Thank you for visiting Nature.com. The browser version you are using has limited CSS support. For the best experience, we recommend that you use an updated browser (or disable Compatibility Mode in Internet Explorer). In the meantime, to ensure continued support, we will render the site without styles and JavaScript.
Interest in the analysis of volatile organic compounds (VOCs) in exhaled air has grown over the past two decades. Uncertainties still exist regarding the normalization of sampling and whether indoor air volatile organic compounds affect the exhaled air volatile organic compounds curve. Assess indoor air volatile organic compounds at routine breath sampling sites in the hospital environment and determine if this affects the composition of the breath. The second goal was to study the daily fluctuations in the content of volatile organic compounds in indoor air. Indoor air was collected at five locations in the morning and afternoon using a sampling pump and a thermal desorption (TD) tube. Collect breath samples only in the morning. TD tubes were analyzed by gas chromatography coupled with time-of-flight mass spectrometry (GC-TOF-MS). A total of 113 VOCs were identified in the collected samples. Multivariate analysis showed a clear separation between breathing and room air. The composition of indoor air changes throughout the day, and different locations have specific VOCs that do not affect the breathing profile. The breaths did not show separation based on location, suggesting that sampling can be done at different locations without affecting the results.
Volatile organic compounds (VOCs) are carbon-based compounds that are gaseous at room temperature and are the end products of many endogenous and exogenous processes1. For decades, researchers have been interested in VOCs because of their potential role as non-invasive biomarkers of human disease. However, uncertainty remains regarding the standardization of the collection and analysis of breath samples.
A key area of standardization for breath analysis is the potential impact of background VOCs in indoor ambient air. Previous studies have shown that background levels of VOCs in indoor ambient air affect the levels of VOCs found in exhaled air3. Boshier et al. In 2010, selected ion flow mass spectrometry (SIFT-MS) was used to study the levels of seven volatile organic compounds in three clinical settings. Different levels of volatile organic compounds in the environment were identified in the three regions, which in turn provided guidance on the ability of widespread volatile organic compounds in indoor air to be used as disease biomarkers. In 2013, Trefz et al. The ambient air in the operating room and the breathing patterns of the hospital staff were also monitored during the working day. They found that levels of exogenous compounds such as sevoflurane in both room air and exhaled air increased by 5 by the end of the working day, raising questions about when and where patients should be sampled for breath analysis to reduce to minimize the problem of such confounding factors. This correlates with the study by Castellanos et al. In 2016, they found sevoflurane in the breath of hospital staff, but not in the breath of staff outside the hospital. In 2018 Markar et al. sought to demonstrate the effect of changes in indoor air composition on breath analysis as part of their study to assess the diagnostic ability of exhaled air in esophageal cancer7. Using a steel counterlung and SIFT-MS during sampling, they identified eight volatile organic compounds in indoor air that varied significantly by sampling location. However, these VOCs were not included in their last breath VOC diagnostic model, so their impact was negated. In 2021, a study was conducted by Salman et al. to monitor VOC levels in three hospitals for 27 months. They identified 17 VOCs as seasonal discriminators and suggested that exhaled VOC concentrations above the critical level of 3 µg/m3 are considered unlikely secondary to background VOC pollution8.
In addition to setting threshold levels or outright excluding exogenous compounds, alternatives to eliminating this background variation include collecting paired room air samples simultaneously with exhaled air sampling so that any levels of VOCs present at high concentrations in the respirable room can be determined. extracted from exhaled air. Air 9 is subtracted from the level to provide an “alveolar gradient”. Therefore, a positive gradient indicates the presence of endogenous Compound 10. Another method is for participants to inhale “purified” air that is theoretically free of VOC11 pollutants. However, this is cumbersome, time consuming, and the equipment itself generates additional VOC pollutants. A study by Maurer et al. In 2014, participants breathing synthetic air reduced 39 VOCs but increased 29 VOCs compared to breathing indoor ambient air12. The use of synthetic/purified air also severely limits the portability of breath sampling equipment.
Ambient VOC levels are also expected to vary throughout the day, which may further affect the standardization and accuracy of breath sampling.
Advances in mass spectrometry, including thermal desorption coupled with gas chromatography and time-of-flight mass spectrometry (GC-TOF-MS), have also provided a more robust and reliable method for VOC analysis, capable of simultaneously detecting hundreds of VOCs, thus for deeper analysis. air in the room. This makes it possible to characterize in more detail the composition of the ambient air in the room and how large samples change with place and time.
The main objective of this study was to determine the varying levels of volatile organic compounds in indoor ambient air at common sampling sites in the hospital environment and how this affects exhaled air sampling. A secondary objective was to determine whether there were significant diurnal or geographic variations in the distribution of VOCs in indoor ambient air.
Breath samples, as well as corresponding indoor air samples, were collected in the morning from five different locations and analyzed with GC-TOF-MS. A total of 113 VOCs were detected and extracted from the chromatogram. The repeated measurements were convolved with the mean before a principal component analysis (PCA) of the extracted and normalized peak areas was performed to identify and remove outliers. Supervised analysis through partial least squares—discriminant analysis (PLS-DA) was then able to show a clear separation between breath and room air samples (R2Y = 0.97, Q2Y = 0.96, p < 0.001) (Fig. 1). Supervised analysis through partial least squares—discriminant analysis (PLS-DA) was then able to show a clear separation between breath and room air samples (R2Y = 0.97, Q2Y = 0.96, p < 0.001) (Fig. 1). Затем контролируемый анализ с помощью частичного дискриминантного анализа методом наименьших квадратов (PLS-DA) смог показать четкое разделение между образцами дыхания и комнатного воздуха (R2Y = 0,97, Q2Y = 0,96, p <0,001) (рис. 1). Then controlled analysis with partial least squares discriminant analysis (PLS-DA) was able to show a clear separation between breath and room air samples (R2Y=0.97, Q2Y=0.96, p<0.001) (Figure 1).通过偏最小二乘法进行监督分析——判别分析(PLS-DA) 然后能够显示呼吸和室内空气样本之间的明显分离（R2Y = 0.97，Q2Y = 0.96，p < 0.001）（图1）。通过 偏 最 小 二乘法 进行 监督 分析 分析 判别 判别 分析 分析 (PLS-DA) 然后 能够 显示 呼吸 室内 空气 样本 的 明显 （（（（（（（（， ， q2y = 0.96 ， p <0.001） （1）。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。。 Контролируемый анализ с помощью частичного дискриминантного анализа методом наименьших квадратов (PLS-DA) затем смог показать четкое разделение между образцами дыхания и воздуха в помещении (R2Y = 0,97, Q2Y = 0,96, p <0,001) (рис. 1). Controlled analysis with partial least squares discriminant analysis (PLS-DA) was then able to show a clear separation between breath and indoor air samples (R2Y = 0.97, Q2Y = 0.96, p < 0.001) (Figure 1). Group separation was driven by 62 different VOCs, with a variable importance projection (VIP) score > 1. A complete list of the VOCs characterizing each sample type and their respective VIP scores can be found in Supplementary Table 1. Group separation was driven by 62 different VOCs, with a variable importance projection (VIP) score > 1. A complete list of the VOCs characterizing each sample type and their respective VIP scores can be found in Supplementary Table 1. Разделение на группы было обусловлено 62 различными VOC с оценкой проекции переменной важности (VIP) > 1. Полный список VOC, характеризующих каждый тип образца, и их соответствующие оценки VIP можно найти в дополнительной таблице 1. Grouping was driven by 62 different VOCs with a Variable Importance Projection (VIP) score > 1. A complete list of VOCs characterizing each sample type and their respective VIP scores can be found in Supplementary Table 1.组分离由62 种不同的VOC 驱动，变量重要性投影(VIP) 分数> 1。组分离由62 种不同的VOC 驱动，变量重要性投影(VIP) 分数> 1。 Разделение групп было обусловлено 62 различными ЛОС с оценкой проекции переменной важности (VIP) > 1. Group separation was driven by 62 different VOCs with a variable importance projection score (VIP) > 1. A complete list of VOCs characterizing each sample type and their respective VIP scores can be found in Supplementary Table 1.
Breathing and indoor air show different distributions of volatile organic compounds. Supervised analysis with PLS-DA showed a clear separation between breath and room air VOCs profiles collected during the morning (R2Y = 0.97, Q2Y = 0.96, p < 0.001). Supervised analysis with PLS-DA showed a clear separation between breath and room air VOCs profiles collected during the morning (R2Y = 0.97, Q2Y = 0.96, p < 0.001). Контролируемый анализ с помощью PLS-DA показал четкое разделение между профилями летучих органических соединений в выдыхаемом воздухе и воздухе в помещении, собранными утром (R2Y = 0,97, Q2Y = 0,96, p <0,001). PLS-DA controlled analysis showed a clear separation between the exhaled and indoor air volatile organic compound profiles collected in the morning (R2Y=0.97, Q2Y=0.96, p<0.001).使用PLS-DA 进行的监督分析显示，早上收集的呼吸和室内空气VOC 曲线明显分离（R2Y = 0.97，Q2Y = 0.96，p < 0.001）。使用 PLS-DA Контролируемый анализ с использованием PLS-DA показал четкое разделение профилей ЛОС дыхания и воздуха в помещении, собранных утром (R2Y = 0,97, Q2Y = 0,96, p <0,001). Controlled analysis using PLS-DA showed a clear separation of the VOC profiles of breath and indoor air collected in the morning (R2Y=0.97, Q2Y=0.96, p<0.001). Repeated measurements were reduced to the mean before the model was built. Ellipses show 95% confidence intervals and centroids of the asterisk group.
Differences in the distribution of volatile organic compounds in indoor air in the morning and afternoon were investigated using PLS-DA. The model identified significant separation between the two timepoints (R2Y = 0.46, Q2Y = 0.22, p < 0.001) (Fig. 2). The model identified significant separation between the two timepoints (R2Y = 0.46, Q2Y = 0.22, p < 0.001) (Fig. 2). Модель выявила значительное разделение между двумя временными точками (R2Y = 0,46, Q2Y = 0,22, p <0,001) (рис. 2). The model revealed a significant separation between the two time points (R2Y = 0.46, Q2Y = 0.22, p < 0.001) (Figure 2).该模型确定了两个时间点之间的显着分离（R2Y = 0.46，Q2Y = 0.22，p < 0.001）（图2）。该模型确定了两个时间点之间的显着分离（R2Y = 0.46，Q2Y = 0.22，p < 0.001）（图2）。 Модель выявила значительное разделение между двумя временными точками (R2Y = 0,46, Q2Y = 0,22, p <0,001) (рис. 2). The model revealed a significant separation between the two time points (R2Y = 0.46, Q2Y = 0.22, p < 0.001) (Figure 2). This was driven by 47 VOCs with a VIP score > 1. VOCs with the highest VIP score characterizing morning samples included multiple branched alkanes, oxalic acid and hexacosane, while afternoon samples presented more 1-propanol, phenol, propanoic acid, 2-methyl-, 2-ethyl-3-hydroxyhexyl ester, isoprene and nonanal. This was driven by 47 VOCs with a VIP score > 1. VOCs with the highest VIP score characterizing morning samples included multiple branched alkanes, oxalic acid and hexacosane, while afternoon samples presented more 1-propanol, phenol, propanoic acid, 2-methyl- , 2-ethyl-3-hydroxyhexyl ester, isoprene and nonanal. Это было обусловлено наличием 47 летучих органических соединений с оценкой VIP > 1. ЛОС с самой высокой оценкой VIP, характеризующей утренние образцы, включали несколько разветвленных алканов, щавелевую кислоту и гексакозан, в то время как дневные образцы содержали больше 1-пропанола, фенола, пропановой кислоты, 2-метил- , 2-этил-3-гидроксигексиловый эфир, изопрен и нонаналь. This was due to the presence of 47 volatile organic compounds with a VIP score > 1. The VOCs with the highest VIP score for morning samples included several branched alkanes, oxalic acid, and hexacosane, while daytime samples contained more 1-propanol, phenol, propanoic acids, 2-methyl-, 2-ethyl-3-hydroxyhexyl ether, isoprene and nonanal.这是由47 种VIP 评分> 1 的VOC 驱动的。这是由47 种VIP 评分> 1 的VOC 驱动的。 Этому способствуют 47 VOC с оценкой VIP > 1. This is facilitated by 47 VOCs with a VIP score > 1. The highest VIP-rated VOCs in the morning sample included various branched alkanes, oxalic acid, and hexadecane, while the afternoon sample contained more 1-propanol, phenol, propionic acid, 2-methyl-, 2-ethyl-3-hydroxyhexyl. ester, isoprene and nonanal. A complete list of volatile organic compounds (VOCs) that characterize daily changes in indoor air composition can be found in Supplementary Table 2.
The distribution of VOCs in indoor air varies throughout the day. Supervised analysis with PLS-DA showed separation between room air samples collected during the morning or during the afternoon (R2Y = 0.46, Q2Y = 0.22, p < 0.001). Supervised analysis with PLS-DA showed separation between room air samples collected during the morning or during the afternoon (R2Y = 0.46, Q2Y = 0.22, p < 0.001). Контролируемый анализ с помощью PLS-DA показал разделение между пробами воздуха в помещении, собранными утром и днем (R2Y = 0,46, Q2Y = 0,22, p < 0,001). Controlled analysis with PLS-DA showed separation between indoor air samples collected in the morning and afternoon (R2Y = 0.46, Q2Y = 0.22, p < 0.001).使用PLS-DA 进行的监督分析显示，早上或下午收集的室内空气样本之间存在分离（R2Y = 0.46，Q2Y = 0.22，p < 0.001）。使用 PLS-DA Анализ эпиднадзора с использованием PLS-DA показал разделение проб воздуха внутри помещений, собранных утром или днем (R2Y = 0,46, Q2Y = 0,22, p < 0,001). Surveillance analysis using PLS-DA showed a separation of indoor air samples collected in the morning or afternoon (R2Y = 0.46, Q2Y = 0.22, p < 0.001). Ellipses show 95% confidence intervals and centroids of the asterisk group.
Samples were collected from five different locations at St Mary’s Hospital in London: an endoscopy room, a clinical research room, an operating room complex, an outpatient clinic and a mass spectrometry laboratory. Our research team regularly uses these locations for patient recruitment and breath collection. As before, indoor air was collected in the morning and afternoon, and exhaled air samples were collected only in the morning. PCA highlighted a separation of room air samples by location through permutational multivariate analysis of variance (PERMANOVA, R2 = 0.16, p < 0.001) (Fig. 3a). PCA highlighted a separation of room air samples by location through permutational multivariate analysis of variance (PERMANOVA, R2 = 0.16, p < 0.001) (Fig. 3a). PCA выявил разделение проб комнатного воздуха по местоположению с помощью перестановочного многомерного дисперсионного анализа (PERMANOVA, R2 = 0,16, p <0,001) (рис. 3а). PCA revealed separation of room air samples by location using permutational multivariate analysis of variance (PERMANOVA, R2 = 0.16, p < 0.001) (Fig. 3a). PCA 通过置换多变量方差分析（PERMANOVA，R2 = 0.16，p < 0.001）强调了房间空气样本的位置分离（图3a）。 PCA PCA подчеркнул локальную сегрегацию проб комнатного воздуха с помощью перестановочного многомерного дисперсионного анализа (PERMANOVA, R2 = 0,16, p <0,001) (рис. 3а). PCA highlighted the local segregation of room air samples using permutational multivariate analysis of variance (PERMANOVA, R2 = 0.16, p < 0.001) (Fig. 3a). Therefore, paired PLS-DA models were created in which each location is compared to all other locations to determine feature signatures. All models were significant and VOCs with VIP score > 1 were extracted with respective loading to identify group contribution. All models were significant and VOCs with VIP score > 1 were extracted with respective loading to identify group contribution. Все модели были значимыми, и ЛОС с оценкой VIP > 1 были извлечены с соответствующей нагрузкой для определения группового вклада. All models were significant, and VOCs with a VIP score > 1 were extracted with appropriate loading to determine the group contribution.所有模型均显着，VIP 评分> 1 的VOC 被提取并分别加载以识别组贡献。所有模型均显着，VIP 评分> 1 的VOC Все модели были значимыми, и VOC с баллами VIP> 1 были извлечены и загружены отдельно для определения групповых вкладов. All models were significant and VOCs with VIP scores > 1 were extracted and uploaded separately to determine group contributions. Our results show that ambient air composition varies with location, and we have identified location-specific features using model consensus. The endoscopy unit is characterized by high levels of undecane, dodecane, benzonitrile and benzaldehyde. Samples from the Clinical Research Department (also known as the Liver Research Department) showed more alpha-pinene, diisopropyl phthalate, and 3-carene. The mixed air of the operating room is characterized by a higher content of branched decane, branched dodecane, branched tridecane, propionic acid, 2-methyl-, 2-ethyl-3-hydroxyhexyl ether, toluene and 2 – the presence of crotonaldehyde. The outpatient clinic (Paterson Building) has a higher content of 1-nonanol, vinyl lauryl ether, benzyl alcohol, ethanol, 2-phenoxy, naphthalene, 2-methoxy, isobutyl salicylate, tridecane, and branched chain tridecane. Finally, indoor air collected in the mass spectrometry laboratory showed more acetamide, 2’2’2-trifluoro-N-methyl-, pyridine, furan, 2-pentyl-, branched undecane, ethylbenzene, m-xylene, o- xylene, furfural and ethylanisate. Various levels of 3-carene were present in all five sites, suggesting that this VOC is a common contaminant with the highest observed levels in the clinical study area. A list of agreed VOCs sharing each position can be found in Supplementary Table 3. In addition, a univariate analysis was performed for each VOC of interest, and all positions were compared against each other using a pairwise Wilcoxon test followed by a Benjamini-Hochberg correction. The block plots for each VOC are presented in Supplementary Figure 1. Respiratory volatile organic compound curves appeared to be location-independent, as observed in PCA followed by PERMANOVA (p = 0.39) (Figure 3b). Additionally, pairwise PLS-DA models were generated between all the different location for the breath samples too, but no significant differences were identified (p > 0.05). Additionally, pairwise PLS-DA models were generated between all the different locations for the breath samples too, but no significant differences were identified (p > 0.05). Кроме того, парные модели PLS-DA также были созданы между всеми разными местоположениями образцов дыхания, но существенных различий выявлено не было (p > 0,05). In addition, paired PLS-DA models were also generated between all different breath sample locations, but no significant differences were found (p > 0.05).此外，在呼吸样本的所有不同位置之间也生成了成对PLS-DA 模型，但未发现显着差异(p > 0.05)。 PLS-DA 模型，但未发现显着差异(p > 0.05)。 Кроме того, парные модели PLS-DA также были сгенерированы между всеми различными местоположениями образцов дыхания, но существенных различий обнаружено не было (p > 0,05). In addition, paired PLS-DA models were also generated between all different breath sample locations, but no significant differences were found (p > 0.05).
Changes in ambient indoor air but not in exhaled air, VOC distribution differs depending on sampling site, unsupervised analysis using PCA shows separation between indoor air samples collected at different locations but not corresponding exhaled air samples. The asterisks denote the centroids of the group.
In this study, we analyzed the distribution of indoor air VOCs at five common breath sampling sites to gain a better understanding of the effect of background VOC levels on breath analysis.
Separation of indoor air samples was observed at all five different locations. With the exception of 3-carene, which was present in all areas studied, the separation was caused by different VOCs, giving each location a specific character. In the field of endoscopy evaluation, separation-inducing volatile organic compounds are mainly monoterpenes such as beta-pinene and alkanes such as dodecane, undecane and tridecane, which are commonly found in essential oils commonly used in cleaning products 13. Considering the frequency cleaning endoscopic devices, these VOCs are likely the result of frequent indoor cleaning processes. In clinical research laboratories, as in endoscopy, the separation is mainly due to monoterpenes such as alpha-pinene, but also probably from cleaning agents. In the complex operating room, the VOC signature consists mainly of branched alkanes. These compounds can be obtained from surgical instruments as they are rich in oils and lubricants14. In the surgical setting, typical VOCs include a range of alcohols: 1-nonanol, found in vegetable oils and cleaning products, and benzyl alcohol, found in perfumes and local anesthetics.15,16,17,18 VOCs in a mass spectrometry laboratory are very different from expected in other areas as this is the only non-clinical area assessed. While some monoterpenes are present, a more homogeneous group of compounds shares this area with other compounds (2,2,2-trifluoro-N-methyl-acetamide, pyridine, branched undecane, 2-pentylfuran, ethylbenzene, furfural, ethylanisate). ), orthoxylene, meta-xylene, isopropanol and 3-carene), including aromatic hydrocarbons and alcohols. Some of these VOCs may be secondary to chemicals used in the laboratory, which consists of seven mass spectrometry systems operating in TD and liquid injection modes.
With PLS-DA, a strong separation of indoor air and breath samples was observed, caused by 62 of the 113 detected VOCs. In indoor air, these VOCs are exogenous and include diisopropyl phthalate, benzophenone, acetophenone and benzyl alcohol, which are commonly used in plasticizers and fragrances19,20,21,22 the latter can be found in cleaning products16. The chemicals found in exhaled air are a mixture of endogenous and exogenous VOCs. Endogenous VOCs consist mainly of branched alkanes, which are byproducts of lipid peroxidation23, and isoprene, a byproduct of cholesterol synthesis24. Exogenous VOCs include monoterpenes such as beta-pinene and D-limonene, which can be traced back to citrus essential oils (also widely used in cleaning products) and food preservatives13,25. 1-Propanol can be either endogenous, resulting from the breakdown of amino acids, or exogenous, present in disinfectants26. Compared to breathing indoor air, higher levels of volatile organic compounds are found, some of which have been identified as possible disease biomarkers. Ethylbenzene has been shown to be a potential biomarker for a number of respiratory diseases, including lung cancer, COPD27 and pulmonary fibrosis28. Compared to patients without lung cancer, levels of N-dodecane and xylene have also been found at higher concentrations in patients with lung cancer29 and metacymol in patients with active ulcerative colitis30. Thus, even if indoor air differences do not affect the overall respiration profile, they can affect specific VOC levels, so monitoring indoor background air may still be important.
There was also a separation between indoor air samples collected in the morning and afternoon. The main features of morning samples are branched alkanes, which are often found exogenously in cleaning products and waxes31. This can be explained by the fact that all four clinical rooms included in this study were cleaned prior to room air sampling. All clinical areas are separated by different VOCs, so this separation cannot be attributed to cleaning. Compared to the morning samples, the afternoon samples generally showed higher levels of a mixture of alcohols, hydrocarbons, esters, ketones, and aldehydes. Both 1-propanol and phenol can be found in disinfectants26,32 which is expected given the regular cleaning of the entire clinical area throughout the day. Breath is collected only in the morning. This is due to many other factors that can affect the level of volatile organic compounds in exhaled air during the day, which cannot be controlled. This includes consumption of beverages and food33,34 and varying degrees of exercise35,36 prior to breath sampling.
VOC analysis remains at the forefront of non-invasive diagnostic development. Standardization of sampling remains a challenge, but our analysis conclusively showed that there were no significant differences between breath samples collected at different locations. In this study, we showed that the content of volatile organic compounds in the ambient indoor air depends on the location and time of day. However, our results also show that this does not significantly affect the distribution of volatile organic compounds in the exhaled air, suggesting that breath sampling can be performed at different locations without significantly affecting the results. Preference is given to including multiple sites and duplicating specimen collections over longer periods of time. Finally, the separation of indoor air from different locations and the lack of separation in exhaled air clearly shows that the sampling site does not significantly affect the composition of human breath. This is encouraging for breath analysis research as it removes a potential confounding factor in the standardization of breath data collection. Although all breath patterns from a single subject were a limitation of our study, it may reduce differences in other confounding factors that are influenced by human behavior. Single-disciplinary research projects have previously been used successfully in many studies37. However, further analysis is required to draw firm conclusions. Routine indoor air sampling is still recommended, along with breath sampling to rule out exogenous compounds and identify specific pollutants. We recommend eliminating isopropyl alcohol due to its prevalence in cleaning products, especially in healthcare settings. This study was limited by the number of breath samples collected at each site, and further work is required with a larger number of breath samples to confirm that the composition of human breath does not significantly affect the context in which the samples are found. In addition, relative humidity (RH) data were not collected, and while we acknowledge that differences in RH can affect VOC distribution, logistical challenges in both RH control and RH data collection are significant in large scale studies.
In conclusion, our study shows that VOCs in ambient indoor air vary by location and time, but this does not appear to be the case for breath samples. Due to the small sample size, it is not possible to draw definitive conclusions about the effect of indoor ambient air on breath sampling and further analysis is required, so it is recommended to take indoor air sampling during breathing to detect any potential contaminants, VOCs.
The experiment took place for 10 consecutive working days at St Mary’s Hospital in London in February 2020. Each day, two breath samples and four indoor air samples were taken from each of the five locations, for a total of 300 samples. All methods were performed in accordance with the relevant guidelines and regulations. The temperature of all five sampling zones was controlled at 25°C.
Five locations were selected for indoor air sampling: Mass Spectrometry Instrumentation Laboratory, Surgical Ambulatory, Operating Room, Evaluation Area, Endoscopic Evaluation Area, and Clinical Study Room. Each region was chosen because our research team often uses them to recruit participants for breath analysis.
Room air was sampled through inert coated Tenax TA/Carbograph thermal desorption (TD) tubes (Markes International Ltd, Llantrisan, UK) at 250 ml/min for 2 minutes using an air sampling pump from SKC Ltd., total Difficulty Apply 500 ml of ambient room air to each TD tube. The tubes were then sealed with brass caps for transport back to the mass spectrometry laboratory. Indoor air samples were taken in turn at each location every day from 9:00 to 11:00 and again from 15:00 to 17:00. Samples were taken in duplicate.
Breath samples were collected from individual subjects subjected to indoor air sampling. The breath sampling process was performed as per the protocol approved by the NHS Health Research Authority—London—Camden & Kings Cross Research Ethics Committee (reference 14/LO/1136). The breath sampling process was performed as per the protocol approved by the NHS Health Research Authority—London—Camden & Kings Cross Research Ethics Committee (reference 14/LO/1136). Процесс отбора проб дыхания проводился в соответствии с протоколом, одобренным Управлением медицинских исследований NHS — Лондон — Комитет по этике исследований Camden & Kings Cross (ссылка 14/LO/1136). The breath sampling process was carried out in accordance with the protocol approved by the NHS Medical Research Authority – London – Camden & Kings Cross Research Ethics Committee (Ref. 14/LO/1136). The breath sampling procedure was carried out in accordance with protocols approved by the NHS-London-Camden Medical Research Agency and the King’s Cross Research Ethics Committee (ref 14/LO/1136). The researcher gave informed written consent. For normalization purposes, the researchers had not eaten or drunk since midnight the previous night. Breath was collected using a custom-made 1000 ml Nalophan™ (PET polyethylene terephthalate) disposable bag and a polypropylene syringe used as a sealed mouthpiece, as previously described by Belluomo et al. Nalofan has been shown to be an excellent respiratory storage medium due to its inertness and ability to provide compound stability for up to 12 hours38. Remaining in this position for at least 10 minutes, the examiner exhales into the sample bag during normal quiet breathing. After filling to the maximum volume, the bag is closed with a syringe plunger. As with indoor air sampling, use the SKC Ltd. air sampling pump for 10 minutes to draw air from the bag through the TD tube: connect a large diameter needle without filter to the air pump at the other end of the TD tube through the plastic tubes and SKC. Acupuncture the bag and inhale breaths at a rate of 250 ml/min through each TD tube for 2 min, loading a total of 500 ml breaths into each TD tube. The samples were again collected in duplicate to minimize sampling variability. Breaths are collected only in the morning.
TD tubes were cleaned using a TC-20 TD tube conditioner (Markes International Ltd, Llantrisant, UK) for 40 minutes at 330°C with a nitrogen flow of 50 ml/min. All samples were analyzed within 48 hours of collection using GC-TOF-MS. An Agilent Technologies 7890A GC was paired with a TD100-xr thermal desorption setup and a BenchTOF Select MS (Markes International Ltd, Llantrisan, UK). The TD tube was initially preflushed for 1 minute at a flow rate of 50 ml/min. Initial desorption was carried out at 250°C for 5 minutes with a helium flow of 50 ml/min to desorb VOCs onto a cold trap (Material Emissions, Markes International, Llantrisant, UK) in a split mode (1:10) at 25°C. Cold trap (secondary) desorption was performed at 250°C (with ballistic heating 60°C/s) for 3 min at a He flow rate of 5.7 ml/min, and the temperature of the flow path to the GC was continuously heated. up to 200 °С. The column was a Mega WAX-HT column (20 m×0.18 mm×0.18 μm, Chromalytic, Hampshire, USA). The column flow rate was set to 0.7 ml/min. The oven temperature was first set at 35° C. for 1.9 minutes, then raised to 240° C. (20° C./min, holding 2 minutes). The MS transmission line was maintained at 260°C and the ion source (70 eV electron impact) was maintained at 260°C. The MS analyzer was set to record from 30 to 597 m/s. Desorption in a cold trap (no TD tube) and desorption in a conditioned clean TD tube were performed at the beginning and end of each assay run to ensure that there were no carryover effects. The same blank analysis was performed immediately before and immediately after desorption of the breath samples to ensure that the samples could be analyzed continuously without adjusting the TD.
After visual inspection of the chromatograms, the raw data files were analyzed using Chromspace® (Sepsolve Analytical Ltd.). Compounds of interest were identified from representative breath and room air samples. Annotation based on VOC mass spectrum and retention index using the NIST 2017 mass spectrum library. Retention indices were calculated by analysing an alkane mixture (nC8-nC40, 500 μg/mL in dichloromethane, Merck, USA) 1 μL spiked onto three conditioned TD tubes via a calibration solution loading rig and analysed under the same TD-GC–MS conditions and from the raw compound list, only those with a reverse match factor > 800 were kept for analysis. Retention indices were calculated by analyzing an alkane mixture (nC8-nC40, 500 μg/mL in dichloromethane, Merck, USA) 1 μL spiked onto three conditioned TD tubes via a calibration solution loading rig and analyzed under the same TD-GC–MS conditions and from the raw compound list, only those with a reverse match factor > 800 were kept for analysis. Retention indices were calculated by analyzing 1 µl of a mixture of alkanes (nC8-nC40, 500 µg/ml in dichloromethane, Merck, USA) in three conditioned TD tubes using a calibration solution loading unit and analyzed under the same TD-GC-MS conditions. и из исходного списка соединений для анализа были оставлены только соединения с коэффициентом обратного совпадения > 800. and from the original list of compounds, only compounds with a reverse match coefficient > 800 were kept for analysis.通过分析烷烃混合物（nC8-nC40，500 μg/mL 在二氯甲烷中，Merck，USA）计算保留指数，通过校准溶液加载装置将1 μL 加标到三个调节过的TD 管上，并在相同的TD-GC-MS 条件下进行分析并且从原始化合物列表中，仅保留反向匹配因子> 800 的化合物进行分析。通过 分析 烷烃 （（nc8-nc40，500 μg/ml 在 中 ， ， merck ， USA） 保留 指数 ， 通过 校准 加载 装置 将 1 μl 到 三 调节 过 的 的 管 ， 并 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在 在800 的化合物进行分析。 Retention indices were calculated by analyzing a mixture of alkanes (nC8-nC40, 500 μg/ml in dichloromethane, Merck, USA), 1 μl was added to three conditioned TD tubes by calibrating the solution loader and added there. выполненных в тех же условиях TD-GC-MS и из исходного списка соединений, для анализа были оставлены только соединения с коэффициентом обратного соответствия > 800. performed under the same TD-GC-MS conditions and from the original compound list, only compounds with an inverse fit factor > 800 were retained for analysis. Oxygen, argon, carbon dioxide and siloxanes are also removed. Finally, any compounds with a signal to noise ratio < 3 were also excluded. Finally, any compounds with a signal to noise ratio < 3 were also excluded. Наконец, любые соединения с отношением сигнал/шум <3 также были исключены. Finally, any compounds with a signal-to-noise ratio <3 were also excluded.最后，还排除了信噪比< 3 的任何化合物。最后，还排除了信噪比< 3 的任何化合物。 Наконец, любые соединения с отношением сигнал/шум <3 также были исключены. Finally, any compounds with a signal-to-noise ratio <3 were also excluded. The relative abundance of each compound was then extracted from all data files using the resulting compound list. Compared to NIST 2017, 117 compounds have been identified in breath samples. Picking was performed using MATLAB R2018b software (version 9.5) and Gavin Beta 3.0. After further examination of the data, 4 more compounds were excluded by visual inspection of the chromatograms, leaving 113 compounds to be included in the subsequent analysis. An abundance of these compounds were recovered from all 294 samples that were successfully processed. Six samples were removed due to poor data quality (leaky TD tubes). In the remaining datasets, Pearson’s one-sided correlations were calculated among 113 VOCs in repeated measurements samples to assess reproducibility. The correlation coefficient was 0.990 ± 0.016, and the p value was 2.00 × 10–46 ± 2.41 × 10–45 (arithmetic mean ± standard deviation).
All statistical analyzes were performed on R version 4.0.2 (R Foundation for Statistical Computing, Vienna, Austria). The data and code used to analyze and generate the data is publicly available on GitHub (https://github.com/simonezuffa/Manuscript_Breath). The integrated peaks were first log-transformed and then normalized using total area normalization. Samples with repeated measurements were rolled up to the mean value. The “ropls” and “mixOmics” packages are used to create unsupervised PCA models and supervised PLS-DA models. PCA allows you to identify 9 sample outliers. The primary breath sample was grouped with the room air sample and was therefore considered an empty tube due to sampling error. The remaining 8 samples are room air samples containing 1,1′-biphenyl, 3-methyl. Further testing showed that all 8 samples had significantly lower VOC production compared to the other samples, suggesting that these emissions were caused by human error in loading the tubes. Location separation was tested in PCA using PERMANOVA from a vegan package. PERMANOVA allows you to identify the division of groups based on centroids. This method has previously been used in similar metabolomic studies39,40,41. The ropls package is used to evaluate the significance of PLS-DA models using random seven-fold cross-validation and 999 permutations. Compounds with a variable importance projection (VIP) score > 1 were considered relevant for the classification and retained as significant. Compounds with a variable importance projection (VIP) score > 1 were considered relevant for the classification and retained as significant. Соединения с показателем проекции переменной важности (VIP) > 1 считались подходящими для классификации и сохранялись как значимые. Compounds with a variable importance projection score (VIP) > 1 were considered eligible for classification and were retained as significant.具有可变重要性投影(VIP) 分数> 1 的化合物被认为与分类相关并保留为显着。具有可变重要性投影(VIP) 分数> 1 Соединения с оценкой переменной важности (VIP) > 1 считались подходящими для классификации и оставались значимыми. Compounds with a score of variable importance (VIP) > 1 were considered eligible for classification and remained significant. Loads from the PLS-DA model were also extracted to determine group contributions. The VOCs for a particular location are determined based on the consensus of paired PLS-DA models. To do so, all locations VOCs profiles were tested against each other and if a VOC with VIP > 1 was constantly significant in the models and attributed to the same location, it was then considered location specific. To do so, all locations VOCs profiles were tested against each other and if a VOC with VIP > 1 was constantly significant in the models and attributed to the same location, it was then considered location specific. Для этого профили ЛОС всех местоположений были проверены друг против друга, и если ЛОС с VIP> 1 был постоянно значимым в моделях и относился к одному и тому же месту, тогда он считался специфичным для местоположения. To do this, the VOC profiles of all locations were tested against each other, and if a VOC with VIP > 1 was consistently significant in the models and referred to the same location, then it was considered location-specific.为此，对所有位置的VOC 配置文件进行了相互测试，如果VIP > 1 的VOC 在模型中始终显着并归因于同一位置，则将其视为特定位置。为 此 ， 对 所有 的 的 voc 配置 文件 了 相互 测试 ， 如果 vip> 1 的 voc 在 中 始终 显着 并 归因 于 一 位置 ， 将 其 视为 特定。。。 位置 位置 位置 位置 位置 位置 位置 位置 位置 位置 位置 位置С этой целью профили ЛОС во всех местоположениях были сопоставлены друг с другом, и ЛОС с VIP> 1 считался зависящим от местоположения, если он был постоянно значимым в модели и относился к одному и тому же местоположению. To this end, VOC profiles at all locations were compared with each other, and a VOC with VIP > 1 was considered location dependent if it was consistently significant in the model and referred to the same location. Comparison of breath and indoor air samples was carried out only for samples taken in the morning, since no breath samples were taken in the afternoon. The Wilcoxon test was used for univariate analysis, and the false discovery rate was calculated using the Benjamini-Hochberg correction.
The datasets generated and analyzed during the current study are available from the respective authors upon reasonable request.
Oman, A. et al. Human volatile substances: Volatile organic compounds (VOCs) in exhaled air, skin secretions, urine, faeces and saliva. J. Breath res. 8(3), 034001 (2014).
Belluomo, I. et al. Selective ion current tube mass spectrometry for targeted analysis of volatile organic compounds in human breath. National protocol. 16(7), 3419–3438 (2021).
Hanna, GB, Boshier, PR, Markar, SR & Romano, A. Accuracy and methodologic challenges of volatile organic compound-based exhaled breath tests for cancer diagnosis. Hanna, GB, Boshier, PR, Markar, SR & Romano, A. Accuracy and methodological challenges of volatile organic compound-based exhaled breath tests for cancer diagnosis. Khanna, GB, Boshire, PR, Markar, SR. and Romano, A. Accuracy and methodological issues of volatile organic compound-based exhaust air tests for cancer diagnosis. Hanna, GB, Boshier, PR, Markar, SR & Romano, A. 基于挥发性有机化合物的呼出气测试在癌症诊断中的准确性和方法学挑战。 Hanna, GB, Boshier, PR, Markar, SR & Romano, A. Accuracy and methodological challenges in cancer diagnosis based on volatile organic compounds. Khanna, GB, Boshire, PR, Markar, SR. and Romano, A. Accuracy and methodological issues of volatile organic compound breath testing in cancer diagnosis. JAMA Oncol. 5(1), e182815 (2019).
Boshier, PR, Cushnir, JR, Priest, OH, Marczin, N. & Hanna, GB Variation in the levels of volatile trace gases within three hospital environments: Implications for clinical breath testing. Boshier, PR, Cushnir, JR, Priest, OH, Marczin, N. & Hanna, GB Variation in the levels of volatile trace gases within three hospital environments: Implications for clinical breath testing. Boshear, PR, Kushnir, JR, Priest, OH, Marchin, N. and Khanna, GB. Differences in levels of volatile trace gases in three hospital settings: significance for clinical breath testing. Boshier, PR, Cushnir, JR, Priest, OH, Marczin, N. & Hanna, GB 三种医院环境中挥发性微量气体水平的变化：对临床呼气测试的影响。 Boshier, PR, Cushnir, JR, Priest, OH, Marczin, N. & Hanna, GB Boshear, PR, Kushnir, JR, Priest, OH, Marchin, N. and Khanna, GB. Changes in levels of volatile trace gases in three hospital settings: significance for clinical breath testing. J. Religious Res. 4(3), 031001 (2010).
Trefz, P. et al. Real-time, continuous monitoring of respiratory gases in clinical settings using time-of-flight mass spectrometry of the proton transfer reaction. anus. Chemical. 85(21), 10321-10329 (2013).
Castellanos, M., Xifra, G., Fernández-Real, JM & Sánchez, JM Breath gas concentrations mirror exposure to sevoflurane and isopropyl alcohol in hospital environments in non-occupational conditions. Castellanos, M., Xifra, G., Fernández-Real, JM & Sánchez, JM Breath gas concentrations mirror exposure to sevoflurane and isopropyl alcohol in hospital environments in non-occupational conditions. Castellanos, M., Xifra, G., Fernandez-Real, J. M. and Sanchez, J. M. Exhaled gas concentrations reflect exposure to sevoflurane and isopropyl alcohol in a hospital setting in a non-occupational setting. Castellanos, M., Xifra, G., Fernández-Real, JM & Sánchez, JM 呼吸气体浓度反映了在非职业条件下的医院环境中暴露于七氟醚和异丙醇。 Castellanos, M., Xifra, G., Fernández-Real, JM & Sánchez, JM Castellanos, M., Xifra, G., Fernandez-Real, J.M. and Sanchez, J.M. Airway gas concentrations reflect exposure to sevoflurane and isopropanol in a hospital setting in a lay setting. J. Breath res. 10(1), 016001 (2016).
Markar S.R. et al. Evaluate non-invasive breath tests for the diagnosis of cancer of the esophagus and stomach. JAMA Oncol. 4(7), 970-976 (2018).
Salman, D. et al. Variability of volatile organic compounds in indoor air in a clinical setting. J. Breath res. 16(1), 016005 (2021).
Phillips, M. et al. Volatile breath markers of breast cancer. Breast J. 9 (3), 184–191 (2003).
Phillips, M., Greenberg, J. & Sabas, M. Alveolar gradient of pentane in normal human breath. Phillips, M., Greenberg, J. & Sabas, M. Alveolar gradient of pentane in normal human breath. Phillips M, Greenberg J and Sabas M. Alveolar pentane gradient in normal human breathing. Phillips, M., Greenberg, J. & Sabas, M. 正常人呼吸中戊烷的肺泡梯度。 Phillips, M., Greenberg, J. & Sabas, M. Phillips M, Greenberg J and Sabas M. Alveolar pentane gradients in normal human breathing. free radicals. storage tank. 20(5), 333–337 (1994).
Harshman S.V. et al. Characterization of standardized breath sampling for offline use in the field. J. Breath res. 14(1), 016009 (2019).
Maurer, F. et al. Flush ambient air pollutants for exhaled air measurement. J. Breath res. 8(2), 027107 (2014).
Salehi, B. et al. The therapeutic potential of alpha- and beta-pinene: nature’s miraculous gift. Biomolecules 9 (11), 738 (2019).
CompTox chemical information panel – benzyl alcohol. https://comptox.epa.gov/dashboard/dsstoxdb/results?search=DTXSID5020152#chemical-functional-use (accessed 22 September 2021).
Alfa Aesar – L03292 Benzyl alcohol, 99%. https://www.alfa.com/en/catalog/L03292/ (accessed 22 September 2021).
Good Scents Company – Benzyl Alcohol. http://www.thegoodscentscompany.com/data/rw1001652.html (accessed 22 September 2021).
The CompTox chemical panel is diisopropyl phthalate. https://comptox.epa.gov/dashboard/dsstoxdb/results?search=DTXSID2040731 (accessed 22 September 2021).
Humans, IARC Working Group on Carcinogenic Risk Assessment. Benzophenone. : International Agency for Research on Cancer (2013).
Good Scents Company – Acetophenone. http://www.thegoodscentscompany.com/data/rw1000131.html#tooccur (accessed 22 September 2021).
Van Gossum, A. & Decuyper, J. Breath alkanes as an index of lipid peroxidation. Van Gossum, A. & Decuyper, J. Breath alkanes as an index of lipid peroxidation. Van Gossum, A. and Dekuyper, J. Alkane respiration as an indicator of lipid peroxidation. Van Gossum, A. & Decuyper, J. Breath 烷烃作为脂质过氧化的指标。 Van Gossum, A. & Decuyper, J. Breath alkanes as an indicator of 脂质过过化的的剧情。 Van Gossum, A. and Dekuyper, J. Alkane respiration as an indicator of lipid peroxidation. EURO. country Journal 2(8), 787–791 (1989).
Salerno-Kennedy, R. & Cashman, KD Potential applications of breath isoprene as a biomarker in modern medicine: A concise overview. Salerno-Kennedy, R. & Cashman, KD Potential applications of breath isoprene as a biomarker in modern medicine: A concise overview. Salerno-Kennedy, R. & Cashman, K.D. Possible applications of isoprene in respiration as a biomarker in modern medicine: a brief review. Salerno-Kennedy, R. & Cashman, KD 呼吸异戊二烯作为现代医学生物标志物的潜在应用：简明概述。 Salerno-Kennedy, R. & Cashman, KD Salerno-Kennedy, R. and Cashman, K. D. Potential applications of respiratory isoprene as a biomarker for modern medicine: a brief review. Wien Klin Wochenschr 117 (5–6), 180–186 (2005).
Kureas M. et al. Targeted analysis of volatile organic compounds in exhaled air is used to differentiate lung cancer from other lung diseases and in healthy people. Metabolites 10(8), 317 (2020).
Post time: Sep-28-2022