The unique feature of the volatilomic concept is that fast and reliable information on normal and abnormal processes occurring in the human organism is obtained via analysis of the human volatile chemical footprints, i.e. volatiles exhaled from or emitted or secreted by the human body into its surrounding environment [1, 2]. In this context, exhaled breath analysis holds a distinguish status. This is because breath samples can be obtained non-invasively, rapidly, taken as often as deemed necessary, and analysed using a range of analytical gas-phase detection techniques [1]. For real-time gas analysis, soft chemical ionization mass spectrometric analytical techniques, e.g. proton transfer reaction mass spectrometry or selected ion flow tube mass spectrometry, are useful. However, these techniques are not ideal for biomarker discovery research owing to their ambiguity in assigning volatiles to given product ions. Owing to the pre-separation of volatiles before mass spectrometric analysis and the greater information contained in the mass spectra, gas chromatography-mass spectrometry (GC-MS) provides the required high chemical specificity to identify compounds with a much higher level of confidence, and is hence considered the gold standard for the analysis of breath chemical fingerprints [1, 2]. GC-MS is, however, an off-site technique, often demanding transport of the breath samples and frequently prolonged periods of breath sample storage prior to any chemical analysis.

Sorbent tubes are commonly used for sample storage [3–5]. In this storage approach, volatiles contained in exhaled breath are retained on the adsorbent materials that fill the tubes. These tubes are then transported to a laboratory and stored until analysis is possible. Preservation of the breath sample volatiles' integrity during transport and storage is arguably one of the most critical challenges in breath gas analysis [6, 7]. Various physical phenomena can lead to serious breath sample quality assurance issues. These include, for examples, compound degradation, background emission of pollutants, retention in storage containers and interactions between sample constituents; all of which can irreversibly modify the original sample composition and consequently distort the results of the chemical analyses.

Owing to its low cost and hydrophobic nature, Tenax® TA is one of the most popular adsorbents used by researchers involved in GC-MS volatile breath analysis. Consequently, the recovery of breath volatiles from this adsorbent material under different storage durations or temperatures has received some attention [4, 8–10]. However, these existing studies are sparse and provide divergent results. For example, Harshman et al [8] monitored the stability of seventy-four breath volatiles over 31 d at three different storage temperatures (4 °C, 21 °C and 37 °C), with the results indicating 4 °C as the optimum storage temperature. The researchers also recommended that the analysis of any breath sample should occur within 14 d of storage to minimize a potential 1–2 standard deviation gain or loss of a given volatile concentration. In contrast, Kang and Thomas [9] investigated the stability of twenty-five breath samples obtained from a single volunteer on dual bed Tenax® TA:Carbograph adsorbent tubes, which were stored at −80 °C over 12.5 months. They reported the maximum storage duration under these conditions of 1.5 months.

The main goal of this note is to report data on the recovery of fifteen selected common volatiles found in exhaled breath from Tenax® TA adsorbent tubes stored at −80 °C for up to 90 d. For the simplicity of collecting the exhaled breath samples into Tenax® TD tubes an Owlstone Medical ReCIVA® Breath Sampler was used. The analysis of these tubes was by TD-GC-MS.

Reference mixtures produced from liquid chemicals were used to build the retention time library. Reagents with stated purities of 95–99.9% were purchased from Merck (Austria) and Fluka (Switzerland). Primary standards were prepared by introducing and evaporating several microliters of a given liquid volatile into evacuated glass bulbs (1 l, Supelco, Canada). Following this, primary standards were diluted by transferring appropriate volumes of the glass bulb mixtures into 3–25 l Tedlar bags (SKC Inc., USA) that were pre-filled with humid air (RH 100% at 34 °C). Effectively, gas mixtures with volatile volume fractions ranging from 5 to 20 parts per billion by volume (ppbv) were used to establish retention times and Kovats retention indices of the volatiles under study.

The identification of the monitored compounds was performed in two steps. First, the spectrum of a peak was checked against the NIST mass spectral library database. Second, the NIST tentative identification was verified by comparing the temperature-programmed Kovats retention indices of the peak of interest with the library of retention indices obtained for the reference mixtures. The retention indices were calculated using the homologous series of n-alkanes (C5–C15).

Stainless steel industry standard thermal desorption tubes (¼-inch outside diameter and 3½ inch long) with SilcoNert™ coating pre-packed with Tenax TA (35/65) were purchased from Markes International Limited (Bridgend, UK). Prior to obtaining a breath sample, the adsorbent tubes were pre-conditioned according to the manufacturer's recommendations.

Four volunteers (1 male and 3 females, ages 25–50 years, and all non-smokers) provided breath samples for the purpose of this study. No requirements regarding food intake or beverage drinking were requested. However, the study subjects were advised to avoid chewing gum and mouth fresheners and that they should refrain from using cosmetics or fragrances on the day of sampling. Exhaled breath samples were taken from each volunteer in a seated position after 10 min of rest using a ReCIVA® Breath Sampler (Owlstone Medical, UK) and using the sampling protocol recommended by that company. A major advantage of the ReCIVA® Breath Sampler is that it can provide a set of four breath samples simultaneously, so that each sample contains the same portion of exhaled breath. Each volunteer provided two sets of four breath samples taken on different days. Sampling was achieved automatically by drawing into each TD tube a volume of 500 ml of exhaled breath at a flow rate of 200 ml min−1 during end-tidal exhalation segments. One randomly chosen sample from each set was analysed immediately after sampling. The peak areas obtained during this measurement were taken as the base for the calculations of recoveries. The remaining sample tubes were closed with brass ¼-inch nuts and immediately stored at −80 °C (Arctiko, ULTF 80). After 15–20 min at −80 °C, the sealing nuts were retightened to account for any loosening resulting from thermal contraction. These samples were analysed after different storage times and the temperature of storage was monitored using the Arctiko internal software. There was no predefined time point of analysis of these samples; however, the recovery of breath samples was monitored over a period of 90 d. In parallel to the breath samples, one room air sample was taken per set using the ReCIVA® sampler and analysed immediately after sampling.

A two-stage thermal desorption process was performed using a UNITY thermal desorber and autosampler TD100 (Markes International Limited, UK). During the primary desorption, Tenax® TA adsorbent tubes were heated up to 280 °C for 6 min under a helium 6.0 flow maintained at a rate of 20 ml min−1. Next, the released volatiles were trapped on a cold trap that was packed with graphitized carbon black and maintained at 5 °C. The injection of the volatiles into the capillary column was performed via the rapid heating of the cold trap to 320 °C for 90 s in the splitless mode.

Volatile analysis relied on an Agilent 7890A/5975C GC-MS system (Agilent, USA). Extracted compounds were separated using a Rxi-624Sil MS column (30 m × 0.32 mm, layer thickness 1.8 μm, Restek, USA) operated in a constant helium flow of 1.5 ml min−1. The column temperature program was as follows: 40 °C for 10 min, followed by heating at 5 °C min−1 up to 150 °C, which was then held for 5 min, then further heating at 10 °C min−1 up to a maximum temperature of 280 °C, at which the column was maintained for a further 5 min.

The volatile chemical analysis was performed using electron ionization (using standard electron energies of 70 eV) and with the mass spectrometer working in a SCAN mode, with the associated m/z ratio ranging from 20 up to 250. The quadrupole, ion source, and transfer line were maintained at 150 °C, 230 °C and 280 °C, respectively. Chlorobenzene-d5 was used as the internal standard (TO-14 A standard, Restek) to provide quality control.

For the purposes of this study, peak areas of the volatiles as determined by GC-MS were used as the parameter to calculate the volatile recoveries. The recovery of the volatiles is expressed as a ratio of the VOC peak area obtained after a certain time of storage and its peak area obtained immediately after sampling. Effectively, three recovery points were obtained per set. For all compounds of interest, recovery triads obtained from all subjects were grouped together. As a stability criterion, the relative expanded uncertainty was used with a coverage factor of two.

In the following, the limit of detection (LOD) is defined as three times the noise amplitude, and only peaks with signal-to-noise ratio larger than nine, i.e. corresponding to the limit of quantification (LOQ) defined as 3 × LOD, were taken into account during the statistical analysis.

The precision of the GC-MS method was evaluated using relative standard deviations (RSDs). Percentage RSDs were calculated using the standard deviation of VOC responses obtained for two sets of breath samples obtained from two volunteers using the ReCIVA® Breath Sampler. The obtained retention indices (RIs) and percentage RSDs of the volatiles investigated in this study are as follows: α-pinene (RI = 920, 5.8%), β-pinene (RI = 968, 6.8%), γ‐terpinene (RI = 1048, 8.7%), isoprene (RI = 562, 6.4%), 2,3-butanedione (RI = 611, 9%), allyl methyl sulphide (RI = 698, 6%), acetoin (RI = 762, 3.8%), toluene (RI = 772, 4.8%), n-octane (RI = 800, 8%), n-nonane (RI = 900, 7.5%), DL-limonene (RI = 1043, 5.5%), 1-(methylthio)-propane (RI = 716, 8.1%), 1-(methylthio)-1-propene (RI = 742, 8.3%), p-xylene (RI = 891, 9%) and p-cymene (RI = 1047, 9%). These percentage RSD values are considered to be satisfactory for the purposes of this study.

This study is focused on blood-borne volatiles (i.e. those with a higher abundance in breath than generally found in inhaled air) with signal to noise ratios greater than 25 at the beginning of the storage. The threshold of 25 was arbitrarily chosen to avoid signals that are close to LOD and therefore are characterized by elevated RSD. Fifteen volatiles that satisfied these requirements were monitored in this study, namely, three hydrocarbons (isoprene, n-heptane, and n-nonane), two aromatics (toluene and p-cymene), three volatile sulphur compounds (allyl methyl sulfide, 1-(methylthio)-propane, and 1-(methylthio)-1-propene), four terpenes (α-pinene, DL-limonene, β-pinene, and γ-terpinene), one ketone (2-pentanone), acetoin and 2,3 butanedione.

The obtained recoveries of these volatiles from Tenax® TA adsorbent tubes stored at −80 °C for 90 d are presented in figures 1–3. From these figures, it can be seen that four of the selected compounds, namely p-cymene, n-nonane, 1-(methylthio)-propane and 1-(methylthio)-1-propene were found to be stable during the first four weeks of storage, whereas the recoveries of 2-pentanone, isoprene, 2,3 butanedione, toluene, α-pinene, DL-limonene, n-heptane, β-pinene and γ-terpinene fulfilled the applied criterion during the same period in all samples collected but one. The reasons for the lowered recoveries in these single samples are unclear. They are not related to the sample procedure, because the lowered recoveries were not observed in the case of the other volatiles in these samples. Some of them may, however, represent values falling within the range of 2 × RSD and 3 × RSD.

Figure 1. Recoveries of 2-pentanone, isoprene, 2,3 butanedione, toluene, α-pinene and DL-limonene from Tenax® TA adsorbent tubes stored at −80 °C over the period of 90 d. The green area corresponds to 2 × RSD.

Download figure:

Figure 2. Recoveries of β-pinene, p-cymene, n-nonane and n-heptane from Tenax® TA adsorbent tubes stored at −80 °C over the period of 90 d. The green area corresponds to 2 × RSD.

Download figure:

Figure 3. Recoveries of γ-terpinene, allylmethylsulfide, 1-(methylthio)-propane, 1-(methylthio)-1-propene and acetoin from Tenax® TA adsorbent tubes stored at −80 °C over the period of 90 d. The green area corresponds to 2 × RSD.

Download figure:

In comparison to all other volatiles, the stability of acetoin was found to be much poorer. Only 50% of the breath samples containing this compound fulfilled the applied criterion during the 90 d of storage. For allyl methyl sulphide, nine out of eleven samples were stable during this period. This can still be considered as satisfactory taking into consideration the expected elevated reactivity of sulphur containing compounds. Only one compound, n-nonane, was found to be stable during the whole of the investigated period. However, it needs to be stressed that the stabilities of the remaining aliphatic hydrocarbons (i.e. n-heptane and isoprene) were also acceptable during the 90 d of storage, with only a few samples having lowered recoveries.

The aim of this paper is to report results on the stability of fifteen commonly exhaled breath volatiles during their storage in Tenax® TA adsorbent tubes maintained at a temperature of −80 °C for up to 90 d. Of the fifteen volatiles investigated, only n-nonane was found to be stable throughout the whole of the investigated period. With the exception of just one of the compounds investigated, namely acetoin, all compounds were found to be stable during the first four weeks of their storage, i.e. p-cymene, n-nonane, 1-(methylthio)-propane, 1-(methylthio)-1-propene, 2-pentanone, isoprene, 2,3 butanedione, toluene, α-pinene, DL-limonene, n-heptane, β-pinene, allyl methyl sulphide and γ-terpinene.

The important outcome of this study is the evidence that has been provided showing that fourteen out of the fifteen key breath volatiles investigated can be stored at −80 °C for up to 4 weeks. These results are in agreement with those obtained by Kang and Thomas [9]. However, the results of our study has two other implications. Firstly, occasionally samples can lose integrity, and, therefore, the collection of multiple samples is advised—at the very least duplicate samples should be taken. Secondly, the results impose transport restrictions on breath samples stored in Tenax® TA adsorbent tubes at −80 °C, namely that they should be shipped on dry ice.

EU HORIZON Innovation Actions HORIZON-CL3-2021-DRS-01-05, Project Number 101073924 (ONELAB) for funding. The authors declare no conflicts of interest.

The study was conducted in accordance with the Declaration of Helsinki. Informed consent was obtained from all subjects involved in the study.