Non-linear associations of blood volatile organic compounds (VOCs) with sleep health and mortality in American adults: data from the National health and nutrition examination survey (NHANES) 2007-2012.
Chen M, Wang Y, et al. • International archives of occupational and environmental health • 2026
VOC exposure may be linked to poor sleep patterns among U.S. adults, with non-linear associations observed between certain blood VOCs and sleep health indicators, and methyl tert-butyl ether identified as a risk factor for all-cause mortality.
Key Findings
Results
No significant linear correlations were found between blood VOC levels and sleep duration or poor sleep patterns.
Weighted linear and logistic regression models showed no significant linear associations between the five VOCs and sleep outcomes.
The five VOCs studied were bromoform (LBXVBF), bromodichloromethane (LBXVBM), chloroform (LBXVCF), dibromochloromethane (LBXVCM), and methyl tert-butyl ether (LBXVME).
Sleep health indicators assessed included trouble sleeping, sleep disorders, and insufficient (<6 h/day) or excessive (>9 h/day) sleep.
Data were drawn from NHANES 2007-2012 participants.
Results
Restricted cubic spline (RCS) regression models identified significant non-linear relationships between certain VOCs and sleep health outcomes.
P for non-linearity was less than 0.05 for certain VOC-sleep health associations.
The RCS model was used to detect non-linear dose-response relationships that linear models would miss.
This finding contrasts with the null results from standard linear/logistic regression models.
Results
Bromoform (LBXVBF) was identified as the most important contributor to poor sleep quality in mixture analyses.
Adjusted quantile-based g-calculation (QGC) analysis highlighted LBXVBF as a crucial factor related to poor sleep quality.
LBXVBF received a weight of 0.733 in the QGC analysis, indicating it was the dominant compound driving the mixture effect.
Weighted quantile sum (WQS) and Bayesian kernel machine regression (BKMR) models were also employed to assess VOC mixture effects.
Results
BKMR analysis showed a positive trend between VOC mixture levels and poor sleep patterns at the 55th to 75th percentiles of exposure.
The positive trend was observed specifically in the 55th to 75th percentile range of the overall VOC mixture.
BKMR models account for potential interactions and non-linearities among multiple exposures simultaneously.
The association suggested that moderate-to-higher VOC exposure levels were linked to worse sleep patterns compared to lower levels.
Results
Methyl tert-butyl ether (LBXVME) was identified as a risk factor for all-cause mortality with a significant non-linear association.
Cox proportional hazards regression combined with restricted cubic spline (COX-RCS) analysis was used for survival analysis.
The non-linear association between LBXVME and all-cause mortality was statistically significant (P for non-linear = 0.0359).
LBXVME was the only VOC among the five examined to show a significant non-linear relationship with mortality.
Results
Sex-based differences in sleep problems were observed at baseline, with women having higher prevalence of trouble sleeping and men more prone to insufficient sleep.
The baseline profile was categorized by sex.
Women showed higher prevalence of trouble sleeping compared to men.
Men were more prone to insufficient sleep (defined as less than 6 hours per day) compared to women.
Methods
The study used multiple analytical methods including GL, RCS, WQS, QGC, and BKMR models to assess VOC-sleep relationships in NHANES 2007-2012 participants.
Generalized linear (GL), restricted cubic spline (RCS), weighted quantile sum (WQS), quantile-based g-calculation (QGC), and Bayesian kernel machine regression (BKMR) models were all employed.
Cox proportional hazards regression was additionally used for survival (mortality) analysis.
The five trihalomethane-related VOCs and methyl tert-butyl ether measured in blood were the exposure variables.
Data source was NHANES 2007-2012, a nationally representative U.S. survey.
What This Means
This research suggests that exposure to certain volatile organic compounds (VOCs) — chemicals found in the environment, drinking water, and various industrial and consumer products — may be associated with poor sleep health among U.S. adults. The study analyzed blood levels of five VOCs in thousands of Americans who participated in a large national health survey (NHANES) between 2007 and 2012, linking those levels to self-reported sleep problems and death records. Importantly, the relationships between VOC exposure and sleep problems were not straightforward or linear; instead, they followed curved patterns that only became apparent when using more sophisticated statistical methods.
Among the five chemicals studied, bromoform stood out as the most strongly associated with poor sleep quality when the effects of the chemical mixture were analyzed together. Additionally, methyl tert-butyl ether (MTBE), a chemical historically used as a gasoline additive, showed a significant non-linear association with risk of death from any cause. The study also found that women were more likely to report trouble sleeping, while men were more prone to getting less than six hours of sleep per night.
This research suggests that routine exposure to VOCs, even at levels found in the general population, may have consequences for sleep health and potentially long-term survival. Because sleep is critical to overall health, these findings highlight the importance of monitoring and potentially reducing population-level VOC exposures. However, because this was an observational study, it cannot definitively prove that VOCs cause sleep problems — only that the two are statistically associated. Further research is needed to understand the biological mechanisms and to establish whether reducing VOC exposure could improve sleep outcomes.
Chen M, Wang Y, He Y, Pang Y, Tang L. (2026). Non-linear associations of blood volatile organic compounds (VOCs) with sleep health and mortality in American adults: data from the National health and nutrition examination survey (NHANES) 2007-2012.. International archives of occupational and environmental health. https://doi.org/10.1007/s00420-025-02194-z