BACKGROUND:Oil refinery workers are exposed to benzene, which is a well-known cause of leukaemia, but results on leukaemia in oil refinery workers have been mixed, and the data on workers' exposure is limited. Oil refinery workers are also exposed to asbestos and several studies have shown increased risk of mesothelioma. AIM:The objective was to investigate cancer incidence, especially leukaemia, at low to moderate exposure to benzene in an update of a previous study of employees at three Swedish oil refineries. METHODS:Cancer incidence was followed up in 2264 men (1548 refinery operators) employed at three oil refineries in Sweden for at least one year. Job types and employment times were collected from complete company files. A retrospective assessment of the benzene exposure was performed by occupational hygienists in collaboration with the refineries using historic measurements as well as detailed information on changes in the industrial hygiene and technological developments. Cases of cancer were retrieved by a linkage with the Swedish Cancer Register through 35-47 years of follow-up and standardized incidence ratios (SIR) with 95% confidence intervals (CI) were calculated. RESULTS:In total, 258 tumors had occurred versus 240 expected (SIR 1.07; 95% CI 0.95-1.21). There were 10 cases of leukaemia, all in refinery operators (SIR 2.4; 95% CI 1.18-4.51). There were three cases of pleural mesothelioma, two of which in refinery operators. The mean estimated cumulative benzene exposure for the cases of leukaemia was 7.9 ppm-years (median 4.9, range 0.1-31.1). DISCUSSION:The study suggests that low to moderate average cumulative benzene exposure increases the risk of leukaemia. Limitations include the modest number of cases and potential misclassification of exposure. CONCLUSION:The present study indicated an increased risk of leukaemia in male oil refinery workers with low to moderate exposure to benzene.

PDS 67: Outdoor air pollution cardiometabolic effects, Exhibition Hall (PDS), Ground floor, August 27, 2019, 4:30 PM - 5:30 PM Background: Air pollution is associated with incidence of cardiovascular morbidity and mortality. A main suggested mechanism is that air pollution accelerates the progression of atherosclerosis, and some studies have found associations with measures of atherosclerosis. For coronary artery atherosclerosis, relevant for myocardial infarctions, the few existing studies have found weak or null results though. We use the Swedish CArdioPulmonary bioImage Study (SCAPIS)-Gothenburg cohort to study associations between long-term air pollution and coronary artery calcium (CAC) score, a measure of calcified plaques. Methods: A cohort was recruited from the general population of Gothenburg during 2013-18 (n=6265, age 50-64) and thoroughly examined for cardiovascular risk factors, including computed tomography of the heart providing a CAC score. Participants were assigned yearly residential exposures to air pollution (source-specific PM10, PM2.5, NOx, BC, 1990-2015) from dispersion models. We used logistic regression to examine associations between long-term (15+ years) exposure to air pollutants and CAC (Agatston) score, adjusting for confounders and road traffic noise. Results: Median exposure to PM2.5 was low by international standards (mean 15-year exposure 8 µg/m3) and decreased during the study period. About 40% of participants had a positive CAC score, and there was a dose-dependent but non-significant increase in CAC at higher air pollution exposure, with 8% higher risk of high (>100) CAC score in the highest PM2.5 exposure quartile compared to the lowest. A positive CAC score was more common, and the association with PM2.5 stronger, in men. Road traffic noise did not affect the association. Conclusions: In this well-examined cohort we found a tendency of an association between long-term exposure to air pollution and coronary artery calcification, but not statistically significant. We aim to use the entire SCAPIS cohort (N=30000), recruited from 6 Swedish cities, to confirm or reject the hypothesis with greater power.

<b>In this general population sample with detailed control for smoking habits there is an association between blood cadmium and emphysema based on lung computed tomography</b>http://bit.ly/2K9S39J

TPS 621: Exposure to metals, Johan Friso Foyer, Floor 1, August 26, 2019, 3:00 PM - 4:30 PM Background/Aim: Urinary excretion of elements varies considerably over the day, so 24h excretion (mass per 24 hours) is a better metric, considered the gold standard which spot samples (often adjusted for creatinine) are compared with. There is, however, normal variability also between days in 24h urinary excretion, but since few studies did repeated sampling there is little information on the relation between within- and between person variability in 24h excretion of elements. Methods: We used a variability biobank with total 24h urine (split into 6 voids) on two separate days (about one week apart) from 60 healthy non-smokers (31 women and 29 men) from Gothenburg, Sweden, aged 21 – 64 years (mean 34 years). The 24h excretions of 22 elements were quantified using ICP-MS. Variability and intra-class correlation (ICC) were calculated using a mixed effects model of log-transformed 24h excretions, including sex as a fixed factor and individual as a random factor. Results: Variability between days was low for Cd, Co, Hg, Pb, Sn, Se, V, and Zn, having within-person variability of only 10 - 25% of total variability (ICC 0.75 – 0.90). It was moderate for As, Br, Cu, Fe, Li, Mn, Mo, Ni, P, S, U, and W (ICC 0.35 – 0.75), and highest for Cr (ICC=0.3) and Sb (ICC=0.18). For Fe, ICC was much lower in women than in men, but for other elements there were small differences. Conclusions: Although 24h excretion is the gold standard for urinary biomonitoring of elements, variability between days is high for some elements. For the toxic elements Pb, Hg, and Cd between-individual variability dominates.

Petroleum refinery workers are exposed to the carcinogens benzene and 1,3-butadiene. Declining exposures have been reported internationally but information on current exposure in the Swedish refinery industry is limited. The aim was to examine refinery workers’ personal exposure to benzene and 1,3-butadiene and increase awareness of exposure conditions by collaboration with involved refineries.

This study aimed to investigate if high levels of blood cadmium at baseline were associated with increased fracture risk during follow-up in middle-aged women. No increased fracture risk was observed during follow-up, but women with higher levels of cadmium had an increased overall mortality.

Wood smoke, a well-known indoor and outdoor air pollutant, may cause adverse health effects through oxidative stress. In this study 8-isoprostane, a biomarker of oxidative stress, was measured in exhaled breath condensate (EBC) and urine before and after experimental exposure to wood smoke. The results were compared with measurements of other biomarkers of oxidative stress and inflammation. Thirteen subjects were exposed first to clean air and then, after 1 week, to wood smoke in an exposure chamber during 4-hour sessions. Exhaled breath condensate, exhaled nitric oxide, blood and urine were sampled before and at various intervals after exposure to wood smoke and clean air. Exhaled breath condensate was examined for 8-isoprostane and malondialdehyde (MDA), while exhaled air was examined for nitric oxide, serum for Clara cell protein (CC16) and urine for 8-isoprostane. 8-isoprostane in EBC did not increase after wood smoke exposure and its net change immediately after exposure was inversely correlated with net changes in MDA (r(s)= -0.57, p= 0.041) and serum CC16 (S-CC16) (r(p)= -0.64, p= 0.020) immediately after the exposure. No correlation was found between 8-isoprostane in urine and 8-isoprostane in EBC. In this study controlled wood smoke exposure in healthy subjects did not increase 8-isoprostane in EBC.

Petroleum refinery workers’ exposure to the carcinogens benzene and 1,3-butadiene has decreased during normal operations. However, certain occupational groups or events at the refineries still involve a risk of higher exposures. The aim of this study was to examine the personal exposure to benzene and 1,3-butadiene at refinery turnarounds and during work in the oil harbour.

OBJECTIVES:To examine to what extent exposure to organic solvents during the working life affects general well-being in the long term, and to explore the relationship between self-reported symptoms and cognitive functioning in previously solvent-exposed floor layers.METHODS:The study included 41 solvent-exposed floor layers and 40 unexposed referents participating in a longitudinal follow-up study 18 years after the baseline assessment. Symptom prevalence and level of spare time activities were studied using the same methods as in the initial study. These include a general health examination, the Q16 symptom questionnaire, and a questionnaire for spare time activities. Relationships between symptoms and cognitive functioning were analysed based on recently published data on cognitive functioning of the participants at follow-up.RESULTS:At follow-up neuropsychiatric symptoms such as need to check things, depressive mood, and abnormal fatigue, were more prevalent among floor layers, particularly the most exposed individuals, than among referents. In addition, the most highly exposed floor layers reported more concentration difficulties and irritability. Fatigue and depressive mood increased over the follow-up time in the most exposed floor layers but not in the referents. Memory difficulties, although more frequent among floor layers than among referents, had decreased in floor layers while increased in referents. Floor layers also reported some negative effects on intimate relations and activity level. Neuropsychiatric symptoms were related to poorer performance chiefly in memory tasks and tests of complex attention and perceptual speed, more seldom in visuospatial tasks.CONCLUSIONS:Findings of exposure-related, long-lasting, partly deteriorating neuropsychiatric complaints indicate that general well-being later in life has been affected in floor layers with past heavy solvent exposure. We also found frequent associations between symptom prevalence and the cognitive functioning. Together with previous findings of dose-related cognitive decrements, the present results strengthen the evidence that long-term heavy occupational solvent exposure may negatively interact with the normal ageing process.

ISEE-560 Objective: Domestic wood burning for heating and pleasure is quite common in many countries with cold climates as Sweden, releasing air pollutants including particulate matter (PM). In an earlier study the trace elements Cl, K, Ca, Mn, Cu, Zn, and Rb were found to be significantly higher in homes using wood burning for space heating. The aim here is to investigate possible factors influencing the indoor levels of PM2.5, BS, and trace elements in the wood burning homes. Materials and Methods: Thirteen households were included in the study, 10 of which had boilers and 3 had fireplaces. GK2.05 (KTL) cyclones equipped with teflon filters and connected to pumps (BGI 400S) were used for the 24-hour indoor sampling. The filters were analyzed for mass and elemental concentration using energy dispersive x-ray fluorescence (EDXRF) spectroscopy and for black smoke (BS) using an EEL43 reflectometer. Information on wood-burning appliances (presence of water storage tank and type of appliance), and wood-burning (frequency of wood replenishments, amount and type of wood and burning time) was collected. Associations between the levels of indoor pollutants (log-transformed data) and the factors mentioned were assessed with stepwise backward multiple regression analysis. P-values <0.05 were considered statistically significant. Results: Larger amounts of wood used and few numbers of replenishments were associated with higher indoor levels of K, PM2.5 and BS, but for Ca and Cu only the amount of wood used. No associations were found for Cl, Zn, and Rb. Conclusions: As expected, larger amounts of wood used contributed to higher indoor levels. Lower emissions may be achieved with frequent replenishments because of the more complete combustion obtained with smaller batches of wood. This is a pilot study and further investigations are needed.

ISEE-251 Objective: The aim of this study was to survey annoyance and worry in people living close to a petrochemical area on the Swedish west coast, and to compare these findings with previous studies made in 1981, 1986, 1992, and 1998. The rationale is that authorities, the industry, and the public are concerned about the annoyance and possible health effects caused by the industry and correct information is needed to guide industry and policy makers. Material and Methods: Questionnaires were distributed to a stratified sample of about 750 residents 18 to 75 years old in different areas close (0.3–2 km) to petrochemical industry in Stenungsund, Sweden, and a control area farther away (25 km). The questions concerned various background factors, perceived health, and annoyance and concern caused by different environmental factors. The response rate was 72% (N = 542). Results for areas at different distances from the industry were compared with each other and with results from previous studies. Results: Respondents in the areas close to the petrochemical industry were significantly more often annoyed than those in the reference areas by industrial noise (about 30% vs. 1%), and by smell from the industry (about 20% vs. 3%). They were also more often “worried sometimes” about the health risks of industrial emissions (about 40% vs. 15%) and industrial accidents (about 30% vs. 10%). The prevalence of annoyance caused by smell from the industry had decreased in the last few decades, whereas annoyance caused by industrial noise had increased. Worry about health risks was at the same level as in 1992 and 1998. Conclusions: In the area close to the petrochemical industry, annoyance caused by noise and smell was common, as was concern about health risks. Smell problems had decreased somewhat, either because of reduced emissions or increased acceptance. There seems to be room for further improvements.

P-737 Introduction: Anesthetic gases such as desflurane, sevoflurane, isoflurane, enflurane and halothane are used on a daily bases in operating theatres. Waste anesthetic gases may have a negative effect on the members of the operating team as well as on the personnel in the resuscitation rooms. There is a need for an easy and convenient procedure for measuring the personal exposure to these gases. SKC 575–002 is a diffusive sampler filled with Anasorb 747. This sampler is widely used in the sampling of volatile gases, but has not been validated experimentally for use with halogenated anesthetic gases. The calculated uptake rates, recommended by SKC Inc., are based on geometry and physical characters. Our aim was to validate the sampler at various conditions. Methods: We validated the samplers in a laboratory dynamic system with standard atmospheres (4–140 mg/m3) of the compounds. Parameters tested were desorption efficiency, concentration, relative humidity, back-diffusion (clean air for 7.5 h after 0.5h of exposure), as well as storage stability for up to 8 weeks. Uptake rates were achieved by comparing with an active sampling approach (Gilian pumps (n=2–4)), providing independent controls for each sampling trial lasting for 1, 4 or 8 hours. Analysis and detection were performed using a gas chromatograph-flame ionization detector (GC/FID). Analysis of variance was used to evaluate influence of time, concentration, dose and relative humidity. Results: The desorption efficiencies ranged from 100 to 108% for all targets compounds. The uptake rates ranged from 11 to 15 mL/min and the levels were lower than those reported by SKC Inc. There were statistically significant effects of relatively humidity for all compounds (approx 10% lower uptake rates at 25% RH as compared with 60% RH). We also found significant effects (10–20%) of time and concentration for all substances with lower uptake rates at higher concentrations and longer sampling times. All samplers were run in replicates (n=5–8) and the relative standard deviations (RSD) were mostly low (<10%). The effects on back-diffusion or losses due to storage up to 8 weeks were insignificant. Discussion and Conclusions: Effects on uptake rates of sampling time, concentration and relative humidity were moderate. The samplers SKC 575–002 with Anasorb 747 could be used for personal monitoring of occupational exposure to halogenated anesthetic gases. Specific uptake rates, taken time and concentration into account, could be used or alternatively the lowest uptake rate found to ensure that levels are not underestimated.

Particulate air pollution is known to increase cardiovascular morbidity and mortality. Proposed mechanisms underlying this increase include effects on inflammation, coagulation factors, and oxidative stress, which could increase the risk of coronary events and atherosclerosis. The aim of this study was to examine whether short-term exposure to wood smoke affects markers of inflammation, blood hemostasis, and lipid peroxidation in healthy humans. Thirteen subjects were exposed to wood smoke and clean air in a chamber during two 4-h sessions, 1 wk apart. The mass concentrations of fine particles at wood smoke exposure were 240-280 mug/m3, and number concentrations were 95,000-180,000/cm3. About half of the particles were ultrafine (< 100 nm). Blood and urine samples were taken before and after the experiment. Exposure to wood smoke increased the levels of serum amyloid A, a cardiovascular risk factor, as well as factor VIII in plasma and the factor VIII/von Willebrand factor ratio, indicating a slight effect on the balance of coagulation factors. Moreover, there was an increased urinary excretion of free 8-iso-prostaglandin2alpha, a major F2-isoprostane, though this was based on nine subjects only, indicating a temporary increase in free radical-mediated lipid peroxidation. Thus, wood-smoke particles at levels that can be found in smoky indoor environments seem to affect inflammation, coagulation, and possibly lipid peroxidation. These factors may be involved in the mechanisms whereby particulate air pollution affects cardiovascular morbidity and mortality. The exposure setup could be used to establish which particle characteristics are critical for the effects.

P-634 Introduction: Atmospheric dispersion modelling is used to estimate levels of various air pollutants. The aim was to determine the accuracy in these models by comparing calculated residential outdoor levels of nitrogen dioxide (NO2) based on emission databases and dispersion modelling, with levels measured by diffusion samplers. Methods: Repeated, simultaneous measurements of personal exposure and residential outdoor levels of NO2 for ten non-smoking subjects were carried out for three measurement campaigns during the winters 2004 and 2005. Each campaign lasted for seven days and the diffusive sampler Willem's badge was used. The sampler is equipped with a glass fiber filter impregnated with triethanolamine and acetone. The uptake rate was 0.46 ml/min and the limit of detection was 0.11 μg/m3. During personal measurements the sampler was worn in the breathing zone and for outdoor measurements it was placed outside the residence. GPS coordinates for the location were noted and the height of the sampler above ground was used in the modelling program. A receptor modelling strategy (ENVIMAN) was used to calculate the residential outdoor NOX concentrations. NOX concentrations were converted to NO2 using the specific relationship between NOX and NO2 obtained from measurements at the urban background station of Gothenburg. A mixed ANOVA model (Proc Mixed, SAS) was used to estimate the relationship between measured and modelled residential outdoor NO2 concentrations. Results: In total, 30 stationary outdoor and 26 personal exposure measurements were successfully performed. There was a statistically significant difference (p<0.05, Wilcoxon signed rank test) between the residential outdoor levels and the personal exposure for each of the three campaigns, where the residential outdoor levels exceed the personal exposure. The median personal exposure was 17, 10 and 8 μg/m3 respectively for the three campaigns. Corresponding median residential outdoor levels were 24, 14 and 14 μg/m3. The intra-class correlation coefficient (ICC) between measured and modelled residential outdoor levels was high, 0.84. The regression equation for measured versus modelled NO2 was: MEASURED = 1.89 + 0.94 * MODELLED. Confidence intervals (95%) for the slope: (0.74–1.14) and intercept: (-2.89–6.67). Discussion and Conclusions: A good agreement was found between modelled and measured residential outdoor levels of NO2 and shows that modelling was feasible for estimating the stationary outdoor levels. The personal exposure, however, was significantly lower than the residential outdoor concentration, as could be expected since levels of NO2 indoors are generally lower than outdoors if no major indoor sources are present.

P-614 Introduction: Urban background stations are generally affected by long-range transported pollution (LDT) from different source areas, but little is known about the impact on personal exposure and indoor concentrations. The aim was to assess the influence of different air masses on the measured trace element concentrations at different places. Methods: Simultaneous personal, indoor and urban background 24 hour samples of PM2.5 were collected simultaneously during the spring and autumn seasons of 2002 and 2003. For the personal and indoor sampling a GK2.05 (KTL) cyclone and a BGI 400S personal sampling pump (4 l min-1) were used. The urban background samples used an EPA-WINS impactor (16.7 l min-1). Trace element concentrations were determined on the Teflon filters by Energy Dispersive X-Ray Fluorescence (EDXRF) analysis. Air mass back trajectories were computed using the NOAA ARL HYSPLIT Model (Draxler and Rolph, 2003) to investigate the effect of LDT. For each sampling day, 96-hour air mass back trajectories were computed at startup time and 6, 12, 18, and 24 hours later. The trajectories were divided into four classes, representing different source areas, namely, continental areas (central and Eastern Europe), marine areas, the Nordic countries (Finland, Norway, and Sweden), and the United Kingdom. Results: Air mass origin strongly affected the measured urban background concentrations of some elements (S, Cl, V, Ni, Br and Pb). These findings were also seen for personal exposures (S, Cl, V and Pb) and indoor levels (S, Cl, V, Ni and Pb). For the crustal elements Ca, Ti, Mn, Fe and Cu, no differences were observed between any of the air masses. Discussion and conclusion: When comparing the different air masses that reached Göteborg, some of the differences between trajectory classes were large enough to have a statistically significant impact also on indoor and personal exposure levels. This was the case for particles from combustion processes (S, V, Ni, and Pb) from industrialized regions (continental areas and the UK) and sea salt particles (Cl) from marine areas. As could be expected, there were no differences between elements of crustal origin. A better understanding of how regional transport processes influence personal exposure is important for interpretation of time series analyses of air pollution and health. Draxler, R.R., Rolph, G.D. 2003. HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model. NOAA Air Resources Laboratory, Silver Spring, MD. www.arl.noaa.gov/ready.html.

P-304 Background: Exposure to polycyclic aromatic hydrocarbons (PAHs) increases the risk of cancer in humans, and benzo(a)pyrene (BaP) and fluoranthene (Flu) are used as indicators of the carcinogenic PAHs. Domestic wood burning is considered to be the major source of PAHs in Sweden together with road traffic. The aim of this study was to measure the indoor and outdoor levels of BaP and Flu in a small Swedish town. Methods: A total of 20 non-smoking subjects living in two different residential areas in Hagfors (5,600 inhabitants), Sweden, were investigated during the winter 2000. One area consisted of single-family houses where wood burning for heating is common (WB), whereas the other consisted of apartments with district heating (DH). Stationary indoor air sampling were performed in the bedrooms for 24 hours using active samplers, each containing a glass fibre filter and two polyurethane foam adsorbents in series, at a flow rate of 2 l/min. Outdoor measurements (3-days sampling) were performed at one site in each area (N=6). Repeated indoor measurements were done in a subgroup of homes (N=10). A mixed-effect model was used to estimate the effect of residential area and variability in indoor levels. The PAHs were determined by GC-MS using electron ionization (EI) and selected ion monitoring (SIM). Results: Flu was found in both the gaseous and the particulate phase, while BaP was found only in the particulate phase. The median indoor level of Flu was significantly higher (p=0.0045) in the DH area compared to WB (3.1 ng/m3, range 1.4–19 ng/m3 versus 0.91 ng/m3, range 0.64–6.4 ng/m3). Medians of BaP indoors were 0.07 ng/m3 (DH; 0.05–0.71 ng/m3) and 0.14 ng/m3 (WB; 0.04–0.80 ng/m3) and the difference was not statistically significant. Outdoor median level of Flu and BaP was higher in WB than in DH (Flu; 3.4 ng/m3, range 2.6–4.0 ng/m3 versus 1.6 ng/m3, range 0.91–2.3 ng/m3 and BaP; 0.47 ng/m3, range 0.03–0.64 ng/m3 versus 0.04 ng/m3, range 0.03–0.06 ng/m3). For BaP the variability of the bedroom levels was dominated by the day-to-day variability (97%), whereas for Flu the between-bedroom variability dominated (89%). Discussion and Conclusions: Levels of BaP around the health-based guideline of 0.1 ng/m3 were found indoors in both residential areas. In the district heating area the indoor levels of Flu exceeded the health-based guideline of 2 ng/m3. Indoor sources may contribute to the higher indoor levels of Flu in the district heating area.