Introduction

Methyl tert-butyl ether (MTBE) is a common oxygenated fuel additive, which is used to enhance gasoline octane levels and reduce emissions of carbon monoxide and other pollutants from combustion engines1. MTBE has been extensively applied in the petroleum refining industry to meet environmental regulations aimed improve air quality2. Despite environmental benefits of this substance, evidence on health concerns due to exposure to it has increased3. MTBE’s physical-chemical properties, including high volatility and water solubility, facilitate its emission into the atmosphere4.The American Conference of Governmental Industrial Hygienists (ACGIH) has recommended a threshold limit value (TLV) of 50 ppm for respiratory exposure of MTBE to protect workers from potential health risks, including the respiratory tract irritation, neurological symptoms, and possible carcinogenicity5. Nevertheless, the International Agency for Research on Cancer (IARC) and other agencies have introduced MTBE as a possible human carcinogen (Group 2B), which has prompted stringent regulatory measures in several countries6. In response to health concerns and environmental contamination, several countries including the United States, Canada, and several European countries have restricted or banned the use of MTBE in gasoline since the 2000s6. However, production and use of MTBE continue in numerous developing countries where there is demands for cost-effective fuel additives, and regulatory enforcement may be less stringent7. Consequently, refinery workers are still potentially exposed to MTBE emissions, which highlight the ongoing importance of comprehensive health risk assessments in such environments8.

Several research have addressed the health risk due to exposure to MTBE in refinery. For instance, in occupational settings, Hsieh et al. (2006) conducted an occupational exposure assessment in Taiwanese petroleum refineries and reported that MTBE concentrations were between undetectable value and 145.6 µg/m3, with hazard quotients (HQ) higher than acceptable limits in some cases9. Similarly, Perbellini et al. evaluated exposure to MTBE among refinery workers in Italy and found that inhalation risks were higher than acceptable limits for certain job categories10. In environmental settings also, Lin et al. (2011) observed increase in MTBE concentrations in ambient air near fuel storage sites, which highlight the potential of community exposure11. These findings emphasize the need for ongoing monitoring and risk assessment in areas with probability of exposure to MTBE12. Despite the studies on health risk assessments of MTBE, a significant limitation in many previous studies is the non-consideration of seasonal variability in risk estimates13. Temperature and humidity, which vary significantly between cold and warm seasons, influence the physicochemical behavior, volatilization rate, atmospheric dispersion, and degradation pathways of MTBE14. During colder months, lower ambient temperatures generally reduce the evaporation, concentration, and health risks of MTBE. Conversely, warmer conditions enhance volatilization, which potentially increase inhalation exposure among refinery workers14. Humidity also affects the adsorption-desorption dynamics of MTBE on surfaces and respiratory tract uptake efficiency and complicates exposure patterns15.

The omission of seasonal factors in many health risk evaluations leads to oversimplified assessments, and it may underestimate peak exposures and risks, particularly during summer months16. For example, Li et al. (2018) demonstrated that non-consideration of seasonal variation in MTBE emissions underestimated the maximum exposure concentrations by up to 25%, which potentially result in insufficient worker protection. Furthermore, Vandenberg et al. (2020) argued that single-season assessments do not consider the full spectrum of exposure scenarios, thereby limiting the effectiveness of risk management strategies17. Although some studies have attempted to incorporate seasonal variations, their approaches often rely on deterministic models that are unable to address the inherent variability and uncertainty in exposure parameters18,19.

To overcome these challenges, probabilistic modeling techniques such as Monte Carlo simulation, have emerged as robust tools for health risk assessment under uncertainty and variability16. Monte Carlo simulation applies repeated random sampling of input variables from defined probability distributions, which produce a range of possible outcomes20. This contrasts with traditional deterministic approaches, which provide single-point risk estimates often unable to consider the true variability in exposure or toxicological response21. Therefore, Monte Carlo methods enable the integration of multiple sources of uncertainty including emission rates, meteorological conditions, and worker activity patterns22.

Developing and implementing an innovative Monte Carlo simulation-based framework for the health risk assessment of exposure to MTBE in various occupations of refinery environments during various seasons can provide a clear view of its health risks. This allows for a more accurate estimation of seasonal fluctuations in health risks during both cold and warm periods. The proposed framework serves as an effective tool to enhance occupational health monitoring and risk management strategies within petroleum refining industries. Therefore, the present study aimed to investigate seasonal changes in health risks due to exposure to MTBE among the workers of a refinery.

Materials and methods

Sampling site

In the oil refinery complex, MTBE (Methyl tert-butyl ether) may be released from several places such as gasoline blending units, storage, catalytic reforming units, wastewater treatment plants, tanks, loading/unloading terminals, and distillation columns. Five groups of people are working in these units: site men, repair men, safety men, loading operator, and supervisors. Due to the deterioration of aging infrastructure and the ongoing wear of storage, transmission, and refining systems, workers may be exposed to MTBE through leaks and the volatilization from tanks, pipelines, and processing equipment. Job groups were defined as similar exposure groups (SEGs) based on task analysis and refinery operational data. The moderate intra-group variability (CV = 15–35%) and clear inter-group differences confirm the validity of this categorization.

Occupational exposure information

A cross-sectional study was performed on employees working in an oil refinery in Iran. Inclusion criteria for worker participation included a minimum of 12 months of career length and possible inhalation exposure to MTBE, as determined by an initial evaluation. The sample size was estimated by the Cochran equation with a confidence level of 95% (α = 0.05), which result in 30 participants. Therefore, the study performed on a total of 30 workers across all job categories (site men, repairmen, safety men, loading operators, and supervisors), not 30 individuals per group. Workers were recruited based on inclusion criteria. These inclusion criteria included minimum 12 months of career length and potential inhalation exposure to MTBE. Recruitment was conducted via random sampling from the refinery employee. Before the study, informed consent obtained from all participants. For each participant, three air samples were gathered during their work shift to calculate the time-weighted average (TWA) exposure. Before the study, all workers signed informed consent forms. Individuals also filled out questionnaires on demographic and occupational data including age, weight, work experience, number of working days per year, and daily exposure duration.

Sampling method

MTBE samples were collected according to the National Institute for Occupational Safety and Health (NIOSH) Method 1615. Sampling campaigns were carried out during winter and summer. Air samples were gathered using two adsorbent tubes containing activated coconut shell charcoal (front section: 400 mg; back section: 200 mg) manufactured by SKC Inc. The tubes were attached to workers’ collars in their breathing zone. Air was drawn through the tubes using calibrated personal sampling pumps (AirChek TOUCH, SKC Inc.) with a flow rate of approximately 100–200 mL/min. Before the main sampling, breakthrough volume tests were conducted. Breakthrough tests confirmed no analyte migration to the back section even under summer high-temperature conditions (mean 40.4 ± 1.89 °C), indicating adequate sampling efficiency. The sampling time also was approximately 20–120 min. Following sample collection, the adsorbent tubes were capped with plastic caps and placed in cool boxes to reduce the potential loss of analytes. Ambient temperature and relative humidity were also measured using a WBGT device (Tenmars Electronics Co., Taiwan). Sampling was conducted in such a way that the collected samples were representative of the overall exposure of workers during their work. For this purpose, three samples per participant were collected on the same workday during an 8-hour shift (8 to 10 am, 10 to 12 am, and 2 to 4 pm), with approximately equal sampling durations (20 to 120 min for each sample) to cover representative exposure periods and tasks. Individuals’ daily tasks were performed in the same way throughout the week. Intra-sample variation (coefficient of variation, CV) averaged 22% per participant (range: 15–35%), reflecting task-specific fluctuations in MTBE levels.

Sample Preparation and analysis

In the laboratory setting, the adsorbent tubes were opened, and the gathered MTBE was extracted by adding 2 mL of carbon disulfide (CS2) into extraction vials. Following a 60-minute desorption period, a 1 µL of the extract was introduced into a gas chromatograph (GC 7890, Agilent Technologies) equipped with a flame ionization detector (FID) and a fused silica capillary column. The carrier gas was helium with a flow rate of 1 to 2 mL/min. Nine blank samples were analyzed as well to monitor possible contamination and to verify the accuracy of the analytical procedures.

Quality assurance / quality control (QA/QC)

After sampling, both air and blank samples were promptly placed in cold boxes with temperatures of approximately 4 °C for shipment. Also, all samples were kept in refrigerators with temperature of 4 °C before analyzing. Identical analytical protocols and equipment were used for both main and blank samples. To assess potential contamination, carbon disulfide solvent blanks were analyzed using GC-FID. The limit of detection (LOD) was calculated using the equation: LOD = 3.3 × (standard deviation of blanks / slope of the calibration curve). Calibration standards for MTBE were prepared and introduced onto pre-cleaned charcoal tubes to evaluate recovery efficiency, which was found to be 94 ± 12%. Analytical results below the limit of detection (LOD = 0.05 mg/m³) were presented as “N.D.” (not detected) in raw data tables. The limit of quantification (LOQ = 0.15 mg/m³, calculated as 10 × (standard deviation of blanks / slope of calibration curve)) was considered, and only values ≥ LOQ were used for mean calculations. Non-detected data (< LOD) were treated as LOD/√2 (0.035 mg/m³) for conservative risk estimation in statistical and Monte Carlo analyses, based on the USEPA guidelines.

Health risk assessment

The quantitative risk assessment approach designed by the United States Environmental Protection Agency (USEPA) is applied to evaluate risks associated with exposure to MTBE. To determine carcinogenic risk and non-carcinogenic risk, the Lifetime Cancer Risk (LCR) and the Hazard Quotient (HQ) indices were computed, respectively.

Non-carcinogenic risk assessment

To assess the non-carcinogenic risk associated with exposure to MTBE, the methodology outlined by the United States Environmental Protection Agency (USEPA) was applied23. The non-carcinogenic risk is quantified using Hazard Quotient (HQ) values, which indicate the maximum daily exposure without adverse health effects. For the inhalation exposure route, Eq. (1) was utilized.

$$\:HQ\hspace{0.17em}=\hspace{0.17em}EC/RfC$$
(1)

In this equation, EC is exposure concentration of the investigated pollutant (mg/m3) and RfC is reference concentration (mg/m3).

To determine the EC value, Eq. 2 was applied24.

$$\:EC\:=\:(C\:\times\:\:ET\:\times\:\:ED\:\times\:\:EF)\:⁄\:AT$$
(2)

In this equation, C is the concentration of pollutant (mg/m3), ET is exposure time (hours/day), ED is exposure duration (years), EF is exposure frequency (days/year), AT is averaging time (ED in years × 365 days/year). Table 1 describes the exposure levels and toxicological parameters utilized in this study.

If HQ values are lower than 1 it indicates the lack of adverse non-carcinogenic health effects, while HQ values equal to and higher than 1 point the presence of non-carcinogenic health effects25.

Carcinogenic risk assessment

The assessment of carcinogenic risk due to exposure to the pollutant was also conducted using the USEPA methodology. Within this approach, the carcinogenic risk is quantified by the lifetime cancer risk (LCR) index. The LCR values were determined by Eq. 3.

$$\:LCR\hspace{0.17em}=\hspace{0.17em}CDI\:\times\:\:SF$$
(3)

In this equation, CDI is chronic daily intake (mg/kg-day) and SF is cancer slope factor ((mg/kg-day)−1). The SF values were obtained from the IRIS toxicological database26,27. The CDI values were calculated by Eq. 424.

$$\:CDI\:=\:\frac{C\:\times\:IR\:\times\:ET\times\:ED\:\times\:EF\:}{BW\:\times\:AT\:}$$
(4)

In this equation, C is concentration of pollutant (mg/m3), IR is inhalation rate (m3/hour), ET is exposure time (hours/day), ED is exposure duration (years), EF is exposure frequency (days/year), BW is body weight (kg), AT is averaging time (days). Averaging time (AT) must be computed based on the working lifetime rather than whole lifetime to determine health risks for exposed workers. This approach follows USEPA Risk Assessment Guidance for Superfund (RAGS) and occupational health standards. Table 1 describes the exposure levels and toxicological parameters utilized in this study.

It shows definite risk if LCR values are higher than 1 × 10− 4, it shows probable risk if those are between 1 × 10− 4 and 1 × 10− 5, it shows possible risk if those are between 1 × 10− 5 and 1 × 10− 6, and it shows negligible risk if those are lower than 1 × 10[− 6 28.

Table 1 Exposure and toxicological parameters used in this study for non-carcinogenic and carcinogenic risk assessment.
Full size table

Monte-Carlo simulations

To obtain reliable and realistic findings considering the inherent uncertainty and variability in the measurements, the use of mathematical modeling is suggested29. The Monte Carlo simulation (MCS) method, as probabilistic and statistical-mathematical techniques, utilizes repeated simulations to quantify uncertainty. This approach utilizes repeated random sampling from defined probability distributions of input variables (e.g., exposure concentration, exposure time, frequency, duration, and body weight) to quantify uncertainty and variability in risk estimates29. This approach provides a distribution of probable outcomes rather than a single point estimate, which allows percentile-based analysis (e.g., 5th, 50th, and 95th percentiles) to better represent real-world variability. In this analysis, 1,000 iterations were performed, and the outcomes were evaluated within a confidence interval ranging from 1% to 99% 30. This analysis was conducted using Crystal Ball software (version 11.1.2.4, Oracle, Inc., USA), with outcomes evaluated within a 95% confidence interval. The input parameters to Monte Carlo simulation were concentration, exposure frequency, exposure time, exposure duration, and averaging times. So, Therefore, risk assessments consider the probability distribution of different input variables to determine the overall probability distribution of the risk30. Sensitivity analysis was also conducted to identify the most influential variables on risk outputs.

Statistical analysis

Additional descriptive statistics were generated using SPSS software. To evaluate potential differences related to sampling duration, MTBE concentrations were compared between shorter (< 60 min) and longer (≥ 60 min) sampling periods. No significant differences were found (p > 0.05 in both seasons), confirming that variable sampling durations did not introduce bias.

Results

Participants

Table 2 represents demographic and occupational characteristics of the participants. The workers with various demographic and occupational characteristics participated in the study.

Table 2 Demographic and occupational characteristics of the participants (n = 30).
Full size table

Climatic conditions of the conducted measurements

In this research, air temperature, relative humidity, and wind speed were simultaneously measured during both seasons. Table 3 reports detailed climatic data measured by WBGT device. The mean and standard deviation values of air temperature, relative humidity, and wind speed were 40.4 ± 1.89 °C, 45.1 ± 5.11%, and 1.88 ± 1.04 m/s in summer (August 2024) and 15.3 ± 1.64 °C, 20.5 ± 2.12%, and 1.96 ± 0.89 m/s in winter (February 2024). Overall, these findings indicate that both temperature and relative humidity at the workplaces were considerably higher in summer compared to winter, while wind speed tended to be greater during the winter than that in summer.

Table 3 Detailed Climatic data during sampling (mean ± SD).
Full size table

MTBE concentrations and threshold limit values

The mean concentrations, non-carcinogenic risks, and carcinogenic risks associated with exposure to MTBE were evaluated in comparison to the acceptable limits across different occupational settings during summer and winter periods (as shown in Fig. 1; Table 4). According to the American Conference of Governmental Industrial Hygienists (ACGIH), the recommended threshold limit value (TLV) for MTBE exposure is 180.27 mg/m³ (50 ppm)31. Also, EPA suggests the acceptable value for non-carcinogenic health risk by 1 and for carcinogenic risk by 1 × 10− 6 28.

The findings revealed that the mean concentrations of exposure to MTBE were lower than TLV among all workers and in various jobs in both seasons. All concentrations are reported in mg/m³. Moreover, the findings indicated that the mean values of non-carcinogenic risk related to exposure to MTBE in total workers and in various jobs were lower than acceptable value in both seasons. Furthermore, the results indicated that the mean values of carcinogenic risk related to exposure to MTBE in total workers were higher than acceptable value in both seasons. These values were greater than acceptable value in jobs of loading operator, repairman, and site man in summer and in job of loading operator and repairman in winter season.

Seasonal changes in MTBE concentrations

Figure 1; Table 4 presented the concentrations of MTBE, as well as the associated non-carcinogenic and carcinogenic risks for workers in different occupations during summer and winter seasons. For all workers, these values were greater during the summer compared to those in the winter. The results revealed that concentrations, non-carcinogenic risks, and carcinogenic risks of exposure to MTBE in site man, repairman, loading operator, and safety man were significantly greater in summer season compered to winter seasons. However, these values in supervisor were significantly higher in winter season compared to summer season. The smaller seasonal difference observed for supervisors and safety men is likely due to their limited direct involvement in high-emission tasks (e.g., loading/unloading terminals or tanks), resulting in lower overall exposure proximity to MTBE sources compared to loading operators, site men, and repairmen. Moreover, it was revealed that the concentrations and health risks related to exposure in the examined occupations were ranked in descending order as follows: loading operator > site men > repairman > safety men > supervisors in summer season and loading operator > repairman > site men > supervisors > safety men in winter season.

Fig. 1
figure 1

The concentrations of MTBE (mg/m³), as well as the associated non-carcinogenic and carcinogenic risks for workers in different occupations during summer and winter seasons.

Full size image
Table 4 Statistical description of concentrations (mg/m³), non-carcinogenic risks, and carcinogenic risks of exposure to MTBE among the workers in the various jobs in summer and winter seasons.
Full size table

Findings of Monte Carlo simulation

The probability distributions and percentiles related to non-carcinogenic risk (HQ) and carcinogenic risk (LCR) of exposure to MTBE for various jobs in summer and winter were calculated using Monte Carlo simulation (Fig. 2). In Table 5 also, the statistical description of HQ and LCR values for MTBE in various jobs in summer winter seasons was presented. The results of the Monte Carlo simulation revealed that the mean values of non-carcinogenic risk and carcinogenic risk in summer season (9.99 × 10− 3 and 2.35 × 10− 6) were higher than those in winter season (2.93 × 10− 3 and 7.18 × 10− 7). Also, the mean values of non-carcinogenic risks were lower than acceptable value (1) in both seasons but the mean values of carcinogenic risks were greater than acceptable value (1 × 10− 6) in summer season.

Figure 3 shows the sensitivity analysis related to the non-cancer and cancer risk assessment of exposure to MTBE clearly demonstrating that the importance of exposure concentration (C), exposure frequency (EF) and exposure time (ET) in summer season and concentration (C), exposure time (ET) and exposure duration (ED) in winter season.

Fig. 2
figure 2

The probability distribution and percentiles associated with non-carcinogenic and carcinogenic risks resulting from exposure to MTBE during summer and winter seasons based on the Monte Carlo simulation.

Full size image
Table 5 Non-carcinogenic risk (HQ) values for exposure to MTBE across different occupations during summer and winter seasons, with the 5th and 95th percentile based on the Monte Carlo simulation.
Full size table
Fig. 3
figure 3

Sensitivity analysis related to non-cancer and cancer risk assessment of exposure to MTBE.

Full size image

Discussion

The present research aimed to evaluate the occupational health risks associated with exposure to methyl tert-butyl ether (MTBE) among workers in various job categories during summer and winter seasons, using both deterministic (HQ, LCR) and probabilistic (Monte Carlo simulation) approaches. Despite mean MTBE concentrations across all jobs were below the threshold limit value (TLV) (180.27 mg/m³), The findings point to chronic carcinogenic risk levels in specific job roles and seasonal conditions. The highest MTBE concentrations were observed among loading operators, followed by site men and repairmen that it may be because of their proximity to fuel transfer operations. There was an occupational trend persisted in both seasons, although overall concentrations were notably higher during summer because of elevated ambient temperatures increasing MTBE volatility. The smaller seasonal difference for supervisors and safety men can be explained by their primarily administrative, oversight, or patrol duties, which involve less time near high-emission sources. These findings corroborate previous research conducted by Zhang et al. They found temperature-sensitive increases in airborne MTBE concentrations in South China, especially during warm months32. In terms of non-carcinogenic risk, all computed HQ values were lower than the threshold of 1, which shows no significant short-term health risk among the studied population. However, the highest HQ values were again observed among loading operators, which reflects higher exposure burdens in this group. These results align with the findings of Tong et al.24. They reported relatively low HQ values in automobile manufacturing despite detectable concentrations of volatile organic compounds (VOCs)24. In terms of carcinogenic risk, the mean LCR for most workers in summer (2.46 × 10⁻⁵) exceeded the acceptable risk limit (1 × 10⁻⁶), which indicates a probable carcinogenic risk. In contrast, the winter LCR value (5.45 × 10⁻⁶) remained closer to or below acceptable thresholds in most jobs except loading operators and repairmen. Moreover, these results reflect the findings reported by Sadeghi-Yarandi et al. (2020). They found elevated LCR values in petrochemical workers exposed to 1,3-butadiene in Iran, which suggests that storage and transfer operations in petrochemical environments are particularly susceptible to cumulative VOC exposure and associated long-term risks25. These results show the fact that TLVs are primarily designed for prevention of acute toxic effects, and those may not be sufficiently protective against cumulative carcinogenic outcomes, especially under chronic low-dose exposure scenarios. This observation is consistent with critiques in the literature, which emphasize the need to supplement TLVs with quantitative risk assessment models for chronic endpoints33. Monte Carlo simulations provided further robustness to the deterministic findings, simulating 1,000 iterations to model the uncertainty and variability in exposure parameters. The probabilistic mean values for HQ and LCR in summer were higher than those in winter, which reinforces the impact of seasonal dynamics on exposure levels. More importantly, the 95th percentile (P95) LCR in summer (9.18 × 10⁻⁶) substantially exceeded the acceptable benchmark, which suggested that a significant proportion of the population may be at elevated risk. Sensitivity analysis in the Monte Carlo model identified pollutant concentration (C), exposure time (ET), and exposure frequency (EF) as the most influential variables during the summer, and C, ET, and exposure duration (ED) as most important variables during winter. These results are consistent with the probabilistic risk models used by Badeenezhad et al. They reported similar patterns for exposure assessments of nitrate in groundwater34. These findings highlight critical points of intervention. For instance, reduced exposure duration or improved process controls during high-exposure tasks can directly influence overall risk. Zhang et al. investigated VOC emissions in petroleum refineries in China. The present study revealed similar seasonal trends and job-specific exposure hierarchies to this research, with the highest risks typically associated with direct-handling roles such as loading operators. However, Zhang et al. noted slightly higher baseline concentrations of MTBE, which may reflect differences in regulatory controls, fuel formulation, or ventilation infrastructure28. Conversely, Tong et al.24 reported negligible cancer risks, likely due to shorter exposure durations and better-engineered controls in their setting24. Unlike studies that focus solely on environmental exposures, the occupational environment examined in the present study provides insights into high-risk subpopulations that may not be adequately protected by general environmental guidelines. This distinction reinforces the importance of specific health risk assessments, especially in petrochemical and fuel-handling industries. However, these findings are associated with several limitations. Firstly, the exposure assessment in this study focused exclusively on the inhalation route. Future studies should expand the exposure model to incorporate all plausible exposure pathways. Secondly, the study relied on estimated or self-reported exposure variables (e.g., exposure time, frequency, and duration), which may be subject to recall bias or inaccuracies due to inconsistent work routines. These limitations could affect the accuracy of the deterministic and probabilistic outputs. Moreover, the study’s temporal design comparing only two seasons limits the generalizability of seasonal trends. Future studies should consider the full spectrum of seasonal exposure variability. Given these limitations, several recommendations can be made for future research. First, longitudinal studies with repeated measurements over time are needed to examine the cumulative health impacts of chronic low-level MTBE exposure and validate risk assessment outcomes with actual health endpoints. Second, interventional studies evaluating the effectiveness of engineering controls (e.g., vapor recovery systems), administrative measures (e.g., task rotation), and personal protective equipment (PPE) in high-risk jobs can provide actionable evidence for occupational health management. Additionally, probabilistic models such as Monte Carlo simulations should be enhanced with greater iteration depths, larger datasets, and scenario-specific inputs. Integration of multi-exposure chemical risk assessment frameworks would also allow us to evaluate combined exposures in real-world petrochemical work environments. Moreover, there is a clear need to harmonize occupational safety standards with modern risk-based approaches that consider long-term cumulative exposure and population susceptibility.

Conclusion

The present study highlights a clear occupational health concern regarding chronic exposure to MTBE, especially in high-risk jobs such as loading operators and repairmen during the summer season, where carcinogenic risk levels exceeded international safety limits. Although non-carcinogenic risks remained within acceptable limits, the increased long-term cancer risk even at concentrations below the TLV emphasizes the need for re-evaluation of current exposure standards in fuel-handling industries. These findings support the implementation of engineering controls such as vapor recovery systems and optimized ventilation in transfer zones and seasonal adjustments in workload or shift duration. From a risk management perspective also, exposure monitoring and preventive measures should be intensified during warmer months. Industries can utilize this risk profile to refine occupational safety protocols and reduce cumulative exposure.