Seafood Contamination after the BP Gulf Oil Spill and Risks to Vulnerable Populations: A Critique of the FDA Risk Assessment

Background: The BP oil spill of 2010 resulted in contamination of one of the most productive fisheries in the United States by polycyclic aromatic hydrocarbons (PAHs). PAHs, which can accumulate in seafood, are known carcinogens and developmental toxicants. In response to the oil spill, the U.S. Food and Drug Administration (FDA) developed risk criteria and established thresholds for allowable levels [levels of concern (LOCs)] of PAH contaminants in Gulf Coast seafood. Objectives: We evaluated the degree to which the FDA’s risk criteria adequately protect vulnerable Gulf Coast populations from cancer risk associated with PAHs in seafood. Discussion: The FDA LOCs significantly underestimate risk from seafood contaminants among sensitive Gulf Coast populations by failing to a) account for the increased vulnerability of the developing fetus and child; b) use appropriate seafood consumption rates; c) include all relevant health end points; and d) incorporate health-protective estimates of exposure duration and acceptable risk. For benzo[a]pyrene and naphthalene, revised LOCs are between two and four orders of magnitude below the level set by the FDA. Comparison of measured levels of PAHs in Gulf seafood with the revised LOCs revealed that up to 53% of Gulf shrimp samples were above LOCs for pregnant women who are high-end seafood consumers. Conclusions: FDA risk assessment methods should be updated to better reflect current risk assessment practices and to protect vulnerable populations such as pregnant women and children.


General Parameters
We applied more health-protective variables used in other historical FDA oil spill seafood risk assessments to describe the acceptable risk level (1 in a million cancer risk) and exposure duration (10 years).

Adult Variables
The FDA risk assessment performed for dioctyl sodium sulfosuccinate, a component of the chemical dispersants, used a more protective body weight for adults, including women, of 60kg (Bolger 2010). To better account for high-end seafood consumption patterns on the Gulf Coast, we used consumption rates that more closely approximate high end consumers. Where available (fish, shrimp, and crab), we calculated the 95 th percentile consumption rates based on the total seafood consumption rates and fraction apportioned to the individual seafood categories reported in Seafood Choices: Balancing Benefits and Risks (IOM 2007). We found very limited data on oyster consumption rates, particularly for high end consumers. For this analysis, we relied on the consumption frequency data collected in a study of seafood consumption rates among Louisiana Recreational Anglers (Lincoln et al. 2011) coupled with a "standard" seafood portion size from the EPA Exposure Factors Handbook. While this study is regionally specific to the Gulf Coast, the respondents were overwhelmingly Caucasian; consumption rates from other coastal communities, such as the Vietnamese fishing community, which has been identified by the US EPA as a population with significantly elevated seafood consumption rates, are likely to be significantly higher.

Early Life Exposure Adjustment Parameters
To account for the increased vulnerability of subpopulations, we evaluated two early life exposure scenarios -Child (exposure between 2-12 years old) and Pregnant woman (exposure during the 3 rd trimester of gestation and continuing for 10 years). We evaluated cancer risks using US EPA and California EPA published methods for childhood and prenatal exposures (US EPA 2005, OEHHA 2009). This includes age-specific doses based on available data on consumption rates and body weight and early-life vulnerability factors. For the child scenario, we used the US EPA Age Dependent Adjustment Factors (ADAFs) and for the pregnant woman scenario, we used the California EPA's Office of Environmental Health Hazard Assessment (OEHHA) median Age Sensitivity Factors (ASFs). Subsequent to publication of the median ASFs, OEHHA utilized an ASF of 3 for the juvenile age group (2 to <16) in a risk assessment for the PAH, benzo(a)pyrene. We incorporated this information by using an ASF of 3 for the later portion of the juvenile age range (6 to <12) while maintaining the ASF 5 for the younger, more vulnerable, (2 to 5) age range. Dividing this age grouping also allowed us to apply more agespecific and relevant exposure variables.
For fish, shrimp, and crabs, we were able to use the consumption rates beginning at 2 years of age from the IOM (2007) report. The US EPA (2008) Child Specific Exposure Factors Handbook reported fish consumption rates for high-end consumers for the 0 to 2 year old interval. To approximate consumption rates for shellfish (shrimp, crabs, and oysters) during this period, we applied a scaling factor based on the consumption distribution for fish (the ratio of consumption rates for the 2-5 compared to 0-2 age groupings) to the 2-5 year old consumption rates for shellfish. Similar to adults, limited information on childhood consumption rates for oysters is available. For this assessment, we used the reported values for Native American Children in the US EPA (2011) 2011 Edition of the Exposure Factors Handbook. Prenatal exposures were estimated as 10% of the maternal dose based on available data from animal studies on PAH toxicity (Perera et al 2005).

Total Shellfish Consumption Risk Profile
It would be expected that high-end seafood consumers are not just eating one type of shellfish; therefore cancer risks should reflect the total shellfish diet. We calculated revised LOCs to reflect this combined shellfish consumption by apportioning the acceptable risk according to the relative fraction of the combined intake of our estimated individual shellfish consumption rates.

Health risks associated with Gulf Coast shellfish tested after the oil spill
We analyzed seafood testing data collected by the FDA and NOAA to determine health risks associated with the reported carcinogenic PAH levels. During the first round of testing, NOAA utilized a different analytical method from FDA and from later sampling events. Therefore, the results are not comparable and are reported separately. Results reported by FDA included detected values, the designation of TR for "trace," and non-detect. In this analysis, we evaluated all detected levels of carcinogenic PAHs. To accomplish this, we assigned the samples marked "trace" with a value equal to half the level of quantification reported for that analyte-seafood type pairing in the FDA's Laboratory Information Bulletin for the analytical method (US FDA 2010b). For each type of seafood, PAH, and responsible agency, we calculated the total number of samples available (from oil-spill impacted areas), the number and percent of samples with detectable levels of PAHs, the mean and 95% Confidence Interval (CI) of the detected values, and the percent of samples that exceeded the revised LOCs (Supplemental Tables 2&3). The 7 PAHs for which established cancer toxicity equivalents have been derived can be summed using the cancer equivalent values to reflect their relative potency as compared to benzo(a)pyrene. We used this method to calculate total BaP-equivalents and compared this value to the revised LOCs.
We also calculated estimated excess cancer risk for our 3 Gulf Coast Vulnerable Populations scenarios associated with the mean (and 95 th CI) values of total BaP-equivalents and naphthalene (Table 4). For this calculation, we averaged the available data for shrimp (FDA and NOAA HPLC analytical method data) and also averaged the concentrations of PAHs in crab meat and crab hepatopancreas. We assumed that a high end consumer is eating all three different types of seafood and therefore calculated cancer risks from combined shellfish exposures and for total cancer risk from BaPe and naphthalene exposures.  Law et al (2002) attempted to standardize the literature using the BaPequivalent approach and reported a range of 0 to 222 ppb BaPe from the US mussel watch monitoring program 1986-1996 data files. 16% 55% * Concentrations of carcinogenic PAHs (including naphthalene) were compared to LOCs revised to incorporate total shellfish consumption. As per FDA procedures for carcinogenic PAHs, samples were flagged where the combined ratio of the concentration to the relevant LOC exceed one.