Adverse effects of COVID-protective face-masks and wearing durations onto respiratory-haemodynamic physiology and exhaled breath constituents


 While protecting against the coronavirus transmission, face-masks may have adverse effects on respiratory-haemodynamic parameters. We investigated immediate and progressive effects of FFP2 and surgical masks on exhaled breath constituents and physiological attributes in 30 healthy volunteers at rest. We continuously monitored exhaled breath profiles in the mask space in elderly (age: 60–80 years) and adults (age: 20–60 years) over a period of 30 min by high-resolution real-time mass-spectrometry (PTR-ToF-MS). Peripheral oxygen saturation, respiratory- and haemodynamic parameters were measured (non-invasively) continuously in parallel. Profound and consistent decrease in SpO2 and increase in pET-CO2 indicates ascending deoxygenation and inadequate ventilation in subjects. Cardiac output and MAP changed as secondary. Exhalation of blood-borne volatile metabolites mirrored behaviour of cardiac output, MAP, SpO2, respiratory rate and pET-CO2. FFP2 masks affected more pronouncedly than surgical masks. Elderly cohort was more vulnerable to those effects. Exhaled humidity increased and exhaled oxygen decreased significantly over time. Breath profiles of endogenous aldehydes, hemiterpene, organosulfur, short-chain fatty acids, alcohols and ketone indicated cross-talks between physio-metabolic effects such as hypoxia, oxidative stress, hypoventilation, compartmental vasoconstriction, altered systemic bacterial activity and energy homeostasis. Concentrations of exogenous VOCs such as aromatics, nitrile and monoterpene depicted compartmental storage and washout. Breathomics allows unique physio-metabolic insights into side effects of face-mask wearing. Mask induced deoxygenation, oxidative stress, CO2 rebreathing, vasoconstriction and blood pressure fluctuations in elderly were clinically concerning (as leading towards hypoxia and hypoventilation). Intelligible global-pandemic policies should reconsider the type and wearing durations of recommended face-masks, based upon age and/or cardio-pulmonary conditions.


Introduction
Since early 2020, face masks have gradually become an integral part of our new-normal lifestyle in order to protect us from air/breath borne transmission of SARS-CoV-2 infection [1,2]. During the second wave of the pandemic, use of surgical and/or FFP2/N95/KN95 masks turned out as strictly mandatory attributes while in public. National and/or global policy makers have recommended even adapting FFP3 masks for further protection considering the emerging CoV-2 variants [3,4]. In Germany, government has recommended use of FFP2 masks up to 75 min at a stretch and use of surgical mask throughout one's presence in the public and the same guidelines are applied for school attending children as well as for elderly (aged > 60 years).
While protecting us from the COVID-19 transmission, masks are inducing variable side effects on our cardiorespiratory physiology [5][6][7], broncho-pulmonary gas-exchange[8] and in vivo metabolic processes. Studies have shown variable effects of surgical masks on cardiopulmonary parameters, O 2 -CO 2 homeostasis, blood pH and thermoregulation etc [9]. Studies have also shown that conditions such as resistive breathing and/or hypoxia driven hyperventilation, respiratory alkalosis and increased oxidative stress can cause immediate immune suppression [10,11] as well as may lead to metabolic alkalosis [12]. A recent pilot observation of mask driven cardiopulmonary effects in 12 healthy subjects (age: 40.8 ± 12.4 years) at rest and during exercise, are interpreted as signi cant but modest [7]. However, it could not offer an insight into such effects in elderly subjects and also into metabolic changes at the downstream level. In order to understand the immediate physio-metabolic effects of face masks, we need to monitor continuous changes in metabolic markers along with simultaneous changes in respiratory and haemodynamic parameters.
In this context, high-resolution mass-spectrometry based real-time analysis of exhaled volatile organic compounds (VOCs) could offer a unique insight into body's immediate physiological [13][14][15][16][17] and metabolic [18,19] status. Continuous and breath-resolved measurements allow us to track changes in exhaled metabolic markers over the durations of mask use. Combining VOC pro ling with simultaneous pulseoximetry, capnography and haemodynamic monitoring could enable an unconventional understanding of clinically relevant effects of face masks.
We applied online high-resolution mass-spectrometry (i.e. proton transfer reaction -time-of-ight -mass-spectrometry / PTR-ToF-MS) based breathomics in parallel to non-invasive measurements of SpO 2 , respiratory rate, pET-CO 2 , exhaled humidity and oxygen, cardiac output, stroke volume, heart rate and blood pressure etc. The physio-metabolic side effects of FFP2 and surgical masks over 15-30 min of use on healthy human subjects aged between 20-80 years will be addressed in detail. Effects from both masks will be compared upon wearing duration and age in order to realize any further needs for reforming the present global-pandemic policies.

Human subjects:
All experiments were conducted according to the amended Declaration of Helsinki guidelines and signed informed consent from 30 subjects (aged between 20 -80 years) were obtained (Approval number: A2021-0012 -issued by the Institutional Ethics Committee of University Medicine Rostock, Germany) prior to inclusion. Subjects were not suffering from any acute diseases/health condition and were not undertaking any special diet and/or medication. Among the subjects above the age of 60, three had mild COPD (one male and two female) and two had chronic bronchitis (female) in the past.

Determination of sample size:
We applied the analysis of variance (ANOVA) test for calculation of sample size. For a minimum detectable difference in mean substance intensities of 450 cps, a standard deviation of 300 was estimated. To attain an alpha value of 0.005 and a test power of 0.99 experimental groups, while considering a population of 100,000, the sample size resulted at 26 (with minimal group size of at least 13 each). In this study, we have included 30 subjects for analysis in order to detect even less than 5% differences in exhaled VOCs up to low parts per trillion by volume (pptV) levels.

Assignment of groups:
We have divided the study population in two groups; namely: adults (<60 years of age) and elderly (>60 years of age). Anthropometric data were con rmed by participants during inclusion and are presented in Table 1.

Page 4/16
Experimental setup: Three devices were synchronized for real-time measurements of several parameters simultaneously (Fig. 1). Continuous monitoring of breath VOCs, exhaled abundances of O 2 , CO 2 and humidity via PTR-ToF-MS, non-invasive measurements of haemodynamic parameters via volume clamp method, SpO 2 monitoring via pulseoximetry. Main-stream capnography (for pET-CO 2 ) was performed immediately before and after the mask use. Data acquisition was initiated in parallel.
Breath sampling protocol: Volunteers rested by sitting on a chair for at least 10 min before actual sampling. Each participant were instructed to maintained the sitting posture [20] and then wore a face mask to breathe orally. They spontaneously inhaled and exhaled only via mouth [21].
The transfer-line of PTR-ToF-MS was connected (via PEEK nger-tight ttings) to a PEEK extension line (i.e. 30 cm long, with outer diameter of 1 mm and inner diameter of 0.75 mm) in order to directly sample breath-resolved VOCs from the mask dead space (Fig. 1). The PTR transfer line was xed (via metal clamps) at the back of subject's head (at a level below the left/right earlobe). The PEEK line was placed along the subject's right/left cheek (following the maxillary line) and was inserted within the mask dead space till the front of subject's lips. The tip of this sampling line was cased within a conical PEEK ferrule in order to avoid any unwanted contact with mask surface or with subject's lips. These extension lines were sterilized for reuse.
In each volunteer, measurements with two different masks (viz. FFP2 and surgical) were conducted on two consecutive days and at the same time. Adults were measured for 30 min and elderly subjects were measured for 15 min. The measurements in elderly subjects were stopped once they attained a SpO 2 level <94%.

PTR-ToF-MS measurements of breath VOCs:
Breath VOCs were measured continuously via a PTR-ToF-MS 8000 (Ionicon Analytik GmbH, Innsbruck, Austria) and with pre-optimized experimental conditions [15,22], i.e. continuous side-stream mode of sampling via a 6m long heated (at 75°C) silico-steel transfer-line connected to a sterile mouthpiece. A continuous sampling ow of 20 ml/min was applied and the time resolution of the PTR-ToF-MS measurements was 200 ms. Thus, data points were generated after every 200 ms and on each data point hundreds of compounds were measured at their trace abundances (in both expiratory-and room air). The ion source current was set to 4 mA and the H 2 O ow was set to 6 ml/min. Drift tube temperature was set to 75°C, voltage was 610 V and the pressure was 2.3 mbar. The resulting E/N ratio was 139 Td. After every minute a new data le was recorded automatically and the mass scale was recalibrated after each run (60 s).
VOC data processing: VOCs were measured in counts per seconds (cps) and corresponding intensities were normalised onto primary ion (H 3 O + ) counts. Raw data was processed via PTR-MS viewer software (version 3.4). As PTR-MS continuously records both exhaled breath and inhaled room-air, the 'breath tracker' algorithm (based on Matlab version 7.12.0.635, R2011a) was applied to identify expiratory and inspiratory phases [15]. Here, acetone was used as the tracker mass as it is an endogenous substance, which has signi cantly higher signal intensity in expiration than in inhalation. As the high mass resolution of PTR-ToF-MS (4000-5000 Δm/m) can assign volatiles upon their measured mass and corresponding sum formula with high precision [21], compound names are used while discussing results. VOCs were quanti ed via multi-component mixture of standard reference substances. Quanti cation process under adapted sample humidity (as in exhaled breath) using a liquid calibration unit (LCU, Ionicon Analytik GmbH, Innsbruck, Austria) is our pre-established state-of-the-art [23].

Selection of VOCs for analysis:
Here we considered compounds with expiratory abundances signi cantly above the inspiratory/room-air abundance. Out of those markers 32 substances were selected. These VOCs are well-known breath markers in clinical breathomics and re ect different origins, physicochemical characters and dependencies on physiology, metabolism, pathology, therapy and lifestyle/habits [18,19,21,24,25]. None of these VOCs were contributed from the applied masks as we examined the mask emissions for direct comparisons.

Continuous haemodynamic monitoring:
Non-invasive measurements of haemodynamic parameters (e.g. cardiac output, stroke volume, pulse rate and mean arterial pressure etc.) were performed via our pre-optimised setup by using volume clamp method ( Statistical analysis: Analytical mean values (of measured parameters) from each participant were calculated over each minute of breath-resolved measurement. Data from every 5 th minute were included for statistical analysis. In case of any non-parametric distribution of data, median values were considered for statistical analysis.
In order to reduce the evident intra-individual variations in measured variables, each participant was used as his/her own control. Thus, variables from each subject were normalised onto the corresponding initial values (of the rst minute). Normalisation was performed separately for each mask types (FFP2 and surgical) and in each age groups (adults and elderly).
As every group mean/median value are contributed by each volunteer (of that group), the relative standard deviations (RSDs) in VOC abundances from each group were also calculated for each substance. The RSDs were calculated (in %) by rating sample standard deviations (SDs) over corresponding sample means.
Statistically signi cant differences within groups were assessed via repeated measurement ANOVA on ranks (Friedman repeated measures analysis of variance on ranks, Shapiro-Wilk test for normal distribution and post hoc Student-Newman-Keuls method for pairwise multiple comparisons between all groups; p-value ≤ 0.005) in SigmaPlot software (version 14).
For all measured variables, from all pairwise comparisons, the differences are presented by referring to the corresponding values at the 1 st minute of each mask and within each age group.
In order to compare the effects of both mask types on both age groups, relative changes (in %) over time (with respect to initial values) were calculated for selected variables. Here, we have selected the principal physio-metabolic denominators and candidate VOCs that are potentially originating from several in vivo metabolic processes. Relative changes were calculated at 15 th and 30 th min in adults and at 15 th min in elderly cohort. The changes in pET-CO 2 values were calculated between immediately before and after mask use. In case of inter-group comparisons, one-way ANOVA was applied due to unequal group size. All groups were compared to each other.
In order to understand the correlations between exhaled VOCs and physiological parameters within each mask type, dimension reduction factor analysis (Factor extraction via principal components method, factor scores via regression method and 1-tailed signi cance at p-value ≤ 0.005) were performed in SPSS. Figure 2 represents heatmaps of relative changes in normalised mean values of physiological parameters such as, partial pressure of the end-tidal CO 2 (pET-CO 2 ), peripheral oxygen saturation (SpO 2 ), respiratory rate (RR), cardiac output (CO), stroke volume (SV), pulse rate (PR), mean arterial pressure (MAP) and relative changes in exhaled alveolar abundances of 32 protonated/charged VOCs of interest during the use of FFP2 and surgical face masks by healthy adults and elderly cohorts. VOC selection criteria is described in method section. Measured variables from each volunteer were normalised onto corresponding median values from the rst minute. The mean of those normalised values from every 5 th minute is presented in the heatmaps. pET-CO 2 values are depicted from immediately before and after the mask use and are placed at the rst and nal minute of heatmaps for direct comparisons. The changes in relative standard deviations (RSDs) of all measured parameters are presented via heatmaps in Supplementary Figure-S1.

Figure 3 (Boxplots) is depicting absolute or normalised values of physiological parameters and of alveolar concentrations of exhaled
VOCs in every 5 th minute (starting from the 1 st min) in four groups. First two groups consist of data from FFP2 masks on adults and elderly cohorts, respectively. Later two groups contain data from surgical masks on adults and elderly cohorts, respectively. pET-CO 2 values are presented from immediately before and after the use of masks. Here, 3 (A) represents the physiological parameters viz. absolute values of SpO 2 , pET-CO 2 and respiratory rate and normalised values of haemodynamics. 3 (B) represents aliphatic aldehydes and organosulfur, 3 (C) represents hemiterpene, ketone and smoking/environment related VOCs, exhaled humidity and oxygen, 3 (D) represents aliphatic acids, alcohols and monoterpene. Absolute values are only considered for parameters, which are less likely to be affected by inter-or intra-day variations within each individual. From all pairwise comparisons, the differential expressions (viz. statistically signi cant at p-value ≤ 0.005) in each variable within each group is indicated with respect to the corresponding '1 st minute' of measurement.
The correlation coe cients and respective p-values between physiological parameters and VOCs of interest are presented in Table 2.
Detailed inter-VOC correlations (with respect to physiological parameters) along with corresponding p-values are presented in Supplementary Table-S1 and S2.  Table 3.

Discussion
FFP2 and surgical face-masks immediately affected physiological and metabolic attributes. Effects were progressed with the course of mask wearing time. Most pronounced effects were observed in case of FFP2 mask and especially in elderly subjects. Profound and consistent decrease in SpO 2 and increase in pET-CO 2 have indicated ascending deoxygenation [26] and deteriorating ventilation in all subjects, which are caused mainly due to rebreathing of CO 2 [27,28] from mask dead space and change in normal breathing patterns.
Haemodynamic parameters such as cardiac output and MAP changed in counter-(homeostatic)-response/secondary to those respiratory effects. Irrespective of the origins, signi cant and substance-speci c changes in exhalation of many blood-borne volatile metabolites took place within minutes. Some changes mirrored the pro les of oxygen saturation, haemodynamics and respiratory parameters. Exhaled oxygen decreased while breath humidity increased over time. Exhalation pro les of potentially endogenous aldehydes, hemiterpene, organosulfur, alcohols, ketones and short-chain fatty acids have indicated in vivo physio-metabolic cross-talks between hypoxia and oxidative stress, hypoventilation and compartmental vasoconstriction, altered systemic bacterial activity and energy homeostasis etc. Exhalation of exogenous aromatics, nitriles and monoterpenes were mainly related to pre-exposers and lifestyle.
While looking at the CO 2 exhalation and accumulation of breath humidity over time, a systematic effect of rebreathing to increase most of the VOCs (with high aqueous solubility or high volatility) could be assumed. Nevertheless, endogenous VOCs with similar physiochemical properties behaved in contrast by clearly indicating more systemic effects on their putative in vivo/metabolic origins.
Hypoxia and deoxygenation facilitates the production of reactive oxygen species (ROS) and thereby promotes acute oxidative stress [29,30]. This further leads to lipid peroxidation and production of α, β-unsaturated aldehydes [31]. In our setup, the instant and gradual increase in endogenous acetaldehyde, butyraldehyde and pentanaldehyde exhalations along with the decreasing SpO 2 in case of FFP2 masks indicates an early onset and progression of oxidative stress. Such oxidative stress was insigni cant in case of surgical masks, even in elderly subjects. As oxidative stress driven disbalance in O 2 and ROS interplay may lead to acute cardiac dysfunction [32], DNA damage [33], oncogene activation and cancers [34] etc. the selection of face-mask should be re-evaluated.
Crotonaldehyde is formed via condensation of acetaldehyde molecules in alkaline medium and thereby cannot be attributed directly to oxidative stress. Acrolein behaved differently due to its exogenous origin from diet, smoking and/or environment [35].
While looking at the exhalation pro les of organosulfur such as DMS, AMS and butanethiol, the effects of both face-masks are re ected on the systemic origin of these substances from microbial anaerobic methylation[36] in the lower gut. Studies have shown that effects of hypoxia acts as an 'invisible pusher' of gut microbiota [37]. Gut ora maintain the hypoxic balance of the intestinal environment in order to regulate the nutrient absorption, gut permeability and immune response [38]. As face-mask externally induces a deoxygenation, the normal gut microbial activity is likely to reduce gradually, resulting in descending production of those organosulfur in colon region. Despite an increase in cardiac output, exhaled abundances of these substances decreased signi cantly in case of both masks in either age groups. This could be due to the fact that the perfusion was distributed to active compartments (tissue/organs) with increased O 2 demand rather than in the gut. Due to its origin from the oral cavity bacteria [21], no systemic effects were observed in the exhalation of H 2 S.
Short-chain fatty acids (SCFAs) e.g. acetic acid and butyric acid are produced by lower gut bacteria during the anaerobic lysis of primarily undigested dietary bres and/or starch [39,40]. Due to their origin from large-intestinal environment, exhaled pro les behaved as the gut originated sulphides. Further to that SCFAs plays important role in energy metabolism, plasma acid-base homeostasis and blood pressure regulation [41,42]. As prolonged hypoxia eventually may leads to anaerobiosis and metabolic acidosis (lowering of plasma pH) [43], SCFAs production is very likely to be reduced. Crotonic acid and formic acid are potentially sourced from cosmetics and disinfectant/sanitizers and similar to our previous observations [14,44], those VOCs re ected washout behaviours.
Despite its origin from carbohydrate metabolizing bacteria of the intestine [45,46], ethanol did not follow the behaviours of organosulfur or SCFAs. In contrast, ethanol exhalations tend to increase and rose most signi cantly in elderly cohort with FFP2 mask by mirroring the pro les of cardiac output and MAP. Evidences have indicated hypoxic switching of metabolic routes that produce more ethanol than lactates in order to regulate blood pH levels [43]. Endogenous ethanol increases the permeability of small-intestinal epithelium and colon in order to increase glucose transport towards hepatic and cellular glycolysis [47]. A consecutives descent in endogenous breath acetone (i.e. the by-product of glycolysis)[48] indicate a decline in carbohydrate metabolism and a demand in glucose uptake for energy metabolism. An elevated MAP and -cardiac output denominate increased perfusion [49] of vital organs to aid the primary source of energy from the compartments such as the small-intestine. Therefore, the increase in ethanol exhalation may be due to its signalling act [50] between intestinal permeability and glucose transport to blood for maintaining the energy homeostasis.
The observed tendencies of various gut-originated VOCs re ect the regional diversity of systemic micro ora within same organ. Phenol and isopropanol behaved differently than ethanol due to their exogenous origin from the dietary intake, beverages and uptake from the ambient environment, disinfectants or sanitizers.
During spontaneous breathing in normal sitting position, exhalation of CO 2 and endogenous isoprene [51] exhalation remain closely related to each other and they positively mirror cardiac output and negatively mirror ventilation [14,20]. Pronounced increase in pET-CO 2 values from before to after use of both masks occurred most likely due to partial rebreathing from the mask dead space and changes in spontaneous respiration (e.g. changes in inspiratory/expiratory time) that may alter alveolar slope [52]. Those effects are expected to elevate breath isoprene as was observed during exhalation of expiratory reserve volume by healthy subjects [14]. Within this setup, isoprene exhalation remained independent of cardiac output and CO 2 exhalation. Previously we have observed that breath holding manoeuvres [15,53] and externally applied upper-airway resistances against respiratory ow [44] had signi cantly increased breath isoprene concentrations. Although respiratory rate tends to decrease in adults and increase in elderly subjects, the changes remained within the normal physiological range. In all cases, isoprene concentrations signi cantly decreased throughout the experiments; most likely due to the sympathetic vasoconstriction (deoxygenation induced) in muscle compartments [54], which are the potential storage of this VOC but stayed inactive while sitting. Previously, we witnessed such decline in breath isoprene (in contrast to cardiac output) in healthy adults during the second minute in standing posture [20]. At that point, cardiac output started to increase but isoprene still decreased as sympatho-adrenergic vasoconstrictions took place in the lower extremities of the body to push up (against gravitation) the peripheral blood volume towards thoracic compartments. This was in order to counter the falling blood pressure and cardiac output while standing. As distribution of blood ow is crucial under hypoxia [55], in the present setup, the same phenomena might have helped to redistribute the available perfusion within active compartments in order to aid the rising O 2 demand. Due to having both haemoglobin and plasma bicarbonate buffer, such effects were not observed in case of pET-CO 2 . Surprisingly, we have observed hyperventilation (pET-CO 2 < 35 mmHg) in most of the elderly subjects even before wearing the masks for our experiment.
This could occur in order to compensate/eliminate the elevated pET-CO 2 from precedent mask wearing (while arriving from elsewhere to our setup) phase. Although we let all subjects sit without any mask for 15 min within our setup before starting experiments, this seem to be not enough to compensate the precedent effects in all subjects above the age of 60 years.
Exogenous monoterpene like limonene is sourced to breath from clinical environment or via recently consumed fruit juice or similar.
Acetonitrile and aromatics such as furan, benzene and toluene etc. are exposed from the environment and/or smoking habits [21,56]. These substances are lipophilic in nature and are stored in the fatty tissues. They mimicked the isoprene exhalation mainly due to having similar physio-chemical properties (e.g. low aqueous solubility, high volatility etc.) and compartmental storage.
While considering the limitations, our pilot study is conducted on a limited sample size and only on healthy adults and elderly subjects. Assuming this setup on a large population of age, gender and BMI matched healthy and sick subjects (including children) could enhance our clinical understanding on the observed adverse effects beyond the everlasting physiological (intra-and inter individual) variations in patient suffering from obstructive and restrictive lung conditions and other respiratory diseases. In elderly cohort, we had to limit the measurements within 15 min as most of them reached a SpO 2 level < 94 by then under FFP2 masks. In our setup, we did not measure the fraction of exhaled nitric oxide (FeNO), which is known to regulate vasomotor tone, blood pressure and is regarded as a marker for oxidative stress [57]. The FeNO in breath increases under acute exogenous hypoxia[58] and pilot ndings have indicated inverse relationship between isoprene and nitric oxide [59]. Thus, future examination of breath NO under the same/improved experimental conditions may reveal its clearer relationship to isoprene and other endogenous VOCs that are associated with oxidative stress and systemic microbial activity.
In conclusion, real-time breathomics has revealed a deeper insight into the physio-metabolic side effects of face-mask wearing. Based on recent pilot observations of cardiopulmonary parameters during exercise and rest in 12 healthy adults [7], researchers have generally recommended the continuous mask use. Within our setup, we have investigated the respiratory-, hemodynamic-and down-stream metabolic changes in adults and elderly subjects. Although we observed signi cant side effects and good compensation/adaptation trends in healthy adults below the age 60 years, those side effects emerged profoundly towards substantial risk in subjects over the  Instrumentation and experimental setup of the study. Customised PEEK extension of the PTR transfer line for direct sampling from the mask dead space is presented at the top. Continuous real-time breathomics via PTR-ToF-MS (1) in parallel to continuous non-invasive monitoring of haemodynamic via ClearSight system (2) and pulseoximetry (3) is presented at the bottom.  Statistical signi cances were tested by means of repeated measurement-ANOVA on ranks (p-value ≤ 0.005). From all pairwise-multiple comparisons, statistically signi cant differences with respect to the '1st minute' are indicated via coloured '#' and '*'. Here, # and * are assigned to FFP2 and surgical masks, respectively. Green and red colours denote adults and elderly cohorts, respectively. Similarly, pET-CO2 values before and after mask use are compared statistically.