L

Sa

LAa

b

a

ARAA

KIIT

1

bsim

dahdwHn[

B

1h

Digestive and Liver Disease 44 (2012) 1027– 1031

Contents lists available at SciVerse ScienceDirect

Digestive and Liver Disease

j our nal ho me page: www.elsev ier .com/ locate /d ld

iver, Pancreas and Biliary Tract

erum endotoxin and inflammatory mediators in patients with cirrhosisnd hepatic encephalopathy

okesh Jaina, Barjesh Chander Sharmaa,∗, Praveen Sharmab, Siddharth Srivastavaa,mit Agrawala, Shiv Kumar Sarinb

Department of Gastroenterology, G.B. Pant Hospital, New Delhi, IndiaDepartment of Hepatology, ILBS, New Delhi, India

r t i c l e i n f o

rticle history:eceived 20 January 2012ccepted 5 July 2012vailable online 9 August 2012

eywords:L-6L-18NF-alpha

a b s t r a c t

Background: Recent observations suggest that inflammatory response may be important in the patho-genesis of hepatic encephalopathy. The aim of the study was to measure arterial ammonia, tumournecrosis factor-alpha, Interleukin-6, Interleukin-18, and serum endotoxin levels and their correlationwith different grades of hepatic encephalopathy.Methods: 120 patients with cirrhosis were enrolled: 20 patients each of cirrhosis with grades I, II, IIIand IV hepatic encephalopathy, cirrhosis with and without minimal hepatic encephalopathy and healthycontrols were tested for arterial ammonia, tumour necrosis factor-alpha, Interleukin-6, Interleukin-18and serum endotoxin levels.Results: Median arterial ammonia, tumour necrosis factor-alpha, Interleukin-6, Interleukin-18 and serumendotoxin levels were significantly higher in patient with hepatic encephalopathy and minimal hepatic

encephalopathy as compared to patients without minimal hepatic encephalopathy and healthy con-trols. Arterial ammonia (r = 0.72, p = 0.03), tumour necrosis factor alpha (r = 0.87, p = 0.02), Interleukin-6(r = 0.50, p = 0.05), Interleukin-18 (r = 0.76, p = 0.02) and serum endotoxin (r = 0.91, p = 0.01) correlatedwith higher grades of hepatic encephalopathy.Conclusion: In hepatic encephalopathy arterial ammonia, inflammatory mediators, and serum endotoxinare elevated and correlate with encephalopathy grade.

Gast

© 2012 Editrice

. Introduction

Hepatic encephalopathy (HE) refers to brain dysfunction causedy liver failure and/or portal – systemic shunt that produces apectrum of neuropsychiatric abnormalities. Recent observationsn patients with liver disease suggest that inflammatory response

ay be important in the pathogenesis of HE [1].Induced hyperammonemia resulted in significantly greater

eterioration in psychometric tests in cirrhotic patients who hadn ongoing infection compared with those in whom the infectionad resolved [2]. The presence and severity of HE were indepen-ent of severity of liver disease and ammonia concentration, butere associated with higher levels of markers of inflammation [3].yperammonemia, endotoxin, systemic inflammation and central

euroinflammation play synergistic role in pathophysiology of HE4–8].

∗ Corresponding author at: Department of Gastroenterology, Room 203, Academiclock, G.B. Pant Hospital, New Delhi 110002, India. Fax: +91 11 23219222.

E-mail address: drbcsharma@hotmail.com (B.C. Sharma).

590-8658/$36.00 © 2012 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevierttp://dx.doi.org/10.1016/j.dld.2012.07.002

roenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.

Significant relationship has been found between tumour necro-sis factor (TNF)-alpha and ammonia in patients with cirrhosis andHE and it is suggested that TNF-alpha could be strongly involved inthe pathogenesis of HE in these patients [9].

High circulating levels of TNF-alpha following an intravenousbolus of endotoxin resulted in reduced cerebral blood flow andthis might explain why inflammation causes neuropsychologicaldeterioration and changes in cerebral blood flow [10].

Several studies have shown that serum levels of TNF-alpha aresignificantly elevated in patients with acute and chronic liver dis-eases, where these elevations are independent of the aetiology ofthe underlying disease [11]. Serum levels of TNF-alpha are sig-nificantly higher in patients with cirrhosis than in those withoutcirrhosis, reaching the highest levels in decompensated cirrhosis[12]. It has also been shown that plasma levels of TNF-alpha corre-late with the severity of hepatic encephalopathy (HE) in fulminanthepatic failure [13].

Odeh et al. [9,14] have shown that serum levels of TNF-alpha and

ammonia correlate positively with the severity of HE in patientswith chronic liver disease. Goral et al. [15] reported that sever-ity of HE was closely related to levels of serum TNF-alpha andother cytokines. It has been shown that in patients with cirrhosis,

Ltd. All rights reserved.

1028 L. Jain et al. / Digestive and Liver Disease 44 (2012) 1027– 1031

Table 1Demographic features and baseline characteristics of patients with and without minimal hepatic encephalopathy, and with overt hepatic encephalopathy.

No MHE (n = 20) MHE (n = 20) Overt HE (n = 80) p-Value

Age (years) £ 42 (16–65) 41 (17–68) 41(15–70) 0.138Sex (M:F) 18:12 19:11 51:29 0.122Hb (g/dl) £ 8.7 (8–13.8) 8.6 (8–13.5) 8.2 (8–14) 0.128Platelets × 103 £ 72 (54–102) 59 (54–109) 60 (54–109) 0.112Bilirubin (mg/dl) £ 1.5 (.8–3) 1.3 (.5–2.5) 2 (.8–3.8) 0.144AST (U/L) £ 65 (32–78) 50 (23–73) 58 (23–88) 0.132ALT (U/L) £ 79 (21–89) 50 (24–76) 57 (21–91) 0.165ALP (U/L) £ 155 (124–167) 167 (114–184) 157 (111–184) 0.157Albumin (g/dl) £ 2.5 (2.5–3.4) 2.5 (2.2–3.2) 2.5 (2.3–3.5) 0.233International normalized ratio £ 1.57 (1.12–1.98) 1.55 (1.11–2.11) 1.57 (1–1.98) 0.212S. creatinine (mg/dl) £ .8 (.6–1.2) .8 (.7–1.1) .8 (.6–1.4) 0.224S. sodium (meq/L) £ 139 (135–145) 139 (134–145) 135 (131–149) 0.128CTP A/B/C 13/12/5 14/12/4 0/21/59 0.03MELD score £ 16 (13–24) 20 (13–28) 25 (13–34) 0.224

£ y; MHM

Imiogttstw

2

P(oisuhpcnp

sApi

TB

£

, Continuous variables expressed with median (range); HE, hepatic encephalopathELD, model for end stage liver disease.

nterleukin (IL)-6 and Interleukin-18 levels correlate with grades ofinimal hepatic encephalopathy (MHE) and are useful to discrim-

nate patients with and without MHE [16]. There is paucity of datan serum inflammatory markers and serum endotoxin in differentrades of HE and also the correlation between serum inflamma-ory markers, serum endotoxin and grade of HE. We conductedhe study to evaluate arterial ammonia, TNF-alpha, IL-6, IL-18, anderum endotoxins levels in different grades of hepatic encephalopa-hy and to correlate inflammatory mediators and serum endotoxinith grades of hepatic encephalopathy.

. Methods

Patients with liver cirrhosis were prospectively enrolled at G.B.ant hospital, a tertiary care institute, over a period of one yearJanuary 2011 to December 2011). Exclusion criteria were: historyf gastrointestinal bleed in the previous 6 weeks, active ongoingnfection, creatinine > 1.5 mg/dL, electrolyte impairment (serumodium < 130 meq/L, serum potassium < 3.5 meq/L or >5.0 meq/L),se of psychotropic drugs in the previous 6 weeks, recent alco-ol use (<6 weeks), previous transjugular intrahepatic shunt (TIPS)lacement or shunt surgery, hepatocellular carcinoma, severeomorbidity such as congestive heart failure, pulmonary disease,eurological and psychiatric problems impairing quality of life, andoor vision precluding neuropsychological assessment.

Diagnosis of cirrhosis was based on clinical, biochemical, ultra-

onographic and liver histological findings whenever available.ll patients of cirrhosis without overt HE were screened for theresence of MHE. Diagnosis and grading of HE was done accord-

ng to West Haven criteria. MHE was diagnosed according to

able 2aseline characteristics of patients with grade I–grade IV hepatic encephalopathy.

Grade I HE n = 20 Grade II HE n =

Age (yr) £ 42 (15–70) 41 (17–68)

Sex (M:F) 13:7 12:8

Hb (g/dl) £ 8.6 + (8–14) 8.6 + (8–13.5)

Platelets × 103 £ 69 (55–105) 59 (54–109)

Bilirubin (mg/dl) £ 1.6 (.6–2.5) 1.3 (.5–2.5)

AST (U/L) £ 46 (24–78) 50 (23–73)

ALT (U/L) £ 49 (25–81) 50 (24–76)

ALP (U/L) £ 148 (111–174) 167 (114–184)

S. albumin (g/dl) £ 2.8 (2.4–3.5) 2.5 (2.2–3.2)

International £ normalized ratio 1.44 (1–1.88) 1.55 (1.11–2.11S. creatinine (mg/dl) £ .8 (.6–1) .8 (.7–1.1)

S. sodium (meq/L) £ 134 (131–137) 139 (134–145)

CTP A/B/C 0/8/12 0/6/14

MELD score £ 19 (14–34) 23 (14–32)

, Continuous variables expressed with median (range); HE, hepatic encephalopathy; Hb,

E, minimal hepatic encephalopathy; Hb, haemoglobin; CTP, Child–Turcotte–Pugh;

Psychometric hepatic encephalopathy score (PHES score). PHESevaluation includes digit symbol test (DST), number connectiontest-A (NCT-A), number connection test-B (NCT-B), serial dottingtest (SDT) and line-tracing test (LTT). Normality tables were con-structed after multiple linear regression for each test.

Expected vs. observed results were compared. A score wasderived as the number of standard deviations of difference betweenthe two values for each test. MHE was diagnosed with the sum ofall scores ≤ −4 points [16]. A written informed consent was takenfrom all the patients or patients’ attendant. Clinical and laboratoryevaluation of patients were done to assess for aetiology of liver dis-ease & severity of liver disease (Child–Pugh [CTP] and MELD score).Study patients were divided into five groups according to grades ofencephalopathy and twenty age- and sex-matched healthy controlswere included.

All patients and healthy controls were investigated for com-plete blood count, liver function tests, serum electrolytes, serumcreatinine, blood sugar, viral markers (HbsAg and Anti HCV anti-body), arterial ammonia, serum endotoxin, TNF-alpha, IL-6 andIL-18.

2.1. Blood samples

To obtain plasma, blood (5 mL) was taken in BD Vacutainertubes containing ethylene diamine tetraacetate and centrifugedat 500 × g for 10 min. The supernatant was collected, and stored

at −80 ◦C in aliquots of 500 �L. To obtain serum, blood (5-mL)was taken in BD Vacutainer tubes without ethylene diamine tetraacetate and centrifuged at 500 × g for 10 min. The supernatant wascollected, and stored at −80 ◦C in aliquots of 500 �L.

20 Grade III HE n = 20 Grade IV HE n = 20 p-Value

42 (16–65) 40 (15–69) 0.23814:6 12:8 0.2228.7 + (8–13.8) 8.2 + (8–11.8) 0.12372 (54–102) 70 (64–90) 0.1222.5 (.8–3) 3.5 (.8–3.8) 0.13465 (32–78) 45 (30–88) 0.12279 (21–89) 69 (20–99) 0.145155 (124–167) 165 (122–169) 0.1672.5 (2.5–3.4) 2.3 (2.5–3) 0.333

) 1.57 (1.12–1.98) 1.77 (1.12–2.98) 0.222.8 (.6–1.2) 1.2 (.6–1.4) 0.124139 (135–145) 134 (135–149) 0.1280/5/15 0/2/18 0.0927 (13–31) 29 (13–31) 0.124

haemoglobin; CTP, Child–Turcotte–Pugh; MELD, model for end stage liver disease.

iver Disease 44 (2012) 1027– 1031 1029

2

mwtKTb5

ao

2

cwpvlleSw

3

ydtgIprwa(pitcp(

T5ecwTs

pe(

a(ppa

onia

, in

flam

mat

ory

med

iato

rs

and

seru

m

end

otox

in

in

con

trol

s

and

cirr

hot

ics

wit

h

and

wit

hou

t m

inim

al

hep

atic

ence

ph

alop

ath

y

and

dif

fere

nt

grad

es

of

ence

ph

alop

ath

y.

Con

trol

s

(n

=

20)

No

MH

E(n

=

20)

p

MH

E(n

=

20)

p G

rad

e

I HE

(n

=

20)

p

Gra

de

II

HE

(n

=

20)

p

Gra

de

III

HE

(n

=

20)

p

Gra

de

IVH

E

(n

=

20)

p

mcg

/dl £

30

(15–

28)

52

(42–

62)

0.02

7

0

(62–

87)

0.01

77

(65–

87)

0.05

88

(65–

88)

0.05

98

(67–

106)

0.01

105

(76–

116)

0.01

l £

8

(6–1

8)

15

(11–

24)

0.01

27.6

(21–

36)

0.02

39

(22–

47)

0.04

53

(38–

62)

0.04

64

(56–

72)

0.02

74

(60–

82)

0.02

£5

(3–9

)

9.5

(4.2

–11.

4)

0.01

18

(11–

24)

0.01

24

(20–

32)

0.03

33

(24–

40)

0.03

45

(35–

55)

0.01

50

(45–

60)

0.03

l £8

(4–1

1)

21

(16–

32)

0.01

43

(32–

54)

0.02

65

(50–

72)

0.01

84

(74–

94)

0.01

99

(83–

102)

0.01

122

(92–

132)

0.01

otox

in, E

U/m

l £

.25

(.25

–.30

)

.35

(.25

–.50

)

0.03

.45

(.3–

.55)

0.04

.55

(.4–

.6)

0.01

.62

(.5–

.7)

0.01

.74

(.5–

.8)

0.01

.9

(.8–

1.0)

0.03

s

vari

able

s

exp

ress

ed

wit

h

med

ian

(ran

ge);

TNF

�, t

um

our

nec

rosi

s

fact

or

alp

ha;

IL, i

nte

rleu

kin

;

MH

E,

min

imal

hep

atic

ence

ph

alop

ath

y;

HE,

hep

atic

ence

ph

alop

ath

y.

L. Jain et al. / Digestive and L

.2. Determination of cytokines, endotoxin and arterial ammonia

The Interlukin studied (IL-6, IL-18) and TNF-alpha were deter-ined using BD OptEIATM Human ELISA Kit II. Serum endotoxinas measured using Toxin Sensor TM Chromogenic LAL Endo-

oxin Assay Kit. Arterial ammonia was measured by ammonia Testit II for the Pocket Chem BA device (Arkay, Inc., Kyoto, Japan).he continuous measurement range is 7–286 �mol/L; the normallood ammonia level for healthy adults for this device is less than4 �mol/L.

The investigator performing laboratory tests of IL-6, IL-18, TNF-lpha, serum endotoxin and arterial ammonia was blinded to statusf HE in the patients.

.3. Statistical analysis

Data were expressed in median and range. Median values werealculated for each group. One-way analysis of variance (ANOVA)as used to compare multiple groups followed by a post hoc com-arison by Bonferroni method. To compare unpaired data betweenarious groups, Mann Whitney U test was done. To explore corre-ations, Spearman’s correlation coefficient test was used. Multipleogistic regression analysis was done to examine predictors ofncephalopathy. All data were processed with the software packagePSS Version 15.0 (SPSS Inc., Chicago, IL) and 2-sided P values < 0.05ere considered significant.

. Results

One hundred and twenty cirrhotic patients [median age 41ears (range, 15–70 years)] were enrolled. Study patients wereivided into five groups according to grades of Encephalopa-hy: grade 0 (n = 40) [with MHE n = 20 and without MHE n = 20],rade I (n = 20), grade II (n = 20), grade III (n = 20) and gradeV (n = 20). Demographic features and baseline characteristics ofatients without MHE, with MHE and with overt HE were compa-able, except CTP status, with Child C more common in patientsith overt HE (Table 1). Demographic features and baseline char-

cteristics were similar in patients with grade I to grade IVTable 2). Demographic features and baseline characteristics ofatients without MHE, with MHE and with overt HE are shown

n Table 1. Aetiology of liver cirrhosis was alcohol (40%), hepati-is B virus infection (25%), hepatitis C virus infection (10%) andryptogenic cirrhosis (25%). Serum sodium levels were similar inatients without MHE, with MHE, and grade I to grade IV HETables 1 and 2).

Median arterial ammonia (89 vs. 52 vs. 30 mcg/dl, p = 0.001),NF-alpha (40 vs. 15 vs. 8 pg/ml, p = 0.001), IL-6 (23 vs. 9.5 vs.

pg/ml, p = 0.001), IL-18 (66 vs. 21 vs. 8 pg/ml, p = 0.001) and serumndotoxin levels (0.8 vs. 0.35 vs. 0.25 EU/ml, p = 0.002) were signifi-antly higher in patients with HE and MHE as compared to patientsith no MHE and healthy controls. On post hoc analysis levels of

NF-alpha, IL-6, IL-18, serum endotoxin and arterial ammonia wereignificantly higher with higher grades of HE (Table 3).

Arterial ammonia (r = 0.72, p = 0.03), TNF-alpha (r = 0.87, = 0.02), IL-6 (r = 0.5, p = 0.5), IL-18 (r = .76, p = 0.02) and serumndotoxin (r = 0.91, p = 0.01) correlated with higher grades of HEFig. 1).

On analysis of factors like arterial ammonia, IL-6, IL-18, TNF-lpha and serum endotoxin as predictors of HE, arterial ammonia

OR – 4.6, 95% C.I., 1.2–1.8, p = 0.031), IL-6 (OR – 3.9, 95% C.I., 1.4–1.9,

= 0.047) and IL-18 (OR – 9.6, 95% C.I., 1.1–1.9, p = 0.001) levels wereredictors for development of overt encephalopathy on multivari-te analysis. Ta

ble

3A

rter

ial a

mm

Am

mon

ia,

TNF

�, p

g/m

IL-6

, pg/

ml

IL-1

8,

pg/

mSe

rum

end

£,

Con

tin

uou

1030 L. Jain et al. / Digestive and Liver Disease 44 (2012) 1027– 1031

F l amme

4

eshiAgcArba[sdT

meafhtbt

ig. 1. Correlation between TNF (panel a), IL-6 (panel b), IL-18 (panel c), arteriancephalopathy in cirrhotic patients.

. Discussion

We have shown that levels of TNF-alpha, IL-6, IL-18, serumndotoxins and arterial ammonia are elevated in patient of cirrho-is with HE and MHE as compared to patients with no MHE andealthy controls. The pathogenesis of HE in chronic liver failure

s still incompletely understood, and seems to be multifactorial.mmonia is considered to be an important factor in the patho-enesis of HE [17]. However, the mechanisms by which ammoniaontributes to the manifestations of HE remain poorly defined.strocytes are thought to be the principal target of ammonia neu-otoxicity [18]. Most but not all patients with HE have elevatedlood ammonia levels, and recovery from encephalopathy is oftenccompanied by declining levels of blood ammonia [19]. Odeh20,21], has shown that cytokines, TNF-alpha and endotoxin aretrongly involved in pathogenesis of HE and may explain many oferangements seen in HE. Serum levels of ammonia, cytokines andNF-alpha correlate with severity of HE [15,16].

The exact mechanism by which inflammation and hyperam-onemia modulate the neuropsychological function remains to be

lucidated. IL-6 has been shown to modulate cognitive function,nd a negative association between circulating IL-6 and memoryunction has been found in low dose endotoxemia [22]. TNF-alpha

as been shown to increase the diffusion of ammonia in cen-ral nervous system endothelial cells, to reduce plasma levels ofranch chain amino acid [22] and also to increase the peripheral-ype benzodiazepine receptors in the rat cortex and striatum [23].

onia (panel d), and serum endotoxin (panel e) with grades I–IV (Gr.) of hepatic

All of these effects are thought to contribute to the pathogenesisof HE.

In normal conditions, intestinal bacterial antigens, includinglipopolysaccharides (LPS) derived from gut flora, are absorbed intothe blood and removed by the reticuloendothilial cells of the liver.In liver disease, the pro-inflammatory cytokines are upregulatedsecondary to liver damage or endotoxemia [19]. Raised levels ofIL-6 have been found in cirrhotics with HE vs. those without MHE[16].

Similarly, elevated TNF-alpha levels have been found elevatedin cirrhotics with varying grades of HE in comparison to thosewithout HE. A reduction in serum TNF-alpha value in associa-tion with a decrease in the severity of HE with therapy has beenobserved by Odeh et al. [9] who suggest a direct role of TNF-alphain the pathogenesis of HE. We observed significantly increasedTNF-alpha, IL-6 and IL-18 in HE patients substantiating the roleof pro-inflammatory cytokines in the pathogenesis of HE in cir-rhotics. Pedersen et al. [24], showed that the infusion of ammoniain normal rats did not lead to any increase in either the intracra-nial pressure (ICP) or cerebral blood flow (CBF); the infusion of LPSalone led to a significant increase in ICP, but not in CBF, whereasthe infusion of both LPS and ammonia led to significantly higherincrease in ICP and also in CBF. Thus, it is the combined insult of

ammonia and pro-inflammatory cytokines that leads to cerebraloedema and encephalopathy. TNF-alpha can produce a variety ofmetabolic derangements that could be involved in the pathogenesisof HE due to chronic liver failure. TNF-alpha may increase ammonia

iver D

ct

ia

bpb

airiitah

mnataol

uHsTiceasw

ael

CW

A

i

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

L. Jain et al. / Digestive and L

oncentration in the brain, and consequently ammonia toxicity tohe central nervous system [25].

While blood ammonia levels have been found to be increasedn most patients with HE, reports of correlation between bloodmmonia levels and severity of HE are not consistent [17].

Ong et al. [25] have demonstrated that arterial, as well as, venouslood levels of ammonia correlate strongly with severity of HE inatients with chronic liver failure. We found significant correlationetween arterial ammonia levels and grade of HE.

Liu et al. [26] have shown that alteration of gut flora by thedministration of probiotics and fibre in patients with cirrhosiss associated with improvement in MHE severity, along with aeduction in venous endotoxin and ammonia levels at day 30,n comparison to the control group. Marini and Broussard [27]njected LPS into normal and hyperammonemic rats and foundhat although cytokine production increased similarly in both thenimal groups, the cognitive defects were stronger and longer inyperammonemic rats.

Rodrigo et al. [28], demonstrated that chronic hyperam-onemia is sufficient to induce microglial activation and

euroinflammation; that contributes to the cognitive and motorlterations that occur during hepatic encephalopathy. Treatmentargeted towards the reduction of inflammation by non-steroidal,nti-inflammatory drugs has been shown to prevent the devel-pment of ammonia induced brain oedema and also to restoreearning ability in rats after portocaval anastomosis [7,8,29].

Agusti et al. [30], showed that reducing neuroinflammation bysing inhibitors of p38 would improve the neurological status inE without inducing secondary effects in the kidney. Our study

hows that patients with cirrhosis with HE have significantly higherNF-alpha, IL-6 and IL-18 levels than patients without HE. Theres a significant, positive correlation between pro-inflammatoryytokines (TNF-alpha, IL-6, IL-18), arterial ammonia and serumndotoxin with grades of encephalopathy. This suggests that hyper-mmonemia and inflammation (elevated TNF-alpha, IL-6, IL-18 anderum endotoxin) play a role in the pathogenesis of HE in patientsith cirrhosis.

In conclusion serum endotoxins, proinflammatory markers andrterial ammonia levels are elevated in patients with HE. Serumndotoxin, proinflammatory markers and arterial ammonia corre-ate well with grades of HE.

onflict of interest statemente have no conflict of interest.

cknowledgements

We are thankful to Dr. Nirupama and Jasvinder Singh for assist-ng in interleukins, endotoxin and TNF measurement.

eferences

[1] Balata S, Damink SW, Ferguson K, et al. Induced hyperammonemia alters neu-

ropsychology, brain MR spectroscopy and magnetization transfer in cirrhosis.Hepatology 2003;37:931–9.

[2] Shawcross DL, Davies NA, Williams R, et al. Systemic inflammatory responseexacerbates the neuropsychological effects of induced hyperammonemia incirrhosis. Journal of Hepatology 2004;40:247–54.

[

isease 44 (2012) 1027– 1031 1031

[3] Shawcross DL, Wright G, Olde SW, et al. Role of ammonia and inflammation inminimal hepatic encephalopathy. Metabolic Brain Disease 2007;22:125–38.

[4] Wright G, Jalan R. Ammonia and inflammation in the pathogenesis of hepaticencephalopathy: Pandora’s box. Heoatology 2007;46:291–4.

[5] Nolan JP. The role of intestinal endotoxin in liver injury: a long and evolvinghistory. Hepatology 2010;52:1829–35.

[6] Butterworth RF. Hepatic encephalopathy: a central neuroinflammatory disor-der. Hepatology 2011;53:1372–6.

[7] Cauli O, Rodrigo R, Piedrafita B, et al. Neuroinflammation contributes to hypoki-nesia in rats with hepatic encephalopathy: ibuprofen restores its motor activity.Journal of Neurosciences Research 2009;87:1369–74.

[8] Cauli O, Rodrigo R, Piedrafita B, et al. Inflammation and hepatic encephalopathy:ibuprofen restores learning ability in rats with portacaval shunts. Hepatology2007;46:514–9.

[9] Odeh M, Sabo E, Srugo I, et al. Relationship between TNF-alpha and ammoniain patients with hepatic encephalopathy due to chronic liver failure. Annals ofMedicine 2005;37:603–12.

10] Lee MD, Zentella A, Pekalla PH, et al. Effect of endotoxin induced monokineson glucose metabolism in the muscle cell line L6. Proceedings of the NationalAcademy of Sciences of the United States of America 1987;84:2590–4.

11] Darling G, Goldstein DS, Stull R, et al. Tumor necrosis factor: immune endocrineinteraction. Surgery 1989;106:1155–60.

12] Nagaki M, Iwai H, Naiki T, et al. High levels of serum interleukin-10 and tumornecrosis factor-� are associated with fatality in fulminant hepatitis. Journal ofInfectious Diseases 2000;182:1103–8.

13] Rolando N, Ellis AJ, De Groote D, et al. Correlation of serial cytokine levelswith progression to coma (grade IV) in patients with acute liver failure (ALF).Hepatology 1995;22:366A.

14] Odeh M, Sabo E, Srugo I, et al. Serum levels of tumor necrosis factor-� corre-late with severity of Hepatic encephalopathy due to chronic liver failure. LiverInternational 2004;24:110–6.

15] Goral V, Atayan Y, Kaplan A. The relation between pathogenesis of liver cir-rhosis, hepatic encephalopathy and serum cytokine levels: what is the role oftumor necrosis factor alpha? Hepato-Gastroenterology 2011;58:943–8.

16] Montoliu C, Piedrafita B, Serra MA, et al. IL-6 and IL-18 in blood may discrim-inate cirrhotic patients with and without minimal hepatic encephalopathy.Journal of Clinical Gastroenterology 2009;43:272–9.

17] Butterworth RF. Pathophysiology of hepatic encephalopathy: a new look atammonia. Metabolic Brain Disease 2002;17:221–7.

18] Schenker S, Brady CE. Pathogenesis of hepatic encephalopathy. In: Conn HO,Bircher J, editors. Hepatic encephalopathy: syndromes and therapies. Bloom-ington, IL: Med-Ed Press; 1994. p. 44–61.

19] Wright G, Davies NA, Shawcross DL, et al. Endotoxemia produces coma andbrain swelling in bile duct ligated rats. Hepatology 2007;45:1517–26.

20] Odeh M. Pathogenesis of hepatic encephalopathy: the tumour necrosis factor-�theory. European Journal of Clinical Investigation 2007;37:291–304.

21] Odeh M. Endotoxin and tumor necrosis factor-[alpha] in the pathogenesis ofhepatic encephalopathy. Journal of Clinical Gastroenterology 1994;19:146–53.

22] Krabbe KS, Reichenberg A, Yirmiya R, et al. Low dose endotoxemia andhuman neuropsychological functions. Brain, Behavior, and Immunity 2005;19:453–60.

23] Duchini A, Govindrajan S, Santucci M, et al. Effects of tumor necrosis factor-�and interleukin-6 on fluid phase permeability and ammonia diffusion in CNSderived endothelial cells. Journal of Investigative Medicine 1996;44:474–82.

24] Pedersen HR, Ring-Larsen H, Olsen NV, et al. Hyperammonemia acts synergisti-cally with lipopolysaccharide in inducing changes in cerebral hemodynamics inrats anesthetized with pentobarbital. Journal of Hepatology 2007;47:245–52.

25] Ong JP, Aggarwal A, Krieger D, et al. Correlation between ammonia levelsand the severity of hepatic encephalopathy. American Journal of Medicine2003;114:188–93.

26] Liu Q, Duan ZP, Ha DK, et al. Synbiotic modulation of gut flora: effecton minimal hepatic encephalopathy in patients with cirrhosis. Hepatology2004;39:1441–9.

27] Marini JC, Broussard SR. Hyperammonemia increases sensitivity to LPS. Molec-ular Genetics and Metabolism 2006;88:131–7.

28] Rodrigo R, Cauli O, Gomez-Pinedo U, et al. Hyperammonemia induces neu-roinflammation that contributes to cognitive impairment in rats with hepaticencephalopathy. Gastroenterology 2010;139:675–84.

29] Ahboucha S, Jiang W, Chatauret N, et al. Indomethacin improves loco-

motor deficit and reduces brain concentrations of neuroinhibitory steroidsin rats following portacaval anastomosis. Neurogastroenterology Motility2008;20:949–57.

30] Agusti A, Cauli O, Rodrigo R, et al. p38 MAP kinase is a therapeutic target forhepatic encephalopathy in rats with portacaval shunts. Gut 2011;60:1572–9.