) 160 68.2 63.2 80.9 79.7

279.4 56.9 75.4 68.6 86.6

PT
500 40.3 82.4 54.4 92.5
IL-10 (pg/ml) 32.3 69.2 73.1 65.7 87.5
49.5 60.8 77.6 57.1 91.1

RI
100 45.7 84.8 47.1 94.9
PCT (ng/ml) 0.5 50.4 74.4 53.7 80.2

SC
1 32.6 84.8 41.2 90.6
2 18.6 88 29.4 95
CRP (mg/L) 20 67.4 43.2 75.7 43.3

U
40 51.2 63.2 61 59.9
100 26.4 84 36 83.8
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
EP
C
AC