1.1.1. Secondhand smoke

During smoking of cigarettes, cigars, pipes and other tobacco products, in addition to the mainstream smoke drawn and inhaled by smokers, a stream of smoke is released between puffs into the air from the burning cone. Once released, this stream (also known as the sidestream smoke) is mixed with exhaled mainstream smoke as well as the air in an indoor environment to form the secondhand smoke to which both smokers and nonsmokers are exposed. A small additional contribution to the smoke issues from the tip of the cigarette and through the cigarette paper during puffing and through the paper and from the mouth end of the cigarette between puffs (IARC, 1986; NRC, 1986; US EPA, 1992). Thus, secondhand tobacco smoke is composed of aged exhaled mainstream smoke and diluted sidestream smoke.

Secondhand tobacco smoke contains a variable proportion of exhaled mainstream smoke ranging from 1 to 43% (Baker & Proctor, 1990). Because of its rapid dilution and dispersion into the indoor environment, secondhand tobacco smoke acquires different physicochemical properties to those of mainstream smoke and sidestream smoke and the concentrations of the individual constituents are decreased. The principal physical change is a decrease in the proportion of smoke constituents found in the particulate phase as opposed to the vapour phase of the smoke. The median particle size of secondhand tobacco smoke is subsequently smaller than that of the particles of mainstream smoke. The principal chemical change is in the composition (i.e. in the relative quantities of the individual constituents present); this is caused by differences in the ways in which individual constituents respond to ventilation and to contact with indoor surfaces. There is some indication that chemical transformation of reactive species also occurs.

The effects of exposure to secondhand tobacco smoke, or involuntary (passive) smoking, cannot be estimated from any individual constituents. Secondhand tobacco smoke is actually a complex mixture, containing many compounds for which concentrations can vary with time and environmental conditions. Cigarette smoking is the main source for involuntary exposure because it is by far the most prevalent form of tobacco smoking although specific patterns may differ between countries. Emissions of sidestream smoke in indoor environments with low ventilation rates can result in concentrations of toxic and carcinogenic agents above those generally encountered in ambient air in urban areas (IARC, 1986; Jenkins et al., 2000).

Studies on the complex composition of secondhand tobacco smoke in ‘real world’ conditions have been limited partly because of the presence of additional sources of secondhand smoke constituents. Therefore compositional and physical studies of secondhand tobacco smoke have often been performed in environmental chambers (also known as a ‘controlled experimental atmosphere’). The disadvantage of the controlled experimental atmosphere is that it does not reflect real life situations. The studies of the chemical composition of secondhand tobacco smoke, either in a controlled experimental atmosphere or in the field, have been limited. This is mainly because there are still no standardized criteria for the development of experimental atmospheres that represent second-hand tobacco smoke (Jenkins et al., 2000).

Respirable suspended particles (only those particles that are small enough to reach the lower airways of the human lung) can exist in many forms in indoor air: those resulting from secondhand tobacco smoke are present in the form of liquid or waxy droplets. They are smaller than the particles in mainstream smoke. The mass median diameter of mainstream smoke particles averages 0.35–0.40 µm. Gravimetric determination indicates that the respirable suspended particles of secondhand tobacco smoke in typically encountered environments may comprise one-third of the respirable suspended particles in indoor air. However, in some environments, this fraction may be as much as two-thirds (Jenkins et al., 2000).

Respirable suspended particle concentrations of 4091 µg/m³ were measured in an experimental room in which 120 cigarettes were smoked during 9 hours in 1 day for the evaluation of exposure to benzene and other toxic compounds (Adlkofer et al., 1990).

Worldwide in indoor environments where people smoke, the mean levels of respirable suspended particles ranged from 24 to 1947 µg/m³. Background levels of respirable suspended particles depend on many factors including local vehicular traffic patterns, quality of ventilation systems and the presence of other sources (e.g. cooking and wood-burning stoves). Comparisons between smoking and nonsmoking locations revealed up to threefold higher concentrations of respirable suspended particles in smoking areas. The US EPA-proposed maximal level for fine particles in outdoor ambient air (65 µg/m³ particulate matter, that is 2.5 µm or smaller in size, for 24 h) is frequently exceeded in indoor situations where people are smoking (Jenkins et al., 2000).

Besides respirable suspended particles and nicotine (see Section 1.2), carbon monoxide (CO) has been the most extensively studied constituent of secondhand tobacco smoke. The contemporary commercial cigarettes in the USA deliver approximately 15 mg CO in mainstream smoke and an additional 50 mg in sidestream smoke (Jenkins et al., 2000). In an indoor environment, CO concentrations are rapidly diluted. The measured mean concentrations reported for CO in offices, other workplaces, functions and public gatherings, transportation, restaurants and cafeterias, bars and taverns where people smoke ranged from 0.2 to 33 ppm. The American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) standard for CO concentration in indoor air is 9 ppm (Jenkins et al., 2000).

The mean levels of nitric oxide (NO) in indoor areas were reported from not detected to 500 ppb and those of nitrogen dioxide (NO2) from not detected to 76 ppb (Jenkins et al., 2000).

There are a number of studies that have addressed the composition of secondhand tobacco smoke beyond the ‘common’ constituents such as nicotine, CO and respirable suspended particles. A few of these are shown in Table 1.1 (Eatough et al., 1989; Löfroth et al. 1989; Higgins et al., 1990; Löfroth, 1993; Martin et al., 1997). The focus of these studies was primarily on vapour-phase constituents. Vapour phase represents the bulk of the mass of secondhand tobacco smoke whereas the respirable suspended particle-related constituents are present at very low concentrations that are very difficult to quantify. For example, if the levels of respirable suspended particles are in the range of 20 to 1000 µg/m³, constituents of the particulate phase present at concentrations of 1–100 ppm in the particles themselves will be present at airborne concentrations from 20 pg/m³ to 100 ng/m³. These are very low concentrations for detection by any sampling and analysis method (Jenkins et al., 2000).

Table 1.1

Concentrations (in µg/m³) of selected constituents of secondhand tobacco smoke in some experimental and real-life situations.

As can be seen from the data in Table 1.1, the field studies show considerable variation in the measured levels of constituents of secondhand tobacco smoke. Similar concentrations of benzene and isoprene to those shown in Table 1.1 were reported in a smoke-filled bar (from 26 to 36 µg/m³ and 80–106 µg/m³, respectively), although the nicotine levels were much lower (22 µg/m³). The concentration of 1,3-butadiene measured in the smoke-filled bar was from 2.7 to 4.5 µg/m³ (Brunnemann et al., 1990). In a field study of 25 homes of smokers, Heavner et al. (1995) estimated that the median fraction of benzene contributed by secondhand tobacco smoke was 13% (ranging from 0 to 63%).

In a study of six homes of smokers, secondhand tobacco smoke was found to make a substantial contribution to the concentrations of 1,3-butadiene (Kim et al., 2001).

The levels of carbonyl compounds measured in an experimental room under extremely high concentrations of secondhand tobacco smoke (Adlkofer et al., 1990) were: formaldehyde, 49 µg/m³; acetaldehyde, 1390 µg/m³ and propionaldehyde, 120 µg/m³. The concentrations of other constituents of secondhand tobacco smoke were: nicotine, 71 µg/m³; benzene, 206 µg/m³; benzo[a]pyrene, 26.7 ng/m³; pyrene, 25 ng/m³ and chrysene, 70.5 ng/m³.

Benzo[a]pyrene was also detected in natural environments containing secondhand tobacco smoke, with concentrations ranging from not detected to 3.6 ng/m³ (or up to 3.35 ng/m³ when the background concentration was subtracted) (Jenkins et al., 2000). Trace levels of some other polycyclic aromatic hydrocarbons (PAHs; such as naphthalene, chrysene, anthracene, phenanthrene and benzofluoranthenes) were also reported. The vapour-phase 2- to 3-ring PAHs predominate quantitatively over the higher-ring system PAHs.

The levels of N-nitrosodimethylamine (NDMA) measured in the field (e.g. in work-rooms, conference rooms, restaurants and bars where people smoked) ranged from less than 10 ng/m³ to 240 ng/m³ (Jenkins et al., 2000). In unventilated offices in which 11–18 cigarettes were smoked during a 2-h period, up to 8.6 ng/m³ N-nitrosodiethylamine (NDEA) and up to 13 ng/m³ N-nitrosopyrrolidine (NPYR) were measured.

The N′-nitrosonornicotine (NNN) concentrations measured in a poorly ventilated office where heavy smoking of cigarettes, cigars and pipes took place ranged from not detected to 6 ng/m³ and those of 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) from not detected to 13.5 ng/m³ (Klus et al., 1992). The upper levels reported by Klus et al. (1992) and by Adlkofer et al. (1990) for the ‘heavily smoked rooms’ (11 cigarettes smoked in 2 h in a 84 m² office) were somewhat lower than those measured by Brunnemann et al. (1992): NNN concentrations ranged from not detected to 22.8 ng/m³ and NNK concentrations between 1.4 and 29.3 ng/m³, measured inside bars, restaurants, trains, and a car, an office and a smoker's home (Brunnemann et al., 1992).

The effects of cigar smoking on indoor levels of CO, respirable suspended particles and particle-bound PAH particles were investigated in an office where several brands of cigar were machine smoked, in a residence where two cigars were smoked by a person and at cigar social events where up to 18 cigars were smoked at a time. The average concentrations of CO at cigar social events were comparable with, or larger than, those measured on a main road during rush-hour traffic. A mass balance model developed for predicting secondhand tobacco smoke was used in this study to obtain CO, respirable suspended particle and PAH emission. These factors show that cigars can be a stronger source of CO than cigarettes. In contrast, cigars may have lower emissions of respirable suspended particles and PAHs per gram of tobacco consumed than cigarettes, but the greater size and longer smoking time of a single cigar results in greater total respirable suspended particle and PAH emission than from a single cigarette (Klepeis et al., 1999). Nelson et al. (1997) tested six brands of cigar and the yields of respirable suspended particles averaged 52 mg/cigar. Yields of CO, nitric oxide (NO) and nitrogen dioxide (NO2) averaged 32, 10.5 and 2.1 mg/cigar, respectively, and that of volatile organic compounds (VOC) (analysed with gas chromatograph-flame ionization detection (FID)) was estimated to be 340 mg/cigar (propane equivalent). Ratios of secondhand tobacco smoke respirable suspended particles to the surrogate standards for the particulate markers ultra-violet particulate matter (UVPM) and fluorescent particulate matter (FPM) were 6.5 and 27.8, respectively. Another particulate marker, solanesol, made up 2.1% of the particles from cigars. For the two gas-phase markers, the ratio of 3-ethenyl pyridine to other gas-phase species was more consistent than ratios involving nicotine.

The comparative analysis of the composition of secondhand tobacco smoke from Eclipse (a cigarette that primarily heats tobacco rather than burning it) and from four commercial cigarettes with a wide range of Federal Trade Commission (FTC) yields is shown in Table 1.2 (Bombick et al., 1998). Eclipse contributed similar amounts of CO to secondhand tobacco smoke to those contributed by burning-tobacco cigarettes but contributed 86–90% less respirable suspended particles. Commercial cigarettes, however, contributed a similar amount of constituents to secondhand tobacco smoke, regardless of their ranking on the FTC scale.

Table 1.2

Mean concentrations (in µg/m³) of selected components of secondhand smoke of four commercial cigarette brands and Eclipse measured in a chamber with a controlled experimental atmosphere.

1.1.2. Exhaled mainstream smoke

Baker and Proctor (1990) estimated that exhaled mainstream smoke contributes 3–11% of CO, 15–43% of particles and 1–9% of nicotine to secondhand tobacco smoke. Non-inhaling smokers can contribute larger amounts. There is little information on how much exhaled mainstream smoke contributes to the overall composition of secondhand tobacco smoke except that the contribution to the particulate phase is more significant than that to the vapour phase (see monograph on tobacco smoke/Section 4 — breath compounds).

1.1.3. Sidestream smoke

The composition of cigarette sidestream smoke is similar to that of mainstream smoke. However, the relative quantities of many of the individual constituents of sidestream smoke are different from those found in mainstream smoke. Also, the absolute quantities of most of the constituents released in sidestream smoke differ from those delivered in mainstream smoke (Jenkins et al., 2000).

Like mainstream cigarette smoke, sidestream smoke contains many compounds that are emitted as gases and particles. The distinction between particle and vapour-phase constituents is appropriate for those constituents that are non-volatile (e.g. high-molecular-weight organic compounds and most metals) and those that are clearly gases (e.g. CO). Constituents with appreciable vapour pressure (i.e. most of the constituents of tobacco smoke) can be found in both the particulate phase and the vapour phase of cigarette smoke. The term ‘semivolatiles’ has been used to describe such constituents. The degree to which these compounds are distributed between the particle and vapour phases is determined by their volatility (and stability) and the characteristics of their environment. These constituents are distributed preferentially in the particulate phase in highly concentrated smokes, such as those inhaled by smokers, and preferentially in the particle and vapour phases in highly diluted smokes, such as those encountered by involuntary smokers. The phase distribution and the ultimate fate of any given constituent released into the ambient environment is likely to differ depending upon ambient conditions and upon the chemical or physical properties of that constituent.

The manner in which cigarettes are smoked greatly influences their mainstream delivery and sidestream emissions. For the different machine-smoking protocols referred to in this monograph, see Table 1.9 in the monograph on tobacco smoke. Particulate matter that is released in mainstream smoke during active smoking enters the respiratory tract largely intact, whereas the particulate matter in sidestream smoke is available for inhalation only after dilution in ambient air and after the physical and chemical changes that occurred during that dilution. However, conventional analysis of sidestream smoke provides only information on the quantities of individual smoke constituents released into the air. Moreover, the methods for analysis of sidestream smoke are not as well defined as those for mainstream smoke (Jenkins et al., 2000).

Table 1.9

Summary of selected studies on the effects of workplace smoking bans and restrictions on exposure to secondhand tobacco smoke.

As they leave the cigarette, sidestream smoke particles are initially slightly smaller than mainstream smoke particles (geometric mean diameter, 0.1 µm versus 0.18 µm; Guerin et al., 1987). The natural dissipation rates of sidestream smoke particles dispersed in an experimental chamber were studied from the standpoint of a static atmosphere and were expressed as half-lives of residence in the air. The half-lives for particles with diameters less than 0.3 µm, 0.3–0.5 µm and 0.5–1 µm were found to be 25.5, 12.8 and 4.9 h, respectively. Total particulate matter decreased by half over 6.2 h (Vu Duc & Huynh, 1987). However in real-life situations the ageing of sidestream smoke over several minutes may lead to an increase in particle size of secondhand tobacco smoke due to the coagulation of particles and the removal of smaller particles that attach to surfaces in the environment. Sidestream smoke is produced at generally lower temperatures and with a very different oxygen flux to that of mainstream smoke.

The ratio of sidestream smoke to mainstream smoke is customarily used to express the distribution of individual constituents between the two smoke matrices. The distribution of specific components is dependent on their mechanism of formation. Higher ratios of sidestream smoke to mainstream smoke between cigarettes or smoking conditions generally reflect a lower mainstream smoke delivery with no significant change in sidestream smoke delivery. Under similar smoking conditions, filter-tipped cigarettes will have lower mainstream smoke yields than their untipped analogues. Sidestream smoke yields will not vary greatly, because they reflect the weight of tobacco burned during smouldering. In general, more tobacco is burned during smouldering than during puffing (Guerin et al., 1987).

Adams et al. (1987) reported that sidestream smoke contains more alkaline and neutral compounds than mainstream smoke (the pH of the sidestream smoke of cigarettes with a wide range of FTC yields averaged 7.5 [7.2–7.7], whereas the pH of mainstream smoke averaged 6.1 [6.0–6.3]). The differences are due to temperature during burning and mechanisms of chemical transfer (release) from unburned tobacco.

Many constituents of sidestream smoke belong to chemical classes known to be genotoxic and carcinogenic. These include the IARC group 1 carcinogens benzene, cadmium, 2-aminonaphthalene, nickel, chromium, arsenic and 4-aminobiphenyl; the IARC group 2A carcinogens formaldehyde, 1,3-butadiene and benzo[a]pyrene; and the IARC group 2B carcinogens acetaldehyde, isoprene, catechol, acrylonitrile, styrene, NNK, NNN and lead among others.

Adams et al. (1987) determined the levels of selected toxic and carcinogenic agents in the mainstream and sidestream smoke of four different types of US commercial cigarette brands — untipped and filter-tipped — with a wide range of FTC yields. In this study, smoke was generated by a machine using the standard FTC method. The concentrations of all agents except NNN were higher in the sidestream than in the mainstream smoke of both untipped and filter-tipped brands. The tar yields in sidestream smoke ranged from 14 to 24 mg per cigarette (similar to the range reported by Ramsey et al., 1990) and were, on average, 5.3 times higher than those in mainstream smoke. The highest sidestream smoke/mainstream smoke ratio for tar was calculated for ultra low-yield cigarettes (ratio, 15.7) and the lowest for untipped cigarettes (ratio, 1.12). The mainstream smoke yields are strongly affected by variables that only slightly affect sidestream smoke yields (Guerin et al., 1987). The ratios of sidestream smoke to mainstream smoke for nicotine, CO and NNK for the ultra low-yield brand were 21.1, 14.9 and 22.3, respectively. The highest emissions of nicotine, NNK and NNN were measured in the sidestream smoke of untipped cigarettes (4.62 mg, 1444 ng and 857 ng, respectively). In mainstream smoke, these values were 2.04 mg, 425 ng and 1007 ng, respectively. The levels of volatile N-nitrosamines in sidestream smoke greatly exceeded those measured in mainstream smoke (e.g. 735 ng versus 31.1 ng NDMA per untipped cigarette and 685 ng versus 4.1 ng NDMA for ventilated filter-tipped cigarettes; average ratio, 95). The average ratio of sidestream smoke to mainstream smoke for the carcinogen NPYR was 10. The authors concluded that the availability of cigarettes with greatly reduced amounts of carcinogens in mainstream smoke had little bearing on the emissions of carcinogens in sidestream smoke.

Chortyk and Schlotzhauer (1989) compared the emissions of various smoke components for 19 low-yield brands of filter-tipped cigarettes with those measured for a reference high-yield untipped brand. It was found that low-yield cigarettes produced large quantities of tar in sidestream smoke, about equal to that of the high-yield cigarette. On an equal weight basis, the low-tar cigarettes emitted more of these hazardous compounds into sidestream and secondhand smoke than did the high-tar cigarette.

The yields of various constituents of sidestream smoke of 15 Canadian cigarette brands measured using the FTC machine-smoking protocol with the exception that cigarettes were smoked to a butt length of 30 mm, ranged as follows: tar, from 15.8 to 29.3 mg per cigarette in non-ventilated brands and from 24.2 mg to 36.0 mg in ventilated brands; nicotine, from 2.7 mg to 4.6 mg in non-ventilated brands and 3.0 mg to 6.1 mg in ventilated brands, and CO, from 40.5 mg to 67.3 mg in non-ventilated brands and 46.5 mg to 63.1 mg in ventilated brands. Yields in sidestream smoke were much higher than those in mainstream smoke for all brands tested. The average ratios for sidestream smoke to mainstream smoke were 3.5, 6.6 and 6.8 for tar, nicotine and CO, respectively. The highest yields from sidestream smoke were obtained from the brands with the lowest mainstream smoke yields (Rickert et al., 1984). The concentrations of carbon monoxide in the sidestream smoke in the Canadian cigarettes were higher than those reported for four different types of American blend cigarettes smoked according to the FTC protocol (Adams et al., 1987). Differences in the tobacco blend may be one explanation for this discrepancy and Canadian cigarettes are made predominantly from flue-cured tobacco.

The average yields of total particulate matter and nicotine in sidestream smoke generated by the machine-smoking of two cigarette brands that are popular among smokers in India (one filter-tipped and one untipped) were 16.51 and 0.9 mg per cigarette, respectively (Pakhale & Maru, 1998). In the sidestream smoke from bidis, these concentrations were 5.5 mg total particulate matter and 0.25 mg nicotine. The sidestream smoke released from chuttas contained 19.8 mg total particulate matter and 2.07 mg nicotine per unit. In all Indian products, the emissions of total particulate matter were much higher in mainstream smoke than in sidestream smoke, which is demonstrated clearly by the ratio of sidestream smoke to mainstream smoke which ranged from 0.13 to 0.49 (see also the monograph on tobacco smoke, Section 1.2.7). This is a modified, more intensive smoking standard (two puffs per minute instead of one) used because of the poor burning properties of the tobacco in Indian products.

In the 1999 Massachusetts Benchmark Study (Borgerding et al., 2000), a subset of 12 brands was analysed for the chemical composition of the sidestream smoke that was generated by machine-smoking using the ‘more intense’ Massachusetts method. The data obtained are summarized in Table 1.3. The concentrations of the constituents of the sidestream smoke determined in this study differed significantly from those obtained using the standard FTC method that had been reported by Adams et al. (1987): the yields of CO, ammonia and benzo[a]pyrene were higher, those of tar and catechol were of the same order of magnitude and those of NNN and NNK were significantly lower. NNN and NNK levels were even lower than those measured in the mainstream smoke generated by the same intense machine-smoking method. The values obtained by the Massachusetts Study for some gaseous compounds such as 1,3-butadiene, acrolein, isoprene, benzene and toluene were also far below those obtained by machine-smoking using the FTC method (Brunnemann et al., 1990; Table 1.4). For the 12 commercial cigarette brands tested by the Massachusetts puffing parameters, the highest median sidestream smoke/mainstream smoke ratios in the Massachusetts study were obtained for ammonia (ratio, 147), 3-aminobiphenyl (ratio, 10.8), formaldehyde (ratio, 14.8), pyridine (ratio, 16.1) and quinoline (ratio, 12.1).

Table 1.3

Average values of 44 smoke constituents in the sidestream smoke of 12 commercial cigarette brands assayed in the 1999 Massachusetts Benchmark Study using Massachusetts smoking parameters.

Table 1.4

Concentrations of selected gas-phase compounds in sidestream smoke of commercial cigarettes.

Often, conflicting results concerning the phase distribution of individual constituents and poor agreement between laboratories for quantitation of sidestream emissions are attributed to different methods used for smoke generation and collection.

Table 1.5 shows the yields of IARC carcinogens in sidestream smoke generated under standard International Organization for Standardization (ISO) and the more intense Health Canada methods, of four popular regular-size Canadian cigarette brands.

Table 1.5

Yields of IARC carcinogens in regular-sized Canadian cigarettes. Comparison of International Organization for Standardization (ISO) and Health Canada (HC) machine-smoking parameters.

On the basis of their mainstream smoke tar yields as measured by the ISO/FTC machine-smoking method, the four cigarette brands may be classified as ‘full flavour’, ‘light’, ‘extra light’ and ‘ultra light’. British Columbia has established the Tobacco Testing and Disclosure Regulation and became the first jurisdiction in the world to require Canadian tobacco manufacturers to disclose on a brand-by-brand basis the contents of cigarettes and tobacco and the levels of potentially toxic chemicals in tobacco smoke.

Among the 44 smoke components reported by the manufacturers on a yearly basis, there are seven IARC group 1 carcinogens (benzene, cadmium, 2-aminonaphthalene, nickel, chromium, arsenic and 4-aminobiphenyl), three IARC group 2A carcinogens (formaldehyde, 1,3-butadiene and benzo[a]pyrene) and eight IARC group 2B carcinogens (acetaldehyde, isoprene, catechol, acrylonitrile, styrene, NNK, NNN and lead).

Of the seven IARC group 1 carcinogens, arsenic yields in sidestream smoke were below the detection limits of both the ISO and Health Canada smoking methods. In general, yields of the six other IARC group 1 carcinogens in sidestream smoke were higher when measured by the ISO than by the Health Canada smoking method. The ISO and Health Canada methods gave similar yields for nickel and chromium in the ‘light’ cigarette and for chromium in the ‘ultra light’ cigarette.

For most IARC group 2A and 2B carcinogens, the yields in sidestream smoke measured by the Health Canada method were 40–80% of corresponding yields measured by the ISO method. Exceptionally, for NNK and lead the yields measured by the Health Canada method were higher than the yields measured by the ISO method, but only for the ‘extra light’ and ‘ultra light’ brands. The yields of NNK measured by the Health Canada method were up to 2.6-fold higher than the yields measured by the ISO method (Government of British Columbia, 2003).

Table 1.5 also allows comparisons of sidestream smoke yields between the brands. There is no significant difference between the total sidestream smoke yields of IARC group 1 carcinogens of ‘full flavour’ and ‘light’, extra light and ultra light cigarettes when measured by either the ISO or Health Canada methods.

The data in Tables 1.3 and 1.5 suggest that during more intense smoking (as employed by the Massachusetts and Health Canada methods: i.e. larger puffs, shorter interval between puffs and partial or complete blockage of ventilation holes), smaller quantities of tobacco are burned during the smouldering of the cigarette, thus affecting the emissions of toxins in sidestream smoke. Therefore, the real-life contribution of sidestream smoke to the overall concentrations of selected components of secondhand tobacco smoke may have been overestimated in the past because most smokers draw smoke from their cigarettes with an intensity more similar to that of the Massachusetts or Health Canada methods than the FTC machine-smoking method (Djordjevic et al., 2000). This concept needs to be investigated more thoroughly, especially in view of the finding that an increase in puff volume from 17.5 mL to 50 mL and in filter ventilation from 0 to 83% failed to reduce the levels of tar, CO and nicotine in the sidestream smoke, whereas the yields in mainstream smoke and subsequently the ratios of sidestream smoke to mainstream smoke changed significantly (Guerin et al., 1987).

In addition to the constituents listed in Table 1.3, some further constituents have been quantified in sidestream smoke since the publication of the 1986 IARC Monograph. These are NDMA (up to 735 ng per cigarette), N-nitrosopiperidine (NPIP, 19.8 ng) and NPYR (up to 234 ng) (Adams et al., 1987); and volatile hydrocarbons, e.g. ethene up to 1200 µg, propene up to 1300 µg, butenes up to 900 µg and pentenes up to 2100 µg. The sidestream smoke emissions of various unsaturated gaseous hydrocarbons were 3–30 times those reported for the mainstream smoke emissions. These compounds constitute a potential health risk as they are metabolized in vivo to reactive genotoxic epoxides (Löfroth et al., 1987; Löfroth, 1989). High-molecular-weight n-alkanes (C₂₇ [66–86.5 µg per cigarette], C₂₉ [28–39 µg per cigarette] C₃₁ [148–197 µg per cigarette], C₃₃ [43.5–62 µg per cigarette]) were also quantified in the sidestream smoke of commercial cigarettes (Ramsey et al., 1990).

The co-mutagenic beta-carbolines, norharman and harman, were quantified in the sidestream smoke condensates of some Japanese cigarette brands. The concentrations per cigarette were 4.1–9.0 µg for norharman and 2.1–3.0 µg for harman (Totsuka et al., 1999).

1.2.1. Measurements of nicotine and particulate matter in indoor air

The report of the US Environmental Protection Agency (US EPA, 1992) summarizes over 25 separate studies that reported concentrations of nicotine in air measured in more than 100 different indoor microenvironments. Hammond (1999) also reported an extensive survey of the concentrations of nicotine in air. Based on the large numbers of measurements made in various indoor environments in the USA between 1957 and 1991, the average concentrations of nicotine in air showed about 100-fold variation, i.e. from 0.3–30 µg/m³ (US EPA, 1992). The average concentrations in homes with one or more smokers typically ranged from 2 to 10 µg/m³, with the highest averages being up to 14 µg/m³. Data from the mid 1970s until 1991 indicate that the nicotine concentrations in offices were similar to those measured in homes, with a large overlap in the range of air concentrations for the two types of environment. The maximum levels of nicotine, however, were considerably higher in offices than in domestic environments (US EPA, 1992; California EPA, 1997). In studies using controlled and field conditions, the concentrations of nicotine in air were found to increase as a function of the number of smokers present and the number of cigarettes consumed (US EPA, 1992).

Jenkins et al. (1996) studied exposure to secondhand tobacco smoke in 16 cities in the USA by sampling personal breathing zone air from about 100 nonsmokers in each city. The demographics of the study subjects were comparable with the population of the USA in general, although more women than men participated in the study. The mean 24-h time-weighted average concentration of nicotine was 3.27 µg/m³ for those exposed to secondhand tobacco smoke both at work and away from work, 1.41 µg/m³ for those only exposed away from work and 0.69 µg/m³ for those who were exposed only at work. The mean 24-h time-weighted average concentration of nicotine in air measured by personal monitoring, for those who were not exposed to secondhand tobacco smoke was 0.05 µg/m³ (Jenkins et al., 1996).

Personal exposure to particulate matter associated with secondhand tobacco smoke was determined using the set of specific markers such as respirable suspended particles, fluorescent particulate matter and solanesol-particulate matter. The ranges of mean concentrations of these particles for workers exposed to secondhand smoke in 11 countries were: respirable suspended particles, from 24 to 112 µg/m³; fluorescent particulate matter, from 5.7 to 57 µg/m³; and solanesol-particulate matter, from 3.6 to 64 µg/m³ (Jenkins et al., 2000). By measuring the levels of solanesol-particulate matter and nicotine, the exposure to secondhand tobacco smoke of office workers living and working with smokers was determined to be higher in winter than in summer (median 24-h time-weighted average concentrations, 25 µg versus 2.4 µg solanesol-particulate matter and 1.3 µg versus 0.26 µg nicotine, respectively) (Phillips & Bentley, 2001).

1.2.2. Population-based measurements of exposure

Most population-based estimates of exposure to secondhand tobacco smoke have been obtained from self-reports. When measuring exposure to secondhand smoke in indoor areas, nicotine or respirable suspended particles can be measured in air sampled using personal monitors. In a few studies, biomarkers such as cotinine have been measured in physiological fluids.

(a) Adults

Some studies suggest that exposure to secondhand tobacco smoke is related to occupation and socioeconomic status, and that higher exposure is more common among adults employed in blue-collar jobs, service occupations and poorly paid jobs and among the less well educated (Gerlach et al., 1997; Curtin et al., 1998; Whitlock et al., 1998). Exposure to secondhand tobacco smoke may also be higher among racial and ethnic minority groups in areas of the USA, although it is unclear if this is due to different socioeconomic status (Gerlach et al., 1997).

Relatively few data are available on the prevalence of nonsmokers' exposure to secondhand tobacco smoke on a population basis, using biomarkers. Survey data from a study in the USA in 1988–91 showed that 37% of adult non-tobacco users lived in a home with at least one smoker or reported exposure to secondhand tobacco smoke at work, whereas serum cotinine levels indicated more widespread exposure to nicotine. Of all the non-tobacco users surveyed including children, 88% had detectable serum cotinine levels, indicating widespread exposure to secondhand tobacco smoke in residents of the USA (Pirkle et al., 1996). Data were recently published on the serum cotinine levels measured in 2263 nonsmokers in 12 locations across the USA (Figure 1.1) (CDC, 2001). As reported previously (Pirkle et al., 1996), exposure to secondhand tobacco smoke tended to be higher among men than among women. Among racial/ethnic groups, blacks had the highest cotinine levels. People younger than 20 years of age had higher cotinine levels than those aged 20 years and older.

Figure 1.1

75th percentile of serum cotinine concentrations for the US nonsmoking population aged 3 years and older, National Health and Nutrition Examination Survey, 1999

Table 1.7 summarizes the data obtained from a number of recent population-based studies that used questionnaires to characterize exposure. Some of these studies were national in scope, e.g. the national samples in Australia, China and the USA, whereas others were from single states or specific localities. Several of the studies incorporated cotinine as a biomarker. Unfortunately, few data are available from developing countries (Samet & Yang, 2001). In a case–control study of lung cancer and exposure to secondhand smoke in 12 centres from seven European countries, 1542 control subjects up to 74 years of age were interviewed between 1988 and 1994 about their exposure to secondhand smoke. Exposure of adults to secondhand smoke from a spouse who smoked was reported by 45% of the subjects (including 2% who were exposed to smoke from cigars or pipes only); an additional 8% were not exposed to spousal smoke, but reported exposure to secondhand smoke produced by cohabitants other than spouses. Exposure to secondhand smoke in the workplace was reported by 71% of men and 46% of women. Combined exposure to secondhand smoke both from the spouse and at the workplace was reported by 78%. Exposure to secondhand smoke in vehicles was reported by 20% and exposure in public indoor settings such as restaurants was reported by 29% (Boffetta et al., 1998). Recent data from Finland illustrate trends in self-reported exposure to secondhand tobacco smoke at work and at home over a 15-year period. In 1985, 25% of employed nonsmoking men and 15% of employed nonsmoking women were exposed to secondhand tobacco smoke at work. In contrast, the figures for exposure to secondhand smoke in 2000 were 8% of men and 4% of women (Jousilahti & Helakorpi, 2002).

Table 1.7

Prevalence of exposure to secondhand tobacco smoke — population-based studies.

In a 1993 review of existing studies, Siegel (1993) noted wide variation in the concentrations of secondhand tobacco smoke by location when measured by weighted mean levels of nicotine in the ambient air of offices (4.1 µg/m³), restaurants (6.5 µg/m³), bars (19.7 µg/m³) and dwellings with at least one smoker (4.3 µg/m³).

The recently published study in the USA of the median serum cotinine concentration measured in nonsmokers aged 3 years and older found a 75% decrease over the period 1991–99 (CDC, 2001), suggesting positive effects of policies for cleaner indoor air.

1.3.1. Policy options

There are a range of options available for the regulation of secondhand tobacco smoke. Of these options, the least effective is designating smoking areas that have no separate ventilation. This option provides only minimal protection to nonsmokers; studies have shown that substantial exposure to secondhand tobacco smoke occurs in workplaces where there are smoking areas without separate ventilation (Repace, 1994). A more effective option is the use of separately ventilated smoking lounges; this protects non-smokers but is costly and may elevate lung cancer risk in smokers (Siegel et al., 1995). Separately ventilated smoking lounges also endanger workers (e.g. waiting staff) who must enter these areas as part of their job. Finally, the most effective alternative is a totally smoke-free workplace (Brownson et al., 2002).

1.3.2. Prevalence of regulations

In 1985, only about 38% of workers in the USA were employed by firms that had policies restricting smoking (Farrelly et al., 1999). Since that time, smoking restrictions have become increasingly common. According to the 1999 National Worksite Health Promotion Survey, 79% of workplaces with 50 or more employees had formal smoking policies that prohibited smoking or allowed it only in separately ventilated areas (US Department of Health and Human Services, 2000). Data showed that, from 1995–96, 64% of indoor workers in the USA were covered by a total ban on smoking in the workplace. The proportion of workers in the USA who work in a smoke-free workplace varies considerably by state — from 84% in Maryland and Utah to 40% in Nevada (Burns et al., 2000). There are few systematic data available on the enforcement of existing policies to restrict smoking in the workplace, although existing studies suggest that compliance is likely to be high (Stillman et al., 1990; Wakefield et al., 1996). National data from the USA also suggest that despite some protective laws, workers in blue-collar and service occupations remain much more likely to be exposed to secondhand tobacco smoke in the workplace than other categories of workers (Gerlach et al., 1997). In the USA, hospitals are the only sector that has voluntarily implemented a nationwide smoking ban. This ban was announced in November 1991 and full implementation was required by December 31, 1993. Two years after implementation, the policy was found to have been successful with 96% of hospitals complying with the smoking ban standard (Longo et al., 1995; Brownson et al., 2002).

A recent overview of the legislation restricting smoking at work in different countries has been provided by the American Cancer Society (Corrao et al., 2000) (Table 1.8). Despite using a wide range of already published sources together with Internet searches, no information could be found for many countries. Thus, absence of an entry in this Table does not imply certainty that no legislation exists in a particular country. Conversely, the existence of a law banning or restricting smoking implies nothing about its enforcement. In general, it appears that voluntary restrictions under control of the employer are more common in developing countries than in developed countries (Brownson et al., 2002). A country that relies on voluntary action to ban or restrict smoking may have quite high rates of worker protection. For example, although Australia banned smoking in all federal government workplaces in 1988, it has been left to individual employers to determine their own policies. Yet, in 1999, 71% of indoor workers in the state of Victoria reported a total ban on smoking at their workplace, 21% reported some restrictions on smoking and 8% reported unrestricted smoking (Letcher & Borland, 2000). As in other countries, employees in small Australian workplaces are less likely to report protection, as are workers in particular types of employment (Wakefield et al., 1996; McMaugh & Rissell, 2000). In a survey of indoor workers, 38% of those employed in a restaurant or hotel, 15% of warehouse/store workers and 17% of those working in a workshop or factory reported unrestricted smoking where they worked, compared with only 3% of workers in open-plan offices (Letcher & Borland, 2000). In the United Kingdom, workplace restrictions are also voluntary. In 1997, 40% of the workforce was estimated to be working in a totally smoke-free environment (Freeth, 1998). A survey of 1500 workplaces in Scotland found that 79% of them had designated nonsmoking areas, but only 22% had banned smoking completely (ASH, 2001). Despite the limitations of the data presented in Table 1.9, it is apparent that most countries have some laws that restrict smoking. However, it is very likely that there is considerable need for improvement in protection of workers from secondhand tobacco smoke in almost all countries (Brownson et al., 2002).

Table 1.8

Variations in workplace smoking policies in selected countries.

A few researchers have begun to examine the prevalence of smoking restrictions in the home because such restrictions are likely to have beneficial effects on the health of children. In 1997, a population-based, cross-sectional telephone survey was conducted using random-digit-dialling asking 6199 adult Oregonians to provide baseline data on tobacco use in Oregon. Seventy per cent of the households were composed of nonsmokers only, and 85% of those had a full ban on smoking inside the home. Of the households containing one or more smokers, 38% had a full household ban on smoking. Fifty per cent of households with at least a smoker and a child present did not have a full ban on indoor smoking (Pizacani et al., 2003). Face-to-face interviews were conducted with 380 rural, low-income Native American and white parents of children aged 1–6 years in Oklahoma. The prevalence of complete smoking bans was 49% in Native American homes and 43% in white homes. Bans on smoking in cars were less common, with 35% of Native American and 40% of white caregivers reporting complete bans (Kegler & Malcoe, 2002). In Victoria, Australia, the percentage of respondents who reported discouraging visitors from smoking in the home rose from 27% in 1989 to 53% in 1997 (Borland et al., 1999), and not smoking in the presence of children rose from 14% in 1989 to 33% in 1996. Similarly, attitudes toward smoking in the home have changed in Ontario, Canada. The percentage of respondents favouring not smoking in homes where there were children increased from 51% in 1992 to 70% in 1996 (Ashley et al., 1998).

Only minimal regulation applies to constituents of cigarettes and tobacco smoke. This covers only the content of tar, nicotine and carbon monoxide (cf. Section 1.4(e) of the monograph on tobacco smoke).

1.3.3. Effectiveness of regulations

Evaluations of the effects and effectiveness of workplace smoking policies have used a wide variety of study designs and measurements of exposure to secondhand tobacco smoke and tobacco use behaviours. Most of the published studies are simple assessments conducted before and after adoption of a workplace policy, although more recent (and more complex) investigations have employed cross-sectional surveys of workers in workplaces operating different policies. Few studies have evaluated or controlled for potential bias and confounding of the observed differences or changes in exposure to secondhand smoke or in tobacco use behaviours (Brownson et al., 2002).

The effectiveness of workplace smoking policies has been measured by differences or changes in perceived air quality in the workplace following a ban or restriction, and by differences or changes in active measurements of nicotine vapour concentrations, metabolites, or levels of particles. Overall, workplace smoking policies have been highly effective in reducing the exposure of nonsmokers to secondhand tobacco smoke. The ‘best evidence subset’ comprised ten studies, including cross-sectional surveys, before-and-after comparisons, different settings or locations (offices, public sector workplaces, medical centres, workplaces community-wide) and different outcome measurements (Table 1.9) (Briss et al., 2000; Hopkins et al., 2001). In nine of ten studies, workplace smoking policies had a significant impact on exposure to secondhand tobacco smoke. In assessments conducted between 4 and 18 months after implementation of the policy, the median relative percentage difference in self-reported exposure to secondhand tobacco smoke was –60%, range +4% to −97%. Workplaces with smoking bans tended to show greater reduction in exposure to secondhand tobacco smoke than did workplaces with smoking restrictions (Hopkins et al., 2001; Brownson et al., 2002).

Hammond (1999) summarized the existing literature on average indoor nicotine concentrations when various workplace smoking policies were enacted. In workplaces with policies that had banned smoking, nicotine concentrations were generally decreased to less than 1 µg/m³. The mean concentrations of nicotine in workplaces that allowed smoking ranged from 2 to 6 µg/m³ in offices, 3 to 8 µg/m³ in restaurants and from 1 to 6 µg/m³ in the workplaces of blue-collar workers. By comparison, studies of nicotine concentrations that included at least 10 homes of smokers and that were sampled for 14 h to 1 week found average nicotine concentrations of between 1 and 6 µg/m³.