We used a self-administered questionnaire to explore predictors of workplace SHS exposure, as measured by personal exposure to vapor-phase nicotine, for a population of trucking industry workers. For nonsmokers, the self-reported duration of time spent in smoking-allowed areas, combined with the average number of people smoking in those areas, was a positive predictor of measured nicotine concentrations, suggesting that the self-reports of SHS exposure among the nonsmokers in this population is relatively reliable. In addition, the absence of a workplace smoking policy (as reported by the workers) was significantly associated with elevated personal nicotine levels among the nonsmokers. However, these factors were not statistically significant predictors of personal nicotine levels among the smokers. Not surprisingly, both smoking and nonsmoking drivers had statistically significantly lower (but detectable) levels of nicotine compared with nondrivers, after controlling for active smoking, because drivers in general spend less time with other workers than nondrivers do. Note that although drivers spend most of their work day alone in their truck cabs, nonsmoking drivers still have an opportunity to be exposed to SHS at the trucking terminals, at delivery docks, and on breaks. The “official” smoking policy, as reported by the terminal managers, and the existence of a state or county smoking ban were not associated with nicotine levels for either nonsmokers or smokers. The “perceived” smoking policy reported by the workers and the official policy agreed only about half of the time, implying either that enforcement and compliance of the smoking policy in these workplaces were poor or that in fact the official policy was not well correlated with actual exposure. It is interesting to note that, among the workers who worked at the terminals without any official smoking policy (n = 20), most of them (95%) perceived that their workplace has some restrictions on cigarette smoking.
Although smokers are certainly exposed to SHS from their own cigarettes and the other smokers around them, it is difficult to tease out predictors of these exposures because of the strong effects of active smoking. In a study of SHS in the home, Leaderer and Hammond (1991)
found a linear relationship between self-reported number of cigarettes smoked and area levels of vapor-phase nicotine. In our data, the amount of self-reported active smoking explained a large proportion of the variation in nicotine concentrations (R2
= 0.44 in the model including only the number of cigarettes smoked in the previous 2 days, age, and trucking terminal) and may have masked the smaller effects of workplace smoking policy or time spent with other smokers.
We did not find a statistically significant difference in nicotine concentration between the nonsmokers who reported an “indoor restricted area policy” and those who reported an “outdoor only policy” (), possibly because of the low power due to small sample size. In addition, this might be due to errors in the reporting of policy or the possibility of differences in behavior of smokers dependent on the policy that would affect the actual exposure levels. Another explanation is that indoor air might be contaminated by smoking right at doorways.
Blue-collar workers and service workers are more likely to be exposed to workplace SHS than are other worker groups (CDC 2006
; U.S. Environmental Protection Agency 1992
), likely due to higher smoking prevalence in these groups and the lower prevalence of smoke-free workplace policies. In the nationwide Current Population Survey (CPS) conducted in 1992–1993, > 70% of blue-collar workers and > 60% of service workers worked in workplaces that had no smoking restriction policy, a significantly higher percentage than among workers in other industries (Gerlach et al. 1997
). In a study examining trends in smoke-free workplace policies from 1992 through 2002 in North Carolina, Plescia et al. (2005)
found that blue-collar and service workers, especially males, were less likely to work in a smoke-free worksite than were white-collar workers throughout this 11-year study period. However, the overall coverage of workplace smoke-free policies increased from 46% to 71% nationwide, and the increasing trend was observed in all work groups (Plescia et al. 2005
). Our study, conducted more recently (2003–2005), found that most of unionized trucking company workers are currently working in worksites with some kind of onsite smoking policy.
In our study, only 23% of nonsmokers and 10% of smokers reported that the policies were always enforced. These numbers suggest a relatively weak policy enforcement in this work setting. In the Plescia et al. (2005)
study only 3% of workers reported that someone had violated the company policy in 2001–2002; however, when stratified by work groups, service and blue-collar workers reported slightly higher prevalence of noncompliance than did white-collar workers. In a CPS study conducted between 1999 and 2002, only 7% of nonsmokers reported that they experienced workplace SHS exposure (Pickett et al. 2006
). It is interesting to note, however, that most of the trucking terminals in our study were located in the states or counties where a smoking ban was not enacted by law at the time we conducted the study. However, recently the number of states with smoking ban regulations increased remarkably. During 2003–2005, only about 15% of the workers in this study worked at the terminals located in the states or counties with smoking ban regulation (); but in 2008, about 66% of these workers (67% for nonsmokers and 64% for smokers) would have been covered by a state or county workplace smoking ban if they still worked at the same terminals.
Studies of serum cotinine are also consistent with the observation that blue-collar workers experience higher exposures to SHS than do white-collar workers. In the Third National Health and Nutrition Examination Survey (NHANES III) conducted from 1988 through 2002, blue-collar workers had higher serum cotinine levels than did other adults (Arheart et al. 2008
; Pirkle et al. 2006
; Wortley et al. 2002
). The most recent study of NHANES III data suggested that although the serum cotinine levels for nonsmokers declined by 76% between 1988 and 2002 in all worker groups, blue-collar and service workers consistently had the highest levels of SHS exposure (Arheart et al. 2008
), mainly because they work in areas with a higher prevalence of smokers.
Although numerous studies have focused on measuring levels of SHS exposure in the workplace in the service sector (e.g., bars and restaurants) (Hyland et al. 2000
; Kiser and Boschert 2001
; Repace et al. 2006
; Weber et al. 2003
), we identified only a few studies conducted in the 1980s that focused on quantifying levels in the transportation industry. In a study of workers in three railroads in 1982–1983, the median level of estimated vapor-phase nicotine (converted from particle-phase nicotine) among nonsmoking nonoffice workers was 0.10 μg/m3
; Schenker et al. 1990
), which was much lower than the median level observed in our study (median of 0.93 μg/m3
for 183 nonsmoking nonoffice trucking workers). In 1983–1984, one of the railroads was revisited and vapor-phase nicotine levels were measured. The median nicotine level of nonsmoking engineers was 0.40 μg/m3
). This level was lower than the median level of 0.87 μg/m3
for nonsmoking truck drivers in our study, who worked at the similar condition as railroad engineers (small spaces, fewer other people). However, the nonsmoking office workers in the railroad were exposed to a greater nicotine level (median, 5.70 μg/m3
) in 1983–1984, compared with the 20 office workers in our trucking population (median, 0.59 μg/m3
In contrast to studies that used biomarkers, such as serum, urine, and salivary cotinine, to validate self-reported exposure to SHS (Emmons et al. 1994
; George et al. 2006
; Jenkins and Counts 1999
; Kemmeren et al. 1994
; Nondahl et al. 2005
; Seccareccia et al. 2003
), we measured vapor-phase nicotine concentrations in the breathing zone because of its ease of collection and because it avoided workplace biological sample collection and storage. In addition, the purpose of this study was to assess the SHS exposure pattern in the trucking industry, so we wanted only a measure of workplace exposure. Because biomarkers integrate exposures from all sources, they would not have been appropriate for this study. Moreover, the correlations between nicotine and commonly used biomarkers have been shown to be relatively good (LaKind et al. 1999
; Leaderer and Hammond 1991
In addition to evaluating the impact of smoking policy on workplace SHS exposure, we also attempted to assess the relationship between self-reported SHS exposure and vapor-phase nicotine levels among the trucking industry workers. Eisner et al. (2001)
conducted a study to validate a SHS exposure survey, using the same personal badge sampling device as used in our study, among 50 nonsmoking asthmatic adults. They found a moderate correlation (r
= 0.47) between self-reported SHS exposure duration and air nicotine concentration (median, 0.03 μg/m3
) in the previous 7 days. A study conducted by O’Connor et al. (1995)
among 415 nonsmoking pregnant women also demonstrated a similar correlation (r
= 0.41) between SHS exposure duration and nicotine concentration (median, 0.1 μg/m3
). These studies did not find a large impact of exposure intensity on the correlation between exposure duration and self-reported SHS exposure. In another study, Coghlin et al. (1989)
found a strong correlation between an SHS exposure score (duration in hours × number of smokers × proximity of smokers) with log-transformed nicotine (r
= 0.91) in 19 nonsmoking volunteers (nicotine level median, 2.0 μg/m3
). In our study (median, 0.87 μg/m3
for nonsmokers), we found that the exposure duration combined with number of smokers were significant predictors of nicotine level after adjustment of confounders. It is possible that the intensity of exposure, as measured by number of smokers, might become more important when the nicotine level is higher.
In our multiple regression models, we found that work shift may also influence SHS exposure. Finally, we also found a statistically significant relationship between educational level and nicotine concentration in the multiple regression model for nonsmokers. Previous studies have suggested that higher prevalence of active smoking is related to lower educational level in the United States (Kanjilal et al. 2006
) and many other countries (Gupta and Ray 2007
), but limited studies focused on the association between educational level and SHS exposure. Kanjilal et al. (2006)
reported that the smoking prevalence of the adults with lower educational level was significantly greater than the prevalence of those with higher educational level in the U.S. general population across 1971–2002. In our study, we did not find a significant association between active smoking and educational level among the smokers, possibly because of the relatively consistent social status in this blue-collar work group. However, we did find an association between lower educational level and workplace SHS exposure measured by personal sampling among the nonsmokers (). This implies that even within a blue-collar and relatively homogeneous population, educational status may still influence a nonsmoker’s exposure to SHS.
In summary, despite state and local movements toward smoke-free laws, this group of blue-collar workers was still exposed to workplace SHS as recently as 2005. Our findings suggest that most workplaces in this segment of the U.S. trucking industry have an official smoking policy. However, the workers’ perceived, rather than the official or state, policy was associated with measured SHS exposure levels among the nonsmokers. The self-reported duration and intensity of SHS exposure are relatively reliable among the nonsmokers in this work setting. In addition, factors such as job duties, work shift, and educational level might also be important predictors of workplace SHS exposure.