Exposure to volatile hydrocarbons and neurologic function among oil spill workers up to 6 years after the Deepwater Horizon disaster

Abstract

Background:

During the 2010 Deepwater Horizon (DWH) disaster, oil spill response and cleanup (OSRC) workers were exposed to toxic volatile components of crude oil. Few studies have examined exposure to individual volatile hydrocarbon chemicals below occupational exposure limits in relation to neurologic function among OSRC workers.

Objectives:

To investigate the association of several spill-related chemicals (benzene, toluene, ethylbenzene, xylene, n-hexane, i.e., BTEX-H) and total petroleum hydrocarbons (THC) with neurologic function among DWH spill workers enrolled in the Gulf Long-term Follow-up Study.

Methods:

Cumulative exposure to THC and BTEX-H across the oil spill cleanup period were estimated using a job-exposure matrix that linked air measurement data to detailed self-reported DWH OSRC work histories. We ascertained quantitative neurologic function data via a comprehensive test battery at a clinical examination that occurred 4–6 years after the DWH disaster. We used multivariable linear regression and modified Poisson regression to evaluate relationships of exposures (quartiles (Q)) with 4 neurologic function measures. We examined modification of the associations by age at enrollment (<50 vs. ≥50 years).

Results:

We did not find evidence of adverse neurologic effects from crude oil exposures among the overall study population. However, among workers ≥50 years of age, several individual chemical exposures were associated with poorer vibrotactile acuity of the great toe, with statistically significant effects observed in Q3 or Q4 of exposures (range of log mean difference in Q4 across exposures: 0.13–0.26 micron). We also observed suggestive adverse associations among those ≥ age 50 years for tests of postural stability and single-leg stance, although most effect estimates did not reach thresholds of statistical significance (p<0.05).

Conclusions:

Higher exposures to volatile components of crude oil were associated with modest deficits in neurologic function among OSRC workers who were age 50 years or older at study enrollment.

Introduction

The Deepwater Horizon (DWH) disaster is the largest marine oil spill in U.S. history (National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling 2011). The explosion of the drilling rig on April 20, 2010 resulted in the release of an estimated 4.9 million barrels of crude oil into the Gulf of Mexico (National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling 2011). To stop the spill and remove the oil from the environment, tens of thousands of workers and volunteers participated in an oil spill response and cleanup (OSRC) operation, which was largely completed by June 2011 (U.S. Coast Guard 2011). While working on response and cleanup, workers were exposed to a range of chemical hazards, including crude oil hydrocarbons and particulate matter from combustion of oil/gas and diesel-powered vessel engines (Kwok et al. 2017).

Several volatile crude oil chemicals, including benzene, toluene, ethylbenzene, xylene, and n-hexane (BTEX-H), are known neurotoxicants (ATSDR 2004). In addition to being components of the crude oil, they are also present in gasolines and used as solvents and industrial raw materials (ATSDR 1999, 2000, 2007a, b, 2010). In animal models and human studies, acute inhalation of high levels of these chemicals has been associated with impaired neurologic function, as demonstrated by overt symptoms of neurotoxicity, changes in electrophysiological measures of peripheral nerve performance, and reduced neurobehavioral test performance (ATSDR 1999, 2000, 2007a, b, 2010). Results of these and other studies informed the development of health-based occupational standards (e.g., permissible exposure limits (OSHA 2022)) and guidelines (threshold limit values) (ACGIH 2012) for short-term exposure to BTEX-H.

Neurological effects of longer-term, lower-level exposure to volatile hydrocarbons have been investigated in occupational studies. Most studies assessed exposure by comparing workers chronically exposed to organic solvents or petroleum fuels with a group with lower or no exposure (Gong et al. 2003; Jovanović et al. 2004; Mutti et al. 1982b; Shih et al. 2011; Stĕtkárová et al. 1993). In these studies, exposure to volatile hydrocarbon mixtures predominated by one or more BTEX-H chemicals at levels below the occupational limit were associated with reduced cutaneous sensitivity to vibration and reduced motor/sensory conduction velocities (Sanagi et al. 1980; Shih et al. 2011), postural instability (Kuo et al. 1996; Sanagi et al. 1980), deficits in visual contrast sensitivity (Gong et al. 2003), and color vision impairment (Cavalleri et al. 2000), with two studies showing effects that persisted many months after exposure (Cianchetti et al. 1976; Mutti et al. 1982b). It is unclear, however, which agent among these mixture components contribute to the neurologic effects. Quantitative exposure estimates to individual BTEX-H chemicals ascertained from air samples or from urinary biomarkers were evaluated in relation to neurologic function in a few studies, with inconsistent results (Bates et al. 2019; Bleecker et al. 1991; Gong et al. 2003; Jovanović et al. 2004; Lee et al. 2007; Mutti et al. 1982b; Neghab et al. 2012; Smith et al. 1997). In addition, most of these studies were small (n<100) or had extensive exclusion criteria, which limited the generalizability of the findings.

The neurologic effects of OSRC work have been explored in previous studies of oil spill workers. Higher prevalence of acute (Ha et al. 2012) and persistent neurological symptoms (Na et al. 2012) was associated with longer durations of exposure among workers responding to the Hebei Spirit oil spill. Following the DWH disaster, self-reported exposures to crude oil were associated with acute neurologic symptoms (Krishnamurthy et al. 2019) and longer-term risk of neurological outcomes among Coast Guard responders who participated in the cleanup (Denic-Roberts et al. 2023). Modestly worse performance in some neurobehavioral tests was observed among DWH OSRC workers with higher single daily maximum exposure to total petroleum hydrocarbons (THC), a composite measure of the volatile components of the crude oil, several years after the spill (Quist et al. 2019). These studies provide some evidence that medium-term or chronic exposure to lower levels of volatile crude oil chemicals may induce adverse, possibly persistent, neurologic effects. The objective of the present study was to evaluate associations of quantitative estimates of crude oil-related volatile hydrocarbon exposures (THC and BTEX-H) with tests of neurologic function among DWH oil spill workers 4–6 years after the spill.

Methods

Study population

The Gulf Long-Term Follow-up Study (GuLF Study) is a prospective cohort study of the potential health effects of the DWH disaster. Study participants included individuals ≥21 years of age who either had participated in OSRC work for at least one day (workers) or had completed safety training but were not hired (non-workers) (Kwok et al. 2017). Enrollment occurred from March 2011 to May 2013 via computer-assisted telephone interviews that collected information on socio-demographics, lifestyle, health, and a detailed work history of DWH OSRC activities (Kwok et al. 2017). A total of 32,608 participants were enrolled in the study. Between August 2014 and July 2016, cohort members who resided within 60 miles of study clinics in New Orleans, LA or Mobile, AL were invited to a follow-up clinical examination during which their neurologic function was evaluated, as described below (Kwok et al. 2017).

Among the 3,401 participants who completed the clinical examination, we restricted our analysis to the 2,873 workers. Oil spill chemical exposures were not assessed for non-workers, as they did not have an opportunity to be occupationally exposed to the crude oil chemicals. We included in our sample the workers who completed at least one test of neurologic function (N=2,868). We excluded workers who reported a peripheral neuropathy (N=32) or diabetes diagnosis (n=142) before the DWH disaster. Finally, we excluded 84 workers with missing covariates needed for analysis and reached a final analytical sample of 2,610 participants. All participants provided written informed consent. The study was approved by the Institutional Review Board of the National Institute of Environmental Health Sciences.

Exposure assessment

Cumulative exposure to THC and five volatile hydrocarbon chemicals (BTEX-H) across each participant’s work period were estimated via a job-exposure matrix that linked air measurement data to OSRC work histories. The primary source of measurement data came from approximately 28,000 full-shift personal air samples collected on a subset of DWH OSRC workers, including some who enrolled in the GuLF Study, during their work shifts from April 2010 to June 2011 using organic vapor passive dosimeters (Stewart et al. 2022). Laboratory assays of these samples resulted in >143,000 measurements of THC and BTEX-H (Stenzel et al. 2022a; Stewart et al. 2022). In addition to these personal measurements, direct-reading volatile organic compound (VOC) area measurements were collected on 38 large vessels. Study industrial hygienists converted these VOC measurements to equivalent full-shift THC and BTEX-H estimates to supplement measurements on days where there were few or no personal measurements on these vessels (Groth et al. 2022a; Ramachandran et al. 2022).

To estimate exposures for each study participant, the industrial hygienists created over 3,000 exposure groups (EGs) based on three exposure determinants: job/activity (e.g., driller, burning oil, skimming), location of work (hot zone, source, offshore, nearshore, and land, separately by Gulf state), and time period (periods of time during the cleanup reflecting changes in oil weathering and OSRC events) (Stenzel et al. 2022b). Each EG was a unique combination of these three determinants and represented workers who were expected to have similar distributions of exposures. The average of the air measurements corresponding to each EG was estimated for the EG (Groth et al. 2017; Groth et al. 2022a; Groth et al. 2022b; Huynh et al. 2022a, b; Huynh et al. 2022c; Ramachandran et al. 2022). Because some EGs contained a large number of measurements below the analytical method’s limit of detection, a left-censored Bayesian framework was used to estimate THC/BTEX-H exposure averages and other exposure statistics for the EGs (Groth et al. 2017; Groth et al. 2022b). More details on the statistical methods used in the exposure assessment have been described in the papers cited above.

Based on their reported DWH OSRC work history, study participants were matched to the appropriate EGs for assigning exposure estimates. Because many participants reported multiple work activities across the cleanup, some of which occurred on the same day, daily average exposure estimates were developed for each day worked by taking the average of the exposure estimates across all jobs/activities performed on a day. To examine the total burden of exposure experienced by each participant during the cleanup, cumulative exposure across all workdays was calculated by summing up the daily average exposure estimates across all of each participant’s workdays within a time period and then across all time periods.

Neurologic Function Testing

The neurologic function testing battery included tests of vibrotactile threshold, visual acuity, visual contrast sensitivity, postural sway, and ability to maintain single leg stance. These tests were administered at the clinical examination by trained clinical examiners who were blinded to each participant’s exposure levels.

The Vibratron II electromechanical vibrometer (Physitemp, Inc.) was used to measure bilateral great toe sensory acuity at a frequency of 120Hz using a standard protocol (Gerr et al. 1990). Examiners manually controlled the delivered vibration amplitude. A total of five vibration threshold values (three descending and two ascending values) were obtained for each great toe. After discarding the first value, the final vibrotactile threshold for each toe was the median of the remaining four values. We report vibrotactile threshold in log microns of peak-to-peak amplitude and examined the single bilateral mean (i.e., average of the two toes) in analyses. Higher values of vibrotactile threshold indicate worse performance.

Visual acuity was assessed binocularly using the Optec 1000 (Optec, Inc.) to determine eligibility for the statistical analysis of visual contrast sensitivity (Campagna et al. 1995). Visual contrast sensitivity was evaluated with the Optec 1000 Functional Assessment of Contrast Sensitivity test. In the test, circular stimuli consisting of alternating light and dark bars were presented to the participants. Nine stimuli of decreasing contrast were presented at each of 5 spatial frequencies (1.5, 3, 6, 12, and 18 cycles per degree). At each spatial frequency, the weakest contrast score correctly identified was selected and examined in analyses. Higher scores of contrast sensitivity indicate better performance.

Postural stability was evaluated with an Advanced Mechanical Technology, Inc. force platform (AMTI) (Letz et al. 1994). Participants were instructed to stand upright on the platform, and four trials were administered, two with eyes open and two with eyes closed. The mean sway speed (millimeters per second) was calculated from force signals recorded by the platform during the trials. We examined the average sway speeds across the eyes-open trials and, separately, the average value across the eyes-closed trials in analyses. Higher postural sway speed indicates less stability.

Trained staff evaluated participants’ ability to stand on one leg and maintain upright balance for 30 seconds (s) (Tyson and Connell 2009). If the participant was unable to maintain balance for the entire 30-s interval, then the procedure was repeated up to two additional times. We modeled the outcome as inability to maintain single-leg stance in any attempt (yes; no).

Statistical modeling

We used multivariable linear regression to estimate the mean differences and 95% confidence intervals (95% CI) in vibrotactile threshold, visual contrast sensitivity, and postural sway associated with increasing quartiles (Q) of exposures among workers. We checked assumptions of linear regression via diagnostic plots of the residuals. Assumptions were met for all continuous outcomes except for postural sway, for which we removed six outliers and ln-transformed mean sway speeds (both eyes open and eyes closed) to improve normality and achieve homoscedasticity of the residuals. For test of maintaining single leg stance, we used multivariable modified Poisson regression to estimate prevalence ratios (PR) and 95% CI with each quartile exposure of interest (Zou 2004). In addition to assessing quartile exposures, we also examined ln-transformed continuous exposure levels in all tests.

We selected covariates to be adjusted for based on a directed acyclic graph (DAG) and included the minimally sufficient adjustment set and predictors of the outcome that are not descendants of the exposure in the model (Greenland et al. 1999) (Figure S1). In the main analysis, we adjusted for continuous age, sex (male; female), self-reported race (White; Black; other/multiracial), cigarette smoking (current; former; never), highest educational attainment (less than high school; high school diploma or general equivalency diploma; some college or 2-year degree; 4-year college graduate or more), current binge drinking (consumption of ≥5 drinks on one occasion ≥12 times in the year before the clinical examination), previous oil industry experience (yes; no), and residential proximity to the spill (living in a coastal county directly affected by the spill; living in a Gulf or non-Gulf state further from the spill) in all models. For analyses, we aggregated participants who self-identified as races other than White or Black and those who identified as multiple races into the “other/multiracial” group because of the small numbers of workers in these groups (see Supplemental Text S1 for all race selection options). We used self-reported race as a proxy for the downstream effects of socioeconomic disparities as well as differential life experiences due to structural racism (e.g., health care access and utilization, neighborhood ambient air pollution exposure) that might have influenced workers’ health and neurologic function (Lane et al. 2022; Saadi et al. 2017; Xu et al. 2016). All covariates, with the exception of binge drinking, which was self-reported during the clinical examination, were ascertained in the enrollment interview. For analysis of vibrotactile threshold, we additionally adjusted for height measured in the clinical examination. For visual contrast sensitivity, we restricted analyses to participants with better than 20/50 visual acuity and additionally adjusted for use of vision correction (i.e., wearing glasses or corrective lenses). For tests of postural stability and ability to maintain single leg stance, we conducted analyses with and without additional adjustment for ever performing OSRC work on water vessels to try to account for baseline vestibular function.

In a sub-analysis, we investigated potential effect measure modification (EMM) of the associations between crude oil exposures and neurologic function by age at enrollment (<50 vs. ≥50 years). We evaluated EMM of the associations by assessing stratified models to determine effects in the subgroups and by performing likelihood ratio tests across strata and reporting the p-values. In another sub-analysis, we examined associations restricted to men; there were too few women in the study population to conduct analysis in this subgroup. In a separate sensitivity analysis, we additionally adjusted in the model for body mass index (BMI) (underweight or normal; overweight; obese) calculated from height and weight measured at the clinical examination, as obesity has been associated with neurologic function with mechanisms beyond insulin dysregulation (Wang et al. 2016). We did not account for BMI in the main analysis because crude oil chemical exposures have been associated with diabetes in this cohort (Jardel et al. 2022) and with obesity in other studies (Lee et al. 2022); thus, BMI ascertained after exposure might act as a mediator in the association of crude oil exposure and neurological outcomes. To address potential confounding from exposure to higher levels of fine particulate matter (PM_2.5) from controlled burning activities (Chen et al. 2022; Pratt et al. 2022), which has been modestly associated with neurological test performance in the same cohort (Norris et al., in press), we repeated analyses excluding 210 workers with higher PM_2.5 exposures from burning-related activities. Because age-related declines in neurologic function have been well-documented (Kalisch et al. 2008; Seidler et al. 2010), with studies showing a loss of mechanoreceptors on the skin, declines in nerve conduction velocity (Kalisch et al. 2008; Seidler et al. 2010), and changes in the structure and biochemistry of the brain (Vassar and Rose 2014) with increasing age, we also assessed whether our estimates were sensitive to how age was specified in the model by investigating restrictive cubic splines as an alternative functional form of age. In addition to examining cumulative exposures, we also examined average exposures, calculated as the ratios of cumulative exposures and exposure durations, as measures of exposure intensity.

All analyses were performed using SAS, version 9.4 (SAS Institute Inc., Cary, NC, USA). An alpha level of 0.05 was considered statistically significant for all analyses.

Results

Around 63% of the analytical sample were younger than 50 years at enrollment, and about three quarters were male (Table 1). Half of all participants self-reported race as White, while 41% of them self-reported as Black. The remaining participants (8%) self-identified as other races or belonging to multiple race categories. At enrollment, less than half of participants had at least some college education, and around one third were current smokers. About 11% of the participants were categorized as binge drinkers. Approximately 16% of the participants had prior experience working in the oil industry, and the vast majority of participants (93%) resided in a Gulf state coastal county/parish directly impacted by the spill at time of enrollment. The highest daily exposure to each BTEX-H chemical was well below the threshold limit values established by the American Conference of Governmental Industrial Hygienists (benzene: 500 ppb; toluene: 20,000 ppb; ethylbenzene: 20,000 ppb; xylene: 20,000 ppb; n-hexane: 50,000 ppb) or the permissible exposure limits established by the Occupational Safety and Health Administration (benzene: 1000 ppb; toluene: 200,000 ppb; ethylbenzene: 100,000 ppb; xylene: 100,000 ppb; n-hexane: 500,000 ppb). Three quarters of workers were exposed for ≤8 months (Table S1).

Table 1.

Characteristics of DWH disaster oil spill workers who completed at least one neurologic function test (N=2,610)

Characteristic	n (%)
Age (y)
<30	427 (16.4)
30–39	542 (20.8)
40–49	677 (25.9)
50–59	650 (24.9)
≥60	314 (12.0)
Sex
Male	2013 (77.1)
Female	597 (22.9)
Race
White	1336 (51.2)
Black	1061 (40.7)
Other/multiracial	213 (8.2)
Education
Less than high school	555 (21.3)
High school diploma/GED	913 (35.0)
Some college/2-year degree	786 (30.1)
4-year college graduate or more	356 (13.6)
Smoking status
Current smoker	896 (34.3)
Former smoker	548 (21.0)
Never smoker	1166 (44.7)
Current binge drinkera
No	2332 (89.4)
Yes	278 (10.7)
Previous oil industry experience
No	2181 (83.6)
Yes	429 (16.4)
Residential county proximity to Gulf of Mexicob
Direct contact	2432 (93.2)
Indirect contact or other residence	178 (6.8)
Vision correctionc
Missing	78
No	1602 (63.3)
Yes	930 (36.7)
Poor visiond
Missing	77
No	2066 (81.6)
Yes	467 (18.4)
Weight classification
Missing	1
Underweight or normal (BMI <25)	572 (21.9)
Overweight (25 ≤ BMI < 30)	828 (31.7)
Obese (BMI ≥ 35)	1209 (46.3)
Median (min, max) height (in)	68.6 (55.4–78.3)

Abbreviations: DWH, Deepwater Horizon; GED, General Equivalency Diploma

^aBinge drinking is defined as self-reporting ≥5 drinks on one occasion ≥12 times during the year before the clinical exam

^bDirect proximity is defined as living in a county directly adjacent to the Gulf of Mexico; indirect proximity or other residence is defined as living in a Gulf or non-Gulf state further from the spill.

^cVision correction includes wearing corrective glasses and contact lenses.

^dPoor vision is defined as binocular visual acuity score of 20/50 or worse.

In tests of vibrotactile threshold, we observed modestly poorer vibrotactile acuity among workers in upper quartiles of exposure to benzene, toluene, ethylbenzene, and n-hexane compared to the respective referent groups (Figure 1, Table S2). However, the increases were not statistically significant, nor was there evidence of a monotonic exposure-response trend. No apparent associations were observed in analyses of continuous exposures.

Figure 1.

Cumulative exposure (quartiles) to crude oil chemicals and vibrotactile threshold (micron) among all participants and stratified by age at enrollment

We did not observe a positive association between the crude oil exposures and visual contrast sensitivity score (Figure 2, Table S3). At a spatial frequency of 1.5 cycles/degree, we observed unexpectedly better performance (i.e., higher contrast sensitivity scores) in contrast sensitivity across all upper quartiles of exposure to THC and BTEX-H compared to the first quartile, with some chemicals showing a monotonic trend. Effect estimates at spatial frequencies of 3, 6, and 12 cycles/degree were modestly above the null, but with no evidence of exposure-response trends. We observed non-monotonic and non-significant decreases in contrast sensitivity score at 18 cycles/degree.

Figure 2.

Cumulative exposure (quartiles) to crude oil chemicals and visual contrast sensitivity performance score among DWH disaster oil spill workers

There was no consistent relationship between crude oil chemical exposures and ln-transformed postural sway speed (Figure 3, Table S4). In analyses of continuous exposures, we observed a statistically significant positive association for xylene when participants performed the test eyes open. Patterns of association were similar when participants performed the test with their eyes closed (Figure 4, Table S5).

Figure 3.

Cumulative exposure (quartiles) to crude oil chemicals and ln-transformed postural sway speed (mm/s) (eyes open) among all participants and stratified by age at enrollment

Figure 4.

Cumulative exposure (quartiles) to crude oil chemicals and ln-transformed postural sway speed (mm/s) (eyes closed) among all participants and stratified by age at enrollment

We observed no associations between THC or benzene exposure and inability to maintain a single leg stance (Figure 5, Table S6). Modest but non-significant increases in PR were observed in Q4 for toluene, ethylbenzene, and xylene compared to the respective referent groups (range of PR: 1.13–1.18); however, there was no evidence of an exposure-response relationship across quartile exposures or in analyses of continuous exposures.

Figure 5.

Cumulative exposure (quartiles) to crude oil chemicals and inability to maintain single leg stance among all participants and stratified by age at enrollment (note that y-axis is on the log scale)

When analyses were stratified by age at enrollment (<50 vs. ≥50 years), we observed heterogeneous associations for vibrotactile threshold (Figure 1, Table S2) and postural sway tests (Figures 3, Tables S4–5), with stronger associations observed among participants ≥50 years of age. When examining vibrotactile threshold within the older subgroup, we observed monotonic trends across quartile exposures to benzene, toluene, and n-hexane and statistically significant effect estimates in the top quartile of these chemicals. Within the younger subgroup, we observed non-significant associations in the inverse direction. When examining the eyes-open postural sway test, we saw elevated effect estimates across higher quartiles of exposure to THC and BTEX-H within the older subgroup, including monotonic trends for toluene and xylene. Within the younger subgroup, effect estimates were close to or below the null. Associations were also stronger within the older subgroup, particularly for benzene, toluene, and n-hexane when participants performed the test with their eyes closed. In the single leg stance test, we observed stronger associations for several chemicals (benzene, toluene, xylene, n-hexane) among the older subgroup compared to the younger group, although likelihood ratio tests did not reach statistical significance (Figure 4, Table S6). There was no evidence of heterogeneity in associations between crude oil exposures and contrast sensitivity by age (Figure S2, Table S7). We also examined associations among workers age 60 years or older; however, because of the relatively small number of these subjects, effect estimates were unstable and consequently difficult to interpret. Therefore, we have not presented these results.

There were no substantive differences in results when we restricted the analyses to men (Figures S3–7, Tables S8–13). We observed similar results in sensitivity analyses in which we additionally adjusted for BMI or modeled age as splines (results not shown) or restricted to workers with lower levels of burning-related PM_2.5 exposure (Figures S3–7, Tables S8–13). Associations were somewhat reduced in analyses examining average exposures and in analyses not adjusted for ever working on water vessels (results not shown).

Discussion

In this study, we examined the relationship of exposure to THC and BTEX-H with several tests of neurological performance among oil spill workers 4–6 years after the DWH disaster. In the subgroup older than 50 years at enrollment, workers with higher exposures showed worse performance in the vibrotactile threshold test and modestly worse performance in postural stability and single-leg stance tests compared to those with lower exposures. The most consistent associations in this age group were observed for vibrotactile threshold, where effect estimates were above the null across all quartiles of exposure compared to the referent group, and several chemicals showed monotonic exposure-response trends. We did not find evidence of adverse neurological effects in the younger subgroup. Among the overall study population, we observed unexpectedly better performance in visual contrast sensitivity among workers with higher exposure, with several monotonic trends observed at a spatial frequency of 1.5 cycles/degree.

Numerous epidemiologic studies have shown adverse neurologic effects among populations exposed to high levels of volatile hydrocarbons (Grisold and Carozzi 2021; Valentine 2020). n-Hexane has been established as a cause of peripheral neuropathy among factory workers and persons with glue-sniffing addiction (Takeuchi 1993). Overt or subclinical changes in peripheral nerves, assessed via electromyography or vibrotactile threshold test, have also been detected among workers exposed to organic solvents containing one or more components of BTEX at concentrations above the occupational guidelines (Baslo and Aksoy 1982; Costa et al. 2012; Demers et al. 1991; Indhushree et al. 2016; Shih et al. 2011).

Only a few studies have examined quantitative measures of peripheral nerve function among populations exposed to lower levels of volatile hydrocarbons. In Bleecker et al. (1991), lifetime-weighted average occupational exposure to total hydrocarbons was associated with higher vibrotactile thresholds among paint manufacturing workers. Significant changes in nerve conduction measures were linked to longer duration of work and higher levels of urinary hippuric acid (a metabolite of toluene), but not with urinary methylhippuric acid (a metabolite of xylene), in a study of paint and lacquer factory workers exposed to sub-occupational limit levels of toluene and xylene (Jovanović et al. 2004). In a small study of shoe factory workers, urinary 2,5-hexanedione (a metabolite of n-hexane) concentration was associated with significant changes in sensory nerve action potential; however, no association was found for time-weighted average n-hexane exposure estimated from breathing zone samples (Mutti et al. 1982b). In our study, we found evidence of adverse associations between THC/BTEX-H exposures and vibrotactile threshold only among workers older than 50 years of age. In contrast to the GuLF Study participants who were exposed to crude oil chemicals for only a few months (median: 5 months, Q1-Q3: 3–8 months), workers in these earlier occupational studies were typically exposed for many years, which might account for the stronger associations observed in these studies. In addition, while the majority of occupational studies assessed neurologic function while workers were still exposed, our study examined neurologic function several years after the cessation of the spill-related occupational exposures. This might have allowed some workers to recover from the neurotoxic insult and may explain the weaker associations observed in our study.

Exposure to toluene, styrene, and mixed solvents at levels above the occupational limit have been associated with color vision loss and/or lower contrast sensitivity among workers (Boeckelmann and Pfister 2003; Costa et al. 2012; Gobba and Cavalleri 2003; Indhushree et al. 2016). Visual dysfunction among workers has also been observed at lower concentrations of volatile hydrocarbons, with several studies reporting impaired color vision among exposed workers (Campagna et al. 2001; Cavalleri et al. 2000; Lee et al. 2007) and one study linking reduced contrast sensitivity to urinary metabolite of xylene among workers (Gong et al. 2003). In our study, we did not observe positive associations between THC/BTEX-H exposure and contrast sensitivity. Unexpectedly, we observed modestly better performance in this test among workers with higher exposure, particularly at spatial frequency of 1.5 cycles/degree, although few associations were statistically significant. This could result from residual confounding by workers’ baseline visual function (“healthier worker effect”) (Chowdhury et al. 2017). Indirect evidence for this comes from a cross-tabulation of job group and use of vision correction, which showed that the proportion of vision correction use was lowest among response (33.3%) and operations workers (31.9%), who experienced the highest crude oil exposures, and highest among support workers (50.9%), who had the lowest exposures. It is plausible that workers with better visual function were selected into jobs that, coincidentally, also exposed them to higher levels of crude oil, and differences in baseline visual function could not be completely accounted for by restricting this analysis to workers with better than 20/50 visual acuity and adjusting for use of vision correction.

Maintenance of posture and balance requires coordination among the peripheral nervous system, the central nervous system, and the musculoskeletal system (Vassar and Rose 2014). The whole process involves the receipt of sensory signals by the visual, vestibular, and somatosensory systems, neural transmission of sensory information to the brain for processing, and coordinated motor responses transmitted from the brain via motor neurons to muscles for postural adjustment (Vassar and Rose 2014). In our study, we evaluated workers’ ability to maintain balance by measuring postural sway speed when participants stood on a platform and, separately, by asking participants to maintain a 30-second single-leg stance. We observed only suggestive associations, and only among workers older than 50 years at enrollment. Although symptoms of dizziness/vertigo, unsteady gait, and muscle weakness have been reported in some studies of workers exposed to organic solvents (Cianchetti et al. 1976; Herbert et al. 1995; Jovanović et al. 2004; Mutti et al. 1982a; Sanagi et al. 1980), few studies have evaluated balance using quantitative tests. In a small study of aircraft maintenance personnel exposed to jet fuel vapor concentrations above the occupational guidelines, postural stability was significantly associated with cumulative exposure to benzene, toluene, and xylene (Smith et al. 1997). The relationship between acute exposure to low concentrations of m-xylene and postural sway was assessed in two studies of volunteers, with positive associations found in one study (Savolainen et al. 1984; Savolainen et al. 1985). Kuo et al. (1996) explored postural sway in a small group of sewer workers chronically exposed to low levels of volatile hydrocarbons and found positive associations with a composite exposure metric measured as “total benzene equivalents”.

Several studies have investigated the neurologic effects of OSRC work. In analyses of Hebei Spirit oil spill workers, higher prevalence of neurological symptoms was reported a few weeks after the spill among workers with inhalation exposure to crude oil (Sim et al. 2010) and longer duration of OSRC work but not with urinary hippuric acid or mandelic acid (Ha et al. 2012). In addition, a persistent but non-significant association was observed between cleanup duration and vestibular symptoms one year after that spill (Na et al. 2012). In the DWH disaster, increasing frequency of inhalation exposure to crude oil was associated with higher prevalence of neurologic symptoms during the OSRC cleanup among a group of Coast Guard responders who participated in the DWH spill cleanup (Krishnamurthy et al. 2019). In the same cohort, ever inhalation exposure to crude oil was associated with higher hazards of International Classification of Diseases-coded neurological diagnoses ascertained from military health encounter records up to 5.5 years post-DWH (Denic-Roberts et al. 2023). A previous analysis in the GuLF Study examined participants’ neurobehavioral function 4–6 years after the spill using 16 quantitative neurocognitive tests and found modest decreases in some tests among workers with higher single daily maximum exposure to THC compared to those with lower maximum exposures (Quist et al. 2019). Blood concentrations of benzene and toluene were associated with self-reported neurologic symptoms among participants of the GuLF Study, although the blood samples were collected 2–3 years after the disaster and reflected recent ambient exposures rather than cleanup-related exposures (Werder et al. 2019).

In our study, we observed heterogeneity in associations between crude oil exposures and performance on vibrotactile threshold, postural stability, and single-leg stance tests by age, with effects seen primarily in the workers over 50 years of age. Although few studies have examined EMM of the relationship between volatile hydrocarbons exposure and neurologic function by age, studies of chemotherapies (du Bois et al. 1999) and other neurotoxicants (Albers et al. 1988) have shown higher risk or severity of toxic neuropathy among older participants. Research has shown that the peripheral nerves have the capacity to regenerate, and certain regions in the brain have the ability to compensate for neurotoxic insult (Tilson and Mitchell 1983); however, this capacity decreases with age (Verdú et al. 2000). The length of time in our study between exposure and the clinical examination (median: 4.6 years, range: 4.2–6.1 years) may have allowed the younger participants to recover from any neurotoxic deficits induced by crude oil chemicals, while older workers were less able to recover. It is also possible that older workers are more susceptible to neurotoxicants because they are less able to detoxify and excrete these chemicals. Studies have shown that the metabolic clearance of many drugs decreases with age, likely due to reduced liver and renal blood flow and functions (Mangoni and Jackson 2004; Soejima et al. 2022). Another possibility is that older individuals have less physiological reserves for responding to toxic insults; age-related declines in neurologic function have been well-documented in studies (Kalisch et al. 2008; Seidler et al. 2010). Lastly, the lipophilic nature of BTEX-H chemicals may have allowed them to accumulate more easily among older workers, who likely have a higher body fat percentage from the aging process compared to younger workers (proportion of overweight/obese in our study by age: 75.2% among <40 years; 78.6% among 40–49 years; 80.3% among 50–59 years; 81.2% among ≥60 years) (Palmer and Kirkland 2016).

The mechanisms by which BTEX-H and other volatile hydrocarbons impair neurologic function have been investigated in animal and cellular studies (Tormoehlen et al. 2014). BTEX-H chemicals have been hypothesized to lead to axonopathy by changing the integrity of the cellular membrane or interacting with the membrane proteins to impair cellular functions (ATSDR 1999, 2000, 2007a, b, 2010). Other proposed mechanisms involve generation of reactive oxygen species, induction of neurofilament accumulation that interrupts axonal transport (ATSDR 1999, 2007a; Valentine 2020), and demyelination of axons that leads to impaired signal conduction (Valentine 2020). These processes occur not only in the peripheral nervous system, but also affect the central nervous system (including the optic nerve), as the high lipophilicity and relatively low molecular weight of BTEX-H and their metabolites allow them to readily cross the blood-brain barrier (Cruz et al. 2014; Luo et al. 2021; National Research Council Committee on Acute Exposure Guideline 2010; World Health Organization 2010). Studies have demonstrated the presence of these chemicals in brains of exposed animal models and humans (Ameno et al. 1992; Avis and Hutton 1993; Ito et al. 2002; Kishi et al. 1988). The brain white matter, which contains a high concentration of lipid-rich myelin, may be more susceptible to the toxicity (Filley 2013). Consistent with this mechanism, studies have observed accumulations of neurofilaments in the axoplasm, myelin degeneration, and/or structural and biochemical changes in the brain in animals and humans exposed to these chemicals (Andersson et al. 1981; Bjornaes and Naalsund 1988; Hsieh et al. 1988; Mutti et al. 1988; Puri et al. 2007; Rosenberg et al. 1988; Schaumburg and Spencer 1976).

One strength of our study is the rigorous assessment of participants’ cumulative exposures using a job-exposure matrix based on extensive personal air samples collected on a subset of DWH oil spill workers, including some who enrolled in the GuLF Study, during the spill and detailed DWH spill work histories self-reported by all study participants, which allowed us to investigate individual etiologic agents potentially responsible for any observed neurologic effects among these workers. This study also contributes to the larger occupational literature on volatile hydrocarbons by investigating neurological effects of several chemicals at exposures below occupational limits. Previous studies of these chemicals have examined more highly exposed worker populations, with most relying on crude proxies of exposure (e.g., exposed vs. control group, duration of work). The few studies that examined exposure to individual chemicals below the occupational limits were of modest size (Bleecker et al. 1991; Gong et al. 2003; Jovanović et al. 2004). With a larger cohort, we were better powered to conduct subgroup analyses and found adverse neurologic effects among older workers that persisted several years after exposure. Persistent neurological effects have been observed in a few studies of workers exposed to organic solvents for years (Cianchetti et al. 1976; Mutti et al. 1982b), but we are the first to report such effects after several months of exposure. Another strength of the study was the use of validated quantitative measures to assess neurologic function, including subtle, subclinical effects that may occur earlier in disease progression (Gerr et al. 2000). These tests have been shown to have good test-retest reliability (Gerr and Letz 1988; Hohberger et al. 2007; Letz et al. 1994) and have been used in many other studies (Bates et al. 2019; Indhushree et al. 2016; Starks et al. 2012; Yokoyama et al. 1997). Lastly, because the GuLF Study obtained data on a wide range of factors, we were able to account for numerous potential confounders of the associations to obtain more accurate effect estimates.

Our study has a few limitations. First, study findings need to be interpreted cautiously in the context of the cross-sectional study design. There could also be bias if participants were assigned to specific jobs/activities based on health-related factors at the time of spill (e.g., baseline neurological function) that were predictive of their neurologic function at the clinical examination. We accounted for several indicators of baseline health by excluding participants with pre-spill diabetes or peripheral neuropathy and by adjusting for BMI and, for the visual contrast sensitivity test, vision correction. For tests of postural stability and single-leg stance, we found slightly stronger associations when we additionally adjusted for ever performing OSRC work on a water vessel, a proxy for baseline vestibular function. Still, there could be residual confounding from unmeasured confounders (e.g., other medical conditions that affected neurologic function) or imperfect measurement of existing covariates (e.g., inaccuracy in self-reported medical diagnoses and smoking/drinking) in the models, which may in part explain the apparent better performance observed for visual contrast sensitivity. We lacked the data to adjust for certain other potential exposures experienced during the oil spill cleanup (Middlebrook et al. 2012), although we accounted for PM_2.5 from the controlled burning, which has been modestly associated with some neurologic function test performance in this cohort (Norris et al., in press) and found similar results. We did not collect data on participants’ occupational exposures after the end of their spill cleanup. It is possible that some workers who had higher spill-related exposures continued to have higher occupational exposures after the spill cleanup that could have affected their neurologic function at the time of the clinical exam. We adjusted in the models for pre-spill oil industry work as a proxy for pre-spill crude oil exposures; this variable may also serve as a reasonable indicator of participants’ propensity to work in the oil industry after the spill. Additionally, participants who reported working in the oil industry accounted for <18% of participants in the highest quartile of THC and BTEX-H exposures. Thus, continued occupational exposures from oil industry work is insufficient to explain the associations observed in our study.

Second, the exposure estimates assigned to workers contain some degree of uncertainty, (Stewart et al. 2022). Exposure group average exposures were estimated using a left-censored Bayesian method (Groth et al. 2017; Groth et al. 2022b; Huynh et al. 2016) shown to have low bias and imprecision for most exposure groups, but more uncertainty for those with extreme censoring or few measurements (Stenzel et al. 2022a). However, this is unlikely to have substantively biased our risk estimates in analyses using exposure quartiles because the majority of workers in these low-exposure groups fell into the lowest exposure category.

Third, there could be a selection bias if workers who participated in the clinical examination differed in important ways from workers eligible for the clinical examination (based on residential proximity to a study clinic) (Kwok et al. 2017) who did not participate. It is possible that workers who had greater neurological symptoms were more motivated to participate in the clinical examination; alternatively, any adverse neurological or other health effects from OSRC-related exposures might have reduced their ability to travel to the clinic. To explore the potential for selection bias, we compared characteristics, including neurologic symptoms, ascertained at enrollment between eligible workers who participated in the clinical examination and those who did not (Table S14). We found minimal differences between the two groups, mitigating concerns about this potential bias.

Our study showed modest associations between cumulative exposures to several volatile components of crude oil and several tests of neurologic function 4–6 years after the oil spill among spill workers who were 50 years of age or older. We observed no apparent associations among the younger workers, perhaps because younger workers were less susceptible to the neurotoxic exposures or because they experienced greater recovery of neurologic function after cessation of the exposure. The positive associations were consistent with findings from previous studies of individuals exposed to oil spills and workers exposed to organic solvents and provide evidence that exposure to these agents at levels below current occupational guidelines may induce adverse neurologic effects in a susceptible population. Additional follow-up time for these workers would help determine whether the observed associations persist beyond 4–6 years after the spill. Research is also needed in other populations and settings to confirm these adverse neurologic effects observed for crude oil chemical exposures below occupational limits.

Highlights.

• The Deepwater Horizon oil spill workers were exposed to crude oil chemicals.

• Exposures to BTEX and n-hexane were elevated, but below occupational limits.

• Neurologic function was assessed via a battery of four tests at a clinical exam.

• The clinical exam occurred 4–6 years after the spill.

• Exposures were linked to worse neurologic function among workers ≥50 years of age.

Abbreviations:

BMI

body mass index

BTEX-H

benzene, toluene, ethylbenzene, xylene, and n-hexane

CI

confidence interval

DWH

Deepwater Horizon

EMM

effect measure modification

OSRC

oil spill response and cleanup

PR

prevalence ratio

THC

total petroleum hydrocarbons