Content uploaded by Raquel Fernández del Río
Author content
All content in this area was uploaded by Raquel Fernández del Río on Nov 12, 2018
Content may be subject to copyright.
Volatile biomarkers in breath associated with liver
cirrhosis — comparisons of pre- and post-liver
transplant breath samples
Fernandez del Rio, Raquel; O'Hara, Margaret; Holt, Andrew; Pemberton, P.; Shah, T.;
Whitehouse, T.; Mayhew, Christopher
DOI:
10.1016/j.ebiom.2015.07.027
License:
Creative Commons: Attribution (CC BY)
Document Version
Publisher's PDF, also known as Version of record
Citation for published version (Harvard):
Fernández Del Río, R, O'Hara, M, Holt, A, Pemberton, P, Shah, T, Whitehouse, T & Mayhew, C 2015, 'Volatile
biomarkers in breath associated with liver cirrhosis — comparisons of pre- and post-liver transplant breath
samples' EBioMedicine, vol. 2, no. 9, pp. 1243-1250. DOI: 10.1016/j.ebiom.2015.07.027
Link to publication on Research at Birmingham portal
Publisher Rights Statement:
Eligibility for repository : checked 20/10/2015
General rights
Unless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or the
copyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposes
permitted by law.
•Users may freely distribute the URL that is used to identify this publication.
•Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of private
study or non-commercial research.
•User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?)
•Users may not further distribute the material nor use it for the purposes of commercial gain.
Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document.
When citing, please reference the published version.
Take down policy
While the University of Birmingham exercises care and attention in making items available there are rare occasions when an item has been
uploaded in error or has been deemed to be commercially or otherwise sensitive.
If you believe that this is the case for this document, please contact UBIRA@lists.bham.ac.uk providing details and we will remove access to
the work immediately and investigate.
Download date: 17. Oct. 2018
Research Paper
Volatile Biomarkers in Breath Associated With Liver Cirrhosis —
Comparisons of Pre- and Post-liver Transplant Breath Samples
R. Fernández del Río
a
,M.E.O'Hara
a,
⁎,A.Holt
b
, P. Pemberton
c
, T. Shah
b
, T. Whitehouse
c
,C.A.Mayhew
a
a
School of Physics and Astronomy, University of Birmingham, Birmingham B15 2TT, UK
b
Department of Hepatology, University HospitalBirmingham NHS Trust, Birmingham B15 2TH, UK
c
Critical Care and Anaesthesia, University Hospital Birmingham NHS Trust, Birmingham B15 2TH, UK
abstractarticle info
Article history:
Received 9 June 2015
Received in revised form 17 July 2015
Accepted 20 July 2015
Available online 26 July 2015
Keywords:
Breath analysis
Cirrhosis
Diagnosis limonene
Liver transplant
PTR-MS
Volatile organic compounds
Background: The burden of liver disease in the UK has risen dramatically and there is a need for improved diagnostics.
Aims: To determine which breath volatiles are associated with the cirrhotic liver and hence diagnostically useful.
Methods: A two-stage biomarker discovery procedure was used. Alveolar breath samples of 31 patients with cirrhosis
and 30 healthy controls were mass spectrometrically analysed and compared (stage 1). 12 of these patients had their
breath analysed after liver transplant (stage 2). Five patients were followed longitudinally as in-patients in the post-
transplant period.
Results: Seven volatiles were elevated in the breath of patients versus controls. Of these, five showed statistically signif-
icant decrease post-transplant: limonene, methanol,2-pentanone,2-butanoneandcarbondisulfide. On an individual
basis limonene has the best diagnostic capability (the area under a receiver operating characteristic curve (AUROC) is
0.91), but this is improved by combining methanol, 2-pentanone and limonene (AUROC curve 0.95). Following trans-
plant, limonene shows wash-out characteristics.
Conclusions: Limonene, methanol and 2-pentanone are breath markers for a cirrhotic liver. This study raises the poten-
tial to investigate these volatiles as markers for early-stage liver disease. By monitoring the wash-out of limonene fol-
lowing transplant, graft liver function can be non-invasively assessed.
© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
1. Introduction
The publication of the 2014 Lancet Commission on liver disease has
highlighted how the burden of liver disease in the UK has risen sharply
over the past few decades and that it poses a major public health issue
(Williams et al., 2014). It is the only major cause of mortality and mor-
bidity which is on the increase in England, while at the same time de-
creasing in most other European countries, with cirrhosis accounting
for 83% of deaths (Davies, 2012). It is the third biggest cause of prema-
ture mortality, with three quarters of liver deaths due to alcohol
(Williams et al., 2014). Liver disease has a widespread effect not only
to the patient, encompassing physical and psychological morbidity
and mortality, but also incurring significant societal costs. One of the
main difficulties is that often patients do not present symptoms or
signs until the disease is advanced. Even then diagnosis is difficult and
the symptoms and signs are often general and can be mistaken for
other pathologies. Non-invasive diagnostic techniques currently used,
namely serum biomarkers and transient elastography (TE) are not
ideal. Serum biomarkers are not liver specific and TE results require an
expert clinician for interpretation (Castera et al., 2015).
Among the ten key recommendations in a recent Lancet report is to
strengthen the detection of early-stage liver disease, which is essential
to reduce disease progression (Williams et al., 2014). Analysis of vola-
tiles in the breath has the potential to deliver this, but only if chemical
compounds can be found that are unambiguously associated with a
diseased liver.
To date, the use of breath volatiles for medical diagnosis has met
with limited success. Confounding factors, such as volatiles present in
the environment, contamination in the sampling procedures and poor
sampling methods, have meant that there is a great deal of uncertainty
in volatile discovery (Kwak and Preti, 2011). Problems of bias and false
EBioMedicine 2 (2015) 1243–1250
Abbreviations: AID, autoimmune liver disease; ALD, alcoholic liver disease; AUROC,
area under receiver operator curve; CD, cryptogenic disease; GC, gas chromatography;
HBV, hepatitis B virus; HCC, hepatocellular cancer; HCV, hepatitis C virus; ITU, intensive
treatment unit; LQ, lower quartile; MS, mass spectrometry; OPU, out-patient clinic; PBC,
primary biliary cirrhosis; PSC, primary sclerosing cholangitis; ppbv, parts per billion by
volume; ppmv, parts per million by volume; PTR-MS, proton transfer reaction mass spec-
trometry; ROC, Receiver operating characteristics; TAC, transplant assessment clinic; TE,
transient elastography; UKELD, United Kingdom model for end-stage liver disease; UQ,
upper quartile; VOC, volatile organic compounds; VMR, volume mixing ratio; BMI, body
mass index.
⁎Corresponding author.
E-mail addresses: r.fernandezdelrio@bham.ac.uk (R. Fernández del Río),
M.E.OHara@bham.ac.uk (M.E. O'Hara), Andrew.holt@uhb.nhs.uk (A. Holt),
pembo@doctors.ork (P. Pemberton), Tahir.Shah@bham.ac.uk (T. Shah),
TonyWhitehouse@uhb.nhs.uk (T. Whitehouse), c.mayhew@bham.ac.uk (C.A. Mayhew).
http://dx.doi.org/10.1016/j.ebiom.2015.07.027
2352-3964/© 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Contents lists available at ScienceDirect
EBioMedicine
journal homepage: www.ebiomedicine.com
discovery in biomarker discovery research have been widely reviewed
(Broadhurst and Kell, 2006; Ransohoff, 2005).
Previous studies investigating breath volatiles in patients suffering
with liver disease have proposed a large number of possible biomarkers
(Millonig et al., 2010; Morisco et al., 2013; Sehnert et al., 2002; Solga
et al., 2006; Tangerman et al., 1983; Van den Velde et al., 2008;
Dadamio et al., 2012; Friedman et al., 1994; Hanouneh et al., 2014;
Khalid et al., 2013, 2014; Shimamoto et al., 2000; Vollnberg et al.,
2009; Verdam et al., 2013), but generally different studies report
different volatiles. Various GC–MS studies found raised levels of many
volatiles in the breath of patients with liver disease, including dimethyl
sulfide, acetone, 2-butanone, 2-pentanone, β-pinene, α-pinene,
and limonene (Van den Velde et al., 2008; Dadamio et al., 2012;
Friedman et al., 1994; Khalid et al., 2013). Studies using soft chemical
ionization mass spectrometric techniques have reported volatiles such
as acetaldehyde, ethanol, isoprene, benzene, methanol, 2-butanone,
2- or 3-pentanone, heptadienol, and a monoterpene (limonene)
(Moriscoet al., 2013). Although theresults of these studies are extreme-
ly encouraging, few volatile organic compounds (VOCs) are common to
more than two or three studies and it is not useful to have hundreds of
putative markers. Furthermore some volatiles, which have been
proposed as biomarkers for liver disease, such as isoprene, acetone
and ethanol, are not specific enough because they are possible bio-
markers for other diseases or arise from numerous normal metabolic
processes. If breath analysis is to progress to clinical utility, then
markers must be definitively associated with the disease in question.
All previous studies can be regarded as hypothesis-generating, in
that they do not follow up in a second group to confirm the putative
biomarkers. We report here a two-stage breath biomarker discovery
process: breath samples from a group of patients suffering from liver
disease are first compared to breath samples from healthy controls;
post-transplant breath samples are then compared with a sub-cohort
of these patients who went on to have a liver transplant. A set of
putative volatile markers is first determined by comparing patients
with controls, and then pre- and post-transplant breath samples are
examined to look for intra-individual differences in these volatiles. In
this way, this study is hypothesis-led and uses patients as their own
controls, thereby reducing the risk of false discovery. Furthermore, the
use of patients' companions as controls and of room air samples
minimizes the influence of any exogenous volatiles present in the
home and hospital as confounding factors.
2. Methods
2.1. Patients, Controls and Hospital Room Air
Patients were recruited at the University Hospital Birmingham from
either the transplant assessment clinic or in wards after being admitted
with hepatic encephalopathy. 31 patients suffering from liver disease
participated in the pre-transplant measurements (F/M 8/23, mean age
55 years, min–max 27–71 years). There were a number of etiologies
and 11 patients had more than one condition: alcoholic liver disease
(N = 13), hepatocellular cancer (N = 10), cryptogenic (N = 4), hepati-
tis C (N = 5), primary sclerosing cholangitis (N= 4), primary biliary cir-
rhosis (N = 2), autoimmune liver disease (N = 1), hepatitis B (N = 3),
non-alcoholic steatohepatitis (N = 1), and non-alcoholic fatty liver dis-
ease (N = 1). Of these 31 patients, 12 went on to have a liver transplant
(F/M 4/8, meanage 48 years, min–max 27–58 years). Oneadditional pa-
tient (F, age 53 years) was recruited into the post-transplant study.
Table 1 summarises the details of the 13 patients who had a liver
transplant (12 from the pre-group), with patients being identified by
sex (F or M) and a number. In addition to pre-transplant diagnostics, de-
finitive diagnoses by histopathological examination of the explanted
liver are provided. All but one (F2) were diagnosed with cirrhosis by
standard liver function tests and biopsy. F2 was admitted suffering
with hepatic encephalopathy, and histopathology of the explanted
liver gave a diagnosis of severe hepatitis with multiacinar necrosis.
Supplementary Table 1 shows demographic information for all
patients including medications they were taking at the time of the
pre-transplant sample.
For 28 pre-transplant measurements, breath samples from the
patients' companions were taken. For the other three, two came alone
to clinic andthe other's companion declined to take part. Two additional
controls were therefore recruited, one was a ward nurse and the other
was a visitor to the hospital in ITU. These controls, while not related to
the patient, had been in the same room for several hours prior to sam-
pling so that confounding factors associated with volatiles present in
the room environment were taken into consideration. In total 30 con-
trols (F:M 23:7, mean age 44 years, min–max 20–75 years) took part
in the study. The larger number of females in the control group arose
due to the tendency of men coming to the clinic accompanied with
their wives. While this means the control group is not ideally matched,
there is no consistent evidence of dependences of volatile breath
composition on sex (Kwak and Preti, 2011; Ellis and Mayhew, 2014).
In confirmation of this we also found no correlation between sex and
VOCs in either our control or patient groups. We consider that inhaled
VOCs have a greater potential to confound biomarker discovery. As
the majority of the companions were livingwith the patients, they pro-
vided an ideal control for exposure to exogenous volatiles in the home
environment. VOCs inhaled at home, or in transit, may well still be
present in breath for hours or days after inhalation and the biological
half-life of inhaled VOCs is not well known (Beauchamp, 2011).
All study subjects were asked to complete a detailed questionnaire
which included details on their home environment, diet, smoking
status, health and medications. Participants were asked if they had con-
sumedfruitandfruitjuicesandfruitflavoured drinksas a normal partof
their diet, and, if so, to provide details on quantity and how long before
the breath sampling these had been consumed.
Hospital room air was collected every time breath samples were
taken so that any exogenous volatiles, such as isopropanol coming
from hand gels resulting in product ions at m/z 43 and m/z 61, could
be taken into consideration.
2.2. Breath Sampling Protocol
There is no agreed standard for the collection of breath for volatile
analysis and uncontrolled breath sampling has been shown to be
unreliable (Schubert et al., 2001; O'Hara et al., 2008). Therefore,
capnography controlled sampling was used to collect only the alveolar
phase of the breath. Subjects were in a relaxed state throughout the
measurements and were either in a seated or lying position. They
were asked to breathe normally into a gas tight respiratory system
(Intersurgical Limited) containing an in-line CO
2
mainstream sensor
connected to a fast-time response capnometer (Capnogard 1265
Novametrix Medical Systems Inc.). A 100 ml glass syringe (Sigma-
Aldrich) was coupled to the tubing using a 3-way luer-lock stopcock
(Braun Medical Limited). When the alveolar plateau on the capnograph
was observed, a breath sample was manually drawn from the subject's
breath stream into the syringe. Three to four breaths samples were col-
lected for each 100 ml syringe, and four replicates of these were taken
for each subject. Glass syringes were used, because our tests showed
that they have no contaminating volatiles. Fig. 1 schematically shows
the sampling system used.
After collection, the syringes were sealed using the luer lock
fitting. They were transported from hospital to laboratory (a 10 minute
outdoor walk) in an opaque storage box. Once at the laboratory, the
syringes were placed inside an incubator set at 40 °C.
All samples were mass spectrometrically analysed within 2 h of col-
lection. For the measurements, syringes were taken out of theincubator
and immediately placed into a purpose designed heating bag (Infroheat,
Wolverhampton) maintained at a constant temperature of 40 °C in
order to limit condensation, which could otherwise lead to volatile
1244 R. Fernández del Río et al. / EBioMedicine 2 (2015) 1243–1250
loss (Beauchamp et al., 2008). The luer stopcock was coupled to a
Swagelok fitting and connected directly to the inlet of the analytical de-
vice, a proton transfer reaction mass spectrometer (PTR-MS). The inlet
flow was set at 10–15 ml/min and the drift tube and inlet lines were
maintained at 45 °C. The syringes are gas tightand have minimal friction
such that atmospheric pressure is sufficient to push the plunger in
smoothly so that the breath sample is being drawn into the instrument
at a constant flow.
2.3. Analytical Measurements
PTR-MS is a platform technology designed to detect low concentra-
tions of volatiles (less than parts per billion by volume). Hence it has
found use in many analytical applications ranging from drug detection
through to industrial pollution (Ellis and Mayhew, 2014; Lindinger
et al., 1998; Agarwal et al., 2011; de Gouw et al., 2011; Jurschik et al.,
2012). Details of the instrument used, a PTR-Quad-MS (IONICON
Analytik GmbH), and how it operates are described in detail in the
literature (Ellis and Mayhew, 2014; O'Hara et al., 2008; Lindinger
et al., 1998). In brief, it exploits the reactions of protonated water with
neutral volatiles (M), usually leading to a protonated parent (MH
+
). If
dissociative proton transfer occurs then it is not extensive in terms of
the number of resulting product ions. Operational parameters used
for this investigation were those previously reported (O'Hara et al.,
2009a,b). Namely, the drift-tube was maintained at a pressure of
2.07 ± 0.01 mbar and temperature of 45 ± 1 °C. The voltage across
the drift-tube was set at 600 V, which is sufficiently high to reduce
water clustering to reagent and product ions by collision induced
dissociation.
Am/z range of 20 to 200 amu was scanned with a dwell time of 0.5 s
per atomic mass unit. Mass spectra of the breath samples wererecorded
from the average of three cycles for each of the four syringes, for every
participant. These four spectra were averaged to provide one data set
for each subject with the uncertainty expressed as the standard error
of the mean for the four syringes.
The intensities of the product ion(s) associated with a given volatile
were converted to volume mixing ratios (VMR) in unitsof nmol/mol by
use of a standard procedure that relies on a calculated, compound-
specific, collisional reaction rate coefficient, determined using the effec-
tive translational temperature of the reagent ions (Ellis and Mayhew,
2014).
To help identify product ions, pure samples of key volatiles were
individually measured using PTR-MS to establish the m/z values of the
product ions.
2.4. Data Analysis and Statistics
Room air contamination is a potential confounding factor in breath
analysis and has been the subject of much discussion. If the intensity of
aproductionhasasignificant contribution coming from room air, then
care must be taken when using it as a biomarker. There is no simple cor-
rection which can be applied to account for inhaled volatile concentra-
tions, and it has been shown that simply subtracting the room air
concentration is too simplistic (Schubert et al., 2005; Spanel et al., 2013;
Pleil et al., 2013). For this study, a filter was applied such that only ion sig-
nal intensities in the breath sample that were at least twice that in the
room air samples in at least half of the patients were retained for analysis.
This resulted in a set of 40 product ions for analysis. (The m/z values and
normalised counts per second are provided in Supplementary Table 2,
many of which are well known including m/z 33 (methanol), m/z 45 (ac-
etaldehyde), m/z 47 (ethanol), m/z 59 (acetone) and m/z 69 (isoprene).
The data sets for each volatile of interest were assessed using a
Shapiro–Wilks test and were found not to be normally distributed so
non-parametric tests were used. IBM SPSS version 22 was used for all
statistical analysis. Mann–Whitney U-tests determined which m/z
values differed between the patients and controls. A Wilcoxon signed
Table 1
Liver transplant patient details, including sex (female F, male M), age, initial diagnosis,histopathological results, location of pre-transplant and post-transplant breath sampling, and the number of days prior to andafter transplant when breath sam-
ples were collected. Diseases include autoimmune liver disease (AID), alcoholic liver disease (ALD), cryptogenic disease (CD), hepatitis B (HBV), hepatitis C (HCV), hepatocellular cancer (HCC), primary biliary cirrhosis (PBC) and primary sclerosing
cholangitis (PSC). Breath samples were taken at various locations including the intensive treatment unit (ITU), out-patient clinic (OPC), transplant assessment clinic (TAC), and in wards.
Patient
ID
Age
(yr)
Initial
diagnosis
Histopathological results Location of pre-transplant
breath sample
Pre-transplant breath
sample: days before
transplant
Location of post-transplant
breath sample
Post-transplant breath
samples: days after
transplant
F1 27 AID Severe steatohepatitis (PSC) TAC 54 OPC 65
F2 49 Liver Failure Severe hepatitis with multiacinar necrosis, seronegative hepatitis ITU 0
1
OPC 3, 5, 130
F3 53 PBC Cirrhosis (PBC) TAC 74 OPC 45
F4 58 PSC Cirrhosis (PSC) TAC 83 Ward 5–8, 11–15, 18, 58
F5 53 ALD Hepatocellular carcinoma, liver cirrhosis.(ALD, HCV) ––Ward 2–6, 9–12
M1 54 ALD Severe steatohepatitis Ward 47 OPC 33
M2 45 ALD Cirrhosis (ALD) TAC 97 OPC 22
M3 53 ALD Cirrhosis (ALD/HCV)
Hepatocellular carcinoma (grade2)
TAC 179 Ward 4, 7, 48
M4 53 ALD, HBV, HCV Cirrhosis (ALD, HCV, HBV) TAC 21 OPC 126
M5 56 ALD, HCV, HCC Cirrhosis (ALD, HCV, HCC) TAC 125 OPC 61
M6 53 CD Cirrhosis with mild steatohepatitis
(aetiology possibly NASH but uncertain)
TAC 154 OPC 22
M7 36 CD Cirrhosis of uncertain aetiology TAC 180 Ward 2, 3, 6–8, 55
M8 67 ALD Cirrhosis (ALD) ITU 14 Ward 6
The pre-transplant breath sample for patient F2 was taken approximately 10 min before the patient went into surgery. This patient was admitted with liver failure and hepatic encephalopathy.
1245R. Fernández del Río et al. / EBioMedicine 2 (2015) 1243–1250
rank test was used to determine which volatile concentrations differed
between pre- and post-transplant breath samples. To compare blood
chemistry values with breath volatiles Kendall's tau-b correlation
coefficients were measured. Receiver Operating Characteristic (ROC)
curves were used to determine the diagnostic accuracy of volatiles.
3. Results
For the first stage of our study, a Mann–Whitney U-test with a
significance level of 95% was used to compare the 40 product ion signal
intensities used in our analysis between patients and controls. Of these,
eight showed significant differences in intensities between patients and
controls. Their m/z values and significance (p value in brackets) are 33
(b0.001), 73 (0.004), 77 (0.035), 81 (b0.001), 87 (b0.001), 89 (0.03),
135 (0.019) and 137 (b0.001). In the second phase, pre- and post-
transplant intensities of these eight ions were compared using a
Wilcoxon Signed Rank Test for paired samples with a significance of
95%. This eliminated m/z 89 and 135 from the putative marker set. Of
the remaining ions, m/z 33 is assigned to be protonated methanol
(CH
3
OH
2
+
)(Lindinger et al., 1997). Based on previous GC and GC–MS
studies (Sehnert et al., 2002; Van den Velde et al., 2008; Dadamio
et al., 2012), we tentatively identify m/z 73 as protonated 2-butanone
(C
4
H
8
OH
+
), m/z 77 as protonated carbon disulfide (CS
2
H
+
), m/z 87 as
protonated 2-pentanone (C
5
H
10
OH
+
), and m/z 81 and 137
as limonene. (m/z 81 is a fragment ion (C
6
H
9
+
) resulting from
dissociative proton transfer and m/z 137 is protonated limonene
(C
10
H
17
+
).) VMRs of these volatiles in room air, patient and control sam-
ples are shown in Fig. 2 for (a) methanol, (b) carbon disulfide, (c) 2-
butanone, (d) 2-pentanone, and (e) limonene (sum of the intensities
of the m/z 81 and m/z 137 product ions). The median, mean, lower quar-
tile (LQ), and upper quartile (UQ) of the VMRs in units of nmol/mol for
each volatile are shown. It is clear from Fig. 2 that the presence of the
volatiles in room air has a negligible effect on the concentrations in
the breath samples. Furthermore, the analysis is based on comparisons
and the patients and controls should be affected similarly by the room
air contaminations.
The pre-transplant and post-transplant VMRs for methanol, carbon
disulfide, 2-butanone, 2-pentanone and limonene for all of our
participants who underwent liver transplants (4 females (F1–F4) and
8males(M1–M8)) are provided in Table 2. It should be noted that
the number of days between collecting the pre-transplant and post-
transplant breath samples is variable, because it is not possible to
control when subjects are available or when a donor liver would be
found. Only for one of the patients, F2, were we able to collect a pre-
transplant breath sample just prior to surgery. However, and indepen-
dent of when the pre- and post-transplant breath samples were taken,
the results in Table 2 clearly demonstrate that the pre-transplant
concentrations of these volatiles are, for the majority of patients, higher
than the post-transplant levels for most patients. Limonene shows
the largest average decrease and also decreased in all patients post-
transplant. Post-transplant concentrations of limonene dropped to
within the normal control range (median (LQ, UQ) being 2.3 nmol/mol
(1.9, 3.0)) within a number of days for all but one of the patients, M4,
for whom limonene was found to be high even some months after
transplant.
In order to gain an insight on how the methanol, carbon disulfide,
2-butanone, 2-pentanone and limonene breath VMRs changed over a
period of timeafter transplant, five patients (F2, F4, F5, M3 and M7) par-
ticipated in a longitudinal study. The key result is that limonene VMRs
dropped gradually following transplant surgery, as illustrated in Fig. 3.
(The same data for limonene presented as normalised to the highest
intra-individual value are shown in Supplementary Fig. 1.) This concen-
tration time dependence was not observed for methanol, carbon
disulfide, 2-butanone and 2-pentanone. Their VMRs were found to
have dropped to within the normal range by the time of the first post-
transplant measurement.
Taken together, the box plots (Fig. 2), the ratio of the pre- and post-
transplant VMR values and the significance values given above imply
that ions at m/z 33, 81, 87 and 137 are the ones that are most diagnosti-
cally useful. This is confirmed by ROC curve analyses. Individually,
limonene is found to provide the most predictive power (AUROC —
0.91 (standard error 0.04)). However, the best accuracy is achieved by
combining the data from methanol, 2-pentanone and limonene. The
VMRs for limonene, methanol and 2-pentanone were normalised to
the highest patient value for that volatile. These normalised fractions
were simply added with no weightings. Fig. 4 shows a ROC curve
for the combined data. The AUROC is 0.95 (standard error 0.03) and
achieves a sensitivity of 97% with a specificity of 70%.
Clinical chemistry data for the patients for whom blood data were
available were analysed for possible correlations with limonene, metha-
nol and 2-pentanone. Correlations were checked for alanine aminotrans-
ferase, alkaline phosphatase, aspartate transferase, albumin, total
bilirubin, creatinine, neutrophils, platelets, potassium, prothrombin/
international normalised ratio, and the United Kingdom Model for End-
Stage Liver Disease (UKELD). Kendall's tau-b analysis showed only one
correlation with a significance score below 0.05. This was for methanol
with UKELD which had a Kendall's tau-b coefficient of 0.237 (significance
0.042). Over 33 correlations were tested and no multiple testing correc-
tionwasappliedsoitispossiblethatthisisacoincidentalfinding.
Volatile concentrations were examined for correlations with disease
etiology. Owing to the small sample size and large number of etiologies,
this was only feasible for the 13 patients with ALD versus the other 18
patients. Limonene was higher (p = 0.020) in the ALD group than the
rest, with median (LQ, UQ) of 19.7 nmol/mol (9.2, 63.9) for ALD versus
6.1 nmol/mol (2.9, 16.6) for all other etiologies. Methanol, 2-pentanone,
2-butanone and carbon disulfide showed no statistically significant
difference.
Correlations between the 7 volatiles of interest in the putative marker
set were examined both within the patient and the control group using a
Kendall's tau-b test. In the patient group, there were 8 correlations with a
significance score of b0.05. Results for all 21 correlations are shown in
Supplementary Table 3. In the control group, only two were significant,
limonene with m/z 135 (p = 0.016) and 2-butanone with carbon disul-
fide (p b0.001). These were also found in the patient group. Limonene
correlated significantly with 2-butano ne (p = 0.004), carbon
Fig. 1. Schematic of the breath sampling device. Breath samples are only drawn into the
glass syringe once the capnograph shows that the alveolar phase of the exhaled breath
has been reached. Typically 3–4 breaths are needed to fill a syringe to 100 ml.
Fig. 2. Boxplots showing in units of nmol/mol lower quartile (LQ), median, mean and upper quartile (UQ) calculated volume mixing ratios (VMRs) for (a) methanol, (b) 2-butanone,
(c) carbon d isulfide, (d) 2-pentanone, and (e)limonene for 31 patients with livercirrhosis, 30 controls and room air samples. Whiskers are 1.5times the inter-quartile range and outliers
are depicted by a star.
1246 R. Fernández del Río et al. / EBioMedicine 2 (2015) 1243–1250
Patients Controls Room Air
0
20
40
60
80
100
120
140
160
180
Median = 1.1
LQ = 0.8
UQ = 1.7
Median = 2.3
LQ = 1.9
UQ = 3.0
Median = 13.6
LQ = 3.9
UQ = 47.0
Limonene VMR (nmol/mol)
(e) Limonene
Patients Controls Room Air
0
100
200
300
400
500
600
700
800
900
Median = 25
LQ = 21
UQ = 39
Median = 147
LQ = 114
UQ = 176
Median = 200
LQ = 162
UQ = 404
Meth ano l VMR (nmo l/mo l)
(a) Methanol
Patients Controls Room Air
0
10
20
30
40
50
60
Median = 6.6
LQ = 4.3
UQ = 8.9
Median = 15.9
LQ = 12.9
UQ = 22.5
Median = 21.1
LQ = 16.6
UQ = 31.0
2-butanone VMR (nmol/mol)
(b) 2-butanone
Patients Controls Room air
0
5
10
15
20
Median = 1.0
LQ = 0.7
UQ = 1.4
Median = 4.1
LQ = 3.1
UQ = 6.9
Median = 6.1
LQ = 4.3
UQ = 9.2
)lom/lomn(RMVediflusidnobraC
(c) Carbon disulfide
Patients Controls Room Air
0
10
20
30
40
50
60
Median = 1.4
LQ = 1.0
UQ = 2.6
Median = 6.8
LQ = 5.8
UQ = 7.9
Median = 12.9
LQ = 9.8
UQ = 20.9
2-pentanone VMR (nmol/mol)
(d) 2-pentanone
1247R. Fernández del Río et al. / EBioMedicine 2 (2015) 1243–1250
disulfide (p = 0.034), m/z 89 (p = 0.001) and m/z 135 (p b0.001)
but not with methanol or 2-pentanone. This suggests that the mecha-
nisms for the presence of limonene and that of methanol and 2-
pentanone are independent. 21 correlations were examined with no
multiple-testing correction applied, therefore some correlations may
be coincidental.
Correlations between the concentrations of volatiles and demo-
graphic markers such as age, BMI and sex were also checked. No
significant correlations were found.
4. Discussion
A major aim of this study is to determine the viability of breath anal-
ysis as a non-invasive technique for monitoring/diagnosing liver disease
by identifying volatiles in the breath which are a consequence of the
disease. In this investigation, the monitoring of volatiles in breath
following a dramatic change in the condition of the patient, namely a
liver transplant, has provided a method to attribute three diagnostically
useful volatiles to the cirrhotic organ itself. These are methanol,
2-pentanone, and limonene, with that of limonene being the most
significant.
Limonene has been found in previous breath volatile studies to be
elevated in the breath of patients with cirrhosis compared with controls
(Morisco et al., 2013; Dadamio et al., 2012; Friedman et al., 1994). It has
also been observed in the breath of healthy volunteers; limonene levels
found in our control group are comparable to those previously observed
in healthy human volunteers (Mochalski et al., 2013). Limonene is not
produced in the human body. It is a common compound naturally
found in many foods and drinks; hence it would be difficult to avoid
ingesting. Within the control and patient groups, we found no association
between breath limonene and diet and no correlation between having a
self-reported large amount of fruit consumption and breath limonene
concentrations. Once in the blood stream, limonene is metabolised by
the P450 enzymes CYP2C9 and CYP2C19 to the metabolites perillyl alco-
hol, trans-carveol and trans-isopiperitenol (Miyazawa et al., 2002). It has
been found that levels of the enzyme CYP2C19 are reduced in patients
with cirrhosis and that levels inversely correlate with severity of cirrhosis.
Moreover, of four P450 enzymes tested in patients with liver disease, me-
tabolism by CYP2C19 was found to decrease at the earliest stage of dis-
ease (Frye et al., 2006). This is suggestive that the observed raised
concentrations of limonene in breath arise from the inability of a cirrhotic
liver to produce the appropriate metabolic enzyme (Morisco et al., 2013).
Patient M4 is anomalous in this respect, as his breath limonene concen-
trations do not drop to the normal range post-transplant. Although his
graft liver function blood tests were found to be normal, our results sug-
gest that this patient's new liver is not producing sufficient enzyme to
fully metabolise limonene.
Owing to its lipophilic properties, we propose that limonene which
is not metabolised by the liver accumulates in the fat of patients suffer-
ing from liver disease. Limonene has a blood/air partition coefficient of
36 and an olive oil/blood partition coefficient of 140 (Falk et al., 1990).
Assuming that the olive oil/blood partition coefficient is close to a
body fat/blood partition coefficient, a breath concentration of 1 part
per billion by volume (ppbv) would translate to a fat concentration of
approximately 5 parts per million by volume (ppmv). Our highest re-
corded breath VMR is 170 nmol/mol which implies a concentration in
fat of the order of 850 ppmv. A study involving women with early-
stage breast cancer taking a high oral dose of limonene (2 g/daily for
2–6 weeks before surgery) found that mean limonene concentration
in breast tissue was 41.3 ± 49.9 μg/g which is much higher than that
found in a control group (0.08 ± 0.13 μg/g) (Miller et al., 2013). Breast
tissue is primarily composed of fat (Boston et al., 2005). This supports
our hypothesis that unmetabolised limonene accumulates in fat tissue.
Following transplant, the metabolism of limonene increases, but it
takes time for the limonene to be released from the fat into the blood
stream. This, we propose, explains the observed time dependence on
limonene VMRs in the breath after transplant. A similar wash-out be-
haviour is not observed for methanol and 2-pentanone presumably
owing to their low solubilities in fat (Griffin et al. , 1999; Sangster, 1989).
Some medications are known to be CYP2C9 and CYP2C19
substrates and inhibitors. We therefore looked into the possibility
that medications could affect limonene concentrations. Twenty
patients were taking a CYP2C19 substrate (lansoprazole, omeprazole,
propanalol, esomeprazole), 2 were taking a CYP2C9 substrate
(naproxen, carvedilol), 6 were taking both a CYP2C9 and CYP2C19 sub-
strate and 1 was taking a CYP2C9 inhibitor (sulfamethaoxazole). Our
results show no associations of any medications which are CYP2C9
and CYP2C19 substrates or inhibitors with VMRs of breath limonene.
Moreover, nine patients who were taking enzyme substrates before
transplant were still taking them after transplant.
Of interest is the correlation between limonene and m/z 135,
because this product ion may come from perillyl alcohol (C
10
H
16
O), a
metabolite of limonene. Studies by us (unpublished) have shown that
the reaction of H
3
O
+
with perillyl alcohol leads to a dominant product
ion C
10
H
15
+
resulting from dehydration of the protonated parent. Dehy-
dration following protonation is a common reaction process observed
with many alcohols (Brown et al., 2010). Morisco et al.(2013) also
noted an ion at m/z 135 from patients with cirrhosis, but they assigned
this to a terpene related compound. The fact that a correlation of m/z
135 and limonene (p b0.001) is also significant in the control group
lends support to our assignment, because one would expect levels of a
compound and its metabolite to be correlated in a group with well-
functioning livers. It is also of interest to note that the correlation
between limonene and m/z 89 has a very low p-value, but that m/z 89
Table 2
Calculatedmean volume mixingratios (VMRs) for methanol, 2-butanone, carbon disulfide, 2-pentanone, and limonene, in units of nmol/molfor pre- and post-transplantbreath samples.
The post-transplant values correspond to thosefor the last post-transplant sample given in Table 1. Measurement uncertainties are provided in brackets. The ratios of pre- topost-trans-
plant concentrations are also provided.
Patient Mean methanol VMRs
nmol/mol
Mean 2-butanone VMRs
nmol/mol
Mean carbon disulfide VMRs
nmol/mol
Mean 2-pentanone VMRs
nmol/mol
Mean limonene VMRs
nmol/mol
Pre Post Pre/post Pre Post Pre/post Pre Post Pre/post Pre Post Pre/post Pre Post Pre/post
F1 200 (6) 190 (3) 1.1 (0.04) 18 (1.2) 11 (0.2) 1.6 (0.1) 4.7 (0.4) 2.0 (0.4) 2.4 (0.5) 9.0 (0.8) 5.3 (0.6) 1.7 (0.2) 7.5 (0.6) 3.5 (0.6) 2.1 (0.4)
F2 90 (9) 230 (6) 0.4 (0.04) 16 (1.1) 15 (1.9) 1.1 (0.2) 2.4 (0.5) 20 (2.9) 0.1 (0.03) 8.3 (0.8) 7.3 (0.6) 1.1 (0.1) 60 (4.7) 1.1 (0.1) 54 (6)
F3 530 (2) 78 (5) 6.8 (0.4) 21 (1.7) 12 (2.6) 1.8 (0.4) 11 (1.3) 2.3 (0.3) 4.8 (0.8) 10 (1.2) 5.7 (0.3) 1.8 (0.2) 14 (0.7) 2.3 (0.2) 6.1 (0.6)
F4 560 (18) 71 (3) 7.9 (0.4) 19 (1.2) 12 (0.4) 1.6 (0.1) 4.7 (0.2) 2.3 (0.4) 2.0 (0.4) 25 (3.3) 6.1 (0.3) 4.1 (0.6) 11 (0.5) 2.3 (0.1) 4.8 (0.3)
M1 170 (2) 120 (15) 1.4 (0.2) 26 (1.4) 14 (1.4) 1.9 (0.2) 9.3 (0.6) 2.1 (0.4) 4.4 (0.9) 10 (0.5) 17 (2) 0.6 (0.1) 32 (0.7) 3.3 (0.5) 9.7 (1.5)
M2 430 (36) 81 (1) 5.3 (0.4) 13 (2.2) 15 (1.7) 0.9 (0.2) 3.1 (0.1) 2.4 (0.5) 1.3 (0.3) 23 (0.6) 6.2 (1.5) 3.7 (0.9) 94 (3.4) 1.6 (0.2) 59 (8)
M3 320 (32) 86 (7) 3.7 (0.5) 55 (3.1) 46 (4.2) 1.2 (0.1) 18 (2.0) 1.8 (0.2) 10 (1.6) 29 (1.0) 13 (1.2) 2.2 (0.2) 170 (1) 5.7 (0.2) 30 (1)
M4 490 (37) 93 (6) 5.3 (0.5) 38 (5.9) 23 (7.5) 1.7 (0.6) 8.5 (1.0) 2.6 (0.5) 3.3 (0.7) 38 (4.3) 9.9 (0.9) 3.8 (0.6) 110 (6) 55 (4) 2.0 (0.2)
M5 190 (3) 320 (1) 0.6 (0.01) 15 (0.5) 17 (1.3) 0.9 (0.1) 4.0 (0.4) 3.6 (0.1) 1.1 (0.1) 21 (1.3) 6.7 (0.5) 3.1 (0.3) 9.2 (0.5) 4.1 (0.2) 2.2 (0.2)
M6 230 (12) 79 (5) 2.9 (0.2) 30 (2.2) 12 (3.3) 2.5 (0.7) 13 (1.9) 2.1 (0.4) 6.2 (1.5) 13 (1.3) 5.1 (0.2) 2.5 (0.3) 120 (8) 4.8 (0.3) 25 (2)
M7 510 (21) 160 (8) 3.2 (0.2) 19 (0.7) 24 (4.2) 0.8 (0.1) 4.9 (0.9) 4.5 (0.7) 1.1 (0.3) 14 (2.6) 7.5 (0.3) 1.9 (0.4) 47 (2) 2.6 (0.1) 18 (1)
M8 180 (5) 86 (2) 2.1 (0.07) 21 (1.0) 12 (0.4) 1.8 (0.1) 6.8 (0.6) 1.9 (0.3) 3.6 (0.7) 8.8 (0.3) 13 (1.6) 0.7 (0.1) 7.7 (0.6) 3 (0.2) 2.6 (0.3)
1248 R. Fernández del Río et al. / EBioMedicine 2 (2015) 1243–1250
shows no discrimination between pre- and post-liver transplant. This is
suggestive that m/z 89 arises from an independentprocess related to the
patient's illness, but is not related to the cirrhosis itself.
The enhanced levels of methanol and 2-pentanone in pre-transplant
patients could come from a number of sources, including diet. Elevated
levels of methanol have been reported following consumption of alco-
hol or large quantities of fruit (Lindinger et al., 1997). It is a product of
the degradation of pectin by colonic bacteria (Siragusa et al., 1988),
and of metabolism of the sweetener aspartame (Stegink et al., 1989).
However, and in agreement with Morisco et al. (2013) we find that
fruit consumption cannot explain the increased methanol concentra-
tions in the breath of liver patients compared to controls. Alcoholic
drinks are a source of methanol, but only one patient reported that he
had drunk alcohol within the 24 h prior to the breath sampling.
Methanolis metabolised in humansin the liver, mainly by alcohol dehy-
drogenase (Skrzydlewska, 2003), soit is possible that this mechanism is
impaired when a liver becomes cirrhotic. Morisco et al.(2013), also
found elevated methanol in cirrhotic patients versus healthy controls.
The source of 2-pentanone in breath is unknown (King et al., 2010). It
has been found in human breath, faeces, skin and urine (de Lacy et al.,
2014), and it has been suggested that lung cells produce 2-pentanone
(Filipiak et al., 2010). 2-pentanone was suggested as a biomarker for
liver disease by three previous studies (Morisco et al., 2013; Van den
Velde et al., 2008; Vollnberg et al., 2009).
In conclusion, we have performed a two-stage study which com-
pares volatiles in the breath of pre-transplant cirrhotic patients with
controls followed by pre- and post-transplant breath samples. This has
resulted in an assignment of methanol, 2-pentanone and limonene as
markers in exhaled breath for the cirrhotic liver. We have demonstrated
that limonene can also be used for assessing liver function following
transplant by monitoring wash-out. Our study links limonene with the
diseased organ itself, rather than simply the diseased patient as a
whole. Breath volatiles have the advantage of offering non-invasive
testing, but also offer the opportunity to assess the global function of
the liver, rather than a localised test such as a biopsy. Our study raises
the possibility of a pharmacokinetic-based test for assessing liver func-
tion which could be used for diagnosing liver disease, i.e. where a
known quantity of limonene is administered and its wash-out in breath
is assessed over time. Importantly, this study provides a set of bio-
markers which can be used in future studies to assess the potential of
breath analysis for the diagnosis of early-stage liver disease.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.ebiom.2015.07.027.
Conflicts of Interest
All authors have declared no conflicts of interest.
Contributions
MEOH and CAM in discussion with AH and TS were responsible for
the conception of this study. RFdR was responsible for the design and
development of the breath sampling system, taking the majority of
the breath samples, for proposing the longitudinal study, and for
undertaking most of the mass spectrometric measurements. RFdR and
MEOH were responsible for the data analysis of the mass spectrometric
files. MEOH was responsible for obtaining ethical approval, patient
recruitment, study management, statistical analyses of the data and
assisted with some of the sample collection and measurement. CAM,
MEOH and RFdR wrote the paper with input from PP, AH, TS, and TW.
PP, AH, TS, and TW were responsible for the provision of patients and
for providing the medical context. The final paper has been approved
by all authors.
Ethics Approval
The regional ethics committee of Camden and Islington, London
approved this study (REC reference: 13/LO/0952). Informed consent
was obtained from each volunteer.
Funding Sources
This work was in part funded through the Proton Ionisation Molec-
ular Mass Spectrometry (PIMMS) Initial Training Network which is
supported by the European Commission 7th Framework Programme
under Grant Agreement Number 287382. MEOH thanks the Daphne
Jackson Trust for a fellowship and the Engineering and Physical Sciences
Research Council and University of Birmingham for her sponsorship.
We thank the Wellcome Trust (grant number 097825/Z/11/B) for an
Institutional Strategic Support Fund. We acknowledge the support of
the National Institute of Health Research Clinical Research Network
(NIHR CRN).
Acknowledgements
Preliminary work dealing with liver disease and breath volatiles by
the Molecular Physics Group was published in a PhD thesis by Brown,
024681012141640 80 120
0
20
40
60
80
100
F2
F4
F5
M3
M7
Limonene VMR (nmol/mol)
Days after transplant
Fig. 3. Longitudinal changes in volumemixing ratios (VMRs)in nmol/mol for limonene at
given daysafter liver transplant for patients F2,F4, F5, M3, and M7. Thedata point at day 0
for F2 was taken just before transplant surgery.
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.2
0.4
0.6
0.8
1.0
Sensitivity
1-specificity
Fig. 4. Receiver operating characteristic curvefor a combination of methanol, 2-pentanone
and limonene data in the study groups.
1249R. Fernández del Río et al. / EBioMedicine 2 (2015) 1243–1250
P.A., (2012) (http://etheses.bham.ac.uk/3839/). We thank Sister Diana
Hull, Senior Research Nurse, funded through the National Institute for
Health Research, for her considerable assistance in completing case sup-
port forms, providingpatient data and for generally facilitating the stud-
ies. We alsothank Sister Carmel Maguire for her help in the out-patient
clinic and Sister Samantha Howell for her assistance in transplant as-
sessment clinic. Finally, we wish to thank the patients and controls for
their cooperation in participating in this study.
References
Agarwal, B., Petersson, F., Jurschik, S., Sulzer, P., Jordan, A., Mark, T.D., et al., 2011. Use of
proton transfer reaction time-of-flight mass spectrometry for the analyticaldetection
of illicit and controlled prescription drugs at room temperature via direct headspace
sampling. Anal. Bioanal. Chem. 400 (8), 2631–2639.
Beauchamp, J., 2011. Inhaled today, not gone tomorrow: pharmacokinetics and
environmental exposure of volatiles in exhaled breath. J. Breath Res. 5 (3), 037103
(Epub 2011/06/10).
Beauchamp, J., Herbig, J., Gutmann, R., Hansel, A., 2008. On the use of Tedlar(R) bags for
breath-gas sampling and analysis. J Breath Res 2 (4), 046001. http://dx.doi.org/10.
1088/1752-7155/2/4/046001 Epub 2008/12/01.
Boston, R.C., Schnall, M.D., Englander, S.A., Landis, J.R., Moate, P.J., 2005. Estimation of
the content of fat and parenc hyma in breast tissue using MRI T1 histograms and
phantoms. Magn. Reson. Imaging 23 (4), 591–599 (Epub 2005/05/28).
Broadhurst, D.I., Kell, D.B., 2006. Statistical strategies for avoiding false discoveries in
metabolomics and related experiments. Metabolomics 2 (4), 171–196.
Brown, P., Watts, P. , Mark, T.D., Mayhew, C.A., 2010. Prot on transfer reaction mass
spectrometry investigations on the effects of reduced electric field and reagent ion
internal energy on product ion branching ratios for a series of saturated alcohols.
Int. J. Mass Spectrom. 294 (2–3), 103–111.
Castera, L., Lik Yuen Chan, H., Arrese, M. (Chairs), Afdhal, N., Bedossa, P., Friedrich-Rust,
M., Han, K.H., Pinzani, M. (Clinical Practice Guidelines Panel members), 2015. EASL-
ALEH Clinical Practice Guidelines: Non-invasive tests for evaluation of liver disease
severity and prognosis. J. Hepatol. 63, 237–264.
Dadamio, J., Van den Velde, S., Laleman, W., Van Hee, P., Coucke, W., Nevens, F., et al.,
2012. Breath biomarkers of liver cirrhosis. J. Chromatogr. B Anal. Technol. Biomed.
Life Sci. 905, 17–22 (Epub 2012/08/28).
Davies, S., 2012. CMO Annual Report: Volume One, 2011 ‘On The State of the Public's Health’.
Department of Health (Available from: http://webarchive.nationalarchives.gov.uk/
20130107105354/http://www.dh.gov.uk/health/2012/11/cmo-annual-report/).
de Gouw, J.A., Middlebrook, A.M., Warneke, C., Ahmadov, R., Atlas, E.L., Bahreini, R., et al.,
2011. Organic aerosol formation downwind from the deepwater horizon oil spill.
Science 331 (6022), 1295–1299.
de Lacy, Costello B., Amann, A., Al-Kateb, H., Flynn, C., Filipiak, W., Khalid, T., et al.,2014. A
review of the volatiles from the healthy human body. J. Breath Res. 8 (1), 014001
(Epub 2014/01/15).
Ellis, A.M., Mayhew, C.A., 2014. Proton Transfer Reaction Mass Spectrometry Principles
and Applications. John Wiley & Sons Ltd., UK.
Falk, A., Gullstrand, E., Lof, A., Wigaeus-Hjelm, E., 1990. Liquid/air partition coefficients of
four terpenes. Br. J. Ind. Med. 47 (1), 62–64 (Epub 1990/01/01).
Filipiak, W., Sponring, A., Filipiak, A., Ager, C., Schubert, J., Miekisch, W., et al., 2 010.
TD–GC–MS analysis of volatile metabolites of human lung cancer and normal cells
in vitro. Cancer Epidemiol. Biomark. Prev. 19 (1), 182–195.
Friedman, M.I., Preti, G., Deems, R.O., Friedman, L.S., Munoz, S.J., Maddrey, W.C., 1994.
Limonene in expire d lung air of patients wi th liver-disease. Dig. Dis. Sci. 39 (8 ),
1672–1676.
Frye, R.F., Zgheib, N.K.,Matzke, G.R., Chaves-Gnecco, D., Rabinovitz, M., Shaikh, O.S., et al.,
2006. Liver disease selectively modulates cytochrome P450—mediated metabolism.
Clin. Pharmacol. Ther. 80 (3), 235–245 (Epub 2006/09/06).
Griffin, S., Wyllie, S.G., Markham, J., 1999. Determination of o ctanol–water partition
coefficient for terpenoids using reversed-phase high-performance liquid chromatog-
raphy. J. Chromatogr. A 864 (2), 221–228.
Hanouneh, I.A., Zein, N.N., Cikach, F., Dababneh, L., Grove, D., Alkhouri, N., et al., 2014.
The breathprints in patients with liver disease identify novel breath biomarkers in
alcoholic hepatitis. Clin. Gastroenterol. Hepatol. 12 (3), 516–523.
Jurschik, S., Agarwal, B., Kassebacher, T., Sulzer, P., Mayhew, C.A., Mark, T.D., 2012. Rapid
and facile detection of four date rape drugs in different beverages utilizing proton
transfer reaction mass spectrometry (PTR-MS). J. Mass Spectrom. 47 (9), 1092–1097.
Khalid, T.Y., Costello, B., Ewen, R., White, P., Stevens, S., Gordon, F., et al., 2013. Breath vol-
atile analysis from patients diagnosed with harmful drinking, cirrhosis and hepatic
encephalopathy: a pilot study. Metabolomics 9 (5), 938–948.
Khalid, T., Richardson, P., Probert, C.S., 2014. The liver breath! breath volatile organic
compounds for the diagnosis of liver disease. Clin. Gastroenterol. Hepatol. 12 (3),
524–526.
King, J., Mochalski, P., Kupferthaler, A., Unterkofler, K., Koc, H., Filipiak, W., et al., 2010.
Dynamic profiles of volatile organic compounds in exhaled breath as determined
by a coupled PTR-MS/GC–MS study. Physiol. Meas. 31 (9), 1169–1184.
Kwak, J., Preti, G., 2011. Volatile disease biomarkers in breath: a critique. Curr. Pharm.
Biotechnol. 12 (7), 1067–1074.
Lindinger, W., Taucher, J., Jordan, A., Hansel, A., Vogel, W., 1997. Endogenous production
of methanol after the consumption of fruit. Alcohol. Clin. Exp. Res. 21 (5), 939–943.
Lindinger, W., Hansel, A., Jordan, A., 1998. On-line monitoring of volatile organic com-
pounds at pptv lev els by means of proton-transfer-reaction mass spectrometr y
(PTR-MS) —medical applications, food control and environmental research. Int.
J. Mass Spectrom. 173 (3), 191–241.
Miller, J.A., Lang, J.E., Ley, M., Nagle, R., Hsu, C.H., Thompson, P.A., et al., 2013. Human
breast tissue disposition and bioactivity of limonene in women with early-stage
breast cancer. Cancer Prev. Res. (Phila.) 6 (6), 577–584 (Epub 2013/04/05).
Millonig, G., Praun, S., Netzer, M., Baumgartner, C., Dornauer, A., Mueller, S., et al., 2010.
Non-invasive diagnosis of liver diseases by breath analysis using an optimize d
ion-molecule reactionmass spectrometry approach: a pilot study. Biom arkers 15
(4), 297–306.
Miyazawa, M., Shindo, M., Shimada, T., 2002. Metabolism of (+)- and (−)-limonenes
to respective carveols and perillyl alcohols by CYP2C9 and CYP2C19 in human liver
microsomes. Drug Metab. Dispos. 30 (5), 602–607.
Mochalski, P., King, J., Klieber, M., Unterkofler, K., Hinterhuber, H., Baumann, M., et al.,
2013. Blood and breath lev els of selected volat ile organic compounds in healthy
volunteers. Analyst 138 (7), 2134–2145.
Morisco, F., Aprea, E.,Lembo, V., Fogliano,V., Vitaglione, P., Mazzone, G., etal., 2013. Rapid
“breath-print”of liver cirrhosis by proton transfer reaction time-of-flight mass
spectrometry. a pilot study. PloS One 8 (4).
O'Hara,M.E., O'Hehir, S.,Green, S., Mayhew, C.A., 2008. Development of a protocol to mea-
sure volatile organic compounds in human breath: a comparison of rebreathing and
on-line single exhalationsusing proton transfer reaction mass spectrometry. Physiol.
Meas.29(3),309–330.
O'Hara, M.E., Clutton-Brock, T.H., Green, S., O'Hehir, S., Mayhew, C.A., 2009a. Mass
spectrometric investigations to obtain the first direct comparisons of endogenous
breath and blood volatile organic compound concentrations in healthy volunteers.
Int. J. Mass Spectrom. 281 (1–2), 92–96.
O'Hara, M.E., Clutton-Brock, T.H., Green, S., Mayhew, C.A., 2009b. Endogenous volatile
organic compounds in breath and blood of healthy volunteers: examining breath
analysis as a surrogate for blood measurements. J. Breath Res. 3 (2).
Pleil, J.D., Stiegel, M.A., Risby, T.H., 2013. Clinical breath analysis: discriminating between
human endogeno us compounds and exogenous (environmental) chemical con-
founders. J. Breath Res. 7 (1), 017107 (Epub 2013/03/01).
Ransohoff, D.F., 2005. Bias as a threat to the validity of cancermolecular-marker research.
Nat. Rev. Cancer 5 (2), 142–149 (Epub 2005/02/03).
Sangster, J., 1989. Octanol–water partition-coefficients of simple organic-compounds.
J. Phys. Chem. Ref. Data 18 (3), 1111–1229.
Schubert, J.K., Sp ittler, K.H., Braun, G., Geiger, K., Guttmann, J., 2001. CO
2
-controlled
sampling of alveolar gas in mechanically ventilated patients. J. Appl. Physiol. 90 (2),
486–492.
Schubert, J.K., Miekisch, W., Birken, T., Geiger, K., Noldge-Schomburg, G.F.E., 2005. Impact
of inspired substance concentrations on the results of breath analysis in mechanically
ventilated patients. Biomarkers 10 (2–3), 138–152.
Sehnert, S.S., Jiang, L., Burdick, J.F., Risby, T.H., 2002. Breath biomarkers for detection of
human liver diseases: preliminary study. Biomarkers 7 (2), 174–187.
Shimamoto, C., Hirata, I., Katsu, K., 2000. Breath and blood ammonia in liver cirrhosis.
Hepato-Gastroenterology 47 (32), 443–445.
Siragusa,R.J., Cerda, J.J., Baig, M.M., Burgin, C.W., Robbins, F.L., 1988. Methanol production
from the degradation of pectin by human colonic bacteria. Am. J. Clin. Nutr. 47 (5),
848–851.
Skrzydlewska, E., 2003. Toxicological and metabolic consequences of methanol poisoning.
Toxicol. Mech. Methods 13 (4), 277–293.
Solga, S.F., Alkhuraishe, A., Cope, K., Tabesh, A., Clark, J.M., Torbenson, M., et al., 2006.
Breath biomarkers and non-alcoholic fatty liver disease: preliminary observations.
Biomarkers 11 (2), 174–183.
Spanel, P., Dryahina, K., Smith, D., 2013. A quantitative study of the influence of inhaled
compounds on their concentrations in exhaled breath. J. Breath Res. 7 (1), 017106
(Epub 2013/ 03/01).
Stegink, L.D., Filer Jr., L.J., Bell, E.F., Ziegler, E.E., Tephly, T.R., 1989. Effect of repeated inges-
tion of aspartame-sweetened beverage on plasma amino acid, blood methanol, and
blood formate concentrations in normal adults. Metabolism 38 (4), 357–363.
Tangerman, A., Meuwesearends, M.T., Vantongeren, J.H.M., 1983. A new sensitive assay
for measuring volatile sulfur-compounds in human breath by Tenax trapping and
gas-chromatography and its application in liver-cirrhosis. Clin. Chim. Acta 130 (1),
103–110.
Van den Velde,S., Nevens, F., Van Hee,P., van Steenberghe,D., Quirynen, M., 2008.GC–MS
analysis of breath odor compounds in liver patients. J. Chromatogr. B Anal. Technol.
Biomed. Life Sci. 875 (2), 344–348.
Verdam, F.J., Dallinga, J.W., Driessen, A., de Jonge, C., Moonen, E.J., van Berkel, J.B., et al.,
2013. Non-alcoholic steatohepatitis: a non-invasive diagnosis by analysis of exhaled
breath. J. Hepatol. 5 8 (3), 543–548 (Epub 2012/11/13).
Vollnberg, B.O., Gunther, R., Buch, S., Schafmayer, C., Braun, F., Hinrichsen, H.M., et al., 2009.
Metabolomic analysis of breath air and tongue swabs using gas chromatography/mass
spectrometry identifies novel metabolic markers of hepatic encephalopathy. Gastroen-
terology 136 (5), A829–A830.
Williams, R., Aspinall, R., Bellis, M., Camps-Walsh, G., Cramp, M., Dhawan, A., et al., 2014.
Addressing liver disease in the UK: a blueprint for attaining excellence in health care
and reducing premature mortality from lifestyle issues of exces s consumption of
alcohol, obesity, and viral hepatitis. Lancet 384 (9958), 1953–1997 (Epub 2014/12/01).
1250 R. Fernández del Río et al. / EBioMedicine 2 (2015) 1243–1250
