Secondhand Tobacco Smoke Exposure in Open and Semi-Open Settings: A Systematic Review

Background: Some countries have recently extended smoke-free policies to particular outdoor settings; however, there is controversy regarding whether this is scientifically and ethically justifiable. Objectives: The objective of the present study was to review research on secondhand smoke (SHS) exposure in outdoor settings. Data sources: We conducted different searches in PubMed for the period prior to September 2012. We checked the references of the identified papers, and conducted a similar search in Google Scholar. Study selection: Our search terms included combinations of “secondhand smoke,” “environmental tobacco smoke,” “passive smoking” OR “tobacco smoke pollution” AND “outdoors” AND “PM” (particulate matter), “PM2.5” (PM with diameter ≤ 2.5 µm), “respirable suspended particles,” “particulate matter,” “nicotine,” “CO” (carbon monoxide), “cotinine,” “marker,” “biomarker” OR “airborne marker.” In total, 18 articles and reports met the inclusion criteria. Results: Almost all studies used PM2.5 concentration as an SHS marker. Mean PM2.5 concentrations reported for outdoor smoking areas when smokers were present ranged from 8.32 to 124 µg/m3 at hospitality venues, and 4.60 to 17.80 µg/m3 at other locations. Mean PM2.5 concentrations in smoke-free indoor settings near outdoor smoking areas ranged from 4 to 120.51 µg/m3. SHS levels increased when smokers were present, and outdoor and indoor SHS levels were related. Most studies reported a positive association between SHS measures and smoker density, enclosure of outdoor locations, wind conditions, and proximity to smokers. Conclusions: The available evidence indicates high SHS levels at some outdoor smoking areas and at adjacent smoke-free indoor areas. Further research and standardization of methodology is needed to determine whether smoke-free legislation should be extended to outdoor settings.

Secondhand smoke (SHS) is a complex mixture of thousands of compounds includ ing particulate matter emitted by the com bustion of tobacco products and from smoke exhaled by smokers [International Agency for Research on Cancer (IARC) 2004]. It contains > 50 chemicals recognized as known and probable human carcinogens, other ani mal carcinogens, and many toxic and irri tant agents (U.S. Department of Health and Human Services 2006). Over the past two decades, scientific evidence has accumu lated linking SHS exposure to adverse health outcomes, including respiratory outcomes in children and adults, acute cardiovascular effects, and lung cancer (IARC 2004;Ott et al. 2006;U.S. Department of Health and Human Services 2006). Most of this evi dence is based on longterm SHS exposure research (IARC 2004). Some recent studies have also reported evidence of effects follow ing shortterm exposure to tobacco smoke, such as eye irritation and respiratory irrita tion among nonsmokers (Junker et al. 2001). Even brief and shortterm exposures to SHS may generate significant adverse effects on the human respiratory system, as discussed in a recent review (Flouris and Koutedakis 2011). Finally, Pope et al. (2001) suggested that effects of acute exposure to tobacco smoke on cardiac autonomic function may contribute to pathophysiological mechanisms linking exposure to SHS to increased risk of cardio vascular mortality.
Smokefree policies have been expand ing worldwide since the World Health Organization (WHO) encouraged countries to follow Article 8 of the Framework Convention on Tobacco Control (FCTC) (WHO 2003) to protect people from SHS (Globalsmokefree Partnership 2009). Legislation has been widely implemented in indoor public places, work places, and public transportation (WHO 2009). Since the implementation of indoor smokefree environ ments, several studies have demonstrated important reductions of SHS exposure, including an 80-90% decrease in previously highexposure settings, such as workplaces and hospitality venues such as bars and restaurants (IARC 2008). However, indoor smoking bans may increase the likeli hood that smokers will gather at convenient outdoor locations such as public areas near building entrances (Kaufman et al. 2010a). In 2007, a revision of the FCTC Article 8 guidelines further recommended that quasi outdoor and outdoor public places should be smokefree under some circumstances, and called upon countries to "adopt the most effective protection against exposure wher ever the evidence shows that hazard exists" (WHO 2009). Recently, some countries have extended smoking bans to some out door locations (Globalsmokefree Partnership 2009;Repace 2008), particularly health care centers and settings where children are pres ent (Globalsmokefree Partnership 2009). However, there remain some outdoor loca tions close to smokefree areas where people may be exposed to SHS, such as terraces and patios in hospitality venues and near entrances to smokefree buildings (Globalsmokefree Partnership 2009).
Some controversy exists regarding whether smoking should be prohibited in outdoor set tings (Chapman 2008; Thomson et al. 2008). Health concerns about SHS exposure, nui sance from SHS, litter, fire hazards, concern about establishing positive smokefree mod els for youth, and reducing youth opportu nities to smoke (Bloch and Shopland 2000;Brennan et al. 2010;Cameron et al. 2010;Chapman 2008;Repace 2008;Thomson et al. 2008Thomson et al. , 2009) exemplify the reasons why smoking should be banned in selected out door locations. Outdoor smoking bans might also support smokers who are trying to quit by limiting their overall cigarette consump tion (Williams et al. 2009). Selected outdoor smoking bans should also help to denormal ize smoking in outdoor areas (Thomson et al. 2008). In a number of jurisdictions, the major ity of the public supports restricting smoking in various outdoors settings, and this support appears to be increasing over time (Thomson et al. 2009). However, those who oppose out door smoking bans argue that it is ethically unsustainable because it does not respect the principle of freedom and autonomy of indi viduals, and that there is insufficient evidence that SHS in these environments has an impact on health (Chapman 2000(Chapman , 2008. SHS exposure has been commonly stud ied in different indoor locations, especially in workplaces such as hospitality venues or health care centers (IARC 2009); however, outdoor SHS has been scarcely evaluated. It has been hypothesized that the introduction of indoor smoking bans has led to a relocation of smok ers to outdoor areas, with a subsequent increase of tobacco smoke levels in outdoor places (Sureda et al. 2012). The aim of the pres ent study is to review research on objectively assessed SHS levels in outdoor settings, includ ing information on indoor and outdoor SHS concentrations, the effect of smoking bans on indoor and outdoor SHS levels, the relation between outdoor and indoor SHS levels, fac tors that influence outdoor and indoor SHS concentrations, and whether measured SHS levels comply with the air quality standards established by the WHO (2005).

Methods
We conducted several different searches in PubMed (http://www.ncbi.nlm.nih. gov/pubmed) for papers published before September 2012 to identify papers on SHS assessment in outdoor settings. We combined different terms as follows: (("Secondhand smoke" OR "environmental tobacco smoke" OR "passive smoking" AND "out door") OR ("Tobacco Smoke Pollution"[Mesh] AND "outdoor")) AND ( The search was more sensitive than specific; therefore, we arrived at the first selection of manuscripts by checking the results of every search and reading titles and abstracts. We then obtained the selected papers and read them carefully. Finally, we completed our search by checking the references of the papers and conducting similar searches in Google Scholar (http://www.scholar.google. com/; with search terms in English).
Our final selection included studies whose main objectives were to measure SHS or tobacco smoke exposure in outdoor set tings using a tobacco biomarker or airborne marker. Outdoor areas included completely open spaces and quasioutdoor areas with temporary or permanent structures, such as a roof or side walls, that would impede upward or lateral airflow, respectively.
We excluded articles that studied SHS exposure indoors but not outdoors and articles that studied air pollution outdoors, but not specifically SHS. We were able to consider papers in English, French, German, Italian, Spanish, and Catalan.

Results
Our initial searches identified 263 papers; after checking the titles, 67 abstracts were reviewed ( Figure 1). Of these, 51 were determined not to meet eligibility criteria. We read the remain ing 16 papers in full, plus 6 additional papers identified from references. We finally identi fied 18 articles and reports that satisfied the inclusion criteria, including 15 published in peerreview journals and 3 academic reports available on the Internet. One report was a pilot study for which we obtained data from the subsequently published study (Klepeis et al. 2007). We included only results related to SHS in outdoor areas from another report [California Air Resources Board (CARB) 2005] concerning SHS exposure in California.
The studies included between 2 and 127 locations. Depending on the specific study objectives, different locations were tested. Nine studies were conducted in hospitality venues (Table 1) such as pubs, restaurants, bars, cafés, and outdoor dining areas. Six stud ies measured SHS in other locations such as entrances to buildings and the adjacent indoor area and transportation settings, including an airport, parks, streets, university campuses, and one junior college campus (Table 2). Three studies assessed SHS in both hospital ity and nonhospitality venues. Most stud ies were observational studies, with only two experimental studies. All included papers were written in English.
SHS in outdoor smoking areas. Mean PM 2.5 concentrations reported for outdoor smoking areas at hospitality venues ranged from 8.32 µg/m 3 (Stafford et al. 2010) to 124 µg/m 3 (Wilson et al. 2007) when smok ers were present (Table 2). In nonhospitality venues, mean PM 2.5 concentrations reported for outdoor settings ranged from 4.60 µg/m 3 (Boffi et al. 2006) to 17.80 µg/m 3 (Boffi et al. 2006) (Figure 2). Klepeis et al. (2007) obtained an overall PM 2.5 mean of 30 µg/m 3 for the observational data for hospitality   (Klepeis et al. 2007). Three studies Parry et al. 2011;Stafford et al. 2010) that compared outdoor SHS measurements dur ing smoking and nonsmoking periods reported that particulate concentrations were significantly higher during active smoking. Two studies reported that PM 2.5 concentra tions in outdoor smoking areas were higher than background PM 2.5 levels similarly mea sured in nearby, smokefree, outdoor air (St. Helen et al. 2011;Travers et al. 2007). An additional study (Boffi et al. 2006) reported high PM 2.5 concentrations both outdoors and indoors during 1 day in a conference center where smoking was permitted.
One study used salivary cotinine to evalu ate SHS exposures among nonsmokers before and after they spent 6 hr at smoking areas of outdoor bars or outdoor restaurants, or at an outdoor control site without smoking (Hall et al. 2009). Median increases in salivary coti nine from pretest to posttest were approxi mately 162%, 102%, and 16% for the bar, restaurant, and control sites, respectively. A similar study measured salivary cotinine in saliva and NNAL in urine samples from nonsmokers before and after being at an out side bar or restaurant or at a control site (St. Helen et al. 2012). Cotinine in samples col lected both immediately after and the morning after 3hr visits to the outside bar and restau rant sites were significantly higher than in the control samples, and NNAL was significantly higher in first morning urine samples after bar and restaurant site visits. Another study used airborne nicotine to assess SHS exposure; the mean 8hr concentrations ranged from 0.013 to 3.1 µg/m 3 (higher than the mean 8hr back ground concentrations of 0.009-0.12 µg/m 3 ) (CARB 2005).

Factors influencing outdoor SHS levels.
Atmospheric conditions, including wind direc tion, wind speed, and atmospheric stability, can modify outdoor SHS levels. Other factors are the density and distribution of the smok ers and the structure of the outdoor location (completely open or semiopen). All of the studies that evaluated possible modifiers of SHS concentrations reported that the den sity of smokers and/or number of lit cigarettes predicted outdoor SHS Cameron et al. 2010;CARB 2005;Edwards and Wilson 2011;Kaufman et al. 2010b;Klepeis et al. 2007;López et al. 2012;Parry et al. 2011;Repace 2005;St.Helen et al. 2011Stafford et al. 2010;Sureda et al. 2012). Most of these studies also found the degree of enclosure of the outdoor area as a determi nant factor Cameron et al. 2010;López et al. 2012;Parry et al. 2011;Stafford et al. 2010;Sureda et al. 2012;Travers et al. 2007). For example, Cameron et al. (2010) reported that PM 2.5 increased by approximately 30% with each additional active smoker within 1 m of the point of mea surement, and by 50% if measured under an overhead cover.
Some studies on wind conditions (speed and direction) and proximity to smokers found that these were not associated with SHS levels (Kaufman et al. 2010b;Travers et al. 2007). However, the CARB study (2005) and two experimental studies (Klepeis et al. 2007;Repace 2005) in public outdoor locations that controlled smoking activity at precise distances from monitored positions reported that out door SHS levels were highly dependent on wind direction and source proximity. Klepeis et al. (2007) demonstrated that upwind PM 2.5 concentrations are likely to be very low, whereas downwind levels during periods of active smoking can be very high. They also reported that PM 2.5 levels decreased by half or more as the distance from a lit cigarette increased from 0.25-0.5 m to 1-2 m, and that levels were generally close to background. However, Repace (2005) reported that out door PM 3.5 and PAH concentrations did not approach background levels until about 7 m.
Outdoor smoking areas and indoor air quality. PM 2.5 concentrations in indoor set tings where smoking was banned but near outdoor smoking areas varied from 4 µg/m 3 (Kaufman et al. 2010b) to 120.51 µg/m 3 ; both studies were carried out in hospitality venues. Indoor PM 2.5 levels far away from outdoor tobacco sources were lower (Sureda et al. 2012;Wilson et al. 2011).
Two studies specifically examined SHS in main entrances of public buildings. Kaufman et al. (2010b) simultaneously measured PM 2.5 concentrations inside and outside of 28 office building entrances. Outdoor SHS levels within 9 m of building entrances were significantly higher in the presence of smoking (11 µg/m 3 with 1-4 cigarettes, and 16 µg/m 3 with ≥ 5 cigarettes) compared to occasions when there was no smoking (8 µg/m 3 ). PM 2.5 median indoor concentrations ranged from 4 to 6 µg/m 3 . Sureda et al. (2012) showed higher median PM 2.5 concentrations in the presence of smoking, both outdoors near main entrances (17.16 µg/m 3 ) and in indoor halls near outdoor smoking areas (18.20 µg/m 3 ), compared with those in control locations without smoking, both indoors (10.40 µg/m 3 ) and outdoors (13.00 µg/m 3 ).
Several articles reported positive associations between SHS levels (PM 2.5 concentrations) measured indoors and outdoors Edwards and Wilson 2011;Kaufman et al. 2010b;López et al. 2012;Sureda et al. 2012;Wilson et al. 2011). Indoor SHS levels are higher when smoking occurs in the adja cent outdoor setting, especially when the out door area is semienclosed. For example, Sureda et al. (2012) showed that PM 2.5 concentrations in indoor halls were more closely correlated with outdoor concentrations measured near main entrances (outdoors) than with the indoor control (a nonsmoking area far from the main entrance). Brennan et al. (2010) estimated that a 100% increase in the geometric mean of the outdoor PM 2.5 concentration was associated with a 36.1% rise in the geometric mean of the indoor PM 2.5 concentration in smokefree pubs and bars.
Factors influencing indoor SHS from outdoor areas. Factors such as wind speed and direction that modify outdoor SHS levels also may influence indoor air quality. The effects of structural barriers between outdoor smoking areas and indoor locations were also considered in some articles   2007) included hospitality and non-hospitality venues without distinguishing the mean value between them, and hence it has been included both in "hospitality venues" and "other venues." Wilson et al. (2011) and Edwards and Wilson (2011) provided the individual values for each measurement, and we have computed the arithmetic mean for the figure. Brennan et al. (2010) and López et al. (2012) provided mean and median values, respectively, for venues before and after a smoking ban. We have computed the average values for each study to include them in the figure. observed that open access between indoors and outdoors was associated with lower PM 2.5 levels indoors. However, an Australian study  showed higher indoor PM 2.5 concentrations when doors to outdoor smoking areas were left open.

Smoking bans and SHS exposures.
One study evaluated the impact of laws prohibit ing indoor smoking ) by measuring PM 2.5 concentrations before and after indoor smoking bans were implemented in pubs and bars that had at least one indoor area with an adjacent semienclosed outdoor eating/drinking area, and showed reduced PM 2.5 concentrations both indoors and out doors (65.5% and 38.8%, respectively) from preban to postban. Two other studies eval uated indoor and outdoor SHS in different settings after the implementation of indoor smoking bans (Wilson et al. 2007. Both reported higher concentrations of fine particulates in outdoor smoking areas, espe cially those that were partly enclosed, as well as indoor areas adjacent to outdoor smoking areas compared to other smokefree indoor settings. Finally, a multicenter study carried out in hos pitality venues of eight European countries compared SHS concentrations between venues where indoor smoking was allowed and venues where it was banned . The authors reported that median indoor PM 2.5 and airborne nicotine concentrations were sig nificantly higher in venues where smoking was allowed than in those where it was banned. Conversely, the outdoor nicotine concentra tion was significantly higher for venues where indoor smoking was banned than outdoor areas of venues where indoor smoking was allowed . Tobacco smoke levels compared to background levels. Maximum mean or median out door PM 2.5 concentrations ranged from 128 µg/m 3 (Sureda et al. 2012) to 496 µg/m 3 (Kaufman et al. 2010b), with some point measurements exceeding 1,000 µg/m 3 (Klepeis et al. 2007;Travers et al. 2007). The maxi mum peak indoor PM 2.5 concentra tion reported for a smoke free setting was 239 µg/m 3 . In contrast, mean or median back ground PM 2.5 concen trations varied from 6 µg/m 3 (Travers et al. 2007) to 20.4 µg/m 3 (St.Helen et al. 2011).
SHS markers other than PM 2.5 . Three studies evaluated different SHS markers to determine which would be most appropri ate to describe SHS levels in outdoor areas. Sureda et al. (2012) reported a Spearman cor relation coefficient between outdoor PM 2.5 and airborne nicotine concentrations of 0.365 (95% CI: 0.009, 0.650). Hall et al. (2009) reported that the number of smokers pres ent had a strong positive association with outdoor PM 2.5 concentrations but not CO concentrations. Moreover, CO levels mea sured outside restaurants and bars did not differ significantly from concentrations mea sured at a control location, in contrast with findings for PM 2.5 concentrations. Other studies used biological markers such as coti nine or NNAL to show SHS exposure (Hall et al. 2009;St.Helen et al. 2012).

Discussion
We found only 18 studies that met our cri teria, but these indicated that SHS levels in some outdoor smoking areas are not negligi ble, especially in areas that are semienclosed.
SHS levels and air quality standards. In general, SHS levels measured in out door smoking areas were high, particularly in hospitality venues where PM 2.5 concen trations ranged from 8.32 µg/m 3 (Stafford et al. 2010) to 182 µg/m 3 (Hall et al. 2009) when smokers were present. SHS levels were also increased in indoor areas adjacent to out door smoking areas. Hall et al. (2009) andSt.Helen et al. (2012) reported that saliva cotinine concentrations were higher in study participants following exposure to SHS at outdoor bars and restaurants when smoking was allowed than after exposure to smoke free terraces. These results suggest that hos pitality workers and patrons may be exposed to high SHS levels under certain conditions. Although outdoor SHS levels are more tran sient than indoor levels, and can quickly drop to background levels in the absence of active smoking, potential health effects of these exposures merit consideration and need to be further studied.
According to the WHO, there is no safe level of SHS (WHO 2000). The WHO guidelines indicate that the lower range of concentrations at which adverse health effects have been demonstrated is not greatly above background concentrations (estimated at 3-5 µg/m 3 in the United States and Western Europe for PM 2.5 ). In the updated WHO Air Quality Guidelines, an annual outdoor aver age value of 10 µg/m 3 for PM 2.5 was selected as the lower end of the range over which sig nificant effects on survival have been observed (Gorini et al. 2005;WHO 2000WHO , 2005. These are the lowest levels at which total, cardiopulmonary, and lung cancer mortality have been shown to increase with more than 95% confidence in response to PM 2.5 . Most of the reviewed studies of PM 2.5 concentra tions in outdoor smoking areas reported levels higher than the annual mean guideline value of 10 µg/m 3 recommended by WHO Influences of outdoor SHS on indoor air quality. Indoor smokefree areas near out door smoking areas showed higher levels than smokefree indoor areas that were farther away from outdoor SHS sources, suggest ing that SHS from outdoor smoking areas can enter adjacent buildings. Some findings also suggested that although outdoor SHS concentrations dropped immediately to back ground levels when the SHS sources were extinguished, indoor SHS concentrations persisted at relatively high levels and slowly decayed over several hours until doors were opened to ventilate the building (Klepeis et al. 2007). SHS levels in outdoor locations are more susceptible to variation due to the proximity of active smoking and wind condi tions. During periods of active smoking, out door SHS levels can be comparable to levels in indoor smoking areas, but outdoor levels dropped rapidly after smoking activity ceased.
Other factors influence SHS levels. Some factors can influence SHS levels both indoors and outdoors Cameron et al. 2010;Edwards and Wilson 2011;Kaufman et al. 2010b;Klepeis et al. 2007;López et al. 2012;Repace 2005;St.Helen et al. 2011Stafford et al. 2010;Sureda et al. 2012). Smoker density and enclosure of the outdoor locations are determinant modifi ers. Some studies also suggest that wind speed and direction, as well as proximity to smok ers, are associated with SHS levels outdoors.
SHS airborne markers other than PM 2.5 . Particulate matter was the most common air borne marker used in the presently reviewed articles. However, PM 2.5 is not a specific marker; markers such as airborne nicotine are specific to SHS (Gorini et al. 2005;Ott et al. 2006). Biological markers have been scantily used. However, cotinine has been proposed as a very sensitive and specific biological marker of SHS exposure (Benowitz 1999), and total NNAL has been used to characterize human exposure to carcinogenic tobaccospecific nitrosamines among nonsmokers exposed to SHS (Anderson et al. 2001). Further research is necessary to evaluate which SHS marker would be most appropriate to measure SHS levels in outdoors settings and whether it would be necessary to combine more than one marker.
Limitations. Some of the reviewed studies did not control for important factors that can influence SHS levels, such as wind conditions, the structural characteristics of outdoor area (semienclosed vs. totally open), or proxim ity to active smokers. Future studies should control for these factors to enable a better understanding of the results. Additionally, some studies used PM 2.5 concentrations to estimate SHS levels in outdoor areas, but did not control for other sources of PM 2.5 , such as cooking or trafficrelated air pollution (Gorini et al. 2005). Further studies should record the presence of other sources of combustion, such as cooking facilities, proximity to roadways, or traffic density; measure and report back ground levels of PM 2.5 ; and/or use specific SHS markers such as airborne nicotine.
volume 121 | number 7 | July 2013 • Environmental Health Perspectives Publication bias is a potential source of error in systematic reviews. We searched the available literature in PubMed, the main bio medical database, and Google Scholar and checked references to identify documents not published in academic journals. However, we cannot rule out the possibility that some unpublished manuscripts or other documents addressing the topic of interest may have been missed. Direct comparisons of results among studies were hampered by the use of differ ent statistics (medians, means, or geometric means) and sampling strategies; the use of standardized methods could strengthen the validity of results and facilitate comparisons among different populations and locations. Furthermore, the number of venues mea sured in each study was limited. Future stud ies should consider including representative samples of locations selected using standard statistical sampling procedures and sample size computations.
Strengths. The reviewed studies included a variety of venue types (e.g., entrances to public buildings, hospitality venues, transpor tation settings) and characteristics. Most of the reviewed studies were observational, and thus provide information that reflects smok ing behaviors and exposures under normal reallife conditions. However, experimental studies provide the opportunity to control for unpredictable variables, such as the proxim ity of smokers or wind conditions. The use of realtime monitoring permits determination of the precise magnitude of extremely transient (shortterm) concentrations and exposures, while retaining the flexibility of exploring con centrations and exposure across a variety of averaging times and time series and calculating mean concentrations and exposures (Klepeis et al. 2007).

Conclusion
Only limited evidence is available regarding SHS exposure in outdoor settings as deter mined by environmental and biological mark ers; therefore, the existing evidence must be interpreted carefully. However, our review clearly indicates the potential for high SHS exposures at some outdoor settings and indoor locations adjacent to outdoor smoking areas. This review shows that high smoker density, highly enclosed outdoor areas, low wind con ditions, and close proximity to smokers gen erate higher outdoor SHS concentrations. Accounting for these factors is important for future studies on the relationship between outdoor SHS exposure and health outcomes.
The WHO Framework Convention on Tobacco Control has concluded that 100% smokefree environments are required to adequately protect the public's health from the harmful effects of SHS (WHO 2003). The present review indicates that further research using standardized methodology is needed to better characterize outdoor SHS exposure levels and determine whether smokefree legislation should be extended to outdoor areas.
Future studies should include repre sentative samples of different locations; use standardized statistical analyses and report multiple measures of central tendency and measures of variability (standard errors, con fidence intervals, or quartiles); and consider potential modifiers of SHS levels including smoker density, degree of enclosurement of outdoor locations, wind speed and direc tion, and proximity to smokers. Finally, further research is needed to determine the most appropriate marker or combination of markers to assess SHS exposure, which may include more specific environmental and individual markers of exposure (e.g., airborne nicotine and cotinine in saliva) in addition to PM 2.5 concentration.