Non-alcoholic steatohepatitis: A non-invasive diagnosis by analysis
of exhaled breath
Froukje J. Verdam1,2, Jan W. Dallinga3, Ann Driessen4, Charlotte de Jonge1,2, Edwin J.C. Moonen3,
Joep B.N. van Berkel3, Jakobus Luijk3, Nicole D. Bouvy1, Wim A. Buurman1, Sander S. Rensen1,⇑,
Jan Willem M. Greve2, Frederik Jan van Schooten3
1Department of General Surgery, Nutrition and Toxicology Research Institute Maastricht (NUTRIM), Maastricht University Medical Centre,
Maastricht, The Netherlands;2Department of General Surgery, Atrium Medical Centre Parkstad, Heerlen, The Netherlands;3Department
of Toxicology, Nutrition and Toxicology Research Institute Maastricht (NUTRIM), Maastricht University Medical Centre, Maastricht,
The Netherlands;4Department of Pathology, Maastricht University Medical Centre, Maastricht, The Netherlands
See Focus, pages 407–408
Background & Aims:Histological evaluation of a liver biopsy is
the current gold standard to diagnose non-alcoholic steatohepa-
titis (NASH), but the procedure to obtain biopsies is associated
with morbidity and high costs. Hence, only subjects at high risk
are biopsied, leading to underestimation of NASH prevalence,
and undertreatment. Since analysis of volatile organic com-
pounds in breath has been shown to accurately identify subjects
with other chronic inflammatory diseases, we investigated its
potential as a non-invasive tool to diagnose NASH.
Methods:Wedge-shaped liver biopsies from 65 subjects (BMI
24.8–64.3 kg/m2) were obtained during surgery and histologi-
cally evaluated. The profile of volatile organic compounds in
pre-operative breath samples was analyzed by gas chromatogra-
phy–mass spectrometry and related to liver histology scores and
plasma parameters of alanine aminotransferase (ALT) and aspar-
tate aminotransferase (AST).
Results:Three exhaled compounds were sufficient to distinguish
subjects with (n = 39) and without NASH (n = 26), with an area
under the ROC curve of 0.77. The negative and positive predictive
values were 82% and 81%. In contrast, elevated ALT levels or
increased AST/ALT ratios both showed negative predictive values
of 43%, and positive predictive values of 88% and 70%, respec-
tively. The breath test reduced the hypothetical percentage of
undiagnosed NASH patients from 67–79% to 10%, and of misdiag-
nosed subjects from 49–51% to 18%.
Conclusions:Analysis of volatile organic compounds in exhaled
air is a promising method to indicate NASH presence and
absence. In comparison to plasma transaminase levels, the breath
test significantly reduced the percentage of missed NASH
patients and the number of unnecessarily biopsied subjects.
? 2012 European Association for the Study of the Liver. Published
by Elsevier B.V. All rights reserved.
Non-alcoholic fatty liver disease (NAFLD) is the most prevalent
liver disease worldwide, affecting one in three adults, and one
in ten adolescents in the USA [1,2]. NAFLD is present in the
majority of patients with metabolic risk factors such as obesity
and type 2 diabetes mellitus (T2DM) . While steatosis, the
early stage of NAFLD, is considered to be benign and reversible,
progression towards more advanced stages often occurs. These
advanced stages, referred to as non-alcoholic steatohepatitis
(NASH), are characterized by inflammation [3,4]. Importantly,
NASH is in turn associated with the development of hepatic fibro-
sis, cirrhosis, hepatocellular carcinoma, and an increased risk of
liver failure and liver-related mortality [3,4]. It is therefore clini-
cally relevant to differentiate between patients with sole hepatic
steatosis and those suffering from NASH, at an early stage.
Currently, a liver biopsy remains necessary to accurately diag-
nose NASHand to assess itsseverity[5,6]. However, the procedure
discomfort, costs, and morbidity; significant complications are
encountered in 0.5% of cases [7,8]. In order to optimize the risk-
patients with clinical risk factors, and a per-operative biopsy from
all morbidly obese patients undergoing abdominal surgery .
tate aminotransferase (AST) to ALT (AST/ALT ratio), insulin resis-
tance, hypertension, sleep apnea, and increased plasma levels of
Journal of Hepatology 2013 vol. 58j543–548
Keywords: NAFLD; NASH; Obesity; Liver biopsy; Exhaled air; Volatile organic
Received 29 February 2012; received in revised form 10 October 2012; accepted 31
October 2012; available online 7 November 2012
qDOI of original article: http://dx.doi.org/10.1016/j.jhep.2012.12.010.
⇑Corresponding author. Address: Department of Surgery, Maastricht University
Medical Centre, PO Box 616, 6200 MD Maastricht, The Netherlands. Tel.: +31
433881496; fax: +31 433884154.
E-mail address: email@example.com (S.S. Rensen).
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase;
BMI, body mass index; NAFLD, non-alcoholic fatty liver disease; NASH, non-alc-
oholic steatohepatitis; VOC, volatile organic compounds; HbA1c, glycated hemo-
globin; LR, likelihood ratio; PV, predictive value; GC–MS, gas chromatography–
triglycerides . However, performing liver biopsy procedures
based upon these risk factors leads to a selection bias in clinical
practice, especially since mainly the presence of obesity and
plasma levels of aminotransferases are taken into account. On
the one hand, these plasma levels are often maintained within
the normal range despite advanced disease , resulting in an
underestimation of NASH prevalence, as well as undertreatment.
upon elevated aminotransferase levels, a large proportion of biop-
elevated AST and/or ALT levels are not specific for the presence of
this liver disease.
In view of (1) the clinical relevance of NASH, (2) the difficul-
ties of selecting the appropriate population to biopsy, and (3)
the biopsy-related burden, a less invasive method to identify
patients with NASH is urgently required. Such a method could
be the analysis of volatile organic compounds (VOC) in exhaled
breath. VOC are considered as markers of oxidative stress and
can indicate the presence of reactive oxygen species that derive,
for example, from peroxidation of polyunsaturated fatty acids
. Components in exhaled air have been previously shown to
reflect the presence of inflammatory diseases affecting the air-
ways [13,14] and liver [15–17]. Hence, analysis of VOC in exhaled
air may be useful for predicting NASH presence. In this pilot
study, we found that subjects with NASH can be accurately dis-
tinguished from those without NASH based upon analysis of
VOC in exhaled breath.
Patients and methods
Sixty-five subjects were included consecutively between October 2007 and May
2011, before they underwent laparoscopic abdominal surgery; either cholecystec-
tomy or primary bariatric surgery. Subjects ranged from overweight to severely
obese with a BMI range of 24.8–64.3 kg/m2. The laparoscopic abdominal surgery
was performed either at the Maastricht University Medical Centre or the Atrium
Medical Centre Parkstad by the same surgeon (JWG). Exclusion criteria were acute,
diseases (such as viral hepatitis), consumption of >10 g alcohol daily, and use of
medication associated with NAFLD (e.g., steroids, amiodarone, valproate, metho-
and sleep apnea were evaluated. This study was approved by the Medical Ethical
Sample collection and analysis
ysis have been published previously [13,18]. Briefly, all subjects exhaled into a
resistance-free 5 L plastic bag (Tedlar bag, SKC Ltd, Dorset, UK). The VOC were
trapped within 24 h after sampling, by deflating the bag into a sorption tube filled
with carbograph 1TD/Carbopack X (Markes International Inc, Cincinnati, OH). For
flight mass spectrometry (Tempus Plus, Thermo Fischer Scientific) .
Preoperative fasting venous blood samples could be obtained on the morning of
the surgery from 61 subjects and were processed as previously described .
Plasma levels of C-reactive protein (CRP), ALT, AST, glucose, insulin, total choles-
terol, HDL, LDL, triglycerides, free fatty acids and HbA1cwere measured according
to the protocol of the Department of Clinical Chemistry of the Maastricht Univer-
sity Medical Centre. The upper limit of normal ALT levels was 35 IU/L for women
and 45 IU/L for men , while an AST/ALT ratio >1 was considered to be elevated
Wedge-shaped liver biopsies of at least 15 by 10 mm were obtained intra-oper-
atively from all patients, by the same surgeon (JWG), and processed as previously
described . All biopsies contained at least five portal tracts to allow for correct
evaluation of the hepatic architecture. No overt pathologic condition other than
NAFLD was observed. Steatosis, hepatocellular ballooning, lobular and portal
inflammation, Mallory’s hyaline, and fibrosis were scored according to both the
Brunt scoring system  and the NAS activity score according to Kleiner et al.
, by an experienced liver pathologist (AD) blinded to the clinical context and
laboratory parameters. Liver biopsies that were evaluated as healthy or steatotic
did not show any sign of portal or lobular inflammation, hepatocyte ballooning, or
fibrosis (n = 26). In contrast, livers showing signs of steatosis and inflammation
were defined as NASH (n = 39) and were further evaluated according to the Brunt
classification and NAS activity score according to Kleiner et al. [5,6].
Data processing and analysis
Processing of data
Detailed descriptions of the data handling procedures have been previously
reported . Briefly, gas chromatography and mass spectrometry (GC–MS) chro-
matograms of all breath samples were recorded. Retention times were normal-
ized by calculating retention indices, relative to toluene and using easily
recognizable component peaks, to correct for chromatographic drifting. The
beginning and end of each run (retention index either <0.15 or >2.8) were
removed because of noisy mass spectra at the beginning of the chromatograms
and column bleeding at the end of each run. The remaining data, containing
almost 4800 different chromatographic peaks as determined by retention time
and mass spectrum combined with a relative intensity, were transformed into
Excel files. The measured mass spectra were compared to one another at the same
retention time. The resemblance of the original spectra determines whether or
not peaks at the same retention time represent the same component. Intensities
under the detection limit were set at 0%.
Statistical analysis of the GC–MS data
The data matrix was analyzed by a stepwise discriminant analysis by a leave-one-
out cross-over approach, using Statistical Package for Social Sciences 19.0.0 (IBM
SPSS Software Inc., Chicago, IL). All but one of the chromatograms were included
to construct the discriminant function. The one that was left out was subse-
quently used to predict the group it belonged to. This was repeated until every
chromatogram had been left out once; all samples have been classified. Based
upon 33 components, the discriminant functions that are optimal in terms of dif-
ferentiation between both groups are not necessarily the best predictors for
unknown samples, because of obvious overfitting. Therefore, the number of vari-
ables was gradually diminished until a reasonable small number of components
with sufficient predictive power remained. This reduction in components was
reached by repeating the analysis from the original large dataset by leaving the
least informative components out, one by one.
Statistical analysis of clinical data
Statistical analysis was performed using SPSS and Prism 5.0 (GraphPad Software
Inc., San Diego, CA) for Windows. Differences between groups were analyzed by
the Mann Whitney U test or the Kruskal Wallis-test followed by Dunn’s post-test-
ing. A p <0.05 was considered statistically significant. Data are presented as
mean ± standard error of the mean.
The body mass index (BMI) ranged from 24.8 to 64.3 kg/m2
(mean 43.7 kg/m2), population characteristics are summarized
in Table 1. NASH was diagnosed in 39 subjects (60%). The average
plasma ALT and AST levels were higher in subjects with NASH
Journal of Hepatology 2013 vol. 58j543–548
compared to subjects without NASH (Table 1). Importantly,
parameters such as gender, age, BMI, and HbA1cdid not differ
significantly. The 39 NASH patients were scored according to
the scoring systems of Brunt and Kleiner; most subjects suffered
from relatively mild NASH. As depicted in Table 2, the majority of
subjects categorized in grade 1 or 2 and in stage 0 or 1 according
to Brunt, whereas according to the NAS activity score by Kleiner
et al. most subjects showed steatosis grade 2, ballooning grade 1,
lobular inflammation grade 1, and no or little fibrosis.
Assessment based on volatile organic compounds
Analysis of VOC showed that subjects with NASH could be
distinguished from those without NASH with a sensitivity of
90% and a specificity of 69%, based upon a combination of three
components. These three most discriminating compounds were
(1) n-tridecane, (2) 3-methyl-butanonitrile and (3) 1-propanol
as identified by means of their mass spectrum. When more com-
ponents were taken into account, the prediction became more
accurate, and as expected, when less compounds were taken into
account, the specificity of breath analysis diminished accordingly
(Fig 1A). The optimal relation between sensitivity and specificity
was calculated at various cut-off values for the discriminant func-
tion, as visualized in the receiver operating characteristic (ROC)
curve. For the three described components, the area under the
curve was 0.77 ± 0.07 (Fig. 1B). In order to interpret the clinical
value of this test, likelihood ratios (LR) and predictive values
(PV) were calculated [23,24]. The positive LR is the probability
that a positive test result actually reflects the presence of the
disease. This was 2.90 for the three components (Table 3),
whereas the negative LR, a measure of the probability that a neg-
ative test result reflects the absence of NASH, was 0.15. The actual
prevalence was taken into account by calculating the positive PV,
the percentage of subjects with a positive test result who actually
suffer from NASH, which was 81%. The negative PV, the propor-
tion of patients with a negative test result that do not suffer from
NASH, was 82% (Table 3). If the decision to obtain a liver biopsy in
our study population had been based upon the three mentioned
components in exhaled breath, 43 patients would have been
biopsied, of whom eight did not have NASH (19%), while four
subjects with NASH (10%) would have been missed. In total, 12
out of the 65 subjects (18%) would have been misdiagnosed.
Assessment of NASH based on elevated plasma transaminase levels
Currently, the decision to obtain a liver biopsy in clinical practice
is largely based upon either elevated ALT levels or increased AST/
ALT ratio. In order to assess the clinical value of the breath test,
we compared the obtained data to the predictive properties of
our current clinical procedure. Hence, we calculated the propor-
tion of the population that would have undergone a liver biopsy
based upon plasma levels of transaminases. Plasma could not be
obtained from four subjects (of whom two suffered from NASH
and two did not), therefore the remaining 61 subjects were ana-
lyzed. If the indication to obtain a biopsy had been based upon
elevated ALT levels, eight biopsies (of seven patients with and
one without NASH) would have been obtained. Perhaps more
importantly, 30 out of the 37 remaining subjects with NASH
(81%) would have been missed (Fig. 2). Increased ALT levels
showed a sensitivity of 19% and a specificity of 96%. The positive
PV for increased ALT was 88% while the negative PV was 43%
Next, AST/ALT ratios were calculated. If a liver biopsy had
been obtained from subjects with an increased AST/ALT ratio,
17 biopsies would have been performed; 12 of subjects with
NASH and 5 of subjects without NASH. Moreover, 25 patients
Table 1. Characteristics of the study population.
Subjects without NASH Subjects with NASH
45 ± 2
41.0 ± 1.3
6.4 ± 0.3
6.0 ± 0.44
22.9 ± 2.0
18.8 ± 1.5
0.9 ± 0.1
No. of patients
44 ± 2
45.2 ± 1.4
6.8 ± 0.3
6.7 ± 2.1
29.6 ± 2.5*
27.1 ± 2.2*
1.1 ± 0.1
Data are presented as mean ± SEM.
⁄ALT and AST levels differed significantly between the groups (for both p <0.05),
whereas no significant differences were observed for all other parameters.
Table 2. Activity, stages and grades of NASH.
NASH subjects (n = 39)
NASH subjects (n = 39)
JOURNAL OF HEPATOLOGY
Journal of Hepatology 2013 vol. 58j543–548
or 68% of all subjects with NASH would have been missed (Fig. 2).
The sensitivity of an increased AST/ALT ratio was 32%, the speci-
ficity 79%, the positive PV 70%, and the negative PV 43%, respec-
tively (Table 3).
The consequences of assessing NASH presence based upon
breath in comparison to plasma in the current study population
are striking. The percentage of subjects that would have under-
gone liver biopsies based upon either VOC, increased plasma
ALT, or AST/ALT ratio is depicted in Fig. 2. It can be concluded that
the diagnostic value of VOC is much higher than that of plasma
transaminases, resulting in less misdiagnosed patients. More spe-
cifically, the prediction did not match the actual histological
hepatic evaluation in 18% of the subjects (12 out of 65) based
upon VOC, in 51% based upon ALT (31 out of 61) and in 49%
(30 out of 61) based upon AST/ALT ratio. The inconsistent predic-
tive properties of plasma transaminases were further emphasized
by the observation that only one patient had both elevated ALT
levels and an increased AST/ALT ratio. Patients who were cor-
rectly identified with NASH by means of plasma parameters were
also identified by means of the breath test. In addition, the breath
test identified a considerable proportion of NASH patients with
normal plasma parameters. In order to assess the potential addi-
tive value of considering other known risk factors, multivariate
analysis was performed including BMI, age, gender, hypertension,
apnea, and HbA1c, plasma ALT, AST, triglycerides, glucose, and
insulin. Only the discriminative value of VOC reached statistical
significance (p <0.05), all other parameters did not show any
additive value to predict NASH presence (Table 4).
In order to find a less invasive method to diagnose NASH, VOC in
exhaled breath were examined in relation to the evaluation of
wedge-shaped liver biopsies in a cohort of 65 overweight and
obese subjects. The feasibility to predict the presence of NASH
based upon three components in exhaled breath (n-tridecane,
3-methyl-butanonitrile, and 1-propanol) was found to be high
compared to plasma and clinical parameters. Moreover, by means
of these three exhaled breath components, almost all of the sub-
jects who would have been missed based upon plasma ALT and
AST/ALT ratio, could be identified. The fact that many subjects
are missed in case of risk assessment based upon ALT levels or
AST/ALT ratio is supported by others. It has been shown previ-
ously that plasma aminotransferase levels do not correlate well
with disease activity. For example, patients with NASH can still
have normal AST/ALT ratios and normal transaminase levels
despite advanced disease .
In order to enhance chances of cure and to prevent potential
progression into fibrosis and cirrhosis, it is essential to diagnose
NASH at an early stage, especially since recent longitudinal stud-
ies showed that patients with NASH already run a high risk to
develop liver-related complications and to suffer from liver-
related mortality [25,26]. The Non-alcoholic Steatohepatitis Clin-
ical Research Network recently stated that clinical and laboratory
parameters are insufficient to reliably diagnose NASH . How-
ever, given the associated discomfort, complications and costs,
the burden of screening for this emerging liver disease by means
of percutaneous liver biopsies would be disproportional to the
benefit. The necessity to find minimal invasive cost-effective
ways of screening will further intensify, given the rising inci-
dence of obesity and NASH. Current research therefore focuses
on less invasive methods to differentiate between the various
stages of NAFLD. For example, plasma markers of epithelial cell
damage, e.g., cytokeratin-18 and its fragments, are promising to
assess NASH presence [27,28]. Radiological modalities such as
ultrasonography  and transient elastography [30,31] may also
assist in identifying subjects with NASH. Transient elastography
(tissue elasticity measurement based on ultrasound technology)
was found to aid in predicting the presence of fibrosis  and
0.0 0.2 0.4 0.6 0.8 1.0
0 2 4 6 8 10 12
in exhaled breath
Fig. 1. Analysis of the number of breath components and ROC curve. (A)
Sensitivity and specificity of exhaled breath related to the number of components
incorporated in the discriminant function. With 1 or 2 components, the
sensitivity of exhaled breath to predict NASH presence in overweight and obese
subjects is 100% with a specificity of 23% and 39% respectively, whereas the
optimum is with three components (sensitivity 90%, specificity 69%). (B) ROC
curve for discriminant function based on n-tridecane, 3-methyl-butanonitrile,
and 1-propanol, the area under the curve is 0.77 ± 0.07 (95% CI 0.64–0.89).
Table 3. Test accuracy of VOC analyses and elevated plasma parameters.
Hypothetical % of subjects
with NASH without biopsy
Hypothetical % of biopsies
from subjects without NASH
Hypothetical % of subjects
who were misdiagnosed
Fig. 2. Number of subjects who are (mis)diagnosed based upon breath or
plasma transaminase levels. (A) Percentage of patients falsely predicted not to
have NASH by VOC (4/39 or 10%), increased ALT values (30/37 or 81%), and
increased AST/ALT ratio (25/37 or 68%). (B) The percentage of obtained biopsies
from patients without NASH would be 7/29 (16%) for VOC, 1 out of 8 (13%) for
ALT, and 5 out of 17 (29%) for AST/ALT, if the indication to obtain a biopsy had
been based hereupon. (C) The percentage of misdiagnosed patients was 12/65
based upon VOC (18%), 31/61 (51%) based upon increased plasma ALT, and 30/61
(49%) based upon AST/ALT ratio >1.
Journal of Hepatology 2013 vol. 58j543–548
hepatic inflammation . With respect to ultrasound, differenti-
ation between steatosis and NASH is reported as challenging; its
sensitivity diminishes even further in case of less than 40%
hepatic steatosis . In addition, the accuracy of radiological
techniques is operator-dependent and subject to intra- and
inter-observer variability .
The classical way to diagnose NASH, by means of histopathol-
ogical examination of a percutaneous needle biopsy consisting of
a tiny portion of the total liver mass (an estimated 1/50,000), is
vulnerable to variability . To reduce this variability, relatively
large wedge-shaped liver biopsies containing at least five portal
tracts were used in this pilot study, allowing a thorough histolog-
ical evaluation . Occasionally, these biopsies may have been
derived from a relatively more subcapsular region, in which
fibrosis is more often detected than in needle biopsies of the dee-
per parenchyma. However, this emphasizes that analysis of
exhaled breath can identify NASH presence at an early stage,
and early identification in a mild stage is pivotal to enhance the
chances of cure. Furthermore, whereas a small part of the liver
is considered in the evaluation of biopsies, the breath test used
in this study non-invasively reflects total liver function.
Previously, other breath tests have been used to indicate the
extent of liver disease. For example, the so-called caffeine breath
test was reported to predict the presence of hepatic fibrosis in a
study of 48 subjects . VOC analysis in exhaled breath has pre-
viously been used to discriminate between patients with and
without inflammatory airway diseases [13,14]. Based upon find-
ings in these studies, we here investigated the potential of VOC
analysis to diagnose NASH, and to differentiate subjects with
and without NASH in a high-risk population with relatively mild
to moderate NASH, emphasizing the discriminative potential of
exhaled breath. The three volatile breath components n-tridec-
ane, 3-methyl-butanonitrile, and 1-propanol were strongly
related to NASH presence, and the absence of NASH was more
accurately predicted than by use of plasma transaminase levels.
Even though the exact origin of these compounds still needs to
be investigated, it is tempting to speculate that they are related
to inflammatory processes and/or lipid peroxidation. On the
one hand, NASH is characterized by inflammation , and inflam-
mation has also been found to be reflected by 1-propanol in pul-
monary cancer . On the other hand, NASH has been
associated with lipid products [3,38], which have also been
detected in exhaled breath in the context of cardiac surgery .
Further studies will not only focus on the exact origin of these
compounds, but also on validation of these data in an indepen-
dent group of patients. The diagnostic value of this breath test
may be further enhanced by a combination with other plasma
markers (e.g., the previously mentioned epithelial damage mark-
ers cytokeratin-18 and its fragments). Analysis of breath could
aid as a first line screening tool for NASH. In addition, it would
be of interest to test the potential of VOC analysis to evaluate
the effect of treatment of this serious and emerging liver disease.
The current cornerstones of this treatment are lifestyle modifica-
tions  and bariatric surgery, whereas novel pharmacological
therapies are under investigation . Bariatric surgery has been
shown to improve NAFLD in 92% and NASH in 81% of cases, and to
induce complete remission of NASH in 69% of cases . Possibly,
VOC analysis can be used to prioritize patients with NASH over
those without NASH to undergo bariatric surgery. In conclusion,
this pilot study indicates that analysis of exhaled breath can
become a suitable non-invasive method to accurately predict
which subjects suffer from NASH. This method has the potential
to diminish the number of missed diagnoses of NASH.
This research was supported by the Senter Novem Innovation
Oriented Research Program on Genomics, grant IGE05012 and a
Transnational University Limburg (TUL) grant.
Conflict of interest
The authors who have taken part in this study declared that they
do not have anything to disclose regarding funding or conflict of
interest with respect to this manuscript.
The authors thank Yanti Slaats for her contribution to the sample
collection at the start of this study and Jelena Arsenijevic for her
assistance in the statistical analyses.
 Angulo P. Obesity and nonalcoholic fatty liver disease. Nutr Rev 2007;
 Vernon G, Baranova A, Younossi ZM. Systematic review: the epidemiology
and natural history of non-alcoholic fatty liver disease and non-alcoholic
steatohepatitis in adults. Aliment Pharmacol Ther 2011;34:274–285.
 Brunt EM. Pathology of fatty liver disease. Mod Pathol 2007;20 (Suppl
 Tilg H, Moschen AR. Evolution of inflammation in nonalcoholic fatty liver
disease: the multiple parallel hits hypothesis. Hepatology 2010;52:
 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;94:2467–2474.
 Kleiner DE, Brunt EM, Van Natta M, Behling C, Contos MJ, Cummings OW,
et al. Design and validation of a histological scoring system for nonalcoholic
fatty liver disease. Hepatology 2005;41:1313–1321.
Table 4. Multivariate analysis.
⁄Only VOC shows statistical significance in multivariate analysis (p <0.05).
The b-coefficient reflects the importance of the parameter potentially predicting
NASH presence, whereas the Wald value is used to test the true value of this
importance; whether it is statistically significant. Apart from the volatile organic
compounds, none of these parameters reached statistical significance; therefore
none of them facilitate the diagnosis of NASH.
JOURNAL OF HEPATOLOGY
Journal of Hepatology 2013 vol. 58j543–548
 Bravo AA, Sheth SG, Chopra S. Liver biopsy. N Engl J Med 2001;344:495–500. Download full-text
 Piccinino F, Sagnelli E, Pasquale G, Giusti G. Complications following
percutaneous liver biopsy. A multicentre retrospective study on 68,276
biopsies. J Hepatol 1986;2:165–173.
 Junior WS, Nonino-Borges CB. Clinical predictors of different grades of
nonalcoholic fatty liver disease. Obes Surg 2012;22:248–252.
 Neuschwander-Tetri BA, Clark JM, Bass NM, Van Natta ML, Unalp-Arida A,
Tonascia J, et al. Clinical, laboratory and histological associations in adults
with nonalcoholic fatty liver disease. Hepatology 2010;52:913–924.
 Mofrad P, Contos MJ, Haque M, Sargeant C, Fisher RA, Luketic VA, et al.
Clinical and histologic spectrum of nonalcoholic fatty liver disease associ-
ated with normal ALT values. Hepatology 2003;37:1286–1292.
 Paredi P, Kharitonov SA, Barnes PJ. Analysis of expired air for oxidation
products. Am J Respir Crit Care Med 2002;166:S31–S37.
 Dallinga JW, Robroeks CM, van Berkel JJ, Moonen EJ, Godschalk RW, Jobsis Q,
et al. Volatile organic compounds in exhaled breath as a diagnostic tool for
asthma in children. Clin Exp Allergy 2010;40:68–76.
 Van Berkel JJ, Dallinga JW, Moller GM, Godschalk RW, Moonen EJ, Wouters
EF, et al. A profile of volatile organic compounds in breath discriminates
COPD patients from controls. Respir Med 2010;104:557–563.
 Millonig G, Praun S, Netzer M, Baumgartner C, Dornauer A, Mueller S, et al.
Non-invasive diagnosis of liver diseases by breath analysis using an
optimized ion-molecule reaction-mass spectrometry approach: a pilot
study. Biomarkers 2010;15:297–306.
 Netzer M, Millonig G, Osl M, Pfeifer B, Praun S, Villinger J, et al. A new
ensemble-based algorithm for identifying breath gas marker candidates in
liver disease using ion molecule reaction mass spectrometry. Bioinformatics
 Solga SF, Alkhuraishe A, Cope K, Tabesh A, Clark JM, Torbenson M, et al.
Breath biomarkers and non-alcoholic fatty liver disease: preliminary
observations. Biomarkers 2006;11:174–183.
 Van Berkel JJ, Dallinga JW, Moller GM, Godschalk RW, Moonen E, Wouters
EF, et al. Development of accurate classification method based on the
analysis of volatile organic compounds from human exhaled air. J Chroma-
togr B Analyt Technol Biomed Life Sci 2008;861:101–107.
 Verdam FJ, Greve JW, Roosta S, van Eijk H, Bouvy N, Buurman WA, et al.
Small intestinal alterations in severely obese hyperglycemic subjects. J Clin
Endocrinol Metab 2011;96:E379–E383.
 Dufour DR, Lott JA, Nolte FS, Gretch DR, Koff RS, Seeff LB. Diagnosis and
monitoring of hepatic injury. I. Performance characteristics of laboratory
tests. Clin chem 2000;46:2027–2049.
 Pratt DS, Kaplan MM. Evaluation of abnormal liver-enzyme results in
asymptomatic patients. N Engl J Med 2000;342:1266–1271.
 Verdam FJ, Rensen SS, Driessen A, Greve JW, Buurman WA. Novel evidence
for chronic exposure to endotoxin in human nonalcoholic steatohepatitis. J
Clin Gastroenterol 2011;45:149–152.
 Bossuyt X. Clinical performance characteristics of a laboratory test. A
practical approach in the autoimmune laboratory. Autoimmun Rev
 Camp BW. What the clinician really needs to know: questioning the clinical
usefulness of sensitivity and specificity in studies of screening tests. J Dev
Behav Pediatr 2006;27:226–230.
 Adams LA, Lymp JF, St Sauver J, Sanderson SO, Lindor KD, Feldstein A, et al.
The natural history of nonalcoholic fatty liver disease: a population-based
cohort study. Gastroenterology 2005;129:113–121.
 Ekstedt M, Franzen LE, Mathiesen UL, Thorelius L, Holmqvist M, Bodemar G,
et al. Long-term follow-up of patients with NAFLD and elevated liver
enzymes. Hepatology 2006;44:865–873.
 Feldstein AE, Wieckowska A, Lopez AR, Liu YC, Zein NN, McCullough AJ.
Cytokeratin-18 fragment levels as noninvasive biomarkers for nonalcoholic
steatohepatitis: a multicenter validation study. Hepatology 2009;50:
 Joka D, Wahl K, Moeller S, Schlue J, Vaske B, Bahr MJ, et al. Prospective
biopsy-controlled evaluation of cell death biomarkers for prediction of liver
fibrosis and nonalcoholic steatohepatitis. Hepatology 2012;55:455–464.
 Williams CD, Stengel J, Asike MI, Torres DM, Shaw J, Contreras M, et al.
Prevalence of nonalcoholic fatty liver disease and nonalcoholic steatohep-
atitis among a largely middle-aged population utilizing ultrasound and liver
biopsy: a prospective study. Gastroenterology 2011;140:124–131.
 Myers RP, Pomier-Layrargues G, Kirsch R, Pollett A, Duarte-Rojo A, Wong D,
et al. Feasibility and diagnostic performance of the FibroScan XL probe for
liver stiffness measurement in overweight and obese patients. Hepatology
 Chen J, Talwalkar JA, Yin M, Glaser KJ, Sanderson SO, Ehman RL. Early
detection of nonalcoholic steatohepatitis in patients with nonalcoholic fatty
liver disease by using MR elastography. Radiology 2011;259:749–756.
 Mottin CC, Moretto M, Padoin AV, Swarowsky AM, Toneto MG, Glock L, et al.
The role of ultrasound in the diagnosis of hepatic steatosis in morbidly obese
patients. Obes Surg 2004;14:635–637.
 Strauss S, Gavish E, Gottlieb P, Katsnelson L. Interobserver and intraobserver
variability in the sonographic assessment of fatty liver. AJR Am J Roentgenol
 Ratziu V, Charlotte F, Heurtier A, Gombert S, Giral P, Bruckert E, et al.
Sampling variability of liver biopsy in nonalcoholic fatty liver disease.
 Arun J, Jhala N, Lazenby AJ, Clements R, Abrams GA. Influence of liver biopsy
heterogeneity and diagnosis of nonalcoholic steatohepatitis in subjects
undergoing gastric bypass. Obes Surg 2007;17:155–161.
 Park GJ, Wiseman E, George J, Katelaris PH, Seow F, Fung C, et al. Non-
invasive estimation of liver fibrosis in non-alcoholic fatty liver disease using
the (13) C-caffeine breath test. J Gastroenterol Hepatol 2011;26:1411–1416.
 Ligor M, Ligor T, Bajtarevic A, Ager C, Pienz M, Klieber M, et al. Determination
of volatile organic compounds in exhaled breath of patients with lung cancer
using solid phase microextraction and gas chromatography mass spectrom-
etry. Clin Chem Lab Med 2009;47:550–560.
 Seki S, Kitada T, Yamada T, Sakaguchi H, Nakatani K, Wakasa K. In situ
detection of lipid peroxidation and oxidative DNA damage in non-alcoholic
fatty liver diseases. J Hepatol 2002;37:56–62.
 Pabst F, Miekisch W, Fuchs P, Kischkel S, Schubert JK. Monitoring of
oxidative and metabolic stress during cardiac surgery by means of breath
biomarkers: an observational study. J Cardiothorac Surg 2007;2:37.
 Lebovics E, Rubin J. Non-alcoholic fatty liver disease (NAFLD): why you
should care, when you should worry, what you should do. Diabetes Metab
Res Rev 2011;27:419–424.
 Zein CO, Yerian LM, Gogate P, Lopez R, Kirwan JP, Feldstein AE, et al.
Pentoxifylline improves nonalcoholic steatohepatitis: a randomized pla-
cebo-controlled trial. Hepatology 2011;54:1610–1619.
 Mummadi RR, Kasturi KS, Chennareddygari S, Sood GK. Effect of bariatric
surgery on nonalcoholic fatty liver disease: systematic review and meta-
analysis. Clin Gastroenterol Hepatol 2008;6:1396–1402.
Journal of Hepatology 2013 vol. 58j543–548