The petrochemical industry involves the processing and transformation of petroleum-derived chemical raw materials, where high temperatures are essential for operations like catalytic cracking and hydrocracking (1, 2). Thus, occupational heat exposure represents a significant workplace hazard. In recent years, climate change-induced extreme heat events have further exacerbated the burden of workplace heat exposure (3), which may significantly increase health risks among petrochemical workers (4, 5). The kidneys, which play a vital role in maintaining fluid and electrolyte homeostasis and in excreting metabolic waste products, are particularly vulnerable to heat stress (6). Under heat stress, peripheral vasodilation and sweating lead to substantial fluid loss, resulting in dehydration and reduced urinary excretion, which in turn elevates serum uric acid levels. Moreover, heat stress induces systemic oxidative stress and inflammatory responses that further impair renal function (7), thereby contributing to the development of both acute kidney injury (AKI) and chronic kidney disease (CKD) (8).

An increasing body of evidence has linked occupational heat exposure to renal dysfunction (9–11). Archived case reports from the US Occupational Safety and Health Administration (OSHA) indicate that heat-related acute kidney injury (HR-AKI) occurs across diverse occupational settings, both indoors and outdoors (12). However, research focusing on the renal health effects of heat exposure among petrochemical workers remains limited. Our preliminary analysis identified hyperuricemia as the most common abnormality related to impaired kidney function among petrochemical workers. As the end product of purine metabolism, uric acid contributes to intracellular oxidation, endothelial dysfunction, renal fibrosis, and glomerulosclerosis (13, 14). Furthermore, hyperuricemia is an established independent risk factor for the development of incident CKD and rapid renal function decline, as well as a biomarker for early-stage renal dysfunction (15, 16), being closely associated with declining eGFR, albuminuria, and kidney failure (17). Evidence also suggests that treating hyperuricemia in its early stage may delay or even prevent the onset of CKD (18). Therefore, hyperuricemia was selected as the primary indicator of early renal dysfunction in this study. Existing epidemiological studies have reported a higher prevalence of chronic kidney disease (17.9%) among residents near refineries compared to the general population (12.3%) (19), along with elevated uric acid levels in oilfield workers and local wildlife (20, 21). These findings highlight the need to investigate the association between occupational heat exposure and hyperuricemia among petrochemical workers to better inform targeted heat-related health protection strategies.

Occupational exposures in the petrochemical industry often involve co-exposure to multiple hazards, such as benzene, H₂S and other chemicals, which may interact synergistically and amplify health risks. Existing evidence suggests that the combination of thermal and chemical exposures in petrochemical operations contributes to increased vulnerability to a range of diseases (22, 23). Traditional single-exposure statistical models are often inadequate for capturing the complex interactions of multiple exposures. By contrast, WQS regression can generate exposure indices that identify the most influential risk factors while accounting for interaction effects among exposures (24). Furthermore, the integration of machine learning with SHAP values offers robust methodologies for investigating nonlinear and high-dimensional relationships inherent in chemical and environmental mixtures (25, 26). Therefore, by combining multiple statistical models and machine learning algorithms, this study may yield valuable insights for developing prevention strategies and promoting the occupational health of petrochemical workers. In this study, we analyzed the association between occupational heat exposure and hyperuricemia among petrochemical workers using multiple statistical models and machine learning methodologies, with aims to provide evidence-based guidance for occupational health management and mitigate the burden of heat-related renal impairment in high-risk industrial settings.

This study examined the association between occupational heat exposure and hyperuricemia among petrochemical workers. Our findings suggested that workplace heat exposure was associated with an increased risk of hyperuricemia. Using WQS regressions, we identified heat exposure as the most influential occupational hazard contributing to hyperuricemia risk. Additionally, a machine learning-based prediction model demonstrated high accuracy in forecasting hyperuricemia risk, with SHAP analysis confirmed the strong contribution of heat exposure.

Exposure to high temperatures has emerged as a significant public health and occupational health concern, particularly due to its association with renal impairment. Prolonged exposure to process-generated heat exacerbates thermal strain and predisposes workers to dehydration and kidney injury (35). Heat stress is a potential contributing factor to chronic kidney disease of unknown etiology (CKDu) in agricultural communities, which is increasingly recognized as a major cause of chronic kidney disease (36). This CKDu, also observed in Central America, often referred to as Mesoamerican nephropathy, has been epidemiologically and experimentally linked to recurrent dehydration and chronic heat exposure (37). Mounting evidence suggests that this condition may represent a distinct form of heat stress nephropathy (HSN), potentially exacerbated by global warming (38, 39). Furthermore, a myriad of chronic kidney disease cases has surfaced among agricultural workers and other individuals engaged in manual labor across various regions of the world (40). The disease is strongly correlated with the conditions of working and living in hot environments (41). Therefore, focusing on early indicators of renal impairment, such as hyperuricemia, may provide a valuable approach for early detection and prevention of heat-related kidney injury.

Petrochemical workers were routinely exposed to multiple occupational hazards such as heat, noise, and chemical toxins during the production and processing stages. Those occupational hazards may impact workers’ health (42). Salem et al. has reported elevated levels of urea, creatinine, and uric acid among Libyan petrochemical workers, suggesting subclinical or overt kidney dysfunction (43). Consistent with these findings, our results also revealed a high prevalence of hyperuricemia among petrochemical workers, which were significantly associated with heat exposure. Several factors may contribute to this high hyperuricemia prevalence. First, the study population consists of frontline occupational workers (male-dominated), who may inherently represent a higher-risk group and generally have higher uric acid levels. Second, the petrochemical industrial park is located in a coastal region, where dietary patterns often involve high consumption of seafoods (e.g., fish, shrimp, crab, and shellfish), leading to increased purine intake and elevated serum uric acid levels. Uric acid, the final product of purine metabolism in the body, is primarily excreted by the kidneys. Under normal physiological conditions, serum uric acid levels are maintained within a certain range. However, heat exposure may disrupt this balance through dehydration, reduced renal perfusion, and impaired excretory function (44), thereby increasing the risk of hyperuricemia. In workplaces that strictly follow occupational production norms and supervision of safety departments, the likelihood of petrochemical workers experiencing rapid and severe renal impairment is relatively low, while hyperuricemia is one of the early manifestations of renal dysfunction (45). Accordingly, attention to changes in uric acid levels and the onset of hyperuricemia may hold greater practical significance for safeguarding kidney health within this occupational population.

In this study, results of multistage GLM regression suggested a significant association between heat exposure and hyperuricemia, consistent with findings from salt pan workers, sugarcane field workers, and construction workers (46, 47). Despite differences in climate and work processes, heat exposure consistently elevates uric acid levels through biological mechanisms involving inflammation, oxidative stress, altered metabolism, and cytokine activation (48–50). Experimental evidence also supports this link, as animal studies have demonstrated that chronic heat exposure induces elevated serum uric acid (51). Interestingly, workers with hyperuricemia in our study tended to be younger and had shorter lengths of service. This unexpected finding may be attributed to a healthy worker survivor bias, whereby individuals unable to adapt to hazard exposures often leave high-exposure jobs earlier, and hiring restrictions might create a workforce with inherently higher health baseline levels (52, 53). Additionally, younger workers are frequently assigned to frontline positions, leading to higher exposure intensity despite shorter tenures (54), and longer-tenured workers may possess greater self-protective awareness due to extended training and job rotations, potentially reducing their risk of hyperuricemia.

In assessing joint effects of combined exposure to heat and other occupational hazards on hyperuricemia, significant multiplicative interaction effects were observed. Interestingly, the combined exposure to heat and gasoline or H₂S, which both associated with higher hyperuricemia risk, despite their opposing individual effects. However, positive additive interactions were not observed, this finding consistent with Chen Y et al. (29), their research show that a multiplicative interaction rather than additive interaction between heat exposure and dust exposure in the development of hyperuricemia in steelworkers. A possible explanation of our study may lie in interaction mechanisms that modify physiological stress responses. While gasoline exposure alone may appear inversely associated with hyperuricemia, potentially due to residual confounding by work location, better ventilation, or health-based job selection, its co-occurrence with heat exposure may intensify oxidative and metabolic stress. Heat exposure promotes dehydration, reduced renal perfusion, and accumulation of reactive oxygen species (ROS) (55), which can impair uric acid excretion. Combined exposure to gasoline, containing volatile organic compounds (VOCs), may further exacerbate oxidative damage and inflammatory pathways, thereby amplifying renal tubular dysfunction (56, 57). This synergistic effect could result in a multiplicative increase in hyperuricemia risk, even when gasoline alone appears protective. Additionally, workers in high-temperature units often handle multiple chemical processes simultaneously, leading to greater overall exposure intensity and cumulative metabolic burden, which may also contribute to this combined effect. Krishnamurthy et al. found the potential synergistic effect between oil exposure and heat (58). The similar joint effect was observed for the combined exposure to heat and H₂S. This apparent paradox may reflect complex biological interactions and workplace exposure patterns. At low concentrations, H₂S can exert transient antioxidant effects by reduce the production of ROS (59). However, under conditions of heat stress, dehydration, and hypoxia, H₂S metabolism may shift toward pro-oxidative and cytotoxic pathways, enhancing oxidative stress and renal tubular injury (60). Furthermore, workers exposed to both heat and H₂S are typically engaged in frontline production units with higher overall exposure intensity and workload, which could amplify metabolic strain and renal burden (61). Therefore, combined exposure to heat and additional occupational hazards may further increase hyperuricemia risks among workers.

Traditional single-variable regression approaches often fall short in capturing the combined effects of multifactorial exposures. In contrast, the WQS model allows for estimation of cumulative effects while weighing each exposure’s relative contribution (62). Previous studies have used the WQS model to analyze the effects of mixed factors on hyperuricemia and to identify the most influential contributing factors (63). In this study, our WQS analysis identified heat and methanol exposure as the dominant contributors to hyperuricemia risk. The WQS model provides a better understanding of the relative impact of individual hazards on health outcomes and offers more precise strategies for pollutant control and management. Given the complexity of occupational exposures in petrochemical environments, predictive modeling plays an increasingly important role in occupational health surveillance. Machine learning models, particularly those that capture non-linear relationships and interactions, offer substantial advantages over traditional statistical approaches (64–66). Anttila et al. has already explored the risk prediction modeling in petrochemical workers, with findings indicating that exposure to hydrocarbons in crude oil may increase the risk of kidney cancer by up to threefold (67). In this study, we comparison nine supervised learning models to predict hyperuricemia risk. After cross-validation and performance evaluation, the LightGBM model was selected for its superior balance of accuracy and generalizability.

The LightGBM model demonstrated strong predictive performance, but the AUC results suggest potential overfitting. In practical terms, this predictive model may be incorporated into occupational health surveillance systems to support dynamic risk stratification and provide early indications of workers at higher risk of hyperuricemia. Such integration could help inform timely preventive measures. Nevertheless, further external validation and real-world implementation studies are warranted before real-world application. While machine learning algorithms offer strong predictive performance, they are often criticized for their lack of transparency (68). To address this, we incorporated SHAP analysis, which has been widely adopted for interpreting machine learning models. By calculating the marginal contribution of each feature to the prediction outcome, SHAP effectively quantified the impact of individual variables (69). SHAP results confirmed that BMI and heat exposure are dominant risk factors, offering actionable insights for workplace surveillance and intervention prioritization.

Although this study identified an association between occupational heat exposure and hyperuricemia among petrochemical workers, several limitations should be acknowledged. First, this research was derived from retrospective health examination records, and we adopted a complete-case analysis approach for all main exposure and outcome variables, which may reduce sample size. The inability to infer causality is also inherently limited. Second, our study did not capture information on dietary habits, hydration patterns, genetic predisposition, medication use, metabolic factors, use of personal protective equipment, medication use, and acclimatization, which may influence the assessment of heat-hyperuricemia association. Third, heat exposure in this study primarily refers to process-generated heat from petrochemical production. Workers’ occupational health examinations were conducted during the warmer months (March to August) each year, which may potentially overestimate the association between heat exposure and hyperuricemia. Nevertheless, the uniform health examination period across all workers helps to minimize between-individual seasonal variability and thus strengthens internal comparability. However, it should be noted that the lack of year-round health examination data may limit our ability to fully characterize seasonal variations in the heat–hyperuricemia relationship. Fourth, during the additive interaction analysis in this study, some Synergy Index (SI) values appeared as NaN, likely because the odds ratio (OR) for a single exposure equaled 1, resulting in a zero denominator in the SI formula. To obtain valid estimates, we followed the software recommendation and set recode = TRUE to reverse-code the exposures. This adjustment ensured computational validity but may have altered the directionality of the additive interaction estimates (negative). Therefore, the opposite directions observed between additive and multiplicative interactions in this study should be interpreted with caution, as they may partly reflect a computational artifact rather than a genuine antagonistic interaction on the additive scale. Finally, the absence of creatinine/eGFR, albuminuria, temperature monitoring during the production process as well as relevant information regarding the workers’ production departments, has precluded a more refined assessment of the association between heat exposure (in terms of intensity and duration) and additional biomarkers of early renal dysfunction.

To address these limitations, our ongoing prospective cohort study has been designed to incorporate structured occupational and dietary questionnaires to systematically collect data on exposure histories and lifestyle factors, as well as surveying nearby residents as a control group to better evaluate the health effects of occupational exposures. Furthermore, we are also actively collaborating with government monitoring agencies to obtain accurate measurements of exposure levels and duration for various occupational hazards, and model dose–response relationships using methods such as restricted cubic splines. These efforts aim to enhance the robustness of future analyses and to inform evidence-based strategies for protecting the health of petrochemical workers.

Occupational heat exposure is a prevalent hazard in the petrochemical industry and is significantly associated with hyperuricemia. By integrating epidemiological and machine learning methods, this study quantified both the independent and combined effects of heat and other hazards on kidney health and identified key predictors of hyperuricemia risk. Findings of this study highlight the need for targeted preventive measures, including heat stress reduction and proactive monitoring high-risk workers. Future research should explore the biological mechanisms underlying these associations and externally validate predictive models in diverse occupational settings to better prevent heat-related early renal dysfunction.