Gulf Longitudinal Follow-up (GuLF) Study

National Institute Of Environmental Health Sciences

Investigators

Linked publications, trials & patents

Abstract

The Gulf Long-term Follow-up Study (GuLF STUDY, www.gulfstudy.nih.gov) was initiated in response to the April 2010 explosion of the Deepwater Horizon (DWH) drilling rig and resulting oil spill in the Gulf of Mexico, the largest marine oil spill in U.S. history. A total of 32,608 oil spill response and clean-up workers completed an enrollment questionnaire (2011-2013. Over 11,000 from the Gulf states completed a home visit that included collection of biospecimens and anthropometric and lung function measurements. In response to community concerns about reports of oil-related chemicals in the blood of some people years after the spill, we analyzed samples from 1000 participants collected at home visit for volatile organic compounds including BTEX-h and styrene as well as metals. Levels of these chemicals, apart from styrene, a chemical unrelated to the oil spill, were not different from those in a nationally representative and local sample of NHANES participants once cigarette smoking was considered. Follow-up data were collected from 21,256 participants between 2013-2016. Mental health trajectories were assessed among those who had symptoms of depression, PTSD, or anxiety at enrollment and a random sample of the remainder in four mental health interviews over two years (n=2,969). Participants (n=3,403) living within 60 miles of New Orleans, LA or Mobile, AL completed a clinical exam (2014-2016) that included additional lung function tests, assessment of neurological function, and biological sample collection. A second follow-up interview was completed in May 2021 with a response rate of 61% among those who completed the first follow-up. Participants are being followed via linkage to cancer registries and the National Death Index. Extensive efforts were made to characterize the exposures of study participants. We used questionnaire and measurement data that had been taken during the spill to develop job exposure matrices for a wide range of possible exposures. This work involved recalibrating exposure measurements taken during cleanup, identifying numerous distinct exposure groups performing like-tasks defined by timing of work relative to the capping of the well, activities, and location, and development of new statistical approaches to deal with values below the LOD and other censoring. Ordinal exposure metrics (for THC, hierarchical job class, and exposure to burning and dispersants) have now been supplemented with quantitative measures of specific oil spill chemicals. The metrics have largely focused on inhalation exposures, but measures of potential dermal exposure are being finalized. A series of peer-reviewed papers describing these efforts have been assembled as a monograph in the Annals of Work Exposures and Health. Several papers have already appeared online; the remainder will be published later in 2021. We continue to explore respiratory health effects. In a subset who completed spirometry at both the home visit and the clinic exam, lung function decrements seen shortly after the spill were no longer apparent 4-6 years later, with the greatest improvement seen among those with the highest exposure. Despite reduced lung function at 1-3 years, at the 4-6-year exam, workers with total hydrocarbon (THC) exposure 1-2.99 ppm and 3 ppm compared to those with 0.29 ppm exhibited higher FEV1. Compared with support workers, those in jobs with the highest likelihood of exposure and those with the highest estimated total hydrocarbon exposure showed the greatest improvement between visits. A possible explanation for this unexpected finding is that those with higher exposures tended to be oil industry professionals or from other groups that may have had better lung function before the spill (Lawrence, et al., 2020). We recently evaluated modeled PM2.5 exposure estimates from burning and flaring of natural gas and oil during cleanup, finding that compared to workers not involved in or near the burning, those with higher cumulative PM2.5 exposure had substantially reduced FEV1, FVC and FEV1/FVC ratio (Chen et al., 2021). Similar dose-response trends were seen for average PM2.5 exposure and in analyses stratified by smoking and time from exposure to spirometry and when we restricted to workers without pre-spill lung disease. In field studies of lung function, a small but potentially informative subset often fails to meet spirometry quality criteria. We evaluated factors associated with spirometry test failure (Lawrence et al., 2021 in revision) to assess potential selection bias. We also studied the impact of neighborhood disadvantage and individual-level SES on lung function. We found that the Area Deprivation Index (ADI), a measure incorporating economic and other stressors, was associated with lower FEV1 and FVC, with magnitudes of associations reaching clinically meaningful levels, with increasing quartiles of ADI associated with lower FEV1 (Lawrence et al., 2021). Associations with FVC were similar and associations persisted even after adjustment for individual-level SES factors. In contrast, despite our findings of adverse mental health outcomes in exposed oil spill workers and studies reporting associations of anxiety and depression with poor lung function, perceived stress, PTSD, depression, and anxiety were not associated with reduced lung function in this cohort (Lawrence et al. in press). We also studied neighborhood deprivation in relation to obesity and diabetes. We found that living in neighborhoods with the highest level of deprivation was associated with prevalence of obesity and (Hu et al., 2021). We have also recently linked other external geospatial data on climate (National Risk Index and Spatial Hazard Events and Losses Database for the United States (SHELDUS)) and social disadvantage/vulnerability (CDC/ATSDRs Social Vulnerability Index and Social Vulnerability Index (Hazards & Vulnerability Research Institute) to the GuLF Study cohort. Using data from 9278 home visit participants who completed questionnaires on mental health symptoms, we linked 2005-2010 county-level data from SHELDUS, a database of loss-causing events, to participant's home address. The total count of loss events and the total duration of disaster days was positively associated with perceived stress but not other outcomes. However, disaster severity approximated by cumulative fatalities/injuries was associated with anxiety and stress and cumulative property/crop loss was associated with anxiety, depression, and PTSD (Hu et al., 2021). Associations with ADI and SHELDUS were independent of associations with oil spill exposures and did not explain those associations. In an analysis using data from the subgroup whose blood levels of chemicals were measured in enrollment samples, we found little evidence for an association between blood metals (cadmium, lead, mercury, manganese, and selenium) and either hypertension or blood pressure levels (Xu et al., 2021). A recently submitted paper evaluates incident hypertension in relation to oil spill exposures. We also investigated blood metals in relation to subclinical neurologic endpoints, considering the metals individually and as a mixture, and potential interactions among exposures. The highest quartile of cadmium was associated with a 50% higher prevalence of central nervous system symptoms, with stronger associations among nonsmokers and those with low selenium (Werder, et al., 2020). Selenium also modified associations between lead and peripheral nervous system symptoms, with increased symptoms in the low selenium group at all quartiles of exposure. Conversely, those with the highest co-exposure to mercury and selenium had reduced neurologic symptoms (PR = 0.73, 95% CI = 0.55, 0.96). Results of the mixture analysis were consistent with single chemical results.