Related Experiment Video
Updated: Oct 3, 2026
Vinyl Chloride and High-Fat Diet as a Model of Environment and Obesity Interaction
Published on: January 12, 2020
Association between butadiene exposure and leukemia mortality in synthetic rubber industry workers
Lynne Pavlic Marshall1, Christine R Langton1, Lindsey A Wood1
1Valeo Sciences LLC, Ladera Ranch, CA, USA.
Objectives:
Regulatory agencies across the USA and EU utilize analyses of a cohort of workers in the styrene-butadiene rubber (SBR) industry to assess excess cancer risk and make policy decisions on occupational and environmental exposure limits for 1,3-butadiene (BD). The objective of this study was to perform an analysis on the SBR data, including explorations of covariates and worker inclusion criteria, to evaluate the effects on the model parameters for estimating leukemia mortality.
Material And Methods:
Exposure and mortality information for 21 069 male and female workers in 6 North American SBR plants were obtained. Cox proportional hazards models were used to evaluate the association between BD and leukemia mortality. All Cox models used number of days from hire date to person-day observation as the time scale and included cumulative BD as a time-varying exposure.
Results:
Most workers were exposed to BD (66.4%). There were 132 deaths due to leukemia. The Cox proportional hazards model excluding all workers with <1 year of employment (N = 1801) and adjusted for age, BD high intensity tasks, and year of hire resulted in a statistically significant association between cumulative BD exposure and leukemia mortality (β = 2.14 × 10-4, 95% CI: 0.40-3.87 × 10-4; HR = 1.000214, 95% CI: 1.000040-1.000387). The same model including all workers in the dataset had similar results (β = 2.21 × 10-4, 95% CI: 0.49-3.93 × 10-4; HR = 1.000221, 95% CI: 1.000049-1.000393).
Conclusions:
The authors' exploration of covariates and worker inclusion criteria to evaluate the quantitative effects on the relationship between BD and leukemia mortality demonstrates little variation in model parameters despite covariates and worker groups included. Using a conservative 95% upper bound of the β-coefficient to estimate excess risk resulted in a greater BD exposure estimate compared to regulatory bodies, suggesting a weaker overall exposure risk relationship. Int J Occup Med Environ Health. 2026;39(4):378-89.
