A total of 40 subjects covering a wide age range participated in the study. They were recruited through information on the website of our institute. Exclusion criteria were absolute and relative contraindications for CPET according to the American Thoracic Society18. Subjects underwent a baseline examination consisting of medical history, recruitment questionnaire, physical examination, routine laboratory tests, electrocardiogram, a pulmonary function test and an initial CPET.
Specific IgE antibodies (sIgE) to ubiquitous aeroallergens (atopy screen sx1, Phadiatop) were measured with the ImmunoCAP 250 system (ThermoFisher Scientific, Phadia AB, Uppsala, Sweden). A positive atopic status was assumed in case of a sIgE concentration to sx1 ≥ 0.35 kU/L.
The prospective, randomized, crossover study design was described in detail in a previous manuscript dealing with the influence of masks on cardiopulmonary performance16. All parts of the study, including recruitment, were conducted between September 2020 and July 2021.
Each subject was tested with four mask situations: no mask (NM) as reference, surgical mask (SM; Type II, MedicalCare & Serve industry®, Wilfried Rosbach GmbH, Willich, Germany), community mask (CM; van Laack® GmbH, Mönchengladbach, Germany), and a filtering face piece (FFP2; Dräger X-plore® 1920 NR D, Dräger® Safety AG, Lübeck, Germany) in randomized order. The cross-over study consisted of two modules where normally worn masks (including leakage) were examined during physical exertion on a bicycle ergometer (ergometry) and under normal working conditions (workplace examination). In a third module (CPET) at identical physical exertion as in ergometry, mask material was presented to subjects in a double-blinded setting using a special mask adapter16. In brief, a round sample of the tested mask (or nothing for the no mask situation) was placed in a commercially available opened, empty bacterial filter and airtight closed with adhesive tape and a metal clamp. This mask adapter was then placed between the silicone CPET mask and the measuring devices (Fig. S1). The order of the three modules, which all took place at our institute, was also randomized. For physical exertion (ergometry, CPET), a maximum of two mask situations per day were tested, with a sufficiently long regeneration time in between. The workplace examinations with the four mask situations were carried out on four different days. In all modules, each session was conducted at a comparable time of day.
Individually determined load levels resulting in a minute ventilation of 10 L/min (rest), 30 L/min (exercise (E1)), 50 L/min (E2), > 60 L/min (E3) and 10 L/min (post)—each lasting six minutes—were used for the physical exertion during ergometry and CPET. According to the German Social Accident Insurance this corresponds to light (rest and post), moderate (E1), heavy (E2), and very heavy (E3) work16. During the 4-h workplace examination the masks were normally worn during light/moderate work in the office or laboratory.
Cardiopulmonary parameters and blood gases were measured and perceived exertion (Borg scale) was requested as described earlier16. Temperature and relative humidity were recorded using a climate data logger (PeakTech 5185®, Ahrensburg, Germany), which was fixed with adhesive tape between nose and mouth.
Identification of “sensitive” individuals
During the baseline examination, the participants filled out questionnaires including questions on health status and specific sensitivity scales enabling a retrospective characterization of the study group in terms of identifying “sensitive” individuals. The latter had no relevance for the in- or exclusion of individuals.
One of the sensitivity scales was the Discomfort Intolerance Scale (DIS) that measures the level of agreement with statements about tolerance of discomfort19. Two distinct sub-factors entitled Intolerance of Discomfort or Pain (DIS-I) (2 items; e.g. “I can tolerate a great deal of physical discomfort”-reverse scored), and Avoidance of Physical Discomfort (DIS-A) (3 items; e.g. “I take extreme measures to avoid feeling physically uncomfortable”) were calculated aside from the global DIS-score. Also included were a questionnaire on environmental worry (Environmental Worry Scale (EWS), 5 items) constructed to express possible negative thoughts and associations towards negative effects and personal threats by environmental factors like for example “I often think about the fact that I am taking pollutants into my body”20, and the Positive and Negative Affectivity Schedule (PANAS) measuring the extent to which the subject generally experiences positive or negative emotions21. For the purpose of the present study, only the negative affectivity subscale (PANAS-NA) containing 10 negative items (upset, guilty, scared, hostile, irritable, ashamed, nervous, jittery, afraid, distressed) was used to assess the trait-like tendency to experience negative affect states. In addition, recruitment tools included two questionnaires on self-reported sensitivity to chemicals. The Chemical Odor Sensitivity Scale (COSS) is an 11-item scale for assessing trigeminal (e.g. shortness of breath, coughing, sickness, and nausea) and olfactory (e.g. perceived unpleasantness) mediated responses when exposed to chemicals such as paint or everyday odors like perfume22. The main focus of the Chemical Sensitivity Scale (CSS, 21 items) is on affective reactions and behavioral disruptions by chemicals23.
From these data a relative cut-off score was calculated to divide subjects into either lower or higher “sensitive” persons. Except for the EWS, the sensitivity factors were median-split: MD(DIS) ≥ 17; MD(DIS-I) ≥ 4; MD(DIS-A) ≥ 8; MD(PANAS-NA) ≥ 16; MD(COSS) ≥ 10; MD(CSS) ≥ 50. Regarding the EWS, subjects were counted as subjects with heightened environmental worry if they partly or fully agreed on at least one of the five items24.
Experimental measurements and questionnaires during the study
Comfort Score and Symptom Score were assessed with mask (or without in the NM situation) before (pre) and after (post) ergometry and CPET, and Comfort Score also within the last 20 s of each load level. During the 4-h workplace examination, Comfort Score and Symptom Score were assessed 30 min before mask wearing (pre), 30, 60, 90, 120, 150, 210, and 240 min during mask wearing, as well as 30 min (post) after the end of mask wearing. Additionally, cognitive performance was assessed at the beginning and at the end of the workplace examination.
Comfort Score questionnaire
The Comfort Score questionnaire with ten items (humidity, heat, breathing resistance, itchiness, tightness, saltiness, feeling unfit, odor, fatigue, and overall discomfort) was used in German translation to quantify the perception of comfort/discomfort of wearing a mask. Sensations had to be rated on a 10-point rating scale with 1 representing ‘‘not at all’’, 5 representing ‘‘mildly’’ and 10 representing ‘‘strongly’’13, 25. The sum of the Comfort Score was the sum of all items except overall discomfort. The Comfort Score was used to determine the direct impact of the masks on the mouth-nose region and on the breathing comfort.
Symptom Score questionnaire
The occurrence and intensity of more general impairments caused by wearing a mask (e.g. headache, dizziness) was recorded with the Symptom Score questionnaire in German translation. The Symptom Score consists of 16 complaints and 4 dummy complaints that have been shown to be sensitive to CO2 inhalation effects26. Each of these complaints was rated on a 5-point graded scale (1 = not at all, 2 = slightly, 3 = medium, 4 = strong and 5 = very strong). The total complaints score (sum of Symptom Score) was a sum of these 16 complaints (16—80).
After the beginning and before the end of the 4-h workplace examination subjects performed a math and a spelling test to examine possible effects of mask wearing on cognitive performance. The duration of each test, which ran automatically on a computer, was 11 min and the spelling test was administered after the math test.
After the instruction, which was shown for 5 s, each of the 93 tasks was visible for 7 s. The elapsed time was displayed in the form of a progress bar below the task. Subjects were asked to answer as quickly and correctly as possible. In the math test, the subjects had to perform simple mental arithmetic. The tasks consisted of multiplying two numbers between 1 and 10 and subtracting a one or two-digit number from this product (e.g. (9 × 5) − 17 = 28). The subjects had to decide, if the correct calculation result was smaller, larger or equal compared with the given solution number. In the spelling test, subjects had to recognize misspelled words. These words had one or two mistakes or were correctly spelled.
If subjects did not respond within the allotted time, the next task was automatically presented, and the task was scored as error (omission). The number of correct and false answers, omissions, and the mean response time were counted.
Data (raw scores) of Comfort and Symptom Scores are expressed as median (minimum–maximum) and interquartile range (IQR 75–25) and presented by boxplots (box: median, 25th-75th percentile; whiskers: 5–95 percentile).
A generalized linear mixed (GLM) model, along with generalized estimating equations (GEE), was used on the logarithmized sum values of Comfort Score and Symptom Score as dependent variable. The GEE procedure extends the GLM to allow for analysis of repeated measurements. Here, load level and time of measuring were included as influencing factors. This approach allows intraindividual comparison at the different examination times (i.e., each subject is compared to him/herself). Least Squares Means were calculated based on these models. The situation without mask at each load level (pre, E1, E2, E3, post) or each time of measurement (pre, 30, 60, 90, 120, 150, 210, 240, post) was used as reference. Also considered were influencing factors such as sex, age (per 10 years) and height (per 10 cm). Further influencing factors on the sum of Comfort Score and cognitive function were tested by including them individually as potential factors to the model.
Since the GLM model was only applicable to the sum of Comfort and Symptom Score, we processed the data for the single questions using analysis of variance (Friedman test, Dunn’s multiple comparisons test as post-hoc test). For Comfort Score, at each load level (pre, E1, E2, E3, post) or each time of measurement (pre, 30, 60, 90, 120, 150, 210, 240, post) the data with mask were compared with the respective non-mask situation (reference).
Pearson’s chi-squared test was used to examine possible differences between men and women related to the different sensitivity scales. The Spearman rank correlation was calculated to predict the monotone association between parameters for correlations. This is further visualized by contrasting the parameters in a heat map.
A p-value of < 0.05 was considered statistically significant. Analyses were performed using SAS 9.4 (SAS Institute, Cary, NC, USA). Figures were drafted with SAS 9.4 and GraphPad Prism version 9 (GraphPad Software, San Diego, CA, USA).
The Ethics Committee of the medical faculty of the Ruhr-University Bochum gave approval to perform the study (Reg. No.: 20–7024) and all subjects gave written informed consent. The person shown in Fig. S1 gave written informed consent for publication of identifying images in an open access-online publication. All methods were performed in accordance with the relevant guidelines and regulations. The study was conducted in accordance with the latest revision of the ethical standards set down by the Declaration of Helsinki.