University of Groningen Fatty liver disease Edens, Mireille ...Eur Radiol 2008; 18(4):806-813. 7 MRI-determined fat content of human liver, pancreas and kidney. 149 Sijens PE, Edens

University of Groningen

Fatty liver diseaseEdens, Mireille Angélique

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2011

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Edens, M. A. (2011). Fatty liver disease: pathophysiology & assessment. s.n.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 29-07-2021

https://research.rug.nl/en/publications/fatty-liver-disease(bd24c291-4172-45c0-90ad-a7846c136d90).html

Fatty Liver Disease Pathophysiology

& Assessment

Mireille Edens

Mireille A. Edens

Fatty Liver Disease: Pathophysiology & Assessment ISBN/EAN: 978-90-9025906-2 Layout and design: Mireille Edens Printed by: Gildeprint Drukkerijen, Enschede

Printing of this thesis was financially supported by:

- University of Groningen - Groningen University Institute for Drug Exploration

- TOSHIBA MEDICAL SYSTEMS NEDERLAND

Copyright © M. A. Edens

All rights reserved. No part of this thesis may be reproduced, distributed, or transmitted in

any form or by any means, without permission of the author.

Email: [email protected]

RIJKSUNIVERSITEIT GRONINGEN

Fatty Liver Disease Pathophysiology & Assessment

Proefschrift

ter verkrijging van het doctoraat in de Medische Wetenschappen

aan de Rijksuniversiteit Groningen op gezag van de

Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op

woensdag 19 januari 2011 om 16:15 uur

door

Mireille Angélique Edens

geboren op 15 september 1980 te Beerta

Promotor: Prof. dr. R.P. Stolk

Beoordelingscommissie: Prof. dr. F. Kuipers

Prof. dr. M. Oudkerk

Prof. dr. M.H. Hofker

Paranimfen: Li Qin Simona Budulac

6

CONTENTS

Chapter Page 1 General introduction: Historical perspectives on fatty liver disease. 9

Edens MA. 2 Fatty liver disease and cardiovascular risk in the general population of 25

East Anglia: The Fenland Study.

Edens MA, Forouhi NG, Emanuella de Lucia-Rolfe, Wareham NJ, Stolk RP, and other investigators of the Fenland Study; Authors to be determined. In preparation/ awaiting enlargement of database.

3 Non-alcoholic fatty liver disease is associated with cardiovascular disease 55

risk markers. Edens MA, Kuipers F, Stolk RP. Obes Rev 2009; 10(4):412-419 _ Review Paper

4 Pathogenesis of fatty liver disease: A theory on lipid content, inhibited 75

metabolism, and inflammation.

Edens MA, Groen AK, Stolk RP. Submitted/ under review _ Review Paper

5 Ultrasonography to quantify hepatic fat content: Validation by 109

1H Magnetic Resonance Spectroscopy.

Edens MA, van Ooijen PM, Post WJ, Haagmans MJ, Kristanto W, Sijens PE, van der Jagt EJ, Stolk RP.

Obesity (Silver Spring) 2009; 17(12):2239-2244.

6 Assessment of the variations in fat content in normal liver using a fast 131

MR imaging method in comparison with results obtained by

spectroscopic imaging.

Irwan R, Edens MA, Sijens PE. Eur Radiol 2008; 18(4):806-813.

7 MRI-determined fat content of human liver, pancreas and kidney. 149

Sijens PE, Edens MA, Bakker SJ, Stolk RP. World J Gastroenterol 2010; 16(16):1993-1998.

8 General discussion:

8.1 Evidence on screening for fatty liver disease: Future perspectives. 165

Edens MA, Stolk RP. Submitted/ under review _ Review Paper 8.2 Additional remarks and recommendations for future research. 191

7

Summaries 197 - English summary 198

- Nederlandse samenvatting (Dutch summary) 202

List of abbreviations 203 Dankwoord (Word of gratitude) 207 Curriculum Vitae 211 - About the author 212 - List of publications 213

8

Chapter 1. General introduction

Historical perspectives on fatty liver disease

Mireille A. Edens

Department of Epidemiology

University Medical Center Groningen University of Groningen

Groningen, the Netherlands

Chapter 1.

10

INTRODUCTION

Development of diagnosis modalities and execution of many studies during the last and

present decade, have substantially increased our knowledge on fatty liver disease (FLD).

Epidemiological studies have revealed a worldwide incidence of obesity 1 and FLD 2, 3. The

present high prevalence of FLD may delineate a major public health risk CHAPTER 2. The

aim of this chapter is to give an overview on the historical development of FLD, and the

contribution of this thesis to our understanding of FLD.

DISCOVERY OF LIVER FAT

Excess fat should be stored in adipocytes (subcutaneous fat), where it functions as producer

of several factors 4. In the case of dietary overflow or pathology, lipids can be stored in the

peritoneal cavity (visceral fat), retro-peritoneal (peri-renal fat), or ectopically, i.e. inside

myocytes and organs (e.g. the liver) as well 5. The presence of fat within the human liver

was first reported by Hartley (1907 and 1909) 6, 7 [source 8].

THE HISTOLOGICAL FATTY LIVER DISEASE SPECTRUM

Classification

Cross-sectional and follow-up studies by histology have revealed that FLD is a broad

spectrum, which includes many histological parameters with varying severity.

Parameters within the spectrum

Younossi ea. (1998) reviewed the literature on histologic FLD parameters, and identified

nineteen histological features, which they grouped into four categories: 1) ‘steatosis’, 2)

‘inflammation’, 3) ‘liver cell injury’, and 4) ‘fibrosis’ 9, which has resulted in the four FLD

subtypes published by Matteoni ea. (1999) 10. Observer variability analysis revealed

however that scoring of ‘inflammation’ was unreliable 9. In the same year, Brunt ea. (1999)

published a semi-quantitative scoring protocol for distinguishing non-alcoholic steatosis

(NAS) and non-alcoholic steatohepatitis (NASH), based on 10 predominant histological

features of NAFLD activity 11. Harrison ea. (2003) presented a modification of the Brunt

protocol, i.e. they divided inflammation score into ‘inflammation’ and ‘degeneration and

necrosis’, similar to the liver cell injury group by Younossi ea., but reliability of this

method was not reported 12. Kleiner ea. (2005) published an update of the Brunt protocol,

Historical perspectives

11

based on 14 histological features 13. Similar to Younossi ea. 9, reliability of scoring was

found to be less than perfect 13.

Terminology

FLD is currently the denominator used to indicate all its subtypes, although the “D” has

been under debate 14. Throughout history many synonyms have been used to indicate the

same subtypes. Table І shows a classification of the FLD spectrum, including terminology

used throughout history.

FAT DISTRIBUTION AND CARDIOVASCULAR DISEASE

In 1947, Vague published about a difference between the cardiovascular disease (CVD) risk

of android (based on male sex) and gynoid obesity (based on female sex) 15.

Liver fat and cardiovascular disease risk

A ‘marker’ for CVD risk includes the metabolic syndrome (MetS). The MetS is associated

with an incidence of CVD events, as reported by a recently published meta-analysis 16. A

close association between non-alcoholic FLD (NAFLD) and the MetS was first published

by Huth ea. (1992) 17. In 2001, Davis & Hui published the following paper: “atherosclerosis

is a liver disease of the heart” 18. Currently, NAFLD is considered the hepatic component of

the MetS 19.

Liver fat, insulin resistance and cardiovascular disease risk

The last two decennia provided some evidence that insulin resistance may be a cornerstone

in fat distribution-associated CVD risk. In 1993, Hotamisligil ea. published: “Adipose

expression of tumour necrosis factor-alpha: direct role in obesity-linked insulin resistance”

20. In CHAPTER 3 21 we review the NAFLD-associated overproduction of CVD risk

markers, including the role of tumour necrosis factor-alpha (TNFα) and hepatic insulin

resistance.

Chapter 1.

12

Table І

. C

lass

ific

atio

n o

f th

e fa

tty liv

er d

isea

se s

pec

trum

, in

cludin

g ter

min

olo

gy u

sed thro

ughout his

tory

EN

TIR

E F

LD

SP

EC

TR

UM

Dia

bet

ic liv

er d

isea

se (

1955)

22

Dia

bet

ic f

atty

liv

er (

1971)

23

FL

D (

1999)

10

Fat

ty liv

er (

2003)

14

Ste

atohep

atit

is

Fat

ty h

epat

itis

(1933)

24

Ste

atonec

rosi

s (1

955)

25

Infl

amm

atory

ste

atonec

rosi

s (1

973)

26

Fat

ty liv

er h

epat

itis

(1979)

27

SH

(1980)

28

Dia

bet

ic h

epat

itis

(1985)

29

Type

2 F

LD

(2002)

30

SH

sco

re ≥

5 (

2005)

13

Cir

rhosi

s w

ithout fa

t

and infl

amm

atio

n

Burn

ed-o

ut F

LD

(2001)

31

Bla

nd C

(2004)

32

HC

C (1

990)

33

C, ci

rrhosi

s; F

, fi

bro

sis;

FL

D, fa

tty liv

er d

isea

se; H

CC

, hep

atoce

llula

r ca

rcin

om

a; M

T, M

atte

oni ty

pe;

S, st

eato

sis;

SH

, st

eato

hep

atitis

Fat

wit

h s

pec

ific

infl

amm

atio

n

and f

ibro

sis

or

cirr

hosi

s

Fat

ty C

(1938)

34

Lae

nnec

’s C

, w

ith e

nla

rged

liv

er (

1939)

35

Nutr

itio

nal

C (

1948)

36

MT

4 (

1999)

10

SH

sta

ge

2; w

ith F

(2009)

37

HC

C (

2005)

37,

38

Fat

wit

h s

pec

ific

infl

amm

atio

n

MT

3 (

1999)

10

SH

sta

ge

1; w

ithout F

(2009)

37

Ste

atosi

s

Hyper

trophic

S (

1935)

39

Type

1 F

LD

(2002)

30

SH

sco

re <

5 (

2005)

13

Fat

wit

h n

on-s

pec

ific

infl

amm

atio

n

MT

2 (

1999)

10

Inte

rmed

iate

SH

(2003)

40

Bord

erli

ne

SH

(2005)

13

Fat

only

Sim

ple

S (

1975)

41

Pure

S (

1991)

42

Pure

fat

ty liv

er (

1995)

43,

44

MT

1 (

1999)

10

Bla

nd S

(2007)

45


13

ETIOLOGY/ PATHOGENESIS

Historically, the destructive effect of alcohol on the liver has been well known, which is

deducible by many publications on alcoholic liver disease [source 46]. Therefore alcoholic

FLD (AFLD) and non-alcoholic FLD (NAFLD) are often distinguished, usually using an

ethanol cut-off value of 20 g/d 19, as this value may non-significantly elevate

aminotransferases only 47. However, alcohol is only one of the many etiologic factors of

FLD. In CHAPTER 4 we propose to modify current theories on pathogenesis and

arrangements of risk factors for FLD into the following three categories: 1) risk factors for

hepatic lipid content, 2) risk factors for inhibited hepatic metabolism, and 3) risk factors for

hepatic inflammation.

DIAGNOSIS

Diagnosis modalities for FLD include: biochemical analysis and histological analysis (both

following liver biopsy), and imaging modalities, i.e. magnetic resonance spectroscopy (1H

MRS), magnetic resonance imaging (MRI), computed tomography (CT), and

ultrasonography. Table ІІ shows an overview on these diagnosis modalities.

Historically, 5% liver fat by biochemical analysis (triglycerides in mg 48/ wet liver weight

in g) is considered a threshold for steatosis 49. For 1H MRS, 5.56% based on the 95%

percentile of a low risk group population is considered a cut-off value, which equals 55.6

mg/g 50. Histologically, steatosis is diagnosed when ≥5% of hepatocytes are infiltrated by a

lipid vesicle 13.

Ultrasonography

Ultrasonography has a special place in this thesis. The first publication on ultrasound in its

infancy [source 51] is by Lynn ea. (1942) 52. Several ultrasound scoring measures for diffuse

parenchymal liver disease, including FLD amongst others, have been presented throughout

history. Originally, these measures were based on ‘beam penetration/attenuation’,

‘echogenicity’, ‘vascularity’, and ‘hepatomegaly/live size’ 53. The echogenicity measure

includes hyperechogeneity of liver parenchyma 53 and hyperechogeneity of liver

parenchyma compared to the adjacent structures 53. The vascularity measure includes

decreased vascularity of the liver (blurring or featureless appearance) 53, and increased

vascularity of the liver 53, 54.

Chapter 1.

14

Ultrasonography for assessment fatty liver disease

Ultrasonography is the most often used diagnosis modality for assessment of FLD, and has

been under development for over thirty years. ‘High intensity echoes of a well defined

pattern’ was introduced by Taylor ea. (1976) 55, and later changed to ‘bright liver with

closely packed echoes’ by Joseph ea. (1979) 56, which was then named ‘the bright liver

echo pattern’ 56-58. Thus, the presence of hyperechogeneity with fine, tightly packed echoes

is considered a criterion for steatosis 57, 58. Theoretically, this fine, tightly packed echo

pattern (i.e. increased pixel density) causes the ‘human eye’ to perceive a ‘bright liver’,

even though pixel intensity itself might not be increased 59, 60. Additionally, degree of

steatosis can be assessed by combinations of increased echogenicity as compared to

adjacent structures, loss of echoes of portal vein walls, and posterior beam attenuation 57, 58.

Fat (simple steatosis) is reflected as a fine, tightly packed echo pattern 57, 58. Fat

accompanied by fibrosis is reflected as coarse pin-head echoes within the fine, tightly

packed echo pattern belonging to fat 57, 58. Besides by an irregular outline of the liver

surface 58, 59, cirrhosis is (just like simple steatosis) reflected as bright liver, but portal wall

veins are preserved and posterior beam attenuation is absent 57, 58. Ultrasonography is

usually used as a qualitative method, but in CHAPTER 5 61 we developed and validated a

method to quantify liver fat content by ultrasonography, based on both texture and

attenuation indices, using multi-voxel magnetic resonance spectroscopy as gold standard.


15

Table ІІ. Historical overview on diagnosis modalities, selection of the literature

Time period References Purpose GS

Since ≤1846 [source 62]

Histology

Visual/microscopical scoring by microscope: Brunt ea. (1999) 11, updated by Kleiner ea. (2005) 13 Digital calculation: Fiorini ea. (2004) 63, Franzen ea. (2005) 64, Liquori (2009) 65

Method

Method

NA

NA


Biochemical analysis

Bligh & Dyer (1959) 48 Method NA


Magnetic Resonance

Magnetic Resonance Spectroscopy (since ≤1961 [source 68])

Thomsen (1994) 69 Szczepaniak (1999) 70

Validation Validation

H BA

Sijens ea. (2006) 71; multi-voxel Method -

Magnetic Resonance Imaging (since ≤1968 [source 72])

Dixon (1984) 73 Irwan ea. (2008), i.e. CHAPTER 6 74

Method Validation

- 1H MRS


Ultrasonography

Selection of studies is shown in table ΙΙΙ


Computed tomography

Phelps ea. (1975) 76 Method -

Mendler ea. (1998); single-energy and dual-energy 77 Validation HMA

Davidson ea. (2006) 78 Method -

Duman ea. (2006) 79 Validation H

BA, biochemical analysis; GS, gold standard; H, histological analysis (qualitative); HMA, histological morphometric analysis (number of hepatocytes infiltrated by lipid vesicles, divided by the total number of hepatocytes); 1H MRS, magnetic resonance spectroscopy; NA, not applicable (gold standard itself).

Table ІІІ. V

alid

atio

n s

tudie

s on u

ltra

sound-d

eter

min

ed f

atty

liv

er d

isea

se, se

lect

ion o

f th

e lite

ratu

re

Pop.

ExCr

GS

Liver state

by the GS (n)

I M

USG m

easures

Alloc.

USG

Validity

qualitative

Validity

quantitative

H

E

L

H

E

L

A

F

T

P

E

D

P

B

A

D

E

W

P

I E

W

P

SM

V

SE

SP

r

Sav

erym

utt

u (

1986)

57

SL

D

H

(ql)

A

bn. (6

7),

Nor.

(18)

Fat

(48),

not Fat

(37)

S

Y&

Y

Y&

Y

Y

Im

pr.

O

rd

87%

94%

89%

84%

Nee

dle

men

(1986)

54

SL

D

H

(ql)

A

bn. (8

8),

Nor.

(22)

FFp (

75),

no F

Fp (

35)

S

Y

Y

Y

Y

Y

Y

Y

Y

Y

Impr.

O

rd

89%

92%

86%

-

Jose

ph (

1991)

58

SL

D

H

(ql)

Fat

(36),

not Fat

(14)

S

Y&

Y

Y&

Y

Y

Im

pr.

O

rd

92%

93%

Gra

if (

2000)

80

SL

D

H

(ql)

G

ener

al (

-)

Ste

atosi

s (-

) S

Y

Y

Y

Y

Impr.

O

rd

82%

100%

$

66%

60%

$

Ham

aguch

i (2

007)

81

SL

D

A, L

D

H; M

Ts (q

l)

FL

D (

64),

Nor.

(30)

I Y

&

Y

Y&

Y

Y

C

um

O

rd

91.7

%

100%

r=

.87, p<

.001

#, obes

e st

udy p

opula

tion;

$, sc

ore

by the

bes

t re

vie

wer

.

USG m

easures:

&, co

uple

d m

easu

res;

HE

L, hyper

echogen

eity

of

liver

par

ench

ym

a; H

EL

A, hyper

echogen

eity

of

liver

par

ench

ym

a co

mpar

ed to the

adja

cent st

ruct

ure

s;

FT

PE

, tightly p

acked

ech

oes

; D

PB

A, ra

te o

f dee

p p

ost

erio

r bea

m a

tten

uat

ion; D

EW

P, dec

reas

ed e

chogen

eity

of

the

wal

ls o

f port

al v

eins;

IE

WP, in

crea

sed e

chogen

eity

of

wal

ls o

f port

al v

eins.

Abbreviations:

A,

alco

holics

; A

bn., a

bnorm

al l

iver

; A

UC

, ar

ea u

nder

the

curv

e; C

um

, cu

mula

tive;

FFp,

fatty-f

ibro

tic

pat

tern

; G

S,

gold

sta

ndar

d;

H,

his

tolo

gy;

1H

MR

S,

mag

net

ic r

esonan

ce s

pec

trosc

opy;

I, i

mag

e(s)

; IM

, im

agin

g m

ethod;

Impr.

, over

all

impre

ssio

n o

f th

e liver

; L

D,

liver

dis

ease

oth

er t

han

FL

D;

MT

, M

atte

oni

types

; N

or.

, norm

al liv

er; O

rd, ord

inal

var

iable

; Pop.: s

tudy p

opula

tion; ql, q

ual

itat

ive;

qt, q

uan

tita

tive;

r, co

rrel

atio

n c

oef

fici

ent; S

, sc

ans;

SE

, se

nsi

tivity; SL

D, su

spec

ted liv

er d

isea

se;

SM

, sc

ori

ng m

ethod; SP, sp

ecif

icity; U

SG

, ultra

sonogra

phy; V

, var

iable

type;

Y, yes

.

16

Chapter 1.


17

FAT IN THE LIVER COMPARED TO OTHER TISSUES

In CHAPTER 6 74 we modified and validated a MRI method, using multi-voxel 1H MRS.

This MRI method enabled us to simultaneously compare hepatic fat content with fat

contents of other tissues in CHAPTER 7 82.

SHOULD WE SCREEN FOR FATTY LIVER DISEASE?

In the general discussion of this thesis, i.e. CHAPTER 8.1, we discuss if there is enough

evidence on screening for FLD.

Finally, in CHAPTER 8.2 some additional remarks and recommendations for future research

are given.

Chapter 1.

18

Reference List (1) Friedman JM. Obesity in the new millennium. Nature 2000 April

6;404(6778):632-4.

(2) Kojima S, Watanabe N, Numata M, Ogawa T, Matsuzaki S. Increase in the prevalence of fatty liver in Japan over the past 12 years: analysis of clinical background. J Gastroenterol 2003;38(10):954-61.

(3) Imamura Y, Uto H, Oketani M, Hiramine Y, Hosoyamada K, Sho Y, Hiwaki T, Baba Y, Tahara K, Kubozono O, Kusano K, Tsubouchi H. Association between changes in body composition and the increasing prevalence of fatty liver in Japanese men. Hepatol Res 2008 November;38(11):1083-6.

(4) Avram MM, Avram AS, James WD. Subcutaneous fat in normal and diseased states: 1. Introduction. J Am Acad Dermatol 2005 October;53(4):663-70.

(5) Snijder MB, van Dam RM, Visser M, Seidell JC. What aspects of body fat are particularly hazardous and how do we measure them? Int J Epidemiol 2006 February;35(1):83-92.

(6) Hartley P. On the nature of the fat contained in the liver, kidney, and heart. J Physiol 1907 August 27;36(1):17-26.

(7) Hartley P. On the nature of the fat contained in the liver, kidney and heart: Part II. J Physiol 1909 June 15;38(5):353-74.

(8) www.ncbi.nlm.nih.gov/pubmed. Search term: ("liver"[MeSH Terms] OR "liver"[All Fields]) AND "fat"[All Fields]. PubMed database, accessed 13-07- 2010 July 13.

(9) Younossi ZM, Gramlich T, Liu YC, Matteoni C, Petrelli M, Goldblum J, Rybicki L, McCullough AJ. Nonalcoholic fatty liver disease: assessment of variability in pathologic interpretations. Mod Pathol 1998 June;11(6):560-5.

(10) Matteoni CA, Younossi ZM, Gramlich T, Boparai N, Liu YC, McCullough AJ. Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology 1999 June;116(6):1413-9.

(11) Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander-Tetri BA, Bacon BR. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999 September;94(9):2467-74.

(12) Harrison SA, Torgerson S, Hayashi PH. The natural history of nonalcoholic fatty liver disease: a clinical histopathological study. Am J Gastroenterol 2003 September;98(9):2042-7.


19

(13) Kleiner DE, Brunt EM, Van NM, Behling C, Contos MJ, Cummings OW, Ferrell LD, Liu YC, Torbenson MS, Unalp-Arida A, Yeh M, McCullough AJ, Sanyal AJ. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology 2005 June;41(6):1313-21.

(14) Neuschwander-Tetri BA, Caldwell SH. Nonalcoholic steatohepatitis: summary of an AASLD Single Topic Conference. Hepatology 2003 May;37(5):1202-19.

(15) Vague J. La differénciation sexuelle, facteur determinant des formes de l'obésité. Presse Med 1947 May 24;55(30):339-40.

(16) Gami AS, Witt BJ, Howard DE, Erwin PJ, Gami LA, Somers VK, Montori VM. Metabolic syndrome and risk of incident cardiovascular events and death: a systematic review and meta-analysis of longitudinal studies. J Am Coll Cardiol 2007 January 30;49(4):403-14.

(17) Huth K, Burkard M, Goebel T. [Dyslipoproteinemia and diabetes mellitus in a metabolic syndrome]. Fortschr Med 1992 April 20;110(11):200-4.

(18) Davis RA, Hui TY. 2000 George Lyman Duff Memorial Lecture: atherosclerosis is a liver disease of the heart. Arterioscler Thromb Vasc Biol 2001 June;21(6):887-98.

(19) Kotronen A, Yki-Jarvinen H. Fatty liver: a novel component of the metabolic syndrome. Arterioscler Thromb Vasc Biol 2008 January;28(1):27-38.

(20) Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis factor-alpha: direct role in obesity-linked insulin resistance. Science 1993 January 1;259(5091):87-91.

(21) Edens MA, Kuipers F, Stolk RP. Non-alcoholic fatty liver disease is associated with cardiovascular disease risk markers. Obes Rev 2009 July;10(4):412-9.

(22) Jorke D. [The reaction of the pyruvic acid blood level after a galactose test in diabetic liver disease.]. Z Gesamte Inn Med 1955 May 1;10(9):440-3.

(23) Jorke D, Unbehaun G. [Studies on the differentiation of diabetic and hepatogenic fatty liver]. Z Gesamte Inn Med 1971 November 1;26(21):Suppl-91.

(24) Baumel J, Guibert H. [Cavernous hemangioma of the liver with fatty hepatitis revealed by puncture biopsy.]. Arch Mal Appar Dig Mal Nutr 1951 April;40(4):407-11.

(25) Butturini U, Mancini AM, Barnchelli A, Casa G. [Metabolic functions of tocopherols (alpha & gamma) and comparison with methylene blue in experimental steatonecrosis induced by carbon tetrachloride.]. G Clin Med 1955 August;36(8):1005-90.

Chapter 1.

20

(26) de Vos M, Roels H, Demeulenaere L. [Inflammatory steatonecrosis of the alcoholic liver. Personal experience and review of the literature (author's transl)]. Acta Gastroenterol Belg 1973 April;36(4):239-48.

(27) Adler M, Schaffner F. Fatty liver hepatitis and cirrhosis in obese patients. Am J

Med 1979 November;67(5):811-6.

(28) Ludwig J, Viggiano TR, McGill DB, Oh BJ. Nonalcoholic steatohepatitis: Mayo Clinic experiences with a hitherto unnamed disease. Mayo Clin Proc 1980 July;55(7):434-8.

(29) Batman PA, Scheuer PJ. Diabetic hepatitis preceding the onset of glucose intolerance. Histopathology 1985 February;9(2):237-43.

(30) Angulo P. Nonalcoholic fatty liver disease. N Engl J Med 2002 April 18;346(16):1221-31.

(31) Ong J, Younossi ZM, Reddy V, Price LL, Gramlich T, Mayes J, Boparai N. Cryptogenic cirrhosis and posttransplantation nonalcoholic fatty liver disease. Liver Transpl 2001 September;7(9):797-801.

(32) Caldwell SH, Crespo DM. The spectrum expanded: cryptogenic cirrhosis and the natural history of non-alcoholic fatty liver disease. J Hepatol 2004 April;40(4):578-84.

(33) Powell EE, Cooksley WG, Hanson R, Searle J, Halliday JW, Powell LW. The natural history of nonalcoholic steatohepatitis: a follow-up study of forty-two patients for up to 21 years. Hepatology 1990 January;11(1):74-80.

(34) Connor CL. Fatty infiltration of the liver and the development of cirrhosis in diabetes and chronic alcoholism. Am J Pathol 1938 May;14(3):347-64.

(35) Debre R. Familial Hepatitis and Chronic Jaundice: (Section for the Study of Disease in Children). Proc R Soc Med 1939 July;32(9):1173-84.

(36) Shorr E, Zweifach BW. Hepato-renal factors in circulatory homeostasis; VEM and VDM mechanisms in nutritional cirrhosis in rats. Fed Proc 1948 March;7(1 Pt):115.

(37) Chagas AL, Kikuchi LO, Oliveira CP, Vezozzo DC, Mello ES, Oliveira AC, Cella LC, Herman P, Bachella T, Caldwell SH, Alves VA, Carrilho FJ. Does hepatocellular carcinoma in non-alcoholic steatohepatitis exist in cirrhotic and non-cirrhotic patients? Braz J Med Biol Res 2009 October;42(10):958-62.

(38) Cuadrado A, Orive A, Garcia-Suarez C, Dominguez A, Fernandez-Escalante JC, Crespo J, Pons-Romero F. Non-alcoholic steatohepatitis (NASH) and hepatocellular carcinoma. Obes Surg 2005 March;15(3):442-6.


21

(39) Ellis RW. Hepatomegaly in an Infant. (? Hepatomegalia Glycogenica. ? Hypertrophic Steatosis). Proc R Soc Med 1935 May;28(7):833-6.

(40) Neuschwander-Tetri BA, Brunt EM, Wehmeier KR, Oliver D, Bacon BR. Improved nonalcoholic steatohepatitis after 48 weeks of treatment with the PPAR-gamma ligand rosiglitazone. Hepatology 2003 October;38(4):1008-17.

(41) Pagis de Micco C, Chapoy P, Perrimond H, Lebreuil G. [Reye's syndrome: pathological study of 2 cases]. Sem Hop 1975 March 14;51(13):891-9.

(42) Nussbaum MS, Fischer JE. Pathogenesis of hepatic steatosis during total parenteral nutrition. Surg Annu 1991;23 Pt 2:1-11.

(43) Teli MR, James OF, Burt AD, Bennett MK, Day CP. The natural history of nonalcoholic fatty liver: a follow-up study. Hepatology 1995 December;22(6):1714-9.

(44) Teli MR, Day CP, Burt AD, Bennett MK, James OF. Determinants of progression to cirrhosis or fibrosis in pure alcoholic fatty liver. Lancet 1995 October 14;346(8981):987-90.

(45) Saluena I, Ortega L, Devesa MJ, Lopez-Alonso G, Taxonera C, Diaz-Rubio M, Ladero JM. [Utility of liver biopsy in the etiologic diagnosis of biochemical liver abnormalities of unknown cause]. Gastroenterol Hepatol 2007 June;30(6):325-30.

(46) www.ncbi.nlm.nih.gov/pubmed. Search term: ("ethanol"[MeSH Terms] OR "ethanol"[All Fields] OR "alcohol"[All Fields] OR "alcohols"[MeSH Terms] OR "alcohols"[All Fields]) AND ("liver"[MeSH Terms] OR "liver"[All Fields]). PubMed database, accessed 13-07 2010 July 13.

(47) Ruhl CE, Everhart JE. Joint effects of body weight and alcohol on elevated serum alanine aminotransferase in the United States population. Clin Gastroenterol

Hepatol 2005 December;3(12):1260-8.

(48) Bligh EG, Dyer WJ. A rapid method of total lipid extraction and purification 9. Can J Biochem Physiol 1959 August;37(8):911-7.

(49) Cairns SR, Peters TJ. Biochemical analysis of hepatic lipid in alcoholic and diabetic and control subjects. Clin Sci (Lond) 1983 December;65(6):645-52.

(50) Szczepaniak LS, Nurenberg P, Leonard D, Browning JD, Reingold JS, Grundy S, Hobbs HH, Dobbins RL. Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population. Am

J Physiol Endocrinol Metab 2005 February;288(2):E462-E468.

(51) www.ncbi.nlm.nih.gov/pubmed. Search term: ("ultrasonography"[Subheading] OR "ultrasonography"[All Fields] OR "ultrasound"[All Fields] OR

Chapter 1.

22

"ultrasonography"[MeSH Terms] OR "ultrasound"[All Fields] OR "ultrasonics"[MeSH Terms] OR "ultrasonics"[All Fields]) AND ("liver"[MeSH Terms] OR "liver"[All Fields]). PubMed database, accessed 26-07 2010.

(52) Lynn JG, Zwemer RL, Chick AJ, Miller AE. A new method for the generation and use of focused ultrasound in experimental biology. J Gen Physiol 1942 November 20;26(2):179-93.

(53) Gosink BB, Lemon SK, Scheible W, Leopold GR. Accuracy of ultrasonography in diagnosis of hepatocellular disease. AJR Am J Roentgenol 1979 July;133(1):19-23.

(54) Needleman L, Kurtz AB, Rifkin MD, Cooper HS, Pasto ME, Goldberg BB. Sonography of diffuse benign liver disease: accuracy of pattern recognition and grading. AJR Am J Roentgenol 1986 May;146(5):1011-5.

(55) Taylor KJ, Carpenter DA, Hill CR, McCready VR. Gray scale ultrasound imaging. The anatomy and pathology of the liver. Radiology 1976 May;119(2):415-23.

(56) Joseph AE, Dewbury KC, McGuire PG. Ultrasound in the detection of chronic liver disease (the "bright liver"). Br J Radiol 1979 March;52(615):184-8.

(57) Saverymuttu SH, Joseph AE, Maxwell JD. Ultrasound scanning in the detection of hepatic fibrosis and steatosis. Br Med J (Clin Res Ed) 1986 January 4;292(6512):13-5.

(58) Joseph AE, Saverymuttu SH, al Sam S, Cook MG, Maxwell JD. Comparison of liver histology with ultrasonography in assessing diffuse parenchymal liver disease. Clin Radiol 1991 January;43(1):26-31.

(59) Joseph AE, Saverymuttu SH. Ultrasound in the assessment of diffuse parenchymal liver disease. Clin Radiol 1991 October;44(4):219-21.

(60) Smith-Levitin M, Blickstein I, Albrecht-Shach AA, Goldman RD, Gurewitsch E, Streltzoff J, Chervenak FA. Quantitative assessment of gray-level perception: observers' accuracy is dependent on density differences. Ultrasound Obstet

Gynecol 1997 November;10(5):346-9.

(61) Edens MA, van Ooijen PM, Post WJ, Haagmans MJ, Kristanto W, Sijens PE, van der Jagt EJ, Stolk RP. Ultrasonography to quantify hepatic fat content: validation by 1H magnetic resonance spectroscopy. Obesity (Silver Spring) 2009 December;17(12):2239-44.

(62) www.ncbi.nlm.nih.gov/pubmed. Search term: "Microscope"[Journal] OR "microscope"[All Fields]. PubMed database, accessed 02-09- 2010.

(63) Fiorini RN, Kirtz J, Periyasamy B, Evans Z, Haines JK, Cheng G, Polito C, Rodwell D, Shafizadeh SF, Zhou X, Campbell C, Birsner J, Schmidt M, Lewin D,


23

Chavin KD. Development of an unbiased method for the estimation of liver steatosis. Clin Transplant 2004 December;18(6):700-6.

(64) Franzen LE, Ekstedt M, Kechagias S, Bodin L. Semiquantitative evaluation overestimates the degree of steatosis in liver biopsies: a comparison to stereological point counting. Mod Pathol 2005 July;18(7):912-6.

(65) Liquori GE, Calamita G, Cascella D, Mastrodonato M, Portincasa P, Ferri D. An innovative methodology for the automated morphometric and quantitative estimation of liver steatosis. Histol Histopathol 2009 January;24(1):49-60.

(66) www.ncbi.nlm.nih.gov/pubmed. Search term: "chemistry"[Subheading] OR "chemistry"[All Fields] OR "chemistry"[MeSH Terms]. PubMed database,

accessed 13-08- 2010.

(67) www.ncbi.nlm.nih.gov/pubmed. Search term: "magnetic resonance spectroscopy"[MeSH Terms] OR ("magnetic"[All Fields] AND "resonance"[All Fields] AND "spectroscopy"[All Fields]) OR "magnetic resonance spectroscopy"[All Fields] OR ("magnetic"[All Fields] AND "resonance"[All Fields]) OR "magnetic resonance"[All Fields]. PubMed database, accessed 13-08- 2010.

(68) www.ncbi.nlm.nih.gov/pubmed. Search term: "magnetic resonance spectroscopy"[MeSH Terms] OR ("magnetic"[All Fields] AND "resonance"[All Fields] AND "spectroscopy"[All Fields]) OR "magnetic resonance spectroscopy"[All Fields]. PubMed database, accessed 13-08 2010.

(69) Thomsen C, Becker U, Winkler K, Christoffersen P, Jensen M, Henriksen O. Quantification of liver fat using magnetic resonance spectroscopy. Magn Reson

Imaging 1994;12(3):487-95.

(70) Szczepaniak LS, Babcock EE, Schick F, Dobbins RL, Garg A, Burns DK, McGarry JD, Stein DT. Measurement of intracellular triglyceride stores by H spectroscopy: validation in vivo. Am J Physiol 1999 May;276(5 Pt 1):E977-E989.

(71) Sijens PE, Smit GP, Borgdorff MA, Kappert P, Oudkerk M. Multiple voxel (1)H MR spectroscopy of phosphorylase-b kinase deficient patients (GSD IXa) showing an accumulation of fat in the liver that resolves with aging. J Hepatol 2006 December;45(6):851-5.

(72) www.ncbi.nlm.nih.gov/pubmed. Search term: "magnetic resonance imaging"[MeSH Terms] OR ("magnetic"[All Fields] AND "resonance"[All Fields] AND "imaging"[All Fields]) OR "magnetic resonance imaging"[All Fields]. PubMed database, accessed 02-09- 2010.

(73) Dixon WT. Simple proton spectroscopic imaging. Radiology 1984 October;153(1):189-94.

Chapter 1.

24

(74) Irwan R, Edens MA, Sijens PE. Assessment of the variations in fat content in normal liver using a fast MR imaging method in comparison with results obtained by spectroscopic imaging. Eur Radiol 2008 April;18(4):806-13.

(75) www.ncbi.nlm.nih.gov/pubmed. Search term: "tomography, x-ray computed"[MeSH Terms] OR ("tomography"[All Fields] AND "x-ray"[All Fields] AND "computed"[All Fields]) OR "x-ray computed tomography"[All Fields] OR ("computed"[All Fields] AND "tomography"[All Fields]) OR "computed tomography"[All Fields]. PubMed database, accessed 02-09- 2010.

(76) Phelps ME, Hoffman EJ, Ter Pogossian MM. Attenuation coefficients of various body tissues, fluids, and lesions at photon energies of 18 to 136 keV. Radiology 1975 December;117(3 Pt 1):573-83.

(77) Mendler MH, Bouillet P, Le Sidaner A, Lavoine E, Labrousse F, Sautereau D, Pillegand B. Dual-energy CT in the diagnosis and quantification of fatty liver: limited clinical value in comparison to ultrasound scan and single-energy CT, with special reference to iron overload. J Hepatol 1998 May;28(5):785-94.

(78) Davidson LE, Kuk JL, Church TS, Ross R. Protocol for measurement of liver fat by computed tomography. J Appl Physiol 2006 March;100(3):864-8.

(79) Duman DG, Celikel C, Tuney D, Imeryuz N, Avsar E, Tozun N. Computed tomography in nonalcoholic fatty liver disease: a useful tool for hepatosteatosis assessment? Dig Dis Sci 2006 February;51(2):346-51.

(80) Graif M, Yanuka M, Baraz M, Blank A, Moshkovitz M, Kessler A, Gilat T, Weiss J, Walach E, Amazeen P, Irving CS. Quantitative estimation of attenuation in ultrasound video images: correlation with histology in diffuse liver disease. Invest

Radiol 2000 May;35(5):319-24.

(81) Hamaguchi M, Kojima T, Itoh Y, Harano Y, Fujii K, Nakajima T, Kato T, Takeda N, Okuda J, Ida K, Kawahito Y, Yoshikawa T, Okanoue T. The severity of ultrasonographic findings in nonalcoholic fatty liver disease reflects the metabolic syndrome and visceral fat accumulation. Am J Gastroenterol 2007 December;102(12):2708-15.

(82) Sijens PE, Edens MA, Bakker SJ, Stolk RP. MRI-determined fat content of human liver, pancreas and kidney. World J Gastroenterol 2010 April 28;16(16):1993-8.

Chapter 2

Fatty liver disease and cardiovascular risk in the

general population of East Anglia: The Fenland Study

In preparation/ Awaiting enlargement of database

Mireille A. Edens 1

Nita G. Forouhi 2

Emanuella de Lucia-Rolfe 2

Nickolas J. Wareham 2

Ronald P. Stolk 1

Other investigators of the Fenland Study 2

Department of Epidemiology 1



MRC Epidemiology Unit 2

Institute of Metabolic Science Addenbrooke’s Hospital

Cambridge, United Kingdom

Chapter 2.

26

ABSTRACT

Objectives Limited population data on fatty liver disease (FLD) in Caucasians have been

published. The main aim of the present study was to determine FLD prevalence and its

association with cardiovascular disease risk in the general population of East Anglia using

the Fenland Study.

Design Population based study, with cross-sectional results.

Setting 3 research centers in Cambridgeshire.

Participants People registered at surgeries in Cambridgeshire aged 30 to 58, excluding

participants with known diabetes mellitus, terminal illness or inability to walk unaided, and

(possible) pregnancy.

Measurements Participants underwent abdominal ultrasonography, anthropometric

measurements, and blood (aminotransferases a.o.) and urine analysis was performed.

Additionally participants filled in several questionnaires.

Main outcome measures Liver fat by ultrasound scored in a cumulative fashion, several

anthropometric measures, metabolic syndrome (ATP ΙΙΙ, IDF, and cumulative Z-score), and

the 10-year Framingham risk score.

Results Liver fat scores were obtained in 762 participants. Overall prevalence of FLD and

non-alcoholic FLD (NAFLD) was 38.5% and 30.5%, respectively. FLD was more prevalent

in men than women (p<0.001). By multiple logistic regression analysis on the presence of

FLD, only BMI (p<0.05, OR 1.3) was significantly associated in men, and BMI (p<0.001,

OR 1.6), waist circumference (p<0.01, OR 1.1), and hip circumference (p<0.01, OR .9)

were significantly associated in women. With increasing liver fat category, the number of

metabolic syndrome components (p<0.001 both ATP III and IDF), the cumulative Z-score

(p<0.001) and the 10-year Framingham risk score (p<0.001) increased as well.

Conclusions This study shows a striking prevalence of FLD in East Anglia, particularly in

men. As FLD is associated with several CVD risk estimates, this striking prevalence may

delineate an increased CVD risk in this population.

The Fenland Study

27

INTRODUCTION

Excessive accumulation of fat within hepatocytes results in a wide progressive spectrum

usually referred to as fatty liver disease (FLD) 1. FLD is characterized by several

histological changes which can roughly be divided into steatosis (fat accumulation) and

steatohepatitis (fat accumulation with inflammation, with or without fibrosis) 2. As

alcoholism is a well know risk factor for FLD, alcoholic FLD (AFLD) and non-alcoholic

FLD (NAFLD) are often discriminated using an ethanol value of 20 g/d (140 g/w) 3. FLD,

at least NAFLD, is associated with an increase in (liver-derived) CVD risk markers 3, 4.

Additionally, FLD increases the risk for cirrhosis 5 and likely hepatocarcinoma as well 6, 7.

In epidemiological studies, FLD is usually estimated by serum markers, i.e. the

aminotransferases alanine aminotransferase (ALT) and/or gamma-glutamyl

aminotransferase (GGT). When an elevation of systemic aminotransferases cannot be

explained by the presence of other liver diseases, FLD is generally suspected. However,

ALT isoforms 8, 9 and GGT isoforms 10, 11 are produced and present in numerous organs,

and could therefore reflect pathology other than liver pathology, e.g. neuromuscular

pathology 12. Therefore, sensitivity and specificity of aminotransferases for diagnosing FLD

are questionable 13-17. In the relatively low ALT range, the entire histologic NAFLD

spectrum, including non-alcoholic steatohepatitis (NASH) with advanced fibrosis, can be

seen 13-15. In steatohepatitis, aminotransferase levels might even be lower compared to

steatosis 16, 17. Of morbidly obese patients with steatohepatitis, 46% present with ALT ≤40

U/L 14. In addition to aminotransferases, data on insulin resistance or diabetes may be

needed in order to improve prediction of NAFLD 15.

Currently, five Western general population-based studies on FLD using imaging modalities

have been performed. These are the Dionysos Nutrition and Liver Study (DNLS) 18, 19, the

Study of Health in Pomerania (SHIP) 20, the Dallas Heart Study (DHS) 21, 22, the Study on

Echinococcus Multilocularis and Internal Diseases in Leutkirch (EMIL) 23, and the Finnish

prevention program for type 2 diabetes (FIN-D2D) 24. These studies used ultrasonography

18-20, 23 and magnetic resonance spectroscopy 21, 22, 24, and revealed a FLD prevalence of

approximately 30%, in the years 1997 – 2002 20-23. No general population based study on

FLD has been performed in the United Kingdom (UK). Objectives of the current study

were to: (i) estimate the prevalence of FLD and NAFLD in a UK population, and (ii)

Chapter 2.

28

investigate the association of FLD and aminotransferases, with metabolic CVD risk

estimates, i.e. the metabolic syndrome and Framingham risk score.

Figure І. Schematic display of the study area, including the three research centres

, Fenland; , East Cambridgeshire. Shown within the boundaries of Cambridgeshire.

Wisbech

Ely

Cambridge

The Fenland Study

29

METHODS

THE STUDY

The Fenland Study is an ongoing population-based cross-sectional study of 10,000 men and

women born within the years 1950 – 1975 and resident in Cambridgeshire, UK (figure І) at

the time of study recruitment. All individuals born within this birth cohort and registered at

family doctor practices in Cambridgeshire were eligible for inclusion. Exclusion criteria

were a known history of diabetes, terminal illness or the inability to walk unaided, and

pregnant women were also excluded. The aim of the study is to examine the interactions

between lifestyle and genetic factors on the risk of obesity and related quantitative

metabolic traits. The study commenced in 2005 and aims to complete by 2014.

STUDY POPULATION

The present study includes Fenland study data, collected in the period 24 January 2005 to

15 August 2008. In this period 3843 people participated in any measurement. Liver scans

were not scored at the same rate as the rest of the data collection, which has caused a major

back log. Particularly this back log, and possibly also some difficulties in scanning obese

people 25, 26, has caused 762 participants with liver fat scores and 3081 participants without

liver fat scores. Percentage of females was 58.2% versus 53.4% (p<0.05), age was 45.7

(±7.0) year versus 45.7 (±7.2) year (p=ns), BMI was 25.4 (17.3–49.4) versus 26.6 (16.9–

63.5) kg/m2 (p<0.001), and alcohol units per week were 5 (0 – 67) versus 4.5 (0–100)

(p=ns), for the population with and without liver fat scores, respectively.

DATA COLLECTION

Data collected during the visit to the research facility

The participants underwent several measurements during a single visit of three hours to a

research facility, after an overnight fast. The research staff received systematic training

prior to measurements.

Anthropometry

Participants were measured in barefoot wearing light clothing. Collected anthropometric

data included weight (to the nearest 200 g, using a calibrated electronic scale; TANITA

model BC-418 MA), height (to the nearest 0.1 cm, using a wall-mounted stadiometer;

SECA 240, Birmingham, UK), and waist circumference and hip circumference (both to the

Chapter 2.

30

nearest 0.1 cm using a non-stretchable fibre-glass insertion tape; D loop tape, Chasmors

Ltd, UK). Waist circumference was measured midway between the lower rib margin and

iliac crest. Hip circumference was measured as the widest level over the major trochanters.

Biochemistry analysis of venous blood

ALT and GGT were divided into quintiles based upon the total population. ALT and GGT

cut-off values were both retrieved from the literature if available 27 and determined in the

present study as well. Optimum cut-off values were established by adding up sensitivity

and specificity. The aminotransferase value corresponding with the highest sum was

considered the overall optimal cut-off value.

Ultrasound imaging

Hepatic ultrasonography was performed using a LOGIQ Book XP ultrasound system with a

3C-RS curved transducer; GE Healthcare Bedford, UK. Four different sweeps (cine-loops)

encompassing the entire liver were recorded: (1) starting at the right lobe in a longitudinal

scan plane a sweep from lateral to medial was made, i.e. showing the liver and long axis of

the right kidney, (2) starting at the dome of the right liver lobe in a transversal/oblique scan

plane a sweep from cranial to caudal was made, i.e. showing the hepatic veins, the portal

and gallbladder, (3) starting at the left lobe in a longitudinal scan plane a sweep from lateral

to medial over the gall bladder area was made, and (4) starting at the left lobe in a

transverse scan plane a sweep from cranial to caudal was made, i.e. showing the hepatic

veins and pancreas. In a small part of participants scans with an acceptable quality could

not be established, and were therefore excluded.

Ultrasound scoring criteria and allocation of scores

The recorded cine-loops were scored retrospectively by two operators, who received both

theoretical and practical training from a senior radiographer. Liver fat scoring criteria were:

(i) increased echo reflectivity of liver parenchyma, (ii) decreased visualization of intra-

hepatic vasculature, (iii) attenuation of ultrasound beam. Each criterion was scored on 4-

point scale (i.e. as 1, 2, 3, or 4) and added, resulting in cumulative liver fat score ranging

from 3 to 12. Liver fat score ≤4 was allocated as normal liver, score 5 – 7 was allocated as

mild fatty liver, 8 – 10 was allocated as moderate fatty liver, and score ≥11 was allocated as

severe fatty liver. Quality of the two operators was ensured by comparison with the senior

radiographer, who independently scored the scans. Correlation coefficients were .97

(p<0.001) and .86 (p<0.001) for operator 1 and 2, respectively.

The Fenland Study

31

Questionnaire

The general questionnaire included demographics (e.g. age, sex, ethnicity 28), medication

use including dosages and treatment indication, and lifestyle (e.g. alcohol consumption).

Age was divided into the following classes: lowest ≤39 years, middle 40–49 years, and

highest ≥50 years. Alcohol consumption was reported in units a week. One unit was

equivalent to ½ a pint of beer, 1 glass of wine, 1 single measure of spirits, and 1 glass of

sherry, and contains approximately 8 grams of ethanol. In order to approach the usual

ethanol threshold of 140 g/w 3, a cut-off value of ≤17 units a week (i.e. 8 grams * 17 units =

136 grams) was used to define non-alcoholism. Due to logistic issues, some data were

missing.

METABOLIC CARDIOVASCULAR RISK ESTIMATES

The association of liver fat content with several estimates of metabolic cardiovascular

disease (CVD) risk was studied. Both the NCEP ATP ΙΙΙ definition 29, and IDF definition 30

of the metabolic syndrome (MetS) were studied. In addition, the standardized sex-specific

cumulative Z-score, and the standardized sex-specific cumulative Z-score by Franks et al. 31

were studied. Finally, the association with the 10-year Framingham risk score 32 was

investigated. Details of these metabolic CVD risk estimates are shown in schematic form

by appendix Ι.

STATISTICAL ANALYSIS

Differences between two groups were compared using the unpaired t-test or Mann-Whitney

U-test. Differences between more groups were tested by one-way ANOVA or Kruskal-

Wallis test. Differences of dichotomous variables between two or more groups were tested

using Chi-square test. Association of variables was estimated by Pearson’s rho (r),

explained variance (R2), and/or receiver operating characteristic curve (ROC-curve) with

area under the curve (AUC). Logistic regression analysis was used to find variables that

significantly predict the presence of FLD. To establish the independent predictors,

univariately significant variables were explored in a multiple logistic regression model.

Statistical analysis was performed using SPSS version 16.0.

Chapter 2.

32

RESULTS

LIVER FAT

Overall prevalence

Figure ІІ shows results on distribution of liver fat score and prevalence of FLD. Median

liver fat score was 4, with a mean of 4.67 (95% confidence interval 4.54 – 4.79). Of all 762

participants, 469 had normal liver (61.5%), 208 had mild fatty liver (27.3%), 85 had

moderate fatty liver (11.2%), and none of the participants had severe fatty liver. Thus, 293

participants had elevated hepatic fat content due to any etiology, corresponding to a FLD

prevalence of 38.5%. Table І shows the characteristics of the study population, stratified for

liver fat category.

Prevalence in non-alcoholics and alcoholics

Of all 459 participants with low alcohol consumption, 319 had normal liver (69.5%), 96

had mild fatty liver (20.9%), and 44 had moderate fatty liver (9.6%). These results

correspond with a NAFLD prevalence of 30.5%. Of all 76 participants with high alcohol

consumption, 40 had normal liver (52.6%), 23 had mild fatty liver (30.3%), and 13 had

moderate fatty liver (17.1%). These results correspond with an AFLD prevalence of 47.4%.

Sex

Figure ІІb shows that liver fat significantly differs between sexes (p<0.001). Of all 242

men, 108 (44.6%) had a fatty liver. Of all 337 women, 75 (22.3%) had a fatty liver.

Because of this large sex difference all further analysis were performed for men and women

separately.

Alcohol

Figure ІІc shows the prevalence’s of AFLD and NAFLD. Of all 169 non-alcoholic men, 29

(46.2%) had a fatty liver. Of all 59 alcoholic men, also 29 (49.2%) had a fatty liver. Of all

290 non-alcoholic women, 62 (21.4%) had a fatty liver. Of all 17 alcoholic women, 7

(41.2%) had a fatty liver.

Age class

With increasing age class, FLD increases as well with p<0.05 for both men and women. In

men, FLD prevalence was 27.5%, 46%, and 53.8% in the lowest, middle, and highest age

class, respectively. In women, FLD prevalence was 13.2%, 20.3%, and 30.5% in the

lowest, middle, and highest age class, respectively.

The Fenland Study

33

Men

Women

a. Distribution of liver fat score ≤4, normal; 5-7, mild; 8-10, moderate; ≥11, severe

b. Prevalence stratified for sex p<0.001, by chi-square test

≤17 units/week

>17 units/week

≤17 units/week

>17 units/week

Men Women

c. Prevalence stratified for alcohol use p=ns for both men nor women, by chi-square test

Figure ІІ. Distribution and prevalence of liver fat by ultrasound, including stratification for

sex and alcohol use.

ns, not significant.

34

Table І. Characteristics of the study population, stratified for liver fat category

Normal Mild Moderate P Total

Women (w/m, w%) (262/134) 66.2 (49/75) 39.5 (26/33) 44.1 p<.001 (337/242) 58.2

Age (year) 44.8 (7.1) 46.6 (6.9) 49.5 (5.3) p<.001 45.7 (7.0)

Ethnicity - White - Mixed - Asian or Asian British - Black or Black British - Other ethnic category

327 2 1 4 2

105 0 2 0 2

49 0 0 0 0

p=ns

481 2 3 4 4

Household income - <£20,000/£20,000-£40,000/ >£40,000

43/143/198

11/54/58

9/23/26

p=ns

63/220/282

Test site (n) - Ely/Wisbech/Cambridge

385/7/52

170/1/19

69/1/3

p=ns

624/9/74

Height (cm) 168.8 (8.9) 171 (9.7) 171 (9.5) p<.05 169.5 (9.2)

Weight (kg) 69.3 (12.2) 83.9 (12.9) 96 (17) p<.001 75.2 (15.8)

BMI (kg/ m2) 24.2 (3.3) 28.7 (3.5) 31.5 (24.1–49.4) p<.001 25.4 (17.3–49.4)

Waist (cm) 83.9 (10.0) 97.3 (9.8) 108.0 (11.2) p<.001 89.2 (13.1)

Hip (cm) 99.6 (7.5) 106.4 (7.3) 114.8 (11.3) p<.001 102.6 (9.4)

WHR .8 (.1) .9 (.1) .9 (.1) p<.001 .9 (.1)

SBP (mmHg) 117.4 (14.2) 126.6 (15.4) 128.9 (12.7) p<.001 120.6 (15.0)

DBP (mmHg) 70.9 (9.3) 77.6 (10.4) 81.5 (9.2) p<.001 73.4 (10.2)

Liver fat score 3 (3–4) 6 (5–7) 8 (8–10) p<.001 4 (3–10)

Glucose (mmol/L) 4.7 (3.4–13.1) 4.9 (3.7–10.3) 5.1 (4.2–16.9) p<.001 4.7 (3.4–16.9)

Triglycerides (mmol/L) .8 (.2–5.8) 1.2 (.3–8.8) 1.8 (.8) p<.001 .9 (.2–8.8)

HDL cholesterol (mmol/L) 1.6 (.4) 1.3 (.7–3) 1.3 (.3) p<.001 1.6 (.4)

Total cholesterol (mmol/L) 5.2 (.9) 5.5 (1.2) 5.8 (1.2) p<.001 5.3 (1.1)

Total/ HDL cholesterol (mmol/L) 3.1 (1.7–8.6) 4 (1.3) 4.7 (1.4) p<.001 3.4 (1.7–9.0)

ALT (U/L) 21 (7–135) 29.5 (7–89) 37 (4–108) p<.001 24 (4–135)

GGT (U/L) 21 (9–433) 31.5 (10–324) 38 (17–176) p<.001 23 (9–433)

Bilirubin (Umol/L) 9 (2–44) 9 (3–27) 10.4 (5.2) p=ns 9 (2–44)

Alkaline Phosphatase (U/L) 71 (19.2) 78 (28–205) 83.9 (22.7) p<.001 74.5 (20.9)

Albumin (G/L) 41.9 (2.8) 42.3 (2.9) 41.2 (3.1) p<.05 41.9 (2.9)

Calcium (mmol/L) 2.2 (1.5–2.5) 2.2 (.1) 2.2 (.1) p<.05 2.2 (1.5–2.5)

Alcohol status (n) - never/ex/current/not known

23/5/320/48

10/0/101/13

3/1/51/4

p=ns

36/6/472/65

Alcohol status (n) - non-alcoholic; ≤17 units per week - alcoholic; >17 units per week

319 40

93 23

44 13

p<.05

459 76

Alcohol units a week 5 (0–52) 6 (0–60) 4 (0–67) p=ns 5 (0–67)

Smoking status (n) - never/ex/current/not known

239/105/40/12

70/38/12/4

27/23/6/3

p=ns

336/166/58/19

Units in cigarette equivalents a day 10 (0 – 40) 10.9 (8.0) 14.8 (8.2) p<.05 11.3 (8.1)

On any medication (y/n, y%) 141/252 (35.9) 52/68 (43.3) 31/28 (52.5) p<.001 224/348 (39.2)

On lipid medication (n) * On blood pressure medication (n) *

5 14

5 13

2 9

12 36

Data represent mean (sd) when normal distribution, median (min-max) when skewed distribution, or percentage when dichotomous. *, as specified in appendix Ι; ALT, alanine aminotransferase; GGT, gamma glutamyl aminotransferase; m, men; ns, not significant; w, women.

Chapter 2.

The Fenland Study

35

AGREEMENT OF LIVER FAT AND AMINOTRANSFERASES

ALT correlated significantly with liver fat score in both men (r=.37 with p<0.001) and

women (r=.36 with p<0.001). Explained variances were 13.4% in men and 12.9% in

women. Also, GGT correlated significantly with liver fat score in both men (r=.18 with

p<0.01) and women (r=.37 with p<0.001). Explained variances were 3.2% in men and

13.4% in women. Figure ІІІ shows ROC curves of ALT and GGT on diagnosis of fatty

liver.

Sensitivity(‘truepositives’)

1 – Specificity (‘false positives’)

20

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

40

30

30

19

25

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

32

36


1 – Specificity (‘false positives’)

a. ALT

In men, the ROC curve of ALT has an AUC of .693.

Optimum cut-off value is 25 (i.e. ≤25, >25), with SE

87%, SP 41%, PPV 54.3%, and NPV 79.7%.

In women, the ROC curve of ALT has an AUC of .681.


73.3%, SP 61%, PPV 35.3%, and NPV 88.8%.

b. GGT

In men, the ROC curve of GGT has an AUC of .721.


56.5%, SP 76.1%, PPV 65.6%, and NPV 68.5%.

In women, the ROC curve of GGT has an AUC of .713.


40%, SP 92.3%, PPV 60%, and NPV 84.2%.

Figure ІІІ. ROC curves of ALT and GGT on diagnosis of fatty liver

, men; , women. AUC, area under the curve; NPV, negative predictive value; PPV, positive predictive value; ROC, receiver operating characteristic; SE, sensitivity; SP, specificity.

36

ASSOCIATES OF FATTY LIVER

BMI class, alcohol class and medication

Figure ІV shows associations of BMI class, alcohol class and medication use with liver fat

score and ALT and GGT. In both men and women, an increase in BMI class is associated

with an increase in liver fat category, ALT, and GGT (p<0.001 for all). In men, an increase

in alcohol class is not associated with liver fat score and ALT (p=ns for both), whereas it is

associated with an increase in GGT (p<0.001). In women, an increase in alcohol class is

associated with an increase in liver fat score (p<0.05), whereas it is not associated with

ALT and GGT (p=ns for both). In men, medication use of any kind is associated with an

increase in liver fat score (p<0.001), an increase in ALT (p<0.01), and an increase in GGT

(p<0.001). In women, medication use of any kind is not associated with liver fat score,

ALT, and GGT (p=ns for all).

PREDICTORS OF FATTY LIVER

Table ІІ shows the results of univariate and multiple logistic regression analysis on the

presence of FLD. By multiple logistic regression analysis, only BMI (p<0.05, OR 1.3) was

significantly associated in men, and BMI (p<0.001, OR 1.6), waist circumference (p<0.01,

OR 1.1), and hip circumference (p<0.01, OR .9) were significantly associated in women.

Chapter 2.

The Fenland Study

37

p<0.001 for both men and women, by Kruskal-Wallis test



a. BMI class

p=ns for men and p<0.05 for women, by Kruskal-Wallis test

p=ns for both men and women, by Kruskal-Wallis test

p<0.001 for men and p=ns for women, by Kruskal-Wallis test

b. Alcohol class

p<0.001 for men and p=ns for women, by Mann-Whitney U test

p<0.01 for men and p=ns for women, by Mann-Whitney U test

p<0.001 for men and p=ns* for women, by Mann-Whitney U test

c. Medication use

Figure ІV. Associations of BMI class, alcohol class, and medication use

with liver fat score, ALT and GGT, in the population with liver fat scores

, men; , women.

BMI class: 1, BMI≤18.49; 2, BMI 18.5–24.99; 3, BMI 25–29.99; 4, BMI 30–34.99.

Alcohol class: 1, ≤17 units/week; 2, 18–34 units/week; 3, 35–51 units/week; 4, 52–68

units/week.

*, borderline significant (p<0.1).

38

Table ІІ. Logistic regression analysis on the presence of fatty liver

Independant variable(s) #

UNIVARIATE MULTIPLE

B p value ExpB (95% CI) B p value ExpB (95% CI)

Men

Age Household income Alcohol units a week Smoking units a day& Body mass index Waist circumference Hip circumference Triglycerides Total cholesterol HDL cholesterol Glucose SBP DBP ALT GGT Alkaline phosphatase Bilirubin Albumin Calcium Constant

.066 -.167 .010 .020 .410 .144 .137 .871 .230

-.517 .360 .049 .078 .040 .019 .011

-.028 -.023 1.700

p=0.001 p=ns p=ns p=ns

p<0.001

p<0.001

p<0.001

p<0.001 p=ns* p=ns p=ns* p<0.001

p<0.001

p<0.001

p=0.001 p=ns p=ns p=ns p=ns

1.068 (1.028–1.109) .846 (,574–1.248)

1.011 (.990–1.032) 1.020 (.975–1.068)

1.507 (1.343–1.691) 1.155 (1.110–1.203) 1.147 (1.095–1.200) 2.389 (1.629–3.503) 1.258 (.992–1.596) .596 (.272–1.309)

1.433 (.964–2.129) 1.050 (1.027–1.073) 1.081 (1.050–1.113) 1.040 (1.021–1.060) 1.019 (1.007–1.031) 1.011 (.997–1.024) .972 (.928–1.019) .977 (.889–1.074)

5.472 (.221–135.376)

.051

.277

.052 -.024 .188

.024

.007

.007

.004

-16.847

p=ns*

p<0.050 p=ns p=ns p=ns

p=ns p=ns p=ns p=ns

p<0.001

1.052 (.998–1.110)

1.319 (1.058–1.644) 1.053 (.968–1.145) .976 (.895–1.065)

1.207 (.827–1.762)

1.025 (.986–1.065) 1.007 (.956–1.060) 1.007 (.984–1.031) 1.004 (.995–1.014)

Women

Age Household income Alcohol units a week Smoking units a day& Body mass index Waist circumference Hip circumference Triglycerides Total cholesterol HDL cholesterol Glucose SBP DBP ALT GGT Alkaline phosphatase Bilirubin Albumin Calcium Constant

.058 -.170 .020 .013 .399 .162 .140

1.659 .502

-1.711 1.416 .034 .078 .067 .040 .035

-.099 -.071 4.922

p<0.010 p=ns p=ns p=ns

p<0.001

p<0.001

p<0.001

p<0.001

p<0.001

p<0.001

p<0.001

p<0.001

p<0.001

p<0.001

p<0.001

p<0.001

p=0.010 p=ns

p<0.010

1.059 (1.019–1.101) .844 (.583–1.220)

1.021 (.980–1.062) 1.013 (.958–1.072)

1.491 (1.358–1.637) 1.176 (1.132–1.222) 1.150 (1.109–1.192) 5.254 (3.047–9.058) 1.653 (1.281–2.132)

.181 (.084–.389) 4.120 (2.438–6.963) 1.034 (1.016–1.053) 1.081 (1.050–1.112) 1.070 (1.042–1.098) 1.041 (1.023–1.060) 1.036 (1.022–1.049)

.906 (.841–.977) .931 (.851–1.019)

137.248 (4.561–4129.853)

-.019

.452

.106 -.162 .199 .294 -.384 .309 -.013 .039 .021 -.024 .014 -.081

5.002

-20.108

p=ns

p<0.001

p<0.010

p<0.010 p=ns p=ns p=ns p=ns p=ns p=ns p=ns p=ns p=ns p=ns

p=ns*

p<0.010

.981 (.917–1.049)

1.571 (1.219–2.024) 1.111 (1.027–1.203)

.850 (.764–.947) 1.221 (.567–2.626) 1.342 (.826–2.182) .681 (.202–2.294)

1.362 (.616–3.010) .987 (.942–1.034)

1.039 (.967–1.117) 1.021 (.980–1.064) .976 (.946–1.008)

1.014 (.995–1.034) .923 (.837–1.017)

148.781 (.514–43103.494)

#Dependent variable: normal or fatty liver; *, borderline significant (p<0.1); &, cigarette equivalents a day. B, logistic regression coefficient; CI, confidence interval; ExpB, odds ratio; ns, not significant.

Chapter 2.

The Fenland Study

39

ASSOCIATIONS OF LIVER FAT CATEGORY, ALT QUINTILE AND GGT QUINTILE

WITH SEVERAL CARDIOVASCULAR RISK ESTIMATES

Figures V shows associations of liver fat category, ALT quintile, and GGT quintile with

several CVD risk estimates, in the population with liver fat scores. In both men and women,

an increase in liver fat category is associated with an increase in the number of MetS ATP

III components, the number of MetS IDF components, the MetS Z-score (p<0.001 for all),

and 10-year Framingham CVD risk (p=0.001 for men, p<0.001 for women). In men, an

increase in ALT quintile is associated with an increase in the number of MetS ATP III

components, the number of MetS IDF components, and the MetS Z-score (p<0.001 for all

three), but not with 10-year Framingham CVD risk (p=ns). In women, an increase in ALT

quintile is associated with an increase in the number of MetS ATP III components

(p<0.001), the number of MetS IDF components (p<0.01), the MetS Z-score (p<0.001), and

10-year Framingham CVD risk (p<0.01). In both men and women, an increase in GGT

quintile is associated with an increase in the number of MetS ATP III components, the

number of MetS IDF components, the MetS Z-score (p<0.001 for all three/six), and 10-year

Framingham CVD risk (p<0.01 for men, p<0.001 for women). Results for the Z-score by

Franks et al. (data not shown) were similar to the regular Z-score.

ROC curves of liver fat, ALT, and GGT on diagnosis of the metabolic syndrome

Figure VІ shows ROC curves of liver fat, ALT, and GGT for diagnosis of the MetS ATP

III. The figure shows that liver fat by ultrasound has more diagnostic value than ALT and

GGT. Results for the MetS IDF were similar (data not shown).

PREVALENCE OF METABOLIC SYNDROME COMPONENTS PER LIVER FAT

CATEGORY

Figures VІІ shows prevalence of the number of MetS ATP III components per liver fat

category. With increasing liver fat category the number of MetS ATP III components

increased as well. Of all participants with normal liver, most participants had no

component. Of all participants with mild fatty liver, most had 1 (women) or 2 (men)

components. Of all participants with moderate fatty liver, most participants had 3

components (both men and women). Results were similar for the MetS IDF (data not

shown).

a. Liver fat category

p<

0.0

01 f

or

both

men

and w

om

en,

by c

hi-

squar

e te

st

p<

0.0

01 f

or

both

men

and w

om

en,

by c

hi-

squar

e te

st

p<

0.0

01 f

or

both

men

and w

om

en,

by A

NO

VA

p=

0.0

01 f

or

men

and p

<0.0

01 f

or

wom

en,

by K

rusk

al W

alli

s te

st

b. ALT quintile

p<

0.0

01 f

or

both

men

and w

om

en,

by c

hi-

squar

e te

st

p<

0.0

01 f

or

men

and p

<0.0

1 f

or

wom

en,

by c

hi-

squar

e te

st

p<

0.0

01 f

or

both

men

and w

om

en,

by A

NO

VA

p=

ns

for

men

and p

<0.0

1 f

or

wom

en,

by K

rusk

al W

alli

s te

st

c. GGT quintile

p<

0.0

01 f

or

both

men

and w

om

en,

by c

hi-

squar

e te

st

p<

0.0

01 f

or

both

men

and w

om

en,

by c

hi-

squar

e te

st

p<

0.0

01 f

or

both

men

and w

om

en,

by A

NO

VA

p<

0.0

1 f

or

men

and p

<0.0

01 f

or

wom

en,

by K

rusk

al W

alli

s te

st

Figure V. C

om

par

ison o

f liver

fat

cat

egory

, A

LT

quin

tile

, an

d G

GT

quin

tile

on the

asso

ciat

ion w

ith s

ever

al c

ardio

vas

cula

r ri

sk e

stim

ates

, in

the

popula

tion w

ith liv

er f

at s

core

s

, m

en;

, w

om

en.

AL

T q

uin

tile

s*:

1,

≤23 f

or

men

and ≤

15 f

or

wom

en;

2,

24–29 f

or

men

and 1

6–18 f

or

wom

en;

3,

30–36 f

or

men

and 1

9–22 f

or

wom

en; 4, 37–47 f

or

men

and 2

3–28 f

or

wom

en; 5, ≥48 f

or

men

and ≥

29 f

or

wom

en. G

GT

quin

tile

s*: 1, ≤22 f

or

men

and ≤

15 f

or

wom

en; 2, 23–28 f

or

men

and 1

6–18 f

or

wom

en;

3, 29–37 f

or

men

and 1

9–22 f

or

wom

en;

4, 38–54 f

or

men

and 2

3–31 f

or

wom

en;

5, ≥55 f

or

men

and ≥

32 f

or

wom

en. *, in

ord

er to d

eriv

e both

AL

T a

nd G

GT

quin

tile

s an

d the

cum

ula

tive

Z-s

core

dat

a fr

om

the

tota

l popula

tion w

ere

use

d; ns,

not si

gnif

ican

t.

Chapter 2.

40


1 –Specificity(‘falsepositives’)

Liverfat score

Liverfat category

4

4

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%



ALT

ALT quintile

30

40

19

30

24

27

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%



GGT

GGT quintile

22

49

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

a. L

iver

fat

sco

re

In m

en,

the

RO

C c

urv

e of

liver

fat

sco

re h

as a

n A

UC

of

.822 (

.813 f

or

liver

fat

cat

egory

). O

ptim

um

cut-

off

val

ue

is 4

(i.e.

≤4,

>4),

with S

E 9

1.2

%,

SP 6

3%

, PPV

28.7

%, an

d N

PV

97.8

%.

In w

om

en,

the

RO

C c

urv

e of

liver

fat

sco

re h

as a

n

AU

C o

f .8

71 (

.826 f

or

liver

fat

cat

egory

). O

ptim

um

cut-

off

val

ue

is 4

(i

.e.

≤4,

>4),

with SE

77.4

%,

SP

83.4

%, PPV

32.4

%, an

d N

PV

97.3

%.

b. A

LT

In m

en,

the

RO

C c

urv

e of

AL

T h

as a

n A

UC

of

.682

(.671 f

or

AL

T q

uin

tile

). O

ptim

um

cut-

off

val

ue

is 2

7

(i.e

. ≤27,

>27),

w

ith

SE

85.3

%,

SP

41.8

%,

PPV

19.3

%, an

d N

PV

94.6

%.

In w

om

en, th

e R

OC

curv

e of

AL

T h

as a

n A

UC

of

.788

(.761 f

or

AL

T q

uin

tile

). O

ptim

um

cut-

off

val

ue

is 2

4

(i.e

. ≤24,

>24),

w

ith

SE

77.4

%,

SP

74.1

%,

PPV

23.5

%, an

d N

PV

97%

.

c. G

GT

In m

en,

the

RO

C c

urv

e of

GG

T h

as a

n A

UC

of

.693

(.683 f

or

GG

T q

uin

tile

). O

ptim

um

cut-

off

val

ue

is 4

9

(i.e

. ≤49,

>49),

with S

E 5

0%

, SP 8

0.3

%,

PPV

29.3

%,

and N

PV

90.8

%.

In w

om

en,

the

RO

C c

urv

e of

GG

T h

as a

n A

UC

of

.800 (

.794 f

or

GG

T q

uin

tile

). O

ptim

um

cut-

off

val

ue

is

22 (

i.e.

≤22,

>22),

with S

E 8

0.6

%,

SP 6

8.8

%,

PPV

21%

, an

d N

PV

97.2

%.

Figure VІ. R

OC

curv

es o

f liver

fat

, A

LT

and G

GT

on d

iagnosi

s of

AT

P I

II M

etS

, m

en;

, w

om

en.

AU

C, ar

ea u

nder

the

curv

e; N

PV

, neg

ativ

e pre

dic

tive

val

ue;

PPV

, posi

tive

pre

dic

tive

val

ue;

RO

C, re

ceiv

er o

per

atin

g c

har

acte

rist

ic;

SE

,

sensi

tivit

y; SP, sp

ecif

icit

y.

The Fenland Study

41

NORMAL LIV

ER

MIL

D FATTY LIV

ER

MODERATE FATTY LIV

ER

P<

0.0

01 f

or

both

men

and w

om

en, by c

hi-

squar

e te

st

Figure VІІ. P

reval

ence

of

the

num

ber

of

Met

S A

TP I

II c

om

ponen

ts p

er liv

er f

at c

ateg

ory

, m

en;

, w

om

en.

Chapter 2.

42

Table ІІІ. C

om

par

ison o

f W

este

rn g

ener

al p

opula

tion-b

ased

stu

die

s on f

atty

liv

er d

isea

se

Stu

dy

Dat

a

coll

ecti

on

Dia

gnosi

s

modal

ity

USG

E

xcl

usi

on

criter

ia

Incl

usi

on c

rite

ria

n

m

(%)

A

ge

E inta

ke

(g/d

ay)

BM

I T

ota

l

FL

D

A

FL

D

NA

FL

D

Sco

ring

criter

ia

Type

DN

LS 1

st-p

has

e 18

(Ita

ly)

1991 –

1993

1997

USG

1A

1, 1A

2,

2, 3

D

Hep

. B

, C

Cir

rhosi

s

BM

I<25; N

As

67

50.7

46.8

±11.7

5.8

22.3

±1.8

16.4

%

As

69

89.9

49.7

±10.7

71.1

23.3

±1.3

46.4

%

BM

I>30; N

As

66

40.9

47.8

±10.2

6.5

32.3

±2.8

75.8

%

As,

BM

I 55

87.3

51.5

±9.2

77.5

32.5

±2.5

94.5

%

SH

IP 2

0

(Ger

man

y)

1997 –

2001

USG

1A

1, 1A

2

D

Hep

. B

, C

Cir

rhosi

s

4222

49

20–79

29.9

%

39.7

%

27.2

%

DH

S 2

1,

22

(USA

)

2000 –

2002

1H

MR

S

na

na

All

2287

47.2

30–65

≈31%

#

37.6

% ##

Hip

anic

s 401

42.9

41 ±

9

30 ±

7

45%

#

Whit

es

734

51.1

46 ±

9

29 ±

6

33%

#

Bla

cks

1105

45.2

46 ±

10

31 ±

8

24%

#

EM

IL 2

3

(Ger

man

y)

2002

USG

1A

2, 2

D

2187

48%

18–65

27.4

%

DN

LS 2

nd-p

has

e 19

(Ita

ly)

2002 –

2003

USG

1A

1, 1A

2,

2, 3

D

Hep

. B

, C

Cir

rhosi

s

No S

LD

287

54.4

60

11

26.7

±4.7

20%

SL

D

311

62.4

58

9

27.8

±6

25%

Fen

land s

tudy

(UK

) 2005 –

2008

USG

1A

1, 1A

2,

2, 3

C, C

D

M

762

41.8

45.7

±7

0 (

0–72)

25.4

(17.3

–49.4

) 38.5

%

47.4

%

30.5

%

FIN

-D2D

stu

dy 2

4

(Fin

land)

2007

1H

MR

S

na

na

2766

7%

21%

Signs: #

, usi

ng a

cut-

off

val

ue

of

>5.5

% h

epat

ic T

G; ##, usi

ng a

cut-

off

val

ue

of

>5%

hep

atic

TG

.

Scoring criteria:

1A

1,

hyper

echogen

eity

of

liver

par

ench

ym

a; 1

A2,

hyper

echogen

eity

of

liver

par

ench

ym

a co

mpar

ed t

o t

he

adja

cent

stru

cture

s; 2

, dee

p a

tten

uat

ion;

3,

dec

reas

ed v

isib

ility o

f port

al v

eins.

Abbreviations:

A,

alco

holic;

BM

I, b

ody m

ass

index

(kg/m

2);

C,C

, sc

ori

ng i

n a

cum

ula

tive

fash

ion a

nd a

lloca

tion i

nto

cat

egori

es;

D,

dic

hoto

mous;

DH

S,

Dal

las

Hea

rt

Stu

dy; D

NL

S, D

ionyso

s N

utr

itio

n a

nd L

iver

Stu

dy; D

M, know

n d

iabet

es m

ellitu

s; E

, et

han

ol in

take;

EM

IL, Stu

dy o

n E

chin

oco

ccus

Multilocu

lari

s an

d I

nte

rnal

Dis

ease

s in

Leu

tkir

ch; FIN

-D2D

, th

e Fin

nis

h p

reven

tion p

rogra

m f

or

type

2 d

iabet

es; H

ep. B

, C

, hep

atitis

B a

nd C

; 1H

MR

S, pro

ton m

agnet

ic r

esonan

ce s

pec

trosc

opy; m

, per

centa

ge

of

men

; N

A, non-a

lcoholic;

SH

IP, Stu

dy o

f H

ealth in P

om

eran

ia; SL

D, su

spec

ted liv

er d

isea

se; T

G, tr

igly

ceri

de;

USG

, ultra

sonogra

phy.

The Fenland Study

43

Chapter 2.

44

DISCUSSION

The prevalence of FLD due to any etiology we found in the population aged 30 to 58 years

is 38.5%. Men contributed to this high prevalence the most (p<0.001). Importantly, this

high prevalence was found in a relatively healthy general population based study

population (e.g. people with known diabetes excluded). Additionally, BMI in the total

population (median 26.3 with range 16.9–63.5 kg/m2) was higher than BMI in the

population with liver fat scores (median 25.4 with range 17.3– 49.4 kg/m2). Therefore, the

true prevalence of FLD in East Anglia may likely be even higher than the already striking

prevalence reported in this paper.

An increase in liver fat category was associated with an increase in the number of ATP III

MetS components, the number of IDF MetS components, the MetS Z-score, and 10-year

Framingham CVD risk (figure V).

AGREEMENT OF AMINOTRANSFERASES AND LIVER FAT

Aminotransferases ALT and GGT correlated with liver fat, which was also found in other

imaging studies. ALT correlates with liver fat by magnetic resonance spectroscopy in non-

diabetic, non-alcoholic, apparently healthy men (r=0.44–0.62, p<0.0001) and women

(r=0.39–0.49, p<0.0001) 33, 34. The present study is also in line with recently proposed

lowering of ALT cut-off values 27, in order to achieve increased overall diagnostic value for

FLD. A study in blood donors suggested lowering ALT cut-off values from 40 to 30 for

men and from 30 to 19 for women 27. The present study suggests optimal ALT cut-off

values of 25 for men and 20 for women (figure ІІІ). Further lowering cut-off values for

aminotransferases, e.g. for screening studies, will increase sensitivity but cause a decreased

specificity as a consequence.

Additionally, this study suggests that both ALT and GGT are inferior to liver fat by

ultrasound regarding diagnosing the MetS ATP III (and IDF) as shown in figure VІ.

The Fenland Study

45

THE PRESENT STUDY COMPARED TO OTHER STUDIES

FLD prevalence compared to other prevalences

This is the first general population based study on imaging-determined FLD in the UK. Of

all Western general population based studies on FLD using imaging modalities,

chronologically organized in table ІІІ, the present study revealed the highest FLD

prevalence. Reported overall prevalence’s were 29.9% by the SHIP study 20, 31% by the

DHS study (33% in Whites) when using a cut-off value based upon the 95th percentile of a

low risk population 21, 22, 27.4% by the EMIL study 23, and 38.5% in the present Fenland

Study. In the FIN-D2D study using magnetic resonance spectroscopy, a substantial

difference in prevalence of AFLD (7%) and prevalence of NAFLD (21%) was found,

whereas BMI and components of the MetS were similar in both groups. Alcohol is a major

risk factor for advanced FLD, and compared to NAFLD, AFLD is more associated with

advanced FLD 35, 36. It is known that, while the FLD spectrum progresses to its advanced

stages, liver fat content decreases 5, 37. Thus theoretically, the AFLD group could include

more advanced FLD cases which are characterised by low liver fat content, likely causing

magnetic resonance spectroscopy to miss advanced AFLD cases. As cirrhosis (an advanced

stage of FLD) appears as bright liver by ultrasound, advanced FLD cases are most likely

included in the present Fenland Study.

Ultrasonography method compared to other ultrasonography methods

In previous old studies, FLD was scored on a semi-quantitative scale as normal, mild,

moderate, or severe 38, 39. When compared with histology, sensitivity and specificity of this

method range from 89%–94% and 84%–93%, respectively 38, 39. In the present study, liver

fat content was scored in a cumulative fashion, followed by allocation to liver fat

categories, similar to the scoring system by Hamaguchi et al. (2007) 40. A differences

between the two scoring systems is that, as based on older studies 38, 39, Hamaguchi et al.

considered hyperechogeneity as a mandatory component for diagnosis of FLD 40.

Additionally, the authors scored attenuation on a 3-point scale (compared to the present 4-

point scale) and vessel blurring on a 2-point scale (compared to the present 4-point scale)

40. The authors validated their method using histology in a population excluding other liver

disease and alcoholics (total n=94, including NAFLD n=64) 40. Their method revealed an

AUC of 0.980, a sensitivity of 91.2%–92.6%, a specificity of 100%, an intra-observer

reliability by Cohen’s kappa of 0.95 (p<0.001) and an inter-observer reliability by Cohen’s

Chapter 2.

46

kappa of 0.95 (p<0.001) 40. It should be noted that the current used ultrasonography scoring

method has not been validated yet.

Study design compared to other study designs

Besides period of data collection and imaging methods, there are other differences between

these studies regarding study populations. Some important characteristics are included in

table ІІІ. Although derived from the general population, the DNLS 1st-phase 18, DHS 21, 22

and DNLS 2nd-phase 19 were strongly designed on BMI and alcohol intake 18, ethnicity 21, 22,

and presence and absence of suspected liver disease 19. The DNLS study 18, 19 and SHIP

study 20 excluded hepatitis and other known liver disease, whereas these were not

determined in the DHS 21, 22 and the present Fenland Study. Another important difference

between the present and other studies in that known diabetes mellitus is an exclusion factor

in the present Fenland Study.

Weaknesses of the present study

The population with and without liver fat scores differed with respect to sex (females

58.2% versus 53.4% respectively [p<0.05]) and BMI (median 25.4 versus 26.6 kg/m2

respectively [p<0.001]). Analysis on the total population revealed that overall men are more

obese than women. The median BMI of men (27 kg/m2) and women (25.6 kg/m2) in the

total population differed significantly (p<0.001). The mean waist circumference of men

(98.2 cm) and women (86.7 cm) differed significantly (p<0.001) as well. Despite of the

major back log of liver fat scores compared to other data, this difference in sex and BMI

might be attributable to difficulties in scanning obese people, i.e. mostly men, using an

image modality 25, 26, e.g. ultrasonography 41, 42 and magnetic resonance spectroscopy as

well 22. In the DHS, using magnetic resonance spectroscopy, only participants fitting the

scanner could be included 22. Besides, being sometimes not possible, the diagnostic value of

ultrasonography is decreased in obese people as well 41, 42.

ASSOCIATES OF FATTY LIVER

BMI class, alcohol class and medication

BMI (notably peripheral fat depots) 43, alcohol 44, and certain drugs 44 are associated with

the etiology of FLD. Both figure ІV and results of the logistic regression analyses (table ІІ)

indicate that BMI is more strongly associated with liver fat by ultrasound than alcohol use.

This was also found in the DNLS 18. However, it might be important to distinguish AFLD

The Fenland Study

47

and NAFLD as both may follow different clinical courses 35, 36. Studies by histology

suggest that AFLD is overall more associated with progression of liver damage, i.e.

progression of fibrosis stage 35, 36, which may delineate a regression of liver fat content 5, 37.

NAFLD is overall more associated with CVD 45, but it should be noted that even people

with NAFLD are at increased risk for severe liver pathology 5-7. Strangely, follow-up of

participants from the DNLS suggests that alcohol use is the most important predictor of

both incidence and regression of fatty liver by ultrasound 46. However, theoretically this

regression might be attributable to an increase in fibrosis stage (notably not determined by

histology) accompanied by a decrease in liver fat content 5, 37. The association between

medication use and liver fat score, particularly in men (figure ІVc), suggests that the

participants with higher liver fat score are unhealthier. Whereas in the SHIP study,

increased medication use was largely attributable to diabetes and lipid lowering medication

47 this is not possible in the present non-diabetic study population. Smoking was not

associated with liver fat by ultrasound (table ІІ), as also previously found 23, 48.

CARDIOVASCULAR DISEASE RISK ESTIMATES

This study is in line with the idea of FLD being the hepatic component of the MetS 3. The

optimal cut-off value for diagnosis of the MetS ATP III was a liver fat score of 4 (≤4, >4),

which corresponds with the threshold between normal liver and (mild) fatty liver.

Additionally, liver fat was associated with the 10-year Framingham CVD risk (figure V).

Although there is speculation about the value of the MetS for CVD risk prediction, a meta-

analysis of longitudinal studies (37 studies including 43 cohorts) revealed that participants

with the MetS are at increased risk for incident CVD events 49.

CONCLUSION

This study shows a striking prevalence of FLD in East Anglia, particularly in men. As FLD

is associated with several CVD risk estimates, this striking prevalence may delineate an

increased CVD risk in this population.

Chapter 2.

48

ACKLOWLEDGEMENTS

The Fenland Study is funded by the Wellcome Trust and the Medical Research Council.

We are grateful to all the volunteers for their time and help, and to the General Practitioners

and practice staff for help with recruitment. We thank the Fenland Study Investigators,

Fenland Study Co-ordination team and the Epidemiology Field, Data and Technical teams.

Biochemical assays were performed by the National Institute for Health Research,

Cambridge Biomedical Research Centre, Core Biochemistry Assay Laboratory, and the

Cambridge University Hospitals NHS Foundation Trust, Department of Clinical

Biochemistry.

The Fenland Study

49

Reference List





(5) Clark JM, Diehl AM. Nonalcoholic fatty liver disease: an underrecognized cause of cryptogenic cirrhosis. JAMA 2003 June 11;289(22):3000-4.

(6) Marrero JA, Fontana RJ, Su GL, Conjeevaram HS, Emick DM, Lok AS. NAFLD may be a common underlying liver disease in patients with hepatocellular carcinoma in the United States. Hepatology 2002 December;36(6):1349-54.

(7) Bugianesi E, Leone N, Vanni E, Marchesini G, Brunello F, Carucci P, Musso A, De Paolis P, Capussotti L, Salizzoni M, Rizzetto M. Expanding the natural history of nonalcoholic steatohepatitis: from cryptogenic cirrhosis to hepatocellular carcinoma. Gastroenterology 2002 July;123(1):134-40.

(8) Yang RZ, Blaileanu G, Hansen BC, Shuldiner AR, Gong DW. cDNA cloning, genomic structure, chromosomal mapping, and functional expression of a novel human alanine aminotransferase. Genomics 2002 March;79(3):445-50.

(9) Lindblom P, Rafter I, Copley C, Andersson U, Hedberg JJ, Berg AL, Samuelsson A, Hellmold H, Cotgreave I, Glinghammar B. Isoforms of alanine aminotransferases in human tissues and serum--differential tissue expression using novel antibodies. Arch Biochem Biophys 2007 October 1;466(1):66-77.

(10) Leh H, Courtay C, Gerardin P, Wellman M, Siest G, Visvikis A. Cloning and expression of a novel type (III) of human gamma-glutamyltransferase truncated mRNA. FEBS Lett 1996 October 7;394(3):258-62.

(11) Pettersen I, Andersen JH, Bjornland K, Mathisen O, Bremnes R, Wellman M, Visvikis A, Huseby NE. Heterogeneity in gamma-glutamyltransferase mRNA expression and glycan structures. Search for tumor-specific variants in human liver metastases and colon carcinoma cells. Biochim Biophys Acta 2003 May 30;1648(1-2):210-8.

Chapter 2.

50

(12) Zamora S, Adams C, Butzner JD, Machida H, Scott RB. Elevated aminotransferase activity as an indication of muscular dystrophy: case reports and review of the literature. Can J Gastroenterol 1996 October;10(6):389-93.

(13) Mofrad P, Contos MJ, Haque M, Sargeant C, Fisher RA, Luketic VA, Sterling RK, Shiffman ML, Stravitz RT, Sanyal AJ. Clinical and histologic spectrum of nonalcoholic fatty liver disease associated with normal ALT values. Hepatology 2003 June;37(6):1286-92.

(14) Gholam PM, Flancbaum L, Machan JT, Charney DA, Kotler DP. Nonalcoholic Fatty liver disease in severely obese subjects. Am J Gastroenterol 2007 February;102(2):399-408.

(15) Fracanzani AL, Valenti L, Bugianesi E, Andreoletti M, Colli A, Vanni E, Bertelli C, Fatta E, Bignamini D, Marchesini G, Fargion S. Risk of severe liver disease in nonalcoholic fatty liver disease with normal aminotransferase levels: a role for insulin resistance and diabetes. Hepatology 2008 September;48(3):792-8.

(16) Assy N, Bekirov I, Mejritsky Y, Solomon L, Szvalb S, Hussein O. Association between thrombotic risk factors and extent of fibrosis in patients with non-alcoholic fatty liver diseases. World J Gastroenterol 2005 October 7;11(37):5834-9.

(17) Wieckowska A, Papouchado BG, Li Z, Lopez R, Zein NN, Feldstein AE. Increased hepatic and circulating interleukin-6 levels in human nonalcoholic steatohepatitis. Am J Gastroenterol 2008 June;103(6):1372-9.

(18) Bellentani S, Saccoccio G, Masutti F, Croce LS, Brandi G, Sasso F, Cristanini G, Tiribelli C. Prevalence of and risk factors for hepatic steatosis in Northern Italy. Ann Intern Med 2000 January 18;132(2):112-7.

(19) Bedogni G, Miglioli L, Masutti F, Tiribelli C, Marchesini G, Bellentani S. Prevalence of and risk factors for nonalcoholic fatty liver disease: the Dionysos nutrition and liver study. Hepatology 2005 July;42(1):44-52.

(20) Volzke H, Robinson DM, Kleine V, Deutscher R, Hoffmann W, Ludemann J, Schminke U, Kessler C, John U. Hepatic steatosis is associated with an increased risk of carotid atherosclerosis. World J Gastroenterol 2005 March 28;11(12):1848-53.

(21) Browning JD, Szczepaniak LS, Dobbins R, Nuremberg P, Horton JD, Cohen JC, Grundy SM, Hobbs HH. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. Hepatology 2004 December;40(6):1387-95.

(22) Szczepaniak LS, Nurenberg P, Leonard D, Browning JD, Reingold JS, Grundy S, Hobbs HH, Dobbins RL. Magnetic resonance spectroscopy to measure hepatic

The Fenland Study

51

triglyceride content: prevalence of hepatic steatosis in the general population. Am


(23) Haenle MM, Brockmann SO, Kron M, Bertling U, Mason RA, Steinbach G, Boehm BO, Koenig W, Kern P, Piechotowski I, Kratzer W. Overweight, physical activity, tobacco and alcohol consumption in a cross-sectional random sample of German adults. BMC Public Health 2006;6:233.

(24) Kotronen A, Yki-Jarvinen H, Mannisto S, Saarikoski L, Korpi-Hyovalti E, Oksa H, Saltevo J, Saaristo T, Sundvall J, Tuomilehto J, Peltonen M. Non-alcoholic and alcoholic fatty liver disease - two diseases of affluence associated with the metabolic syndrome and type 2 diabetes: the FIN-D2D survey. BMC Public

Health 2010;10:237.

(25) Uppot RN, Sahani DV, Hahn PF, Kalra MK, Saini SS, Mueller PR. Effect of obesity on image quality: fifteen-year longitudinal study for evaluation of dictated radiology reports. Radiology 2006 August;240(2):435-9.

(26) Uppot RN, Sahani DV, Hahn PF, Gervais D, Mueller PR. Impact of obesity on medical imaging and image-guided intervention. AJR Am J Roentgenol 2007 February;188(2):433-40.

(27) Prati D, Taioli E, Zanella A, Della Torre E, Butelli S, Del Vecchio E, Vianello L, Zanuso F, Mozzi F, Milani S, Conte D, Colombo M, Sirchia G. Updated definitions of healthy ranges for serum alanine aminotransferase levels. Ann Intern

Med 2002 July 2;137(1):1-10.

(28) National Statistics interim standard classifications for presenting ethnic and national groups data. Office for National Statistics 2001 2008 August 5;http://www.ons.gov.uk/about-statistics/classifications/archived/ethnic-interim/presenting-data/index.html?

(29) Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III). JAMA 2001 May 16;285(19):2486-97.

(30) Alberti KG, Zimmet P, Shaw J. The metabolic syndrome--a new worldwide definition. Lancet 2005 September 24;366(9491):1059-62.

(31) Franks PW, Ekelund U, Brage S, Wong MY, Wareham NJ. Does the association of habitual physical activity with the metabolic syndrome differ by level of cardiorespiratory fitness? Diabetes Care 2004 May;27(5):1187-93.

(32) D'Agostino RB, Sr., Vasan RS, Pencina MJ, Wolf PA, Cobain M, Massaro JM, Kannel WB. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation 2008 February 12;117(6):743-53.

Chapter 2.

52

(33) Westerbacka J, Corner A, Tiikkainen M, Tamminen M, Vehkavaara S, Hakkinen AM, Fredriksson J, Yki-Jarvinen H. Women and men have similar amounts of liver and intra-abdominal fat, despite more subcutaneous fat in women: implications for sex differences in markers of cardiovascular risk. Diabetologia 2004 August;47(8):1360-9.

(34) Kotronen A, Westerbacka J, Bergholm R, Pietilainen KH, Yki-Jarvinen H. Liver fat in the metabolic syndrome. J Clin Endocrinol Metab 2007 September;92(9):3490-7.

(35) Mills SJ, Harrison SA. Comparison of the natural history of alcoholic and nonalcoholic fatty liver disease. Curr Gastroenterol Rep 2005 February;7(1):32-6.

(36) Dam-Larsen S, Franzmann MB, Christoffersen P, Larsen K, Becker U, Bendtsen F. Histological characteristics and prognosis in patients with fatty liver. Scand J

Gastroenterol 2005 April;40(4):460-7.





(41) Mottin CC, Moretto M, Padoin AV, Swarowsky AM, Toneto MG, Glock L, Repetto G. The role of ultrasound in the diagnosis of hepatic steatosis in morbidly obese patients. Obes Surg 2004 May;14(5):635-7.

(42) Moura Almeida A, Cotrim HP, Barbosa DB, de Athayde LG, Santos AS, Bitencourt AG, De Freitas LA, Rios A, Alves E. Fatty liver disease in severe obese patients: diagnostic value of abdominal ultrasound. World J Gastroenterol 2008 March 7;14(9):1415-8.

(43) Donnelly KL, Smith CI, Schwarzenberg SJ, Jessurun J, Boldt MD, Parks EJ. Sources of fatty acids stored in liver and secreted via lipoproteins in patients with nonalcoholic fatty liver disease. J Clin Invest 2005 May;115(5):1343-51.

The Fenland Study

53

(44) Fromenty B, Pessayre D. Inhibition of mitochondrial beta-oxidation as a mechanism of hepatotoxicity. Pharmacol Ther 1995;67(1):101-54.

(45) Ekstedt M, Franzen LE, Mathiesen UL, Thorelius L, Holmqvist M, Bodemar G, Kechagias S. Long-term follow-up of patients with NAFLD and elevated liver enzymes. Hepatology 2006 October;44(4):865-73.

(46) Bedogni G, Miglioli L, Masutti F, Castiglione A, Croce LS, Tiribelli C, Bellentani S. Incidence and natural course of fatty liver in the general population: the Dionysos study. Hepatology 2007 November;46(5):1387-91.

(47) Baumeister SE, Volzke H, Marschall P, John U, Schmidt CO, Flessa S, Alte D. Impact of fatty liver disease on health care utilization and costs in a general population: a 5-year observation. Gastroenterology 2008 January;134(1):85-94.

(48) Chavez-Tapia NC, Lizardi-Cervera J, Perez-Bautista O, Ramos-Ostos MH, Uribe M. Smoking is not associated with nonalcoholic fatty liver disease. World J

Gastroenterol 2006 August 28;12(32):5196-200.

(49) Gami AS, Witt BJ, Howard DE, Erwin PJ, Gami LA, Somers VK, Montori VM. Metabolic syndrome and risk of incident cardiovascular events and death: a systematic review and meta-analysis of longitudinal studies. J Am Coll Cardiol 2007 January 30;49(4):403-14.

(50) Ford ES. Prevalence of the metabolic syndrome defined by the International Diabetes Federation among adults in the U.S. Diabetes Care 2005 November;28(11):2745-9.

Appen

dix Ι. E

stim

ates

of

met

abolic

card

iovas

cula

r ri

sk, w

ith s

pec

ific

atio

ns

Metabolic

risk

variable

Components/variables

Specification

NC

EP A

TPΙΙ

Ι# 2

9

1. E

levat

ed w

aist

>

102 c

m (

m),

>88 c

m (

f)

2. E

levat

ed T

Gs

TG

s ≥1.7

mm

ol/L

A

nd/o

r tr

eatm

ent w

ith s

tatins

and/o

r fi

bra

tes*

3. D

ecre

ased

HD

L

HD

L <

1.0

3 m

mol/L

(m

) or

<1.2

9 m

mol/L

(f)

4. E

levat

ed b

lood p

ress

ure

E

ither

SB

P≥130 m

mH

g, an

d/o

r D

BP≥85 m

mH

g, an

d/o

r an

ti-h

yper

tensi

ves**

5. E

levat

ed f

asting p

lasm

a glu

cose

≥6.1

6 m

mol/L

IDF## 3

0

1. E

levat

ed w

aist

(m

andatory component)

• For

‘Whites

’, ‘

Bla

cks’

, an

d ‘

mix

ed’,

e.g

. ‘B

lack

Bri

tish

’:

≥94 c

m (

m)

and ≥

80 c

m (

f) 3

0, 50.

• For

‘Asi

ans’

, ‘S

outh

Asi

ans’

, an

d ‘

mix

ed’,

e.g

. ‘A

sian

Bri

tish

’:

≥90 c

m (

m)

and ≥

80 c

m (

f) 3

0, 50.

2. E

levat

ed T

Gs

TG

>1.7

mm

ol/L

A

nd/o

r or

trea

tmen

t w

ith s

tatins

and/o

r fi

bra

tes*

3. D

ecre

ased

HD

L

HD

L <

1.0

3 m

mol/L

(m

) or

<1.2

9 m

mol/L

(f)

4. E

levat

ed b

lood p

ress

ure

E

ither

SB

P≥130 m

mH

g, an

d/o

r D

BP≥85 m

mH

g, an

d/o

r an

ti-h

yper

tensi

ves**

5. E

levat

ed f

asting p

lasm

a glu

cose

≥

5.6

mm

ol/L

Z-s

core

1. B

MI;

2. W

aist

; 3. T

Gs;

4. In

ver

ted H

DL

;

5. SB

P; 6. D

BP; 7. G

luco

se

All v

aria

ble

s w

ere

tran

sform

ed u

sing:

σµ

−Χ

=Ζ

(for

mal

es a

nd f

emal

es s

epar

atel

y),

wher

e X

is

the

ori

gin

al

val

ue,

an

d µ

an

d σ

are

mea

n

and

stan

dar

d

dev

iation

of

the

study

popula

tion,

resp

ectivel

y.

All indiv

idual

Z-s

core

wer

e ad

ded

to f

orm

a c

um

ula

tive

Z-s

core

.

Z-s

core

, bas

ed o

n

Fra

nks

et a

l. 3

1

1. Z

-obes

ity: (B

MI+

wai

st)/

2

2. Z

-dysl

ipid

eam

ia: (T

Gs+

inver

ted H

DL

)/2

3. Z

-hyper

tensi

on: (S

BP+

DB

P)/

2,

4. Z

-glu

cose

: G

luco

se

All Z

-sco

res

wer

e st

andar

diz

ed u

sing:

Ζ

Ζ−

Ζ=

Ζσµ

Fra

nks

(fo

r m

ales

and f

emal

es s

epar

atel

y).

All

indiv

idual

Z-F

ranks

wer

e ad

ded

to f

orm

a c

um

ula

tive

Z-F

ranks

score

10-y

ear

FR

S 3

2

1. A

ge;

2. T

ota

l-C

; 3. H

DL

-C;

4. SB

P w

ith/w

ithout an

ti-h

yper

tensi

ves

**;

5. Sm

oker

(yes

/no);

6. D

iabet

ic (

yes

/no)

Alloca

tion o

f poin

ts a

ccord

ing to D

’Agost

ino e

t al

. (2

008);

ta

ble

s 5 a

nd 7

32

Signs:

#, th

e N

CE

P A

TP

ΙΙΙ

Met

S i

s ‘d

iagnose

d’

if a

t le

ast th

ree

com

ponen

ts a

re p

rese

nt;

##, th

e ID

F M

etS is

‘dia

gnose

d’

if e

levat

ed w

aist

plu

s at

lea

st t

wo o

ther

com

ponen

ts a

re p

rese

nt;

*, B

NF-c

ode:

2.1

2 (

lipid

-reg

ula

ting d

rugs,

i.e.

sta

tins

and f

ibra

tes)

; **,

BN

F-c

odes

: 2.4

(bet

a-ad

renore

cepto

r blo

ckin

g d

rugs)

, 2.5

.1 (

vas

odil

ator

anti

hyper

tensi

ve

dru

gs)

, 2.5

.2 (

centr

ally

act

ing a

nti

hyper

tensi

ve

dru

gs)

, 2.5

.4 (

alpha-

adre

nore

cepto

r blo

ckin

g d

rugs)

, 2.5

.5.1

(angio

tensi

n-c

onver

ting e

nzy

me

inhib

itors

), 2

.5.5

.2 (

angio

tensi

n-I

I-re

cepto

r an

tagonis

t), 2.6

.1 (

nit

rate

s), 2.6

.2 (

calc

ium

-chan

nel

blo

cker

s), 2.6

.3 (

pota

ssiu

m c

han

nel

act

ivat

ors

), a

nd 2

.6.4

(per

ipher

al v

asodil

ators

and r

elat

ed d

rugs)

.

Abbreviations:

BM

I, b

ody m

ass

index

(kg/m

2);

BN

F,

Bri

tish

Nat

ional

Form

ula

ry;

C-c

hole

ster

ol; D

BP,

dia

stolic

blo

od p

ress

ure

; (f

), c

ut-

off

val

ue

for

fem

ales

; FR

S,

Fra

min

gham

ris

k s

core

; H

DL

, hig

h d

ensi

ty l

ipopro

tein

; ID

F,

inte

rnat

ional

dia

bet

es f

eder

atio

n; (m

), c

ut-

off

val

ue

for

mal

es; M

etS, m

etab

oli

c sy

ndro

me;

NC

EP A

TPΙΙ

Ι, N

atio

nal

chole

ster

ol ed

uca

tion p

rogra

m. A

dult

tre

atm

ent pan

el 3

; SB

P, sy

stoli

c blo

od p

ress

ure

; T

G, tr

igly

ceri

des

.

Chapter 2.

54

Chapter 3

Non-alcoholic fatty liver disease is associated with

cardiovascular disease risk markers

Obes Rev 2009; 10(4):412-419 _ Review Paper

Mireille A. Edens 1

Folkert Kuipers 2

Ronald P. Stolk 1

Department of Epidemiology 1 Laboratory of Pediatrics; CLDMD 2



Chapter 3.

56

ABSTRACT

Background: Recognition of the link between non-alcoholic fatty liver disease (NAFLD)

and cardiovascular disease (CVD) has boosted research in this area.

Main objective: To review the literature on NAFLD in the context of CVD, focussing on

underlying mechanisms and treatment.

Results: Besides excessive fatty acid influx, etiologic factors may include components of

the metabolic syndrome, cytokines, and mitochondrial dysfunction. NAFLD is associated

with both hepatic and systemic insulin resistance. In the case of NAFLD, the liver

overproduces several atherogenic factors, notably inflammatory cytokines, glucose,

lipoproteins, coagulation factors, and factors increasing blood pressure. Intervention studies

on diet and laparoscopic surgery revealed improvements of hepatic fat content and CVD

risk profile. Pharmacological approaches with potential benefit have been developed as

well, but effects are often confounded by weight change.

Conclusions: NAFLD is associated with an increased CVD risk profile (and hepatic risk).

In order to improve CVD risk profile, prevention and treatment of NAFLD seems

advisable. However, well designed intervention studies, randomised clinical trials, and

long-term follow-up studies are scarce.

Cardiovascular risk of the fatty liver

57

INTRODUCTION

The global increase of overweight and obesity is alarming 1, as obesity is a risk factor for

many diseases including cardiovascular disease (CVD) 2. Obesity has the highest CVD risk

when fat is located in the abdominal region 3.

In the case of obesity accompanied by insulin resistance, triglycerides (TGs) are often

excessively stored ectopically, i.e. in organs and muscles rather than in adipocytes. When

TGs accumulate within hepatocytes (HC in figure І), a pathological condition usually

referred to as fatty liver disease (FLD) will develop. FLD includes a wide spectrum, which

can broadly be divided into steatosis and steatohepatitis 4. Non-alcoholic FLD (NAFLD) is

used to describe FLD in a person who drinks no or little alcohol prior to diagnosis. In the

literature, the amount of ethanol that is allowed for the diagnosis NAFLD varies greatly,

but is maximally 20 g/day for women and 30 g/day for men. The prevalence of NAFLD in

the general adult Western population is relatively high, i.e. 20% 5, 6, whereas the prevalence

of total FLD, including both NAFLD and alcoholic FLD (AFLD), is approximately 30% 5.

In obese non-diabetic Western adults, the prevalence of NAFLD ranges from 80.4% to

97.9% 7, 8.

Although there is a hepatic risk for patients with NAFLD, notably cirrhosis 9 and

hepatocarcinoma 10, the CVD risk for patients with NAFLD may be higher 10. Few studies

have revealed evidence of an association between NAFLD and early CVD markers 11, CVD

events 12, 13 and CVD mortality 10. Follow-up of patients with NAFLD showed a higher

incidence of CVD compared to controls 10, 13. A study by Hamaguchi et al. (2007) revealed

that NAFLD was an independent predictor, even stronger than the metabolic syndrome, for

first time CVD events 13.

The recognition of the CVD risk of NAFLD has boosted research in this area during the

recent years. We have reviewed the literature on NAFLD in the context of CVD risk profile

markers. In this paper, after a short description of the etiology of NAFLD, we provide an

overview of the cardiovascular risk of NAFLD. Finally, the potential need for prevention

and treatment of NAFLD in order to improve CVD risk profile is addressed.

Chapter 3.

58

ETIOLOGY

The etiology of the hepatic lipid imbalance underlying the pathophysiology of NAFLD has

been increasingly unravelled. Besides excessive non-esterified fatty acid (NEFA) influx,

mediating factors may include: i] components of the metabolic syndrome 14, 15, ii] cytokines

16-24, and iii] mitochondrial dysfunction 25 (figure І). A study on the incidence of NAFLD

suggested that the metabolic syndrome precedes NAFLD 14, with insulin resistance as a

cornerstone 15. Animal models, reviewed by Diehl et al. (2005) 16, revealed involvement of

the cytokine tumour necrosis factor alpha (TNFα) and its possible antagonist adiponectin 17

in the pathogenesis of NAFLD. This is reinforced by cross-sectional studies showing

changes in messenger ribonucleic acid (mRNA) of both TNFα receptors (increase) and

adiponectin receptors (decrease) in NAFLD 18-20. TNFα, recently reviewed by Ryden and

Arner (2007), may have numerous mediating actions, amongst others increasing insulin

resistance and inhibiting fatty acid oxidation 22. Several animal studies on adiponectin,

recently reviewed by Lafontan and Viguerie (2006), suggest opposite actions of adiponectin

on energy metabolism, i.e. increasing insulin sensitivity and stimulation of fatty acid

oxidation 23. Additionally, adiponectin might contribute to inhibition of lipogenesis 24.

Polymorphisms in the gene encoding adiponectin receptor 1 are associated with the

presence of high hepatic fat content (and insulin resistance) 21. Mitochondrial dysfunction

(in NASH), recently reviewed by Begriche et al. (2006), can be caused by oxidative stress

amongst others, and may result in TG accumulation and eventually cell death, i.e. necrosis

25.


59

FATTY LIVER DERIVED CVD RISK FACTORS

The liver secretes numerous CVD risk factors, notably cytokines, glucose, lipoproteins,

coagulation factors, and factors increasing blood pressure. In the case of NAFLD,

production of several of these risk factors is altered (figure І).

Systemiccytokineinflux, a.o. TNFα↑/ adiponectin↓:

•From↑white adiposetissue, notablyvisceralfat

De novo

lipogenesis↑

Mitochondrial

dysfunction↑

SystemicNEFA influx↑:

•From↑nutrition

•From↑visceralfat

•From↑peripheralfat

•From↑hepaticfat

Oxidative

stress↑

TG

storage↑

GLUT-4↓

Insulin

signalling↓

I-R↓

Insulinuptake↓and resistance↑:

a.o. C-peptide↑, Insulin↑a,

HOMA-IR↑, QUICKI↑, OGTT↓

Glucose uptake↓:

Glucose↑, AGEs↑

Cytokines↑:

a.o. TNFα↑, IL-6↑

Inflammation↑

Lypolysis↑:

VLDL-apo-B↑, VLDL-TG↑,

LDL size↓, NEFA↑, HDL↓

Glycolysis↑:

Glucose↑, AGEs↑

Heamostaticfactors↑:

a.o. Fibrinogen↑b, Factor VΙΙ↑, PAI-1↑

Anti-heamostaticfactors:

a.o. AntithrombinΙΙΙ↓, ProteinC↑b, ProteinS↑b

HC

KC

HSC

Inflammationmarkers↑:

a.o. CRP↑, SAA↑

Neurotransmitters↑:

Angiotensinogen↑→AngiotensinΙΙ↑→BP↑

Transaminases↑b:

AST↑, ALT↑, GGT↑

Receptors:

•TNF-R1↑

•sTNF-R2↑

•Adipo-R1↓

•Adipo-R2↓ Cytokines↑:

a.o. TNFα↑/ adiponectin↓, IL-6↑

NEFA-FF

•LDL-R

•HDL-R

TNF-α↑/

adiponectin↓

Organell&

celldamage↑

→necrosis↑

NEFA-TP

Figure І. M

odel

on the

card

iovas

cula

r ri

sk o

f non-a

lcoholic

fatty liv

er d

isea

se

Sig

ns:

↑ =

incr

ease

, ↓ =

dec

reas

e.

Abbre

via

tions:

a, d

epen

den

t on p

ancr

eatic

β-c

ell fu

nct

ion; b, m

ay d

ecre

ase

in a

dvan

ced N

AFL

D; A

dip

o-R

1, ad

iponec

tin r

ecep

tor

1; A

dip

o-R

2, ad

iponec

tin r

ecep

tor

2;

AG

Es,

advan

ced g

lyca

tion e

ndpro

duct

s; G

LU

T-4

, glu

cose

tra

nsp

ort

er 4

; H

C, H

epat

ocy

te; H

DL

-R, hig

h d

ensi

ty lip

opro

tein

rec

epto

r; H

SC

, H

epat

ic S

tellat

e C

ells

;

I-R

, in

sulin r

ecep

tor;

KC

, K

upff

er C

ell;

LD

L-R

, lo

w d

ensi

ty lip

opro

tein

rec

epto

r; N

EFA

-FF, non-e

ster

ifie

d f

atty

aci

d f

lip-f

loppin

g; N

EFA

-TP, non-e

ster

ifie

d f

atty

aci

d tra

nsp

ort

pro

tein

; SA

A, se

rum

am

ylo

id A

; sT

NF-R

2, so

luble

tum

our

nec

rosi

s fa

ctor

alpha

rece

pto

r 2; T

NF-R

1, tu

mour

nec

rosi

s fa

ctor

alpha

rece

pto

r 1.

Chapter 3.

60

INFLAMMATION

In both patients with non-alcoholic steatosis (NAS) and non-alcoholic steatohepatitis

(NASH), analysis of liver biopsies revealed hepatic distribution (mRNA) of the

inflammatory cytokine TNFα with its receptors 18, 20, and the anti-inflammatory

adipocytokine adiponectin with its receptors 20. As suggested by animal models, the

increased amount of fatty acids present in the case of NAFLD may mediate hepatic

production of TNFα, causing increased levels of systemic TNFα 26. Upon hepatocyte

damage activated liver-specific macrophages ‘Kupffer Cells’ (KCs) will secrete more

cytokines into the blood, amongst others TNFα 18 and interleukin 6 (IL-6) 16, 27, 28. TNFα

and IL-6 are considered to induce hepatic production of the acute phase protein ‘C-reactive

protein’ (CRP) 29.

Hepatic expression of TNFα mRNA is significantly higher in patients with NASH

compared to NAS 18, 20, and hepatic expression of adiponectin mRNA is significantly lower

in patients with NASH compared to NAS 20. Systemic TNFα concentration is significantly

elevated in both patients with NAS and NASH 19. Both hepatic IL-6 mRNA and systemic

IL-6 concentration are elevated in patients with NAS and the highest in NASH 28. Elevated

CRP is present in 25% of controls compared to 60% of NAFLD patients (p=.003) 30.

Fasting adiponectin concentration predicts hepatic fat content 31, 32 and is significantly

lower in both patients with NASH and NAS compared to controls 19, 32, for both men and

women 19. Fasting adiponectin is inversely correlated with hepatic fat content in healthy

non-diabetic subjects 31, 33, but non-significantly in patients with type 2 diabetes mellitus

(n=10) 31. TNFα, IL-6 and CRP may contribute to the inflammatory CVD milieu,

predisposing to atherosclerosis and CVD 34-36. On the contrary, studies in both humans and

animals 23, 37 have revealed both anti-inflammatory 23 and anti-thrombotic 37 properties of

adiponectin, enabling a protective association between adiponectin and CVD 38.

In addition to a direct predisposition to atherosclerosis, cytokines may have an indirect

effect as well, as cytokines may mediate insulin resistance (figure І). Animal models 22, 39

and both in vitro- 40 and in vivo 41 human studies provided evidence that not fat

accumulation itself, but fat-derived cytokines play a role in (obesity related) insulin

resistance. The pleiotropic cytokine TNFα interferes with the hepatic insulin receptor and

the intra-hepatocellular insulin signalling cascade 22, 40, causing both hepatic and systemic

insulin resistance 22, 40. Additionally, administration of human recombinant TNFα in human


61

cancer patients, resulted in increased levels of very low density lipoprotein (VLDL) and

TG, and decreased levels of high density lipoprotein (HDL) 41, which are features of

diabetic dyslipideamia 42.

HYPERGLYCAEMIA AND DIABETIC DYSLIPIDEAMIA

In the healthy situation, insulin stimulates hepatic and peripheral glucose uptake and

suppresses hepatic glucose production. In patients with NAFLD, hepatic glucose uptake

may be less effectively stimulated by insulin, contributing to elevated plasma glucose

concentrations 40, 43. Moreover, hepatic glucose production might be less effectively

suppressed by insulin 40, but not in non-diabetic NAFLD patients 43, 44.

Hepatic fat content correlates positively with fasting glucose 45, glucose levels after an oral

glucose tolerance test 46, fasting C-peptide 45, fasting insulin 45, and insulin resistance by

homeostatic model assessment (HOMA-IR) 31. In the general population, the odds ratio of

NAFLD compared to normal liver is 9.1 for hyperglycaemia 6. Additionally, the odds ratios

of NAFLD, compared to normal liver, increase with increasing insulin quartile (4.2, 5.9 and

20.0; for the 2nd, 3rd, and 4th quartile, respectively) and HOMA-IR quartile (2.3, 4.4 and

16.7; for the 2nd, 3rd, and 4th quartile, respectively) 6. Plasma glucose and its advanced

glycation endproducts (AGEs) are considered atherogenic 47, 48 and predispose to CVD 49.

In the healthy situation, insulin suppresses hepatic production and secretion of VLDL 50. In

patients with NAFLD, VLDL secretion is less effectively suppressed by insulin, causing

increased systemic VLDL-TG concentrations 31. Hepatic fat content correlates positively

with VLDL1-TG and VLDL1-apolipoprotein-B secretion rates 31. Besides the presence of

NAFLD, the altered hepatic lipid composition present in the case of NAFLD 51, 52 might

play a role in both altered VLDL secretion rates and altered lipoprotein composition as

well. An in vitro study suggested that the presence of different types of fatty acids in the

liver results in both different VLDL-apolipoprotein-B secretion rates and lipoprotein

composition 53.

In patients with NAFLD, VLDL concentration is increased, VLDL particles are larger,

small dense LDL particles predominate, whereas large HDL particle concentration is

decreased 54. Hepatic fat content correlates inversely with fasting HDL cholesterol

concentration 45. In the general population, the odds ratios of NAFLD compared to normal

liver are 6.3 for low HDL cholesterol concentration and 3.5 for hypertriglyceraemia 6.

Chapter 3.

62

Systemic lipids, i.e. TGs 48, 55 and cholesterol (HDL excluded) 47, 55, have been considered

atherogenic and predispose to CVD 48, 55.

COAGULATION

Many coagulation factors are synthesised by hepatocytes 56, 57. The limited data on the

association between obesity-independent markers of NAFLD, amongst others coagulation

factors, has recently been reviewed by Kotronen and Yki-Jarvinen (2008) 45. In patients

with NAFLD, the liver overproduces several factors (figure І), of which plasminogen

activator inhibitor 1 (PAI-1) has direct atherogenic effects 47. However, many factors

(fibrinogen, protein C and protein S) are increased in NAS, but tend to be lower in NASH

58. This may suggest that initially increased factors will decrease while NAS progresses to

advanced NASH. In advanced liver disease, bleeding problems are well-known to occur 56,

57.

BLOOD PRESSURE

Hepatocytes produce angiotensinogen 59, 60, a precursor of angiotensin ΙΙ. Upon hepatocyte

damage activated Hepatic Stellate Cells even synthesize and secrete mature angiotensin ΙΙ

61. Angiotensin ΙΙ is a major proatherogenic and vasoconstrictive peptide/neurotransmitter

47, 61, considered to predispose to elevated blood pressure 59-61 and possibly CVD 62. Both

systolic (only in females) and diastolic blood pressure correlate with hepatic fat content 45.

In the general population, univariate odds ratios of NAFLD, compared to normal liver are

2.0 for systolic hypertension and 1.7 for diastolic hypertension 6.


63

PREVENTION AND TREATMENT

As obesity 1, including childhood obesity 63, has been increasing in the general population,

an earlier peak prevalence of NAFLD and an increased CVD risk profile may be expected

in the future, delineating the need for prevention.

Once NAFLD is present, treatment seems advisable to improve CVD risk profile and

hepatic risk. Prevention and treatment trials of NAFLD should focus on etiologic factors,

notably systemic NEFA influx, components of the metabolic syndrome, cytokines and

mitochondrial dysfunction (figure І). Several studies have investigated the effect of

lowering hepatic fat content on improving CVD risk profile markers, however, no

longitudinal studies on lowering CVD events have been performed yet.

For an overview on diagnosis modalities for NAFLD, the reader is referred to recent review

papers 45, 64. As serum levels of transferases are of limited use 45, 64, 65, the intervention

studies included in table І, have been limited to studies with a diagnosis by histology or

imaging.

Dietary (plus exercise) and laparoscopic surgery interventions (3 – 15 months) revealed that

overall loss of adipose tissue, determined by a decreased BMI (2.6 – 18.2 BMI-points),

promotes loss of hepatic fat content and results in an improved CVD risk profile 66-70. This

suggests that weight loss should be pursued in patients with NAFLD.

Patients unable to lose weight or non-overweight patients might benefit from drug

treatment. However, no medication is currently licensed for NAFLD treatment even though

some have shown potential benefit. Additionally, as weight loss is associated with an

improvement of NAFLD 66-70, statistical adjustment for the amount of weight change as

potential confounder (table І) should be considered when determining true effects of

pharmacological interventions. For an overview on potential treatment modalities, the

reader is referred to recent review papers 45, 64.

Some of the most often prescribed medicine are based on improving CVD risk profile

(components of the metabolic syndrome) by acting on hepatic lipid metabolism, e.g. statins

and fibrates 71, 72, and by acting on hepatic glucose metabolism, e.g. metformin 73. Although

often prescribed, studies on statins 74, fibrates 75 and metformin 76, 77 in the case of NAFLD

are few, often accompanied by weight loss, and inconclusive. Studies on the angiotensin ΙΙ

receptor antagonist losartan revealed beneficial histological results 74, 78.

Chapter 3.

64

As the possible antagonists TNFα and adiponectin 17 may be involved in both the etiology

of NAFLD 16-24 and worsening of CVD risk profile 23, 34-38, TNFα-lowering in NAFLD

might be beneficial to improve both NAFLD and CVD risk profile, as suggested by studies

on pentoxifylline 74, 79.

CONCLUSION

NAFLD is associated with an increased CVD risk profile (and hepatic risk). In order to

improve CVD risk profile, prevention and treatment of NAFLD seems advisable. However,

well designed intervention studies, randomised clinical trials and long-term follow-up

studies are scarce.

Table І. Sel

ecti

on o

f in

terv

enti

on s

tudie

s on N

AF

LD

and C

VD

ris

k m

arker

s, lim

ited

to s

tudie

s w

ith a

dia

gnosi

s by h

isto

logy o

r im

agin

g

Studied treatm

ent modality

Patients

n

Length

HFC Tot

BMI

Ref

BE

FO

RE-A

FT

ER

ST

UD

IES

Moderate W

L (

2.6

BM

I-poin

ts) by diet

T2D

M, obes

e *

8

1-3

m

↓

na

↓

66

Moderate W

L (

≈2.7

BM

I-poin

ts) by diet

HF

C>

5%

, G

DM

11

3–6 m

↓

na

↓

67

HF

C<

5%

, G

DM

12

↓

na

↓

Severe WL

(10.4

BM

I-poin

ts) by diet

Morb

idly

obes

e *

41

9 m

↓

- -

69

Severe WL

(17

BM

I-poin

ts) by L-surgery

NA

FL

D

70

15 m

↓

↓

↓

70

Atorvastatin (

10 m

g)

NA

SH

, ↑lipid

s 10

9 m

↓

=

=

74

Metform

in (

max

imum

2 g

) N

AS

H, ↑A

LT

17

12 m

↓

↓

↓

76

Losartan

(50 m

g)

NA

SH

, hyper

tensi

on

12

9 m

↓

↓

=

74

Pentoxifylline

(2*400 m

g)

NA

SH

, T

2D

M

13

9 m

↓

↓

=

74

Pentoxifylline

(3*400 m

g)

N

AS

H, ↑A

LT

9

12 m

-

↓

=#

79

NO

N-R

AN

DO

MIS

ED

CO

NT

RO

LL

ED

TR

IAL

S

Moderate W

L (

3 B

MI-

poin

ts) by diet/ exc. V

S. no d

iet/ e

xc.

N

AF

LD

, A

FL

D *

25

3 m

↓

- ↓

68

UDCA

(13–15 m

g/k

g)

VS. c

lofi

bra

te (

2*1 g

) N

AS

H, C

hole

lith

iasi

s 40

12 m

↓

- =

75

NA

SH

, ↑T

G

n-3 PUFA ethyl ester

(1 g

) VS. n

o n

-3 P

UF

A e

thyl es

ter

NA

FL

D *

56

6–12 m

↓

na

=

80

RA

ND

OM

ISE

D C

LIN

ICA

L T

RIA

LS

Roziglitazone

(2*4 m

g)

VS. m

etfo

rmin

(2*1 g

) T

2D

M

20

4 m

↓

na

↑

77

Sig

ns:

↑ =

sig

nif

ican

t in

crea

se a

nd/

or

signif

ican

tly i

nfe

rior,

↓ =

sig

nif

ican

t dec

reas

e an

d/

or

signif

ican

tly s

uper

ior,

‘=

’ =

no c

han

ge,

- =

mis

sing v

alue,

* =

eth

anol

inta

ke

unre

port

ed o

r hig

her

than

allow

ed f

or

NA

FL

D,

# =

p ≤

0.1

0.

Abbre

via

tions: e

xc.

= e

xer

cise

, G

DM

= g

esta

tional

dia

bet

es m

ellitu

s, H

FC

= h

epat

ic f

at c

onte

nt, L

-surg

ery =

lap

arosc

opic

surg

ery, m

= m

onth

s, n

= a

mount of

subje

cts

who

had

both

pre

and p

ost

mea

sure

men

ts,

na

= n

ot

applica

ble

, PU

FA

= p

oly

unsa

tura

ted f

atty

aci

ds,

Ref

= r

efer

ence

, T

ot. =

tota

l N

ASH

sco

re/

activity i

ndex

, T

2D

M =

type

2

dia

bet

es m

ellitu

s, U

DC

A =

urs

odeo

xych

olic

acid

, VS. =

ver

sus,

WL

= w

eight lo

ss.


65

Chapter 3.

66

Reference List (1) Friedman JM. Obesity in the new millennium. Nature 2000 April

6;404(6778):632-4.

(2) Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW, Jr. Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med 1999 October 7;341(15):1097-105.

(3) Yusuf S, Hawken S, Ounpuu S, Bautista L, Franzosi MG, Commerford P, Lang CC, Rumboldt Z, Onen CL, Lisheng L, Tanomsup S, Wangai P, Jr., Razak F, Sharma AM, Anand SS. Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case-control study. Lancet 2005 November 5;366(9497):1640-9.




(7) Venturi C, Zoppini G, Zamboni C, Muggeo M. Insulin sensitivity and hepatic steatosis in obese subjects with normal glucose tolerance. Nutr Metab Cardiovasc

Dis 2004 August;14(4):200-4.

(8) Colicchio P, Tarantino G, del Genio F, Sorrentino P, Saldalamacchia G, Finelli C, Conca P, Contaldo F, Pasanisi F. Non-alcoholic fatty liver disease in young adult severely obese non-diabetic patients in South Italy. Ann Nutr Metab 2005 September;49(5):289-95.



(11) Targher G. Non-alcoholic fatty liver disease, the metabolic syndrome and the risk of cardiovascular disease: the plot thickens. Diabet Med 2007 January;24(1):1-6.


67

(12) Targher G, Bertolini L, Poli F, Rodella S, Scala L, Tessari R, Zenari L, Falezza G. Nonalcoholic Fatty liver disease and risk of future cardiovascular events among type 2 diabetic patients. Diabetes 2005 December;54(12):3541-6.

(13) Hamaguchi M, Kojima T, Takeda N, Nagata C, Takeda J, Sarui H, Kawahito Y, Yoshida N, Suetsugu A, Kato T, Okuda J, Ida K, Yoshikawa T. Nonalcoholic fatty liver disease is a novel predictor of cardiovascular disease. World J Gastroenterol 2007 March 14;13(10):1579-84.

(14) Hamaguchi M, Kojima T, Takeda N, Nakagawa T, Taniguchi H, Fujii K, Omatsu T, Nakajima T, Sarui H, Shimazaki M, Kato T, Okuda J, Ida K. The metabolic syndrome as a predictor of nonalcoholic fatty liver disease. Ann Intern Med 2005 November 15;143(10):722-8.

(15) Qureshi K, Abrams GA. Metabolic liver disease of obesity and role of adipose tissue in the pathogenesis of nonalcoholic fatty liver disease. World J

Gastroenterol 2007 July 14;13(26):3540-53.

(16) Diehl AM, Li ZP, Lin HZ, Yang SQ. Cytokines and the pathogenesis of non-alcoholic steatohepatitis. Gut 2005 February;54(2):303-6.

(17) Maeda N, Shimomura I, Kishida K, Nishizawa H, Matsuda M, Nagaretani H, Furuyama N, Kondo H, Takahashi M, Arita Y, Komuro R, Ouchi N, Kihara S, Tochino Y, Okutomi K, Horie M, Takeda S, Aoyama T, Funahashi T, Matsuzawa Y. Diet-induced insulin resistance in mice lacking adiponectin/ACRP30. Nat Med 2002 July;8(7):731-7.

(18) Crespo J, Cayon A, Fernandez-Gil P, Hernandez-Guerra M, Mayorga M, Dominguez-Diez A, Fernandez-Escalante JC, Pons-Romero F. Gene expression of tumor necrosis factor alpha and TNF-receptors, p55 and p75, in nonalcoholic steatohepatitis patients. Hepatology 2001 December;34(6):1158-63.

(19) Hui JM, Hodge A, Farrell GC, Kench JG, Kriketos A, George J. Beyond insulin resistance in NASH: TNF-alpha or adiponectin? Hepatology 2004 July;40(1):46-54.

(20) Kaser S, Moschen A, Cayon A, Kaser A, Crespo J, Pons-Romero F, Ebenbichler CF, Patsch JR, Tilg H. Adiponectin and its receptors in non-alcoholic steatohepatitis. Gut 2005 January;54(1):117-21.

(21) Stefan N, Machicao F, Staiger H, Machann J, Schick F, Tschritter O, Spieth C, Weigert C, Fritsche A, Stumvoll M, Haring HU. Polymorphisms in the gene encoding adiponectin receptor 1 are associated with insulin resistance and high liver fat. Diabetologia 2005 November;48(11):2282-91.

Chapter 3.

68

(22) Ryden M, Arner P. Tumour necrosis factor-alpha in human adipose tissue -- from signalling mechanisms to clinical implications. J Intern Med 2007 October;262(4):431-8.

(23) Lafontan M, Viguerie N. Role of adipokines in the control of energy metabolism: focus on adiponectin. Curr Opin Pharmacol 2006 December;6(6):580-5.

(24) Xu A, Wang Y, Keshaw H, Xu LY, Lam KS, Cooper GJ. The fat-derived hormone adiponectin alleviates alcoholic and nonalcoholic fatty liver diseases in mice. J Clin Invest 2003 July;112(1):91-100.

(25) Begriche K, Igoudjil A, Pessayre D, Fromenty B. Mitochondrial dysfunction in NASH: causes, consequences and possible means to prevent it. Mitochondrion 2006 February;6(1):1-28.

(26) Feldstein AE, Werneburg NW, Canbay A, Guicciardi ME, Bronk SF, Rydzewski R, Burgart LJ, Gores GJ. Free fatty acids promote hepatic lipotoxicity by stimulating TNF-alpha expression via a lysosomal pathway. Hepatology 2004 July;40(1):185-94.

(27) Choi S, Diehl AM. Role of inflammation in nonalcoholic steatohepatitis. Curr

Opin Gastroenterol 2005 November;21(6):702-7.

(28) Wieckowska A, Papouchado BG, Li Z, Lopez R, Zein NN, Feldstein AE. Increased hepatic and circulating interleukin-6 levels in human nonalcoholic steatohepatitis. Am J Gastroenterol 2008 June;103(6):1372-9.

(29) Blake GJ, Ridker PM. Inflammatory bio-markers and cardiovascular risk prediction. J Intern Med 2002 October;252(4):283-94.

(30) Brea A, Mosquera D, Martin E, Arizti A, Cordero JL, Ros E. Nonalcoholic fatty liver disease is associated with carotid atherosclerosis: a case-control study. Arterioscler Thromb Vasc Biol 2005 May;25(5):1045-50.

(31) Adiels M, Taskinen MR, Packard C, Caslake MJ, Soro-Paavonen A, Westerbacka J, Vehkavaara S, Hakkinen A, Olofsson SO, Yki-Jarvinen H, Boren J. Overproduction of large VLDL particles is driven by increased liver fat content in man. Diabetologia 2006 April;49(4):755-65.

(32) Targher G, Bertolini L, Rodella S, Zoppini G, Scala L, Zenari L, Falezza G. Associations between plasma adiponectin concentrations and liver histology in patients with nonalcoholic fatty liver disease. Clin Endocrinol (Oxf) 2006 June;64(6):679-83.

(33) Westerbacka J, Corner A, Tiikkainen M, Tamminen M, Vehkavaara S, Hakkinen AM, Fredriksson J, Yki-Jarvinen H. Women and men have similar amounts of liver and intra-abdominal fat, despite more subcutaneous fat in women:


69

implications for sex differences in markers of cardiovascular risk. Diabetologia 2004 August;47(8):1360-9.

(34) Pearson TA, Mensah GA, Alexander RW, Anderson JL, Cannon RO, III, Criqui M, Fadl YY, Fortmann SP, Hong Y, Myers GL, Rifai N, Smith SC, Jr., Taubert K, Tracy RP, Vinicor F. Markers of inflammation and cardiovascular disease: application to clinical and public health practice: A statement for healthcare professionals from the Centers for Disease Control and Prevention and the American Heart Association. Circulation 2003 January 28;107(3):499-511.

(35) Pai JK, Pischon T, Ma J, Manson JE, Hankinson SE, Joshipura K, Curhan GC, Rifai N, Cannuscio CC, Stampfer MJ, Rimm EB. Inflammatory markers and the risk of coronary heart disease in men and women. N Engl J Med 2004 December 16;351(25):2599-610.

(36) Lizardi-Cervera J, Chavez-Tapia NC, Perez-Bautista O, Ramos MH, Uribe M. Association among C-reactive protein, Fatty liver disease, and cardiovascular risk. Dig Dis Sci 2007 September;52(9):2375-9.

(37) Kato H, Kashiwagi H, Shiraga M, Tadokoro S, Kamae T, Ujiie H, Honda S, Miyata S, Ijiri Y, Yamamoto J, Maeda N, Funahashi T, Kurata Y, Shimomura I, Tomiyama Y, Kanakura Y. Adiponectin acts as an endogenous antithrombotic factor. Arterioscler Thromb Vasc Biol 2006 January;26(1):224-30.

(38) Dekker JM, Funahashi T, Nijpels G, Pilz S, Stehouwer CD, Snijder MB, Bouter LM, Matsuzawa Y, Shimomura I, Heine RJ. Prognostic value of adiponectin for cardiovascular disease and mortality. J Clin Endocrinol Metab 2008 April;93(4):1489-96.

(39) Grefhorst A, Hoekstra J, Derks TG, Ouwens DM, Baller JF, Havinga R, Havekes LM, Romijn JA, Kuipers F. Acute hepatic steatosis in mice by blocking beta-oxidation does not reduce insulin sensitivity of very-low-density lipoprotein production. Am J Physiol Gastrointest Liver Physiol 2005 September;289(3):G592-G598.

(40) Gupta D, Varma S, Khandelwal RL. Long-term effects of tumor necrosis factor-alpha treatment on insulin signaling pathway in HepG2 cells and HepG2 cells overexpressing constitutively active Akt/PKB. J Cell Biochem 2007 February 15;100(3):593-607.

(41) Sherman ML, Spriggs DR, Arthur KA, Imamura K, Frei E, III, Kufe DW. Recombinant human tumor necrosis factor administered as a five-day continuous infusion in cancer patients: phase I toxicity and effects on lipid metabolism. J Clin

Oncol 1988 February;6(2):344-50.

(42) Adiels M, Boren J, Caslake MJ, Stewart P, Soro A, Westerbacka J, Wennberg B, Olofsson SO, Packard C, Taskinen MR. Overproduction of VLDL1 driven by

Chapter 3.

70

hyperglycemia is a dominant feature of diabetic dyslipidemia. Arterioscler

Thromb Vasc Biol 2005 August;25(8):1697-703.

(43) Seppala-Lindroos A, Vehkavaara S, Hakkinen AM, Goto T, Westerbacka J, Sovijarvi A, Halavaara J, Yki-Jarvinen H. Fat accumulation in the liver is associated with defects in insulin suppression of glucose production and serum free fatty acids independent of obesity in normal men. J Clin Endocrinol Metab 2002 July;87(7):3023-8.

(44) Bugianesi E, Gastaldelli A, Vanni E, Gambino R, Cassader M, Baldi S, Ponti V, Pagano G, Ferrannini E, Rizzetto M. Insulin resistance in non-diabetic patients with non-alcoholic fatty liver disease: sites and mechanisms. Diabetologia 2005 April;48(4):634-42.


(46) Thamer C, Machann J, Haap M, Stefan N, Heller E, Schnodt B, Stumvoll M, Claussen C, Fritsche A, Schick F, Haring H. Intrahepatic lipids are predicted by visceral adipose tissue mass in healthy subjects. Diabetes Care 2004 November;27(11):2726-9.

(47) Bansilal S, Farkouh ME, Fuster V. Role of insulin resistance and hyperglycemia in the development of atherosclerosis. Am J Cardiol 2007 February 19;99(4A):6B-14B.

(48) Heine RJ, Dekker JM. Beyond postprandial hyperglycaemia: metabolic factors associated with cardiovascular disease. Diabetologia 2002 April;45(4):461-75.

(49) Goff DC, Jr., Gerstein HC, Ginsberg HN, Cushman WC, Margolis KL, Byington RP, Buse JB, Genuth S, Probstfield JL, Simons-Morton DG. Prevention of cardiovascular disease in persons with type 2 diabetes mellitus: current knowledge and rationale for the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial. Am J Cardiol 2007 June 18;99(12A):4i-20i.

(50) Sparks JD, Sparks CE. Insulin modulation of hepatic synthesis and secretion of apolipoprotein B by rat hepatocytes. J Biol Chem 1990 May 25;265(15):8854-62.

(51) Araya J, Rodrigo R, Videla LA, Thielemann L, Orellana M, Pettinelli P, Poniachik J. Increase in long-chain polyunsaturated fatty acid n - 6/n - 3 ratio in relation to hepatic steatosis in patients with non-alcoholic fatty liver disease. Clin Sci (Lond) 2004 June;106(6):635-43.

(52) Puri P, Baillie RA, Wiest MM, Mirshahi F, Choudhury J, Cheung O, Sargeant C, Contos MJ, Sanyal AJ. A lipidomic analysis of nonalcoholic fatty liver disease. Hepatology 2007 October;46(4):1081-90.


71

(53) Mitmesser SH, Carr TP. Trans fatty acids alter the lipid composition and size of apoB-100-containing lipoproteins secreted by HepG2 cells. J Nutr Biochem 2005 March;16(3):178-83.

(54) Cali AM, Zern TL, Taksali SE, de Oliveira AM, Dufour S, Otvos JD, Caprio S. Intrahepatic fat accumulation and alterations in lipoprotein composition in obese adolescents: a perfect proatherogenic state. Diabetes Care 2007 December;30(12):3093-8.

(55) Gotto AM, Jr. Triglyceride: the forgotten risk factor. Circulation 1998 March 24;97(11):1027-8.

(56) Peck-Radosavljevic M. Review article: coagulation disorders in chronic liver disease. Aliment Pharmacol Ther 2007 November;26 Suppl 1:21-8.

(57) Thachil J. Relevance of clotting tests in liver disease. Postgrad Med J 2008 April;84(990):177-81.

(58) Assy N, Bekirov I, Mejritsky Y, Solomon L, Szvalb S, Hussein O. Association between thrombotic risk factors and extent of fibrosis in patients with non-alcoholic fatty liver diseases. World J Gastroenterol 2005 October 7;11(37):5834-9.

(59) Ron D, Brasier AR, Habener JF. Transcriptional regulation of hepatic angiotensinogen gene expression by the acute-phase response. Mol Cell

Endocrinol 1990 December 21;74(3):C97-104.

(60) Massiera F, Bloch-Faure M, Ceiler D, Murakami K, Fukamizu A, Gasc JM, Quignard-Boulange A, Negrel R, Ailhaud G, Seydoux J, Meneton P, Teboul M. Adipose angiotensinogen is involved in adipose tissue growth and blood pressure regulation. FASEB J 2001 December;15(14):2727-9.

(61) Bataller R, Sancho-Bru P, Gines P, Lora JM, Al-Garawi A, Sole M, Colmenero J, Nicolas JM, Jimenez W, Weich N, Gutierrez-Ramos JC, Arroyo V, Rodes J. Activated human hepatic stellate cells express the renin-angiotensin system and synthesize angiotensin II. Gastroenterology 2003 July;125(1):117-25.

(62) Silventoinen K, Magnusson PK, Neovius M, Sundstrom J, Batty GD, Tynelius P, Rasmussen F. Does obesity modify the effect of blood pressure on the risk of cardiovascular disease? A population-based cohort study of more than one million Swedish men. Circulation 2008 October 14;118(16):1637-42.

(63) Patton HM, Sirlin C, Behling C, Middleton M, Schwimmer JB, Lavine JE. Pediatric nonalcoholic fatty liver disease: a critical appraisal of current data and implications for future research. J Pediatr Gastroenterol Nutr 2006 October;43(4):413-27.

Chapter 3.

72

(64) Preiss D, Sattar N. Non-alcoholic fatty liver disease: an overview of prevalence, diagnosis, pathogenesis and treatment considerations. Clin Sci (Lond) 2008 September;115(5):141-50.

(65) Kunde SS, Lazenby AJ, Clements RH, Abrams GA. Spectrum of NAFLD and diagnostic implications of the proposed new normal range for serum ALT in obese women. Hepatology 2005 September;42(3):650-6.

(66) Petersen KF, Dufour S, Befroy D, Lehrke M, Hendler RE, Shulman GI. Reversal of nonalcoholic hepatic steatosis, hepatic insulin resistance, and hyperglycemia by moderate weight reduction in patients with type 2 diabetes. Diabetes 2005 March;54(3):603-8.

(67) Tiikkainen M, Bergholm R, Vehkavaara S, Rissanen A, Hakkinen AM, Tamminen M, Teramo K, Yki-Jarvinen H. Effects of identical weight loss on body composition and features of insulin resistance in obese women with high and low liver fat content. Diabetes 2003 March;52(3):701-7.

(68) Ueno T, Sugawara H, Sujaku K, Hashimoto O, Tsuji R, Tamaki S, Torimura T, Inuzuka S, Sata M, Tanikawa K. Therapeutic effects of restricted diet and exercise in obese patients with fatty liver. J Hepatol 1997 July;27(1):103-7.

(69) Andersen T, Gluud C, Franzmann MB, Christoffersen P. Hepatic effects of dietary weight loss in morbidly obese subjects. J Hepatol 1991 March;12(2):224-9.

(70) Mattar SG, Velcu LM, Rabinovitz M, Demetris AJ, Krasinskas AM, Barinas-Mitchell E, Eid GM, Ramanathan R, Taylor DS, Schauer PR. Surgically-induced weight loss significantly improves nonalcoholic fatty liver disease and the metabolic syndrome. Ann Surg 2005 October;242(4):610-7.

(71) Chalasani N. Statins and hepatotoxicity: focus on patients with fatty liver. Hepatology 2005 April;41(4):690-5.

(72) Paumelle R, Staels B. Cross-talk between statins and PPARalpha in cardiovascular diseases: clinical evidence and basic mechanisms. Trends Cardiovasc Med 2008 April;18(3):73-8.

(73) Inzucchi SE. Oral antihyperglycemic therapy for type 2 diabetes: scientific review. JAMA 2002 January 16;287(3):360-72.

(74) Georgescu EF, Georgescu M. Therapeutic options in non-alcoholic steatohepatitis (NASH). Are all agents alike? Results of a preliminary study. J Gastrointestin

Liver Dis 2007 March;16(1):39-46.

(75) Laurin J, Lindor KD, Crippin JS, Gossard A, Gores GJ, Ludwig J, Rakela J, McGill DB. Ursodeoxycholic acid or clofibrate in the treatment of non-alcohol-induced steatohepatitis: a pilot study. Hepatology 1996 June;23(6):1464-7.


73

(76) Bugianesi E, Gentilcore E, Manini R, Natale S, Vanni E, Villanova N, David E, Rizzetto M, Marchesini G. A randomized controlled trial of metformin versus vitamin E or prescriptive diet in nonalcoholic fatty liver disease. Am J

Gastroenterol 2005 May;100(5):1082-90.

(77) Tiikkainen M, Hakkinen AM, Korsheninnikova E, Nyman T, Makimattila S, Yki-Jarvinen H. Effects of rosiglitazone and metformin on liver fat content, hepatic insulin resistance, insulin clearance, and gene expression in adipose tissue in patients with type 2 diabetes. Diabetes 2004 August;53(8):2169-76.

(78) Yokohama S, Yoneda M, Haneda M, Okamoto S, Okada M, Aso K, Hasegawa T, Tokusashi Y, Miyokawa N, Nakamura K. Therapeutic efficacy of an angiotensin II receptor antagonist in patients with nonalcoholic steatohepatitis. Hepatology 2004 November;40(5):1222-5.

(79) Satapathy SK, Sakhuja P, Malhotra V, Sharma BC, Sarin SK. Beneficial effects of pentoxifylline on hepatic steatosis, fibrosis and necroinflammation in patients with non-alcoholic steatohepatitis. J Gastroenterol Hepatol 2007 May;22(5):634-8.

(80) Capanni M, Calella F, Biagini MR, Genise S, Raimondi L, Bedogni G, Svegliati-Baroni G, Sofi F, Milani S, Abbate R, Surrenti C, Casini A. Prolonged n-3 polyunsaturated fatty acid supplementation ameliorates hepatic steatosis in patients with non-alcoholic fatty liver disease: a pilot study. Aliment Pharmacol

Ther 2006 April 15;23(8):1143-51.

74

Chapter 4

Pathogenesis of fatty liver disease: A theory on lipid content, inhibited metabolism, and inflammation

Submitted/ Under review _ Review Paper

Mireille A. Edens 1

Albert K. Groen 2

Ronald P. Stolk 1

Department of Epidemiology 1 Laboratory of Pediatrics; CLDMD 2


Groningen, The Netherlands

Chapter 4.

76

ABSTRACT

Background and aims Fatty liver disease (FLD) is the most prevalent hepatic condition

worldwide. Thorough understanding of risk factors and the pathogenesis of FLD is

therefore warranted. Few complete theories on its pathogenesis have been proposed. The

aim of this paper was to critically discuss present theories on the pathogenesis of FLD, and

to arrange risk factors in an easily recognisable manner.

Methods The literature, including behavioural, genetic, and environmental factors

associated with FLD was reviewed, together with their underlying role in the pathogenesis

of FLD. Risk factors for FLD were arranged according to pathogenesis.

Results The following groups of risk factors for FLD were developed: 1) ‘risk factors for

hepatic lipid content’, 2) ‘risk factors for inhibited hepatic metabolism’, and 3) ‘risk factors

for hepatic inflammation’. Some risk factors can be placed in more than one category and

might therefore have a greater effect on FLD than others. These three categories do not

stand alone but interact, forming a downward spiral onto the development and progression

of FLD.

Conclusions We propose to modify current arrangements of risk factors for FLD. The

present arrangement of risk factors may be useful to identify people at high risk for FLD

and to initiate interventions.

Pathogenesis of fatty liver disease

77

INTRODUCTION

The histological spectrum of fatty liver disease (FLD) encompasses a wide range of various

degrees of fatty infiltration, inflammation and fibrosis 1. FLD is broadly dividable into

steatosis (fat accumulation, with or without nonspecific inflammation) and steatohepatitis

(fat accumulation plus inflammation, with or without fibrosis) 2. A liver containing both

fatty infiltration and inflammation which is neither steatosis nor severe enough to be

categorised as steatohepatitis, has been referred to as ‘intermediate’ 3 or ‘borderline’ 2

steatohepatitis. Importantly, while the FLD spectrum progresses in severity (fibrosis stage),

hepatic fat content decreases 4, 5.

Historically, alcoholism is a well known risk factor for FLD. As alcoholic FLD (AFLD)

and non-alcoholic FLD (NAFLD) may follow different clinical courses 6, both are usually

distinguished using an alcohol intake cut-off value of 20 g/d 7. Although non-alcoholic

steatosis (NAS) is often considered a benign condition, at least in the short-term, it leads to

fibrosis in an estimated 8.3% of cases 8. NASH progresses in fibrosis stage in a estimated

37.6% of cases 9. NAFLD is associated with an increased hepatic risk, i.e. cryptogenic

cirrhosis (maximum fibrosis score 2) predisposing to end stage liver disease 5, and/or

hepatocarcinoma 10. Additionally, NAFLD is associated with cardiovascular disease (CVD)

risk factors 11, 12. Longitudinal studies revealed that overall survival of patients with

NAFLD 13, 14 and NASH 14 is significantly reduced, as compared to reference populations.

CVD-related death was found to be the primary cause of death in patients with NAFLD 13,

14. CVD-related death was 7.5% in a reference population, slightly increased to 8.6% in

patients with NAS (p=ns) and significantly increased to 15.5% in patients with NASH

(p<.05 compared to the reference population) 14. Additionally, also liver-related death in

patients with NAFLD 1 and NASH 14 is significantly higher as compared to reference

populations. As the current prevalence of FLD (±30%) and its subtype NAFLD (±20%) in

the general adult Western population are alarmingly high 15, 16 and potentially increasing 17,

18, a good understanding of risk factors and the pathogenesis of FLD is needed.

The reason(s) why some patients progress while others do not is yet unclear, hence

cryptogenic. In an attempt to explain the pathogenesis of FLD a number of hypothesis have

been developed previously 19-23. The aim of this paper was to modify previous theories on

the pathogenesis of fatty liver disease.

Chapter 4.

78

RISK FACTORS AND THEORIES ON PATHOGENESIS

Previously presented risk factors

Historically, conditions associated with FLD were simply presented and not arranged, e.g.

by Lee (1989) 24. Later, several authors presented an arrangement of etiologic factors, e.g.

Mach (2000) 25 and Angulo (2002) 26. These arrangements include: ‘toxic factors’,

‘nutritional factors’, ‘endocrine factors and metabolic diseases’, and ‘other rare causes’ 25

and ‘nutritional’, ‘drugs’, ‘metabolic or genetic’, and ‘other’ 26. Another more often used

arrangement is the division between ‘primary causes’ (e.g. obesity) and ‘secondary causes’

(e.g. alcoholism or hepatitis) 27, 28. However, none of these arrangements is based on the

pathogenesis of FLD.

Previously presented theories

Previously developed theories on the pathogenesis of FLD include: the 2 step model by

Wanless ea. (1989) 19, the ‘2 hits theory’ by Day and James (1998) 20, the modified ‘2 hits

theory’ by Day (2002) 21, and the ‘four-step model’ by Wanless and Shiota (2004) 22.

Insulin is considered the first ‘hit’ or ‘step’ causing steatosis in all theories 22. Day and

James hypothesised that lipid peroxidation 20, later modified to lipid peroxidation and/or

direct lipotoxicity 21, is a second ‘hit’ that causes steatohepatitis. Wanless and Shiota

proposed that intracellular lipid toxicity or lipid peroxidation, or both, modified by alcohol,

drugs, and ischemia can be considered a second ‘hit’ that causes necrosis 22. Additionally,

they added a third and a fourth ‘step’. Step 3 being release of bulk lipid from hepatocytes

into the interstitium leading to direct and inflammatory injury to hepatic veins, and step 4

being venous obstruction with secondary collapse and ultimately fibrous septation and

cirrhosis 22. In 2007, the idea of ‘at least 3 hits’ was published, which suggests that the

pathogenesis of FLD is a complex interaction between behaviour, environment, and genetic

susceptibility 23.

Currently presented arrangement of risk factors and theory on pathogenesis

After reviewing the literature on risk factors associated with FLD, including behaviour, the

environment, and genes 23, we propose to modify current theories on the pathogenesis of

FLD. We propose to arrange risk factors in the following three categories: 1) risk factors


79

for hepatic lipid content, 2) risk factors for inhibited hepatic metabolism, and 3) risk factors

for hepatic inflammation. Figure І (simplified version) and appendix І (detailed version

including references) show an overview on risk factors arranged in these three categories.

Figure ІІ is a model on intra-hepatic pathogenesis. For each category of risk factors a brief

description of pathogenesis will be provided.

Figure І. Overview on risk factors for fatty liver disease, arranged in the following three categories: 1) risk factors for hepatic lipid

content, 2) risk factors for inhibited hepatic m

etabolism

, and 3) risk factors for hepatic inflam

mation.

For a

detailed list of risk factors and their

references, the reader is referred to appendix І.

Arrows: Closed arrows, stimulatory; interrupted

arrows, inhibitory.

Abbreviations:

BFD,

body fat

distribution;

Dets., determinants; HI; hepatic inflam

mation;

HLC, hepatic lipid content; iHM, inhibited

hepatic

metabolism

; P-I

complex,

pharmacological-interaction

complex;

SAT,

subcutaneous

adipose tissue; SIBO,

small

intestinal bacterial overgrowth; VAT, visceral

adipose tissue.

Chapter 4.

80

Figure ІІ. M

odel on risk factors for fatty liver disease showing intra-hepatic interaction

Arrows: Closed arrows, stimulatory; interrupted

arrows, inhibitory.

Signs: 1, Substrate for microsomes include both

saturated and unsaturated fatty acids; 2, Substrate

for

mitochondria

include

short-chain (<C

8),

medium-chain (C

8-C

11), long-chain (C

12-C

20),

and very long-chain (>C

20)

fatty acids; 3,

Substrate for peroxisomes include straight-chain

saturated fatty acyl-CoAs (>C

20), dicarboxylic

acids, prostanoids, C

27 bile acid interm

ediates;

▲, increase; ▼, decrease.

Abbreviations: B

H, ballooning of hepatocytes;

CARBS,

carbohydrates;

DNL,

de

novo

lipogenesis; ERS, endoplasm

atic reticular stress;

HFFAs, hepatic free fatty acids; HSC, hepatic

stellate cell; HSL, horm

one

sensitive

lipase;

IBD, inflam

matory bowel disease; IR

, insulin

receptor; L

S, lipolysis and secretion from lipid

vesicle; LP, fatty acid-induced production of

TNFα

via

a lysosomal

pathway;

MF,

‘metaflammation’; OS, oxidative

stress; PLs,

phospholipids; PPARα, peroxisome proliferator-

activated receptor alpha; SIBO, sm

all intestinal

bacterial

overload;

TGs,

triglycerides;

TI,

traditional inflam

mation response; TPs, toxic

products origination from oxidation; UO, up-

regulation of oxidation; VLDL, very low density

lipoprotein.


81

Table Ι. G

enet

ic r

isk f

acto

rs (

incl

. sy

stem

ic a

nd inte

stin

al)

thro

ughout th

e N

AFL

D s

pec

trum

Variable

Correlation

with liver fat

Increasing severity of the NAFLD spectrum →

NAS

vs.

contr

ols

NAFLD

&

vs.

contr

ols

NASH

vs.

contr

ols

NASH

vs.

non-N

ASH

#

NASH

vs.

NA

S

LIPID

Hep

atic

LPL

mR

NA

+ 2

9

▲29

Hep

atic

PPA

Rγ

mR

NA

=●

29

=●

30

Hep

atic

PPA

Rγ2

mR

NA

+

29

▲29

Hep

atic

PG

C1 m

RN

A

▼

29

Hep

atic

PPA

Rδ m

RN

A

=

●29

Hep

atic

LX

Rα m

RN

A

▲31

Hep

atic

SR

EB

P-1

c m

RN

A

=●

30 o

r ▲

31

Hep

atic

ChR

EB

P m

RN

A

▼31

Hep

atic

AC

AC

A m

RN

A

=

●29

Hep

atic

FA

S m

RN

A

=●

30,

31

Hep

atic

AC

C1 m

RN

A

▲30,

31

Hep

atic

AC

SL

4 m

RN

A

+ 2

9

▲29

Hep

atic

DG

AT

1 m

RN

A

=●

30

Hep

atic

AD

RP m

RN

A

=●

30

Hep

atic

HSL

mR

NA

▼

30

Hep

atic

SC

D1 m

RN

A

=

●29

METABOLISM

Hep

atic

PPA

Rα m

RN

A

=

○29

▼30

Hep

atic

AM

PK

mR

NA

=

●31

Hep

atic

AC

C2 m

RN

A

=●

30

Hep

atic

CPT

1a

mR

NA

=●

29

▼30

Hep

atic

UC

P2 m

RN

A *

=

●30

Hep

atic

LC

AD

mR

NA

▲

30

Hep

atic

HA

DH

α m

RN

A

▲30

Hep

atic

AC

OX

mR

NA

=

●30

Hep

atic

BO

X m

RN

A

▲30

Hep

atic

CY

P2E

1 m

RN

A

▲30

Hep

atic

CY

P4A

11 m

RN

A

▲30

Hep

atic

Cat

alas

e m

RN

A

▲30

Hep

atic

SO

D m

RN

A

▲30

Hep

atic

GSS m

RN

A

=30

INFLAMMATION

Hep

atic

TN

Fα m

RN

A

▲32

▲33,

34

SA

T T

NFα m

RN

A

▲32

Syst

emic

TN

Fα lev

el

=

35

▲33

▲35,

36

=

33 o

r ▲

35

Hep

atic

TN

Fα-R

1 m

RN

A

▲32

= 3

4

Hep

atic

TN

Fα-R

2 m

RN

A

= 3

2

= 3

4

Syst

emic

sT

NFα-R

1 lev

el

=

●35

▲

35

Syst

emic

sT

NFα-R

2 lev

el

=

●35

=

●35

Hep

atic

adip

onec

tin m

RN

A *

X 3

7

X

37

▼

34 o

r X

37

VA

T a

dip

onec

tin m

RN

A

=

○37

▼

37

=

○37

SA

T a

dip

onec

tin m

RN

A

=

○37

=

○37

=

●37

Syst

emic

adip

onec

tin lev

el

− 3

8

=○

37

▼

37

=

○34,

37

Hep

atic

adip

o-R

1 m

RN

A

=

●29 =

●37

▲

37

=

●37 o

r =

○34

Hep

atic

adip

o-R

2 m

RN

A

=

●29 =

●37

▲

37

▲

34 o

r =

●37

Chapter 4.

82

Hep

atic

MC

P1 m

RN

A

+ 2

9

▲29

Syst

emic

MC

P1 lev

el

▲

35

=

35

Hep

atic

MIP

1α m

RN

A

+ 2

9

=●

29

Hep

atic

CD

68 m

RN

A

=

●29

Syst

emic

LP

S lev

el

▲33

SIB

O

▲39

▲36

Inte

stin

al p

erm

eabil

ity

▲39

=●

36

Syst

emic

endoto

xin

lev

els

=36

Signs: +

, si

gnif

ican

t posi

tive

corr

elat

ion; −

, si

gnif

ican

t in

ver

se c

orr

elat

ion;

▲,

signif

ican

tly i

ncr

ease

d;

▼,

signif

ican

tly d

ecre

ased

; =

, si

milar

/non-

signif

ican

t; =

●,

sim

ilar

/non-s

ignif

ican

t in

crea

se;

=○

, si

milar

/non-s

ignif

ican

t dec

reas

e; X

, undet

ecta

bly

low

mR

NA

; &,

NA

FL

D i

ncl

udes

both

NA

S a

nd

NA

SH

; * u

sual

ly n

ot ex

pre

ssed

in the

liver

; #, non-N

ASH

incl

udes

both

contr

ols

and N

AS

.

Abbreviations: A

CA

CA

, ac

etyl-

coen

zym

e A

car

bxyla

se-α

; A

CC

1, ac

etyl-

CoA

car

boxyla

se 1

; A

CC

2, ac

etyl-

CoA

car

boxyla

se 2

; A

CS

L4, ac

yl-

CoA

synth

etas

e lo

ng-c

hai

n f

amily m

ember

4; A

CO

X, st

raig

ht-

chai

n a

cyl-

CoA

oxid

ase;

AD

RP

, ad

ipose

dif

fere

ntiat

ion-r

elat

ed p

rote

in (

adip

ophilin

); A

MP

K,

AM

P-a

ctiv

ated

pro

tein

kin

ase;

AT

P, ad

enosi

ne

trip

hosp

hat

e; B

OX

, bra

nch

ed-c

hai

n a

cyl-

CoA

oxid

ase;

CD

68, cl

ust

er o

f dif

fere

ntiat

ion 6

8; C

hR

EB

P,

carb

ohydra

te r

esponsi

ve

elem

ent bin

din

g p

rote

in; C

PT

1a,

car

nitin

e pal

mitoylt

ransf

eras

e 1; C

YP

2E

1, cy

toch

rom

e P

450 2

E1; C

YP

4A

11, cy

toch

rom

e

P450 4

A11; D

AG

, dia

gly

cero

l; D

GA

T1, dia

cylg

lyce

rol O

-acy

ltra

nsf

eras

e 1; E

R, en

dopla

smic

ret

iculu

m; F

AS

, fa

tty a

cid s

ynth

ase;

GS

S, glu

tath

ione

synth

etas

e; H

AD

Hα, L

3-h

ydro

xyac

yl-

CoA

deh

ydro

gen

ase

alpha;

HF

FA

, hep

atic

fre

e fa

tty a

cids;

HO

MA

, hom

eost

asis

model

ass

essm

ent; H

SL

,

horm

one

sensi

tive

lipas

e; incl

., incl

udin

g; L

CA

D, lo

ng-c

hai

n a

cyl-

CoA

deh

ydro

gen

ase;

LP

L, lipopro

tein

lip

ase;

LP

S, lipopoly

sach

arid

e-bin

din

g

pro

tein

, L

XR

α, li

ver

X r

ecep

tor

alpha;

MC

P1, m

onocy

te c

hem

oat

trac

tant pro

tein

1; M

IP1α, m

acro

phag

e in

flam

mat

ory

pro

tein

1α; m

RN

A, m

esse

nger

ribonucl

eic

acid

; n/a

, not ap

plica

ble

; P

GC

1, P

PA

Rγ

coac

tivat

or

1; P

PA

Rα, per

oxis

om

e pro

life

rato

r-ac

tivat

ed r

ecep

tor

alpha;

PP

AR

δ, per

oxis

om

e

pro

life

rato

r-ac

tivat

ed r

ecep

tor

del

ta; P

PA

Rγ2

, per

oxis

om

e pro

life

rato

r-ac

tivat

ed r

ecep

tor

gam

ma

isofo

rm 2

; S

CD

1, st

earo

yl-

coen

zym

e A

des

atura

se;

SO

D, su

per

oxid

e dis

muta

se; S

RE

BP

-1c,

ste

rol re

gula

tory

ele

men

t bin

din

g p

rote

in 1

c; U

CP

2, unco

upling p

rote

in 2

; vs.

, ver

sus.


83

Chapter 4.

84

PATHOGENESIS OF FATTY LIVER DISEASE

CATEGORY 1. RISK FACTORS FOR HEPATIC LIPID CONTENT

Generally, in the case of more hepatic lipid availability (uptake and de novo synthesis) on

one hand, compared to hepatic lipid disposal (oxidation, and excretion & secretion) on the

other hand, lipids are stored within hepatocytes as neutral triglycerides (TGs).

Uptake and de novo synthesis

Lipid sources and pathways

There are three nutrition-related sources regarding hepatic lipid origin with four pathways:

1) hepatic uptake of dietary fats via chylomicrons (a) and the NEFA pool, i.e. spillover (b),

2) hepatic uptake of dietary carbohydrates followed by de novo lipogenesis, and 3) hepatic

uptake of lipids from the NEFA pool with a peripheral origin (“old fat”) 40. Endotoxins,

which are complex lipopolysaccharides, could be considered another likely minor source of

lipids and carbohydrates 41, 42, but will be discussed in category 3.

Studies on stable isotopes administered in food and/or infused into the blood, have revealed

insight in hepatic lipid origin in the case of suspected NAFLD (by aminotransferases) 40.

After four days of labelling followed by a liver biopsy in the fasting state, the contributions

of peripheral NEFA, de novo lipogenesis, and dietary fats to hepatic fat content were 59%

(range 45.1% - 74.3%), 26.1% (range 12.7% - 37%), and 14.9% (range 4.3% - 28%),

respectively 40. Thus on average, NEFA originating from peripheral fat depots might be the

most important lipid source for hepatic fat content, followed by de novo lipogenesis and

thereafter dietary fat 40. However, the wide ranges may reflect large inter-individual

differences.

Peripheral lipolyis

The first step in lipolysis from adipose tissue is catalysed by adipose triglyceride lipase

(ATGL), leading to formation of diaglycerol (DAG) and free fatty acid. The DAG is

subsequently cleaved by hormone sensitive lipase (HSL), which is inhibited by insulin 43.

Access of HSL is controlled by perilipin 43. Perilipin is a member of endoplasmatic

reticular-derived PAT protein family (PAT is named after perilipin, adipophilin/ADRP, and

TIP47, and additionally includes S3-12) 43. Thus in the case of impaired insulin production

(e.g. pancreatic β-cell dysfunction or fasting), insulin resistance, and/or impaired perilipin

more lipolysis from adipose tissue may take place. Both visceral adipose tissue (VAT) and


85

subcutaneous adipose tissue (SAT) correlate with hepatic fat content, but only VAT is

independently associated with hepatic TG content, at least in non-diabetic subjects 38. These

results are in agreement with the ‘portal hypothesis’ which suggests that peripheral fat

depots draining in the portal vein, notably VAT, contribute to hepatic fat content the most

44. Moreover, the fatty acid flow from SAT to the liver is intercepted by muscles, enabling a

protective role of exercise for FLD 45-47.

Dietary composition and de novo lipogenesis

Total energy intake is associated with NAFLD 48. The role of dietary composition for

NAFLD was recently reviewed by Le and Bortolotti (2008) 49. The authors state that the

role of carbohydrates in the pathogenesis of NAFLD has been clearly shown, but that the

role of lipids remains controversial 49. This might be attributable to de novo lipolysis on

average having a larger contribution than dietary fat 40.

Expression of several genes related to de novo lipogenesis (LXRα, SREBP-1c, ChREBP) is

changed in the case of NAFLD (table I). Stable isotope studies have provided evidence that

de novo lipogenesis is chronically elevated in subjects with NAFLD. In controls, the

contribution of de novo lipogenesis to VLDLTG is <5% in the fasting state 50, 51 with

significant postprandial elevation 50. In the case of NAFLD, the contribution of de novo

lipogenesis to VLDLTG is 15% - 25% in the fasting state 40, 51, with no postprandial

elevation 40, 51.

Excretion, secretion and storage

Hepatic free fatty acids (HFFAs) can be excreted in bile as part of phospholipid, secreted

by VLDL, or oxidised. Little is known about the magnitude of lipid excretion by bile in

human subjects with NAFLD 52, and oxidation will be discussed in category 2.

Lipid secretion in VLDL

Due to hepatic insulin resistance, secretion of VLDL (containing particularly TGs but also

cholesterol and phospholipids) may be increased 12, however, this increased secretion may

not be high enough to prevent FLD. In the case of rare storage and secretion disorders, FLD

is likely to occur 53, 54.

Hepatic TG synthesis and storage

HFFAs which are not oxidised, excreted in bile, or secreted in VLDL may be stored as

neutral TGs. Molecular mechanisms underlying TG synthesis and storage have extensively

been reviewed previously 43, 55. Expression of several genes related to TG synthesis (FAS,

Chapter 4.

86

ACC1, DGAT1) and TG storage (ADRP) is changed in the case of NAFLD (table I). Lipids

are stored as neutral TGs while isolated from the cytoplasm by a phospholipid monolayer

forming ‘lipid vesicles’, which can be either microvesicular or macrovesicular 43.

Consequences of elevated hepatic lipid levels

Lipotoxicity and lipoapoptosis

Elevated cellular lipid levels predispose to lipotoxicity (metabolically relevant cellular

dysfunction) which could lead to necrosis, and lipoapoptosis (programmed cell-death) 56.

When discussing lipotoxicity and lipoapoptosis it may be relevant to distinguish HFFAs

and fatty acids stored as neutral TGs within lipid vesicles. HFFAs may exert direct toxicity

by uncoupling phosphorylation and indirect toxicity via production of reactive oxygen

species (ROS) 57, 58, and may induce the production of inflammatory cytokines as well 59-62.

Hepatic microcirculation

Increased TG storage within lipid vesicles may cause ballooning of hepatocytes, which is

hypothesised to pinch off hepatic microcirculation 63. Additionally, in the case of NAFLD

secretion of several atherogenic factors (e.g. lipids and glucose) may be increased 12

predisposing to impaired hepatic microcirculation as well. Impaired microcirculation may

have consequences regarding both transport to hepatocytes (notably supply of oxygen) and

from hepatocytes (draining off any unfavourable products).

CATEGORY 2. RISK FACTORS FOR INHIBITED HEPATIC METABOLISM

Molecular mechanisms of hepatic lipid oxidation have been extensively reviewed

previously 55, 58, thus will be described here very briefly. Oxidation in the liver takes place

in mitochondria, peroxisomes, and endoplasmatic reticular microsomes. The latter two are

usually minor oxidation systems but may become more important in the case of FLD 55.

Many key enzymes involved in fatty acid oxidation are regulated by peroxisome

proliferator-activated receptor alpha (PPARα), excluding enzymes of the non-classical

peroxisomal pathway 55.

Metabolic burden and toxic by-products

In the case of NAFLD, PPARα is downregulated 30, however several enzymes involved in

fatty oxidation localised in mitochondria, peroxisomes and microsomes are upregulated

(table Ι). However, given the increased amount of hepatic substrate, oxidation may be

substantial. Oxidation delineates production of toxic by-products, e.g. ROS (superoxide and


87

hydrogen peroxide, by all oxidation systems), malondialdehyde and 4-hydroxynonenal (by

peroxisomes and microsomes), and dicarboxylic acids (by microsomes). The higher the

metabolic burden to oxidation systems the higher the amount of toxic by-products. ROS are

neutralised by anti-oxidants, e.g. glutathione, and enzymes; superoxide dismutase (SOD)

and catalase. Dicarboxylic acids, which are toxic for mitochondria, are oxidised by

peroxisomes 55. Some genes related to neutralisation pathways (glutathione synthetase,

SOD, catalase) are upregulated in the case of NAFLD (table Ι) 30, which suggests increased

oxidation and oxidative stress.

Drugs

The liver plays an important role in metabolism and excretion of many drugs and toxins.

Although uncommon, hence idiosyncratic, drug-induced liver injury can be caused by both

prescription drugs 64, 65 (iatrogenic-induced) and over-the-counter drugs 64. Idiosyncratic

drug-induced liver injury may occur particularly in those drugs that have significant hepatic

metabolism 65. Some herbals 64 and environmental toxins and petro-chemicals 66 are

associate with liver injury as well, but their mechanisms are unknown.

Drug excretion

Xenobiotics and endobiotics can be excreted out of hepatocytes into bile or blood (followed

by urinary excretion) by efflux transporters 67. Efflux transporters include e.g. multidrug

resistance-associated proteins (Mrps), bile salt excretory protein (Bsep), and breast cancer

resistance protein (Bcrp) 67. Little is known about the magnitude of drug excretion in

human NAFLD. An animal study on high fat diet-induced NAS and methionine- and

choline-deficient diet-induced NASH and additional Acetaminophen administration may

suggest changes in drug excretion 67. Expression of several efflux transporters was changed,

and excretion shifted from biliary to urinary.

Organelles

Mitochondria

Mitochondrial fatty acid oxidation primarily involves the progressive shortening of straight-

chain fatty acids into acetyl-CoA for generation of adenosine triphosphate (ATP) 55.

Hepatocytes are dependent on aerobic mitochondrial metabolism for production of ATP 58,

which necessitates a proper hepatic microcirculation 63. Although PPARα and transport of

lipids into mitochondria by CPT1a are downregulated, several genes related to

mitochondrial oxidation (LCAD, HADHα, UCP2) are upregulated in the case of NAFLD

Chapter 4.

88

(table Ι). A severely impaired recovery of fructose-induced ATP-depletion in subjects with

NASH compared to controls has been reported 68. NASH, might in some subjects be

associated with structural mitochondrial changes, i.e. enlargement (mega mitochondria) and

development of crystalline inclusions similar to brown adipose tissue 58. Several factors are

believed to underlie inhibition of mitochondrial ATP production, i.e. metabolic burden (via

oxidative stress), TNFα (via increasing mitochondrial membrane permeability), alcohol and

several drugs 57, 69, and impaired hepatic microcirculation (impaired oxygen supply, and

impaired discharge of metabolites) 63. The role of microcirculation in NAFLD was

reviewed by Farrell ea. (2008) 63. The authors state that there is evidence for impaired

perfusion in NAFLD 63, likely caused by ballooning/swelling of hepatocytes 63 and/or the

atherosclerotic milieu 12.

Peroxisomes

Peroxisomes metabolize relatively more toxic and biologically active molecules and hereby

play a role in detoxification 55. Although PPARα is downregulated, ACOX (regulated by

PPARα) and BOX (not regulated by PPARα) are upregulated in the case of NAFLD (table

Ι).

Endoplasmatic reticular microsomes

Microsomes usually play a minor role in oxidation 55. Several genes related to microsomal

oxidation (CYP2E1, CYP4A11) are upregulated in the case of NAFLD (table Ι) 30. Besides

the increased presence of substrate for microsomal oxidation, this might be attributable to

drug use as well. The pathogenesis of drug-induced liver injury is largely unknown, but

phase Ι and/or phase ΙΙ reactions in microsomes are believed to play a role 65. Phase Ι

reactions are catalysed by CYP 450 4A enzymes, predisposing to reactive drug metabolites

65. Phase ΙΙ reactions are involved in detoxifying these toxic metabolites 65.

Consequences of (impaired) hepatic metabolism

Decreased ATP production may predispose to lipid storage, and liver injury and/or necrosis

68. If toxic by-products (either lipid-derived or drug-derived) outweigh neutralisation/

detoxification, organelle damage may occur. Organelle damage may lead to decreased lipid

oxidation or even necrosis. As a response a ‘traditional inflammation’ response will set in,

which will be described in the inflammation paragraph.


89

CATEGORY 3. RISK FACTORS FOR HEPATIC INFLAMMATION

Inflammation is a key feature of obesity and type 2 diabetes 70, and by definition a

characteristic of advanced NAFLD, i.e. steatohepatitis. Two subclasses of inflammation

distinguished are ‘traditional inflammation’ (short-term response to deal with injury) and

‘metaflammation’ (metabolically triggered inflammation or chronic low-grade

inflammation) 70. Infections may trigger an inflammatory response as well. In addition,

several drugs may induce allergic reactions according to the hapten and prohapten concept

(binding of small pharmacological molecules to larger carriers to gain immunogenicity)

and/or the pharmacological interaction (PI) concept (direct binding of drugs with immune

receptors) 71.

‘Metaflammation’

The primary transcription factor nuclear factor kappa-beta (NF-κB) and its upstream

activator IκB kinase (IKKβ) are considered a ‘master-switch’ for coordination of

‘metaflammation’ in the liver, peripheral adipose tissue, and the central nervous system 59.

IKKβ/NF-κB controls the production of several inflammatory cytokines and other

inflammation mediators 59. Additionally, IKKβ/NF-κB is considered to play a vital role in

inhibition of apoptosis induced by tumour necrosis factor alpha (TNFα) 72, 73.

Pro-inflammatory imbalance

Of all inflammatory cytokines, TNFα has been studied the most and was recently reviewed

70, 74. Cytokines including TNFα are potent activators of IKKβ/NF-κB 59, 75, suggesting

auto-up regulation. Of all anti-inflammatory adipocytokines, the adipose tissue-derived

hormone adiponectin has been studied the most and was recently reviewed by several

authors 76-79. Some studies in animals 80 and humans 81, 82 have provided evidence that

TNFα down regulates adiponectin (and vice versa) at both genetic and cellular level.

Cytokine origin and uptake

TNFα is produced by both cells of the immune system 83, 84 and by fatty acids in either

peripheral adipose tissue 32, 85 or ectopic tissue 60, 61. Likely, most TNFα is produced when

tissues are infiltrated by macrophages. Additionally, certain TNFα polymorphisms have

been associated with NAFLD 86. Both polymorphisms encoding adiponectin receptor 1 87

and adiponectin receptor 2 88 have been found to be associated with hepatic fat content.

Chapter 4.

90

Hepatic-derived cytokines

Hepatic levels of TNFα and adiponectin expression are shown in table І. Throughout the

spectrum, the liver increasingly upregulates production of TNFα 32-34. Adiponectin

expression is either low 34 or undetectable 37. Adiponectin expression is lower in NASH

compared to NAS 34. Additionally, hepatic fat content correlates with hepatic levels of

MCP1 and MIP1α 29, predisposing to hepatic infiltration by monocytes and macrophages.

Peripheral adipose tissue-derived cytokines

The total amount of peripheral fat is increased in the case of NAFLD 38. TNFα expression

is much higher in VAT than SAT 37. In contrast adiponectin expression is much higher in

SAT than VAT 37. However, even in SAT, TNFα expression in the case of NASH is

increased 32. Macrophage infiltration is twice as high in VAT than SAT 85. Additionally,

the number of macrophages infiltrated in VAT, and not SAT, is significantly associated

with steatosis grade, NASH score, fibrosis stage, and fibro-inflammation index 84.

Systemic levels, receptor polymorphisms, and hepatic uptake

Hepatic and systemic levels of TNFα (receptor) and adiponectin (receptor) expression are

shown in table І. In the case of NAFLD, systemic TNFα levels are elevated 33, 35, 36, whereas

systemic adiponectin levels tend to be decreased 34, 37. Systemic soluble TNFα-R1 and

soluble TNFα-R2 tend to be slightly elevated 35. Hepatic TNFα-R1 and TNFα-R2 tend to be

upregulated in NAFLD 32, 34. Results on hepatic adipo-R1 and adipo-R2 are inconclusive 34,

37.

Infections

Endotoxins and the intestinal barrier

The intestine hosts a wide variety of bacteria 41, 42. Gram-negative bacteria have an outer

membrane encorporating endotoxins, which are complex lipopolysaccharides. Depending

on the intestinal barrier and amount of lipopolysacharide-binding protein (a hepatic-derived

acute phase protein), endotoxins are transported to the liver via the portal vein. The

intestinal barrier may be affected by e.g. alcohol 89 and inflammatory bowel disease 39,

predisposing to increased intestinal permeability likely caused by disruption of intestinal

tight junctions 39. Lipopolysacharide-binding protein is a hepatic-derived acute phase

protein. Animal models have shown that the liver is vulnerable to endotoxin-induced liver

damage 73. The damage occurred despite induction of NF-κB, having several anti-apoptotic


91

transcriptional targets 72, 73, indicating that endotoxin-induced liver damage is likely caused

by necrosis.

Subjects with NAFLD and NASH have an increased prevalence of small intestinal bacterial

overload (table Ι) 36, 39. Subjects with NAFLD 39 and NASH 36 have a non-significantly 36 or

significantly 39 increased intestinal permeability, by lactulose-rhamnose test and by urinary

excretion test, respectively. Subjects with increased intestinal permeability had an increased

prevalence of moderate or severe steatosis, instead of mild steatosis 39. Additionally,

systemic levels of lipopolysacharide-binding protein are elevated in subjects with NAFLD

33. Systemic endotoxin levels were similar in NASH compared to controls, however this

could be attributed by hepatic uptake (figure Ι) 36.

Consequences of hepatic inflammation

Whereas adiponectin has pleiotropic favourable effects 76-79, its down regulator TNFα 80-82

has pleitropic unfavourable opposite effects 70, 74. Adiponectin is associated with insulin

sensitivity and stimulates oxidation by sequential activation of AMPK, MAPK and PPARα

(determined in muscle cells) 90. TNFα causes insulin resistance 91, 92 and impairs oxidation

as well via several pathways. Firstly, TNFα down regulates adiponectin 80-82 and therefore

PPARα 90. Secondly, TNFα increases mitochondrial membrane permeability which

predisposes to cytochrome C-induced apoptosis 72, 73.

Chapter 4.

92

DISCUSSION

This paper reviews the pathogenesis of FLD, leading to the proposition to modify present

theories on the pathogenesis of FLD and to arrange risk factors for FLD based on the

pathogenesis of FLD. Risk factors for FLD are arranged into the following three categories:

1) risk factors for hepatic lipid content, 2) risk factors for inhibited hepatic metabolism, and

3) risk factors for hepatic inflammation. Some risk factors can be placed in more than one

category and might therefore have a greater effect on FLD than others. Importantly, these

three categories do not stand alone but interact, forming a downward spiral onto the

development and progression of FLD.

COMPARISON WITH OTHER THEORIES ON PATHOGENESIS

All previously presented theories 19-22 propose that insulin is a first ‘hit’ that causes steatosis

22. Insulin does facilitate transport of fat and glucose (followed by de novo lipogenesis)

across hepatic cell membranes, but there are non-insulin dependent ways as well. NEFA,

e.g. origination from peripheral fat depots 40, can flip-flop across cell membranes amongst

others 55, 93. Glucose can pass through the non-insulin responsive glucose transporter 2

(GLUT-2). Additionally, FLD is associated with insulin resistance (and hyperinsulinaemia)

12, which may induce the overproduction of VLDL in the case of FLD 94.

Regarding the previously presented second ‘hit(s)’ causing necrosis 19-22, we firstly believe

that the amount of factors should be expanded with inflammation (TNFα) and hepatic

microcirculation 63. Secondly, we believe that particularly in the case of ‘non-toxic’

dosages, initially metabolism will inhibited (rather than immediate necrosis), predisposing

to increased fat storage. Only in the case of prolonged influence of risk factors on

organelles, or in the case of severely toxic factors on organelles, necrosis will set in.

Therefore the present paper suggests that the pathogenesis of the FLD spectrum (i.e. in the

cases with non-severe toxic risk factors) is a gradual process, rather than an abrupt

consequence of separate ‘hits’. A gradual progression would also be in agreement with the

existence of ‘intermediate’ or ‘borderline’ NASH 2, 3.

Regarding the previously presented third and fourth step 22, the present paper proposes that

(besides collapse of veins 22) atherosclerosis causes reduced transport and draining causing


93

damage/necrosis to the liver. This is followed by an inflammatory response (‘traditional

inflammation’) 70, which results in repair by means of fibrosis.

HEPATIC ADAPTATION TO INCREASED FAT LOAD

Several factors may be produced and/or expressed (mRNA) in fatty livers, e.g. crystalline

inclusions 58, UCP2 58, and adiponectin 34 (table І), whereas these are usually found in

adipose tissue only. These findings may imply that the fat laden hepatocyte both

microscopically and functionally tends to resemble an adipocyte 58. This also strengthens

the hypothesis of the insulin responsive GLUT-4 being present in the case of NAFLD (in

stead of the usual GLUT-2), previously hypothesized to explain the hepatic insulin

resistance present in the case of NAFLD 12. Thus, the potential of the liver to adapt to its

environment 58 might underlie the increased CVD risk 11, 12.

LIMITATIONS OF THIS REVIEW

The studies presented in this paper include various study populations, various study

designs, and various diagnosis modalities. All these study characteristics are given in the

legend to appendix І.

Study populations and designs

Appendix І includes animal models and human populations (both healthy subjects and

patients). This could cause problems regarding ‘generalisation’ to other populations.

Appendix І includes cross-sectional and longitudinal (both prospective and retrospective)

studies. In order to determine a ‘cause-effect’, prospective longitudinal studies may be

needed.

Diagnosis modalities

The entire FLD spectrum can currently be diagnosed by histological analysis only. The first

proposal for scoring histological FLD lesions was published in 1999 95 and updated in 2005

2. Hepatic fat content (thus not inflammation and fibrosis), can be determined by imaging,

but it is important to note that low hepatic fat content can indicate absence or an early place

within the FLD spectrum, but also a progressed place within the FLD spectrum 5, 96.

Aminotransferases are produced and present in many tissues 97, 98, thus are not liver

specific, which means that elevated levels can be attributed to other than liver disease 99, 100.

Chapter 4.

94

CONCLUSION

We propose to modify current theories on the pathogenesis of FLD and to modify current

arrangements of risk factors for FLD. The present arrangement of risk factors may be useful

to identify people at high risk for FLD and to initiate interventions.


95

Reference List (1) Matteoni CA, Younossi ZM, Gramlich T, Boparai N, Liu YC, McCullough AJ.

Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology 1999 June;116(6):1413-9.


(3) Neuschwander-Tetri BA, Brunt EM, Wehmeier KR, Oliver D, Bacon BR. Improved nonalcoholic steatohepatitis after 48 weeks of treatment with the PPAR-gamma ligand rosiglitazone. Hepatology 2003 October;38(4):1008-17.




(7) Bedogni G, Bellentani S. Fatty liver: how frequent is it and why? Ann Hepatol 2004 April;3(2):63-5.


(9) Argo CK, Northup PG, Al Osaimi AM, Caldwell SH. Systematic review of risk factors for fibrosis progression in non-alcoholic steatohepatitis. J Hepatol 2009 August;51(2):371-9.

(10) Bugianesi E, Leone N, Vanni E, Marchesini G, Brunello F, Carucci P, Musso A, De Paolis P, Capussotti L, Salizzoni M, Rizzetto M. Expanding the natural history of nonalcoholic steatohepatitis: from cryptogenic cirrhosis to hepatocellular carcinoma. Gastroenterology 2002 July;123(1):134-40.



Chapter 4.

96

(13) Adams LA, Lymp JF, St Sauver J, Sanderson SO, Lindor KD, Feldstein A, Angulo P. The natural history of nonalcoholic fatty liver disease: a population-based cohort study. Gastroenterology 2005 July;129(1):113-21.







(19) Wanless IR, Bargman JM, Oreopoulos DG, Vas SI. Subcapsular steatonecrosis in response to peritoneal insulin delivery: a clue to the pathogenesis of steatonecrosis in obesity. Mod Pathol 1989 March;2(2):69-74.

(20) Day CP, James OF. Steatohepatitis: a tale of two "hits"? Gastroenterology 1998 April;114(4):842-5.

(21) Day CP. Pathogenesis of steatohepatitis. Best Pract Res Clin Gastroenterol 2002 October;16(5):663-78.

(22) Wanless IR, Shiota K. The pathogenesis of nonalcoholic steatohepatitis and other fatty liver diseases: a four-step model including the role of lipid release and hepatic venular obstruction in the progression to cirrhosis. Semin Liver Dis 2004 February;24(1):99-106.

(23) Charlton M. Noninvasive indices of fibrosis in NAFLD: starting to think about a three-hit (at least) phenomenon. Am J Gastroenterol 2007 February;102(2):409-11.


97

(24) Lee RG. Nonalcoholic steatohepatitis: a study of 49 patients. Hum Pathol 1989 June;20(6):594-8.

(25) Mach T. Fatty liver--current look at the old disease. Med Sci Monit 2000 January;6(1):209-16.

(26) Angulo P. Nonalcoholic fatty liver disease. N Engl J Med 2002 April 18;346(16):1221-31.

(27) Sanyal AJ. AGA technical review on nonalcoholic fatty liver disease. Gastroenterology 2002 November;123(5):1705-25.

(28) Adams LA, Lindor KD. Nonalcoholic fatty liver disease. Ann Epidemiol 2007 November;17(11):863-9.

(29) Westerbacka J, Kolak M, Kiviluoto T, Arkkila P, Siren J, Hamsten A, Fisher RM, Yki-Jarvinen H. Genes involved in fatty acid partitioning and binding, lipolysis, monocyte/macrophage recruitment, and inflammation are overexpressed in the human fatty liver of insulin-resistant subjects. Diabetes 2007 November;56(11):2759-65.

(30) Kohjima M, Enjoji M, Higuchi N, Kato M, Kotoh K, Yoshimoto T, Fujino T, Yada M, Yada R, Harada N, Takayanagi R, Nakamuta M. Re-evaluation of fatty acid metabolism-related gene expression in nonalcoholic fatty liver disease. Int J

Mol Med 2007 September;20(3):351-8.

(31) Higuchi N, Kato M, Shundo Y, Tajiri H, Tanaka M, Yamashita N, Kohjima M, Kotoh K, Nakamuta M, Takayanagi R, Enjoji M. Liver X receptor in cooperation with SREBP-1c is a major lipid synthesis regulator in nonalcoholic fatty liver disease. Hepatol Res 2008 November;38(11):1122-9.

(32) Crespo J, Cayon A, Fernandez-Gil P, Hernandez-Guerra M, Mayorga M, Dominguez-Diez A, Fernandez-Escalante JC, Pons-Romero F. Gene expression of tumor necrosis factor alpha and TNF-receptors, p55 and p75, in nonalcoholic steatohepatitis patients. Hepatology 2001 December;34(6):1158-63.

(33) Ruiz AG, Casafont F, Crespo J, Cayon A, Mayorga M, Estebanez A, Fernadez-Escalante JC, Pons-Romero F. Lipopolysaccharide-binding protein plasma levels and liver TNF-alpha gene expression in obese patients: evidence for the potential role of endotoxin in the pathogenesis of non-alcoholic steatohepatitis. Obes Surg 2007 October;17(10):1374-80.

(34) Kaser S, Moschen A, Cayon A, Kaser A, Crespo J, Pons-Romero F, Ebenbichler CF, Patsch JR, Tilg H. Adiponectin and its receptors in non-alcoholic steatohepatitis. Gut 2005 January;54(1):117-21.

Chapter 4.

98

(35) Abiru S, Migita K, Maeda Y, Daikoku M, Ito M, Ohata K, Nagaoka S, Matsumoto T, Takii Y, Kusumoto K, Nakamura M, Komori A, Yano K, Yatsuhashi H, Eguchi K, Ishibashi H. Serum cytokine and soluble cytokine receptor levels in patients with non-alcoholic steatohepatitis. Liver Int 2006 February;26(1):39-45.

(36) Wigg AJ, Roberts-Thomson IC, Dymock RB, McCarthy PJ, Grose RH, Cummins AG. The role of small intestinal bacterial overgrowth, intestinal permeability, endotoxaemia, and tumour necrosis factor alpha in the pathogenesis of non-alcoholic steatohepatitis. Gut 2001 February;48(2):206-11.

(37) Nannipieri M, Cecchetti F, Anselmino M, Mancini E, Marchetti G, Bonotti A, Baldi S, Solito B, Giannetti M, Pinchera A, Santini F, Ferrannini E. Pattern of expression of adiponectin receptors in human liver and its relation to nonalcoholic steatohepatitis. Obes Surg 2009 April;19(4):467-74.

(38) Westerbacka J, Corner A, Tiikkainen M, Tamminen M, Vehkavaara S, Hakkinen AM, Fredriksson J, Yki-Jarvinen H. Women and men have similar amounts of liver and intra-abdominal fat, despite more subcutaneous fat in women: implications for sex differences in markers of cardiovascular risk. Diabetologia 2004 August;47(8):1360-9.

(39) Miele L, Valenza V, La Torre G, Montalto M, Cammarota G, Ricci R, Masciana R, Forgione A, Gabrieli ML, Perotti G, Vecchio FM, Rapaccini G, Gasbarrini G, Day CP, Grieco A. Increased intestinal permeability and tight junction alterations in nonalcoholic fatty liver disease. Hepatology 2009 June;49(6):1877-87.

(40) Donnelly KL, Smith CI, Schwarzenberg SJ, Jessurun J, Boldt MD, Parks EJ. Sources of fatty acids stored in liver and secreted via lipoproteins in patients with nonalcoholic fatty liver disease. J Clin Invest 2005 May;115(5):1343-51.

(41) Cani PD, Delzenne NM, Amar J, Burcelin R. Role of gut microflora in the development of obesity and insulin resistance following high-fat diet feeding. Pathol Biol (Paris) 2008 July;56(5):305-9.

(42) Musso G, Gambino R, Cassader M. Gut microbiota as a regulator of energy homeostasis and ectopic fat deposition: mechanisms and implications for metabolic disorders. Curr Opin Lipidol 2010 February;21(1):76-83.

(43) Anderson N, Borlak J. Molecular mechanisms and therapeutic targets in steatosis and steatohepatitis. Pharmacol Rev 2008 September;60(3):311-57.

(44) Abate N, Garg A. Heterogeneity in adipose tissue metabolism: causes, implications and management of regional adiposity. Prog Lipid Res 1995;34(1):53-70.

(45) Perseghin G, Lattuada G, De Cobelli F, Ragogna F, Ntali G, Esposito A, Belloni E, Canu T, Terruzzi I, Scifo P, Del Maschio A, Luzi L. Habitual physical activity


99

is associated with intrahepatic fat content in humans. Diabetes Care 2007 March;30(3):683-8.

(46) Zelber-Sagi S, Nitzan-Kaluski D, Goldsmith R, Webb M, Zvibel I, Goldiner I, Blendis L, Halpern Z, Oren R. Role of leisure-time physical activity in nonalcoholic fatty liver disease: a population-based study. Hepatology 2008 December;48(6):1791-8.

(47) Kantartzis K, Thamer C, Peter A, Machann J, Schick F, Schraml C, Konigsrainer A, Konigsrainer I, Krober S, Niess A, Fritsche A, Haring HU, Stefan N. High cardiorespiratory fitness is an independent predictor of the reduction in liver fat during a lifestyle intervention in non-alcoholic fatty liver disease. Gut 2009 September;58(9):1281-8.

(48) Vilar L, Oliveira CP, Faintuch J, Mello ES, Nogueira MA, Santos TE, Alves VA, Carrilho FJ. High-fat diet: a trigger of non-alcoholic steatohepatitis? Preliminary findings in obese subjects. Nutrition 2008 November;24(11-12):1097-102.

(49) Le KA, Bortolotti M. Role of dietary carbohydrates and macronutrients in the pathogenesis of nonalcoholic fatty liver disease. Curr Opin Clin Nutr Metab Care 2008 July;11(4):477-82.

(50) Timlin MT, Parks EJ. Temporal pattern of de novo lipogenesis in the postprandial state in healthy men. Am J Clin Nutr 2005 January;81(1):35-42.

(51) Diraison F, Moulin P, Beylot M. Contribution of hepatic de novo lipogenesis and reesterification of plasma non esterified fatty acids to plasma triglyceride synthesis during non-alcoholic fatty liver disease. Diabetes Metab 2003 November;29(5):478-85.

(52) Trauner M, Claudel T, Fickert P, Moustafa T, Wagner M. Bile acids as regulators of hepatic lipid and glucose metabolism. Dig Dis 2010;28(1):220-4.

(53) Decarlis S, Agostoni C, Ferrante F, Scarlino S, Riva E, Giovannini M. Combined hyperlipidaemia as a presenting sign of cholesteryl ester storage disease. J Inherit

Metab Dis 2009 February 15.

(54) Sankatsing RR, Fouchier SW, de Haan S, Hutten BA, de Groot E, Kastelein JJ, Stroes ES. Hepatic and cardiovascular consequences of familial hypobetalipoproteinemia. Arterioscler Thromb Vasc Biol 2005 September;25(9):1979-84.

(55) Musso G, Gambino R, Cassader M. Recent insights into hepatic lipid metabolism in non-alcoholic fatty liver disease (NAFLD). Prog Lipid Res 2009 January;48(1):1-26.

Chapter 4.

100

(56) Kusminski CM, Shetty S, Orci L, Unger RH, Scherer PE. Diabetes and apoptosis: lipotoxicity. Apoptosis 2009 May 8.

(57) Fromenty B, Pessayre D. Inhibition of mitochondrial beta-oxidation as a mechanism of hepatotoxicity. Pharmacol Ther 1995;67(1):101-54.

(58) Caldwell SH, Chang CY, Nakamoto RK, Krugner-Higby L. Mitochondria in nonalcoholic fatty liver disease. Clin Liver Dis 2004 August;8(3):595-617, x.

(59) Cai D. NFkappaB-mediated metabolic inflammation in peripheral tissues versus central nervous system. Cell Cycle 2009 August 15;8(16):2542-8.

(60) Feldstein AE, Werneburg NW, Canbay A, Guicciardi ME, Bronk SF, Rydzewski R, Burgart LJ, Gores GJ. Free fatty acids promote hepatic lipotoxicity by stimulating TNF-alpha expression via a lysosomal pathway. Hepatology 2004 July;40(1):185-94.

(61) Li Z, Berk M, McIntyre TM, Gores GJ, Feldstein AE. The lysosomal-mitochondrial axis in free fatty acid-induced hepatic lipotoxicity. Hepatology 2007 December 20.

(62) Malhi H, Gores GJ. Molecular mechanisms of lipotoxicity in nonalcoholic fatty liver disease. Semin Liver Dis 2008 November;28(4):360-9.

(63) Farrell GC, Teoh NC, McCuskey RS. Hepatic microcirculation in fatty liver disease. Anat Rec (Hoboken ) 2008 June;291(6):684-92.

(64) Navarro VJ, Senior JR. Drug-related hepatotoxicity. N Engl J Med 2006 February 16;354(7):731-9.

(65) Lammert C, Bjornsson E, Niklasson A, Chalasani N. Oral medications with significant hepatic metabolism at higher risk for hepatic adverse events. Hepatology 2009 September 14.

(66) Cotrim HP, De Freitas LA, Freitas C, Braga L, Sousa R, Carvalho F, Parana R, Santos-Jesus R, Andrade Z. Clinical and histopathological features of NASH in workers exposed to chemicals with or without associated metabolic conditions. Liver Int 2004 April;24(2):131-5.

(67) Lickteig AJ, Fisher CD, Augustine LM, Aleksunes LM, Besselsen DG, Slitt AL, Manautou JE, Cherrington NJ. Efflux transporter expression and acetaminophen metabolite excretion are altered in rodent models of nonalcoholic fatty liver disease. Drug Metab Dispos 2007 October;35(10):1970-8.

(68) Cortez-Pinto H, Chatham J, Chacko VP, Arnold C, Rashid A, Diehl AM. Alterations in liver ATP homeostasis in human nonalcoholic steatohepatitis: a pilot study. JAMA 1999 November 3;282(17):1659-64.


101

(69) Labbe G, Pessayre D, Fromenty B. Drug-induced liver injury through mitochondrial dysfunction: mechanisms and detection during preclinical safety studies. Fundam Clin Pharmacol 2008 August;22(4):335-53.

(70) Hotamisligil GS. Inflammation and metabolic disorders. Nature 2006 December 14;444(7121):860-7.

(71) Pichler WJ. Pharmacological interaction of drugs with antigen-specific immune receptors: the p-i concept. Curr Opin Allergy Clin Immunol 2002 August;2(4):301-5.

(72) Wullaert A, Heyninck K, Beyaert R. Mechanisms of crosstalk between TNF-induced NF-kappaB and JNK activation in hepatocytes. Biochem Pharmacol 2006 October 30;72(9):1090-101.

(73) Yang S, Lin H, Diehl AM. Fatty liver vulnerability to endotoxin-induced damage despite NF-kappaB induction and inhibited caspase 3 activation. Am J Physiol

Gastrointest Liver Physiol 2001 August;281(2):G382-G392.

(74) Ryden M, Arner P. Tumour necrosis factor-alpha in human adipose tissue -- from signalling mechanisms to clinical implications. J Intern Med 2007 October;262(4):431-8.

(75) Baeuerle PA, Henkel T. Function and activation of NF-kappa B in the immune system. Annu Rev Immunol 1994;12:141-79.

(76) Lafontan M, Viguerie N. Role of adipokines in the control of energy metabolism: focus on adiponectin. Curr Opin Pharmacol 2006 December;6(6):580-5.

(77) Mendez-Sanchez N, Chavez-Tapia NC, Zamora-Valdes D, Uribe M. Adiponectin, structure, function and pathophysiological implications in non-alcoholic fatty liver disease. Mini Rev Med Chem 2006 June;6(6):651-6.

(78) Guerre-Millo M. Adiponectin: an update. Diabetes Metab 2008 February;34(1):12-8.

(79) Shetty S, Kusminski CM, Scherer PE. Adiponectin in health and disease: evaluation of adiponectin-targeted drug development strategies. Trends Pharmacol

Sci 2009 May;30(5):234-9.

(80) Maeda N, Shimomura I, Kishida K, Nishizawa H, Matsuda M, Nagaretani H, Furuyama N, Kondo H, Takahashi M, Arita Y, Komuro R, Ouchi N, Kihara S, Tochino Y, Okutomi K, Horie M, Takeda S, Aoyama T, Funahashi T, Matsuzawa Y. Diet-induced insulin resistance in mice lacking adiponectin/ACRP30. Nat Med 2002 July;8(7):731-7.

Chapter 4.

102

(81) Bruun JM, Lihn AS, Verdich C, Pedersen SB, Toubro S, Astrup A, Richelsen B. Regulation of adiponectin by adipose tissue-derived cytokines: in vivo and in vitro investigations in humans. Am J Physiol Endocrinol Metab 2003 September;285(3):E527-E533.

(82) Komai N, Morita Y, Sakuta T, Kuwabara A, Kashihara N. Anti-tumor necrosis factor therapy increases serum adiponectin levels with the improvement of endothelial dysfunction in patients with rheumatoid arthritis. Mod Rheumatol 2007;17(5):385-90.

(83) Trayhurn P, Wood IS. Signalling role of adipose tissue: adipokines and inflammation in obesity. Biochem Soc Trans 2005 November;33(Pt 5):1078-81.

(84) Tordjman J, Poitou C, Hugol D, Bouillot JL, Basdevant A, Bedossa P, Guerre-Millo M, Clement K. Association between omental adipose tissue macrophages and liver histopathology in morbid obesity: influence of glycemic status. J Hepatol 2009 August;51(2):354-62.

(85) Cancello R, Tordjman J, Poitou C, Guilhem G, Bouillot JL, Hugol D, Coussieu C, Basdevant A, Bar Hen A, Bedossa P, Guerre-Millo M, Clement K. Increased infiltration of macrophages in omental adipose tissue is associated with marked hepatic lesions in morbid human obesity. Diabetes 2006 June;55(6):1554-61.

(86) Valenti L, Fracanzani AL, Dongiovanni P, Santorelli G, Branchi A, Taioli E, Fiorelli G, Fargion S. Tumor necrosis factor alpha promoter polymorphisms and insulin resistance in nonalcoholic fatty liver disease. Gastroenterology 2002 February;122(2):274-80.

(87) Stefan N, Machicao F, Staiger H, Machann J, Schick F, Tschritter O, Spieth C, Weigert C, Fritsche A, Stumvoll M, Haring HU. Polymorphisms in the gene encoding adiponectin receptor 1 are associated with insulin resistance and high liver fat. Diabetologia 2005 November;48(11):2282-91.

(88) Kotronen A, Yki-Jarvinen H, Aminoff A, Bergholm R, Pietilainen KH, Westerbacka J, Talmud PJ, Humphries SE, Hamsten A, Isomaa B, Groop L, Orho-Melander M, Ehrenborg E, Fisher RM. Genetic variation in the ADIPOR2 gene is associated with liver fat content and its surrogate markers in three independent cohorts. Eur J Endocrinol 2009 April;160(4):593-602.

(89) Tamai H, Horie Y, Kato S, Yokoyama H, Ishii H. Long-term ethanol feeding enhances susceptibility of the liver to orally administered lipopolysaccharides in rats. Alcohol Clin Exp Res 2002 August;26(8 Suppl):75S-80S.

(90) Yoon MJ, Lee GY, Chung JJ, Ahn YH, Hong SH, Kim JB. Adiponectin increases fatty acid oxidation in skeletal muscle cells by sequential activation of AMP-activated protein kinase, p38 mitogen-activated protein kinase, and peroxisome proliferator-activated receptor alpha. Diabetes 2006 September;55(9):2562-70.


103

(91) Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis factor-alpha: direct role in obesity-linked insulin resistance. Science 1993 January 1;259(5091):87-91.

(92) Gupta D, Varma S, Khandelwal RL. Long-term effects of tumor necrosis factor-alpha treatment on insulin signaling pathway in HepG2 cells and HepG2 cells overexpressing constitutively active Akt/PKB. J Cell Biochem 2007 February 15;100(3):593-607.

(93) Schaffer JE. Fatty acid transport: the roads taken. Am J Physiol Endocrinol Metab 2002 February;282(2):E239-E246.




(97) Yang RZ, Blaileanu G, Hansen BC, Shuldiner AR, Gong DW. cDNA cloning, genomic structure, chromosomal mapping, and functional expression of a novel human alanine aminotransferase. Genomics 2002 March;79(3):445-50.




(101) Valenti L, Fracanzani AL, Bugianesi E, Dongiovanni P, Galmozzi E, Vanni E, Canavesi E, Lattuada E, Roviaro G, Marchesini G, Fargion S. HFE genotype,

Chapter 4.

104

parenchymal iron accumulation, and liver fibrosis in patients with nonalcoholic fatty liver disease. Gastroenterology 2010 March;138(3):905-12.

(102) Sokol RJ, Twedt D, McKim JM, Jr., Devereaux MW, Karrer FM, Kam I, von Steigman G, Narkewicz MR, Bacon BR, Britton RS, . Oxidant injury to hepatic mitochondria in patients with Wilson's disease and Bedlington terriers with copper toxicosis. Gastroenterology 1994 December;107(6):1788-98.

(103) Targher G, Bertolini L, Rodella S, Zoppini G, Zenari L, Falezza G. Associations between liver histology and cortisol secretion in subjects with nonalcoholic fatty liver disease. Clin Endocrinol (Oxf) 2006 March;64(3):337-41.

(104) Volzke H, Schwarz S, Baumeister SE, Wallaschofski H, Schwahn C, Grabe HJ, Kohlmann T, John U, Doren M. Menopausal status and hepatic steatosis in a general female population. Gut 2007 April;56(4):594-5.

(105) Lonardo A, Carani C, Carulli N, Loria P. 'Endocrine NAFLD' a hormonocentric perspective of nonalcoholic fatty liver disease pathogenesis. J Hepatol 2006 June;44(6):1196-207.

(106) Ludescher B, Leitlein G, Schaefer JE, Vanhoeffen S, Baar S, Machann J, Claussen CD, Schick F, Eschweiler GW. Changes of body composition in bulimia nervosa: increased visceral fat and adrenal gland size. Psychosom Med 2009 January;71(1):93-7.

(107) Kumpf VJ. Parenteral nutrition-associated liver disease in adult and pediatric patients. Nutr Clin Pract 2006 June;21(3):279-90.

(108) Watkins PB, Kaplowitz N, Slattery JT, Colonese CR, Colucci SV, Stewart PW, Harris SC. Aminotransferase elevations in healthy adults receiving 4 grams of acetaminophen daily: a randomized controlled trial. JAMA 2006 July 5;296(1):87-93.

(109) Sutinen J, Hakkinen AM, Westerbacka J, Seppala-Lindroos A, Vehkavaara S, Halavaara J, Jarvinen A, Ristola M, Yki-Jarvinen H. Increased fat accumulation in the liver in HIV-infected patients with antiretroviral therapy-associated lipodystrophy. AIDS 2002 November 8;16(16):2183-93.

(110) Argo CK, Loria P, Caldwell SH, Lonardo A. Statins in liver disease: a molehill, an iceberg, or neither? Hepatology 2008 August;48(2):662-9.

(111) Lonardo A, Adinolfi LE, Loria P, Carulli N, Ruggiero G, Day CP. Steatosis and hepatitis C virus: mechanisms and significance for hepatic and extrahepatic disease. Gastroenterology 2004 February;126(2):586-97.


105

(112) Tost JR, Vidal R, Cayla J, Diaz-Cabanela D, Jimenez A, Broquetas JM. Severe hepatotoxicity due to anti-tuberculosis drugs in Spain. Int J Tuberc Lung Dis 2005 May;9(5):534-40.

(113) Laine L, Goldkind L, Curtis SP, Connors LG, Yanqiong Z, Cannon CP. How common is diclofenac-associated liver injury? Analysis of 17,289 arthritis patients in a long-term prospective clinical trial. Am J Gastroenterol 2009 February;104(2):356-62.

(114) Kent PD, Luthra HS, Michet C, Jr. Risk factors for methotrexate-induced abnormal laboratory monitoring results in patients with rheumatoid arthritis. J Rheumatol 2004 September;31(9):1727-31.

(115) Gisbert JP, Luna M, Gonzalez-Lama Y, Pousa ID, Velasco M, Moreno-Otero R, Mate J. Liver injury in inflammatory bowel disease: long-term follow-up study of 786 patients. Inflamm Bowel Dis 2007 September;13(9):1106-14.

(116) Al Chalabi T, Underhill JA, Portmann BC, McFarlane IG, Heneghan MA. Effects of serum aspartate aminotransferase levels in patients with autoimmune hepatitis influence disease course and outcome. Clin Gastroenterol Hepatol 2008 December;6(12):1389-95.

(117) Kuo CF, Yu KH, Luo SF, Chiu CT, Ko YS, Hwang JS, Tseng WY, Chang HC, Chen HW, See LC. Gout and risk of non-alcoholic fatty liver disease. Scand J

Rheumatol 2010 June 21.

(118) Shimizu I. Impact of oestrogens on the progression of liver disease. Liver Int 2003 February;23(1):63-9.

(119) Bruno S, Maisonneuve P, Castellana P, Rotmensz N, Rossi S, Maggioni M, Persico M, Colombo A, Monasterolo F, Casadei-Giunchi D, Desiderio F, Stroffolini T, Sacchini V, Decensi A, Veronesi U. Incidence and risk factors for non-alcoholic steatohepatitis: prospective study of 5408 women enrolled in Italian tamoxifen chemoprevention trial. BMJ 2005 April 23;330(7497):932.

(120) Jamerson PA. The association between acute fatty liver of pregnancy and fatty acid oxidation disorders. J Obstet Gynecol Neonatal Nurs 2005 January;34(1):87-92.

(121) Reichlin S. Neuroendocrine-immune interactions. N Engl J Med 1993 October 21;329(17):1246-53.

Appen

dix Ι.

Ris

k f

acto

rs f

or

fatty l

iver

dis

ease

, ar

ranged

in t

he

follow

ing t

hre

e ca

tegori

es:

1)

risk

fac

tors

for

hep

atic

lip

id c

onte

nt, 2

) ri

sk

fact

ors

for

inhib

ited

hep

atic

met

aboli

sm, an

d 3

) ri

sk f

acto

rs f

or

hep

atic

infl

amm

atio

n *

Cat

egory

1

Cat

egory

2

Cat

egory

3

RISK FACTORS FOR

HEPATIC

LIPID

CONTENT

RISK FACTORS FOR

INHIB

ITED H

EPATIC

METABOLISM

RISK FACTORS FOR

HEPATIC

INFLAMMATIO

N

G _ STORAGE AND

SECRETIO

N DISEASE

Storage diseases, e.g.:

Lyso

som

al a

cid lip

ase

def

icie

ncy

(W

olm

an d

isea

se o

r C

hole

ster

ol es

ter

stora

ge

dis

ease

) 53

Secretion disease, e.g.:

Hypobet

alip

opro

tein

emia

54


Hae

moch

rom

atosi

s (I

ron a

ccum

ula

tion)

101


Wil

son’s

dis

ease

(C

opper

acc

um

ula

tion)

102

A _ THE LIV

ER

Pathophysiology within the liver is shown in figure ІІ and

table І


table І


table І

B _ (LIPID

) SOURCES

AND EXERCISE

Total calorie intake

48 and lipid sources:

1. Lypolysis from peripheral fat depots

40 particularly in the

case of:

• In

suli

n r

esis

tance

• Fas

ting/r

apid

wei

ght lo

ss/s

tarv

atio

n 2

6

• U

nfa

voura

ble

body f

at d

istr

ibuti

on (

VA

T↑)

cause

d b

y e

.g.:

- L

ipoat

rophy/l

ipodyst

rophy 2

6

- C

ort

isol

103 (

stre

ss-i

nduce

d g

luco

cort

icoid

) -

Est

rogen

def

icie

ncy

(m

ale

sex 3

8, m

enopau

se 1

04 o

r T

urn

er’s

syndro

me

105)

- B

uli

mia

Ner

vosa

106

Reactive oxygen species

55

Fat depot-induced inflammation:

• In

flam

mat

ory

cyto

kin

e pro

duct

ion b

y p

erip

her

al f

at d

epots

par

ticu

larl

y in:

- U

nfa

voura

ble

body f

at d

istr

ibuti

on (

VA

T↑ a

nd S

AT

↓)

32,

37

- M

acro

phag

e in

filtra

tion in a

dip

ocy

tes,

whic

h o

ccurs

par

ticu

larl

y in V

AT

84,

85

2. Dietary carbohydrates

40 (+ DNL)

3. Dietary fat

40

Decreased physical activity 4

5-4

7

C _ SPECIF

IC TYPES O

F

NUTRIT

ION &

INTAKE

Lip

opoly

sacc

har

ides

↑ 4

2

Alcohol

57

Alc

ohol →

inte

stin

al p

erm

eabil

ity↑ 8

9 →

SIB

O/e

ndoto

xin

s/lipopoly

sacc

har

ides

↑ 8

9

Lip

opoly

sacc

har

ides

↑ 4

2

Total parenteral nutrition 2

6, 107 →

SIB

O/e

ndoto

xin

s/lipopoly

sacc

har

ides

107

Lip

opoly

sacc

har

ides

↑ 4

2

Gluten &

Celiac disease (untreated)

→ tig

ht ju

nct

ions↓

39 →

in

test

inal

per

mea

bil

ity↑ 3

9 →

SIB

O/e

ndoto

xin

s/lipopoly

sacc

har

ides

↑ 3

9

Consumption of oxidants e.g. metals 1

01, 102

Consumption of antioxidants e.g. vitamins

Toxic m

ushrooms

26

Hypoglycin (

unusu

al a

min

o a

cid)

57

Cocaine

26

E _ ENVIR

ONMENT

Several (petro)chem

icals e.g.:

• B

enze

ne,

xyle

ne,

vin

yl ch

lori

de

66

Potential P-I response

71

• N

ickel

71

Chapter 4.

106

F _ CARDIO

-VASCULAR

SYSTEM

Anti-arrhythmic dru

g:

• A

mio

dar

oneD

■ 5

7,

69

Several calcium channel blockers 26

Several anti-platelet drugs e.g.:

• A

spir

in 5

7,

69


71

F _ CENTRAL NERVOUS

SYSTEM

Several selective serotonin re-uptake inhibitors e.g.:

• A

min

epti

ne

57,

69

• T

ianep

tine

57,

69

P-I response 7

1

Non-opioid analgesics, e.g.:

• Par

acet

amol (a

ceta

min

ophen

) 69, 108

• Sev

eral

2-A

rylp

ropio

nic

aci

ds,

e.g

.:

- Ib

upro

fen 5

7, 69

Strong opioids e.g.:

• B

upre

norp

hin

e■ 6

9

Drug for control of epilepsy:

• V

alpro

ic a

cid

■ 5

7,

69

Drug for control of Parkinson e.g.:

• T

olc

apone■

69 (

cate

chol-

O-m

ethylt

ransf

eras

e in

hib

itor)

Prostaglandine synthase inhibitor

• C

elec

oxib

71

Local analgesics:

• L

idoca

in 7

1

• M

epiv

acai

n 7

1

Drugs for control of epilepsy:

• C

arbam

azep

ine

71

• L

amotr

igin

e 71

D&F _ INFECTIO

NS

HAART-induced lipodystrophy in H

IV 1

09

Several anti-H

IV drugs:

• Sev

eral

Did

eoxynucl

eosi

des

57

• Sev

eral

NR

TIs

■ 6

9, e.

g.:

- D

idan

osi

ne

26

Hepatitis drug:

• In

terf

eron a

lpha

57

Viral infections e.g.:

• H

epat

itis

, par

ticu

larl

y type

C 1

10, 111

Potential P-I response? 7

1

Lip

opoly

sacc

har

ides

↑ 4

2

Several tetracyclines (antibiotics)

57,

69

Anti-TBC e.g.:

• R

ifam

pin

& P

yra

zinam

ide

112

• M

ethotr

exat

e &

Dic

lofe

nac

113, 114

Bacterial infections e.g.:

• SIB

O 3

6,

39 +

inte

stin

al p

erm

eabil

ity↑ 3

9

• T

BC

(dru

g tre

ated

) 112-1

14

Potential P-I response 7

1

Fungal infections

Burn injury

→ infe

ctio

ns↑

D&F _ AUTOIM

MUNE

DISEASE &

MUSCULOSCELETAL/

JOIN

T DISEASE

Glucocorticoid drugs

26 (

ster

oid

horm

ones

) →

unfa

voura

ble

body f

at d

istr

ibution

Several NSAID

s e.g.:

• A

spir

in 5

7,

69

• D

iclo

fenac

69

• N

imes

uli

de

69

IBD drug:

• T

hio

puri

ne

115

Several RA drugs:

• M

ethotr

exat

e &

Dic

lofe

nac

113, 114

• E

tori

coxib

113, 114

Autoim

mune and atopic inflammatory disease e.g.:

• A

uto

imm

une

hep

atit

is 1

16

• IB

D (

dru

g tre

ated

) 115 →

inte

stin

al tig

ht

ju

nct

ions↓

39 →

inte

stin

al p

erm

eabil

ity↑ 3

9

• R

A (

dru

g tre

ated

) 113, 114

Potential P-I response? 7

1

Gout and cytotoxic-induced hyperuricaem

ia drug:

• B

enzb

rom

arone■

69

Gout

117


71

D _ SKIN

Adipose tissue inflammation e.g.:

Web

er-C

hri

stia

n s

yndro

me

26


107

F _ G

YNAECOLOGY

Pote

ntial

chan

ge

in b

ody f

at d

istr

ibuti

on

Natural

118 and synthetic 5

7 estrogens


1

D&F _ M

ALIG

NANT

DISEASE

Pote

ntial

chan

ge

in b

ody f

at d

istr

ibuti

on

Colorectal cancer dru

g:

• Ir

inote

can 6

9 (

topois

om

eras

e I

inhib

itor)


1

Several horm

one antagonists in breast cancer i.e.:

• T

amoxif

en■ 6

9,

119

• T

ore

mif

ene

69

H _ RARE SYNDROMES

ASSOCIA

TED W

ITH FLD

Reye’s syndrome

57

Acute fatty liver of pregnancy

57,

120

UNDERLYIN

G:

NERVOUS AND

ENDOCRIN

E SYSTEM

Hypothyroidism →

body f

at↑ 1

05

Overactivity of the HPA axis:

• H

yper

glu

coco

rtic

oid

ism

, e.

g. hyper

cort

icolism

26 →

unfa

voura

ble

body f

at d

istr

ibuti

on 1

05

Neuroendocrine-im

munomodulation 1

21

• In

crea

sed infl

amm

atio

n r

esponse

• D

ecre

ased

infl

amm

atio

n r

esponse

Underactivity of the HPG axis:

• D

ecre

ased

ovar

ial pro

duct

ion o

f es

trogen

s →

unfa

voura

ble

body f

at d

istr

ibution 3

8,

104

• D

ecre

ased

tes

ticu

lar

pro

duct

ion o

f te

stost

eron →

unbal

ance

bet

wee

n f

at a

nd lea

n m

ass

105

Underactivity of the HPG axis:

• D

ecre

ased

ovar

ial pro

duct

ion o

f es

trogen

s →

dec

reas

ed a

nti

-oxid

ants

118

*, th

is tab

le m

ay v

ery lik

ely b

e a

par

tial

lis

t. F

or

all dru

gs

poss

ibly

involv

ed w

ith the

inci

den

ce a

nd p

rogre

ssio

n o

f FL

D the

read

er is

refe

rred

to

the

foll

ow

ing p

aper

s: i

) dru

g-i

nduce

d m

itoch

ondri

al d

ysf

unct

ion 5

7,

69,

ii)

dru

g-

and h

erbal

-induce

d h

epat

oto

xic

ity 6

4,

iii)

dru

g-i

nduce

d

infl

amm

atio

n 7

1, iv

) en

vir

onm

enta

l to

xin

s 66, an

d v

) gen

eral

rev

iew

26.

Study population:

hum

ans

36, 38, 40, 45-4

8, 53, 54, 66, 102-1

04, 106, 108, 109, 112-1

17, 119; an

imal

model

(s)

only

73, 89.

Review papers

26, 42, 55, 57-5

9, 69, 71, 105, 107, 110, 111, 118, 120, 121.

Study design:

pro

spec

tive

40, 47, 53, 73, 89, 108, 113, 115, 116, 119; cr

oss

-sec

tional

36, 38, 39, 45, 46, 54, 102-1

04, 109, 117; re

trosp

ecti

ve

48, 66, 106, 112, 114.

FLD diagnosis modality:

his

tolo

gy 3

9,

48,

66,

102,

103;

1H

MR

S 3

8,

45,

47,

109; ult

raso

nogra

phy 4

6,

53,

54,

66,

104,

115,

117; liver

funct

ion tes

t(s)

only

40,

89,

108, 112-1

14, 116, 119.

Signs:

↑, in

crea

se; ↓, dec

reas

e; →

, le

adin

g to;

■, m

arket

ed d

rug that

has

rec

eived

a b

lack

box w

arnin

g 6

9; D, ca

use

s phosp

holi

pid

osi

s.

Abbreviations:

DN

L, de

novo lip

ogen

esis

; H

AA

RT

, hig

hly

act

ive

antire

trovir

al tre

atm

ents

; H

IV, hum

an im

munodef

icie

ncy

vir

us;

1H

MR

S,

mag

net

ic re

sonan

ce sp

ectr

osc

opy;

HP

A ax

is,

hypoth

alam

o-p

ituitar

y-a

dre

nal

ax

is;

HP

G ax

is,

hypoth

alam

o-p

ituit

ary-g

onad

al ax

is;

IBD

, in

flam

mat

ory

bow

el d

isea

se;

NR

TI,

nucl

eosi

de

rever

se t

ransc

ripta

se i

nhib

itor;

NSA

ID,

non-s

tero

idal

anti-i

nfl

amm

atory

dru

g;

P-I

com

ple

x,

phar

mag

olo

gic

al-i

nte

ract

ion co

mple

x;

PPA

Rα,

per

oxis

om

e pro

life

rato

r-ac

tivat

ed re

cepto

r al

pha;

R

A,

rheu

mat

oid

ar

thri

tis;

SIB

O,

smal

l in

test

inal

bac

teri

al o

ver

gro

wth

; T

BC

, tu

ber

culo

sis;

TN

Fα, tu

mour

nec

rosi

s fa

ctor

alpha;

VA

T, vis

cera

l ad

ipose

tis

sue.

Chapter 4.

108

Chapter 5

Ultrasonography to quantify hepatic fat content:

validation by 1H Magnetic Resonance Spectroscopy

Obesity (Silver Spring) 2009; 17(12):2239-2244

Mireille A. Edens 1

Peter M. van Ooijen 2

Wendy J. Post 1

Mark J. Haagmans 2

Wisnumurti Kristanto 2

Paul E. Sijens 2

Erik J. van der Jagt 2

Ronald P. Stolk 1

Department of Epidemiology 1

Department of Radiology 2 University Medical Center Groningen

University of Groningen Groningen, the Netherlands

Chapter 5.

110

ABSTRACT

An abundance of fat stored within the liver, or steatosis, is the beginning of a broad

hepatological spectrum, usually referred to as fatty liver disease (FLD). For studies on

FLD, quantitative hepatic fat ultrasonography would be an appealing study modality.

Objective of the present study was to developed a technique for quantifying hepatic fat

content by ultrasonography and validate this using proton magnetic resonance spectroscopy

(1H MRS) as gold standard. Eighteen White volunteers (BMI range 21.0 to 42.9) were

scanned by both ultrasonography and 1H MRS. Altered ultrasound characteristics, present

in the case of FLD, were assessed using a specially developed software program. Various

attenuation and textural based indices of FLD were extracted from ultrasound images.

Using linear regression analysis, the predictive power of several models (consisting of both

attenuation and textural based measures) on log 10-transformed hepatic fat content by 1H

MRS were investigated. The best quantitative model was compared with a qualitative

ultrasonography method, as used in clinical care. A model with four ultrasound

characteristics could modestly predict the amount of liver fat (adjusted explained variance

43.2%, p=0.021). Expanding the model to seven ultrasound characteristics increased

adjusted explained variance to 60% (p=0.015), with r=0.789 (p<0.001). Comparing this

quantitative model with qualitative ultrasonography revealed a significant advantage of the

quantitative model in predicting hepatic fat content (p<0.001). This validation study shows

that a combination of computer-assessed ultrasound measures from routine ultrasound

images can be used to quantitatively assess hepatic fat content.

Ultrasonography to quantify hepatic fat content

111

INTRODUCTION

A continuous accumulation of lipids in the liver may result in a broad hepatological

spectrum, usually referred to as fatty liver disease (FLD) 1, 2. An abundance of fat within the

liver, or steatosis, can progress to steatohepatitis (fat and inflammation, with or without

fibrosis) and cirrhosis (maximum fibrosis score) 3, and has also been associated with

hepatocarcinoma 4. Additionally, FLD, particularly non-alcoholic FLD (NAFLD), is an

underlying condition for cardiovascular disease 5, 6. As alcoholic FLD (AFLD) and NAFLD

are histologically indistinct 7, distinction between both is neither possible nor relevant in

relation to measurement of hepatic fat content. The estimated FLD prevalence is one third

of the general adult Western population 8-10, and may have been increasing in parallel with

the global increase of obesity 11.

Hepatic fat content can be determined by histological 2 or biochemical 12, 13 analysis of liver

tissue by biopsy, magnetic resonance techniques 14, computed tomography 15 and

ultrasonography 16, 17. Ultrasonography is, in contrast to other diagnostic modalities, an

appealing method for large population studies on FLD, as is it non-invasive (painless, no

harmful radiation), portable and relatively inexpensive. In the case of parenchymal liver

disease, reflections of liver tissue by ultrasonography are altered 16, 17. In clinical care,

ultrasonography is the most often used diagnostic modality, but in a qualitative way.

Steatosis can be qualitatively assessed by: i] hyperechogenity of liver tissue (‘bright liver’)

as often compared to hypoechogenity of the kidney cortex, ii] fine, tightly packed echoes,

iii] fall of echo amplitude with depth (posterior beam attenuation), iiii] loss of echoes from

the walls of the portal veins (featureless appearance) 16, 17. As this is a qualitative scoring

method and also subjective 18, quantitative approaches for identification of liver disease

have been suggested. However, these methods have never been validated by an appropriate

quantitative gold standard. The purpose of the present study was to develop and validate

quantitative analysis of ultrasonography images, for assessment of hepatic fat content, using

proton magnetic resonance spectroscopy (1H MRS) as gold standard.

Chapter 5.

112

METHODS AND PROCEDURES

VOLUNTEERS

Volunteers were recruited by advertisement, and a heterogenic study population was strived

after. Exclusion criteria were current presence of hepatic pathology, previous hepatic or

renal surgery, and standard MR-contraindications. The volunteers underwent both hepatic

ultrasonography and 1H MRS, and a short physical examination. All volunteers gave

written informed consent. This study was approved by the Medical Ethics Committee of the

University Medical Center Groningen.

ULTRASONOGRAPHY

Ultrasonography was performed using a Philips ATL ultrasound machine (Philips, Best, the

Netherlands), with a 5 – 2 MHz curved array transducer.

Quantitative ultrasonography

Imaging

In one ultrasound image both liver and right kidney were visualised 17, as shown by figure

І. Imaging was performed by an experienced radiologist. One standard image, with ‘persist’

(med), ‘2D opt’ (gen), ‘frame rate’ (high), ‘gain’ (40) and ‘image depth’ (14.7 cm), was

used for analysis.

Analysis

Images were analysed by an operator (operator 1) twice, with a one month interval, and the

average values by operator 1 were used in this study. In order to study inter-operator

reliability, another operator (operator 2) analysed the images, while untrained for the

method and blinded for all study outcomes.

Data extraction and data

Data were extracted from ultrasound images using a modified version of a specially

developed software program (dept. of BME, Technion IIT, Haifa, Israel) in the MATLAB

programming environment, previously described by Gaitini et al. 19. Figure І shows an

example of data extraction. According to a standard protocol, regions of interest and

attenuation lines were interactively placed in the liver images in order to calculate several

attenuation indices and several textural indices. Figure ІІ shows a scheme on quantitative

ultrasonography measures, including the presently validated indices in the white boxes.


113

Both a region of interest (quadrangle) and an attenuation

line (closed line) were placed according to a standard

protocol.

The region of interest had to be placed in a bright area,

while avoiding large artefacts like rib shadows and large

blood vessels, at a depth of 4 to 6 cm. The attenuation line

had to be placed in a bright pathway, while avoiding large

artefacts, at a straight line from the ‘origin’ of ultrasound

(intermittent line).

The region of interest served for the determination of

several textural indices. The attenuation line was used for

determining attenuation estimates 19.

Figure І. Ultrasonography image analysis

Qualitative ultrasonography

In addition to the quantitative approach, the radiologist made an ultrasound image with

optimum settings, as used in clinical care. This image was qualitatively scored by the

radiologist, according to standard qualitative criteria 16, 17, 26, while blinded for all study

outcomes.

[b]

Texturalbased

[e]

First-order

greyscalestatistics

[k]

•Slope

• •Lenghtof line

[c]

Statisticalmethods

[f]

Second-order

greyscalestatistics

[m]

•Meangreylevel

[i]

Spatial grey level

dependence matrices

[j]

Grey level

difference matrices

[n]

•Co AngSec Mom #

•Co Correlation

•Co DiffEntropy

•Co Energy#

•Co Entropy

•Co Inertia

•Co Loc Homo

•Co SumEntropy

[o]

•DiffAngSec Mom

•DiffContrast

•DiffCovariance

•DiffEntropy

•DiffMean

&

•FP1 &

•HP1

•InvDiffMom

[a]

Attenuationbased

[h]

Linearregression

[g]

Texture feature

coding method

[p]

•TFCM Coarseness

•TFCM Code Entropy

•TFCM Code Similarity

•TFCM Homogenity*

•TFCM Meanconvergence

•TFCM Variance

Quantitative

ultrasonography

[d]

Structuralmethods

[l]

•Offset

Figure ІІ. S

chem

e an

d c

lass

ific

atio

n o

n q

uan

tita

tive

ult

raso

nogra

phy, in

cludin

g the

pre

sently v

alid

ated

mea

sure

s in

the

white

boxes

[k]

to [

p]

Mea

sure

s sh

arin

g s

uper

scri

pts

had

the

sam

e outc

om

e, *

= o

utc

om

e w

as a

lway

s 0.

For

a det

aile

d e

xpla

nat

ion o

n t

he

pre

sently v

alid

ated

quan

tita

tive

ultra

sonogra

phy

mea

sure

s 19, th

e re

ader

is

refe

rred

to the

follow

ing r

efer

ence

s: [

a] 1

9, 20, [b

] 21, [c

] 19, 21, 22, [d

] 19, 21, 22, [e

] 22, [f

] 22-2

4, [g

] 23, [h

] ap

pen

dix

І, [i

] 25, [j

] 25, [k

]and[l

] ap

pen

dix

І, [m

]

19, [n

] ap

pen

dix

ІІ,

[o]

appen

dix

ІІ,

[p]

23.

Chapter 5.

114


115

MULTI-VOXEL PROTON MAGNETIC RESONANCE SPECTROSCOPY

In general, by means of radiofrequency transmission and reception, a magnetic resonance

scanner detects resonance signals of both hepatic lipids (mainly methylene, i.e. CH2, from

fatty acyl chains) and hepatic water 27. As previously described in detail 14, 28, 1H MRS was

performed, using a 1.5 Tesla whole-body scanner (MAGNETOM Avanto; Siemens

Medical Solutions, Erlangen, Germany) equipped with gradients of up to 40 mTm-1

(maximal slew rate = 200 mT m-1/ms) and a six-channel spine array coil. Subjects were in

supine position with a large flex coil placed over the liver, which was simultaneously used

with the spine array coil as receiver. T1-weighted gradient-echo images were recorded to

assess the anatomy of liver. Using a field of view of 16×16 cm2 and a volume of interest of

5×8×4 cm3 positioned within the liver, hybrid 2D-spectroscopic imaging (chemical shift

imaging or CSI), point resolved spectroscopy (PRESS) with a repetition time (TR) of 5000

ms and an echo time (TE) of 30 ms was performed. The CSI measurement lasted 16×16×5

= 1280 s, corresponding to approximately 21 min. Shimming was automated and water

suppression was not applied in order to be able to calculate the fat-water ratio distributions

in the liver directly. At the used TR of 5 s, T1 saturation of the water and fat signals is

negligible, i.e. TR > 5T1. At the used TE of 30 ms the correction applied to our data to

compensate for the fact that the fat signal has a longer T2 (78 ms) than water (60 ms) was

12.2 %. Hepatic fat content was calculated by the peak CH2 signal (at 1.3 parts/ million)

divided by the sum of the peak CH2 signal and peak H2O signal (at 4.7 parts/ million), using

water as an internal reference 14, 28. 1H MRS has been validated, by comparison with both

histological and biochemical analysis of liver tissue by biopsy 27, 29, 30.

A hepatic fat content of 5.56% by 1H MRS is used as cut-off value for diagnosing FLD,

based on the 95th percentile hepatic fat distribution of a low risk group 10.

STATISTICS

Univariate analysis and multiple regression analysis

As distribution of hepatic fat content by 1H MRS was skewed, values were log 10

transformed. Plotting and correlation (Pearson) was used to explore univariate concordance

with log 10 1H MRS. The classification of variables in figure ІІ (white boxes), followed by

‘backward selection’, was used for variable selection in a linear regression model. Firstly,

the variables from separate boxes of figure ІІ were assessed, i.e. separate ultrasonography

Chapter 5.

116

aspects. Secondly, variables from combinations of boxes of figure ІІ were assessed, i.e.

information from several ultrasonography aspects.

Evaluation and bootstrap

Models were evaluated on adjusted explained variance (adj. R2) and explained variance

(R2). By means of bootstrap, 95% confidence intervals were estimated for regression

coefficients, and adj. R2 and R2. Moreover, a 95% prediction interval was calculated.

Quantitative versus qualitative ultrasonography

The Chi-square test was used to test differences between the two methods. Additionally,

sensitivity and specificity of both methods were calculated. In addition to the 95%

prediction interval of the quantitative method, a 95% prediction interval was calculated for

the qualitative method as well.

Reliability

Both intra-observer and inter-observer reproducibility of algorithms were studied, using the

Bland & Altman method 31.

Statistical analysis was performed using software programs SPSS version 14 and R version

2.6.2.


117

RESULTS

STUDY POPULATION

Twenty apparently healthy White volunteers were examined. One volunteer was excluded

because of hepatic pathology (haemangioma), and one volunteer was excluded because of

rib shadows over the liver-kidney image.

The study population (n = 18) consisted of 10 males and 8 females, with a mean ±sd age of

46.0 ±14.1 year, body mass index of 28.7 ±6.4 kg/m2 (range 21.0 to 42.9), and waist to hip

ratio of 0.93 ±0.11. Hepatic fat content by 1H MRS ranged from 0.32% – 18.55%, with a

median of 1.75%.

UNIVARIATE ANALYSIS

Plots and correlation coefficients revealed no associations. Only slope and co-entropy were

borderline significantly associated with log 10 1H MRS, with r=−0.423 (p=0.081) and

r=−0.418 (p=0.084), respectively.

MULTIPLE ANALYSIS

Information from separate ultrasonography aspects, i.e. separate boxes from figure ІІ,

revealed no associations with log 10 1H MRS (data not shown). Combining information

from different ultrasonography aspects, i.e. by combining boxes from figure ІІ, was

associated with hepatic fat content by log 10 1H MRS (model 1, table І). Including more

ultrasonography characteristics, further improved the association (model 2, table І).

Algorithms

The algorithms derived from these models are:

Algorithm 1:

log 10 1H MRSpred. = –37.67 –0.07*offset –0.78*slope

−3.85*co entropy + 3.56*co sum entropy.

Algorithm 2:

log 10 1H MRSpred. = −72.68 −0.07*offset −0.81*slope

−3.63*co entropy + 3.34*co sum entropy

−0.20*diff contrast + 84.84*inv diff mom + 5.98*FP1.

Chapter 5.

118

A scatter plot of algorithm 2 with log 10 1H MRS, including a prediction interval, is shown

in figure ІІІa. This means that by applying our algorithm to ultrasound images of new

volunteers, 95% of their predicted values will fall within this interval.

Table І. Prediction by combinations of ultrasound characteristics

Model

ID

Independent

variables a B [95% CIb]

p-

value

Model

p-value

Model adj. R2

[95% CIb]

Model R2

[95% CIb]

1

(constant)

offset

slope

co entropy

co sum entropy

−37.67 [−57.45 – −18.07]

−.07 [−0.10 – −0.03]

−.78 [−1.19 – −0.37]

−3.85 [−5.95 – −1.79]

3.56 [0.81 – 6.31]

.007

.009

.007

.008

.052

.021

43.2%

[−2.9% –

97.2%]

56.5%

[21.3% – 97.9%]

2

(constant)

offset

slope

co entropy

co sum entropy

diff contrast

inv diff mom

FP1

−72.68 [−96.70 – −48.76]

−.07 [−0.10 – −0.04]

−.81 [−1.17 – −0.44]

−3.63 [−5.33 – −1.93]

3.34 [1.24 – 5.40]

−.20 [−0.30 – −0.09]

84.84 [40.42 – 129.65]

5.98 [2.97 – 9.01]

.001

.012

.009

.011

.039

.019

.019

.016

.015

60.0%

[49.1% –

99.6%]

76.5%

[70.0% – 99.8%]

a = dependent variable is log 10 1H MRS, b = bootstrapped 95% confidence interval.

Adj. R2 = adjusted explained variance, B = regression coefficients, R2 = explained variance.

QUANTITATIVE VERSUS QUALITATIVE ULTRASONOGRAPHY

Validity of algorithm 2, in comparison with a qualitative ultrasonography method used in

clinical care, is shown in figure ІІІ. Using cut-off value 0.75 log 10 1H MRS, which is

5.56% by 1H MRS 10, 3 people had FLD. Quantitative ultrasonography was significantly

better associated with the presence of FLD than qualitative ultrasonography (χ 2 = 32.8,

with df = 1, p<0.001). Sensitivity and specificity were 66.7% and 100%, respectively, for

quantitative ultrasonography. Sensitivity and specificity were 100% and 40%, respectively,

for qualitative ultrasonography. Additionally, the prediction interval of the quantitative

method was much narrower in comparison to the prediction interval of the qualitative

method (figure ІІІ).


119

a. Quant. ultrasonography r = .789 (p<.001) R2 = 62.3%

b. Qual. ultrasonography r = .353 (p=.151) R2 = 12.4%

1H MRS

FLD no FLD

Quant.

FLD 2 0

no FLD 1 15

Qual.

FLD 3 9

no FLD 0 6

c. Comparison Quantitative cut-off: ≥.22 Qualitative cut-off: ≥1

Figure ІІІ. Quantitative versus qualitative ultrasonography, using correlation coefficients,

sensitivity and specificity, and prediction intervals

Qual. = qualitative, Quant. = quantitative, r = Pearson’s correlation coefficient, R2 =

explained variance.

RELIABILITY

Intra-operator and inter-operator reproducibility, regarding algorithm 2, are shown in figure

ІV. If we tolerate an operator difference smaller than 0.5, as shown by the interrupted lines,

5 people had a larger intra-operator difference of 0.55 to 2.66 log 10 1H MRSpred. Eleven

people had a larger inter-operant difference of 0.53 to 1.98 log 10 1H MRSpred. These

differences were independent of hepatic fat content (figure ІV).

Chapter 5.

120

a. 1st attempt by op. 1 minus 2nd attempt by op. 1

b. 1st attempt by op. 1 minus attempt by op. 2

Figure ІV. Reliability

op. = operator


121

DISCUSSION

This study shows that combinations of quantitative ultrasonography measures are

significantly associated with hepatic fat content by (log 10) 1H MRS (table І), and even

better than a qualitative method currently used in clinical care (figure ІІІ). Reliability was

reasonably well (figure ІV).

These results suggest that combinations of computer-assessed ultrasonography measures

quantify the ultrasonographic characteristics of FLD 16, 17, i.e. i] hyperechogenity, by offset

in a certain degree, ii] fine, tightly packed echoes in the case of hepatic fat, and coarse pin

head echoes in the case of fibrosis, by the textural based measures, and iii] fall of echo

amplitude with depth, by slope, and by offset in lesser degree (table І).

VALIDITY

Previously, ‘slope’ 19, 20 and ‘offset’, ‘mean grey level’, ‘co entropy’ and ‘co sum entropy’

19, revealed discriminative power in the FLD spectrum. In the present study, none of the

measures were univariately associated with hepatic fat content by 1H MRS. The attenuation

based measure ‘slope’ did not significantly predict hepatic fat content, whereas the slope

previously did reveal power for discriminating pure fatty livers (steatosis) from healthy

livers, with an area under the curve of 1 19, 20. However, the slope lost discriminative power

in the total FLD spectrum 19, 20. It is known that both hepatic fat content 3, 32, 33 and

(therefore) attenuation 17 are decreased in the case of (advanced) fibrosis and cirrhosis,

which might have caused the fall in discriminative power of the slope in the total FLD

spectrum 19, 20. Fibrosis itself does not produce attenuation 34. This may also explain why, in

models, the attenuation based indices must be accompanied by textural indices of

coarseness/ fibrosis, in particular ‘co entropy’ 19, 35.

In this apparently healthy study population, we obviously did not perform hepatic biopsy

for histological scoring of fibrosis stage, nor magnetic resonance elastography (MRE),

which determines liver stiffness as a marker of fibrosis 36. Therefore, it was not possible to

verify the effect of fibrosis on ultrasonography algorithms. Additionally, because of both

inclusion of the right kidney 17 and rib shadows in ultrasound images, it was not always

possible to draw the attenuation line to the bottom of the liver for realization of a ‘far field’

slope 19, 20. However, inclusion of length of the attenuation line in models on slope did not

Chapter 5.

122

lead to an improvement in the prediction of hepatic fat content by 1H MRS (data not

shown). Adjusting for focus depth and frequency was not possible, because of the small

ranges.

A limitation of the present study is the small study population (n = 18), however, bootstrap

of model 2 revealed good 95% confidence intervals (table І). Additionally, model 1 with

only 4 variables already showed significant results.

Because the performance of a test depends on the prevalence of an underlying disorder, e.g.

FLD, the sensitivity of quantitative ultrasonography may be lower in a clinical population.

RELIABILITY

Intra-operator difference was reasonably well, but inter-operator difference was less (figure

ІV). This may be explained by operator 1 being experienced, while operator 2 was not.

Retrospective analysis of the outliers from figure ІV (print screens’ of analysed images

were saved), revealed insight in the differences in analysis.

FUTURE RESEARCH

While histological scoring of liver tissue by biopsy is often considered the gold standard for

diagnosis of FLD, 1H MRS is more reliable, and may be more valid for quantification of

hepatic fat content. Reproducibility of qualitative histological analysis regarding steatosis

grade is good as weighted kappa scores range from 0.64 to 0.90 2, 37, whereas

reproducibility regarding hepatic fat content by 1H MRS is excellent with reported

correlation coefficients of up to 0.99 (p<0.001) 10. For future studies, 1H MRS combined

with MRE 36 would be an interesting gold standard.

CONCLUSION

This is the first in vivo validation study on quantitative hepatic fat ultrasonography, using

an excellent quantitative gold standard, i.e. multi-voxel 1H MRS 14, 28. Therefore, we feel

that the method needs to be improved before used as clinical diagnosis modality.

This validation study shows that a combination of computer-assessed ultrasound measures

from routine ultrasound images can be used to quantitatively assess hepatic fat content.

Reliability should be improved by more protocolized procedures and training of operators.

Please also see ‘additional remarks and recommendations for future research’


123

ACKNOWLEDGEMENTS

The authors thank dr. H Azhari, who enabled this study by sharing the software program,

developed by the Department of Biomedical Engineering, Technion Israel Institute of

Technology, Haifa, Israel 19.

The authors thank I Willeboordse, AM van Tienhoven, JH Potze, and P Kappert for

magnetic resonance scanning.

Chapter 5.

124

Reference List

(1) Matteoni CA, Younossi ZM, Gramlich T, Boparai N, Liu YC, McCullough AJ.

Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology 1999 June;116(6):1413-9.



(4) Smedile A, Bugianesi E. Steatosis and hepatocellular carcinoma risk. Eur Rev Med

Pharmacol Sci 2005 September;9(5):291-3.








(11) Friedman JM. Obesity in the new millennium. Nature 2000 April 6;404(6778):632-4.

(12) Araya J, Rodrigo R, Videla LA, Thielemann L, Orellana M, Pettinelli P, Poniachik J. Increase in long-chain polyunsaturated fatty acid n - 6/n - 3 ratio in relation to


125

hepatic steatosis in patients with non-alcoholic fatty liver disease. Clin Sci (Lond) 2004 June;106(6):635-43.

(13) Puri P, Baillie RA, Wiest MM, Mirshahi F, Choudhury J, Cheung O, Sargeant C, Contos MJ, Sanyal AJ. A lipidomic analysis of nonalcoholic fatty liver disease. Hepatology 2007 October;46(4):1081-90.


(15) Duman DG, Celikel C, Tuney D, Imeryuz N, Avsar E, Tozun N. Computed tomography in nonalcoholic fatty liver disease: a useful tool for hepatosteatosis assessment? Dig Dis Sci 2006 February;51(2):346-51.



(18) Smith-Levitin M, Blickstein I, Albrecht-Shach AA, Goldman RD, Gurewitsch E, Streltzoff J, Chervenak FA. Quantitative assessment of gray-level perception: observers' accuracy is dependent on density differences. Ultrasound Obstet

Gynecol 1997 November;10(5):346-9.

(19) Gaitini D, Baruch Y, Ghersin E, Veitsman E, Kerner H, Shalem B, Yaniv G, Sarfaty C, Azhari H. Feasibility study of ultrasonic fatty liver biopsy: texture vs. attenuation and backscatter. Ultrasound Med Biol 2004 October;30(10):1321-7.

(20) Graif M, Yanuka M, Baraz M, Blank A, Moshkovitz M, Kessler A, Gilat T, Weiss J, Walach E, Amazeen P, Irving CS. Quantitative estimation of attenuation in ultrasound video images: correlation with histology in diffuse liver disease. Invest

Radiol 2000 May;35(5):319-24.

(21) Zhang J, Tan T. Brief review of invariant texture analysis methods. Pattern

Recognition 2002;35:735-47.

(22) Haralick RM. Statistical and structural approaches to texture. Proc IEEE 1979 May;67(5):786-804.

(23) Horng MH, Sun YN, Lin XZ. Texture feature coding method for classification of liver sonography. Comput Med Imaging Graph 2002 January;26(1):33-42.

Chapter 5.

126

(24) Gotlieb C, Kreyszig H. Texture descriptors based on co-occurence matrices. Computer vision, graphics, and image processing 1990;51:70-86.

(25) Prakash KN, Ramakrishnan AG, Suresh S, Chow TW. An investigation into the feasibility of fetal lung maturity prediction using statistical textural features. Ultrason Imaging 2001 January;23(1):39-54.

(26) Mendler MH, Bouillet P, Le Sidaner A, Lavoine E, Labrousse F, Sautereau D, Pillegand B. Dual-energy CT in the diagnosis and quantification of fatty liver: limited clinical value in comparison to ultrasound scan and single-energy CT, with special reference to iron overload. J Hepatol 1998 May;28(5):785-94.

(27) Szczepaniak LS, Babcock EE, Schick F, Dobbins RL, Garg A, Burns DK, McGarry JD, Stein DT. Measurement of intracellular triglyceride stores by H spectroscopy: validation in vivo. Am J Physiol 1999 May;276(5 Pt 1):E977-E989.



Imaging 1994;12(3):487-95.

(30) Vuppalanchi R, Cummings OW, Saxena R, Ulbright TM, Martis N, Jones DR, Bansal N, Chalasani N. Relationship among histologic, radiologic, and biochemical assessments of hepatic steatosis: a study of human liver samples. J Clin Gastroenterol 2007 February;41(2):206-10.

(31) Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986 February 8;1(8476):307-10.


(33) Adams LA, Sanderson S, Lindor KD, Angulo P. The histological course of nonalcoholic fatty liver disease: a longitudinal study of 103 patients with sequential liver biopsies. J Hepatol 2005 January;42(1):132-8.

(34) Joseph AE, Saverymuttu SH. Ultrasound in the assessment of diffuse parenchymal liver disease. Clin Radiol 1991 October;44(4):219-21.

(35) Mazzone AM, Urbani MP, Picano E, Paterni M, Borgatti E, De Fabritiis A, Landini L. In vivo ultrasonic parametric imaging of carotid atherosclerotic plaque by videodensitometric technique. Angiology 1995 August;46(8):663-72.


127

(36) Yin M, Woollard J, Wang X, Torres VE, Harris PC, Ward CJ, Glaser KJ, Manduca A, Ehman RL. Quantitative assessment of hepatic fibrosis in an animal model with magnetic resonance elastography. Magn Reson Med 2007 August;58(2):346-53.

(37) Merriman RB, Ferrell LD, Patti MG, Weston SR, Pabst MS, Aouizerat BE, Bass NM. Correlation of paired liver biopsies in morbidly obese patients with suspected nonalcoholic fatty liver disease. Hepatology 2006 October;44(4):874-80.

Chapter 5.

128

Online Appendix І. Realization of attenuation based indices

0 ,6 5

0 ,7 5

0 ,8 5

0 ,9 5

1 ,0 5

1 ,1 5

0 1 2 3 4 5

Depth →

Grey level →

Slope

Offset

Schematic display of offset and slope realization

Along the attenuation line (figure І), each pixel was automatically

selected.

For each pixel under the line, the grey level value was averaged with the

grey level values of 3 pixels to the left and 3 pixels to the right

horizontally, and stored together with depth information. A linear

regression line was applied, using the least-squares approximation, and

its corresponding slope (grey level units/ mm) and offset (grey level)

were generated 19.

Additionally, the length of the attenuation line (number of pixels) was

generated.


129

Online Appendix ІІ. Realization of textural based indices

Indices of the spatial grey level dependence matrix 19, 21, 22, 24, 25, 35_ The aim of co-occurrence features is to capture

texture characteristics, i.e. heterogeneity. Elements of the co-occurrence matrix (algorithm [1]) designate the

probability that two pixels located within a region, separated by distance d along direction θ , have grey level

values of i and j:

Co-occurrence: ( ) ( ) ( )( ) ( ) ( )( ) [ ]yxd

LLnmlknmlk

djiP ,,,,,,,

,, ∈Ν

Ν=|

θθ , [1]

where θ = 0º and d = 4 pixels 19, θdΝ is the number of pixel pairs, and Ν is the number of grey level

transitions, in a region (Lx, Ly).

• Co-occurrence entropy: ( ) ( )( ) ,|,log, |,1

0

1

0

θθ djiPdjiPGG

ji

∑∑−Ν

=

−Ν

=

⋅ , [2]

• Co-occurrence sum entropy: ( ) ( )( )∑ ⋅k

sumsum kPkP log , [3]

where ( ) ( )θ,|,1

0

1

0

djiPkPGG

ji

sum ∑∑−Ν

=

−Ν

=

= for i + j = k.

Indices of the grey level difference matrix 22, 24, 25_ The aim of difference features is to capture texture

characteristics, i.e. homogeneity. Elements of the difference matrix (algorithm [4]) designate the probability that

after a displacement along vector δ within a region, pixels will have grey level value i:

o Difference: ( ) ( )( ) [ ]yx LLiyxIPif ,,|' ∈== δδ , [4]

where ( )yx ∆∆= ,δ and ( ) ( ) ( ) |,-yx,|, yyxxIIyxI ∆+∆+=δ , in a region (Lx, Ly).

• Difference contrast: ( )δ|'1

0

2if

G

ii∑

−Ν

=

[5]

• Inverse difference moment: ( )

∑−Ν

= +

1

02

1

|'G

i i

if δ [6]

• FP1:

MN

yxIM

x

N

y

∑∑= =1 1

),(δ

, where M and N are column and row, respectively. [7]

130

131

Chapter 6

Assessment of the variations in fat content in

normal liver using a fast MR Imaging method in

comparison with results obtained by Spectroscopic

Imaging

Eur Radiol 2008; 18(4):806-813.

Roy Irwan 1

Mireille A. Edens 2

Paul E. Sijens 1

Department of Radiology 1 Department of Epidemiology 2



Chapter 6.

132

ABSTRACT

A recently published Dixon-based MRI method for quantifying liver fat content using dual-

echo breath-hold gradient echo imaging was validated by phantom experiments and

compared with results of biopsy in two patients [Radiology 2005;237:1048-1055] 1. We

applied this method in ten healthy volunteers and compared the outcomes with the results

of MR spectroscopy (1H MRS), the gold standard in quantifying liver fat content. Novel

was the use of spectroscopic imaging yielding the variations in fat content across the liver

rather than a single value obtained by single voxel 1H MRS. Compared with results of 1H

MRS, liver fat content according to MRI was too high in nine subjects (range 3.3-10.7% vs.

0.9-7.7%) and correct in one (21.1 vs. 21.3%). Furthermore, in one of the ten subjects the

MRI fat content according to the Dixon-based MRI method was incorrect due to a (100-x)

versus x percent lipid content mix-up. The second problem was fixed by a minor

adjustment of the MRI algorithm. Despite systematic overestimation of liver fat contents by

MRI, Spearman's correlation between the adjusted MRI liver fat contents with 1H MRS was

high (r = 0.927, P < 0.001). Even after correction of the algorithm, the problem remaining

with the Dixon-based MRI method for the assessment of liver fat content, is that, at the

lower end range, liver fat content is systematically overestimated by 4%.

MRI to quantify hepatic fat content

133

INTRODUCTION

Substantial accumulation of fat in the liver (steatosis) can progress into steatohepatitis and

cirrhosis, a spectrum range called fatty liver disease 2, 3. In the general adult Western

population, the prevalence of fatty liver disease (alcoholic plus non-alcoholic) ranges from

30% to 34% 4-6. Evidence of non-alcoholic fatty liver disease (NAFLD) being an early

marker of cardiovascular disease is increasing, and the condition might even be an early

mediator in cardiovascular disease 7. Additionally, an accurate knowledge of liver fat

content can be important in case surgeons want to make sure that steatosis in donor livers

does not exceed 20% 8.

The quantification of liver fat fraction using magnetic resonance imaging (MRI) has gained

attention in the last decade 1, 9, 10, as ultrasound imaging is not considered to be a

sufficiently quantitative tool and computed tomography leads to undesired radiation

exposure in examinations in patients who may have no complications from their fatty liver

11. Among the existing MRI methods for quantifying liver fat content, the recently

presented adaptation of the two-point Dixon method using in-phase (IP) and out-phase (OP)

images, appears accurate 1. However, this fast method for quantifying liver fat content by

the dual-echo breath-hold gradient echo imaging method has not yet been validated in vivo,

apart from being compared with the results of histology in two patients 1. We used 1H-

Magnetic Resonance Spectroscopy (1H MRS), a more time consuming method allowing for

direct quantification of water and fat signals in a selected volume of interest. 1H MRS is

considered to be the gold standard for liver fat quantification in vivo 12 and has recently

been improved by implementing multi-voxel rather than single-voxel measurements 13.

In this study we aimed to compare the above mentioned fast MRI method for quantifying

liver fat content 1 with the results of multi-voxel 1H MRS, obtained in ten healthy

volunteers. Novel is the assessment of the variations in fat content across the liver by

spectroscopic imaging acquisition of a plane of voxels, yielding a plane of water-to-fat

ratios indicative of the heterogeneity in liver fat distributions, rather than using single voxel

1H MRS for sampling a small single volume. In the process, we found a limitation in the

algorithm proposed for estimating fat content causing an invalid value in one of our

volunteers. We propose a simple modification in the algorithm to deal with this problem

Chapter 6.

134

and show that the results of MRI, while correlating well with those of 1H MRS,

systematically overestimate liver fat content in the lower end range.

Figure І. Transverse T1-weighted MR

image of volunteer 9, showing a field of

view of 16×16 cm2 and a volume of interest

(VOI) of 5×8×4 cm3 positioned inside the

liver (the measurements were restricted to

the 32 entire voxels within the blue line).

Figure ІІ. Example of a ROI which was

drawn on both intermediate weighted and

T1-weighted dual-echo image pairs

matching to the whole VOI of the CSI (the

same volunteer as figure І).

Figure ІІІ. Spectral map showing water and fat peaks in the 32 quantified voxels in the

same volunteer as figures І and ІІ.


135

MATERIALS AND METHODS

VOLUNTEER STUDY

Studies were conducted on ten healthy volunteers (all males; mean age = 47 years; range:

22 – 58 years). The Body Mass Index of the volunteers ranged from 20 to 33.5 kg/m2 with a

mean value of 26.4 kg/m2. The studies were performed with informed consent and medical

ethical committee approval.

All volunteers were examined by MRI and 1H MRS in one measurement session, using a

1.5 Tesla whole-body scanner (MAGNETOM Avanto; Siemens Medical Solutions,

Erlangen, Germany) equipped with high performance gradients (maximal gradient strength

= 40 mTm-1; maximal slew rate = 200 mT m-1/ms) and a six-channel spine array coil in the

supine position with the spine along the symmetry axis of the coil system. In both MRI and

1H MRS a large flex coil placed over the liver was used simultaneously with the spine array

coil as receiver. To assess the anatomy of the liver, T1-weighted gradient-echo images were

recorded.

MR IMAGING

MRI of the liver was performed by using a breath-hold dual-echo T1 weighted gradient

echo sequence with a 6 mm slice thickness, section gap 0 mm, matrix 256x160 and a

repetition time (TR) of 155 ms. Dual-echo spoiled gradient recalled images were acquired

with TE = 2.4 ms (OP) and TE = 4.8 ms (IP) and flip angles of 70o and 20o to generate T1-

weighted and intermediate-weighted images, respectively. These images were corrected for

T2* decay using

*2Tcorrected eSS

τ

= (equation 1)

where τ is the echo time difference between IP and OP images, and S represents the signal

intensity in a defined region of interest (ROI) 1. Under these conditions τ = 2.4 ms

combined with T2* = 19.44, calculated from the mean spectral line width of the water peak

in human liver measured by 1H MRS in the ten volunteers (details are given in a next

paragraph), gave a correction factor of 1.13 for Sip relative to Sop.

Chapter 6.

136

The published algorithm for estimating fat content by Hussain et al. 1 consists of the

following three steps:

(step 1) adjustment for T2* relaxation using equation 1.

(step 2) calculation of the apparent fat content using the following equation:

( )%100

2% ×

−=

IP

OPIP

S

SSfat

(equation 2)

for both intermediate or hydrogen density weighted (%fatHwt at 20o flip angle)

and T1-weighted (%fatT1wt at 70o flip angle) image pairs.

(step 3) if %fatHwt ≤ %fatT1wt, then %fat = %fatHwt; otherwise, %fat = 100% - %fatHwt

In this study we propose to modify the third step in the Hussain algorithm for estimating fat

content to:

(step 3corrected) if fatHwt AND %fatT1wt ≤ 20%, then %fat = MIN[%fatHwt, %fatT1wt],

if fatHwt AND %fatT1wt > 20% and %fatHwt ≤ %fatT1wt, then %fat = %fatHwt,

otherwise, %fat = 100% - %fatHwt,

where AND is a logical operator, and MIN[a, b] is a mathematical

operator computing the minimum value between a and b.

MR SPECTROSCOPY

The flex large and spine coils used for MRI were also used for 1H MRS. The subjects were

scanned in the supine position. Movement of the liver with breathing was not suppressed as

in the past we have shown that the quality of liver MR spectra is not affected by this 13.

Hybrid 2D-spectroscopic imaging (chemical shift imaging, CSI) 14, point resolved

spectroscopy (PRESS) with a repetition time (TR) of 5000 ms and an echo time (TE) of 30

ms, was performed using a field of view of 16×16 cm2 and a volume of interest of 5×8×4

cm3 positioned inside the liver (figure І). The CSI measurement lasted 16×16×5 = 1280 s or

approximately 21 min. Shimming was automated and water suppression was not applied in

order to be able to calculate the fat-water ratio distributions in the liver directly.


137

The water-fat analysis was restricted to the transverse plane of 4×8 = 32 entire voxels. Fat

to water ratios, were defined as usual 6, 12, 15-17. The ratio of the curve fitted -CH2- lipid

signal peak area (1.3 ppm) divided by the sum of the same lipid signal and that of H2O (4.7

ppm), is equal to the weight fat/(fat+water) ratio because the relative hydrogen contents of

water and fat are very similar (approximately 11%) 12, 15. The obvious variation of B1

sensitivity over the CSI field affected the water and lipid signals similar, and therefore was

not influencing the

1H MRS results expressed in water-to-fat ratio plus (inter-voxel) standard deviation. Using

standard postprocessing-software provided by the manufacturer, the water and fat peak

areas of all 32 voxels were fitted in the frequency domain with assumption of Lorentzian

line shapes preceded by phase correction and, occasionally, adjustment of the base line.

Determination of the fat contents for each of the above mentioned 32 1H MRS voxels thus

led to estimates of the mean value and heterogeneity (standard deviation) in the liver fat

content of the volunteers. At the used TR of 5 s, T1 saturation of the water and fat signals is

negligible, that is TR > 5T1 12, and at the used TE of 30 ms the correction applied to our

data to compensate for the fact that the fat signal has a longer T2 (78 ms) than water (60 ms)

was 12.2 % 15. The CSI series lasted approximately 21 minutes as compared with less than

5 minutes required for the two series yielding the four sets of images used for liver fat

quantification.

The T2* used for correcting Sip in the above mentioned MRI algorithm, was calculated from

the CSI spectral maps by fitting the line widths of the liver water signals of 16 voxels from

the centers of the volumes of interest to obtain a mean water line width for every subject.

The mean water line width (∆ν) for the group of 10 was 17.05 ± 3.74 Hz.

νπ ∆⋅=

1*2T

(equation 3)

Using equation 3, a mean T2* of 19.4 ± 4.1 ms, was calculated. This value was used for

correcting all MRI fat contents of Table 1. As a double check we also used a gradient-

recalled echo sequence with two IP echoes (4.5 and 18 ms) in analogy to a method applied

by Hussain et al. 1, obtained similar mean T2* values in the ten volunteers (20 ± 3 ms).

Chapter 6.

138

DATA PROCESSING

To co-register the 1H MRS data to the MR images, the coordinates of the anterior right

(AR) and posterior left (PL) at the same height position were used as regional checkpoints.

These coordinates can be determined from the Siemens viewing workstation. Figure І

illustrates a T1-weighted image overlaid by CSI voxels with the starting voxel in the upper

left corner of the blue VOI (i.e. right anterior in the liver) matching the first quantified 1H

MRS voxel. Once the AR and PL were defined, an ROI was drawn on both intermediate

weighted and T1-weighted dual-echo image pairs, as demonstrated in figure ІІ.

Statistical Analysis

Measured MRI data were imported in an Excel Spreadsheet presented as mean ± standard

deviation (SD) of 7 slices. The value for each slice was obtained by averaging 4 samples

per dual-echo image pair. The distributions of the obtained liver fat contents were not

normal, so non-parametric correlations with 1H MRS determined liver fat content were

evaluated by assessment of Spearman's correlation coefficient (with 2-tailed testing of

significance). Note that the standard deviations in the MRI data reflect the spread

(heterogeneity) amongst the seven MRI slice sections contained in the 8x4x4 cm3 1H MRS-

analysis volume of interest, whereas the standard deviations in the 1H MRS determined

liver fat contents reflect the spread amongst 32 1H MRS voxel subvolumes of 1x1x4 cm3.

Automated matching of the 7 summarized MRI slices to the dimensions of the 1H MRS

voxels to obtain the IP and OP values for each 1H MRS voxel was unfortunately not

available for use with the Avanto MRI system. It was therefore not possible to obtain MRI

and 1H MRS standard deviations for the same sets of VOI subvolumes or assess the

MRI/1H MRS correlations for all 320 evaluated voxels rather than for the 10 averaged sets

of results per subject.


139

Table І. MRI fat measurements (means ± SD of 7 slices; T2* corrected) according to the intermediate and T1 weighted gradient echo sequences, Hussain’s algorithm 1, our corrected algorithm and the gold standard 1H MRS (mean ± SD of 32 voxels; T2 corrected). Results shown in ascending order of 1H MRS measurements.

Liver fat content (%)

Volunteer

Intermediate

weighted

MRI (20˚)

T1

weighted

MRI (70˚)

Hussain

algorithm 1

Corrected

algorithm 1H MRS

1 4.0 ± 2.8 4.7 ± 2.1 4.0 ± 2.8 4.0 ± 2.8 0.9 ± 0.3

2 3.3 ± 0.4 3.9 ± 0.5 3.3 ± 0.4 3.3 ± 0.4 1.2 ± 1.2

3 3.9 ± 0.3 5.1 ± 0.3 3.9 ± 0.3 3.9 ± 0.3 1.2 ± 0.8

4 8.3 ± 2.8 8.4 ± 1.2 8.3 ± 2.8 8.3 ± 2.8 2.2 ± 0.8

5 7.9 ± 1.9 9.2 ± 1.4 7.9 ± 1.9 7.9 ± 1.9 2.3 ± 0.7

6 11.6 ± 1.1 10.7 ± 2.6 89.3 ± 2.6 10.7 ± 2.6 5.1 ± 1.5

7 10.9 ± 1.8 12.5 ± 1.0 10.9 ± 1.8 10.9 ± 1.8 5.3 ± 1.4

8 11.1 ± 0.9 11.4 ± 2.3 11.1 ± 0.9 11.1 ± 0.9 5.8 ±1.9

9 10.2 ± 0.6 14.5 ± 0.6 10.2 ± 0.6 10.2 ± 0.6 7.7 ± 2.3

10 21.1 ± 0.7 29.8 ± 0.3 21.1 ± 0.7 21.1 ± 0.7 21.3 ± 5.5

mean

(mean SD) 9.23 (1.33) 11.02 (1.23) 17.0 (1.48) 9.14 (1.48) 5.40 (1.64)

Spearman's

correlation with 1H MRS

rs = 0.839

P = 0.002

rs = 0.967

P = 0.000

rs = 0.790

P = 0.007

rs = 0.927

P = 0.000

-------

-------

Chapter 6.

140

RESULTS

An example of a volunteer CSI measurement is shown in figure 3. Each of the 32 quantified

1H MRS voxels reveals a large water peak and a small (-CH2-)n fat peak. The B1 field,

strongest in the center of the volume of interest, affects the water and fat signals similarly.

The variation in the calculated water-to-fat ratios amongst the 32 1H MRS voxels in the ten

volunteers is reflected by the standard deviations in column 6 of table І. The MRI

determined fat contents according to the pairs of intermediate- and T1 weighted gradient

echo sequences measured at TE's of 2.4 and 4.8 ms are shown in columns 2 and 3,

respectively (table І). Application of the algorithm published by Hussain et al. 1 yields

column 4 and our modification of the Hussain algorithm, described in 'Materials and

Methods', yields column 5. In one case the intermediate weighted MRI (20˚) result was

higher than the T1 weighted MRI (70˚) result (volunteer 6), resulting in a liver fat content

of 89.3% for the Hussain algorithm as opposed to the corrected algorithm yielding a content

of 10.7%.

In the 1-21% liver lipid content range (by 1H MRS) of our volunteer study, correlation

between our corrected MRI algorithm results and 1H MRS determined liver fat content is

high (r = 0.927), but the correlation between the T1 weighted sequence results alone with 1H

MRS fat content is still better (r = 0.967). In the range 1-10%, the MRI determined liver fat

contents (corrected algorithm) are systematically higher, on average 4% (range: 2.1% -

6.1%), than those obtained with 1H MRS (table І).

Figure ІV shows that the linear fit lines for the intermediate weighted series and of the

corrected algorithm are similar, and that it makes little difference whether or not the mean

T2* (19.4 ms) or the individual T2* of each subject (range: 12.7-26.5 ms) is used for the

correction of Sip (compare figure ІVa and ІVb).

y = 1,2107x + 4,6033

y = 0,8061x + 4,8678

y = 0,8055x + 4,9606

05

10

15

20

25

30

35

05

10

15

20

25

30

35

Liver fat content according to MRS (%)

Liver fat content according to MRI (%)

y = 1,2926x + 4,2593

y = 0,9119x + 4,4367

y = 0,9114x + 4,5296

05

10

15

20

25

30

35

05

10

15

20

25

30

35

Liver fat content according to MRS (%)

Liver fat content according to MRI (%)

a. M

RI

dat

a ca

lcula

ted u

sing T

2* =

19.4

4 m

s b.

MR

I dat

a ca

lcula

ted

usi

ng

the

exper

imen

tally

det

erm

ined

T2* o

f ea

ch i

ndiv

idual

subje

ct (

range

12.7

- 26.5

ms)

Figure ІV. Sca

tter

plo

t of

the

MR

I det

erm

ined

fat

per

centa

ges

as

a fu

nct

ion o

f liver

fat

conte

nt

acco

rdin

g t

o 1

H M

RS (

table

1),

incl

udin

g

linea

r fi

t lines

.

o =

T1-w

eighte

d M

RI

seri

es (

70°)

; □

= I

nte

rmed

iate

wei

ghte

d M

RI

seri

es (

20°)

; ∆

= C

orr

ecte

d a

lgori

thm

usi

ng T

1-w

eighte

d a

nd I

nte

rmed

iate

wei

ghte

d

MR

I se

ries

.


141

Chapter 6.

142

DISCUSSION

The promising algorithm of Hussain et al. for liver fat quantification using a fast two-point

Dixon method using in-phase and out-phase images had not yet been validated in vivo,

apart from being compared with the results of biopsy in two patients 1. In this study we

used 1H MRS, a more time-consuming method allowing for direct quantification of water

and fat signals in a selected volume of interest and considered to be the gold standard for

liver fat quantification in vivo 12, to assess the above mentioned MRI method in ten healthy

subjects. Novel was the assessment of the variations in fat content across the liver by

chemical shift imaging acquisition of a plane of voxels, yielding a plane of water-to-fat

ratios indicative of the heterogeneity in liver fat distributions, rather using single voxel 1H

MRS for sampling a small single volume.

We found that the Hussain algorithm for estimating fat content gave an invalid value in one

of ten subjects (table І: volunteer 6). In this subject with a liver fat content of 5.1%

according to 1H MRS, the intermediate weighted value was larger than the T1 weighted

value, in contrast to the other nine subjects. Obviously, Hussain et al. had patients with

higher liver fat contents in mind when they published the original article 1. Figures 3 and 4

of that publication are characteristic since the range of 0-10% was either not measured or

the estimated fat content deviated substantially. This does not reduce the importance of that

publication. However, studies in nutritional physiology and diabetes research often deal

with volunteers who have much lower liver fat content than the patients envisioned by

Hussain et al. These studies in volunteers are extremely important for our physiological and

pathophysiological knowledge and by our algorithm modification we have therefore

improved the Hussain method in an important range of liver fat content.

An important difference between the current and the Hussain report is that the latter showed

both theoretically and experimentally that the intermediate-weighted double echo set yields

a better estimate of liver fat content than the more T1-weighted data set, whereas we report

the opposite (correlation with 1H MRS: r=0.84 and r=0.97, respectively). This appears to

mainly reflect that in our case the lower-end liver fat range of 0-15% is overrepresented.

Note that in their data the IP and OP signals also approach each other when liver fat content

decreases towards 10% and even appear to cross near 3% (figure 4 in 1). Discussed below is

our proposed correction for water/lipid mix-ups and the, regardless of the proposed


143

correction, overestimation of liver fat content by dual echo breath-hold MRI as compared

with 1H MRS.

Correcting the MRI algorithm for (100-x)% versus x% lipid mix-ups

The introduction of the (100-x) versus x swap by Hussain et al. 1 served to compensate that

for high lipid content (˜50%) the dominant frequency might be swapped. This was resolved

based on the insight that the high flip angle acquisition gives stronger T1-weighting, leading

to higher lipid signal since fat T1 is shorter than water T1. Therefore the high flip angle

sequence serves to discriminate the two compartments in each case 1. We have shown that

for low lipid content the algorithm of Hussain at al. may fail (volunteer 6). This is probably

mainly due to the fact that at low fat concentration the signal is dominated by water and the

standard deviation of the water signal might be easily larger than the small fat distribution.

This can be understood by considering the standard deviations in the liver fat contents listed

in table І. In the MRI results the standard deviations refer to the spread amongst the 7 MRI

slices adding up to the liver volume analyzed by 1H MRS. The MRI standard deviations

vary between approximately 5% of the mean value in the subjects with the highest fat

content (volunteers 9,10) up to 70% in the subject with the lowest fat content (volunteer 1).

In the intermediate weighted and T1 weighted MRI fat contents of Volunteers 1, 2, 4, 5, 6, 7

and 8 the standard deviations are of similar magnitude as the fat content difference,

suggesting that with some bad luck the Hussain algorithm would have failed in 7 out of 10

volunteers having liver fat contents of 0.9 to 5.8% according to 1H MRS. Although

theoretically at the lower end range the intermediate weighted (Hwt) should yield a lower

fat content than the T1 weighted series 1, this is not always the case in clinical practice.

Equation 2 actually indicates that with decreasing liver fat contents, Sip and Sop approach

each other, leading to calculated fat contents of decreasing accuracy. To summarize, in

subjects with low liver fat contents tissue heterogeneity and/or measurement accuracy

limitations can cause the intermediate series to come up with a higher fat content than the

T1-weighted series.

We have modified the Hussain algorithm in order to eliminate (100-x)% lipid versus x%

lipid mix-ups encountered in the lower end range of liver fat contents. The validity of the

algorithm should now extend to the range 0-80%, actually covering any liver fat content

encountered in healthy volunteers as well as patients in studies reported to date. With

Chapter 6.

144

regard to the clinical impact of our modification correcting for (100-x)% lipid versus x%

lipid mix-ups in the Hussain method of liver fat quantification by MRI 1 one could argue

that radiologists who interpret abdominal images would not confuse a patient with 10%

lipid and 90% lipid. Although this is correct in view of the knowledge that lipid contents of

90% are encountered in subcutaneous fat but never in the liver 6, it is beyond doubt that

quantitative methods yielding correct outcomes are preferable to those that come up with

invalid results once in a while. Our approach is empirical in that there is no physical basis

for the modified formula. In other words, the 20% cut-off is an ad-hoc threshold value that

can be adjusted downward with improving signal-to-noise ratios in the IP and OP gradient

echo data (as can be achieved by scanning longer, with more sensitive RF coils, at higher

magnet field strength, etc.).

Overestimation by MRI of the liver fat content in the lower end range

Another issue is why the MRI assessed liver fat contents are systematically higher than the

1H MRS results (+ 4%), at least in the range 1-10%. Note that this problem is not solved by

our proposed correction of the Hussain algorithm. This is best visualized in figure ІV

showing all experimental MRI fat contents above the dotted line and linear fit lines crossing

the y-axis near 4.5% rather than 0%. The considerable standard deviations in the 1H MRS

data, roughly 30% and higher in the subjects with the lowest fat contents, though no doubt

mainly representing the heterogeneity in the distributions of liver fat contents, will to some

extent also reflect decreasing signal-to-noise ratios in the curve fitted fat signals and

possibly displacement of the liver by breathing resulting in spectra not exactly coming from

the specified volumes of interest. Nevertheless the discrepancy appears to be systematic,

resembling the effect of incorrect T2* correction illustrated by simulations (figures 1 and 2)

in the publication of Hussain et al. 1. If T2* is taken too small, this results in underestimated

liver fat contents in the lower end range. This leads to the question whether our

overestimations of MRI fat contents in the same range of fat contents might thus reflect our

1H MRS determined T2* values being not representative for the true T2* in the MRI series.

We excluded this possibility by using a gradient-recalled echo sequence with two IP echoes

(4.5 and 18 ms) in analogy to a method applied by Hussain et al. 1, to obtain similar mean

T2* values in the ten volunteers (20 ± 3 ms according to MRI versus 19 ± 4 ms according to

1H MRS).


145

So an incorrect T2* correction can not explain the 4.3% overestimation of liver fat content

by MRI. Noted is that the linear coefficient of correlation between corrected MRI fat

content and 1H MRS fat content is high (r = 0.927; P < 0.001), though not better than the

correlation of T1 weighted MRI alone with 1H MRS fat content (r = 0.967; P < 0.001).

Recently Biglands et al. 18 showed a similar discrepancy between the results of a Dixon-

based MRI method resembling that of Hussain et al 1 and liver fat contents according to

histology data: in the liver fat range of 0-10% the MRI fat fractions systematically were 3%

too high. The too high liver fat contents obtained by MRI might reflect an overestimation in

the MRI quantification due to T1 effects. Another plausible hypothesis might be that

imaging and CSI look at not entirely the same compartments. 1H MRS, although considered

to be the gold standard for liver fat quantification in vivo 12, only detects narrow line widths

and might thus fail to quantify broad very short T2* lipid components which vanish in the

baseline, whereas imaging is sensitive for a larger range of T2*. The fat image, on the other

hand, will also include some broad signal from water resonances leaking into the fat signal.

Most importantly, the method presented here and in the Hussain paper combines signal

intensities measured at a given location in space during two different breath-hold periods in

a single algorithm. Thus, spatial miss-registration between the two data sets is expected to

lower the actual resolution of any fat-content map by one order of magnitude compared to

that of the original images - at least in clinical routine with patients.

Imprecise fat-content estimations by the in-/opposed-phase method at both the low (0%)

and high (100%) end are inherent to the method and have been repeatedly pointed out.

Several studies have indicated that other MRI methods of liver fat determination, which do

not suffer from (100-x)% lipid versus x% lipid mix-ups in need of correction by

implementation of system dependent threshold values, might be more accurate. Machan et

al. used chemically selective saturation of liver water and fat signals obtained good

correspondence between MRI fat contents in the range 0-50% and the results of 1H MRS 10.

Cotler et al. used a similar MRI method and obtained good correspondence between MRI

fat contents in the lower end range (0-25%) and the results of optical image analysis 19.

Chapter 6.

146

CONCLUSION

A correction to the Hussain algorithm for estimating liver fat content is proposed to prevent

invalid outcomes in subjects who have higher fat contents according to the intermediate

weighted gradient echo MRI series than according to the T1 weighted MRI series. After

implementation of the correction, the correlation between the MRI and 1H MRS outcomes

is very good, indicating that the method can also be used for determining liver fat contents

in subjects with low or moderately elevated liver fat contents.

Further studies on larger study populations are needed to confirm the validity of this

method in a broader range of liver fat content. An advantage in clinical practice of being

able to use the improved Hussain method rather than multiple voxel 1H MRS would be that

the MRI method is faster (5 minutes versus 21 minutes). A problem remaining with the

Dixon-based MRI method is that at the lower and range the liver fat contents are

systematically overestimated as compared with 1H MRS (present study) or results obtained

at histology 18. Alternative MRI methods using chemically selective saturation might be

preferable 10, 19 and should in future studies be compared with the results of spectroscopic

imaging.


147

Reference List

(1) Hussain HK, Chenevert TL, Londy FJ, Gulani V, Swanson SD, McKenna BJ, Appelman HD, Adusumilli S, Greenson JK, Conjeevaram HS. Hepatic fat fraction: MR imaging for quantitative measurement and display--early experience. Radiology 2005 December;237(3):1048-55.





(5) Browning JD, Szczepaniak LS, Dobbins R, Nuremberg P, Horton JD, Cohen JC, Grundy SM, Hobbs HH. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. Hepatology 2004 December;40(6):1387-95.



(7) Targher G, Bertolini L, Padovani R, Zenari L, Zoppini G, Falezza G. Relation of nonalcoholic hepatic steatosis to early carotid atherosclerosis in healthy men: role of visceral fat accumulation. Diabetes Care 2004 October;27(10):2498-500.

(8) Kim SH, Lee JM, Han JK, Lee JY, Lee KH, Han CJ, Jo JY, Yi NJ, Suh KS, Shin KS, Jo SY, Choi BI. Hepatic macrosteatosis: predicting appropriateness of liver donation by using MR imaging--correlation with histopathologic findings. Radiology 2006 July;240(1):116-29.

(9) Kawamitsu H, Kaji Y, Ohara T, Sugimura K. Feasibility of quantitative intrahepatic lipid imaging applied to the magnetic resonance dual gradient echo sequence. Magn Reson Med Sci 2003 April 1;2(1):47-50.

(10) Machann J, Thamer C, Schnoedt B, Stefan N, Haring HU, Claussen CD, Fritsche A, Schick F. Hepatic lipid accumulation in healthy subjects: a comparative study using spectral fat-selective MRI and volume-localized 1H-MR spectroscopy. Magn Reson Med 2006 April;55(4):913-7.

Chapter 6.

148

(11) Valls C, Iannacconne R, Alba E, Murakami T, Hori M, Passariello R, Vilgrain V. Fat in the liver: diagnosis and characterization. Eur Radiol 2006 October;16(10):2292-308.


Imaging 1994;12(3):487-95.


(14) Sijens PE, van den Bent MJ, Nowak PJ, van Dijk P, Oudkerk M. 1H chemical shift imaging reveals loss of brain tumor choline signal after administration of Gd-contrast. Magn Reson Med 1997 February;37(2):222-5.

(15) Longo R, Ricci C, Masutti F, Vidimari R, Croce LS, Bercich L, Tiribelli C, Dalla Palma L. Fatty infiltration of the liver. Quantification by 1H localized magnetic resonance spectroscopy and comparison with computed tomography. Invest Radiol 1993 April;28(4):297-302.

(16) Tiikkainen M, Bergholm R, Vehkavaara S, Rissanen A, Hakkinen AM, Tamminen M, Teramo K, Yki-Jarvinen H. Effects of identical weight loss on body composition and features of insulin resistance in obese women with high and low liver fat content. Diabetes 2003 March;52(3):701-7.


(18) Biglands JD, Wilson D, Ward J, Treanor D, Gurthry A, Nijhawan A, Smith J, Wyatt J, Robinson P. Comparison of MRI and histopathalogic methods of quantifying hepatic fat fraction. In Proc Intl Soc Magn Reson Med , May 19-25,

2007, Berlin p2703 2007.

(19) Cotler SJ, Guzman G, Layden-Almer J, Mazzone T, Layden TJ, Zhou XJ. Measurement of liver fat content using selective saturation at 3.0 T. J Magn Reson

Imaging 2007 April;25(4):743-8.

Chapter 7

MRI determined fat content of human liver,

pancreas and kidney

World J Gastroenterol 2010; 16(16):1993-1998.

Paul E. Sijens 1

Mireille A. Edens 2

Stephan J. Bakker 3

Ronald P. Stolk 2

Department of Radiology 1

Department of Epidemiology 2 Department of Internal Medicine 3



Chapter 7.

150

ABSTRACT

Aim: To assess and correlate the lipid contents of various organs in obese subjects and in

persons with a normal body weight.

Methods: Magnetic resonance spectroscopy and a previously validated gradient echo

magnetic resonance imaging method with Dixon’s two point technique were used in this

study to quantify fat in liver, pancreas as well as kidney.

Results: In 36 volunteers with body mass index (BMI) ranging from 20.0 to 42.9 kg/m² the

median fat contents of liver, pancreas and kidney were 2.3% (interquartile range: 0.2%-

7.8%), 2.7% (1.0%-6.5%) and 0.7%(0.1%-1.4%), respectively. BMI and subcutaneous fat

correlated significantly with liver and pancreas fat contents. Shown for the first time is

significant correlation of the liver and pancreas fat contents in healthy controls (r = 0.43, P

< 0.01). These observations are related to body weight as measured by BMI and the amount

of subcutaneous fat. Kidney fat content is very low and correlates with none of the other fat

depots.

Conclusion: Renal lipid accumulation, unlike the coupled accumulations of fat in liver and

pancreas, is not observed in obese subjects. Unlike what has been suggested in previous

studies, renal lipid accumulation appears not to be involved in the pathogenesis of renal

disease in humans.

MRI-determined fat content of human tissues

151

INTRODUCTION

Storage of fat within the peritoneal cavity (visceral fat) and within other tissues (ectopic fat)

rather than in subcutaneous adipocytes, is accompanied by adverse metabolic and lipotoxic

effects 1, 2. Both visceral fat 1 and liver fat 3 are known for secreting numerous atherogenic

factors into the blood system, e.g. lipids and inflammatory cytokines, predisposing to

cardiovascular disease, amongst others atherosclerotic renovascular disease (ARVD) 4, 5. It

has been hypothesized that in the case of obesity and metabolic syndrome, ectopic fat

causes lipotoxic damage to organs 4, 5, e.g. liver (steatohepatitis and cirrhosis) 6, pancreas

(β-cell dysfunction) 7 and lipotoxicity of the kidney 4.

Whereas liver steatosis has been quantified in dozens of MRI studies, and pancreatic

lipomatosis 8-10 and muscle fattening 8-11 in some, quantitative documentation of the content

of fat in human kidneys in situ appears to be lacking entirely. In rats, however, steatosis of

the kidney was recently associated with an alteration in renal acidification 12. In that study,

the fat accumulated in the renal cortex as shown quantitatively by enzymatic triglyceride

measurement and qualitatively by oil red O staining 12. This fits the notion that, within the

cortex, the proximal tubule is vulnerable to lipid accumulation due to its role in the

reabsorption of free fatty acids bearing albumin 13, 14.

The most accurate, and therefore preferable, method for quantifying hepatic fat content is

metabolic MRI, i.e. multivoxel MR spectroscopy (1H MRS) 15-17. Other MRI methods are

less reproducible or have systematic errors. In a previously published comparison of

subjects examined by 1H MRS and by a Dixon-based dual-echo breath hold gradient echo

method, we found strong correlation between the liver fat contents according to both

methods and concluded that MRI can be used for quantifying fat content, provided that the

small systematic overestimation of fat contents at the lower end range is corrected for 17.

The comparison is now extended to 36 volunteers, varying from lean to obese, with

application of the Dixon-based MRI fat quantification method to pancreas, kidney as well

as liver. Only the liver was also examined by 1H MRS. Our approach was to apply the

equation converting the liver MRI fat values to the 1H MRS values, also to the pancreas and

kidney MRI values. Additional 1H MRS examinations of pancreas and kidney would have

made the total MRI examination times too lengthy. Furthermore, reliable 1H MRS of the

Chapter 7.

152

renal cortex, a thin layer of tissue sandwiched between the peripheral fat and the fatty core

of the kidney, is technically not feasible.

High hepatic lipid contents are common in obese subjects who are otherwise healthy, i.e.

overweight persons not suffering from metabolic disease 15. In this study we examined 36

volunteers with a wide range of body mass index (BMI) values, by previously validated

MRI methods to quantify the fat contents of liver, pancreas as well as kidney. The purpose

was to quantify the lipid contents of these organs in obese subjects and in persons with a

normal body weight. Our hypothesis was that in obese persons the lipid contents of liver,

pancreas and kidney would be higher than in thin persons. Our comparatively large study

was also intended to assess possible relationships between the fat contents in different

organs.

Іa. Liver and pancreas Іb. Kidney and subcutaneous tissue measurement

Іc. Liver 1H MRS spectral map, showing an array of

water and fat peaks

Figure І. Regions of interest on MRI (OP series with 70º pulse angle and TE=2.4 ms) and

the corresponding 1H MRS.


153

MATERIALS AND METHODS

Thirty-six adult volunteers were recruited by advertisement, to undergo abdominal MRI

and liver 1H MRS. These volunteers had a body mass index (BMI) ranging from 20.0 to

42.9 kg/m2 with a mean value of 27.5 kg/m2 and were average in that they were neither

profound exercisers nor sedentary. Eight out of 36 subjects were obese, i.e. they had BMI

values exceeding a value of 30. Two third of the participants were men, the median age was

39 years (range: 22 – 64 years). Everyone was interviewed to assess the status of his health

and cases of any disease, including diabetes, were excluded. All volunteers gave informed

consent. The studies were approved by the medical ethics committee.

MR STUDIES

MRI and 1H MRS studies took place in one measurement session, using a 1.5 Tesla whole-

body scanner (MAGNETOM Avanto; Siemens Medical Solutions, Erlangen, Germany)

equipped with gradients of up to 40 mTm-1 (maximal slew rate = 200 mT m-1/ms) and a six-

channel spine array coil. Subjects were in the supine position. In both MRI and 1H MRS a

large flex coil placed over the liver was used simultaneously with the spine array coil as

receiver. T1-weighted gradient-echo images were recorded to assess the anatomy of liver,

kidneys and pancreas.

DIXON’S TWO POINT MRI TECHNIQUE

Breath-hold T1 weighted gradient echo MR images, a dual flip angle adaptation of the wide

spread gradient recalled echo MRI method based on Dixon’s two point technique 18, were

acquired with 6 mm slice thickness, section gap 0 mm, matrix 256x160 and a repetition

time (TR) of 155 ms, and TEs of 2.4 ms (OP) and 4.8 ms (IP). Flip angles of 20o and 70o

were used to generate intermediate weighted and T1-weighted images, respectively. Images

were corrected for T2* decay:

*2Tcorrected eSS

τ

=,where τ is the echo time difference

between IP and OP images, and S represents the signal intensity in a defined region of

interest (ROI) 19. Under these conditions τ = 2.4 ms combined with T2* = 19.44, calculated

from the mean spectral line width of the water peak in human liver measured by 1H MRS in

Chapter 7.

154

the 36 volunteers (details are given in a next paragraph), gave a correction factor of 1.13 for

Sip relative to Sop.

The algorithm used for estimating fat content, modified to prevent occasional mix-ups of

water and fat signals 17, consists of:

(a) adjustment for T2* relaxation as described above,

(b) calculation of the apparent fat content using the following equation

( )%100

2% ×

−=

IP

OPIP

S

SSfat

(equation 1)

for both intermediate or hydrogen density weighted, (%fatHwt at 20o flip angle) and

T1-weighted (%fatT1wt at 70o flip angle) image pairs,

(c) if fatHwt AND %fatT1wt ≤ 20%, then %fat = MIN[%fatHwt, %fatT1wt],

if fatHwt AND %fatT1wt > 20% and %fatHwt ≤ %fatT1wt, then %fat = %fatHwt,

otherwise, %fat = 100% - %fatHwt,

where AND is a logical operator, and MIN[a, b] is a mathematical operator computing the

minimum value between a and b. A T2* of 19.4 ms for liver tissue was adopted from the

results of a previous study 17 for correcting Sip in the above mentioned MRI algorithm. This

value was used for correcting all MRI fat contents.

LIVER 1H MRS FOR CORRECTING ABDOMINAL MRI’S

Hybrid 2D-spectroscopic imaging (chemical shift imaging, CSI), point resolved

spectroscopy (PRESS) with a repetition time (TR) of 5000 ms and an echo time (TE) of 30

ms, was performed using a field of view of 16×16 cm2 and a volume of interest of 5×8×4

cm3 positioned inside the liver (figure 1c) 20. The CSI measurement lasted 16×16×5 = 1280

s or approximately 21 min. Shimming was automated and water suppression was not

applied in order to be able to calculate the fat-water ratio distributions in the liver directly

17. At the used TR of 5 s, T1 saturation of the water and fat signals is negligible, that is TR >

5T1 21, and at the used TE of 30 ms the correction applied to our data to compensate for the

fact that the fat signal has a longer T2 (78 ms) than water (60 ms) was 12.2 % 22. In our 1H

MRS method respiratory compensation was not applied as this was previously shown not to

affect the quality and composition of our liver spectra 17.


155

The liver 1H MRS data were co-registered to the MR images in the liver by using the

coordinates of the anterior right (AR) and posterior left (PL) at the same height position as

regional checkpoints. The kidney and pancreas MRI fat contents were derived from a single

slice centered on the organ involved. Figure І shows the regions of interest for liver and

pancreas (a), kidney and subcutaneous fat (b) and the with figure 1a matching liver 1H

MRS spectral map (c) for one of the volunteers examined. In the analysis of kidney data the

results for left and right kidney volumes of interest containing medulla and renal column

tissue were averaged. 1H MRS measurements of the renal cortex and pancreas were not

performed because, apart from measurement time considerations, accurate quantification of

small tissue volumes adjacent to the fatty tissue surrounding these organs is impossible.

The thickness of the layer of subcutaneous fat was measured at the two most left and right

points on the same MRI slice to obtain a measure of the amount of subcutaneous fat tissue

(figure Іb). A more precise method 23 was not feasible in our study, as the MRI field of

view was often too small to depict the entire abdominal body cross section.

The pancreas measurements were confined to the tail (cauda) because that tended to be

easier to depict than caput and corpus (figure Іa). The difference between the liver fat

content according to 1H MRS, the gold standard 17, and liver fat content according to MRI

was used to correct all MRI contents determined in this study (kidney, liver, pancreas). This

approach is reasonable because with the combined use of the flex and spine array as

receiver MRI coils, the B1 field in the positions corresponding with the locations of these

organs did not show significant differences. In other words, the outcome of phantom

experiments gave us reason to expect that equivalent Dixon’s two point technique MRI

signal intensities in liver, pancreas and kidney should correspond to the same fat contents.

STATISTICAL ANALYSIS

The distributions of the obtained fat contents were not normal, thus non-parametric

correlations were calculated using Spearman's correlation coefficients (with 2-tailed testing

of significance).

Chapter 7.

156

RESULTS

MRI CORRECTION

Mean liver fat content according to 1H MRS was 3.2% lower than that according to MRI

(4.4% vs 7.6%). The linear correlation between liver fat content according to MRI and 1H

MRS is shown in figure ІІ (r = 0.95; P < 0.0001). The obtained slope (y=1.216x-4.82) was

used to correct all MRI determined tissue fat contents.

y = 1,216x - 4,8186

R2 = 0,898

0

5

10

15

20

25

30

0 5 10 15 20 25

MRI fat content (%)

MRS fat content (%

)

MRI FAT MEASUREMENTS

After correction for this systematic overestimation, which is inherent to the MRI method 13,

calculated fat contents of liver and pancreas and kidney were 2.3% (inter-quartile range:

0.2%-7.8%), 2.7% (1.0%-6.5%) and 0.7%(0.1%-1.4%), respectively. In the kidney fat

measurements the variance between subjects was lower (inter-quartile range: 1.8%) than in

liver (7.6%) and pancreas (5.5%). Illustrated by figure ІІІ is that, compared with the normal

Figure ІІ. Liver fat content

according to 1H MRS plotted

against the fat content of the

same volume of interest

according to MRI.


157

body weight subjects, the subgroup of obese subjects tended to have comparatively high fat

in the liver (median 4.57% versus 1.11%), slight elevation in the pancreas (3.60% vs.

2.26%) and equally low fat contents in the kidney (1.35 versus 0.64). The inter-quartile

ranges for these organs in obese and normal weight persons showed great overlap and were

of comparable magnitude, i.e. 2.93-7.70 (0.18-7.76), 2.32-8.88 (0.62-5.45) and 0.61-1.49

(0.10-1.17), respectively.

Figure ІІІ. Tissue fat contents, median values and interquartile ranges, according to MRI

(i.e. after correction by 1H MRS) for subjects with normal body weights and for obese

subjects.

Normal body weight (n=28) 1 = Liver; 2 = Pancreas; 3 = Kidney

0

1

2

3

4

5

6

7

8

9

10

1 2 3

Fat content (%

)

Obese (n=8) 1 = Liver; 2 = Pancreas; 3 = Kidney

0

1

2

3

4

5

6

7

8

9

10

1 2 3

Fat content (%

)

Chapter 7.

158

Table І shows that BMI, a measure of overall fat deposits in the human body, correlated

significantly with the amount of subcutaneous fat (r = 0.77, P < 0.01), liver fat content (r =

0.52, P < 0.01) and pancreas fat content (r = 0.35, P < 0.05). The amount of subcutaneous

fat also correlated significantly with liver fat content and pancreas fat content (r = 0.45 and

r = 0.44, respectively; P < 0.01). Liver and pancreas fat contents were significantly

correlated (r = 0.43; P < 0.01). Kidney fat content was comparatively low, showed little

inter-subject variation and correlated with none of the other fat depots.

Table І. Spearman’s correlation coefficients between measures of obesity, and MRI fat

measurements according to Dixon’s two point technique in kidney, liver and pancreas after

correction by 1H MRS.

(n = 36) Correlations

BMI Subc. fat (cm) Liver fat (%) Pancreas fat (%) Kidney fat (%)

BMI ------- 0.765** 0.517** 0.349* 0.264

Subc. fat (cm) 0.765** ------- 0.447** 0.442** 0.256

Liver fat (%) 0.517** 0.447** ------- 0.428** 0.231

Pancreas fat (%) 0.349* 0.442** 0.428** ------- 0.081

Kidney fat (%) 0.264 0.256 0.231 0.081 -------

* P < 0.05, ** P < 0.01

BMI, body mass index; Subc., subcutaneous


159

DISCUSSION

The healthy volunteers, who varied in body weight from lean to obese, had kidney fat

contents that were comparatively low and showed little variation between subjects. In the

kidney fat content correlated with none of the other fat depots, indicating that in otherwise

healthy subjects obesity does not affect kidney fat content. In other words, renal lipid

accumulation, unlike the accumulation of fat in liver and pancreas, is not observed in obese

healthy subjects.

In this study we used a previously validated gradient echo MRI method 17 for determining

the fat contents in liver pancreas and kidney. In our liver measurements we compared the

MR spectroscopy results in the thirty-six volunteers with the MRI liver contents to refine

the correction equation needed to adapt the MRI method for quantitative use. That is, we

determined the coefficient of correlation between the two-point Dixon-based MRI method

and the 1H MRS results in liver and used the result to correct all MRI data for systematic

overestimation at the lower end range, a phenomenon also observed by others 24 and most

probably related to Rician-noise distribution related overestimation of the magnitude

images of weak fat signals. Thus we have used our liver data for correcting all MRI data

based on the assumption that the MRI method works out the same for liver and other organs

(pancreas, kidney). This is the best one can do, considering that 1) 1H MRS examinations of

small fat embedded organs are inaccurate and that 2) it is no option to collect tissue

biopsies from healthy volunteers. Furthermore, as stated in Materials and Methods, control

experiments had made sure that with the radiofrequency coils used the MRI signal

intensities in positions corresponding with those of liver, pancreas and kidney corresponded

to the same fat contents.

Renal lipotoxicity and its role in the pathogenesis of renal disease are not fully understood 4,

5. It has been assumed that renal disease progression is promoted by the accumulation of

lipids in the kidneys, a phenomenon in which triglyceride-rich lipoproteins, free fatty acids

and their metabolites, and albumin-loaded free fatty acids appear to be involved.

Quantitative documentation of the content of fat in human kidneys in situ appears not to

exist, so we cannot compare our MRI-determined kidney fat contents to literature data. In a

bovine growth hormone transgenic mouse line, kidney triglycerides, while lower than those

found in liver, recently showed a similar trend of reduced levels as compared with non

Chapter 7.

160

transgenic littermate controls showing overall fat mass increases 25. In rats, also, steatosis of

the kidney was recently associated with an alteration in renal acidification 12. The fat

accumulated in the renal cortex as shown quantitatively by enzymatic triglyceride

measurement and qualitatively by oil red O staining 12. That fitted the notion that, within

the cortex, the proximal tubule is vulnerable to lipid accumulation due to its role in the

reabsorption of free fatty acids bearing albumin 13, 14. Our result in humans, the finding of

very low fat contents (up to 2%) in the cortex of kidneys of both lean and obese individuals,

is not in line with the findings in the above experimental studies.

Our observation of significant correlation between the liver and pancreas fat contents

according to MRI (r = 0.43; P < 0.01) also is in disagreement with a recent study of a

smaller number of volunteers than included by us (17 versus 36) 10, and with another small

study (n = 15) in which the existence of significant correlation between liver and pancreas

fat content is not mentioned 8. The small scale of both earlier studies in humans probably

explains their failure to demonstrate the correlation between the fat contents in liver and

pancreas. The presence of such a correlation does fit a previous demonstration that obese

nondiabetic subjects have increased fat in the pancreas 8. It is also in line with a notion that

lipomatosis of the pancreas reflect early events in the pathogenesis of diabetes 9, conform

what has been shown in numerous publications dealing with the fattening of the liver.

The significant correlations of BMI with both liver and pancreas fat content in this study

are in agreement with a previous report 8. Why would obese subjects tend to accumulate fat

in liver and pancreas and not in the kidney, despite the observation that obesity and the

metabolic syndrome are involved with initiation of chronic kidney disease 4? It seems that

due to specific requirements such as albumin loading needed to facilitate the uptake of fat

into kidney tissue, obese but otherwise healthy subjects do not accumulate fat in the

kidneys.

In conclusion, this is the first demonstration of the use of a MRI method for determining

kidney fat content. Observed for the first time are significantly correlated liver and pancreas

fat contents in (otherwise) healthy persons varying in body weight from lean to obese.

These observations are related to body weight as measured by BMI and the amount of

subcutaneous fat. The amount of fat in the kidney in obese persons is small and not related

to the amount of body fat or the fat content of liver and pancreas. We have thus shown that

in obesity, the first step in the pathogenesis of renal disease, lipid is not accumulated in the


161

kidney. Therefore the role of lipid accumulation in the pathogenesis of renal disease,

diabetes and metabolic disease in humans should be reconsidered. That is not to say that fat

metabolites rather than the triglyceride levels detected here by MRI, may have profound

effects. In future studies we propose to examine patients suffering from the above illnesses

in order to validate our current hypothesis that the lipid content is always low in the

kidneys.

Chapter 7.

162

Reference List (1) Snijder MB, van Dam RM, Visser M, Seidell JC. What aspects of body fat are

particularly hazardous and how do we measure them? Int J Epidemiol 2006 February;35(1):83-92.

(2) Unger RH. Minireview: weapons of lean body mass destruction: the role of ectopic lipids in the metabolic syndrome. Endocrinology 2003 December;144(12):5159-65.


(4) Wahba IM, Mak RH. Obesity and obesity-initiated metabolic syndrome: mechanistic links to chronic kidney disease. Clin J Am Soc Nephrol 2007 May;2(3):550-62.

(5) Weinberg JM. Lipotoxicity. Kidney Int 2006 November;70(9):1560-6.


(7) Poitout V, Robertson RP. Glucolipotoxicity: fuel excess and beta-cell dysfunction. Endocr Rev 2008 May;29(3):351-66.

(8) Kovanlikaya A, Mittelman SD, Ward A, Geffner ME, Dorey F, Gilsanz V. Obesity and fat quantification in lean tissues using three-point Dixon MR imaging. Pediatr Radiol 2005 June;35(6):601-7.

(9) Raeder H, Haldorsen IS, Ersland L, Gruner R, Taxt T, Sovik O, Molven A, Njolstad PR. Pancreatic lipomatosis is a structural marker in nondiabetic children with mutations in carboxyl-ester lipase. Diabetes 2007 February;56(2):444-9.

(10) Schwenzer NF, Machann J, Martirosian P, Stefan N, Schraml C, Fritsche A, Claussen CD, Schick F. Quantification of pancreatic lipomatosis and liver steatosis by MRI: comparison of in/opposed-phase and spectral-spatial excitation techniques. Invest Radiol 2008 May;43(5):330-7.

(11) Sinha S, Misra A, Rathi M, Kumar V, Pandey RM, Luthra K, Jagannathan NR. Proton magnetic resonance spectroscopy and biochemical investigation of type 2 diabetes mellitus in Asian Indians: observation of high muscle lipids and C-reactive protein levels. Magn Reson Imaging 2009 January;27(1):94-100.

(12) Bobulescu IA, Dubree M, Zhang J, McLeroy P, Moe OW. Effect of renal lipid accumulation on proximal tubule Na+/H+ exchange and ammonium secretion. Am

J Physiol Renal Physiol 2008 June;294(6):F1315-F1322.


163

(13) Gekle M. Renal tubule albumin transport. Annu Rev Physiol 2005;67:573-94.

(14) Riazi S, Khan O, Tiwari S, Hu X, Ecelbarger CA. Rosiglitazone regulates ENaC and Na-K-2Cl cotransporter (NKCC2) abundance in the obese Zucker rat. Am J

Nephrol 2006;26(3):245-57.

(15) Sijens PE. Parametric exploration of the liver by magnetic resonance methods. European Radiology 2009;19(11):2594-607.



(18) Dixon WT. Simple proton spectroscopic imaging. Radiology 1984 October;153(1):189-94.

(19) Hussain HK, Chenevert TL, Londy FJ, Gulani V, Swanson SD, McKenna BJ, Appelman HD, Adusumilli S, Greenson JK, Conjeevaram HS. Hepatic fat fraction: MR imaging for quantitative measurement and display--early experience. Radiology 2005 December;237(3):1048-55.


(21) Longo R, Ricci C, Masutti F, Vidimari R, Croce LS, Bercich L, Tiribelli C, Dalla Palma L. Fatty infiltration of the liver. Quantification by 1H localized magnetic resonance spectroscopy and comparison with computed tomography. Invest Radiol 1993 April;28(4):297-302.


Imaging 1994;12(3):487-95.

(23) Fishbein MH, Mogren C, Gleason T, Stevens WR. Relationship of hepatic steatosis to adipose tissue distribution in pediatric nonalcoholic fatty liver disease. J Pediatr Gastroenterol Nutr 2006 January;42(1):83-8.

(24) Kim H, Taksali SE, Dufour S, Befroy D, Goodman TR, Petersen KF, Shulman GI, Caprio S, Constable RT. Comparative MR study of hepatic fat quantification using

Chapter 7.

164

single-voxel proton spectroscopy, two-point dixon and three-point IDEAL. Magn

Reson Med 2008 March;59(3):521-7.

(25) Palmer AJ, Chung MY, List EO, Walker J, Okada S, Kopchick JJ, Berryman DE. Age-related changes in body composition of bovine growth hormone transgenic mice. Endocrinology 2009 March;150(3):1353-60.

Chapter 8.1: General discussion - part 1

Evidence on screening for fatty liver disease: Future perspectives

Submitted/ under review _ Review Paper

Mireille A. Edens

Ronald P. Stolk

Department of Epidemiology



Chapter 8.1: General discussion - part 1.

166

ABSTRACT

Background and Aim Fatty liver disease (FLD) is currently (one of) the most prevalent

hepatic condition(s) worldwide and likely increasing. Since FLD initially gives no clinical

symptoms, but does increase morbidity risk, screening could be useful. The aim of this

paper is to discuss if screening for FLD would be effective.

Methods The world health organisation screening criteria, organised into the categories:

‘condition’, ‘diagnosis’, ‘treatment’, and ‘cost’ & ‘screening program’, are discussed.

Results FLD is associated with increased risk for severe liver pathology (cirrhosis and

hepatocellular carcinoma) and cardiovascular pathology (components of the metabolic

syndrome). Compared to reference populations, survival of both patients with non-alcoholic

FLD (NAFLD; including both steatosis and steatohepatitis) and patients with non-alcoholic

steatohepatitis (NASH) is reduced. Moreover, both liver-related mortality and

cardiovascular-related mortality are increased. An estimated 8.3% of patients with NAS

will develop fibrosis, and an estimated 37.6% of patients with NASH will progress in

fibrosis stage. Ultrasonography is an acceptable method for assessing FLD. Most FLD

cases have a behaviour-related etiology, which provides opportunity for treatment. Studies

on cost-effectiveness of screening for FLD lack, but prospects are promising given the

increased costs (both financial and quality of life) of patients with FLD.

Conclusion Based on the evidence presented in this paper, we conclude that screening for

FLD is advisable. However, cost effectiveness studies on screening for fatty liver disease

are yet to be performed.

Screening for fatty liver disease

167

PREFACE

In CHAPTERS 2 and 3 we found that fatty liver disease is highly prevalent, and associated

with an increased cardiovascular disease risk. Therefore, in this chapter, we investigated

whether screening for fatty liver disease would be beneficial.

INTRODUCTION

Excess fat should be stored in adipocytes (subcutaneous fat), where it functions as an

‘adipose organ’ 1. In the case of dietary overflow, lipids can be stored in the peritoneal

cavity (visceral fat), retro-peritoneal (peri-renal fat), or ectopically, i.e. inside myocytes and

organs (e.g. the liver) as well 2. Hepatic free fatty acids (HFFAs), i.e. not oxidated, secreted

as very low density lipoprotein, or excreted as phospholipids into bile, are stored as neutral

triglycerides (TGs) within lipid vesicles [CHAPTER 4] 3. Continuous accumulation of intra-

hepatocellular TGs will result in fatty liver disease (FLD).

FLD, which includes both alcoholic FLD (AFLD) and non-alcoholic FLD (NAFLD; by

≤20g ethanol a day 4), refers to a broad spectrum. Repeated biopsies have revealed that non-

alcoholic steatosis (NAS; fat, with or without non-specific inflammation) can progress to

non-alcoholic steatohepatitis (NASH; fat + inflammation, without of with fibrosis) 5.

Fibrosis progression in the case of NASH, delineating loss of liver fat and inflammation 6,

results in cirrhosis (fibrosis stage 4 7), also referred to as ‘burned out NASH’ 8. It has been

known for many years that cryptogenic cirrhosis can be associated with hepatocellular

carcinoma (HCC) 6, but it only recently discovered that active NASH without cirrhosis can

be accompanied by HCC as well 9, 10. In the case of a NASH background, HCC can be

multi-focal 10. Besides a hepatic risk, NAFLD may also increase cardiovascular disease

(CVD) risk, as a (non-alcoholic) fatty liver overproduces several CVD risk markers

[CHAPTER 3] 11. NAFLD is usually asymptomatic. Non-specific complaints of weakness

and fatigue 12, 13, and a vague right upper abdominal discomfort or epigastric pain are

reported by some patients 6, 13-15. Clinical signs of chronic liver disease are absent, apart

from hepatomegaly in some patients 12, 14, 15. NASH is generally discovered incidentally

during investigation of other (unrelated) medical conditions 12, 15 or after health screening 15.


168

These risks, may warrant aggressive screening on steatosis 10, 16, in order to prevent

progression to NASH and symptomatic disease.

FLD and NAFLD are highly prevalent worldwide. The estimated current prevalence of

NAFLD by imaging in the Asia-Pacific region was recently reviewed by Amarapurkar ea.

(2007) and ranges from 16% to 42% 17. Two studies in Japan have revealed an incidence of

FLD as well, i.e. from 12.6% in 1989 to 30.3% in 2000 18, and from 33.3% in 2000 to

38.5% in 2005 in men, while the prevalence remained similar in women 19. The estimated

prevalence of FLD and NAFLD by imaging in the Western population varies from 27.4%

to 31% 20-22 and 27% 20, respectively.

However, despite currently being the most common hepatic condition worldwide and likely

increasing, screening on FLD has not been recommended. The World Health Organisation

(WHO) has published screening criteria, originally introduced by Wilson and Jungner 23

and expanded later 24. The aim of this paper is to discuss if screening for FLD would be

effective, by discussing the WHO screening criteria, organised into the following

categories: ‘condition’, ‘diagnosis’, ‘treatment’, ‘cost’ & ‘screening program’ (table І).

Table І. O

rgan

isat

ion o

f sc

reen

ing c

rite

ria

CATEGORY 25

NR

WIL

SON &

JUNGNER CRIT

ERIA

23, 24

EXPANSIO

N 24

Conditio

n

1a

The

conditio

n so

ught

should

be

an im

port

ant

hea

lth pro

ble

m fo

r th

e in

div

idual

and c

om

munity

1b

The

nat

ura

l his

tory

of

the

conditio

n,

incl

udin

g d

evel

opm

ent

from

lat

ent

to

dec

lare

d d

isea

se, sh

ould

be

adeq

uat

ely u

nder

stood

1c

Ther

e sh

ould

be

a re

cogniz

able

lat

ent or

earl

y s

ym

pto

mat

ic s

tage

1d

T

her

e sh

ould

be

a def

ined

tar

get

popula

tion.

Dia

gnosi

s

2a

The

test

should

be

acce

pta

ble

to the

popula

tion

2b

Ther

e sh

ould

be

a su

itab

le tes

t or

exam

inat

ion

2c

Fac

ilitie

s fo

r dia

gnosi

s (a

nd tre

atm

ent)

should

be

avai

lable

Tre

atm

ent

3a

Ther

e sh

ould

be

an a

ccep

ted tre

atm

ent fo

r pat

ients

with r

ecognis

ed d

isea

se

3b

Ther

e sh

ould

be

an a

gre

ed p

olicy

on w

hom

to tre

at a

s pat

ients

3c

Fac

ilitie

s fo

r (d

iagnosi

s an

d)

trea

tmen

t sh

ould

be

avai

lable

Cost

4

The

cost

of

case

-fin

din

g (

incl

udin

g d

iagnosi

s an

d t

reat

men

t of

pat

ients

dia

gnose

d)

should

be

econom

ical

ly

bal

ance

d

in

rela

tion

to

poss

ible

ex

pen

diture

on m

edic

al c

are

as a

whole

Scr

eenin

g

pro

gra

m

5a

Cas

e-fi

ndin

g s

hould

be

a co

ntinuin

g p

roce

ss a

nd n

ot

a ‘o

nce

and f

or

all’

pro

ject

5b

Should

res

pond to a

rec

ognis

ed n

eed

5c

T

he

obje

ctiv

es o

f sc

reen

ing s

hould

be

def

ined

at th

e outs

et

5d

T

he

over

all ben

efits

of

scre

enin

g s

hould

outw

eigh the

har

m

5e

T

her

e sh

ould

be

scie

ntifi

c ev

iden

ce

of

scre

enin

g

pro

gra

m

effe

ctiv

enes

s

5f

T

he

pro

gra

m

should

in

tegra

te

educa

tion,

test

ing,

clin

ical

se

rvic

es, an

d p

rogra

m m

anag

emen

t

5g

T

her

e sh

ould

be

qual

ity

ensu

rance

, w

ith

mec

han

ism

s to

m

inim

ize

pote

ntial

ris

ks

of

scre

enin

g

5h

T

he

pro

gra

m sh

ould

en

sure

in

form

ed ch

oic

e, co

nfi

den

tial

ly,

and r

espec

t fo

r au

tonom

y

5i

T

he

pro

gra

m s

hould

pro

mote

equity a

nd a

cces

s to

scr

eenin

g f

or

the

entire

tar

get

popula

tion

5j

Pro

gra

m e

val

uat

ion s

hould

be

pla

nned

fro

m the

outs

et


169


170

THE CONDITION

FATTY LIVER DISEASE-ASSOCIATED HEALTH PROBLEMS?

Besides occasional non-specific vague complaints 6, 12-15 and hepatomegaly 12, 14, 15, FLD is

associated with various health problems. Physical problems include liver failure caused by

cirrhosis 6, 26 and/or HCC 6, 9, 10, and increased CVD risk [CHAPTER 3] 4, 11. Additionally,

decreased quality of life has been reported 27.

Overall survival compared to reference populations

A meta-analysis on survival of patients with NAFLD has recently been published 16.

Because of its relevance for screening (detection of an early latent stage), this section

focuses specifically on histological-determined FLD subtypes and age (table ІІ). Compared

to reference populations, overall survival of patients with NAS is non-significantly reduced

28-30, whereas survival of both patients with NAFLD 13, 29 and NASH 29, 30 are significantly

reduced. Additionally, survival of patients with AFLD is significantly reduced as well 30.

Causes of death

The primary cause of death in patients with NAFLD is CVD-related death 13, 29-31. CVD-

related death was 7.5% in a reference population, slightly increased to 8.6% in patients with

NAS (p=ns compared to the reference population) and significantly increased to 15.5% in

patients with NASH (p<.05 compared to the reference population) 29. Compared to

reference populations, patients with NASH 29, 32 have a increased risk for liver-related

death. In severe NASH, i.e. half of the population with cirrhosis, infection is the primary

cause of death 33.

Table ІІ.

Longit

udin

al s

tudie

s on s

urv

ival

of

fatt

y liv

er d

isea

se s

ubty

pes

ALIV

E AT BASELIN

E

FOLLOW

-UP PERIO

D

DECEASED

FLD subtypes, i.e. n and age

mean (±sd) & m

edian (range)

N (N%)

Cause of death

NA

S

NA

SH

A

FL

D

E

xtr

a-hep

atic

H

epat

ic

U

Car

dio

vas

cula

r dis

ease

M

O

LB

D

HC

C

Car

dia

c O

V

IHD

O

C

Mat

teoni ea

. (1

999)

32

Unit

ed S

tate

s of

Am

eric

a

N=

59 (

2C

) A

ge

T1: 53 (

±15)

Age

T2: 46 (

±12)

8.9

(±5.5

) &

NR (

0.4

–17.8

)

48 (

36%

) 9

2

2

10

10

11

1

3

N

=73 (

18C

) A

ge

T3: 49 (

±15)

Age

T4: 56 (

±11)

7.8

(±5.3

) &

NR (

0.1

–18.2

)

Dam

-Lar

sen e

a.

(2004)

34

Den

mar

k

N=

109

Age:

39 (

19–80)

16.7

(0.2

–21.9

) 27 (

25%

)

1

N=

106

§

Age:

50 (

26–72)*

**

9.2

(0.6

–23.1

) 79 (

75%

)

22

Dam

-Lar

sen e

a.

(2005)

28†

Den

mar

k

N=

170

Age:

39 (

19–84)

19 (

0.2

–27.1

) N

R

1

N=

247

§

Age:

50 (

26–76)*

**

12.8

(0.1

–27.1

) N

R

54

Ekst

edt ea

. (2

006)

29

Sw

eden

N=

58

Age:

47 (

±12)

13.7

(±1.3

) &

NR (

10.3

–16.3

)

7 (

12%

) 5

1

1

N

=71 (

4C

) A

ge:

55 (

±12)*

**

19 (

27%

) 11

4

2

1

1

San

yal

ea.

(2006)

33

Unit

ed S

tate

s of

Am

eric

a

N=

152

‡ (

74C

) A

ge

CT

P-A

: 55 (

±N

R)

Age

CT

P-B

: 52 (

±N

R)

Age

CT

P-C

: 60 (

±N

R)

10

29 (

19%

) 2

6

5

14

2

Raf

iq e

a. (

2009)

31

Unit

ed S

tate

s of

Am

eric

a

N=

101

Age:

49 (

±15)

18.5

(≥

5–28.5

) 78 (

49%

) 22

14

2

2

7

N

=57

Age:

52 (

±13)

10

Soder

ber

g e

a.

(2010)

30

Sw

eden

N=

67 (

4C

) A

ge:

45 (

±12)

21 (

±7.7

) &

24 (

0.5

–28)

23 (

34%

) 7

5

4

4

2

1

N

=51 (

5C

) A

ge:

49 (

±N

R)

24 (

47%

) 7

8

6

3

N=

25 (

2C

) ¶

Age:

NR

20 (

80%

) 8

1

4

5

2

***,

p<

.001;

†,

enla

rgem

ent

of

thei

r st

udy f

rom

2004;

‡,

n=

74 C

TP

_A

, n=

43 C

TP

_B

, n=

35 C

TP

_C

); §

, al

coholic

stea

tosi

s; ¶

, al

coholic

fatty l

iver

dis

ease

; C

, ci

rrhotics

; C

TP

, C

hild-T

urc

otte-

Pugh s

core

; H

CC

, hep

atoce

llula

r ca

rcin

om

a; L

BD

, liver

and b

ilia

ry d

isea

se (

excl

udin

g H

CC

); M

, ex

tra-

hep

atic

mal

ignan

cy;

NA

FL

D,

non-a

lcoholic

fatty l

iver

dis

ease

; N

AS,

non-a

lcoholic

stea

tosi

s; N

ASH

, non-a

lcoholic

stea

tohep

atit

is;

NR,

not

report

ed;

OC

, oth

er c

ardia

c ca

use

; O

, oth

er e

xtr

a-hep

atic

cau

se; O

V, oth

er v

ascu

lar

even

ts; U

, unknow

n c

ause

.


171

Table ІІІ. L

ongitudin

al s

tudie

s on f

ibro

sis

pro

gre

ssio

n o

f non-a

lcoholic

fatt

y liv

er d

isea

se, by s

eria

l his

tolo

gic

al a

nal

ysi

s

Country of

origin

First biopsy

Follow up period

Last biopsy

NAS

NAFLD

NASH

mean (±sd) & m

edian (range)

Fibrosis

regression

Fibrosis

stable

Fibrosis progression

↑F

→C

Lee

(1989)

12

US

A

N=

8

2.8

(±1.6

) &

2.4

(1.2

–6.1

)

N=

3

N=

3

N=

2

Pow

ell ea

. (1

990)

6

AU

S

N=

13

4.5

(±1.8

) &

4 (

1–8)

N=

1

N=

8

N=

3

N=

1 &

→H

CC

Bac

on e

a. (

1994)

14

US

A

N=

2

5.5

(±2.1

) &

5.5

(4–7)

N

=1

N

=1

Tel

i ea

. (1

995)

15

UK

N

=12

NR (

7.6

–16)

NR

NR

N=

1

Evan

s ea

. (2

002)

35

UK

N

=7

8.2

(±2.6

) &

7 (

5.5

–11.9

)

N=

3

N=

4

Har

riso

n e

a. (

2003)

5

US

A

N

=22 (

3 N

AS

, 19 N

AS

H)

5.7

(N

R)

& N

R (

1.4

–15.7

) N

=4

N=

11

N=

6

N=

1

Fas

sio e

a. (

2004)

36

AR

G

N=

22

5.3

(±2.7

) &

4.3

(3–14.3

) N

=4

N=

11

N=

7

Hui ea

. (2

005)

37

CH

INA

N=

17 (

3 N

AS

, 14 N

AS

H)

5.8

(±1.4

) &

6.1

(3.8

–8)

N

=8

N=

8

N=

1

Adam

s ea

. (2

005)

38†

US

A

N

=103 (

7 N

AS

, 96 N

AS

H)

3.2

(±3)

& N

R (

0.7

–21.3

) N

=30

N=

35

N=

29

N=

9

†, th

is s

tudy is

a cl

inic

al tri

al b

ut th

e ‘i

nte

rven

tion’

had

no e

ffec

t on h

isto

logy; ↑, in

crea

se to h

igher

sta

ge;

→, pro

gre

ssio

n to.

AR

G, A

rgen

tina;

C, ci

rrhosi

s; F

, fi

bro

sis;

HC

C, hep

atoce

llula

r ca

rcin

om

a; N

AF

LD

, non-a

lcoholic

fatt

y liv

er d

isea

se; N

AS, non-a

lcoholi

c

stea

tosi

s; N

AS

H, non-a

lcoholic

stea

tohep

atitis

; N

R, not re

port

ed; U

K, U

nit

ed K

ingdom

; U

SA

, U

nit

ed S

tate

s of

Am

eric

a; A

US, A

ust

ralia


172


173

NATURAL HISTORY OF FATTY LIVER DISEASE?

Several studies on fibrosis progression in the case of (non-alcoholic) fatty liver disease have

been performed 5, 6, 12, 14, 15, 35-38 and summarised in table ІІІ. NAS could be considered a

latent stage of FLD as it is relatively non-progressive 15, 28, 29. An estimated one-twelfth

(8.3%) of patients with NAS will develop fibrosis 15. Fibrosis progression in patients with

NASH has recently been systematically reviewed by Argo ea. (2009) 39. Of all NASH

patients (n=221 in total) having a second biopsy after 5.3 (±4.2) year, 37.6% progressed to

a higher fibrosis stage, 41.6% had no change, and 20.8% regressed to a lower fibrosis stage

39.

Fibrosis progression

Mean fibrosis progression rate was 0.03 (±0.53) stages/year 39. However, when cirrhotics

are excluded (as they cannot progress further), mean fibrosis progression will be higher 13.

Importantly, fibrosis progression delineates regression of both fat 6, 36, 38 and inflammation 5

,

6, 38. Figure І shows a theoretical model on the FLD spectrum.

Associates of fibrosis progression

Multivariate analysis, with advanced fibrosis (stage 3 and 4) as dependent variables,

revealed that both age and inflammation on initial biopsy are independent predictors of

progression to advanced fibrosis 39. Although this implies important prognostic value these

results are not surprising, as injury yields an inflammatory response 40 resulting in repair of

the injury by means of fibrosis [CHAPTER 4] 3. If inflammation on initial biopsy (and

cirrhotics) is excluded, low initial fibrosis stage, BMI, and diabetes are predictors of

fibrosis progression 38.

DEFINED TARGET POPULATION?

The aim of screening for FLD would be to detect an increased CVD risk and/or hepatic

risk, thereby increasing opportunity for intervention, in order to prevent progression of

CVD risk and hepatic risk, i.e. secondary prevention. The most efficient/effective approach

would be to screen people with an already increased hepatic risk and/or CVD risk, e.g.

diabetes. Patients with type 2 diabetes and NAFLD (by ultrasonography) have a

significantly higher incidence of CVD events compared to patients with type 2 diabetes

without NAFLD 41. Diabetes predisposes to hormone sensitive lipase-induced breakdown


174

of TG vesicles, releasing more toxic HFFAs [CHAPTER 4] 3, and is a predictor of fibrosis

progression 38.

Screening high risk groups does raise the question of population screening, i.e. primary

prevention 42. Epidemiological studies have revealed that NAFLD is most common at

middle age and decreases at higher ages 43-45.

Fat

Inflammation

Fibrosis

Infiltration:

INFILTRATION

PROGRESSION OF THE FATTY LIVER DISEASE SPECTRUM (SPEED OF PROGRESSION DEPENDS ON THE AMOUNT OF DETERMINANTS, PER INDIVIDUAL)

Normal NASS NASni NASH-f NASH+f/c Cirrhosis

Increased HCC risk

Figure 1. Schematic model on the natural history of fatty liver disease

The scale of natural progression was estimated from progression throughout the various FLD stages is based on long-term follow up studies (appendix

ІІb). The starting stage may vary between individuals.HCC, hepatocellular carcinoma; NASH-f, non-alcoholic steatohepatitis without fibrosis; NASH+f,

non-alcoholic steatohepatitis with fibrosis, NASni, non-alcoholic steatosis with non-specific inflammation; NASS, non-alcoholic simple steatosis.


175

DIAGNOSIS

METHOD ACCEPTABLE TO THE POPULATION?

A diagnosis modality for screening should firstly be non-invasive, which excludes liver

biopsy and computed tomography (harmful radiation). Secondly, a diagnosis method

should be low-cost, which excludes magnetic resonance spectroscopy and magnetic

resonance imaging. Additionally, a diagnosis method should be liver-specific, which

excludes aminotransferases as aminotransferases are produced and present in numerous

tissues 46, 47. An elevation of aminotransferases can therefore be attributed to other diseases

as well 48, 49. The remaining diagnosis method is ultrasonography.

IS ULTRASONOGRAPHY SUITABLE FOR FATTY LIVER DISEASE?

Within the FLD spectrum, fat can both occur separately and coexist with inflammation

and/or fibrosis (figure І). Therefore, in order to estimate the place within the FLD spectrum,

fat, inflammation, and fibrosis are to be determined ideally.

Fat A selection of ultrasound validation studies is shown in table ΙV. Qualitative

ultrasonography is a valid and reliable method for diagnosing FLD, i.e. an abundance of

liver fat 50-52. Validity is decreased in (morbidly) obese people 53, 54. A quantitative

ultrasonography method has recently been developed as well [CHAPTER 5] 55.

Inflammation Ultrasonography is (currently) unable to determine hepatic inflammation,

which means that it is unable to distinguish steatosis and steatohepatitis. Perhaps it might

be of value to estimate splenomegaly 56, but validation studies using ultrasonography-

determined splenomegaly have not been published.

Fibrosis Ultrasonography can estimate hepatic fibrosis. Two meta-analysis on transient

ultrasonographic elastography (FibroScan), i.e. in all liver diseases 57 and in FLD

specifically 16 have recently been published and show acceptable results (table ΙV). The

success rate (applicability) of ultrasonographic elastography is not perfect, which is mostly

attributed to obesity 16, 57. This suggests that for obese patients an extra large transducer

may be required 58. Additionally, it should theoretically be able to estimate fibrosis using

ultrasonographic texture analysis [CHAPTER 5] 55, but more validation work is needed.

Table ΙV. V

alid

ity o

f ultra

sonogra

phy f

or

asse

ssin

g f

atty

liv

er d

isea

se, se

lect

ion o

f th

e lite

ratu

re

Country of

origin

Gold

Standard

Population

Hepatic fat

Hepatic fibrosis

SE (%)

SP (%)

AUC

r SE (%)

SP (%)

AUC

Qualitative ultrasonography

Sav

erym

utt

u e

a. (

1986)

50,

and J

ose

ph e

a. (

1991)

51

Unit

ed K

ingdom

H

(ql.)

(S)L

D

89–94

84–93

- -

NA

N

A

NA

Ham

aguch

i ea

. (2

007)

52

Japan

H

(ql.)

Excl

. A

s an

d v

H

91.2

–92.6

100

0.9

8

- N

A

NA

NA

Quantitative ultrasonography

Eden

s ea

. (2

009)

[C

HAPTER 5

] 55

The

Net

her

lands

1H

MR

S (

qnt.)

Gen

. (n

o k

now

n

LD

excl

. F

LD

) 66.7

100

- .7

89*

- -

-

Transien

t ultrasonographic elastography †

‡

Met

a-an

alysi

s:

Fri

edri

ch-R

ust

ea.

(2008)

57

All

countr

ies

(n=

50 s

tudie

s)

H (

ql.)

F≥2

NA

N

A

NA

N

A

56 –

100

18 –

100

0.6

8 –

1.0

0

F≥3

NA

N

A

NA

N

A

58 –

95

78 –

97

0.7

2 –

0.9

7

F=

4 (

i.e.

C)

NA

N

A

NA

N

A

73 –

100

78 –

98

0.8

1 –

0.9

9

Wes

tern

reg

ion

(n=

46 s

tudie

s)

F≥2

NA

N

A

NA

N

A

56 –

100

18 –

100

0.6

8 –

1.0

0

F≥3

NA

N

A

NA

N

A

58 –

95

85 –

97

0.7

2 –

0.9

7

F=

4 (

i.e.

C)

NA

N

A

NA

N

A

73 –

100

81 –

98

0.8

7 –

0.9

9

Asi

a-P

acif

ic

regio

n

(n=

4 s

tudie

s)

F≥2

NA

N

A

NA

N

A

79 –

90

78 –

88

0.7

7 –

0.8

1

F≥3

NA

N

A

NA

N

A

86 –

95

78 –

92

0.7

9 –

0.9

3

F=

4 (

i.e.

C)

NA

N

A

NA

N

A

80 –

86

78

0.8

1 –

0.8

8

Met

a-an

alysi

s:

Muss

o e

a. (

2010)

16

All

countr

ies

(n=

6 s

tudie

s)

H (

ql.)

F≥2

NA

N

A

NA

N

A

79 –

100

74 –

93

0.8

4 –

0.9

9

F≥3

NA

N

A

NA

N

A

75 –

100

81 –

100

0.9

0 –

1.0

0

Wes

tern

reg

ion

(n=

2 s

tudie

s)

F≥2

NA

N

A

NA

N

A

81 –

100

78 –

92

0.8

6 –

0.9

9

F≥3

NA

N

A

NA

N

A

100

100

1.0

0

Asi

a-P

acif

ic

regio

n

(n=

4 s

tudie

s)

F≥2

NA

N

A

NA

N

A

79 –

100

74 –

93

0.8

4 –

0.9

9

F≥3

NA

N

A

NA

N

A

75 –

100

81 –

93

0.9

0 –

0.9

9

*, p<

.001;

†, cu

t-off

val

ues

dif

fer

per

stu

dy;

‡, su

cces

s-ra

tes

dif

fer

per

stu

dy.

As,

alc

oholics

; A

UC

, ar

ea u

nder

the

curv

e ra

nge;

C,

cirr

hosi

s; D

VS,

theo

reti

cally p

oss

ible

, but

in n

eed o

f fu

rther

dev

elopm

ent

and

val

idat

ion s

tudie

s; E

xcl

., e

xcl

udin

g; F, fi

bro

sis

stag

e as

det

erm

ined

by M

ET

AV

IR a

nd o

ther

sco

ring s

yst

ems;

FL

D, fa

tty liv

er d

isea

se; G

en.,

gen

eral

popula

tion; H

, his

tolo

gy;

1H

MR

S; M

agnet

ic R

esonan

ce S

pec

trosc

opy; L

D, li

ver

dis

ease

; N

A, not ap

pli

cable

; ql., qual

itat

ive;

qnt.,

quan

tita

tive;

r, co

rrel

atio

n c

oef

fici

ent

by P

ears

on;

SE

, se

nsi

tivity r

ange;

(S)L

D, (s

usp

ecte

d)

liver

dis

ease

; SP, sp

ecif

icity r

ange;

vH

, vir

al

hep

atit

is.


176


177

TREATMENT

ACCEPTED TREATMENT?

When discussing treatment, it is relevant to distinguish cases with a behavioural etiology

and cases with a non-behavioural etiology. Patients with a behaviour-related etiology can

potentially be treated (‘is controllable’), in contrast to patients with a more complicated

etiology like genetic, drug-related, and/or environment-related (‘partially controllable’). In

order to determine a suitable treatment for the individual patient, etiologic factors per

individual patient should be determined. Risk factors for FLD can be arranged in the

following three categories: 1] risk factors for hepatic lipid content, 2] risk factors for

inhibited hepatic metabolism, and 3] risk factors for hepatic inflammation [CHAPTER 4] 3.

Behavioural etiology, i.e. controllable

Given the link with overnutrition & low physical activity, and alcohol (AFLD) [CHAPTER

4] 3, it is to be expected that most of the FLD cases have a behaviour-related etiology.

These behaviour-related cases may be naturally resolved by breaking off etiologic

behaviour. Weight loss interventions resulting in weight loss of ≥2.6 body mass index-

points, were effective in improving NAFLD [CHAPTER 3] 4, 11. This can be achieved

through diet 59, exercise alone 60, diet & exercise 59, 60, or bariatric surgery 59. Weight loss

should not go too fast, given the risk of accentuation or development of focal fatty change

in liver segment ІV 61. A part of the patients may be able to achieve and maintain weight

loss themselves 62. If needed, general practitioners, nutritionists 63, and cognitive behaviour

therapists 64, 65, may be able to assist people with their weight loss.

Non-behavioural etiology, i.e. partially controllable

In the case of a non-behaviour-related etiology, etiologic factors should be abolished, e.g.

use of hepatotoxic drugs if possible 66, 67 and protection for or change of the environment 68.

Otherwise, the non-natural cases may be in need of pharmacological treatment to either

resolve FLD itself [CHAPTER 3] 11, 59, 69 or to resolve FLD associated pathology [CHAPTER

3] 4, 11, 70. A systematic clinical review has recently been published, which describes the

actions and effects of potential pharmacological treatment modalities for FLD 59. It should

be noted that effects of pharmacological approaches are often confounded by weight loss

[CHAPTER 3] 11. If pharmacological treatment coincides with weight loss, it is difficult to

determine the independent effect of the pharmacological intervention. A meta-analysis on


178

randomised clinical trials for FLD was recently published as well, reporting that

randomised trials of adequate size and duration using histological endpoints are needed, to

assess long-term safety and efficacy 69.

AGREED POLICY WHO TO TREAT?

Policy decision guidelines are based on increased absolute risks 42. This likely means that

people with known hepatic and/or CVD risk factors should be offered screening and

treatment for FLD.


179

COST AND SCREENING PROGRAM

COST OF FLD?

It has been shown that patients with FLD (by ultrasonography and alanine

aminotransferase) have more consults from specialists and use more medication compared

to subjects without FLD 70, suggesting that screening and treatment might potentially cut

down health care expenditure 70. The increased medication use in FLD was largely

attributable to diabetes and lipid lowering medication 70.

COST OF A SCREENING PROGRAM FOR FLD?

To the best of our knowledge, there are no published studies on the cost-effectiveness of

ultrasonographic screening for FLD. It is important to realise, that with ultrasonographic

screening for FLD, screening for other (un)expected hepatic lesions, e.g. HCC, will

automatically be included as well. Screening for HCC has been reported to be effective, at

least in a high risk group 71. Cost-effectiveness was not studied, but the authors did report

that the estimated costs of the screening program were similar to their national breast

cancer screening program 71.

AFTER FLD DIAGNOSIS?

After having diagnosed FLD it may be wise to determine etiology and/or comorbidity of

FLD. The design of a potential screening program, including potential management and

treatment of discovered cases, is shown in appendix І.

Etiology

As described in the treatment section, it may be relevant to determine etiology for each

individual patient in order to determine the most appropriate action(s). A behavioural

etiology could simply be determined by questionnaires on behaviour, anthropometry, and

ultrasonography. A non-behavioural etiology could be identified by questionnaires on

behaviour, (familiar) diseases, and environment.

Comorbidity

Given the strong association of FLD with components of the MetS [CHAPTER 3] 4, 11,

components of the MetS should be assessed as well.


180

DISCUSSION

Taken together, the screening criteria regarding ‘condition’, ‘treatment’, ‘diagnosis’, and

‘screening program’ (table І) seem to be reasonable met or possible to meet. Currently cost-

effectiveness studies on screening for FLD lack, but prospects are promising given the

increased cost (both financial 70 and quality of life 27) of patients with FLD. Before

screening prime time, general practioners and specialists should bear FLD in mind and

stimulate weight loss.

STUDIES USED IN THIS PAPER

Histological analysis

Most studies used in this paper used histology as diagnosis modality. It should be noted that

the first histological scoring protocol for internationally uniform scoring of FLD, was

published in 1999 72 and updated in 2005 7. Despite currently being the only modality that

can stage FLD (i.e. distinguish steatosis and steatohepatitis), it has some diagnostic

disadvantages as well. Liver biopsy is subject to sampling bias 73, 74 and histological

analysis is subject to scoring variability (particularly of inflammation) 7, 72, 75.

Risks in the general population

Many studies used in this paper were studies on ‘natural’ NAFLD, which means that,

besides obesity and diabetes, many known etiologic factors were excluded. Obviously,

all/many etiologic factors for FLD [CHAPTER 4] 3 are present in the general population,

which may suggest a (much) faster fibrosis progression rate in the general population.

CONCLUSION

Based on the evidence presented in this paper, we conclude that screening for FLD is

advisable. However, cost effectiveness studies on screening for FLD are yet to be

performed.


181

Reference List (1) Avram MM, Avram AS, James WD. Subcutaneous fat in normal and diseased

states: 1. Introduction. J Am Acad Dermatol 2005 October;53(4):663-70.

(2) Snijder MB, van Dam RM, Visser M, Seidell JC. What aspects of body fat are particularly hazardous and how do we measure them? Int J Epidemiol 2006 February;35(1):83-92.

(3) Edens MA, Groen AK, Stolk RP. Pathogenesis of fatty liver disease: A theory on 'lipid content', 'inhibited metabolism', and 'inflammation'. Submitted 2010.


(5) Harrison SA, Torgerson S, Hayashi PH. The natural history of nonalcoholic fatty liver disease: a clinical histopathological study. Am J Gastroenterol 2003 September;98(9):2042-7.



(8) Ong J, Younossi ZM, Reddy V, Price LL, Gramlich T, Mayes J, Boparai N. Cryptogenic cirrhosis and posttransplantation nonalcoholic fatty liver disease. Liver Transpl 2001 September;7(9):797-801.

(9) Cuadrado A, Orive A, Garcia-Suarez C, Dominguez A, Fernandez-Escalante JC, Crespo J, Pons-Romero F. Non-alcoholic steatohepatitis (NASH) and hepatocellular carcinoma. Obes Surg 2005 March;15(3):442-6.

(10) Chagas AL, Kikuchi LO, Oliveira CP, Vezozzo DC, Mello ES, Oliveira AC, Cella LC, Herman P, Bachella T, Caldwell SH, Alves VA, Carrilho FJ. Does hepatocellular carcinoma in non-alcoholic steatohepatitis exist in cirrhotic and non-cirrhotic patients? Braz J Med Biol Res 2009 October;42(10):958-62.


(12) Lee RG. Nonalcoholic steatohepatitis: a study of 49 patients. Hum Pathol 1989 June;20(6):594-8.


182

(13) Adams LA, Lymp JF, St Sauver J, Sanderson SO, Lindor KD, Feldstein A, Angulo P. The natural history of nonalcoholic fatty liver disease: a population-based cohort study. Gastroenterology 2005 July;129(1):113-21.

(14) Bacon BR, Farahvash MJ, Janney CG, Neuschwander-Tetri BA. Nonalcoholic steatohepatitis: an expanded clinical entity. Gastroenterology 1994 October;107(4):1103-9.


(16) Musso G, Gambino R, Cassader M, Pagano G. Meta-analysis: Natural history of non-alcoholic fatty liver disease (NAFLD) and diagnostic accuracy of non-invasive tests for liver disease severity. Ann Med 2010 November 2.

(17) Amarapurkar DN, Hashimoto E, Lesmana LA, Sollano JD, Chen PJ, Goh KL. How common is non-alcoholic fatty liver disease in the Asia-Pacific region and are there local differences? J Gastroenterol Hepatol 2007 June;22(6):788-93.






(22) Haenle MM, Brockmann SO, Kron M, Bertling U, Mason RA, Steinbach G, Boehm BO, Koenig W, Kern P, Piechotowski I, Kratzer W. Overweight, physical activity, tobacco and alcohol consumption in a cross-sectional random sample of German adults. BMC Public Health 2006;6:233.

(23) Wilson JM, Jungner YG. [Principles and practice of mass screening for disease]. Bol Oficina Sanit Panam 1968 October;65(4):281-393.


183

(24) Andermann A, Blancquaert I, Beauchamp S, Dery V. Revisiting Wilson and Jungner in the genomic age: a review of screening criteria over the past 40 years. Bull World Health Organ 2008 April;86(4):317-9.

(25) Grootendorst DC, Jager KJ, Zoccali C, Dekker FW. Screening: why, when, and how. Kidney Int 2009 October;76(7):694-9.


(27) David K, Kowdley KV, Unalp A, Kanwal F, Brunt EM, Schwimmer JB. Quality of life in adults with nonalcoholic fatty liver disease: baseline data from the nonalcoholic steatohepatitis clinical research network. Hepatology 2009 June;49(6):1904-12.




(30) Soderberg C, Stal P, Askling J, Glaumann H, Lindberg G, Marmur J, Hultcrantz R. Decreased survival of subjects with elevated liver function tests during a 28-year follow-up. Hepatology 2010 February;51(2):595-602.

(31) Rafiq N, Bai C, Fang Y, Srishord M, McCullough A, Gramlich T, Younossi ZM. Long-term follow-up of patients with nonalcoholic fatty liver. Clin Gastroenterol

Hepatol 2009 February;7(2):234-8.


(33) Sanyal AJ, Banas C, Sargeant C, Luketic VA, Sterling RK, Stravitz RT, Shiffman ML, Heuman D, Coterrell A, Fisher RA, Contos MJ, Mills AS. Similarities and differences in outcomes of cirrhosis due to nonalcoholic steatohepatitis and hepatitis C. Hepatology 2006 April;43(4):682-9.

(34) Dam-Larsen S, Franzmann M, Andersen IB, Christoffersen P, Jensen LB, Sorensen TI, Becker U, Bendtsen F. Long term prognosis of fatty liver: risk of chronic liver disease and death. Gut 2004 May;53(5):750-5.

(35) Evans CD, Oien KA, MacSween RN, Mills PR. Non-alcoholic steatohepatitis: a common cause of progressive chronic liver injury? J Clin Pathol 2002 September;55(9):689-92.


184

(36) Fassio E, Alvarez E, Dominguez N, Landeira G, Longo C. Natural history of nonalcoholic steatohepatitis: a longitudinal study of repeat liver biopsies. Hepatology 2004 October;40(4):820-6.

(37) Hui AY, Wong VW, Chan HL, Liew CT, Chan JL, Chan FK, Sung JJ. Histological progression of non-alcoholic fatty liver disease in Chinese patients. Aliment Pharmacol Ther 2005 February 15;21(4):407-13.


(39) Argo CK, Northup PG, Al Osaimi AM, Caldwell SH. Systematic review of risk factors for fibrosis progression in non-alcoholic steatohepatitis. J Hepatol 2009 August;51(2):371-9.

(40) Hotamisligil GS. Inflammation and metabolic disorders. Nature 2006 December 14;444(7121):860-7.

(41) Targher G, Bertolini L, Poli F, Rodella S, Scala L, Tessari R, Zenari L, Falezza G. Nonalcoholic Fatty liver disease and risk of future cardiovascular events among type 2 diabetic patients. Diabetes 2005 December;54(12):3541-6.

(42) Jackson R. Guidelines on preventing cardiovascular disease in clinical practice. BMJ 2000 March 11;320(7236):659-61.

(43) Nomura H, Kashiwagi S, Hayashi J, Kajiyama W, Tani S, Goto M. Prevalence of fatty liver in a general population of Okinawa, Japan. Jpn J Med 1988 May;27(2):142-9.


(45) Chen CH, Huang MH, Yang JC, Nien CK, Yang CC, Yeh YH, Yueh SK. Prevalence and risk factors of nonalcoholic fatty liver disease in an adult population of taiwan: metabolic significance of nonalcoholic fatty liver disease in nonobese adults. J Clin Gastroenterol 2006 September;40(8):745-52.

(46) Leh H, Courtay C, Gerardin P, Wellman M, Siest G, Visvikis A. Cloning and expression of a novel type (III) of human gamma-glutamyltransferase truncated mRNA. FEBS Lett 1996 October 7;394(3):258-62.



185






(53) Mottin CC, Moretto M, Padoin AV, Swarowsky AM, Toneto MG, Glock L, Repetto G. The role of ultrasound in the diagnosis of hepatic steatosis in morbidly obese patients. Obes Surg 2004 May;14(5):635-7.

(54) Moura Almeida A, Cotrim HP, Barbosa DB, de Athayde LG, Santos AS, Bitencourt AG, De Freitas LA, Rios A, Alves E. Fatty liver disease in severe obese patients: diagnostic value of abdominal ultrasound. World J Gastroenterol 2008 March 7;14(9):1415-8.


(56) Suzuki K, Kirikoshi H, Yoneda M, Mawatari H, Fujita K, Nozaki Y, Takahashi H, Abe Y, Inamori M, Shimamura T, Kobayashi N, Kubota K, Saito S, Nakajima A. Measurement of spleen volume is useful for distinguishing between simple steatosis and early-stage non-alcoholic steatohepatitis. Hepatol Res 2010 July 1;40(7):693-700.

(57) Friedrich-Rust M, Ong MF, Martens S, Sarrazin C, Bojunga J, Zeuzem S, Herrmann E. Performance of transient elastography for the staging of liver fibrosis: a meta-analysis. Gastroenterology 2008 April;134(4):960-74.


186

(58) Friedrich-Rust M, Hadji-Hosseini H, Kriener S, Herrmann E, Sircar I, Kau A, Zeuzem S, Bojunga J. Transient elastography with a new probe for obese patients for non-invasive staging of non-alcoholic steatohepatitis. Eur Radiol 2010 June 6.

(59) Kaser S, Ebenbichler CF, Tilg H. Pharmacological and non-pharmacological treatment of non-alcoholic fatty liver disease. Int J Clin Pract 2010 June;64(7):968-83.

(60) Caldwell S, Lazo M. Is exercise an effective treatment for NASH? Knowns and unknowns. Ann Hepatol 2009;8 Suppl 1:S60-S66.

(61) Yoshimitsu K, Irie H, Kakihara D, Tajima T, Asayama Y, Hirakawa M, Ishigami K, Noshiro H, Kakeji Y, Honda H. Postgastrectomy development or accentuation of focal fatty change in segment IV of the liver: correlation with the presence of aberrant venous branches of the parabiliary venous plexus. J Clin Gastroenterol 2007 May;41(5):507-12.

(62) Wing RR, Tate DF, Gorin AA, Raynor HA, Fava JL. A self-regulation program for maintenance of weight loss. N Engl J Med 2006 October 12;355(15):1563-71.

(63) Thomas EL, Brynes AE, Hamilton G, Patel N, Spong A, Goldin RD, Frost G, Bell JD, Taylor-Robinson SD. Effect of nutritional counselling on hepatic, muscle and adipose tissue fat content and distribution in non-alcoholic fatty liver disease. World J Gastroenterol 2006 September 28;12(36):5813-9.

(64) Marchesini G, Suppini A, Forlani G. NAFLD treatment: cognitive-behavioral therapy has entered the arena. J Hepatol 2005 December;43(6):926-8.

(65) Bellentani S, Dalle Grave R, Suppini A, Marchesini G. Behavior therapy for nonalcoholic fatty liver disease: The need for a multidisciplinary approach. Hepatology 2008 February;47(2):746-54.

(66) Huet PM, Villeneuve JP, Fenyves D. Drug elimination in chronic liver diseases. J Hepatol 1997;26 Suppl 2:63-72.

(67) Navarro VJ, Senior JR. Drug-related hepatotoxicity. N Engl J Med 2006 February 16;354(7):731-9.

(68) Cotrim HP, De Freitas LA, Freitas C, Braga L, Sousa R, Carvalho F, Parana R, Santos-Jesus R, Andrade Z. Clinical and histopathological features of NASH in workers exposed to chemicals with or without associated metabolic conditions. Liver Int 2004 April;24(2):131-5.

(69) Musso G, Gambino R, Cassader M, Pagano G. A meta-analysis of randomized trials for the treatment of nonalcoholic fatty liver disease. Hepatology 2010 July;52(1):79-104.


187

(70) Baumeister SE, Volzke H, Marschall P, John U, Schmidt CO, Flessa S, Alte D. Impact of fatty liver disease on health care utilization and costs in a general population: a 5-year observation. Gastroenterology 2008 January;134(1):85-94.

(71) Qian MY, Yuwei JR, Angus P, Schelleman T, Johnson L, Gow P. Efficacy and cost of a hepatocellular carcinoma screening program at an Australian teaching hospital. J Gastroenterol Hepatol 2010 May;25(5):951-6.


(73) Ratziu V, Charlotte F, Heurtier A, Gombert S, Giral P, Bruckert E, Grimaldi A, Capron F, Poynard T. Sampling variability of liver biopsy in nonalcoholic fatty liver disease. Gastroenterology 2005 June;128(7):1898-906.

(74) Ratziu V, Bugianesi E, Dixon J, Fassio E, Ekstedt M, Charlotte F, Kechagias S, Poynard T, Olsson R. Histological progression of non-alcoholic fatty liver disease: a critical reassessment based on liver sampling variability. Aliment Pharmacol

Ther 2007 September 15;26(6):821-30.

(75) Younossi ZM, Gramlich T, Liu YC, Matteoni C, Petrelli M, Goldblum J, Rybicki L, McCullough AJ. Nonalcoholic fatty liver disease: assessment of variability in pathologic interpretations. Mod Pathol 1998 June;11(6):560-5.

(76) Stolk RP, Wink O, Zelissen PM, Meijer R, van Gils AP, Grobbee DE. Validity and reproducibility of ultrasonography for the measurement of intra-abdominal adipose tissue. Int J Obes Relat Metab Disord 2001 September;25(9):1346-51.

(77) Stolk RP, Meijer R, Mali WP, Grobbee DE, van der Graaf Y. Ultrasound measurements of intraabdominal fat estimate the metabolic syndrome better than do measurements of waist circumference. Am J Clin Nutr 2003 April;77(4):857-60.

(78) Lucia Rolfe E, Sleigh A, Finucane FM, Brage S, Stolk RP, Cooper C, Sharp SJ, Wareham NJ, Ong KK. Ultrasound measurements of visceral and subcutaneous abdominal thickness to predict abdominal adiposity among older men and women. Obesity (Silver Spring) 2010 March;18(3):625-31.

(79) Wigg AJ, Roberts-Thomson IC, Dymock RB, McCarthy PJ, Grose RH, Cummins AG. The role of small intestinal bacterial overgrowth, intestinal permeability, endotoxaemia, and tumour necrosis factor alpha in the pathogenesis of non-alcoholic steatohepatitis. Gut 2001 February;48(2):206-11.

(80) Miele L, Valenza V, La Torre G, Montalto M, Cammarota G, Ricci R, Masciana R, Forgione A, Gabrieli ML, Perotti G, Vecchio FM, Rapaccini G, Gasbarrini G,


188

Day CP, Grieco A. Increased intestinal permeability and tight junction alterations in nonalcoholic fatty liver disease. Hepatology 2009 June;49(6):1877-87.

(81) Lammert C, Bjornsson E, Niklasson A, Chalasani N. Oral medications with significant hepatic metabolism at higher risk for hepatic adverse events. Hepatology 2009 September 14.

(82) Pichler WJ. Pharmacological interaction of drugs with antigen-specific immune receptors: the p-i concept. Curr Opin Allergy Clin Immunol 2002 August;2(4):301-5.

(83) Larrey D, Pageaux GP. Genetic predisposition to drug-induced hepatotoxicity. J Hepatol 1997;26 Suppl 2:12-21.


189

USG outcome:

Determ

ineFLD pathogenesisand

pathophysiology[ A]:

Action(s):

Action(s) to beconsidered:

Re-screen:

Norm

alliver

FLD

Fat +

F0-F1

Fat +

F1-F2

Abnorm

al

liver

HCC orotherlesion

Continu lifestyle

Questionnaire onbehaviour&lifestyle:

•Nutrition

•Alcohol and other drugs (current & recent history)

•Energy expenditure

(Fat +)

F3-F4

Questionnaire on complaints and (potential) diseases:

•Knowndiseases

•Known family disease history (e.g. HH)

Questionnaire on drugs/medicine:

•All drugs used currently

•All drugs used in recent history

Change lifestyle:

•Limit calorie intake [E]

•Limit alcohol and usedrugs

•Exersice[F]

Weightloss [G]:

•2.6 BMI-points[ C]

•Gradual, to prevent worsening [H]

These are likelyalreadybeingtreated

Questionnaire on Environment:

Known exposure to toxins at work or elsewhere

Protect for the environment [L]

Changelifestyle:

•Nutritional counselling[M]

•Rehabilitationprogram

•(Cognitive) behavioural therapy [N]

Weight loss:

•Drugs for weight loss [E]

•Bariatric surgery [E]

Screeningfor(family) disease

Examination of obesity, particularly VAT:

•APM for BMI and waist-to-hip ratio

•USG for VAT/SAT ratio [B]

Blood tests for assessment CVD [C] and/or hepatic risk:

•Glucose

•Lipids

•Blood pressure

•Inflammation

•Coagulation

Treatment for insulin resistance [E]

Examination of small intestinal bacterial overload:

Breath test [D]

Treat infection

Examination of lipodystrophy:

•APM/USG for SAT

•Questionnaire/blood tests on menopausal state

•Angiotensinreceptor blockade [E]

•Anti-TNFα[E]*

Examination of liver disease (other than FLD):

•Autoimmunehepatitis

•Viralhepatitis

Take into account genotype [O]

Correct unfavourable fat distribution:

•via Thiazolidinediones[E]

•via horm

ones(HRT)

t = 5

t = 3

t = 2

Appropriate treatm

ent

Refer to a specialist

Refer to a specialist

Fat +

F2-F3

•Eliminate/minimise drug use if possible[I]

•Use drugs without hepatic metabolism [J]

and/or without immunogeneity[K]

•Change nutrition

•Limit alcohol use

Change the environment [L]

Appen

dix І. P

ote

ntial

des

ign o

f a

scre

enin

g p

rogra

m f

or

fatt

y liv

er d

isea

se, w

hic

h is

not ev

aluat

ed y

et, th

us

pure

ly h

ypoth

etic

al

[A],

[CHAPTER 4

] 3; [B],

76-7

8; [C

], [CHAPTER 3

] 11; [D

], 7

9, 80; [E], 5

9; [F], 6

0; [G

], [CHAPTER 3

] 4, 11, 59, 62; [H

], 6

1; [I],

66, 67; [J], 8

1; [K

], 8

2; [L], 6

8; [M

], 6

3; [N

], 6

4, 65; [O

], 8

3.

AP

M,

anth

ropom

etry

; B

MI,

body m

ass

index

; C

VD

, ca

rdio

vas

cula

r dis

ease

; F

, fi

bro

sis

stag

e; F

LD

, fa

tty l

iver

dis

ease

; H

CC

, hep

atoce

llula

r ca

rcin

om

a; H

H,

her

edit

ary h

aem

och

rom

atosi

s; H

RT

, horm

one

repla

cem

ent

ther

apy;

SA

T,

subcu

taneo

us

adip

ose

ti

ssue;

t,

tim

e in

yea

rs;

TN

Fα,

tum

our

nec

rosi

s fa

ctor

alpha;

U

SG

, ult

raso

nogra

phy;

VA

T,

vis

cera

l ad

ipose

ti

ssue.

190

Chapter 8.2: General discussion - part 2

Additional remarks and recommendations for future

research


192

A few additional remarks and recommendations for future research on selected chapters

will be given.

ADDITIONAL REMARKS AND RECOMMENDATIONS TO CHAPTER 2:

“Fatty liver disease and cardiovascular risk in the general population of East Anglia:

The Fenland Study”.

• Regarding prevalence

As 762 participants is a (relatively) small study group, predominantly from the Ely test site,

one could question whether this sample is representative for the general population of

East Anglia. In order to improve representativeness, the distribution of sex and age should

match the distribution of sex and age of the population of East Anglia.

While writing chapter 2, data collection of the Fenland Study continued. At the time of

writing this section, approximately 2000 liver fat scans were available for analysis. The

analysis performed in chapter 2, will be repeated in this extended database, where after it

will be submitted for publication.

• Regarding the ultrasound method

Although the separate ultrasound measures used in this study are based on the literature, the

entire scoring method has not been validated yet. In a new version of chapter 2, based on

the enlarged database (please see point above), results of a validation study will be given.

• Regarding cardiovascular risk markers

The value of the MetS for predicting CVD events is often questioned. However, as

described in the discussion of this chapter, a recent meta-analysis revealed that presence of

the MetS (a dichotomous variable) is a predictor of future incident CVD events 1. It can,

and should, be studied whether the MetS is 1] a risk factor beyond its specific components

2, and 2] is more than the sum of its components (an ordinal variable) 3. Theoretically, more

atherogenic risk factors means a greater risk for CVD events. When predicting future

incident CVD event, scientists should compare in the same database the predictive value of

various risk factors: 1] specific MetS components 2 and other conventional CVD risk

factors, 2] the MetS dichotomous, 3] the MetS ordinal 3, 4] the MetS continuous (Z-score),

5] the 10 year Framingham risk score, and 6] FLD (which might be a better predictor of

CVD than the MetS dichotomous and conventional CVD risk factors 4).

Remarks and recommendations

193


“Ultrasonography to quantify hepatic fat content: validation by 1H magnetic

resonance spectroscopy”.

• Regarding reliability: part 1

As suggested in this chapter, detailed protocols have now been written to improve

reliability. The protocols stress out that this quantitative ultrasonography method is highly

dependant on the quality of the images and the quality of the image analysis. A major part

of the method is based on differences in pixel intensity between pixels. One can imagine

that when the ROI is drawn half in liver parenchyma contaminated by e.g. rib shadow,

this has consequences for validity of the method. Quality control regarding image analysis

has been implemented, in order to verify the quality of the image and image analysis.

Additionally, it would be of value to investigate whether it is possible to develop

computerized methods regarding recognition and exclusion of artefacts.

• Regarding reproducibility by other ultrasound equipment and computer screens

Another point to be discussed is differences in resolution (amount of pixels) between

ultrasound machines and computer screens. Some variables embedded in the algorithm are

based on sum scores, i.e. the sum of spatial distances between all pixels with a different

intensity (indices of the spatial grey level dependence matrix) and the sum of spatial

distances between all pixels with the same intensity (i.e. indices of the grey level difference

matrix). In theory, ultrasound machines and/or computer screens with higher resolutions,

i.e. with more pixels and pixel pairs than used in the present validation study, should

produce higher values of the indices and therefore a (much) higher outcome of the

algorithm. A difference between ultrasound machines is supported by a small experiment

using 3 volunteers and 3 different ultrasound machines, using the same computer screen for

analysis.

Acknowledgement: Jan Visscher for performing ultrasonography.

• Regarding reliability: part 2

While writing the previous section on resolutions, regarding generalisation to other

ultrasound equipment and computer screens, I suddenly realised something. Also in the

case a ROI is drawn larger, more pixels and pixel pairs are included. Thus theoretically,

differences in ROI size could underlie issues regarding intraobserver and interobserver

reliability. When modifying the software, we implemented the amount of pixels of the AL


194

(which did not improve outcome), but did not think about the amount of pixels included

within the ROI. If the theory described above is true, then perhaps both the issues regarding

reproducibility and reliability part 2 might be solvable by implementing the amount of

pixels or pixel pairs in the formula’s of all indices of the co-occurrence matrix. Perhaps

implementing the average value (in stead of the sum) would improve the method, but this is

for future validation work.


“MRI determined fat content of human liver, pancreas and kidney”.

In chapter 7 we concluded that renal lipid accumulation appears not to be involved in the

pathogenesis of renal disease. It should be noted that this chapter was written earlier than

chapter 4 “Pathogenesis of fatty liver disease: A theory on lipid content, inhibited

metabolism, and inflammation”. In chapter 4 is described that the liver transforms severely

toxic fatty acids into little toxic triglycerides, and stores these triglycerides in lipid vesicles

isolated from the cytosol. While writing that, I wondered whether the kidney and pancreas

also possess these protective abilities. Magnetic resonance spectroscopy and imaging do not

distinguish fatty acids and triglycerides. In other words, the little bit of fat measured within

the pancreas and kidney could be of the severely toxic fatty acid type, predisposing to

steatonecrosis. Thus assessment of fat content of the pancreas and kidney might also be

important.

Remarks and recommendations

195

Reference List (1) Gami AS, Witt BJ, Howard DE, Erwin PJ, Gami LA, Somers VK, Montori VM.

Metabolic syndrome and risk of incident cardiovascular events and death: a systematic review and meta-analysis of longitudinal studies. J Am Coll Cardiol 2007 January 30;49(4):403-14.

(2) Inchiostro S, Fadini GP, de Kreutzenberg SV, Citroni N, Avogaro A. Is the metabolic syndrome a cardiovascular risk factor beyond its specific components? J Am Coll Cardiol 2007 June 26;49(25):2465-6.

(3) Kurth T, Logroscino G. The metabolic syndrome: more than the sum of its components? Stroke 2008 April;39(4):1068-9.

(4) Hamaguchi M, Kojima T, Takeda N, Nagata C, Takeda J, Sarui H, Kawahito Y, Yoshida N, Suetsugu A, Kato T, Okuda J, Ida K, Yoshikawa T. Nonalcoholic fatty liver disease is a novel predictor of cardiovascular disease. World J Gastroenterol 2007 March 14;13(10):1579-84.

196

Summaries English summary

Nederlandse samenvatting (Dutch summary)

198

ENGLISH SUMMARY

CHAPTER 1 is the general introduction to this thesis and gives an overview on the history of

fatty liver disease (FLD) and the contribution of this thesis to our knowledge of FLD. The

livers’ ability to store fat has been known for a long time, as the first publication on human

liver fat dates from 1907. FLD is a broad spectrum, consisting of many subtypes, but is

often divided in steatosis (fat accumulation) and steatohepatitis (fat + inflammation, with or

without fibrosis). Throughout history many synonyms for the spectrum and its subtypes

have been published. Subtypes can be determined by histology only. The first proposal

(protocol) for uniformly scoring the histologic FLD lesions was published in 1999, and

updated in 2005. Other diagnosis methods, for liver fat content only include: biochemical

analysis (after liver biopsy), magnetic resonance spectroscopy (1H MRS), magnetic

resonance imaging (MRI), computed tomography (CT), and ultrasonography. As alcohol is

a historically well known risk factor for liver disease, alcoholic FLD (AFLD) and non-

alcoholic FLD (NAFLD) are often distinguished, usually using an ethanol cut-off value of

20 g/d. In 1992, a close association of NAFLD with the metabolic syndrome, a marker of

cardiovascular disease (CVD), was first reported.

CHAPTER 2 explores baseline data of the Fenland Study, which is an ongoing population

based cohort study with currently cross-sectional results only. The chapter describes the

prevalence of FLD and its association with CVD risk in the general population of East

Anglia, aged 30 to 58 years. Exclusion criteria were: diabetes mellitus, terminal illness,

inability to walk unaided, and pregnancy. FLD was assessed by ultrasonography (scored in

a cumulative fashion and attributed as normal, mild, moderate, and severe FLD). CVD risk

was estimated by the metabolic syndrome (MetS), using several definitions, and the 10-year

Framingham CVD risk score. Liver fat scores were obtained in 762 participants. Overall

prevalence of FLD and non-alcoholic FLD (NAFLD) was 38.5% and 30.5%, respectively.

FLD was more prevalent in men than women (p<0.001). By multiple logistic regression

analysis on the presence of FLD, only BMI (OR 1.3, p<0.05) was significantly associated

in men, and BMI (OR 1.6, p<0.001), waist circumference (OR 1.1, p<0.01), and hip

circumference (OR .9, p<0.01) were significantly associated in women. With increasing

liver fat category, the number of metabolic syndrome components (p<0.001 both ATP III

and IDF), the cumulative MetS Z-score (p<0.001) and the 10-year Framingham risk score

199

(p<0.001) increased as well. We conclude that this study shows a striking prevalence of

FLD in East Anglia, particularly in men. As FLD is associated with several CVD risk

estimates, this striking prevalence may delineate an increased CVD risk in this population.

CHAPTER 3 is a review paper on the cardiovascular disease (CVD) risk of non-alcoholic

fatty liver disease (NAFLD). NAFLD is associated with both hepatic and systemic insulin

resistance. In the case of NAFLD, the liver overproduces several atherogenic factors,

notably inflammatory cytokines, glucose, lipoproteins, coagulation factors, and factors

increasing blood pressure. Intervention studies with diet and bariatric surgery revealed

improvements of hepatic fat content and CVD risk profile. Pharmacological approaches to

reduce liver fat have been developed as well, but the effects are often confounded by

weight change. We conclude that NAFLD is associated with an increased CVD risk profile

(and hepatic risk). In order to improve CVD risk profile, prevention and treatment of

NAFLD seems advisable. However, well designed randomised interventions, and long-term

follow-up studies are scarce.

CHAPTER 4 is a review paper on the pathogenesis of FLD. FLD is the most prevalent

hepatic condition worldwide. Thorough understanding of risk factors and the pathogenesis

of FLD is therefore warranted. Few comprehensive theories on its pathogenesis have been

proposed. The aim of this paper was to critically discuss present theories on the

pathogenesis of FLD, and to arrange risk factors in an easily recognisable manner. The

literature, including behavioural, genetic, and environmental factors associated with FLD

was reviewed, together with their underlying role in the pathogenesis of FLD. Risk factors

for FLD were arranged according to pathogenesis. The following groups of risk factors for

FLD were identified: 1) ‘risk factors for hepatic lipid content’, 2) ‘risk factors for inhibited

hepatic metabolism’, and 3) ‘risk factors for hepatic inflammation’. Some risk factors can

be placed in more than one category and might therefore have a greater effect on FLD than

others. These three categories do not stand alone but interact, forming a downward spiral

onto the development and progression of FLD. We propose to modify current arrangements

of risk factors for FLD. The present arrangement of risk factors may be useful to identify

people at high risk for FLD and to initiate interventions.

CHAPTER 5 describes the development of an ultrasonography method for quantitative

assessment of liver fat content, and validation using magnetic resonance spectroscopy (1H

MRS) as gold standard. Eighteen White volunteers (BMI range 21.0 to 42.9) were scanned

200

by both ultrasonography and 1H MRS. Altered ultrasound characteristics, present in the

case of fatty liver disease (FLD), were assessed using a specially developed software

program. Various attenuation and textural based indices of FLD were extracted from

ultrasound images. Using linear regression analysis, the predictive power of several models

(consisting of both attenuation and textural based measures) on log 10-transformed hepatic

fat content by 1H MRS were investigated. The best quantitative model was compared with a

qualitative ultrasonography method, as used in clinical care. A model with four ultrasound

characteristics could modestly predict the amount of liver fat (adjusted explained variance

43.2%, p=0.021). Expanding the model to seven ultrasound characteristics increased

adjusted explained variance to 60% (p=0.015), with r=0.789 (p<0.001). Comparing this

quantitative model with qualitative ultrasonography revealed a significant advantage of the

quantitative model in predicting hepatic fat content (p<0.001). This validation study shows

that a combination of computer-assessed ultrasound measures from routine ultrasound

images can be used to quantitatively assess hepatic fat content.

CHAPTER 6 describes the modification and validation of a recently published Dixon-based

magnetic resonance imaging method (MRI) method/algorithm for quantification of liver fat

content using dual-echo breath-hold gradient echo imaging, with multi-voxel proton

magnetic resonance spectroscopy (1H MRS) as gold standard. Ten men were examined by

MRI and 1H MRS in one measurement session. Using a recently published MRI algorithm,

two problems were encountered: 1) MRI liver fat contents were too high in nine volunteers

(range 3.3-10.7% vs. 0.9-7.7%), and correct in the volunteer with the highest liver fat

content (21.1 vs. 21.3%), and 2) in one of the ten subjects the MRI fat content according to

the Dixon-based MRI method was incorrect due to a (100-x) versus x percent lipid content

mix-up. The second problem was fixed by a minor adjustment of the MRI algorithm.

Despite systematic overestimation of liver fat contents by MRI, Spearman's correlation

between the adjusted MRI liver fat contents with 1H MRS was high (r=0.927, P<0.001).

Even after correction of the algorithm, the problem remaining with the Dixon-based MRI

method for the assessment of liver fat content, is that, at the lower end range, liver fat

content is systematically overestimated by 4%.

CHAPTER 7 describes how we used the MRI method which was developed and validated in

chapter 6 to simultaneously compare fat content in the liver, pancreas, kidney, and

subcutaneous adipose tissue. In 36 volunteers with body mass index (BMI) ranging from

201

20.0 to 42.9 kg/m², the median fat contents of liver, pancreas and kidney were 2.3%

(interquartile range: 0.2%-7.8%), 2.7% (1.0%-6.5%) and 0.7% (0.1%-1.4%), respectively.

BMI and subcutaneous fat correlated significantly with liver and pancreas fat contents.

Shown for the first time is significant correlation of the liver and pancreas fat contents in

healthy controls (r=0.43, P<0.01). These observations are related to body weight as

measured by BMI and the amount of subcutaneous fat. Kidney fat content is very low and

correlates with none of the other fat depots. We conclude that renal lipid accumulation,

unlike the coupled accumulations of fat in liver and pancreas, is not observed in obese

subjects. Unlike what has been suggested in previous studies, renal lipid accumulation

appears not to be involved in the pathogenesis of renal disease in humans.

CHAPTER 8 is the general discussion of this thesis. As FLD is associated with health risks,

notably CVD risk as described in chapters 2 and 3, we addressed the question if it would be

beneficial to screen for FLD. The world health organisation (WHO) screening criteria,

originally introduced by Wilson and Jungner, were organised as ‘the condition’,

‘diagnosis’, ‘treatment’, and ‘cost & screening program’ and discussed stepwise. Compared

to reference populations, survival of both patients with non-alcoholic FLD (NAFLD;

including both steatosis and steatohepatitis) and patients with non-alcoholic steatohepatitis

(NASH) is reduced. Moreover, both liver-related mortality and cardiovascular-related

mortality are increased. An estimated 8.3% of patients with NAS will develop fibrosis, and

an estimated 37.6% of patients with NASH will progress in fibrosis stage. Ultrasonography

is an acceptable method for assessing FLD. Most FLD cases have a behaviour-related

etiology, which provides opportunity for treatment. Studies on cost-effectiveness of

screening for FLD are lacking, but prospects are promising given the increased costs (both

financial and quality of life) of patients with FLD. Based on the evidence presented in this

paper, we conclude that screening for FLD is advisable. However, cost effectiveness studies

on screening for fatty liver disease are yet to be performed.

202

NEDERLANDSE SAMENVATTING (DUTCH SUMMARY)

In het geval van aanhoudend vetopslag in de lever zal fatty liver disease (FLD) ontstaan.

FLD is een breed spectrum dat begint met alleen vetopslag (steatose), wat kan leiden tot

schade en ontstekingen (steatohepatitis). Steatohepatitis kan leiden tot cirrose en/of

levertumoren. FLD is aanwezig bij 25% tot 40% van de Westerse volwassenen. In een

groot bevolkingsonderzoek toonden we aan dat FLD sterk is geassocieerd met markers van

hart- en vaatziekten. Uit ons literatuur overzicht blijkt dat de vervette lever een verhoogde

productie geeft van vele markers van hart- en vaatziekten, zoals lipiden, glucose,

inflammatoire cytokines, stollingsfactoren (productie daalt in het geval van cirrose), en

factoren die de bloeddruk verhogen. Door af te vallen verbeteren zowel FLD als markers

van hart- en vaatziekten. Daarnaast zijn er veelbelovende farmacologische behandelingen.

Op basis van onze studie en de literatuur stellen we dat risicofactoren voor het ontstaan van

FLD bestaan uit drie categorieën: 1] risico’s voor levervet (b.v. calorie inname, beweging),

2] risico’s voor beperkt lever metabolisme (b.v. alcohol, bepaalde medicatie), en 3] risico’s

voor lever ontsteking (b.v. hepatitis, ziekte van Crohn). Beeldvormende technieken voor het

meten van levervet zijn magnetic resonance spectroscopie en imaging, computed

tomografie, en echografie. Doorgaans wordt echografie kwalitatief gebruikt, maar wij

hebben een kwantitatieve methode ontwikkeld en gevalideerd. Gezien de risico’s, de

meetbaarheid, en de behandelbaarheid, concluderen we dat het zinvol is om te screenen op

FLD, door middel van echografie.

Abbreviations

204

LIST OF ABBREVIATIONS

A

AFLD, alcoholic fatty liver disease AGEs, advanced glycation endproducts AL, attenuation line ALT, alanine aminotransferase AR, anterior right ATP ΙΙΙ, adult treatment panel 3 B

BMI, body mass index C

(-CH2-)n, methylene CRP, C-reactive protein CSI, chemical shift imaging CT, computed tomography CVD, cardiovascular disease F

FLD, fatty liver disease (including AFLD and NAFLD) G

GGT, gamma-glutamyl aminotransferase GLUT-4, glucose transporter number 4 H

HC, hepatocyte HCC, hepatocellular carcinoma HDL, high density lipoprotein HFFAs, hepatic free fatty acids 1H MRS, magnetic resonance spectroscopy H2O, water HOMA-IR, insulin resistance by homeostatic model assessment HSC, Hepatic Stellate Cell I

IDF, international diabetes federation IL-6, interleukin 6 IP, in-phase K

KC, Kupffer Cell

205

L

LDL, low density lipoprotein M

MetS, metabolic syndrome MetS, metabolic syndrome MR, magnetic resonance MRI, magnetic resonance imaging mRNA, messenger ribonucleic acid

N

NAFLD, non-alcoholic fatty liver disease (including NAS and NASH) NAS, non-alcoholic steatosis NASH, non-alcoholic steatohepatitis NEFA, non-esterified fatty acids NF-κB, nuclear factor kappa-beta O

OP, out-phase P

PL, posterior left PPAR, peroxisome proliferator-activated receptor ppm, parts per million PRESS, point resolved spectroscopy R

ROI, region of interest S

SAT, subcutaneous adipose tissue T

TE, echo time TGs, triglycerides TNFα, tumour necrosis factor-alpha TR, repetition time V

VAT, visceral adipose tissue VLDL, very low density lipoprotein

206

Dankwoord (Word of gratitude)

208

Hoewel ik zelf natuurlijk de meest significante factor ben geweest in het voltooien van dit

proefschrift ;-) , wil ik een aantal mensen bedanken voor hun significante bijdrage aan het

gehele proces en/of aan afzonderlijke hoofdstukken.

Natuurlijk als eerste, betreffende het gehele proces, mijn promotor Ronald Stolk. Ronald,

bedankt voor alles.

Folkert Kuipers _ Naast je bijdrage aan hoofdstuk 3 waren deze samenwerking en de

‘fatty liver group’ de basis voor een deel van de rest van mijn proefschrift. Waarvoor heel

erg veel dank. Bert Groen _ Dank voor je onmisbare input betreffende hoofdstuk 4.

Bij het ontwikkelen van de kwantitatieve echo methode (hoofdstuk 5), waren nogal een

aantal mensen betrokken. Ik dank Mark Haagmans en prof. van der Jagt voor hun echo

bijdrage, Peter van Ooijen en Wisnu Kristanto voor al hun ideeën en onmisbare bijdrage

betreffende de beeldverwerking van echo’s en het aanpassen van de software, en Wendy

Post voor haar statische bijdrage. Wendy, alias ‘coach Wendy’, ik heb veel van je geleerd

en ik vond het super leuk om met je samen te werken. Tevens dank ik Peter Kappert, Jan-

Hendrik Potze, Irene Willeboordse en Annemarie van Tienhoven voor MR scanning, en

Paul Sijens voor MR analyses. Paul Sijens en Roy Irwan _ dank voor het vragen van

mijn bijdrage bij jullie goede ideeën. Daarnaast bedank ik alle proefpersonen voor hun tijd

en bijdrage aan dit hoofdstuk. Moreover, chapter 5 would not have been possible without

the help of Diana Gaitini and Haim Azhari. Thank you for your pioneer work and your

kind gesture to share your software program with us. A deep bow and big thanks to both of

you and to your team(s). Jan Visscher _ Ook jouw echo bijdragen mogen niet

onopgemerkt blijven. Dank voor al je hulp.

Nita Forouhi _ Thank you for having me over in Cambridge for two weeks, and for your

help with chapter 2. I also thank Ema de Lucia-Rolfe, Richard Powell, Ruhul Amin,

Adam Dickinson, and all other people making the Fenland Study possible. Nick

Wareham _ Thank you for giving me the opportunity to work on the Fenland Study.

Tot slot dank ik de leescommissie: prof. dr. F. Kuipers, prof. dr. M. Oudkerk, en prof.

dr. M. H. Hofker, voor het lezen van het manuscript.

Naast het ‘professionele gedeelte’ van dit dankwoord, wil ik nog een aantal mensen

bedanken.

Allereerst de mensen waar ik tijdens mijn AIO schap het allermeeste mee te maken heb

gehad. Mijn kamergenootjes (my dear roomies), tevens vrienden, in chronologische

volgorde: Hanneke, Josien, Li en Sylvia. Thanks for such a nice ‘home’. Every day was a

fun day with you!

‘Kelderbewoners’ en/of E4 ganggenootjes, tevens vrienden, in alfabetische volgorde:

Carianne, Eryn, Hiltje, Ingrid, en Laura. Naast het zijn van pionier Epi AIO’s, was het

erg leuk om jullie als collega’s te hebben. Daarnaast wil ik jullie bedanken voor alle

activiteiten buiten het werk!

Simona, Sacha, en Hans B, ik kijk met heel veel plezier terug op onze bezoeken aan de

sportschool, onze lunch wandelingen, en onze tafeltennis ‘battles’!

Additionally, I thank all members of the Epi volleyball teams for the nice tournaments, all

(old) floor-E4 colleages (Cleo, Matteusz, Marjan, Karin, Dennis, Judith, and all others) for

the nice chats, and also Aukje and Petra!

Dear paranimfen Li and Simona, thanks for supporting me on the 19th!

Als laatste, maar zeker niet als minste, bedank in mijn vrienden en familie (mam, pap

Edwin, oma’s) voor alle steun!

Mireille Edens

December 2010

209

210

Curriculum Vitae About the author & List of publications

212

ABOUT THE AUTHOR

Mireille Edens was born in 1980 (September 15th) in Beerta, which is a small village in the

North-East of the Netherlands. She received primary education at the Openbare

Basisschool Beerta and secondary education at the Dollard College Winschoten. Starting

with the MAVO, she subsequently climbed up to the HAVO and to the VWO (A-levels).

In 2000 she started the study Movement Sciences at the University of Groningen. She

performed a scientific internship at the Home Mechanical Ventilation Center (Department

of Pulmonology, University Medical Center Groningen), under supervision of dr. P. J.

Wijkstra. This scientific internship has resulted in two papers on which she graduated with

dr. M. H. G. de Greef.

In 2006 she started a PhD at the department of Epidemiology (University Medical Center

Groningen) under supervision of prof. dr. R. P. Stolk. The result of her PhD is currently

lying in front of you ...

213

LIST OF PUBLICATIONS

9. Edens MA, Forouhi NG, de Lucia-Rolfe E, Wareham NJ, Stolk RP, and other investigators of the Fenland Study; Authors to be determined.

Fatty liver disease and cardiovascular risk in the general population of East Anglia: The Fenland Study. In preparation/ awaiting enlargement of database.

8. Edens MA, Leegte MC, ter Veen J, van Weert E, Jager PL, de Greef MH, Wijkstra PJ. The clearance ability of cough assisting techniques in patients with a neuromuscular disease: Results of a pilot study. Submitted/ under review.

7. Edens MA, Stolk RP. Evidence on screening for fatty liver disease: Future perspectives.


6. Edens MA, Groen AK, Stolk RP. Pathogenesis of fatty liver disease: A theory on lipid content, inhibited metabolism, and inflammation.


5. Sijens PE, Edens MA, Bakker SJ, Stolk RP. MRI-determined fat content of human liver, pancreas and kidney. World J Gastroenterol 2010; 16(16):1993-1998.

4. Edens MA, van Ooijen PM, Post WJ, Haagmans MJ, Kristanto W, Sijens PE, van der Jagt EJ, Stolk RP. Ultrasonography to quantify hepatic fat content: Validation by 1H Magnetic Resonance Spectroscopy.

Obesity (Silver Spring) 2009; 17(12):2239-2244.

3. Edens MA, Kuipers F, Stolk RP. Non-alcoholic fatty liver disease is associated with cardiovascular disease risk markers.

Obes Rev 2009; 10(4):412-419 _ Review Paper

2. Irwan R, Edens MA, Sijens PE. Assessment of the variations in fat content in normal liver using a fast MR imaging method in comparison with results obtained by spectroscopic imaging. Eur Radiol 2008; 18(4):806-813.

1. Edens MA, van Son WJ, de Greef MH, Levtchenko EN, Blijham T, Wijkstra PJ. Successful treatment of respiratory dysfunction in cystinosis by nocturnal non-invasive positive pressure ventilation. Clin Nephrol. 2006 Oct;66(4):306-309.

In addition to the papers listed above, Mireille also hopes to publish (parts of) chapter 1, and some

other papers she has been working on.

214

215

216

Documents

University of Groningen Fatty liver disease Edens, Mireille ...Eur Radiol 2008; 18(4):806-813. 7 MRI-determined fat content of human liver, pancreas and kidney. 149 Sijens PE, Edens