Airborne toxic exposure and pulmonary outcomes among petrochemical complex workers

Petrochemical workers are exposed to a variety of airborne toxic compounds which have been associated with increased risk for respiratory outcomes. However, long-term exposure to SO 2 , NO 2 , O 3 , H 2 S and NH 3 in relation to spirometric parameters and self-reported respiratory problems is largely unknown. Airborne concentration levels of SO 2 , NO 2 , O 3 , H 2 S and NH 3 were collected from two xed stations over a 3-year period in a petrochemical complex. We assessed spirometric parameters and respiratory symptoms in the petrochemical workers (n = 200) and in an unexposed group (n = 200). We calculated β-coecients (β) and odds ratios (ORs) with 95% condence intervals (CIs) before and after adjustment for covariates.


Background
Petrochemical workers are exposed to a wide variety of airborne toxic and carcinogenic compounds and there is evidence for increased risk for several diseases such as different types of cancer (1)(2)(3), renal disease (4), hematological changes, and as well as pulmonary disease (5) among workers in petrochemical industries (4,6). Emission of aerial toxic substances into the work environment at petrochemical plants and factories mostly happens through release of chemicals during production processes (e.g., separation, conversion, treating, combustion exhaust), but also from leakage of stored products and during transportation (7)(8)(9).
A recent investigation found that leakages from ammonia-lled transport vessels led to raised aerial ammonia levels in the working environment of an Iranian petrochemical plant (4). Ryerson and colleagues reported that petrochemical plants could be a signi cant source of volatile organic compounds (VOCs), nitrogen oxides (NOx), and ozone (O 3 ) (10). Also, a study of air pollutants within a radius of 5 km around a large complex of oil re neries and petrochemical manufacturing plants showed elevated airborne concentrations of sulfur dioxide (SO 2 ), carbon monoxide (CO), nitrogen dioxide (NO 2 ) and O 3 (11). Hydrogen sul de (H 2 S) is another pollutant that is produced naturally as one of the components of crude petroleum and natural gas. Operation processes including those in petroleum re neries, natural gas plants, and petrochemical plants are the major sources of H 2 S, which is released into the environment (12). A Taiwanese re nery study reported emissions of NO 2, SO 2 and O 3 , the latter as a result of the photochemical reaction of VOCs and NOx in the presence of sunlight (13). Another study in Romania found SO 2 and H 2 S toxicants in the ambient air of almost all production units of a petroleum re nery (14).
Re nery workers have been reported to be exposed to toxic gases in the workplace (4,12,22), which has raised attention towards exposure assessment and health effects in this occupational group. However, long-term monitoring of exposure to SO 2 , NO 2 , O 3 , H 2 S and ammonia (NH 3 ) has, to our knowledge, not been examined in relation to objectively assessed spirometric parameters among petrochemical workers. In the current study, we therefore aimed to describe the long-term aerial concentration levels of SO 2 , NO 2 , O 3 , H 2 S and NH 3 in the ambient air of an Iranian petrochemical complex. Further, we compared spirometric parameters and respiratory symptoms between these petrochemical workers and unexposed controls.

Participants
This retrospective cohort study was counducted in the Iranian petrochemical complex, consisting of 10 re nery and 15 petrochemical plants, in one of the largest industrial sites in the Middle East. Therefore, all workers in this industrial area will, to some extent, experience exposure to toxic chemicals in their work environments (22,23).
Eligibility criteria for the petrochemical worlers were (i) at least one year in their current position and (ii) no history of respiratory diseases or chemical exposures prior to the current position. Using the census sampling method, 200 males engaged in repair and maintenance (n = 182), warehous keeping (n = 9) and ancillary activities (n = 9) were included as petrochemical workers in the study. As unexposed subjects, we included 200 men with the same eligibility critera as the exposed, 100 were sampled from o ce workers and 100 from gas power plant operators, located 200 km away from the petrochemical complex. In total, the study population constituted 400 Iranian males.
The Ethics Committee of Qom University of Medical Sciences approved the protocol of the current study. All subjects gave written informed consent and all data was de-identi ed after collection. collected from one station every second week over 3 years and one station over 1.5 years (total number of valid samples N = 101). Total amount of SO 2 per sample was extracted using ion chromatography (24). To measure NO 2, a triethanolamine coated sampler was used, placed in a polypropylene tube of 9.8 mm internal diameter and 7.35 cm length. The data collection and exposure period was the same as for SO 2 (total number of valid samples N = 103).
The Griess-Saltzman method was applied to determine collected NO 2 (25). To measure O 3 , a sodium nitrite coated diffusive sampler in a polypropylene housing (opening of 20 mm) was used. The samples were collected biweekly from one measurment station over 3 years and analyzed by ion chromatography (26) (total number of valid samples N = 46). H 2 S was measured as molecules who diffused onto cadmium sulphate as an absorbent in a passive sampler. The samples were collected monthly from two measurement stations over 3 years and analyzed by spectrophotometry (27) (total number of valid samples N = 60). To measure NH 3, we collected airborne ammonia by absorption on phosphoric acid, collected monthly from two measurement stations over 3 years (total number of valild samples N = 58). The amount of NH 3 was determined spectrophotometrically by the indophenol method (28).
All sampling devices were protected from rain and wind by placing them in a special shelter. All devices were equipped with a glass bre membrane, supported by a wire net, to reduce wind disturbance. All samples were set up at the middle of each month and removed two weeks-one month later. Because of invalid measurement methods (sampler placement error, using invalid sampler, late or early removal of sampler) or instrument errors (e.g. errors in preparation, collection or analysis) the following proportions of samples had to be excluded from our analysis: 7% of SO 2 samples (101/108), 5% of NO 2 samples (103/108), 36% of O 3 samples (46/72), 17% H 2 S (60/72) and 20% of NH 3 samples (58/72).
Because the exposure assessment methods were based on stationary, as opposed to personal measurements, it was not possible to directly link exposure to each individual worker. Therefore, for the current analyses we assumed that all petrochemical workers were exposed to the same levels of environmental pollutants.
Because the background exposure level of the investigated pollutants might be high in vicinity of the industrial site, we selected the control group of unexposed workers from the city > 200 km away from petrochemical complex.
Levels of average daily exposure for the control population to airborne toxicants were extracted from the literature. Two independent studies reported allmost the same values for public exposure of SO 2 , NO 2 and O 3 from 2006-2015 where the controls were living and working (29,30) while ambient air concentrations of H 2 S and NH 3 was exteracted from other reports (31,32). Two other studies showed that the maximum emmited SO 2 and NO 2 of gas power plants in Iran did not exceed 15 and 5 µg/m 3 , respectively (33, 34).

Spirometry
We measured FVC, FEV1, FEV1/FVC and peak expiratory ow (PEF) at the beginning of each participant's work shift during the period March to December 2017. We used a calibrated spirometry device (Spirolab III, MIR Italy) and performed the measurements according to the American Thoracic Society guidelines (35). Each participant was examined three times and the best test result for each spirometric parameter was used for the analysis (36). Participants were asked to not smoke, or engage in heavy excersice for at least one hour before the test. Additionally, they were asked to abstain from alchol for at least 8 hours before the test. Spirometric parameters were standardized, based on height-, age-and sex-strati ed reference data (general Caucasian population) recommended by the European Respiratory Society and expressed as percent predicted (100%=general population) (37, 38).
Obstructive impairment was de ned as having grade 2 impairment in obstructive parameters.

Respiratory symptoms
All participants completed the European Community Respiratory Health Survey questionnaire II (ECRHS II) to assess prevalence of self-reported respiratory problems (41). The ECRHS II has been widely used to assess prevalence of respiratory symptoms including wheezing, chest tightness, cough, chronic cough, mucus, chronic phlegm and shortness of breath in populations who were occupationally or environmentally exposed to aerial chemical subestanses (42,43). We further generated a binary variable no respiratory problems vs. any respiratory problem based on the answers to the ECRHS II.

Covariates
We assessed the following information by questionnaire: age at study entry, height, weight, working hours per week, length of employment, smoking status (smoker, non-smoker), number of cigarettes smoked per day and number of years smoking, marital status (single, married), education (high school diploma or less, higher education) and residential status (permanent residents, non-permanent residents). We calculated the number of pack-years by multiplying the number of packs of cigarettes smoked per day by the number of years the person has smoked (44).

Data analysis
Statistical analyses were carried out with Stata MP 14.0 (Stata; TX, USA). We used descriptive statistics (means ± SD and numbers with proportions) to present levels of environmental air pollution in vicinity to the the petrochemical complex and the prevalence of lung dysfunction and respiratory symptoms in exposed and unexposed participants. Fisher's exact test and two sample t-test were used to test for signi cant differences between the exposed and unexposed workers. Univariable and multivariable linear regression models were used to analyse the spirometric parameters in relation to exposure status (exposed vs. unexposed). Univariable and multivariable logistic regression models were used to analyze risk of self-reported respiratory problems and obstructive and restrictive lung dysfunction (grade 2 or higher) in relation to exposure status. The multivariable models were adjusted for number of working hours per week, work type, age at study entry, weight, height, marital status,education level, length of employment, current smoking status, and residential status. A p-value < 0.05 was considered statistically signi cant.

Study population
We included 200 workers exposed to SO 2 , NO 2 , O 3 , H 2 S and NH 3 and 200 unexposed workers ( Table 1). The two groups differed by all covariates, except number of cigarette pack-years and educational level. For exposed and unexposed, the respective means were, 37 and 35 years of age when answering the questionnaire, 44 and 48 working hours per week, 10 and 7 years of employment, 8 and 11 cigarettes smoked per day among 26% and 9% current smokers, respectively.
All spirometry parameters were signi cantly (p < 0.001) reduced in the petrochemical workers compared to the unexposed workers when looking at the crude outcomes (Table 1).
Among the petrochemical workers, nine (4.5%) were classi ed with obstructive impairment compared to one among the unexposed (Table 1). In total, 75 (37%) of the petrochemical workers (exposed) and 30 (15%) of the unexposed workers reported at least one respiratory symptom (p < 0.001). The petrochemical workers reported most often shortness of breath followed by mucus. Those two symptoms were reported more often by the petrochemical workers than the unexposed workers (both p < 0.001) ( Table 1). Abbreviations: N, number; SD, standard deviation.
* P-value comparing exposed and unexposed participants ** Diploma or less includes: compulsory school only, high school; higher education includes: bachelor degree, master degree or higher.
*** Mean (SD) calculated in current smokers only (n exposed =52, n unexposed =17) † calculated by multiplying the number of packs of cigarettes smoked per day by the number of years the person has smoked † † Permanent residents: workers who live in the city; Non-permanent residents: workers who work and live temporarily (usually 14 day/month) in the city ‡ "no impairment"=FEV 1 /FVC ≥ 70%-predicted (grade 0), "mild impairment"=FEV 1 /FVC < 70% and FEV 1 ≥ 80%predicted (grade 1), "moderate impairment"=FEV 1 /FVC < 70% and FEV 1 < 80-50%-predicted (grade 2); ‡ ‡ "no or mild imperment"=FVC < 75%-predicted (grade 0 and 1), moderate or higher impairement = FVC < 75%predicted and FEV1/FVC > 0.7 (grade 2 or higher) Table 2 shows the overall average aerial concentration of the ve measured toxicants during the the 3-year sampling period as well as the ambient recommendation levels (   Table 3). After adjusting for covariates, no signi cant difference in obstructive and restrictive impairment was observed between the exposed and unexposed workers.  Table 4). None of the other symptoms were signi cantly associated with exposure from work in the petrochemical facility.

Discussion
In the present study, we examined spirometric parameters and respiratory symptoms among petrochemical workers exposed to airborne toxic pollutants compared to unexposed workers. Concentrations of SO 2 at the petrochemical complex were considerably higher than those recommended by the WHO (48). The average levels of NO 2 and O 3 were within the WHO recommendations but 25% of all samples taken were above the recommendations for both gases. However, the environmental exposure concentrations did not exceed the OSHA-PELs for workers (47). We found signi cant reductions in spirometric parameters (FEV1, FEV1/FVC, PEF) and that higher proportions reported mucus and shortness of breath among petrochemical workers compared to unexposed workers.
The current study demonstrated a higher exposure level of SO 2 , NO 2 , O 3 , H 2 S and NH 3 among petrochemical workers compared to mean daily exposures of the general population. Although the concentration levels of all pollutants in our study were above the daily exposures in the general population, only SO 2 and NO 2 exposure levels exceeded the WHO-ARL (45). This might suggest that SO 2 and NO 2 are more important drivers for respiratory problems among petroleum workers than other pollutants. This nding is in line with other reports observing that SO 2 and NO 2 were among the most common chemical pollutants released into the atmosphere at the petrochemical workplaces (49,50). However, our maximum concentration levels were higher than those of other studies where SO 2 and NO 2 have shown ranges of 105-550 µg/m 3 and 9.6-87 µg/m 3 , respectively (50)(51)(52)(53). This inconsistency in air concentration levels might be due to differences in environmental conditions (temperature, wind, rain, etc.) (54), location of the plant (55), type of production (55), time and period of measurement (56), but also methodological differences in the sampling, the modeling systems, and in the equipment (4).
Although all exposure levels in our study were below the OSHA-PEL, our ndings might suggest that the airborne toxic exposures are harmful even at a lower level if exposed over a long time. The petrochemical workers were on average employed for 10 years in their current position. The reductions we found for FEV1 and PEF among the petrochemical workers are in line with Neghab and colleagues who found a signi cant decrease in FEV1 and FVC among operational petrochemical workers (57). Also, Meo and colleagues observed a signi cant decline in FEV1, FEV1/FVC and PEF but no change in mean FVC among oil re nery workers compared to a control group (8). In another study, FEV1/FVC was signi cantly lower in petrochemical workers than in controls (58). Exposure to SO 2 and NO 2 have been associated with reductions in FVC and FEV1 (59)(60)(61), which suggest that the reductions in spirometric parameters (FEV1, FEV1/FVC, PEF) observed in our study might be due to SO 2 and NO 2 exposure.
Although only nine (< 5%) of the petrochemical workers in our study had obstructive impairment, our ndings are pointing in the same direction as those of others where exposure to high levels of NO 2 and SO 2 have been associated with obstructive impairment, although low concertations of these toxicants may not lead to obstructive impairment (62)(63)(64).
Our nding that risk of any respiratory symptom was elevated among the exposed group, is in line with other studies in petroleum and gas re nery workers (14,65). We also found signi cant increases in reported mucus hypersecretion and shortness of breath among petrochemical workers, which accord with the results of a study in residents near petro-re nery plants where symptoms such as shortness of breath, cough, phlegm and weakness were reported (66).
Mucus hypersecretion and shortness of breath are symptoms of obstructive lung disease (67,68), but the increase we found in these self-reported symptoms deviates from the objective spirometric assessment where only nine petrochemical workers had obstructive impairment (grade 2 or higher), highlighting the importance of an objective assessment in addition to self-reports.
Generally, using personal protective equipment (PPE) is routine in the petrochemical industries of Iran, but not throughout the whole shift and during all operations (69). Therefore, continuous improvement in PPE use throughout the shift, could minimize exposure to toxic chemicals and is warranted. In addition, applying control measures to reduce SO 2 and NO 2 emissions to environmentally acceptable levels are urgent. Our ndings also indicated that the current recommendations for occupational exposure limits (OELs) are not su cient to preserve the workers' lung function and avoid respiratory problems. This suggests that OELs in this occupational setting needs to be revised to minimize adverse health effects.
The current study was conducted in a well-de ned group of petrochemical and unexposed workers with measurement data of gaseous pollutants, objective assessment of spirometric parameters, as well as self-reported respiratory symptoms. A major strength is the long-term measurement of ve pollutants with standardized and high quality sampling techniques and instruments resulting in reliable levels of airborne concentrations in the vicinity of the petrochemical plant. Because the literature suggested a relationship between exposure to low level of any of SO 2 (70), NO 2 (71), O 3 (72), H 2 S (73) and NH 3 (57,74) and respiratory symptoms, the current study focused on these pollutants rather than other petrochemical airborne chemicals such as VOCs. Although air concentrations of gaseous pollutions were measured in a reliable way, we did not have individual exposure data and had to assume a constant level of exposure among all workers in the petrochemical plant.

Conclusion
This study indicated that even exposure to low-level of gaseous pollutants, particularly SO 2 and NO 2 , could adversely affect spirometric parameters and signi cantly increase respiratory problems among petrochemical workers. Availability of data and materials

Abbreviations
The data that support the ndings of this study are available from corresponding author but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of Qom University of Medical Sciences

Competing interests
The authors declare that they have no competing interests Mean value (with 95% CIs) of spirometric parameters in percent predicted by unexposed and exposed to petrochemicals. * p-value <0.001 ** FVC: Forced vital capacity; FEV1: Forced expiratory volume in one second; PEF: Peak expiratory ow