In contrast to other countries, there is an on-going debate but still no smoke-free legislation in Germany. Exposure to environmental tobacco smoke (ETS) in hospitality venues is assumed to be high, but air quality data are lacking. Therefore, the aim of our study was to perform a comprehensive exposure assessment by analysing the indoor air concentration of toxic or carcinogenic ETS compounds in restaurants, pubs, and discotheques. Active sampling of indoor air was conducted for 4 h during the main visiting hours in 28 hospitality venues. Polycyclic aromatic hydrocarbons (PAH), volatile organic compounds (VOC), aldehydes/ketones, and cadmium were analysed. In addition, particle mass concentration was assessed with two different methods and particle number concentration (PNC) was determined. Median nicotine levels were 15 μg/m3 in restaurants, 31 μg/m3 in pubs, and 193 μg/m3 in discotheques. Across these three sampling site categories median levels of 3-ethenylpyridine ranged from 3 to 24 μg/m3, median levels of benzene from 8 to 20 μg/m3, median levels of cadmium from 3 to 10 ng/m3, and median levels of the sum of 16 PAH according to US-EPA from 215 to 375 ng/m3, respectively. Median PM2.5 mass concentration assessed gravimetrically varied between 178 and 808 μg/m3 and PNC between 120,000 and 210,000 particles per cm3 in restaurants, pubs, and discotheques. The majority of the particles had a size of 0.01–0.5 μm. Concentrations of ETS compounds were always highest in discotheques. The strong correlation between ETS-specific markers (nicotine, 3-ethenylpyridine) and PM2.5, PAH, VOC, aldehydes/ketones, and cadmium indicated ETS as main source of these toxic or carcinogenic substances. In conclusion, indoor air concentrations of ETS constituents were high in German hospitality venues and represented a substantial health threat. Effective measures to protect patrons and staff from ETS exposure are necessary from a public health point of view.
Environmental tobacco smoke (ETS) is by far the most significant indoor air quality issue in health terms. ETS is a well-known health threat and has been classified as carcinogenic to humans (WHO IARC, 2004, U.S. Department of Health and Human Services, 2006). In contrast to other countries, a considerable proportion of the population in Germany is exposed to ETS because tobacco consumption is still high in the German population. A recent national telephone health survey among adults revealed smoking prevalences of 37% for men and 28% for women (Lampert and Burger, 2005). In case of adolescents, 32% of 15-year-old boys and 34% of 15-year-old girls reported regular smoking (Langness et al., 2005). Among young adults, smoking prevalence may be even higher with 54% for men aged 18–19 years and 48% for women aged 18–19 years (Schulze and Lampert, 2006).
In addition, there is currently no comprehensive smoking ban in public places, particularly hospitality premises. As a result, exposure to ETS is one of the most important public health issues in Germany.
Data of the German Environmental Survey 1998 indicated that approximately 20% of all adult non-smokers were exposed to ETS at home, in the work place or in other places, when ETS exposure was defined as cotinine excretion in urine (Heinrich et al., 2005). On the basis of the questionnaire data obtained within the German Health Survey 1998, 55% of all adult non-smokers were frequently exposed to ETS. The proportion of ETS-exposed non-smokers may rise up to 90% in case of 18 to 19-year-old men. For all age groups, ETS exposure at other places including restaurants or bars is predominant compared to ETS exposure at home or in the work place (Schulze and Lampert, 2006).
An exposure study in seven European cities demonstrated the highest nicotine concentrations in bars and discotheques compared to other public places such as train stations or universities (Nebot et al., 2005). Pubs and discotheques are essential exposure sites for adolescents and young adults. However, in Germany objective measurement data on the extent of ETS exposure at these public places are still lacking.
Therefore, the aim of our study was to perform a comprehensive exposure assessment by analysing the indoor air concentration of ETS in terms of particulate matter (PM), polycyclic aromatic hydrocarbons (PAH), volatile organic compounds (VOC), aldehydes/ketones, and cadmium in hospitality venues such as restaurants, pubs, and discotheques.
Active sampling of indoor air was performed during 4 h in 28 hospitality venues in the cities Augsburg and Munich, Germany, from April 2005 to May 2006. According to their characteristics the locations were categorised into (1) cafés or restaurants (N=11), (2) pubs or bars (N=7), and (3) discotheques or clubs (N=10). The measurements were performed during the main visiting hours of each individual location as indicated by the manager. Sampling started about noon in five cafés/restaurants, at early evening (about 18:00–20:00) in six cafés/restaurants and seven pubs/bars, and late in the evening in the 10 discotheques/clubs (about 22:00–24:00).
Data on the total guest capacity were given by the manager of each venue. In smaller locations, persons were counted each hour to calculate the average number of persons at one point in time. In larger and more crowded locations such as discotheques, counting of guests was not feasible. In that case, data on total number of guests during the evening or the expected total number were obtained from the manager and 80% of this number was assumed as estimated average number of persons at one point in time.
The proportion of smoking was calculated in case of smaller locations as ratio of the average number of smokers (counted each hour) and of the average number of guests. In larger locations, smoking rate was calculated based on estimations of smoking rates among the guests, mostly ranging from 50% to 60%.
Airborne PM was measured cumulatively by gravimetry and continuously by spectrometry as PM mass concentrations. PM1, PM2.5, and PM10 levels were determined continuously by using an optical laser aerosol spectrometer (dust monitor 1.108, Grimm Technologies, Ainring, Germany). The laser aerosol spectrometer measured particle concentrations in 15 nominal size bins from about 0.3 to 20 μm. Values were stored minute by minute in a data logger. In addition, PM2.5 was measured gravimetrically according to European guideline EN 14907 with a medium volume sampler using a flow controlled pump working with a constant flow of 2.3 m3/h (Leckel, Berlin, Germany), a PM2.5 sampling unit as sample inlet and quartz fibre filter.
For a subset of sampling sites, a TSI model 3034 scanning mobility particle sizer (SMPS, TSI Inc., Shoreview, MN, USA) was used to measure particle number concentrations (PNC) for a discrete size distribution of aerosols within a range of 10–487 nm.
Volatile Organic Compounds
Indoor air samples were collected with a constant flow of 0.02 l/min with Tenax GR as adsorbent in the first tube and Chromosorb 106 in the second tube and analysed using a thermodesorption unit (Gerstel, Muelheim, Germany) coupled to a gas chromatograph/mass spectrometer (GC/MS; gas chromatograph 6890A coupled to MSD 5973N, Agilent, Waldbronn, Germany). Acetonitrile, acrylonitrile, benzene, 2,5-dimethylfuran, 3-ethenylpyridine, nicotine, and 1,3-butadiene were quantified as single compounds. The limit of detection (LOD) for a single compound was 0.2 μg/m3 using a sample volume of 5 l.
The total concentration of volatile organic compounds (TVOC) was determined as toluene equivalents according to the international guideline ISO 16000-6. The total peak area of the total ion chromatogram in the retention time window between n-hexane and n-hexadecane was quantified using the response factor of toluene obtained by external calibration. TVOC values as toluene equivalents are semiquantitative since single compounds in the sample may have response factors that may widely deviate from that of toluene.
Airborne concentrations of aldehydes and ketones were determined by DNPH (2,4-dinitrophenylhydrazone) method. Air was sampled on a Sep-Pak® cartridge (Waters, XPoSure™, Eschborn, Germany) with a constant flow (0.2–0.8 l/min, depending on expected pollution). Elution of aldehyde derivatives was performed by 5 ml acetonitrile (Merck LiChrosolv, Eching, Germany). HPLC measurement was carried out with a system from Shimadzu (pumps LC 10 AD, detector SPD-10AV, controller SCL-10Avp, autosampler SIL-10ADvp, Software Class-VP 7.2.1, Neufahrn, Germany).
A binary gradient (acetonitrile/water) was used to separate the 2,4-DNPH on a RP column SUPELCOSIL LC-18 (l=150 mm, id=4.6 mm, pore size 3 μm). UV wavelength for detection was 360 nm. Quantification was performed via peak areas assessed by integration. Limits of detection were 0.02 μg/sample, resulting in 0.1 μg/m3 (4 h sampling). Formaldehyde, acetaldehyde, butanone, and acrolein were determined.
Polycyclic Aromatic Hydrocarbons
Gaseous and particle-bound PAH were determined by collecting indoor air with a medium volume sampler equipped with a sampling unit consisting of a PM2.5. inlet, a quartz fibre filter and a polyurethane foam. Filter and PU foam was extracted with toluene after addition of deuterated PAH standards, purified on a silica column and analysed by GC/MS.
Naphthalene, acenaphthylene, acenaphthene, fluorene, phenanthrene, anthracene, fluoranthene, pyrene, benz(a)anthracene, chrysene, benzo(b)+benzo(k)-fluoranthene, benzo(a)pyrene, dibenzo(a,h)anthracene, indeno(1,2,3-cd)pyrene, and benzo(g,h,i)perylene were determined as single PAH compounds. Additionally, the sum of these 16 PAH according to US-EPA was calculated.
The LOD was 0.1 ng/m3 using a sample volume of 10 m3.
Samples were collected on 47 mm quartz fibre filters (Pieper, Bad Zwischenahn, Germany) using a medium volume sampler equipped with a PM2.5 sampler as sample inlet operated at a constant flow of 3 m3/h.
Before using these quartz fibre filters in this project, filters from the same production lot were analysed for their heavy metal blank values. After a closed-vessel microwave decomposition of the filter samples using nitric acid and hydrogen peroxide as oxidising agents, cadmium was measured by inductively coupled plasma-mass spectrometry (ICP-MS). With the described analytical approach, a LOD of 0.02 ng cadmium/m3 using a sample volume of 10 m3 was obtained.
Carbon dioxide levels were determined with a continuously monitoring infrared sensor (Testo 445; Testo Inc., Lenzkirch, Germany), using 1-min data logging intervals. A separate sensor connected to the Testo instrument was used to measure humidity and temperature.
Continuously measured particle concentrations were summarized to a site-specific median for each sampling location. Concentrations below the LOD were assigned half of the LOD value.
Correlations between single compounds of ETS and of ETS compounds with characteristics of sampling sites were evaluated with the Spearman's rank correlation coefficient. Kruskal–Wallis test was used to assess differences between the three categories of sampling sites (restaurant/café, pub/bar, and discotheque/club). Owing to the small number of observations, no linear regression model was used including several possible explanatory and confounding variables at the same time. All analyses were performed using the SAS software package version 9.1 (SAS Institute Inc., Cary, NC, USA).
Characteristics of the Sampling Sites
The main characteristics of the sampling sites are given in Table 1. In 15 locations, the sampling site had windows. In all cases, these windows were closed during sampling. In only 7 of the 28 locations there was a door that connected the sampling site to the ambient air outdoors. The majority of hospitality venues were located remote of busy streets. Sampling was performed in the smoking section of each location. In most instances, there were no separate sections for smoker and non-smoker. Only three cafés/restaurants and two pubs/bars provided non-smoking sections, which were not separated by constructional measures from the smoking section. All hospitality venues had a ventilation system available. In one of the discotheques a fog machine was used during measurements.
Particle mass concentrations of PM2.5 ranged from 69 to 6770 μg/m3 assessed gravimetrically and from 55 to 4475 μg/m3 measured spectrometrically (Table 2). PM2.5 concentrations measured either by gravimetry or by aerosol spectrometry correlated very well (Spearman's correlation coefficient r=0.98, P<0.001). Mass concentrations of PM1 varied between 48 and 2615 μg/m3 and of PM10 between 80 and 4806 μg/m3. PM1, PM2.5, and PM10 mass concentrations differed significantly between the three categories of sampling sites (P<0.01) with highest values observed in discotheques/clubs. Omitting one disco with a fog machine did not substantially alter the results (data not shown).
Median PNC ranged from 29,084 to 221,695 particles per cm3 in cafés/restaurants (N=4), from 119,088 to 147,444 particles per cm3 in pubs/bars (N=2), and from 101,309 to 289,928 particles per cm3 in discotheques/clubs (N=7). The PNC in relation to particle size distribution are shown in Figure 1 for three exemplary sampling sites.
VOC and Aldehydes/Ketones
Levels of VOC and aldehydes/ketones in indoor air samples are given in Table 3. In all 28 hospitality venues, levels of tobacco smoke-specific VOC ranged from 0.7 to 450 μg/m3 in case of nicotine and from 0.8 to 40.5 μg/m3 in case of 3-ethenylpyridine. Levels of these tobacco smoke-specific VOC differed significantly between the sampling site categories (P<0.01) in contrast to TVOC (P=0.069). However, the highest concentrations of VOC were always observed in dicotheques/clubs.
Polycyclic Aromatic Hydrocarbons
The sum of all measured 16 gaseous and particle-bound PAH varied between 120 and 840 ng/m3 (Table 4). There was a significant difference of median concentrations of PAH sum across the sampling site categories (P=0.04) with increasing values from restaurants/cafés to discotheques/clubs. Considering individual substances, median concentrations in pubs/bars were not always higher than median concentrations in restaurants/cafés. But for all PAH substances, highest median concentrations were found in dicotheques/clubs.
Cadmium levels ranged from 1.2 to 27.0 ng/m3 with a significant difference between sampling site categories (P<0.01) and highest levels in discotheques. In restaurants/cafés, the median concentration was 2.6 ng/m3 (GM±GSD 2.7±1.2 ng/m3, range 1.2–7.7 ng/m3), in pubs/bars 3.7 ng/m3 (GM±GSD 4.7±1.4 ng/m3, range 1.7–27.0 ng/m3), and in discotheques/clubs 9.7 ng/m3 (GM±GSD 9.4±1.1 ng/m3, range 5.3–16.0 ng/m3).
Spearman's rank correlation coefficients for correlations between ETS marker levels are given in Table 5. Correlation plots for nicotine and 3-ethenylpyridine, PM2.5, TVOC and PAH sum are shown in Figure 2. The strongest correlations were observed for PM2.5 and tobacco smoke-specific markers nicotine and 3-ethenylpyridine. The lowest correlations were found between TVOC and PAH sum.
Significant correlations were observed for PM2.5, nicotine, and 3-ethenylpyridine with characteristics of sampling sites such as total guest capacity and estimated proportion of smoking as given in Table 6. The sum of PAH was moderately correlated with total guest capacity, estimated average number of persons, estimated average number of smokers, and estimated proportion of smoking. Significant correlations existed for cadmium levels and total guest capacity and estimated proportion of smoking. Formaldehyde levels were correlated to all characteristics except for estimated proportion of smoking. There was no correlation of TVOC with characteristics of sampling sites.
Our study was designed against the background of the actual debate on smoke-free legislation on one hand and on the other hand of lacking air quality data for hospitality venues in Germany. By analysing gas-phase constituents and components of ETS PM in indoor air, we demonstrated significant levels of toxic or carcinogenic ETS substances in hospitality venues. The exposure situation was worst in dicotheques.
Nicotine and 3-ethenylpyridine were measured as gas-phase compounds unique to tobacco smoke. The average nicotine concentrations of 21 μg/m3 in restaurants/cafés and of 54 μg/m3 in pubs/bars in our study were in the same order of magnitude as recently published nicotine levels of other countries (Table 7). However, average nicotine concentrations of 227 μg/m3 observed in discotheques were considerably higher than reported before for discotheques in Europe.
Comparable to our study, active sampling for 4 h was performed in a Finnish study (Hyvärinen et al., 2000) and in an international survey performed in France, Switzerland, United Kingdom, Japan, Korea, and United States (Bohanon et al., 2003). In Finland, mean 3-ethenylpyridine levels were 1.4 μg/m3 in restaurants, 1.6 μg/m3 in pubs and 6.3μg/m3 in discotheques (Hyvärinen et al., 2000). Average 3-ethenylpyridine concentrations ranged from 1.3 to 2.7 μg/m3 in restaurants of six countries (Bohanon et al., 2003). In contrast, we observed higher mean 3-ethenylpyridine levels of 4.1 μg/m3 in restaurants/cafés, 10.2 μg/m3 in pubs/bars, and 22.6 μg/m3 in discotheques.
A detailed exposure assessment of further VOC, aldehydes, and gaseous and particle-bound PAH as constituents of ETS has not been described in those recent studies cited in Table 7.
McNabola et al. (2006) measured benzene and 1,3-butadiene in two pubs before the introduction of the smoking ban in Ireland. They observed lower benzene levels of 2.4–6.6 μg/m3 compared to our results in pubs (7.1–64.0 μg/m3), but their results for 1,3-butadiene levels of 2.0–5.8 μg/m3 were in line with our observation of 0.3–5.0 μg/m3 in pubs. Benzene measurements in 26 pubs in Ireland (Goodman et al., 2007) gave a mean concentration of 18.8 μg/m3 in accordance with our result of mean benzene concentration of 17.3 μg/m3 in pubs.
Our data indicated a significant exposure to carcinogenic or probably carcinogenic substances to humans according to the classification of the International Agency for Research on Cancer (IARC) such as benzene, formaldehyde, benz(a)anthracene, benzo(a)pyrene, and cadmium with highest levels in discotheques.
The IARC summarised exposure conditions of extremely high concentrations of ETS in an experimental room with concentrations of 49 μg/m3 formaldehyde, 1390 μg/m3 acetaldehyde, 206 μg/m3 benzene, 26.7 ng/m3 benzo[a]pyrene, 25 ng/m3 pyrene, and 70.5 μg/m3 chrysene (WHO IARC, 2004). These experimental findings exceeded the levels observed in a real-world setting, such as those in our study, and therefore support the plausibility of our findings in terms of significant amounts of carcinogenic substances in hospitality venues where smoking is allowed.
Our results of PM2.5 concentrations corresponded well with findings of recent studies in other countries for restaurants/cafés and pubs/bars (Table 8). In addition, our study demonstrated that PM2.5 exposure in discotheques exceeded considerably levels observed in other hospitality venues. Median PM mass concentrations of about 800 μg/m3 with highest levels above 4000 μg/m3 in single locations as detected in discotheques might be ascribed to failures in sampling and analytical procedures. However, Edwards et al. (2006) reported in their study of pubs PM2.5 levels above 1000 μg/m3 in the smokiest venues. Under experimental conditions, respirable suspended particle concentrations above 4000 μg/m3 have been measured in a room after smoking of 120 cigarettes during 9 h (WHO IARC, 2004). Thus, we evaluated our findings of PM exposure in discotheques plausible since the estimated proportion of smoking ranged from 50% to 60% and the estimated average number of smokers from 90 to 600 depending on the dimensions of the discotheque. Furthermore, we applied gravimetry and aerosol spectrometry as two distinct methods of measurement, which yielded strongly correlated PM2.5 levels.
It might be judged as a limitation of our study that a true random sample of hospitality venues could not be achieved. Reasons for this were predominantly that sampling in rather small and heavily crowded pubs or clubs was not feasible due to the amount of measurement devices. In addition, managers of some selected locations did not give their consent. Among several reasons indicated for declining, one reason was sponsorship by the tobacco industry.
Our study has several advantages: (1) we performed a comprehensive air quality monitoring with a sampling time of 4 h during main visiting hours, thus depicting real life-exposures to ETS components. (2) We performed sampling in three kinds of hospitality venues to assess the wide spectrum of ETS exposure and to compare the situation in pubs and discotheques as major exposure sites of adolescents and young adults with that in restaurants. (3) We measured two ETS-specific markers, nicotine and 3-ethenylpyridine. Correlation analyses indicated that ETS was the main source of PM2.5, TVOC, PAH sum, and cadmium in indoor air of hospitality venues where smoking was not restricted. These findings were consistent with calculations of Repace (2004) that 95% of the indoor particulate PAH could be attributed to tobacco smoke. Recent studies on the effect of a smoking ban demonstrated a substantial decline of ETS components in pubs. In Ireland, a percent change for PM2.5 of 83% from a pre-ban level of 36 μg/m3 to a post-ban level of 6 μg/m3 and for benzene of 79% from a pre-ban level of 19 μg/m3 to a post-ban level of 4 μg/m3 was observed (Goodman et al., 2007). In the United States, a percent change for PM3.5 (RSP) of 96% from a pre-ban level of 179 μg/m3 to a post-ban level of 8 μg/m3 and for particulate PAH of 91% from a pre-ban level of 65 ng/m3 to a post-ban level of 6 ng/m3 was demonstrated (Repace et al., 2006). Data on indoor concentrations of PM2.5 in non-smoking locations in Germany are scarce. Average PM2.5 levels in dwellings are 20–30 μg/m3 according to a recent review (Fromme, 2006).
In conclusion, in contrast to other countries exposure to ETS in hospitality venues is still a public health concern in Germany. Our data of exposure to a variety of toxic or carcinogenic ETS substances in non-negligible concentrations emphasises the need to take action. From a public health point of view, effective measures to protect patrons and staff from ETS exposure are necessary. Our results do not support the exemption of any kind of hospitality venue such as non food-serving pubs and bars from smoke-free legislation.
We thank the managers of the hospitality venues for their cooperation and Andreas Pfaller for excellent analyses of aldehyde and ketone concentrations. The study was funded by the Bavarian State Ministry of the Environment, Public Health and Consumer Protection.
About this article
Archives of Environmental Contamination and Toxicology (2015)