ORIGINAL ARTICLE: HEPATOLOGY

Role of Procalcitonin and C-Reactive Protein

As Biomarkers of Infection in Children With Liver Disease

Rishi Bolia, Anshu Srivastava, yRungmei Marak, Surender K. Yachha, and Ujjal Poddar

ABSTRACT

Objectives: Early and accurate identification of infection in patients with

What Is Known
liver disease is challenging. The present study evaluated the role of
procalcitonin (PCT) and C-reactive protein (CRP) as biomarkers of bacterial  Infection with organ failure is an important cause of
infection in children with liver disease.
Downloaded from https://journals.lww.com/jpgn by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3TDbD+Y6NAIFSoahMi+ihjAak6bYB6wf20OUGu38U4Mw= on 07/10/2018

mortality in patients with liver disease.

Methods: Demographic and clinical data of consecutive children admitted  Early identification of infection in these patients is
with acute liver failure (ALF) or decompensated chronic liver disease
essential, but challenging.
(DCLD) were collected. PCT and CRP were measured within 24 hours
of admission. Blood and urine culture, chest x-ray, and ascitic fluid analysis
What Is New
were done.
Results: One hundred sixty-four children (113 boys, age 76 [0.5–204]  Procalcitonin and C-reactive protein are reliable mar-
months, ALF 69, DCLD 95) were enrolled. Seventy-seven (47%) had
kers of infection and correlate with infection severity
infection. Most common site was ascitic fluid (n ¼ 35), followed by
in children with liver disease, both acute liver failure
urinary tract (n ¼ 26), pneumonia (n ¼ 22), and blood stream infection
and decompensated chronic liver disease.
(n ¼ 16). Twenty-one children had multiple-site infections, 18 had severe  Measuring the biomarkers in children with liver dis-
sepsis, and 36 had systemic inflammatory response syndrome. PCT and CRP
ease helps to identify patients with and without
correlated with infection severity, higher in severe sepsis (8.3 [3.5–38]
infection.
ng/mL and 4.1 [0.3–13.8] mg/dL) than only infection (0.89 [0.1–8] ng/mL  The present information is a useful guide for
and 1.7 [0.32–24] mg/dL) and no infection (0.3 [0.1–6.75] ng/mL and 0.3
judicious use of antibiotics, improving patient out-
[0.1–4.16 mg/dL]). Systemic inflammatory response syndrome was
come and preventing inappropriate antibiotic use
more common in patients with infection (31/77 vs 5/87, P ¼ 0.00). PCT
and resistance.
(>0.5 ng/mL) and CRP (>0.6 mg/dL) performed better in DCLD (AUC of
0.90 and 0.83) compared with patients with ALF (AUC of 0.73 and 0.69).
Conclusions : PCT and CRP are reliable markers of infection and correlate
with infection severity in children with liver disease. Their diagnostic
accuracy is better in DCLD than ALF cases. peripheral white blood cell count are poor indicators of bacterial
infection in patients with liver disease because of their absence
Key Words: children, C-reactive protein, infection, liver disease,
in approximately 30% of cases (2), although at the same time
procalcitonin
identification of a bacterial infection by culture requires 48 to
72 hours to complete. Hence, there is a need for use of
(JPGN 2016;63: 406–411) alternative methods such as biomarkers that can identify infection
quickly, accurately, and also at an earliest stage in patients with

I nfection with organ failure is an important cause of mortality in

patients with liver disease (1). Timely detection of infection
and initiation of appropriate therapy is pivotal for a good outcome.
Identifying patients with infection helps in early initiation of
therapy and possibly a better outcome, whereas ruling out of
infection prevents the inadvertent use of antibiotics and subsequent
development of resistance. Parameters such as fever and increased
Received November 13, 2015; accepted February 23, 2016.
From the Departments of Pediatric Gastroenterology, and yMicrobiology,
Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow,
India.
Address correspondence and reprint requests to Surender K. Yachha, MD,
DM, Department of Pediatric Gastroenterology, Sanjay Gandhi
Postgraduate Institute of Medical Sciences, Raebareli Road, Lucknow
226 014, Uttar Pradesh, India (e-mail: skyachha@yahoo.co.in).
The authors report no conflicts of interest.
Copyright # 2016 by European Society for Pediatric Gastroenterology,
Hepatology, and Nutrition and North American Society for Pediatric
Gastroenterology, Hepatology, and Nutrition
DOI: 10.1097/MPG.0000000000001181
liver disease.
C-reactive protein (CRP) and procalcitonin (PCT) are com-
monly used as diagnostic markers of infection in children; however,
as the liver produces these biomarkers, their diagnostic utility in the
setting of liver dysfunction is a matter of speculation. There is little
information about these markers in children with liver disease.
CRP is an acute-phase reactant synthesized by the liver
within 6 hours of onset of inflammation and tissue necrosis. In a
state of inflammation, CRP binds the phospholipid components of
microorganisms, facilitating their removal by macrophages. It has
been used to identify significant inflammation and infection in
children (3).
PCT is a 116 amino acid protein, a calcitonin precursor,
which in healthy individuals is produced by C-type cells of the
thyroid gland. During inflammation, it is produced from neuro-
endocrine cells in nonthyroidal tissues such as lung, liver, or
kidney. This is because microbial infection induces a ubiquitous
increase of the CALC-1 gene expression resulting in a constitutive
release of PCT from other parenchymal tissues in the body (4). PCT
levels have been found to correlate with the severity of microbial
infection (5).

406 JPGN  Volume 63, Number 4, October 2016

Copyright © ESPGHAN and NASPGHAN. All rights reserved.

JPGN  Volume 63, Number 4, October 2016
Systemic inflammatory response syndrome (SIRS) is the
clinical manifestation of inflammation occurring in the body. It
occurs as a result of cytokines and reactive oxygen species produced
by phagocytes in response to a variety of stimuli. In critically ill
children, it has been found to be a precursor of sepsis in nearly one
quarter of cases (6). In children with liver disease the correlation of
SIRS with infection, however, remains unknown.
PCT concentration has been found to be a better diagnostic
marker of infection than CRP or total leukocyte count (TLC) in
critically ill children (7). It has, however, not been evaluated in
children with acute liver failure (ALF) or decompensated chronic
liver disease (DCLD). The aim of our study was to evaluate the role
of PCT and CRP as serologic biomarkers of bacterial infection in
children with liver disease, both ALF and DCLD.
METHODS
Children with ALF or DCLD admitted in the department of
pediatric gastroenterology at Sanjay Gandhi Postgraduate Institute
of Medical Sciences, Lucknow, India, between March 2013 and
December 2014 were prospectively enrolled in the study. ALF was
defined as per the PALFSG criteria, that is, biochemical evidence of
liver injury, no history of known chronic liver disease, and coagulo-
pathy not corrected by vitamin K administration (INR > 1.5 in
patients with hepatic encephalopathy, or >2.0 in patients without
hepatic encephalopathy) (8). The diagnosis of CLD was based on
liver biopsy or on clinical, imaging, and endoscopic findings
(grade II esophageal varices). Decompensation was defined by
the presence of ascites, encephalopathy, and/or gastrointestinal
bleeding. Child-Pugh score was used to categorize the severity
of liver disease in children with DCLD (9).
Children satisfying the inclusion criteria were evaluated for
infection at admission. Clinical features were recorded. All patients
had a complete hemogram, liver function tests including prothrom-
bin time, renal function tests, serum electrolytes, and blood sugar
testing at admission to our hospital. Blood and urine cultures were
done in all of the patients. Diagnostic paracentesis and ascitic fluid
analysis were done in patients with ascites and x-ray of chest in all
of those with suspected chest infection. Blood samples for PCT and
CRP were drawn at admission and processed immediately; how-
ever, if the patient was admitted after hours, the samples were stored
at 48C and measured within 24 hours of admission (10). PCT
(normal range 0.00–0.5 ng/mL) was determined by a specific and
ultrasensitive immuneluminometric assay (LUMI test, PCT, Brahms
Ag, Berlin, Germany). CRP (normal range 0.00–0.6 mg/dL) was
measured using a turbimetric assay (ILAD900, Instrumentation
laboratory, Milan Italy). TLC was measured using an automated
counter. Standard age related cutoffs were used for evaluating raised
TLC (1 month–1 year of age >17.5  103/mm3, 2–5 years
>15.5  103/mm3, 6–12 years >13.5  103/mm3, and >12 years
>11  103/mm3) (11).
Criteria for the diagnosis of infection were the presence of
any 1 or more of the following: ascitic fluid infection (AFI;
polymorphonuclear cell count in ascitic fluid 250/mm3and/or
positive culture), blood stream infection (BSI; positive blood
culture), urinary tract infection (UTI; presence of pyuria and
100,000 colonies per mL of a single uropathogenic organism),
pneumonia (clinical features and chest x-ray findings). Infections at
other sites were defined by standardized criteria proposed by the
Centers for Disease Control and Prevention proven by microbio-
logical, radiological, or surgical findings (12).
SIRS was defined as per the International pediatric sepsis
consensus conference definition (13). The presence of 2 or more of
the following 4 criteria, 1 of which must be abnormal temperature or
leukocyte count for the diagnosis of SIRS; the 4 criteria are
www.jpgn.org
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Role of Procalcitonin and C-Reactive Protein
abnormal core temperature (38.58C or 368C), tachycardia
(mean heart rate 2 SD above normal for age in the absence of
external stimulus, chronic drugs, or painful stimuli), tachypnea
(mean respiratory rate 2 SD above normal for age or mechanical
ventilation), and TLC elevated or depressed for age. Sepsis was
defined as presence of infection with features of SIRS (12). Severe
sepsis was defined as sepsis plus 1 of the following: cardiovascular
organ dysfunction or acute respiratory distress syndrome or 2 or
more organ dysfunctions (respiratory, renal, neurologic, or hema-
tologic). Organ function was defined as per the criterion in the
consensus statement (13).
Patients with infection were labeled as group I and without
infection as group II. The 2 groups were compared for clinical
features of infection, TLC, PCT, and CRP. Subgroup analyses were
done for type of liver disease (ALF and DCLD), number of sites of
infection (multiple vs single), and severity of infection (infection vs
severe sepsis).
The study was approved by the institute ethics committee
(IEC code: 2013–12-DM-67). Written informed consent was
obtained from enrolled patients.
Statistics
All data are expressed as median with range and percentages.
The Mann-Whitney U test was used to compare continuous vari-
ables between groups, and dichotomous variables were compared
by the chi-square and Fisher exact test. Receiver operating curves
(ROCs) were drawn by the SPSS software version 20.0 (IBM SPSS
Statistics, Armonk, NY) to calculate the area under curve. The
cutoff value was selected depending on (Sp þ Se) max. Sensitivity,
specificity, and likelihood ratio (LR) were calculated to determine
the predictive power of PCT, CRP, TLC, and SIRS in distinguishing
between infected and noninfected patients.
RESULTS
One hundred sixty-four children, (113 boys, age 76 [0.5–
204] months, DCLD 95, ALF 69) were enrolled in the present study.
The indications for admission in patients with DCLD were manage-
ment of hepatic encephalopathy (n ¼ 40 [42%], 24 also had ascites),
ascites (n ¼ 20 [21%]), variceal bleeding (n ¼ 14 [14.8%], 2 also
had ascites), and suspected infection (n ¼ 21 [22.2%], respiratory
tract 7, AFI 7, cholangitis 3, urinary tract 1, no focus 2, all 21 cases
had ascites).
Infection at admission (group I) was detected in 46.9%
(n ¼ 77) cases, and the other 53.1% (n ¼ 87) had no infection (group
II). In group I, 72.7% (56/77) children had single site and 27.3%
(21/77) had multiple-site infections. The sites of infection were AFI
in 35 (45.4%), UTI in 26 (33.8%), pneumonia in 22 (28.6%), and
BSI in 16 (21%) cases. One patient had cholangitis.
Comparing groups I and II, subjects in group I had signifi-
cantly higher PCT (1.7 [0.1–38.07] ng/mL vs 0.3 [0.1–6.75] ng/
mL; P ¼ 0.00) and CRP (1.82 [0.3–24] mg/dL vs 0.33 [0.1–4.16]
mg/dL; P ¼ 0.00) compared with group II cases. Presence or
absence of fever did not differentiate patients with and without
infection (Table 1). The proportion of patients with a raised PCT
(>0.5 ng/mL, 66/77 [85.7%] vs 23/87 [26.4%]; P ¼ 0.00), and
raised CRP (>0.6 mg/dL, 60/77 [77.9%] vs 28/87 [32.1%];
P ¼ 0.00) were significantly higher in group I than group II. When
patients with ALF (n ¼ 69) and DCLD (n ¼ 95) were analyzed
separately, the values of PCT and CRP in group I patients continued
to be significantly higher (Table 1).
Alhough the TLC was higher in patients in group I
compared with group II (14,000 [3600–43800] per mm3 vs 8800
[3800–39600] per mm3; P ¼ 0.00), the proportion of cases with
407
Bolia et al JPGN  Volume 63, Number 4, October 2016

TABLE 1. Comparison of PCT, CRP, SIRS, and presence of fever in infected and noninfected patients with liver disease

Parameter Infection (n ¼ 77) (DCLD 53, ALF 24) No infection (n ¼ 87) (DCLD 42, ALF 45) P

Procalcitonin, ng/mL
All patients 1.70 (0.10–38.07) 0.30 (0.10–6.75) 0.00
DCLD 1.72 (0.10–23.60) 0.21 (0.06–6.16) 0.00
ALF 1.33 (0.10–38.07) 0.46 (0.10–6.75) 0.00
C-reactive protein, mg/dL
Overall 1.82 (0.30–24) 0.33 (0.10–4.16) 0.00
DCLD 2.10 (0.30–24) 0.30 (0.10–4.16) 0.00
ALF 0.85 (0.32–6.32) 0.30 (0.24–4.07) 0.00
Presence of fever
Overall 27 (35%) 20 (23%) 0.10
DCLD 21 (40%) 15 (36%) 0.51
ALF 6 (25%) 5 (11%) 0.17
SIRS
Overall 31 (40%) 5 (5.7%) 0.00
DCLD 17 (32%) 1 (2.3%) 0.00
ALF 14 (58%) 4 (9%) 0.00

ALF ¼ acute liver failure; CRP ¼ C-reactive protein; DCLD ¼ decompensated chronic liver disease; PCT ¼ procalcitonin; SIRS ¼ systemic inflammatory
response syndrome.

high TLC (>age-dependent cutoffs) was not significantly different
between the 2 groups [33/77 (42.8%) vs 22/87 (25.2%); P ¼ 0.77].
The proportion of patients with raised PCT, CRP, and TLC was not
different in subjects with various sites of infection, that is, BSI, AFI,
UTI, or pneumonia (Table 2).
SIRS criteria were fulfilled by 21.9% (36/164) children on
the day of admission. Overall, 18.9% (18/95) patients of DCLD and
26% (18/69) patients of ALF had SIRS. Among patients having
SIRS, 31 of 36 (86%) had coexisting infection (17/18 [94.4%] of
patients with DCLD and 14/18 [78%] of ALF cases). Presence of
SIRS was significantly more common in patients with infection
compared with those without infection (31/77 vs. 5/87 P ¼ 0.00).
PCT and CRP were found to correlate with the severity of
sepsis. Patients with severe sepsis (n ¼ 18) had a higher PCT and
CRP (Fig. 1) compared with patients with infections without organ
dysfunction (n ¼ 59).
A higher PCT (2.9 [0.1–38] vs 1.3 [0.13–23.6]; P ¼ 1.00)
and CRP (2.8 [0.3–24] mg/dL vs 1.7 [0.32–13.8] mg/dL; P ¼ 0.03)
were also observed in patients with infection at multiple sites
compared with those with single-site infection. PCT levels though
higher did not attain statistical significance.
The ROC curves for PCT and CRP for all patients and
separately for patients with ALF or DCLD are shown in
Figure 2. PCT had a higher area under the curve compared with
CRP for all patients and also for ALF and DCLD cases separately
(Fig. 2). The AUCs (Fig. 2) for PCT and CRP obtained for patients
with ALF were lower than that of DCLD patients.
Among the cases with DCLD (n ¼ 95, Child-Pugh score: A
16, B 38, C 41), we also evaluated the performance of PCT and CRP
in relation to the severity of the underlying DCLD applying a
subgroup analysis on the basis of the Child-Pugh score class. The
AUCs in DCLD cases for CRP were 0.79 (0.54–0.99), 0.81 (0.61–1),
and 0.75 (0.63–0.99) in Child A, B, and C, respectively. The AUCs
for PCT were 0.83 (0.74–0.95), 0.82 (0.63–1), and 0.79 (0.61–0.97)
in Child A, B, and C, respectively. We found that AUCs continued to
remain optimal even in advanced liver disease.
The optimal PCT value with the best sensitivity and specificity
was >0.5 ng/mL when patients with ALF or DCLD were analyzed
separately. PCT >0.5 ng/mL, however, had higher sensitivity (86.8%
vs 83.3%) and specificity (91.7% vs 57.8%) in patients with DCLD
compare with those with ALF. The optimal CRP value that gave the
best sensitivity and specificity was >0.6 mg/dL for patients with
DCLD (sensitivity 86.8%, specificity 73.8%) and >0.49 mg/dL for
patients with ALF (sensitivity 75%, specificity 67%).
As a single marker, PCT had the best ability to confirm
infected status among patients with DCLD with a positive like-
lihood ratio (LRþ) of 10.4 (3.5–23.28), compared with 3.31 (1.8–
4.86) for CRP and 1.52 (0.86–2.67) for TLC. We also looked at a
combination of these tests (Table 3) and found that a combination
of CRP and PCT (LRþ of 9.54 [3.2–21.3]) or CRP, PCT and TLC
[LRþ 6.08 (1.9–18.6)], was lower than PCT alone. In patients with
ALF, PCT and CRP had a similar predictive value and combining
the tests did not add much to the diagnostic accuracy (Table 3).
DISCUSSION
In the present prospective study of 164 liver disease
children, 46.9% had an infection. PCT, CRP, and TLC values were

TABLE 2. Relation of various parameters of infection with site of infection

Parameter AFI (n ¼ 35) BSI (n ¼ 16) Pneumonia (n ¼ 22) UTI (n ¼ 26)

PCT (>0.5 ng/mL) 30 (85%) 14 (87.5%) 20 (91%) 21 (81%)

CRP (>0.6 mg/dL) 32 (91%) 13 (81%) 18 (82%) 18 (70%)
SIRS 8 (23%) 8 (50%) 7 (32%) 9 (35%)

No statistical difference between proportions of patients with raised PCT and CRP or presence of SIRS between different infection sites.
AFI ¼ ascitic fluid infection; BSI ¼ blood stream infection; CRP ¼ C-reactive protein; PCT ¼ procalcitonin; SIRS ¼ systemic inflammatory response
syndrome; UTI ¼ urinary tract infection.

408 www.jpgn.org

Copyright © ESPGHAN and NASPGHAN. All rights reserved.

JPGN  Volume 63, Number 4, October 2016 Role of Procalcitonin and C-Reactive Protein

15.00

CRP
PCT

20.00

10.00
ng/mL
mg/dL
10.00
5.00

0.00 0.00

Severe sepsis Infection No infection Severe sepsis Infection No Infection

(n = 18) (n = 59) (n = 87) (n = 18) (n = 59) (n = 87)
8.3 (3.5 – 38) 0.89 (0.1 – 8) 0.3 (0.1 – 6.75) 4.1 (0.3 – 13.8) 1.7 (0.32 – 24) 0.3 (0.1 – 4.1)

FIGURE 1. Relation of PCT and CRP with severity of infection in patients with liver disease. CRP ¼ C-reactive protein; PCT ¼ procalcitonin.

1.0 1.0 1.0 Source of the

curve
CRP
0.8 0.8 0.8
PCT
Reference line
Sensitivity

Sensitivity

Sensitivity
0.6 0.6 0.6

0.4 0.4 0.4

0.2 0.2 0.2

0.0 0.0 0.0

0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
1-Specificity 1-Specificity 1-Specificity
Overall DCLD ALF
CRP–AUC = 0.79 (0.73–0.87) CRP–AUC = 0.83 (0.75–0.92) CRP–AUC = 0.69 (0.56–0.82)
PCT–AUC = 0.82 (0.76–0.88) PCT–AUC = 0.90 (0.83–0.97) PCT–AUC = 0.73 (0.60–0.86)

FIGURE 2. ROCs of PCT and CRP for all patients and ALF and DCLD separately. ALF ¼ acute liver failure; CRP ¼ C-reactive protein; DCLD ¼
decompensated chronic liver disease; PCT ¼ procalcitonin; ROC ¼ receiver operating curve.

TABLE 3. Sensitivity, specificity, and likelihood ratio of CRP, PCT, and TLC in identifying presence of infection in children with liver disease

Sensitivity, % Specificity, % LR-positive LR-negative

Decompensated chronic liver disease

CRP (>0.6 mg/dL) 86.8 73.8 3.31 (1.80–4.86) 0.17 (0.11–0.41)
PCT (>0.5 ng/mL) 86.8 91.7 10.40 (3.50–23.28) 0.15 (0.07–0.29)
TLC 43.4 71.4 1.52 (0.86–2.67) 0.79 (0.58–1.07)
PCT þ CRP 79.2 91.7 9.54 (3.20–21.3) 0.23 (0.13–0.39)
PCT þ CRP þ TLC 43.4 92.8 6.08 (1.90–18.60) 0.61 (0.47–0.78)
ALF
CRP (> 0.49 mg/dL) 75.0 67.0 2.27 (1.12–3.14) 0.37 (0.22–0.98)
PCT (> 0.5 ng/mL) 83.3 57.8 1.97 (1.34–2.90) 0.29 (0.11–0.73)
TLC 41.6 73.3 1.56 (0.79–3.08) 0.79 (0.34–0.94)
PCT þ CRP 62.5 75.5 2.56 (1.40–4.60) 0.50 (0.29–0.85)
PCT þ CRP þ TLC 37.5 88.9 3.37 (1.27–8.94) 0.70 (0.50–0.97)

ALF ¼ acute liver failure; CRP ¼ C-reactive protein; LR ¼ likelihood ratio; PCT ¼ procalcitonin; TLC ¼ total leukocyte count (as per age appropriate
cutoff).

www.jpgn.org 409

Copyright © ESPGHAN and NASPGHAN. All rights reserved.

Bolia et al
significantly higher in children with infection. PCT and CRP had
an AUC of 0.82 and 0.79, respectively, in identifying children
with infection.
Both PCT and CRP were better predictors of infection in
children with DCLD (AUC of 0.90 and 0.83, respectively) com-
pared with ALF (AUC of 0.73 and 0.69 respectively) (Fig. 2). The
possible reason for the poor performance of these biomarkers in
ALF compared with DCLD could be attributed to the fact that CRP
is produced predominantly by hepatocytes and PCT is produced
both by thyroidal and extrathyroidal tissues including the liver in
infected patients. As a result of massive destruction of hepatocytes
in ALF, there is a decrease in their production. Our hypothesis is
supported by the study on 50 ALF cases by Izumi et al (14), in which
an attenuated production of acute-phase proteins in response to
infection was shown.
Applying the same rationale, one would assume that even in
children with DCLD, the performance of these biomarkers would be
affected by the severity of the liver disease. In children with Child C
disease, the AUCs, though lower, however, remained good. Similar
results were shown by Bota et al (15) in adult cirrhotic without any
differences between the levels of CRP and PCT with different
Child-Pugh scores. This suggests that the adequate hepatocyte mass
required for the production of these biomarkers is retained in
patients with advanced DCLD.
PCT was a better predictor of infection compared with CRP
in our patients with DCLD (LR þ 10.40 [3.50– 23.28] vs 3.31
[1.80–4.86]) (Table 3), which is similar to 2 other adult studies
(16,17). In ALF, both CRP and PCT were equally good with PCT
being marginally better especially in terms of sensitivity (83.3% vs
75%) and identifying more cases with infection. These markers
have not been compared in patients with ALF previously.
We found that PCT and CRP levels correlated with the
severity of infection and organ dysfunction. The present obser-
vation can be used clinically. Levels of these markers can stratify
patients, and those with higher PCT and CRP levels can be treated
more aggressively. Similar linear relation between PCT and CRP
values and severity of infection has been demonstrated previously
by Hatherhill et al (7) in 175 children admitted to the pediatric
intensive care unit. Higher PCT and CRP were found in patients
with septic shock compared with those with localized bacterial and/
or viral infection and controls (7).
The best cutoff value of a biomarker is the ‘‘value’’ that gives
optimal sensitivity and specificity, as determined by the ROC curve.
In literature, there is considerable heterogeneity and PCT values
ranging from 0.3 to 8 ng/mL have been used to differentiate
between children with and without infection at various sites
(18,19). This could be due to the varied definitions of infection
and a difference in the underlying disease process. We had 2
homogenous groups of ALF and DCLD, and we used objective
criteria to define sepsis. We derived a cutoff of 0.5 ng/mL from the
ROC curve that is the same as that for children without liver disease
and irrespective of being ALF or DCLD. In adults with chronic liver
disease, PCT cutoffs of 0.3 to 0.49 ng/mL have been derived
(15,20).
CRP cutoff of 0.6 mg/dL gave us the best sensitivity of
86.8% and specificity of 73.8% in patients with DCLD. In adults
with cirrhosis, various values ranging from 1.2 to 5.5 mg/dL have
been proposed (17,21,22). Fernandez et al (23) in a recent review
proposed a cutoff of 2 mg/dL for CRP for identifying infection in
adults with cirrhosis. This cutoff had a sensitivity of 80.4% and
specificity of 80.7% (22). At this value, however, we obtained a
sensitivity of only 41.5% and a specificity of 91.6%. The possible
explanation for a lower CRP cutoff value in children with cirrhosis
compared to adults could be that CRP values are age-dependent and
gradually increase with age (24).
410
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
JPGN  Volume 63, Number 4, October 2016
As discussed above, CRP response is attenuated in patients
with ALF and infections (14,25). As a result, we found the CRP
cutoff to be even lower in this group. A cutoff of 0.49 mg/dL had a
sensitivity of 75% and specificity of 67%.
Almost one quarter of patients without bacterial infection
also had mildly elevated levels of PCT and CRP. We feel that this
may be due to the underlying liver dysfunction, inflammation, and
low level endotoxemia caused by portovenous shunting in these
patients. It has been shown that patients with liver disease, even
without infection, have an increased expression of TNF-receptors
on hepatocytes during acute and chronic liver inflammation (26). As
a result of this, these patients may have mildly elevated acute phase
proteins (27).
Another aspect that we have addressed is SIRS and its
correlation with infection that has not been studied previously in
children with liver disease. In ALF, SIRS may occur as a result of
ALF per se (28) making it interesting to study as to what proportion
of these patients have infection-related SIRS. Overall 26% (18/69)
patients with ALF had SIRS at admission and 78% of those had an
infection. Rolando et al (28), among adult patients with ALF,
showed 38.1% (338/887) fulfilled at least 2 SIRS criteria and
58.5% (198/338) of them developed an infection. Both our study
and that by Rolando et al show that infected patients with ALF more
often developed SIRS. The higher prevalence of SIRS in adults
could be due to difference in the study design of assessing SIRS
during the entire course of illness versus ours at admission alone.
In patients with DCLD, there are many confounders in
identifying SIRS. Fever may occur as a result of endotoxemia
due to portosystemic shunting; tachycardia as a result of hyperdy-
namic circulation and tachypnea due to gross ascites giving rise to a
false positive SIRS. Hypersplenism decreases white blood cell
count and patients receiving beta-blockers have a lower heart rate
resulting in false negative SIRS. In our study, 32% (17/53) of
children with DCLD who had an infection fulfilled the SIRS
criteria. We found that children with SIRS had a significant
correlation with infections (P ¼ 0.005). Although SIRS had a poor
sensitivity (27%), it had a high specificity (94%) to identify
infection. Cazzaniga et al (29) found SIRS in 27.6% (39/141) of
adult cirrhotic patients, 14.1% at admission and other 13.4% during
hospital stay. Eighty-two percent (32/39) of patients with SIRS
developed an infection that is similar to our observation of 95% (17/
18). Our results indicate a high risk for sepsis development in
children with DCLD with SIRS. The higher prevalence of SIRS at
admission in our study could be due to the fact that all our patients
had decompensated CLD who are more prone to infections com-
pared with the adult study that included both compensated and
decompensated patients (50%).
To the best of our knowledge, this is the first study evaluating
the role of PCT and CRP in children with ALF and DCLD. We
carried out the study prospectively in a large sample, which
comprised children with varying grades of severity of liver disease,
with and without any symptoms of infection. A potential limitation
of our study is that the biomarkers were estimated only at admission
and not serially during the entire hospital stay.
Our observations have important clinical implications.
Because symptoms do not indicate the presence or absence of
infection in a majority of patients, performing biomarker levels
routinely in all patients with liver disease at the time of hospital
admission and/or deterioration will help to identify patients with
and without infections. Identifying patients with infections will help
in early initiation of therapy and possibly a better outcome, whereas
ruling out of infections will prevent the inadvertent use of anti-
biotics and subsequent development of resistance.
To conclude, we found that PCT and CRP can help diagnose
bacterial infection among children with ALF and DCLD with a
www.jpgn.org
JPGN  Volume 63, Number 4, October 2016
reasonable accuracy. Both the biomarkers correlate with infection
severity, and their diagnostic accuracy is better in DCLD than ALF
cases. SIRS is significantly more common in infected patients. PCT is
a better serologic biomarker and as a single marker, PCT >0.5 ng/mL
may be used to diagnose bacterial infection in children with
liver disease.
REFERENCES
1. Tandon P, Garcia-Tsao G. Bacterial infections, sepsis, and multiorgan
failure in cirrhosis. Semin Liver Dis 2008;28:26–42.
2. Rolando N, Harvey F, Brahm J, et al. Prospective study of bacterial
infection in acute liver failure: an analysis of 50 patients. Hepatology
1990;11:49–53.
3. Jaye DL, Waites KB. Clinical applications of C-reactive protein in
pediatrics. Pediatr Infect Dis J 1997;16:735–46.
4. Kumar K, Mohindra S, Raj M, et al. Procalcitonin as a marker of sepsis
in alcoholic hepatitis. Hepatol int 2014;8:436–42.
5. Assicot M, Gendrel D, Carsin H, et al. High serum procalcitonin
concentrations in patients with sepsis and infection. Lancet 1993;341:
515–8.
6. Proulx F, Fayon M, Farrell CA, et al. Epidemiology of sepsis and
multiple organ dysfunction syndrome in children. Chest 1996;109:
1033–7.
7. Hatherill M, Tibby SM, Sykes K, et al. Diagnostic markers of infection:
comparison of procalcitonin with C reactive protein and leucocyte
count. Arch Dis Child 1999;81:417–21.
8. Squires RH Jr, Shneider BL, Bucuvalas J, et al. Acute liver failure in
children: the first 348 patients in the pediatric acute liver failure study
group. J Pediatr 2006;148:652–8.
9. Pugh RN, Murray-Lyon IM, Dawson JL, et al. Transection of the
oesophagus for bleeding oesophageal varices. Br J Surg 1973;60:
646–9.
10. Meisner M, Tschaikowsky K, Schnabel S, et al. Procalcitonin—influ-
ence of temperature, storage, anticoagulation and arterial or venous
asservation of blood samples on procalcitonin concentrations. Eur J
Clin Chem Clin Biochem 1997;35:597–601.
11. Ahsan Sama, Noether Julia. Hematology. In: Tscurdy M, Arcara K, eds.
The Harriet Lane Handbook: A Manual for Pediatric House Officers.
19th Edn. Philadelphia: Elsevier Mosby; 2012. 14;323.
12. Garner JS, Jarvis WR, Emori TG, et al. CDC definitions for nosocomial
infections, 1988. Am J Infect Control 1988;16:128–40.
13. Goldstein B, Giroir B, Randolph A, et al. International pediatric sepsis
consensus conference: Definitions for sepsis and organ dysfunction in
pediatrics. Pediatr Crit Care Med 2005;6:2–8.
14. Izumi S, Hughes RD, Langley PG, et al. Extent of the acute phase
response in fulminant hepatic failure. Gut 1994;35:982–6.
www.jpgn.org
Copyright © ESPGHAN and NASPGHAN. All rights reserved.
Role of Procalcitonin and C-Reactive Protein
15. Bota DP, Van Nuffelen M, Zakariah AN, et al. Serum levels of C-
reactive protein and procalcitonin in critically ill patients with cirrhosis
of the liver. J Lab Clin Med 2005;146:347–51.
16. Acevedo J, Fernandez J, Filella X, et al. Diagnostic capacity and
prognostic value of procalcitonin in cirrhotic patients with bacterial
infection. Hepatology 2010;52:905A.
17. Pieria G, Agarwal B, Burroughs AK. C-reactive protein and bacterial
infection in cirrhosis. Ann Gastroenterology 2014;27:113–20.
18. Enguix A, Rey C, Concha A, et al. Comparison of procalcitonin with C
reactive protein and serum amyloid for the early diagnosis of bacterial
sepsis in critically ill neonates and children. Intensive Care Med
2001;27:211–5.
19. Fernandez Lopez A, Luaces Cubells C, Garcia Garcia JJ, et al.
Spanish Society of Pediatric Emergencies Procalcitonin in pediatric
emergency departments for the early diagnosis of invasive bacterial
infections in febrile infants: results of a multicenter study and utility
of a rapid qualitative test for this marker. Pediatr Infect Dis J
2003;22:895–903.
20. Schepis F, Bianchini M, Ferreti I, et al. Procalcitonin is the best
diagnostic and prognostic marker of sepsis in decompensated cirrhotic
patients. J Hepatol 2010;52:S85.
21. Tsiakalos A, Karatzaferis A, Ziakas P, et al. Acute-phase proteins as
indicators of bacterial infection in patients with cirrhosis. Liver Int
2009;29:1538–42.
22. Lin ZY, Chuang WL, Dai CY, et al. Clinical application of
serum C-reactive protein measurement in the detection of bacterial
infection in patients with liver cirrhosis. Kaohsiung J Med Sci
2002;18:121–6.
23. Fernandez J, Arroyo V. Bacterial infections in cirrhosis: a growing
problem with significant implications. Clin Liver Dis 2013;2:102–5.
24. Ribeiro MA. Levels of C-reactive protein in serum samples from healthy
children and adults in São Paulo, Brazil. Braz J Med Biol Res
1997;30:1055–9.
25. Silvestre JP, Coelho LM, Póvoa PM. Impact of fulminant hepatic failure
in C-reactive protein? J Crit Care 2010;25:657.e7–e12.
26. Volpes R, van den Oord JJ, De Vos R, et al. Hepatic expression of type A
and type B receptors for tumor necrosis factor. J Hepatol 1992;14:361–
9.
27. Shiota G, Umeki K, Okano J, et al. Hepatocyte growth factor and acute
phase proteins in patients with chronic liver diseases. J Med 1995;26:
295–308.
28. Rolando N, Wade J, Davalos M, et al. The systemic inflammatory
response syndrome in acute liver failure. Hepatology 2000;32:
734–9.
29. Cazzaniga M, Dionigi E, Gobbo G, et al. The systemic inflammatory
response syndrome in cirrhotic patients: Relationship with their in-
hospital outcome. J Hepatol 2009;51:475–82.
411