Ultrafine Particles in Viennese Gastronomy after Introduction of a National Smoking Ban

Introduction

Since November 2019, the Austrian gastronomy has been smoke-free mainly due to the indoor smoking ban that became effective. Measurements of particle number count (PNC) and lung-deposited surface area (LDSA) can be used to demonstrate compliance with the measures provided by the law, as there is a high correlation between ultrafine particles (UFP) and air nicotine, the latter specifically indicating tobacco smoke [1, 2, 3, 4, 5]. To our knowledge LDSA has not been used before to evaluate air quality after a complete smoking ban in the hospitality industry of other cities.

UFP can be generated from traffic exhaust gases as well as from various smoking and vaporizing sources, such as heating, cooking, cigarette smoke, electronic cigarettes (exhaled nicotine-containing aerosols), candles or the use of deep fryers. In everyday life, fine or ultrafine dust, which is harmful to us, can thus be generated very quickly, e.g. through ordinary activities such as cooking. It has been shown that ultrafine dust has an impact on health and has an increased effect on the well-being of people who already suffer from respiratory tract diseases, since ultrafine dust can reach the depths of the lungs due to its small particle size [1, 6]. Equally at risk are patients with cardiovascular diseases (endothelial dysfunction, vascular stiffness, coagulation promotion) [7, 8, 9].

Larger particles dominate the mass concentrations (PM) and smaller particles dominate the number and surface concentrations. UFP have sizes up to 0.1 µm and dominate PNC [10].

Passive smokers are those persons who not actively smoke cigarettes themselves, but inhale tobacco smoke because they are usually unintentionally surrounded by it. According to a survey by Statistics Austria, more than 25% of non-smokers were occasionally exposed to this passive or second-hand smoke (SHS), with younger age groups being more frequently affected [11]. The 2019 survey found that women and men aged 15 to 29 years were the most likely to be exposed to secondhand smoke, with at least 10% of both genders exposed daily and just over 30% of men and one-fifth of women in this age group exposed at least once a week. In the age group 60-74 years, reported exposure to secondhand smoke decreased, with daily exposure to secondhand smoke occurring in about 5% of respondents [11].

Residues that remain after smoking, e.g. on wallpaper, carpets, upholstered furniture, children’s toys or other surfaces, indoor in private as well as in public locations, are referred to as “cold smoke” or third-hand smoke (THS). This is characterized by long-lasting deposits of up to several months after the last cigarette smoked in the affected rooms. Even by ventilating the room, the cold smoke cannot be completely removed from furniture, carpets and wall paper, is partly changed and released into room air. After a person enters a room after consuming a cigarette in an outdoor area, smoke particles continue to be exhaled for one to two minutes afterwards, and even a few minutes after that, carcinogenic substances continue to be exhaled. Smoking in the presence of or immediately before contact with children as well as infants should be refrained from, as the smoke particles can also settle in their hair and clothing and in house dust. The particles that enter the interior and their compounds are subsequently absorbed through the skin or by inhalation, as well as through the oral mucosa, and can contribute to illness. It is even recommended that all furnishings be replaced when moving into an apartment to reduce exposure to cold smoke as much as possible [12].

For indoor gastronomy environments, it can be assumed that smoking causes an increased concentration of ultrafine dusts both from SHS and THS after recent smoking. This aspect, which is hazardous to health, is to be analyzed in comparison with non-smoking establishments and smoke- free areas.

Objectives

Following the research question of ultrafine particulate matter in the Viennese gastronomy after the legal smoking ban, the focus was on whether and to what extent different UFP parameters can distinguish between different types of venues (with and without smoking before the ban).

Material and Methods

In the first part of the present study, in the course of the data collection, 40 different localities, cafés, discotheques, bars, etc., mainly frequented by younger people, were investigated undercover. For the detection of ultrafine dust, the Miniature Diffusion Size Classifier (MiniDiSC®) was used to measure PNC and LDSA. PNC is dominated by ultrafine particles and indoors by combustion aerosols such as tobacco smoke, suspended for a long time in room air, inhaled deeply into the airways and partially penetrating into the blood via the lungs. LDSA measures the surface area of this ultrafine dust, which can come into contact with body cells and therefore has particular toxicological significance.

Recording of the exposure to ultrafine dust, the surveys were each carried out for a minimum of 20 minutes in order to obtain meaningful and reliable results. As far as possible, the measurements took place to avoid certain interfering factors, such as candles, steam from open kitchens or fog machines. If this was not possible for certain reasons, such as lack of space, or if there was no opportunity for an inconspicuous measurement, this circumstance was recorded. Another condition was that the gastronomic locations were well frequented. The hidden measurement was carried out with the MiniDiSC at about chest height.

Ultrafine particles collected in the localities were transformed into corresponding averaged values using the MiniDiSC data conversion tool© program software. The values for PNC, particle diameter and LDSA collected in this way were collected in an Excel© database and finally converted into a corresponding SPSS© matrix.

Due to a technical problem, measurements in one pub could not be carried out completely. Analyses are based on 39 visited localities between October 2021 and February 2022. The data collection processes were documented pseudonymously, which means that no conclusions can be drawn about individual locations. Only authorized persons can access the sensitive data.

In the second part of the study, these UFP data from nonsmoking sites were compared with corresponding measurements also from 39 locations from smoking sites and nonsmoking areas prior to the ban of November 1st, 2019.

Statistical Analysis

The descriptive and inferential statistical analyses were performed using IBM SPSS® 20.0 statistical software package Parameter and location type n M ± SD min - max Md IQR mean rank PNC pt/cm³ $$ \text{Bar (Disco)} \quad 22 \quad 7 3 7 4 2 \pm 1 3 4 3 5 2 \quad 5 5 8 1 - 5 5 5 0 1 4 \quad 1 9 7 5 1 \quad 1 1 2 6 3; 4 5 6 0 8 \quad 2 0. 1 4 $$ $$ \text{Pub (Restaurant)} \quad 12 \quad 3 1 5 0 8 \pm 1 1 7 4 0 5 \quad 6 1 0 7 - 9 9 7 6 4 \quad 1 9 3 5 7 \quad 1 2 8 8 0; 3 8 7 3 2 \quad 1 9. 8 3 $$ Average diameter (nm) Bar (Disco) 22 47.72±16.51 13.95-77.11 54.17 39.37; 57.91 20.82 (IBM, Armonk, NY). In the context of inferential statistics, the significance level was set at p ≤ .05. Furthermore, to interpret the practical significance of results, the standardized effect size η2 was used according to Cohen’s classification. Considering this, effect size ranges ≥ .01 are considered small, ≥ .06 moderate, and ≥ .14 significant [13].

In accordance with the study design, the differences in ultrafine particle parameters were assessed by means of single-factor, non-parametric methods such as Kruskal- Wallis-testing and, after logarithmization of skewed data, also by means of two-factor parametric methods with respect to the factors type of locality and location status (year of study and smoker/non-smoker location). In addition, a model test was performed to examine the explanatory value of ultrafine dust parameters with respect to the discriminability of non- smoking localities (2021), non-smoking areas (2019) and smoking gastronomies (2019). Finally, ROC functions were used to determine meaningful cut-off values of three ultrafine particle indicators to distinguish smoking and non-smoking localities. UFP parameters as study relevant variables in the measurement sequences, taking into account unit and metric, as well as scale level, are summarized in Table 1.

Results

• Surveys in non-smoking locations in 2021/22, comparing three different types of premises. The assessment of the difference in the extent of the three ultrafine dust parameters in the comparison of the local types showed non-significant results in each case using the corresponding Kruskal-Wallis testing as shown in Table 2.

p-value $$ \text{Café} \quad 5 \quad 9 2 7 6 9 \pm 1 7 0 9 7 0 \quad 8 3 2 5 - 3 9 8 3 7 6 \quad 1 8 8 5 4 \quad 1 2 3 5 0; 2 5 9 3 9 \quad 1 9. 8 $$ $$ \text{total} \quad 39 \quad 6 3 1 8 6. 4 \pm 1 1 7 4 0 5 \quad 5 5 8 1 - 5 5 5 0 1 4 \quad 1 8 8 5 4 \quad 1 2 0 8 1; 4 1 0 8 5 \quad 0. 9 9 6 $$ Café 5 42.61±13.40 19.30-65.22 44.27 39.16; 45.09 15.4

The transformation of the skewed original data into logarithmized (lg10), normally distributed measurement data series was essential in order to be able to perform a multi-factorial comparison. In the figures below, the UFP mean values have been also logarithmized.

Discussion

The sampling and evaluation of the concentrations of UFPs in the present study was carried out in the most frequent Viennese establishments for hospitality, grouped into three types of premises. The data for the present survey was collected in non-smoking establishments, from 09/2021 to 02/2022. At that time, the law on the protection of non-smokers with a smoking ban in all pubs in Austria on 01.11.2019 was already established. For comparison the same type of venues was used which had been sampled by the same methods from April to October 2019, when only partial smoking bans had been in force [3]. The results showed for UFP similar reductions of particle exposure in number and surface after complete smoking bans as had been found before for fine particle mass in Vienna [14, 15], Dublin [16], Aberdeen, Edinburgh [17] and other European cities. Less surveys measured also PNC of ultrafines before and after a ban. PNC is more variable in time and space. In Rome mean PNC decreased in bars from 60,998 to 28,737

three months after the ban, but increased again to 51,069 after one year [18]. In Vienna 3 years after the ban mean PNC significantly dropped in all venues from 105,679 to 63,186 and the median PNC of 18,854 was even lower than the median PNC before the ban in non-smoking rooms, which was 27,776. This indicates that before the ban tobacco smoke from neighbouring smoking rooms was not completely prevented to enter non-smoking rooms and that the legally required separation failed [19]. It could also indicate that particle release from THS in the 2021/22 survey did not play an important role anymore, because of renovation, cleaning and ventilation during the past three years. Some outliers found in certain venues 3 years after the ban could rather be attributable to recent SHS by single smokers still violating the law at times, which makes a continuation of controls necessary in these premises. This conclusion is supported by the study results on mean LDSA, which decreased significantly from 443 to 101 µm²/cm³, but is still higher than reported from other non-occupational environments without smoking and not near kitchens, burning candles, etc [20]. These sources had been omitted both in this study and in the preceding ones [3, 19], but in the hospitality industry UFP from “cooking & eating” microenvironments [21, 22] could not be excluded except by selection of venues and measuring distant from possible other sources like open kitchen or burning candle. More distant possible sources had been registered, but no influence on results could be detected before and after the ban. LDSA was also higher indoors, even in non-smoking rooms, than outdoors from urban traffic [3]. The mean of 34 and median of 25 µm²/cm³ found outdoors in 2019, should be underrun by indoor concentrations without indoor sources like smoking, heating, cooking, etc. From the present study, however, recent smoking in premises of the hospitality industry can only be suspected if mean LDSA exceeds 164 µm²/cm³ and PNC exceeds 34,435 particles per cm³. In winter with low evaporation also a mean aerodynamic particle diameter above 67 µm might give a hint, provided that particles from tobacco smoke had time to coagulate during aging. However, the average diameter plays a subordinate role and could be redundant as a measurement variable. On the other hand PNC and LDSA should be used more frequently as indicators of particulate air pollution from urban traffic [20, 21, 22, 23, 24, 25, 26], as well as from cigarettes [4, 19, 27].

No relevant difference between the different types of premises (bar disco, cafés, pub restaurant) was found any more after the ban, which is encouraging, because bars, clubs and discotheques before the ban had shown the highest exposures [3]. So protection of youth, which is attending these premises most frequently, seems to make progress in Vienna.

Limitations

Like PM1 also PNC and LDSA are highly correlated with air nicotine, but elevated values of UFP concentrations do not necessarily represent reliable evidence of smoking activity in a location. Before the implementation of the general smoking ban in the Viennese gastronomy, the studies could not prove any significant influence of candles, etc. on the surveyed parameters, but further investigations are recommended to confirm the results found. Replicability is not possible, but repeated cross-sectional studies should be able to show continuing improvement of air quality and further reductions of UFP in the hospitality industry.

Acknowledgements and Conflicts of Interests

The Austrian Academy of Science (Commission for Climate and Air Quality) provided all instruments used in this study and covered costs of service and calibration. No other funding was received for this study, but in maintenance of instruments and data collection the authors were supported by students of the Medical University of Vienna (Honoré- Spinler D, Riedmann H) and preceding diploma theses, especially one by Raab M, et al. [3]. The authors declare no conflicts of interests.