The impact of second-hand smoke on nitrogen oxides concentrations in a small interior

Nitrogen oxides (NOx), especially nitrogen dioxide (NO2), are among the most hazardous forms of air pollution. Tobacco smoke is a main indoor source of NOx, but little information is available about their concentrations in second-hand smoke (SHS), particularly in small indoors. This study presents data of NOx and its main components nitric oxide (NO) and NO2 in SHS emitted by ten different cigarette brands measured in a closed test chamber with a volume of 2.88 m3, similar to the volume of vehicle cabins. The results show substantial increases in NOx concentrations when smoking only one cigarette. The NO2 mean concentrations ranged between 105 and 293 µg/m3, the NO2 peak concentrations between 126 and 357 µg/m3. That means the one-hour mean guideline of 200 µg/m3 for NO2 of the World Health Organization was exceeded up to 47%, respectively 79%. The measured NO2 values show positive correlations with the values for tar, nicotine, and carbon monoxide stated by the cigarette manufacturers. This study provides NO2 concentrations in SHS at health hazard levels. These data give rise to the necessity of health authorities’ measures to inform about and caution against NOx exposure by smoking in indoor rooms.


Scientific Reports
| (2021) 11:11703 | https://doi.org/10.1038/s41598-021-90994-x www.nature.com/scientificreports/ Several studies investigated indoor air pollution, including NO x respectively NO 2 concentrations, caused by ovens (primarily gas stoves) or fireplaces for heating or cooking [13][14][15][16][17][18][19] . Some studies focused on tobacco smoke reporting a relatively lower influence on NO x concentrations by burning tobacco products in normal-sized rooms or houses [20][21][22] . The impact of second-hand smoke (SHS) on NO x burden in very small indoors like smoking cabins, telephone cells, or cabins of vehicles, for example, remains widely unclear. This study presents the results of the NO, NO 2 , and NO x (NO + NO 2 ) investigations in SHS of ten different cigarette brands with different strengths, additives, and origin in a 2.88 m 3 measuring chamber. The measurements were realized as part of the Tobacco Smoke Particles and Indoor Air Quality (ToPIQ) studies 23 at two PM investigations. The PM-related results of these studies are already described and published 24,25 .

Material and methods
Tobacco products. The concentrations of NO, NO 2 , and NO x in SHS were measured of nine commercial cigarette brands (named cigarette A to I) and the 3R4F reference cigarette (Kentucky Tobacco Research and Development Center, University of Kentucky, USA) 26 . As the measurements took place during two PM investigations, the reference cigarettes (RC) were termed RC1 24 and RC2 25 . The cigarette brands A, C, and E were bought at the International Airport of Dubai, United Arab Emirates (UAE), while the brands B, D, and F were from the International Airport of Frankfurt, Germany 25 . Additionally, three mentholated cigarette brands (G, H, I) were tested, purchased at the central station, Frankfurt, Germany 24 . The brand names of the cigarettes are given in Table 1. The cigarette brands differ in amounts of tar, nicotine, carbon monoxide (CO), and PM 10 (Table 1). Further information on the cigarettes from Germany is available from the tobacco additives database of the Federal Ministry of Food and Agriculture of Germany 27 .
Test chamber. All measurements took place in a test chamber with an internal volume of 2.88 m 3 . During the experiments, the vents for the supply and exhaust air were closed to minimize air exchange. The test chamber is placed in a laboratory room of the Institute of Occupational Medicine, Social Medicine and Environmental Medicine, Goethe University Frankfurt. The institute is located in an urban area but not near a traffic road. That avoided high NO x concentrations in ambient air by road traffic (see baseline evaluation).

Ambient nitrogen oxide monitor (NO x monitor).
To ascertain the NO, NO 2 , and NO x (NO + NO 2 ) concentrations of SHS, the ambient nitrogen oxide monitor APNA-370 of HORIBA, Ltd. (Kyoto, Japan) was applied 28 . By using a cross-flow modulated semi-decompression chemiluminescence method, NO 2 concentrations were internally calculated from those of NO and NO x . All measurement values were recorded in the unit ppm (parts per million) every three minutes. The not mixed sample air was collected at a point 40 cm above the burning tobacco product and 170 cm above the floor of the test chamber.
Automatic environmental tobacco smoke emitter. SHS 29 . This programmable microprocessor-controlled smoke pump imitated the smoker by moving a 200 ml glass syringe connected with the mouthpiece of the tobacco product via a polyamide tube. Moving the syringe led to puffing the tobacco product. Two valves controlled the air stream and pressed the www.nature.com/scientificreports/ mainstream smoke after each puff into the closed chamber. Between the puffs, the tobacco product smoldered and produced side-stream smoke. In doing so, no person was exposed to the generated SHS.
Smoking protocol. The number (n) of investigated cigarettes of each brand varied between 15 and 41 (Table 1). According to the ToPIQ studies [23][24][25] , all cigarettes were smoked following a modified protocol. Puff volume was 40 ml, and the flow rate was 13 ml/s. After two ignition puffs, each cigarette was smoked in the combustion phase with seven puffs and a frequency of two puffs/min. Subsequently, the post-combustion phase followed after the extinguishing of the cigarette. After ten minutes in total, the chamber was ventilated with outdoor air by an industrial radial fan for at least five minutes to clean the air.
Baseline evaluation. As the NO x monitor detected data permanently, continuous measuring data of 61 h between the two measuring campaigns without SHS generation were chosen to determine the baseline values of NO, NO 2 , and NO x .
Data processing. The NO x monitor provided every three minutes measuring data. Therefore, in the 10-min combustion and post-combustion phase, three values of NO, NO 2 , and NO x , respectively, per investigated cigarette could be taken into account for the following data processing. The mean concentrations (C mean ) of these three measuring values of NO, NO 2 , and NO x were calculated. Additionally, the highest values of NO, NO 2 , and NO x were considered as peak values (C peak ) for each cigarette. For the statistical analysis, all C mean and C peak values were tested for outliers (Grubbs' test). Sixteen outliers were detected and subsequently excluded from further statistical tests. All data were normally distributed. To compare the data of all investigated cigarettes, a one-way analysis of variance (ANOVA) including Tukey's multiple comparison test was performed. The associations of the C mean values of NO, NO 2 , and NO x with concentrations of tar, nicotine, CO, and PM 10 were examined by using correlation analysis (Spearman) and linear regression. PM 10 is classified by the US Environmental Protection Agency (EPA) as inhalable particles ≤ 10 µm and includes the fraction of the fine inhalable particles ≤ 2.5 µm (PM 2.5 ) 30 . The measured PM 10 C mean values of RC1 were lower than those of RC2. Therefore, it was necessary to adjust the PM 10 data of RC1 and the associated cigarette brands G, H, and I on the PM 10 data of RC2 by statistical data transformation (Y = K*Y) using the factor K = 1.47.
Statistical analyses were performed using GraphPad Prism software (version 8 for Windows, GraphPad Software, La Jolla California USA, www. graph pad. com).

Data conversion.
For the comparison of the in this study measured values with common used limit values or guidelines, the data of NO and NO 2 were converted in µg/m 3 using the formula 31 : c concentration, ppb parts per billion, MW molecular weight, MW NO = 30.01 g/mol; MW NO 2 = 46.01 g/mol.
As the NO x monitor display the data in ppm, all values were multiplied by 1000 to convert to the unit ppb (parts per billion). Table 2 and Fig. 1 present the NO, NO 2 , and NO x results of C mean and C peak of all investigated cigarette brands.

Results
Regarding NO, the range of the C mean values was from 132 to 422 ppb (equal to 162 µg/m 3 to 518 µg/m 3 at 1013.25 mbar and 25 °C). For NO 2 , the C mean values ranged from 56 to 156 ppb (equal to 105 µg/m 3 to 293 µg/ m 3 at 1013.25 mbar and 25 °C). Looking at the C mean data of NO x , the values ranged from 247 to 499 ppb.
To compare the measured NO x values in SHS with the usually NO x concentration in indoor air at the study location, we took continuous data of 61 h where no investigation was done into account recorded between the two measurement campaigns. The mean of the thus collected data resulted in the baselines for NO = 0.072 ppb (equal to 0.1 µg/m 3 at 1013.25 mbar and 25 °C), NO 2 = 5.08 ppb (equal to 9.6 µg/m 3 at 1013.25 mbar and 25 °C) and NO x = 5.15 ppb.
A detailed overview of associations of the C mean values of NO, NO 2 , and NO x with concentrations of tar, nicotine, CO, and PM 10 shows Table 3. Additionally, Fig. 2 presents the correlations between NO, NO 2 , and NO x and the stated amounts of tar (A), nicotine (B), CO (C), and the measured values of PM 10 (D). The measured NO data are negatively correlated with the concentrations of tar, nicotine, and CO as specified by the cigarette manufacturers. The concentrations of NO 2 correlates positively with the stated values of tar, nicotine, and CO. NO x correlated negatively with the concentrations of tar, nicotine, and CO, but in the case of CO without significance. The measured PM 10 values show no correlations with the concentrations of NO, NO 2 , and NO x .

Discussion
The measured indoor baseline concentration revealed for NO 2 a mean value of 9.6 µg/m 3 (5.08 ppb). That is in line with previous studies on NO 2 indoor concentrations 20 . Our findings show remarkable rises of NO x in small indoors by the smoke of only one cigarette. Of all ten tested cigarette brands, the measured C mean values of five brands exceeded the WHO one-hour mean guideline of 200 µg/m 3 for NO 2 6 by 9% to 47%. The remaining five brands showed between 1.5% and 47% lower NO 2 values compared to the WHO guideline. The measured NO 2 C mean value of all examined cigarettes was 215 µg/m 3 (119 ppb) and, therefore, exceeded the WHO guideline by www.nature.com/scientificreports/ 8%. Regarding the detected C peak values of NO 2 , six brands exceeded the WHO guideline even in a range from 12 to 79%. Four brands were 12% to 37% below the guideline. The NO 2 C peak mean value of all cigarettes was 264 µg/ m 3 (140 ppb) and consequently 32% higher than the WHO guideline value. The WHO state unambiguously "that NO 2 -at short-term concentrations exceeding 200 µg/m 3 -is a toxic gas with significant health effects" 6 . It should be borne in mind that the WHO annual mean guideline for NO 2 is with 40 µg/m 3 even stricter 4 . There are other guidelines often following the WHO indoor and ambient guidelines for NO 2 , but some differ 32 . At Health Canada, for example, there is an indoor short-term limit value of 170 µg/m 3 and a long-term limit value of 20 µg/m 3 based on toxicological data 33 . Several former studies dealt with NO x emitted by tobacco products in normal-sized indoor rooms. Cyrys et al. 20  C mean NO (ppb) C mean NO (µg/ m 3 ) C peak NO (ppb) C peak NO (µg/ m 3 ) C mean NO 2 (ppb) C mean NO 2 (µg/ m 3 ) C peak NO 2 (ppb) C peak NO 2 (µg/ m 3 ) C mean NO x (ppb) C peak NO x (ppb) BL 0.11 (0.15) 0.1 n/a n/a 5.08 (2.14) 9.6 n/a n/a 5. 15  www.nature.com/scientificreports/ and 15 µg/m 3 , respectively, an increase of 18% in smokers' homes, and an increase of 41% in households using gas for cooking, the main indoor source of NO 2 . Additionally, the authors found that outdoor sources can influence indoor NO 2 levels more than indoor sources depending on the location and season of year. However, they differentiated between smoking and non-smoking in the living room (including the use or non-use of gas in the household) and reported on their influence on indoor NO 2 levels in general. They did not report on the influence of one single combustion event (burning cigarette or gas cooking, e.g.) and how this can boost NO 2 concentration in the indoor air temporarily. Slightly lower NO 2 mean concentrations (2-3 weeks averaged) were found Table 3. Spearman correlations of the C mean values of NO, NO 2 , and NO x with concentrations of tar, nicotine, CO (as specified by the cigarette manufacturers), and PM 10 . P values show the significance of the correlations. ns = not significant (P ≥ 0.05). * = significant (P = 0.01 to 0.05). ** = very significant (P = 0.001 to 0.01). *** = very significant (P < 0.001).  www.nature.com/scientificreports/ in Scottish and Irish homes: 12.8 µg/m 3 (6.82 ppb) in smokers' homes and 16.9 µg/m 3 (9.01 ppb) in households where gas was used for cooking 22 . In a 20 m 2 room with a volume of 57 m 3 , water pipes were smoked in four-hour sessions, and NO and NO 2 concentrations were measured 21 . The authors reported for the smoking sessions a NO mean concentration of 100 ppb and a NO 2 mean concentration of 60 ppb, meaning 123 µg/m 3 and 113 µg/m 3 , respectively. That indicates that smoking significantly increases NO x concentrations also in normal-sized indoor rooms. Our study found remarkable mean values (up to 518 µg/m 3 for NO and 293 µg/m 3 for NO 2 ) and peak values (up to 716 µg/m 3 for NO and 357 µg/m 3 for NO 2 ) caused by smoking of only one cigarette. Admittedly, the measuring chamber with an indoor volume of 2.88 m 3 corresponds more to vehicle indoor volumes than indoor volumes of living rooms. However, it can be assumed that chain-smoking or simultaneously smoking of several cigarettes (by several smokers) will increase NO x concentrations also in larger rooms in a similar way. Therefore, this should be in focus for future studies. The present findings show statistically significant correlations between the strength of a cigarette brand (amount of tar, nicotine, and CO as stated by the manufacturers) and the measured data of NO, NO 2 , and NO x . Interestingly, the lower the tar, nicotine, and CO values, the higher were the measured NO and NO x levels, while the measured NO 2 levels correlated positively with the cigarette strength but with lower significance. The higher the combustion temperature, the more NO x will be generated 4 . Also, the content of bound nitrogen in the tobacco product in the form of nitrate or nitrosamine compounds, e.g., could influence the NO x amount in tobacco smoke 34,35 . Possibly, "lighter" cigarettes burn at smoking with higher combustion temperatures or contain more bound nitrogen. Further investigations on more numerous cigarette brands with various strengths should also examine the burning temperature and the nitrogen amount of the tobacco product. It was reported that NO and NO x concentrations in the mainstream smoke of cigarettes correlated positively with their strength 36 , whereby generation of mainstream smoke took place in a smoking machine following ISO machine-smoking conditions with an as short as possible distance to the NO x analyzer 37 . This set-up is rather comparable to a smokers' NO x exposure inhaling mainstream smoke. In opposite to this, we simulated the situation of a person exposed to SHS near a burning cigarette. Other studies focusing on NO x in mainstream smoke detected mainly NO but almost no NO 2 38-40 . Among others, it was assumed that the reducing conditions near the glowing zone of the cigarette favor the formation of NO as the lower oxide of nitrogen or that reactive volatile organic compounds (VOCs) in the tobacco smoke react with NO 2 38 . Some NO 2 was detected in the mainstream smoke from the initial puff, but not from the following puffs, while NO 2 was detected continuously in the side-stream smoke 40 . Only when using a Cambridge filter pad between the cigarette and the analyzer NO 2 was observed for each puff. The on the pad sampled smoke of the previous puffs could have interacted with NO in the smoke forming NO 2 . As the pad also acted as a barrier between the cigarette and the analyzer, the smoke could age, and, consequently, NO 2 values could increase 40 . That indicates that the more toxic NO 2 is mainly detectable in side-stream smoke and aged smoke but less in mainstream smoke and during a prompt measurement. Since SHS is mainly composed of side-stream smoke (85%) 41 , the detection of NO 2 in SHS is plausible. In addition, the smoke generated in this study had time to age.
A strength of the present study was that the used measuring set-up in the test chamber allowed to create and investigate SHS in a reproducible way without the exposition of any person to tobacco smoke. A methodological limitation was the low frequency of only three measurements per cigarette by the NO x monitor used. Therefore, the real peak values of NO, NO 2 , and NO x could not be detected in all investigated cigarettes. That resulted possibly in slightly too low C mean and C peak measurement values.

Conclusion
In the last years, the discussion about NO x and especially NO 2 generated by diesel vehicles in urban areas has made massive waves. To name is the so-called Dieselgate scandal commenced in September 2015 42 . The focus was set on ambient NO x formation, certainly, influencing also indoor burden by NO x . But, smoking is a not neglecting source of NO x in indoor rooms. The present study provides NO 2 concentrations in SHS generated by smoking cigarettes in small indoors at levels known to be a health hazard. Keeping in mind that the used test chamber (2.88 m 3 ) has a similar volume to vehicle cabins 43 , smoking in cars can lead to a hazardous increase of NO 2 concentration. This risk multiplies accordingly if more than one cigarette is smoked (e.g., chain smoking), there is more than one smoker in the car, or the car is driven with closed windows without sufficient ventilation. Therefore, health authorities' measures are useful and required to inform about and caution against NO x exposure by smoking in cars and other indoor rooms.