Tobacco Documents Online

I I I I I I I I I I I I I 1 I I I I DETERMINATION OF PERSONAL EXPOSURE OF NONSMOKERS TO ENVIRONMENTAL TOBACCO SMOKE IN THE UNITED STATES Roger A. Jenkins, M.A. Palausky, R.W. Counts, M.R. Guerin, A.B. Dindal and C.K. Bayne Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA Introduction In the United States, there is considerable controversy regarding the potential health effects related to exposure to environmental tobacco smoke (ETS). Authors of some epidemiological studies have concluded that there is a small but statistically significant increase in risk of contracting lung cancer to lifetime never-smoking women married to smoking spouses. Others authors have found no such statistically significant increases in relative risk (US EPA, 1992). To date, many studies attempting to quantify ETS exposure in the US population have had to rely on self-reports of exposure (Jenkins, et al., 1992), or extrapolations from determinations of area measurements of ETS levels in locations where cigarettes are actively being smoked (Oldaker, eta l., 1990; Leaderer and Hammond, 1991; Jenkins, et al., 1991; Colett, et al., 1992). Clearly, such is not the same as a direct determination of the exposure of nonsmokers to ETS. However, such direct determinations have been limited to relatively small study populations. The purpose of the study reported here is to directly determine ETS exposures of more than 1000 US nonsmokers. Experimental Study Design The study design consisted of recruiting approximately 100 individual subjects in each of 16 cities distributed geographically around the United States. To determine exposure, each individual wore a sampling pump during the work phase of his/her day, and another pump to collect samples from which to determine ETS exposure away from work. The sampling systems collected both particulate phase and gaseous phase components of ETS. While attempting to create an equally populated 2x2 matrix of subjects living in smoking or nonsmoking homes and working in smoking or nonsmoking workplaces of equal cell population, the difficulties of recruiting individuals living and working in smoking environments were such that the cells were unequally populated. Although all subjects were recruited on the basis of their nonsmoking status, salivary cotinine was used to assess actual smoking status, as well as the relationship between directly measured ETS exposure and salivary cotinine. Subject Recruiting and Itinerary Nearly all of the subjects were recruited through random telephone dialing or marketing research databases. Less than 10% of the subjects were recruited through mall intercept methods. During the initial contact, the subject was required to pass a screening questionnaire (administered by phone by a local marketing research firm). To be included in the study, individuals had to report themselves as not having used tobacco products in the last six months, nor using any form of nicotine patch or gum, be at I

least 18 years of age, and working outside the home on a"regular" (ca. 8 am until 5 pm) shift at a minimum of 35 hours per week. Individuals were excluded from the study if they had "inappropriate" professions for such a study or membership in an advocacy group related to the objectives of the study (eg., no tobacco company workers or members of anti-smoking groups). Following the screening, the subject is assigned to a study cell. On the evening of Day 1 of the subject's involvement, the subject arrives at the test coordination site, and is rescreened to verify the accuracy of the telephone questionnaire. The subject then watches an instructional video with approximately 24 other participants and completes a "first visit" questionnaire concerning his/her lifestyle and details regarding the type of environment in which the subject works. The subject provides a saliva sample and receives his/her sampling systems, after being tested to insure that the subject can actually operate the sampling unit. On the morning of Day 2, the subject begins sampling with the workplace pump upon his/her arrival at work. The sampling apparatus consists of a sound-insulated pump (typically worn over the right shoulder and on the left hip) and a sampling head, containing both particulate and vapor collection devices which is worn in the subject's breathing zone. The subject also completes a workplace diary, recording various smells and observations concerning the use of products which may affect indoor air quality (eg., copying machines, correction fluids, coffee, cigarettes, etc.). Subjects are requested to remain at their work station during the lunch period. At the end of the workday, the subject turns off the workplace sampling pump, completes the workplace pump survey, dons the away-from-work pump (which is outfitted with a larger battery pack to afford sampling for a minimum of 18 hours), and returns home, conducting normal activities, such as shopping, dining, etc, on the way. The subject completes an away-from-work diary on an hourly basis. At bedtime, the subject takes off the pump, and sets it alongside of his/her bed, while the pump continues to sample. The next morning (Day 3), when the subject arrives at work, the away-from-work pump is turned off, and the home pump survey completed. After work that same day, the subject returns to the test coordination center with all of the "take home" materials, completes a last visit survey, provides a second saliva sample, and receives a $100 gratuity. Determination of Exposure Markers Particulate phase ETS air markers were collected on a Fluoropore membrane filter at a flow of approximately 1.7 L/min, while vapor phase markers were collected on XAD-4 resin cartridges (SKC Inc., Eighty Four, PA) at a flow of approximately 0.5 - 0.7 L/min, using a single air sampling pump. Particulate phase markers included respirable suspended particulate matter (RSP, 3.5 µm cut-off), solanesol, ultraviolet absorbing particulate matter (WPM), and fluorescing particulate matter (FPM). ETS gaseous phase markers included nicotine, 3-ethenylpyridine (3-EP), and myosmine. Briefly, RSP was determined gravimetrically (Conner, et al., 1990), and UVPM and FPM were determined by high performance liquid chromatography (HPLC) with UV and fluorescence detectors(Conner, et al., 1990), respectively. Solanesol was also determined using HPLC (Ogden and Maiolo, 1992). All of the vapor phase markers were determined using gas chromatography with thermionic specific (nitrogen selective) detection (Ogden, 1991). Levels of salivary cotinine were determined using radio-immunoassay (Davis and Stiles, 1993) -2- I I I I I I I I I I I I 1 I I I I

Results and Discussion I I 1 i In Table 1 are presented 12 of the 16 urban areas around the United States for which data has been analyzed to date, and for which the results are reported in this manuscript. Samples were collected in these 12 cities from mid-May, 1993 to mid-December, 1993. Cities were chosen based on obtaining a good geographic distribution, weather during the time of year, logistics, lack of pervasive smoking restrictions, and likelihood of high quality field marketing survey research support. Table 1. Urban Areas Selected For Investigation and for which Results are Reported Knoxville, TN Boise, ID Columbus, OH Portland, ME Seattle, WA Buffalo, NY San Antonio, TX Baltimore, MD St. Louis, MO Fresno, CA Daytona Beach, FL Grand Rapids, MI In Table 2 are presented the initial cell assignment populations, based on the initial screening questionnaire results. The relative proportions of participants are indicative of the difficulty of recruiting individuals who live and work in situations where unrestricted smoking occurs. In many cities, thousands of telephone calls were required to locate relatively modest numbers of participants in Cells 1- 3. Table 2. Cell Populations For Cities 1-12 Based on Screening Questionnaire Assignment I I I I I I I I I . -bFumber of Subjects . ... . % of SuLject -po]iltlnnon Cell 1: Smoking Home/Smoking Workplace 121 10.4 Cell 2: Smoking Home/nonsmoking workplace 172 14.8 Cell 3: Nonsmoking home/Smoking workplace 222 19.0 Cell 4: Nonsmoking work and Nonsmoking Workplace 651 55.8 Total 1166 100 When compared with the US population as a whole (corrected for those greater than 18 years of age), extracted from the 1993 Statistical Abstract of the United States (U.S. Department of Commerce, 1993), our study population tended to be younger, have more years of formal education, have a higher median household income, and be comprised of a higher percentage of females. For example, about 68% of the subjects in Cities 1- 12 were female. This may be due to two general observations from marketing survey research: a) women are more likely to answer the telephone in a household; and b) a higher proportion of women are more likely to participate in such a study. When adjusted for the under-18 year old US population, the study is comprised of a slightly larger proportion of younger individuals.. This -3-

I may be due to the requirement that participants work at least 35 hours a week on a regular (ca. 8am - 5 pm) shift, which would tend to exclude individuals who are retired from the study. In general, the participants in the study had higher household incomes. For example, the annual US median household income is approximately $30,000, in contrast to approximately $40,000 for subjects in this study. This may be due to several factors. First, since subject recruiting is conducted by telephone, the selection method excludes those individuals who do not have telephones. Inclusion in the study required the participants to work at least 35 hours a week on a regular shift. A larger fraction of lower household income individuals may not work a full 40 hour work week. Also, higher income households often have two adult workers in the family, and since subjects living in smoking homes were required to be nonsmokers and live with a smoker, the smoking homes may have been selected from the higher income groups. With regards to the occupational distribution, the study contains a lower proportion of individuals in service occupations and those who work in factories. These individuals may have decided not to participate on the basis of safety or appearance concerns for wearing the air sampling pumps. As a result, the study contains a larger proportion of "white collar" workers, who may tend to be more highly paid. In Table 3 is presented a summary of the median exposures of individual participants segregated by those working in smoking and nonsmoking locations, as well as away-from-work settings which include either smoking or nonsmoking homes. (Except where otherwise noted, all of the smoke exposure data has been corrected for those individuals who can be clearly considered smokers: those with salivary cotinine values > 100 ng/mL have been excluded from these tabulations. However, some regular smokers are likely to have salivary cotinine levels which are an order of magnitude lower.) Note that the measured parameter in this table is actual exposure, defined as the average smoke marker concentration in µg/m3, multiplied by the time of exposure, and the estimated breathing rate, in L/min. (The rate of 20 ILmin was taken from the National Research Council report on Environmental Tobacco Smoke [1986].) First, it should be noted that only those individuals who reported consistent exposures (ie. reported observing - or not observing - tobacco products being smoked in their diaries, pump surveys, and last visit surveys) were included in this particular compilation. The justification for this is that many individuals work in locations where they report smoking is permitted, but where no actual tobacco products were observed to have been smoked. Thus, the assignment of such a facility as a "smoking" workplace, when the participant did not observe smoking taking place, seems incongruous, and clouds the interpretation of the data. The same argument can be used for assignment to a cell including a smoking home environment. From the data in Table 3, which are median values, it is clear that those individuals who live and work with smokers are exposed to substantially more ETS components than those who observe no cigarettes, pipes, or cigars being smoked around them. For example, median nicotine exposures for participants in Cell 1(smoking workplaces and an away-from-work categorization which included a smoking home) were more than 50 times greater than those who live, work, shop, and commute in a truly nonsmoking environment (Cell 4 subjects). That exposures to any discernable amount of ETS components occur in environments where no smoking is involved may be indicative of the ubiquitous nature of ETS at trace levels. A comparison of exposures of participants in Cells 2 and 3 provide an indicator of the greatest contributor to ETS exposure. Cell 2 is populated with participants that reported cigarettes being smoked in their presence outside of work, and reported no cigarettes being smoked within their sight or smell in their workplace. In contrast, the participants assigned to Cell 3 confirmed smoking occurring in their workplace, but not observing any cigarettes being smoked outside of work. The comparison indicates that Cell 2 participants are exposed (concentration times duration times breathing rate) to more than four times the amount of ETS componentsas Cell 3 participants, indicating that those locations outside the workplace are a much greater contributor to true ETS exposure. -4- N O O ~ v 00 W ~ c0 i I I I I I I I I I I I I I

a I I I I I I I In Table 4 are presented similar data, but in the more conventional terms of a time averaged concentration of ETS components to which the participant is exposed. That is, the time averaged concentrations are equal to the sum of the concentration/time products for the workplace and away-from- work sampling systems, divided by the total time of measurement of the two sampling systems (ca. 24 hours). The conclusions from the data are the same as those for the data presented in Table 3, however. That is, individuals that work, live, and operate around smokers receive a substantially greater exposure to ETS, and the away-from-work venue appears to be the primary contributor. It is important to note that in general, the levels of ETS to which individuals are actually exposed are substantially lower than those which may be inferred from previous studies of ETS marker levels measured in specific areas over short durations. For example, in Guerin, et al., 1992, several studies of nicotine levels in offices are reported. For the most part, studies reported mean levels of 4 - 14 µg/m3. This compares with mean levels in this study of workplaces in which smoking is unrestricted of ca. 2.7 µg/m3, and median levels of 0.44µg/m3. In Figure 1 are com- pared the distributions of one- hour nicotine levels measured in a study of offices in which smoking was unrestricted in five cities (Oldaker, et at., 1990) with the 8-hour time weighted average levels of nicotine to which office workers were exposed in the 12 cities of our study. The mean one-hour nicotine level was 7.07 ± 8.42 µg/m3, compared with 3.15 ± 5.66 µg/m3 for the actual 8-hour exposure levels. Median values were 4.65 µg/m3 and 0.91 µg/m3, respectively. It is clear from the comparison in Figure I that the distribution of values is dramatically different at the lower concentration levels. There may be at least two explanations for this 60 ~~p, . 3 ~~.1-".I'~5'9 r-1~- 4 6 8 10 12 14 16 iB 20 >21 Nicotine, micrograms per cubic meter Smoking Office Levels ® Exposures in Offices Figure 1. Comparison of distributions of nicotine levels in smoking offices (1 hour measurements) and 8-hour exposure levels of non-smoking workers in smoking offices. difference. First, there has been a general trend in the US society for smokers not to consume their smoking materials in the presence of nonsmokers as much as in previous decades. A second obvious explanation is that nonsmokers, even in environments in which smoking is not restricted, spend only small amounts of time is areas where ETS levels are high. Individuals may believe that they are being exposed to the smoke of many cigarettes because they observe such around them. But in fact, due to either distance from the smoker or the relatively short time in the presence of significant quantities of smoke, the exposures received from the smoke are relatively small in most cases. This data also suggests that short term area measurements may actually overestimate worker exposure in many situations. -5-

Table 3. Comparison of Exposure of Individuals to Environmental Tobacco Smoke Markers Among Cells Away From Work Work Number of - 24-hr Time Averaged Exposure, Kg' . Cell Environment Environment Participants . 3-EP Nicotine Myosmine RSP UVPM FPM Solanesol 1 Smoking a Smoking c 119 Median 20.1 38.4 4.34 881 338 221 3.09 Mean 33.8 88.4 7.77 1211 662 513 13.6 95th %ile 113 252 18.4 3126 1926 1972 49.6 2 Smoking a Nonsmoking d 109 Median 10.3 15.2 L79 675 223 164 1.67 Mean 19.7 36.9 3.41 935 413 320 8.16 95th %ile 65.0 133 11.2 2308 1288 1070 27.9 3 Nonsmoking b Smoking C 163 Median 2.18 2.97 0.400 558 61.5 36.4 0.090 Mean 6.29 13.1 1.47 734 146 101 1.88 95th %ile 26.1 48.3 5.88 1664 529 400 9.23 4 Nonsmoking b Nonsmoking d 497 Median 0.593 0.671 0.024 412 29.9 15.0 0.0412 Mean E 7 E [ 9.09 0.301 492 49.6 31.1 0.134 95th %ile 612 6.90 0.926 1153 153 114 0.494 Noted observations of tobacco products on Home Diary, Home Pump survey, and Last Visit survey for home. Noted no observations of tobacco products on Home Diary, Home Pump survey, and Last Visit survey for home. Noted observations of tobacco products on Work Diary, Work Pump survey, and Last Visit survey for work. Noted no observations of tobacco products on Work Diary, Work Pump survey, and Last Visit survey for work. 'Analytical blank-corrected µg/sample/(Sampling time x Flow rate) = µg/m' per Sample; [(µg/m', Away from work sample x Hours, Away from work sample) + (µg/m', Work sample x Hours, Work sample)]/(I-Iours, Away from work sample + Hours, Work sample) = Time Averaged µg/m'; µg/m' x 1.2 m'/hour (National Academy of Sciences Inhalation Rate) x Hours = µg 'Acmal value was nondetectable; one half of the limit of detection, in µg, and average flow rate, and an 24-hour time were used. -6- £64£8LW0Z M a.. .. ar r..1"Im r-.r r r• .rr rIM .. iar .r r

WON r -M ms mi MM M M M M M ~li r r 00 Table 4. Comparison of Concentrations of Environmental Tobacco Smoke Markers to which Individuals have been Exposed Among Cells Away From Work Work Number of 24-hr Time Averaged Airborne Concentration, µg/m3 ' Cell Environment Environment Participants 3-EP Nicotine Myosmine RSP UVPM FPM Solanesol 1 Smoking a Smoking a 119 Median 0.800 1.46 0.161 32.0 11.9 7.67 0.113 Mean 1.20 3.10 0.275 43.3 23.6 18.3 0.483 95th %ile 3.92 8.81 0.642 116 67.5 57.0 1.77 2 Smoking a Nonsmoking b 109 Median 0.369 0.555 0.068 24.5 8.01 5.89 0.058 Mean 0.719 1.34 0.124 34.5 15.1 11.6 0.295 95th `~ile 2.45 4.79 0.4D1 84.3 46.2 37.5 0.982 3 Nonsmoking b Smoking a 163 Median 0.078 0.114 0.014 20.5 2.26 1.20 0.003 Mean 0.231 0.480 0.053 26.7 5.39 3.78 0.070 95th %ile 0.979 1.80 0.185 60.8 18.7 16.7 0.383 4 Nonsmoking b Nonsmoking b 497 Median 0.022 0.024 0.001 14.9 1.08 0.567 0.0032 Mean 0.059 0.350 0.011 18.1 1.83 1.15 0.005 95th %ile 0.221 0.251 0.035 41.5 5.49 3.97 D.018 a b Noted observations of tobacco products on Home Diary, Home Pump survey, and Last Visit survey for home. Noted no observations of tobacco products on Home Diary, Home Pump survey, and Last Visit survey for home. 'Analytical blank-corrected µg/sample!(Sampling time x Flow rate) = µg/m; per Sample; [(µg/m', Away from work sample x Hours, Away from work sample) + (µg/m', Work sample x Hours, Work sample)]/(Hours, Away from work sample + Hours, Work sample) = Time Averaged µg/m' 2Actual value was nondetectable; one half of the limit of detection, in µg, and average flow rate, and an 24-hour time were used. -7- b66£81480Z

Table 5. Relationship of Household Income and Airborne Concentration of Environmental Tobacco Smoke Markers Household Income Number of Median 24-hr Time Averaged Airborne Concentration, gg/m' ' (in Thousands) Participants 3-EP Nicotine Myosmine RSP. UVPM FPM Solanesol Less than 10 30 0.130 0.146 0.027 27.4 2.60 1.44 0.014 10 - 20 137 0.110 0.125 0.020 19.7 2.15 1.33 0.0032 20 - 30 183 0.068 0.078 0.014 17.3 1.68 0.964 0.0032 30 - 40 226 0.074 0.095 0.014 18.2 1.99 1.16 0.0032 40 - 50 195 0.042 0.047 0.010 17.3 1.63 0.978 0.003z 50 - 75 249 0.035 0.036 0.005 16.9 1.47 0.803 0.0032 75 - 100 66 0.024 0.020 0.002 13.9 1.27 0.659 0.0032 Greater than 100 44 0.019 0.027 0.0062 15.0 1.23 0.631 0.0032 'Analytical blank-corrected µg/sample/(Sampling time x Flow rate) = µg/m3 per Sample; [(µg/m3, Away from work sample x Hours, Away from work sample) +(µg/m3, Work sample x Hours, Work sample)]/(Hours, Away from work sample + Hours, Work sample) = Time Averaged µg/m' 2Actual value was nondetectable; one half of the limit of detection, in µg, and average flow rate, and a 24-hour time were used. 564£SL 6SOZ -8- M" r M Mm"'M ~ r MM W M M Mae M r

I I I I I I I I There are some practical reasons why this and virtually any study which requires voluntary participation of the individual subjects can not be exactly representative of the population of the United States as a whole. One potential criticism of the study design is that it tends to exclude those individuals in lower socio-economic groups. However, the extent to which this may or may not affect the conclusions of the study is difficult to judge. One approach to determining the extent to which socio-economic status impacts study conclusions is to determine the extent to which smoke exposure is correlated with household income. In Table 5 is summarized the median 24-hour time averaged smoke marker levels as a function of self-reported household income. (Note that because the smoke marker levels are 24 hour averages, they are directly proportional to actual ETS exposure over the measured time period.) With the exception of solanesol, for which in many cases the concentrations were at or near detection limits, there is a definite inverse proportionality between household income and smoke exposure. The lowest income groups may receive two to seven times the exposure to ETS marker components as the higher income groups in the study. For example, the median time weighted average (TWA) level of 3-ethenylpyridine (3-EP) was 0.130 µg/m3 for those with household incomes <$10K, compared with 0.019 µg/m3 for those with incomes greater than $100K. Based on the ETS marker data from the first 12 cities, it appears that there may be an important difference between the perception of individuals' exposures to ETS (as judged by self-reported impressions of the number of cigarettes smoked around them) and the reality of that exposure (as judged by the actual levels of smoke constituents collected by the sampling pumps at and away-from work). In Figure 2 is summarized the location of all of the tobacco products to which individuals reported exposure. These data are taken from the last visit survey, which the individuals complete upon their return to the test coordination center in their city. (Note that this is essentially a recollection of the diary data, a count of all of the smoking products, regardless of a subject's classification, and not the data taken directly from the diary.) Greater than 50% of all of the tobacco products observations occur at work. About 25% of the observations are at home, and the remaining 25% are distributed among other locations. The away- from-work cigarettes (in this case, from Cities 1- 12) observed are presented according to a more detailed categorization in Figure 3. Nearly 60% of all of the cigarette "exposures" (actually observations of cigarettes being smoked around the subjects) during the Figure 2. Self-reported Number of Tobacco Products to which Subject was Exposed from Last Visit Survey. away-from-work sampling occur inside the private residence. 1.1 ~  ~ V co 9 '~ co ' W I

From the data presented in Figure 2, one might infer incorrectly that the greatest exposure to ETS occurs in the workplace. However, such an inference is contrary to the data presented in Table 6, which is a comparison of actual exposures (ie., concentration multiplied by duration of exposure and average breathing rate) in micrograms, µg, for three groups of subjects: those that consistently reported observing cigarettes being smoked in their workplace, individuals who consistently reported cigarettes being smoked around them away from work, and those who reported cigarettes being smoked inside their residence in their away-from-work (home) diaries. In Table 7 are presented the TWA airborne concentrations of ETS markers used to compute the exposures. Using the markers nicotine, FPM, and 3-EP, the data indicate that median exposures to ETS are approximately five times greater away from work than in the workplace. The data also indicates that the more highly exposed individuals in each category are exposed to ETS to a greater extent away from work. For example, comparing the 80th percentile cut points for the three categories for nicotine and 3-EP, the exposure cut points for the away-from-work locations which include a smoking home are twice the quantities found in the smoking workplaces. An examination of the Figure 3. Nonwork Tobacco Products Reported on Home Diary. Percent of Total Tobacco Products Reported By Location. data in Table 7 indicates that there are two primary contributors to the greater exposures away from work. First, TWA levels of ETS components are typically two - three times as high, and the average length of exposure is twice as long away from work. O 'I I I I I I , 1 I -10- I I