New Zealand has made significant progress over recent decades with reducing air pollution from tobacco smoke, especially in indoor environments1 and the Government recently announced the ambition to become a smoke-free nation by 2025 (frequently defined as a smoking prevalence below 5%).2 Nevertheless, a number of New Zealand studies on urban pubs1,3,4 and rural pubs5 have found evidence for secondhand smoke (SHS) drift from outdoor smoking areas to indoor areas (via open windows and doors).Studies in other countries have found that the particulate6 and nicotine7 air quality of indoor areas adjacent to outdoor smoking areas was compromised. Similar levels of SHS have been detected in hallways and near outdoor main entrances where smoking is permitted,8 such as entrances to office buildings.9 Likewise, a study measuring airborne nicotine concentrations to monitor SHS in different locations of a hospital, before and after a smoking ban, found the smallest reduction at the hospital main entrance and hallway when compared with all other areas.10Drifting SHS can be a public health concern for both patrons and workers in settings where outdoor smoking is permitted, particularly when levels of smoking are high. Indeed, a study from the United States indicated significant increases in markers for tobacco smoke absorption by non-smokers (salivary cotinine and a urinary marker [NNAL]) following outdoor SHS exposure in the outdoor areas of bar and restaurant settings.11 This is of concern as exposure to SHS has been linked to a number of health consequences such as lung cancer, coronary heart disease, sudden infant death syndrome and stroke.12A recent review of studies on the SHS drift in outdoor areas denotes that most studies reported similar mean PM2.5 concentrations between outdoor areas where smoking was permitted and the nearest smokefree indoor areas.13 Measured levels of SHS were higher where smoker density was high, smokers were in the vicinity of the area being measured, and where the outdoor smoking area was more enclosed.To date, there has been no investigation into the issue of possible SHS drift in restaurant settings in New Zealand, even though it has been found to be a problem in New Zealand pubs1,3,5,14 or in restaurants in other countries where indoor smoking is banned as well.15 Most international studies have looked into this potential drift issue in a number of hospitality settings (e.g. night bars, pubs and restaurants) together.Different hospitality venues might, however, have different background determinants influencing measured PM2.5 concentrations such as cooking smoke in venues where dinner is served.13 Moreover, we believe it is important to study the potential for SHS drift in restaurant settings separately as a higher number of children might be exposed (as compared to pubs or night bars) and the potential for longer duration of exposure (e.g. throughout a meal, rather than while having a beverage).This study therefore aimed to examine the levels of fine particulates (PM2.5) of SHS in outdoor smoking areas and in the adjacent indoor areas to assess possible drift via open windows and doors in a selected sample of restaurants in the urban centre of New Zealands capital, Wellington.In this study we aimed to:(i) Evaluate evidence of possible drift of fine particulates of SHS from: (a) outside selected restaurants (at the outdoor dining area), when compared to (b) the nearest tables indoors to the outdoor area; and (c) as far indoors as possible from the outdoor dining area; and(ii) Collect additional background data (e.g. presence of potential other sources of fine particulates from cooking and lit candles, number of smokers in the outdoor dining/smoking area, time windows/doors were open, wind speed, distance of indoor and outdoor seat to connecting doors/windows, and the extent of enclosure of the outside dining/smoking area (e.g. roofing, walls etc.).MethodsRestaurant selection We took a purposeful sample of eight restaurants in different areas of the Wellington urban core (e.g. Courtenay place area, Cuba street area, Central Business District and along the waterfront). Restaurants were selected if they were: (i) located in the Wellington urban core; (ii) had outdoor tables within five metres of door(s) or window(s) that connected to the restaurant interior; and (iii) were relatively popular among smokers (at least one person smoking upon onset of data collection period). All restaurants were located on streets with car and bus traffic. Traffic levels may have varied slightly. However, restaurants were purposefully selected in the downtown urban area to minimise this variation and ambient measurements were taken while walking between venues.Study period The study was conducted between February and April (late summer) in 2013 on Wednesday and Thursday evenings between 1800 and 2200h.Sampling locations Data were obtained from three different positions inside and outside restaurants simultaneously: (i) the outdoor dining areas (in the most central table available) where smoking was permitted; (ii) within the restaurant (at the table closest to the window or doors connecting with the outdoor dining area); and in a subset of restaurants (n=4) (iii) as far as possible within the restaurant away from the window(s) or door(s). The latter measurement was achieved (where feasible) by standing at the bar or most distant table from the window(s) or door(s) connecting the indoor area to the outdoor smoking area. A minimum of 10 minutes of sampling in each restaurant at each position was established, with a goal of thirty minutes of sampling per restaurant at the indoor and outdoor dining area. A team of two researchers stayed in the outdoor area and another such team spent time in both indoor locations (where feasible). Nevertheless, to avoid affecting occupants behaviour, the researchers behaved discretely and as typical customers (i.e. consumed drinks and meals).Air quality monitors The use of the two air quality monitors followed a protocol modified from and developed for a global air quality monitoring project16 and which has been used in other studies measuring exposure to SHS.1,3-5 In the sampling, fine particulates were measured (PM2.5) using portable real-time airborne particle monitors (i.e. the TSI SidePak AM510 Personal Aerosol Monitor, TSI Inc, St Paul, USA). The air monitors were carried hidden in a bag on the shoulder of one of the observers in each two-person team to sample the ambient air close to the breathing zone. During meals, the bags containing the air monitor were placed on the table or nearest empty seat and faced upward so as to simulate a realistic restaurant experience. A recent review of studies on the SHS drift in outdoor areas denotes that most studies have used fine smoking particulates measurements (PM2.5) as an indicator of SHS.13A calibration factor (0.32) for SHS based on empirical validation studies with the SidePak monitor was applied (i.e. adjusted in each monitors internal settings).17 The monitors were zero-calibrated routinely and internal components were set up as directed by the manufacturer prior to each day of data collection. Monitors were fitted with a 2.5 mcg impactor, had air flow rates of 1.7 L/min and had logging periods of 60 seconds. A length of Tygon\u2122 tubing was attached to the inlet of each monitor, with the other end left protruding slightly outside of the shoulder carry bag. Air quality data were then downloaded from the monitor to a computer, using TrakPro software (Shoreview, MN, USA).Other data collected The number of cigarettes smoked in the outdoor dining area was noted using a 30-second scan every 10-minutes during air quality data collection and reported as total number of patrons seen smoking outside per venue throughout data collection. Additional observational data were collected on the extent to which the outdoor dining area was enclosed (walls and roofing), wind speed (using a pocket air speed and temperature meter), distance in metres to the nearest connecting window/doors to the indoor dining area (using a laser distance meter), the percentage of time that the window/doors were open, and the number of tables in the indoor and outdoor dining area. In all the indoor settings, non-cigarette sources of fine particulates (e.g. burning candles, smoke from cooking areas) were noted. These data were discretely recorded on paper before being compiled in Microsoft Excel 2010 and used as background information when interpreting our results.Analyses Observational data about the restaurant facilities and other potential sources of PM2.5 were analysed in Microsoft Excel and frequencies and percentages for these features were reported. Air monitor data were exported from TrakPro software into Microsoft Excel for compilation. Next, data were exported for analyses in Stata v.12 (College Station, TX, USA). Descriptive statistics for PM2.5 measurements for indoor near entrance, indoor near back of restaurant, and outdoor locations were calculated for each venue and overall. We also ran paired t-tests to test for significant differences between mean PM2.5 measurements between the outdoor versus indoor near the entrance, where we had measurements taken at identical time points. We conducted ANOVAs to test for significant differences between the indoor location near the entrance and the indoor area at the back of the restaurant and between the outdoor dining area versus outdoor ambient air.Ethical approval We obtained approval through the University of Otago (Category B ethics approval process) and were cognisant of the ethical issues involved in this type of research.18ResultsBackground information Of the eight restaurants included in this study, we observed the presence of non-cigarette sources of fine particulates (e.g. burning candles, smoke from cooking area) in the indoor dining areas of six venues (see Table 1).Across all restaurants, we observed an average of nine patrons smoking in the outdoor dining area per venue throughout the air pollution sampling. There were on average 14 tables (range: 6-22) available in the indoor areas and 10 tables (range: 3-25) in the outdoor areas.The majority of outdoor dining areas (67%) were enclosed by either roofing and/or walls (with an average of three sides enclosed). Five of the selected restaurants had doors connecting the indoor and outdoor areas of the restaurants open 100% of the time.Three restaurants only had the doors open for movement of patrons and staff between the inside and outside of the restaurant and were opened between 10-50% of the time. None of the selected venues had windows opened between the indoor and outdoor areas. We also measured the wind speed at outdoor areas for all venues which ranged from calm to a light breeze (0.0-10.0 km/h). Table 1. Observational data by restaurant\r\n \r\n \r\n \r\n Item\r\n \r\n Restaurant\r\n \r\n \r\n \r\n I\r\n \r\n II\r\n \r\n III\r\n \r\n IV\r\n \r\n V\r\n \r\n VI\r\n \r\n VII\r\n \r\n VIII\r\n \r\n \r\n \r\n No. of tables in the outdoor area\r\n \r\n 3\r\n \r\n 5\r\n \r\n 25\r\n \r\n 6\r\n \r\n 21\r\n \r\n 9\r\n \r\n 4\r\n \r\n 15\r\n \r\n \r\n \r\n Total no. of patrons seen smoking in outdoor dining/smoking area during data collection\r\n \r\n 5\r\n \r\n 1\r\n \r\n 14\r\n \r\n 8\r\n \r\n 6\r\n \r\n 23\r\n \r\n 3\r\n \r\n 11\r\n \r\n \r\n \r\n Distance in metres from indoor area to entrance to outdoor dining/smoking area\r\n \r\n 1.78\r\n \r\n 2.03\r\n \r\n 0.96\r\n \r\n 0.05\r\n \r\n 2.00\r\n \r\n 3.50\r\n \r\n 4.80\r\n \r\n 1.30\r\n \r\n \r\n \r\n % of time connecting door(s) was/were open (and number of doors)\r\n \r\n 80 (1)\r\n \r\n 100 (1)\r\n \r\n 100 (4)\r\n \r\n 100 (1)\r\n \r\n 50 (1)\r\n \r\n 100 (1)\r\n \r\n 10 (1)\r\n \r\n 20 (1)\r\n \r\n \r\n \r\n % of time connecting window(s) was/were open\r\n \r\n 0\r\n \r\n 0\r\n \r\n 0\r\n \r\n 0\r\n \r\n 0\r\n \r\n 0\r\n \r\n 0\r\n \r\n 0\r\n \r\n \r\n \r\n % of outdoor dining/smoking area enclosed\r\n \r\n 50%\r\n \r\n 0%\r\n \r\n 80%\r\n \r\n 70%\r\n \r\n 50%\r\n \r\n 90%\r\n \r\n 20%\r\n \r\n 0%\r\n \r\n \r\n \r\n Wind speed (km/h)\r\n \r\n 0.30\r\n \r\n 2.80\r\n \r\n 2.60\r\n \r\n 0.30\r\n \r\n 0.00\r\n \r\n 0.00\r\n \r\n 0.00\r\n \r\n 0.00\r\n \r\n \r\n \r\n Burning candles observed in the indoor dining area (yes/no)\r\n \r\n yes\r\n \r\n no\r\n \r\n yes\r\n \r\n no\r\n \r\n yes\r\n \r\n yes\r\n \r\n no\r\n \r\n yes\r\n \r\n \r\n \r\n Cooking smoke observed in the indoor dining area (yes/no)\r\n \r\n no\r\n \r\n yes\r\n \r\n yes\r\n \r\n no\r\n \r\n yes\r\n \r\n no\r\n \r\n no\r\n \r\n no\r\n \r\n \r\n \r\n PM2.5 measurements We collected fine particulate (PM2.5) measurements at eight restaurants in the indoor and outdoor dining area for a total of 393 minutes (an average of 49 minutes per restaurant). We also collected PM2.5 measurements indoors in the restaurants as far away as possible from the doors/windows connecting the indoor with the outdoor area for a total of 67 minutes and ambient outdoor measurements for a total of 72 minutes.The results indicate a wide range of PM2.5 levels in the different restaurants. When examining results by each venue, very similar PM2.5 levels were observed between the outdoor and adjacent indoor dining area of venue III, where the mean PM2.5 level in the outdoor dining area was 38 mcg/m3, and in the indoor area 41 mcg/m3 (see Table 2). Table 2. Results of air quality monitoring (fine particulates, PM2.5) in different areas of restaurants and ambient air measurements in Wellington\r\n \r\n \r\n \r\n Restaurant\r\n \r\n Minutes measured (n)\r\n \r\n Mean PM2.5 (mcg/m3)\r\n \r\n Min PM2.5 (mcg/m3)\r\n \r\n Max PM2.5 (mcg/m3)\r\n \r\n \r\n \r\n \r\n \r\n Outdoor dining/smoking areas\r\n \r\n \r\n \r\n I\r\n \r\n 79\r\n \r\n 35\r\n \r\n 24\r\n \r\n 64\r\n \r\n \r\n \r\n II\r\n \r\n 31\r\n \r\n 37\r\n \r\n 23\r\n \r\n 178\r\n \r\n \r\n \r\n III\r\n \r\n 44\r\n \r\n 38\r\n \r\n 10\r\n \r\n 264\r\n \r\n \r\n \r\n IV\r\n \r\n 71\r\n \r\n 32\r\n \r\n 18\r\n \r\n 276\r\n \r\n \r\n \r\n V\r\n \r\n 82\r\n \r\n 28\r\n \r\n 8\r\n \r\n 134\r\n \r\n \r\n \r\n VI\r\n \r\n 52\r\n \r\n 74\r\n \r\n 13\r\n \r\n 321\r\n \r\n \r\n \r\n VII\r\n \r\n 13\r\n \r\n 41\r\n \r\n 13\r\n \r\n 170\r\n \r\n \r\n \r\n VIII\r\n \r\n 21\r\n \r\n 20\r\n \r\n 10\r\n \r\n 59\r\n \r\n \r\n \r\n Mean\r\n \r\n 49\r\n \r\n 38\r\n \r\n 15\r\n \r\n 183\r\n \r\n \r\n \r\n Indoor dining areas (at the closest table possible to the door/window connecting to the outdoor dining area)\r\n \r\n \r\n \r\n I\r\n \r\n 79\r\n \r\n 29\r\n \r\n 18\r\n \r\n 55\r\n \r\n \r\n \r\n II\r\n \r\n 31\r\n \r\n 46\r\n \r\n 24\r\n \r\n 106\r\n \r\n \r\n \r\n III\r\n \r\n 44\r\n \r\n 41\r\n \r\n 12\r\n \r\n 165\r\n \r\n \r\n \r\n IV\r\n \r\n 71\r\n \r\n 17\r\n \r\n 7\r\n \r\n 43\r\n \r\n \r\n \r\n V\r\n \r\n 82\r\n \r\n 56\r\n \r\n 5\r\n \r\n 208\r\n \r\n \r\n \r\n VI\r\n \r\n 52\r\n \r\n 26\r\n \r\n 5\r\n \r\n 154\r\n \r\n \r\n \r\n VII\r\n \r\n 13\r\n \r\n 12\r\n \r\n 6\r\n \r\n 20\r\n \r\n \r\n \r\n VIII\r\n \r\n 21\r\n \r\n 15\r\n \r\n 11\r\n \r\n 23\r\n \r\n \r\n \r\n Mean\r\n \r\n 49\r\n \r\n 34*\r\n \r\n 11\r\n \r\n 97\r\n \r\n \r\n \r\n Indoor dining area (as far away as possible from the door/window connecting to the outdoor dining area)\r\n \r\n \r\n \r\n III\r\n \r\n 19\r\n \r\n 24\r\n \r\n 13\r\n \r\n 37\r\n \r\n \r\n \r\n V\r\n \r\n 8\r\n \r\n 51\r\n \r\n 11\r\n \r\n 115\r\n \r\n \r\n \r\n VII\r\n \r\n 30\r\n \r\n 14\r\n \r\n 8\r\n \r\n 27\r\n \r\n \r\n \r\n VIII\r\n \r\n 10\r\n \r\n 18\r\n \r\n 8\r\n \r\n 40\r\n \r\n \r\n \r\n Mean\r\n \r\n 17\r\n \r\n 21\u2020\r\n \r\n 10\r\n \r\n 55\r\n \r\n \r\n \r\n Outdoor ambient air\r\n \r\n \r\n \r\n I & II\r\n \r\n 10\r\n \r\n 33\r\n \r\n 25\r\n \r\n 42\r\n \r\n \r\n \r\n II & III\r\n \r\n 30\r\n \r\n 21\r\n \r\n 17\r\n \r\n 43\r\n \r\n \r\n \r\n VII & VIII\r\n \r\n 32\r\n \r\n 19\r\n \r\n 13\r\n \r\n 108\r\n \r\n \r\n \r\n Mean\r\n \r\n 24\r\n \r\n 22\u2021\r\n \r\n 18\r\n \r\n 64\r\n \r\n \r\n \r\n * We did not detect a significant difference in mean PM2.5 levels between the outdoor dining/smoking areas and the adjacent indoor areas (p=0.149).\u2020 We found significantly higher PM2.5 levels at indoor areas near the entrance compared to indoor areas near the back of the restaurant (p=0.006).\u2021 We found significantly higher PM2.5 levels in the outdoor areas of restaurants where smoking was permitted compared to outdoor ambient levels while walking between venues (p <0.001). Although we observed lit candles and cooking smoke in the indoor area of this venue, the mean fine particulate measurement at the back of the restaurant (24 mcg/m3), as far away as possible from tobacco sources, was considerably lower. The highest mean (74 mcg/m3) and maximum PM2.5 concentration (321 mcg/m3) were observed in the outdoor dining/smoking area of restaurant VI, with a total of 23 patrons seen smoking throughout the data collection period (thus not at a single time point, but total during the course of the meal). The outdoor dining area of this venue was almost entirely enclosed. Although we did observe lit candles in the indoor dining area of this venue, the average indoor dining area measurement was considerably lower (26 mcg/m3), possibly due to a larger distance from the indoor dining area seat to the outdoor tobacco source and only one small door opened between the indoor and outdoor area.We noticed a wide range in the values of the outdoor PM2.5 measurements of the three venues where the highest total number of smokers were counted throughout data collection (restaurant III: 38 mcg/m3 (14 smokers), restaurant VI: 74 mcg/m3 (23 smokers), and restaurant VIII: 20 mcg/m3 (11 smokers). The same pattern was found in the fine particulate levels in the adjacent indoor dining areas near the connecting doors of restaurant III: 41 mcg/m3, restaurant VI: 26 mcg/m3, and restaurant VIII: 15 mcg/m3.Overall results indicate that the highest PM2.5 concentrations were observed in the outdoor dining areas (mean: 38 mcg/m3; range of maximum values: 59-321 mcg/m3) with highest levels being recorded in outdoor areas containing the largest number of smokers throughout data collection. Higher fine particulate matter levels were, however, also observed in the indoor dining areas adjacent to the outdoor dining/smoking area (mean: 34 mcg/m3; range of maximum values: 20-208 mcg/m3).Lowest overall mean levels were measured both at the indoor areas at the back of the restaurant (mean: 21 mcg/m3; range of maximum values: 40-115 mcg/m3) and in the outdoor ambient air (mean: 22 mcg/m3; range of maximum values: 42-108 mcg/m3).The results of the means comparisons indicate significantly higher PM2.5 levels at indoor areas near the entrance compared to indoor near the back of the restaurant (p=0.006) and in the outdoor smoking areas compared to outdoor ambient levels (p <0.001). We did not detect a significant difference in mean PM2.5 levels in outdoor smoking areas and indoor areas near the entrance (p=0.149).DiscussionTo our knowledge this is the first study that examines and compares fine particulates of SHS between outdoor smoking areas and the indoor areas in restaurant settings in New Zealand. Three main conclusions can be drawn from our findings. First, we observed similar concentrations of PM2.5 between the outdoor smoking areas and adjacent indoor areas of restaurants.A recent review of studies on the SHS drift from outdoor to indoor areas also reported that most studies found comparable mean PM2.5 concentrations between outdoor areas where smoking was permitted and the nearest smokefree indoor areas.13 However, in addition to most other studies, we also ran paired t-tests to test for a significant difference between mean PM2.5 measurements between the outdoor versus indoor area near the entrance, where we had measurements taken at identical time points. We did not detect a significant difference in the overall mean levels between these areas.We found significantly higher levels in the front areas of restaurants (near the outdoor smoking/dining area) compared to the back sections and significantly higher levels in the outdoor smoking areas of restaurants compared to ambient air levels. These results at least suggest the potential for SHS drifting through open windows/doors into adjacent indoor areas of restaurants, at levels higher than ambient pollutant levels.Secondly, PM2.5 measurements of the outdoor dining areas in the restaurant varied considerably due to different number of patrons who were smoking, but possibly also due to different design of the outdoor dining areas. Measurements were particularly high when we observed both a large number of smokers and if the outdoor area was enclosed, whereas we found lower measurements with large number of smokers when the outdoor area was designed to be more open (e.g. no roofing, no high walls). Sureda and colleagues (2013) also found measured levels of SHS to be generally higher where smoker density was high, smokers were nearby and where the outdoor smoking area was more enclosed.13 Lower wind speeds are generally associated with higher PM2.5 concentrations in urban settings.19 The low wind speed measurements in our study might have contributed to elevated fine particulate concentrations in the outdoor smoking areas in the places where a large number of smoking patrons were observed.Thirdly, in the case of high PM2.5 measurements in the outdoor dining area, we only found similarly high levels in the adjacent indoor areas if multiple doors were open at all time (but not if only one door was opened or if doors were open only some of the time) and if smokers were seated nearby the door connecting the outdoor with the indoor area. Research has shown that the distance from the source of tobacco smoke plays an important role in SHS exposure, with PM2.5 concentrations decreasing considerably with increasing distance from the tobacco source.20Comparing the mean PM2.5 levels found in our study to other similar studies, showed that the overall mean indoor restaurant PM2.5 measurement (34 mcg/m3) found in the present study was much higher than the average level that was found indoors in Irish pubs in Wellington (9.7 mcg/m3 also collected during late summer).4 This might be explained by other possible determinants that influence fine particulate levels in restaurants such as cooking smoke.13 Our findings of the adjacent indoor dining areas were more similar to the average level found in the indoor areas of bars/cafes collected in Wellington during a similar time period in 2011 (41 mcg/m3).3Our overall mean outdoor measurement (38 mcg/m3) was, however, much lower than the average level found for the outdoor areas of pubs and cafes in this same study (74 mcg/m3).3 However, the latter study selected a purposeful sample of semi-closed outdoor areas, which may explain higher PM2.5 concentrations, whereas the outdoor dining areas of our sample vary from open to completely enclosed outdoor areas with accompanying lower to higher PM2.5 levels resulting in a lower overall average in our study.This study has some methodological limitations, particularly the purposeful sample, and the small sample size. Two restaurants had visible cooking areas with open connection to the dining areas, and in three restaurants cooking smoke smells were reported, so there is the potential for indoor measurements being elevated by this or other non-cigarette sources such as lit candles.21 Despite this we found significantly lower mean measurements deeper inside the restaurants compared to areas near the front of the restaurant, suggesting that cook smoke or lit candles could not be the sole explanation for higher PM2.5 measurements near the front of restaurants (i.e. indoor area near the outdoor smoking/dining area).Although we showed that the overall levels of measured fine particulates were significantly lower while walking between venues compared to outdoor smoking areas of restaurants, indicating that traffic cannot be a sole explanation for fine particulates drifting from outside to indoor dining areas, we did not obtain ambient measurements outdoors directly adjacent to the smoking areas or at the exact same time points as the outdoor smoking measurements.Future studies could collect data during the peak of warm weather, from a wider range of restaurants that have outdoor dining areas (and also more of each type ranging from open to completely enclosed outdoor areas) - including from multiple New Zealand cities, during busier times with more smoking patronage such as in the weekends and for longer sampling time periods.In conclusion, although a future study with a larger sample of restaurants is warranted, our analyses at least suggest that SHS possibly drifts into the indoor dining areas of restaurants through the connecting open doors and windows. This might especially be a problem with high smoking patronage (e.g. during weekends), during peak summer season when generally most restaurants have all doors and windows opened, and when the outdoor dining/smoking area is partially to completely enclosed.To maximise the health p