Available via license: CC BY 4.0
Content may be subject to copyright.
Melanie Fachet*, Simon Lowitzki, Marie-Louise Reckzeh, Thorsten Walles, and Christoph
Hoeschen
Investigation of everyday influencing factors
on the variability of exhaled breath profiles in
healthy subjects
https://doi.org/10.1515/cdbme-2022-1067
Abstract:
Introduction: The human breath is an accurate but complex
read-out of many physiological processes in the organism that
can be monitored via volatile organic compounds (VOCs) in
the exhaled air. However, there are many confounding vari-
ables that limit the transfer and application of breath analysis
to become a clinical procedure.
Method: This work aims to establish a systematic procedure
for sampling and characterization of various everyday influ-
ences of healthy subjects using proton transfer reaction-mass
spectrometry (PTR-MS). In order to limit the influencing fac-
tors on the breath profile, a standard analysis procedure for
sampling and evaluation of the exhaled breath samples was
developed. The correlations between the selected experimen-
tal conditions and the resulting VOC profiles were investigated
using a non-parametric Wilcoxon rank sum test.
Results: In addition to the relevant influence of methodolog-
ical experimental parameters, interesting insights into the ef-
fect of everyday factors on the exhalat gas were obtained and
discussed. Furthermore, subject and condition-specific differ-
ences were found in the exhaled air of male and female sub-
jects.
Conclusion: With a more robust, standardized and repro-
ducible breath sampling protocol, breath analysis is a promis-
ing non-invasive tool towards a system-wide understanding
and personalized diagnosis and treatment of a wide range of
diseases.
Keywords: Breath gas analysis, Breathomics, Proton-
transfer-reaction mass spectrometry, Subject variability, Ev-
eryday influencing factors
*Corresponding author: Melanie Fachet, Simon Lowitzki,
Christoph Hoeschen, Institute for Medical Technology, Chair of
Medical Systems Technology, Otto von Guericke University
Magdeburg, Universitätsplatz 2, 39106 Magdeburg, Germany.
e-mail: melanie.fachet@ovgu.de
Marie-Louise Reckzeh, Thorsten Walles, University Clinic for
Cardiac and Thoracic Surgery, Otto von Guericke University
Magdeburg, Leipziger Str. 44, 39120 Magdeburg, Germany.
1 Introduction
In the last decades, the analysis of volatile organic compounds
(VOCs) in the exhaled human breath has been studied and
developed as a promising technique to identify biomarkers
for the diagnosis and monitoring of various diseases. One of
the main advantages of this procedure is its non-invasive ap-
proach. This is particularly important for populations such as
children and elderly people, and for diseases whose current
standard diagnoses use invasive techniques such as biopsies
and bronchoscopies or imaging methods based on ionizing ra-
diation. Despite of the ongoing advances in this research field,
breath gas analysis is not yet a routine clinical tool. Standard
clinical procedures have not yet been sufficiently established
and validated, especially when it comes to identifying suitable
biomarkers and setting comparative values for healthy volun-
teers [10].
Recent advancements in breath research have led to the iden-
tification of biomarkers in a wide range of diseases such
as lung and breast cancer, COPD, asthma, diabetes, dis-
eases of the skin barrier and many more [2, 4, 6, 9]. Var-
ious methods are available for the measurement of VOCs,
which vary in their detection sensitivity and analysis speed e.g.
gas chromatography-mass spectrometry (GC-MS), electronic
nose, proton transfer reaction-mass spectrometry (PTR-MS),
ion mobility spectrometry (IMS), chemiluminescence or opti-
cal absorption detection techniques [11].
The work presented in this paper uses PTR-MS, which of-
fers the advantages of a rapid response, a soft chemical ion-
ization principle along with the possibility for an absolute
quantification and a high sensitivity down to a parts per tril-
lion (ppt) level [11]. PTR-MS is based on a soft chemical
ionization, where a hydronium ion (H3O+) is used to charge
VOC molecules Rfor proton affinities higher than for water
molecules [1].
It has been shown that a breath sample contains more than
3.500 different VOCs, mostly in the picomolar range. The gas
exchange between the blood system and the external environ-
ment can be monitored in the human breath. This process of
alveolar gas exchange with the blood facilitates oxygen up-
take and releases by-products of metabolic reactions such as
cdbme_2022_8_2.pdf 261 8/29/2022 5:45:49 PM
261
DE GRUYTER
Current Directions in Biomedical Engineering 2022;8(2): 261-264
Open Access. © 2022 The Author(s), published by De Gruyter. This work is licensed under the Creative Commons Attribution 4.0 International License.
M. Fachet et al., Breath gas variability in health volunteers
exhaled breath gas volatiles such as methanol, ethanol, ac-
etaldehyde, acetone and isoprene [8]. The specific composi-
tion of the individual breath pattern is influenced by the sub-
ject’s physiological situation, the lifestyle and state of health
[8]. However, there is currently a lack of standardization in
breath sampling, data acquisition and analysis for the process-
ing of PTR-MS data from clinical study cohorts that are suited
for biomarker identification.
The work presented in this paper aims to investigate the influ-
ence of everyday lifestyle factors and the volatile biomarker
profiles in healthy subjects affecting the reliability, repro-
ducibility and suitability of breath gas analysis for the design
of clinical studies. In this setting, several everyday influenc-
ing factors such as getting up, brushing teeth and food up-
take were investigated for its variability in the VOC abundance
along with possible interferences from ambient air samples.
In this study, the influence of different everyday factors was
determined by measuring the complete range of VOC mass-
to-charge ratios between m/z 20 to m/z 200 [8]. It is impor-
tant to determine which of the various sampling parameters
has a larger influence on the VOC profile and should there-
fore be avoided in clinical studies to minimize inter- and intra-
individual variability of confounding factors.
2 Material and Methods
2.1 Study population
The healthy subject population consisted of 13 voluntary re-
searchers (7 male and 6 female subjects) from the Chair of
Medical Systems Technology and the University Clinic for
Cardiac and Thoracic Surgery. The informed consent was ob-
tained from the study participants and the study was approved
by the institutional ethics committees on human research of
the Otto-von-Guericke University Magdeburg (vote 194/20).
2.2 Breath gas sampling
Breath gas samples were collected in 3 l Tedlar bags. In or-
der to ensure a minimum level of contaminations in the reused
bags, the bags were purged with nitrogen (99.5% purity) twice
and a low background VOC level was verified by additional
measurements. The exhalation volume of 2-2.5 l was typically
reached after 20–30 s containing a mixed fraction of the sub-
ject’s exhaled breath and the samples were subsequently mea-
sured within 2 h after breath collection. Ambient air samples
were taken from laboratories and office rooms where the study
subjects were located in.
2.3 Measurement of samples with
PTR-MS
The breath gas analysis was conducted using a commercial
standard PTR-MS with Time-of-Flight (TOF) mass detec-
tor (PTR-TOF 2000, Ionicon Analytik, Innsbruck, Austria),
which allows very sensitive offline and online measurements
in the low ppb to ppt range. The breath sampling was per-
formed as previously described by [1] and [8]. Briefly, the
PTR-MS measurements were performed with a drift tube pres-
sure of approximately 2.3 mbar. The VOC masses were ana-
lyzed in consecutive scans from a mass-to-charge ratio ranging
from m/z 20 to 200.
2.4 Statistical analysis
The measured VOC profiles of the healthy volunteers were
evaluated using the statistical toolbox in MATLAB (Math-
Works, Version R2021a). Descriptive measures included the
median and interquartile range (IQR) for the selected breath
volatiles. Furthermore, a non-parametric test Wilcoxon rank
sum test was performed for comparison of the two indepen-
dent sampling conditions. This statistical test is well suited for
the small evaluated sample size and because the VOC intensi-
ties are not normally distributed.
3 Results and discussion
3.1 Influence of everyday factors on the
breath profile
The exhaled ethanol concentration (m/z 47) increased after the
condition "eating cake" as shown in Fig. 3. This is attributed
to the function of endogenous ethanol in the carbohydrate
metabolism in the small intestine [7]. The breakdown of com-
plex carbohydrates to glucose in the lower gastrointestinal
tract is mediated by endogenous ethanol which is known to
increase the permability of epithelial and colon cells making
it available for glycolysis [7]. In contrast, the condition "eat-
ing lunch" had no significant influence on the average ethanol
abundance (Fig. 2).
Related to elevated endogenous ethanol abundance that is in-
volved in the intestinal glucose transport, we also found an in-
creased level of breath acetone (m/z 59) as a metabolic byprod-
uct (Fig. 3). We observed a slight decrease in acetone abun-
dances with high physiological variation in both male and fe-
male subjects.
cdbme_2022_8_2.pdf 262 8/29/2022 5:45:49 PM
262
M. Fachet et al., Breath gas variability in health volunteers
Fig. 1: Comparison of differences in inter-individual variations of
selected VOCs after getting up and before brushing teeth related
to subject gender.
Previous studies have shown that endogenous breath isoprene
is a product of the mevalonate pathway related to the biosyn-
thesis of cholesterol [7]. Under all three investigated condi-
tions, a decrease in endogenous isoprene (m/z 69) level com-
pared to the control sample was observed (Fig. 1 - 3). Methanol
(m/z 33) showed the lowest variabilities in its median and is
least sensitive to the influence of the investigated everyday life
influencing factors. In addition, we observed higher average
concentrations of VOCs for male subjects compared to female
subjects. The non-parametric Wilcoxon rank sum test shown
in Table 1 illustrates that, apart from m/z 33 (Methanol), all
other investigated breath components are significantly differ-
ent between the male and the female volunteers. In contrast,
the Wilcoxon rank sum test showed no significant differences
of the investigated volatiles before and after eating lunch.
The inter-subject variability of all measurements was com-
pared by calculating the relative standard deviation for all mea-
surements. The lowest relative standard deviation of 68 % was
observed for acetaldehyde (m/z 45). The endogenous acetone
abundance (m/z 59) had the highest variability with 115 % rel-
ative standard deviation from the mean value.
3.2 Intra- vs. inter-individual differences
of breath biomarkers
Each subject has its own “breath fingerprint”, a characteristic
profile of exhaled VOCs, which is influenced by exogenous
and endogenous factors. To check whether the inter-individual
differences between the subjects are greater than the intra-
individual differences of each individual subject, the vari-
ability between the measurements was analyzed. The results
indicated that the intra-subject differences for isopropanol
(m/z 43), acetaldehyde (m/z 45) and isoprene (m/z 69) predom-
inate, while the inter-subject differences for methanol (m/z 33),
ethanol (m/z 47) and acetone (m/z 69) were larger. Methanol
(m/z 33) and acetone (m/z 59) were found to show higher inter-
than intra-individual differences [8], which can be confirmed
by our results.
3.3 Implications for the design of clinical
studies using PTR-MS
The experimental condition "eating lunch" led to a higher vari-
ability among the study subjects due to the intake of different
meals and should therefore be avoided in clinical studies when
confounding factors should be limited to a minimum. To fur-
ther reduce the variability for ethanol (m/z 47) and acetalde-
hyde (m/z 45), food intake with a high sugar content should be
avoided prior to sampling. The experimental condition "brush-
ing teeth" had only a minor impact on the VOC variability
compared to the condition "getting up" and might not have a
significant impact on the sampling protocol in future clinical
studies (Fig. 1).
4 Summary
In summary, this work provides a tool to systematically evalu-
ate the influence of everyday influencing factors on the breath
profile of healthy subjects. This is important for future clinical
studies to limit the effect of confounding factors and to identify
a robust set of clinically relevant biomarkers for diagnostic and
therapeutic monitoring. With the simple, fast and non-invasive
technique of breath gas analysis based on PTR-MS, it gives the
opportunity to develop a possible application of breathomics
as a diagnostic and therapeutic monitoring tool.
Author Statement
Research funding: The authors state that no funding was in-
volved. Conflict of interest: Authors state no conflict of inter-
est.
cdbme_2022_8_2.pdf 263 8/29/2022 5:45:49 PM
263
M. Fachet et al., Breath gas variability in health volunteers
Fig. 2: Comparison of differences in inter-individual variations of
selected VOCs before and after eating lunch related to subject
gender.
Fig. 3: Comparison of differences in inter-individual variations
of selected VOCs before and after eating cake related to subject
gender.
Tab. 1: Significance testing of breath metabolites for discrimation
between subject gender and influencing factors.
p values for
Ion Tentative female vs. before and
(m/z) compound male volunteers after lunch
33 Methanol 0.14 0.84
43 Isopropanol 0.01 0.38
45 Acetaldehyde 0.02 0.76
45 Ethanol 0.02 0.61
45 Acetone 0.03 0.68
45 Isoprene 0.03 0.44
References
[1] Brunner C, Szymczak W, Höllriegl V, Mörtl S, Oelmez
H, Bergner A, Huber RM, Hoeschen C, Oeh U. Discrim-
ination of cancerous and non-cancerous cell lines by
headspace-analysis with PTR-MS. Anal Bioanal Chem
2010;397:2315–2324.
[2] Duffy E, Jacobs MR, Kirby B, Morrin Aoife. Probing skin
physiology through the volatile footprint: Discriminating
volatile emissions before and after acute barrier disruption.
Experimental Dermatology 2017;10:919-925.
[3] Ehmann R, Boedeker E, Friedrich U, Sagert J, Dippon J,
Friedel G, Walles T. Canine scent detection in the diagnosis
of lung cancer: revisiting a puzzling phenomenon. Eur Respir
J. 2012 Mar;39(3):669-76.
[4] Jiang C, Dobrowolny H, Gescher DM, Meyer-Lotz G, Steiner
J, Hoeschen C, Frodl T. Volatile organic compounds from
exhaled breath in schizophrenia. World J Biol Psychiatry.
2022.
[5] Kim KH, Jahan SA, Kabir E. A review of breath analysis for
diagnosis of human health. Trends in Analytical Chemistry
2012;33:1–8.
[6] Schallschmidt K, Becker R, Jung C, Bremser W, Walles T,
Neudecker J, Leschber G, Frese S and Nehls I. Comparison
of volatile organic compounds from lung cancer patients and
healthy controls—challenges and limitations of an observa-
tional study. J Breath Res 2016;10(4):1–17.
[7] Sukul P, Grzegorzewski S, Broderius C, Trefz P, Mittlmeier
T, Fischer DC, Miekisch W, Schubert J. Physiological and
metabolic effects of healthy female aging on exhaled breath
biomarkers. iScience; 25. 103739.
[8] Thekedar B, Szymczak W, Höllriegl V, Hoeschen C and Oeh
U. Investigations on the variability of breath gas sampling
using PTR-MS. J Breath Res 2009;3(2):1–12.
[9] Trefz P, Schmidt S, Sukul P, Schubert J, Miekisch W, Fischer,
DC. Non-Invasive Assessment of Metabolic Adaptation in
Paediatric Patients Suffering from Type 1 Diabetes Mellitus.
Journal of Clinical Medicine 2019;8:1797.
[10] Walles T. The needle in a haystack. Eur J Cardiothorac Surg.
2016 Apr;49(4):1117-8.
[11] Wang Y, Chengyin S, Jianquan L, Haihe J, Yannan C. Pro-
ton Transfer Reaction Mass Spectrometry (PTR-MS). Mass
Spectrometry Handbook, First Edition. Edited by Mike S.
Lee. 2012 John Wiley & Sons, Inc.
cdbme_2022_8_2.pdf 264 8/29/2022 5:45:49 PM
264
