Biomonitoring of per- and polyfluoroalkyl substances in minority angler communities in central New York State

Abstract

Onondaga Lake in central New York State was listed as a Superfund site in 1994 due to industrial disposal of pollutants. A biomonitoring program was conducted to assess exposure to over 70 legacy contaminants and contaminants of emerging concern in populations disproportionately at risk for exposure residing near Onondaga Lake and to educate these communities on how to reduce exposures. The populations of focus were refugees from Burma and Bhutan and low-income, primarily African American, anglers (urban anglers). These communities consume locally caught fish for economic as well as cultural reasons and therefore may be at higher risk of exposure. This study focuses on assessment of exposure to per- and polyfluoroalkyl substances (PFAS) and associations with local fish consumption. Using respondent driven sampling, 311 refugees and 89 urban anglers were enrolled in the study. Following informed consent, study participants provided blood and urine specimens and completed a questionnaire. Percentiles of locally caught fish meals in the past 12 months by race/ethnicity groups showed that the Burmese participants of Karen ethnicity were the highest consumers, with a median of 135 meals compared to 103 meals for the other Burmese participants, 70 meals for the urban anglers, and 44 meals for the Bhutanese participants. Compared to the National Health and Nutrition Examination Survey (NHANES) 2015–16 sample of the general U.S. population, the Karen participants had markedly elevated perfluorooctane sulfonic acid (PFOS) and perfluorodecanoic acid (PFDA) levels with median serum concentrations 9.5 times greater (41.6 ng/mL vs. 4.4 ng/mL) and 26.9 times greater (2.69 ng/mL vs. 0.10 ng/mL), respectively; the other Burmese participants had moderately elevated levels of PFOS and PFDA with median serum concentrations 3.0 times greater (13.3 ng/mL vs. 4.4 ng/mL) and 7.3 greater times greater (0.73 ng/mL vs. 0.10 ng/mL), respectively; and, PFAS levels were not elevated in the Bhutanese or urban angler cohorts. Male gender was consistently the strongest predictor of PFAS exposure among all study cohorts. A positive association between local fish consumption was indicated only for PFOS among urban anglers. An association between local fish consumption and PFAS was not statistically significant among the refugee cohorts, perhaps due to the lack of ‘lower-end’ exposure or exposure variability. Community events were held by the program staff to present the biomonitoring results and distribute community outreach materials with visual aids specific for the study populations to promote safe fish eating.

1. Introduction

Per- and polyfluoroalkyl substances (PFAS), also known as perfluorochemicals (PFCs), were introduced as industrial chemicals in the 1950s, and the global commercial applications of PFAS became seemingly endless, ranging from non-stick coatings on cookware to industrial surfactants and fire-resistant foams (Lindstrom et al., 2011). Production and use of these chemicals in the United States has decreased over the past decade; however, they persist in the environment and are found worldwide (Giesy and Kannan, 2001; Ahrens and Bundschuh, 2015). Perfluorooctane sulfonic acid (PFOS) levels have been declining in the general U.S. population over time (Calafat et al., 2007; Kato et al., 2011; CDC, 2021). Because of their toxicity and bioaccumulative potential, PFAS have been routinely measured in environmental matrices, fish, wildlife, and humans (Boulanger et al., 2004; Lau et al., 2007; Delinsky et al., 2010; Kato et al., 2011; Ye et al., 2018; Fair et al., 2019). PFOS and perfluorodecanoic acid (PFDA), long-chain PFAS compounds, dominated in frequency of PFAS occurrence in fish samples from U.S. urban rivers and the Great Lakes (Sinclair et al., 2006; Stahl et al., 2014). Several studies have demonstrated that dietary fish is an important source of human exposure to PFOS and PFDA (Christensen et al., 2016; von Stackelberg et al., 2017; Christensen et al., 2017). Drinking contaminated water is another main source of non-occupational PFAS exposure (ATSDR, 2019), particularly for perfluorooctanoic acid (PFOA) (Emmett et al., 2006).

Concern over the health effects of PFAS has recently gained momentum and global attention. Many studies have examined possible relationships between concentrations of specific PFAS in human biological specimens and various health effects. Most studies have focused on PFOS and PFOA, prominent PFAS compounds, which remain in the human body for years (ATSDR, 2021). The epidemiological findings from these studies are not definitive, but suggest a wide spectrum of possible adverse effects including cardiovascular disease, liver damage, thyroid disease, reduced vaccine efficacy, reproductive complications, and low birth weight (ATSDR, 2019; Ballesteros et al., 2017; Fitz-Si-mona et al., 2013; Gallo et al., 2012; Huang et al., 2018; Looker et al., 2014; Maisonet et al., 2012).

The Agency for Toxic Substances and Disease Registry (ATSDR) and the New York State Department of Health (NYSDOH) conducted a biomonitoring program to assess exposure to over 70 legacy contaminants (GLWQB, 1985) and contaminants of emerging concern (Klěcka et al., 2010), including PFAS, on populations consuming fish from the Great Lake Basin’s Onondaga Lake and nearby water bodies in central New York State (NYS). Onondaga Lake was declared a Superfund site in 1994 due to industrial waste dumped directly into the lake for almost a century (EPA, 2019). The restoration of Onondaga Lake has spanned many decades, and the water quality of the lake has improved dramatically. Part of Onondaga Lake remains a Superfund site with fish advisories in place, and swimming from the shore was prohibited at the time of this study. Persistent toxic substances such as mercury, polychlorinated biphenyls (PCBs), and PFAS are still monitored in the lake’s water, sediment, fish and wildlife (NYSDEC, 2014; NYSDEC, 2018). A study conducted in 2004 found elevated concentrations of PFOS in Lake Onondaga surface water (median level 756 ng/L), and PFOS was the most abundant PFAS compound in fish from 20 inland NYS lakes popular with anglers (Sinclair et al., 2006).

Our biomonitoring program focused on two human populations in central New York State at risk for exposure: 1) refugees from Burma (currently Myanmar) and Bhutan, and 2) urban anglers. Study cohorts were low-income and known to eat fish from local waterbodies. Therefore, both may be highly exposed to contaminants in fish and may bear disproportionally high health burdens from the exposure. The biomonitoring data provide important information to guide state and local public health actions to protect people within their jurisdiction. Our study cohorts, particularly the refugee cohorts, offer an opportunity to add novel information to existing data on human exposure to PFAS in very high-end fish consumers residing in the Great Lakes Basin. The overall program follows an earlier (2013–2014) biomonitoring program that included fish eating communities in western New York state, and a comprehensive list of contaminants measured for both biomonitoring programs has been previously published (Savadatti et al., 2019). This article is the first publication for the central NYS program and presents serum PFAS concentrations in the study cohorts and an assessment of potential exposure sources. In this study we examined demographic, behavioral, dietary, and other characteristics of survey participants as predictors of serum PFAS concentrations with a focus on locally caught fish consumption.

2. Material and methods

2.1. Participant recruitment and clinic visit

All study activities were approved by the federal Office of Management and Budget (Control Number 0923–0052) and the NYSDOH Institutional Review Board. To reach these “hidden” populations with no known sampling frame, the NYSDOH used respondent driven sampling (RDS) to recruit participants from each refugee population and the urban angler population (Heckathorn, D.D., 2011; Sabin, 2011; Liu et al., 2018a). The NYSDOH formed an advisory committee and partnered with local non-profit and faith-based organizations which played pivotal roles in program implementation such as identification of community members to serve as “seeds” who initiated the RDS recruitment, advised on the development of program materials, and provided venues for sampling activities and community outreach events. For the Burmese and Bhutanese refugee cohort, the partner organizations also provided trained interpreters to assist with all phases of the program.

RDS uses peer-recruitment within a socially and geographically connected population whereby each participant is asked to become a recruiter of new participants. To be eligible, potential recruits had to be 18–69 years old, have lived in the Syracuse, NY area for at least one year, and have eaten a minimum number of meals in the past year of locally caught fish (12 meals for the refugee cohort and 6 meals for the urban angler cohort). The minimum number of locally caught fish meals, determined per formative sessions with local organizations, represents a subjective minimum number of meals appropriate to capture frequent fish consumers in each subpopulation without being overly exclusive. Fish eaters were included regardless of whether they caught the fish themselves or ate someone else’s catch.

During the formative research phase, “seeds” were identified to participate in the program and to initiate recruitment. Each seed was given three coupons that were uniquely numbered so that they could be linked to his/her future recruits as these could not be a family member. There were no scheduled appointments; referral coupons listed options of event dates and times. A potential recruit brought their coupon to the sampling/data collection venue, was screened for eligibility, and if eligible, was invited to complete the consent form, the study questionnaires, height and weight measurements, and biological specimen collection procedures. Each participant was subsequently offered the opportunity and training to recruit up to three other potential participants.

As referrals arrived at the clinic venue, they were administered the eligibility survey, and if eligible were recruited and completed an informed consent form. Program participation consisted of completing an interviewer administered biomonitoring questionnaire, providing blood and urine samples, and having height and weight measurements taken. The biomonitoring questionnaire collected information on demographics, residential history (including time spent in refugee camps for the Burmese and Bhutanese cohorts), number of births and breastfeeding history for females, employment, education, family income, alcohol and tobacco use, and dietary information on fish and wildlife. Detailed fish consumption data included frequency of locally caught fish eaten in the past 12 months, parts of the fish eaten, and how fish were typically prepared. Participants were asked to identify specific bodies of water (Lake Ontario, Seneca and Oswego Rivers, and Onondaga Lake and tributaries) where the fish were caught (Supplemental Figures 1 and 2). Participants from the refugee cohorts were asked how often they ate fish caught from the specified local waterbodies by season and then were asked which species in the past 12 months they ate and how often. The urban anglers were asked to identify each fish species caught from the specified nearby waterbodies that they ate and how often in the past 12 months. Urban anglers were also asked how many years they have eaten fish out of these bodies of water. For the refugee cohort, culturally specific questions included the use of thanaka powder/cream (a potential source of mercury exposure), frequency of eating homemade or store-bought fish paste, and chewing betel nut (a mild stimulant). Prior to leaving the clinic, participants received a copy of the consent form, fish advisory brochures, and an explanation on healthy fishing and fish consumption. Participants received a gift card worth $75 for their participation. They were given an additional $15 gift card as a token of appreciation for each person they recruited and went on to participate.

2.2. PFAS measurements

The NYSDOH Wadsworth Center Organic Analytical Chemistry Laboratory measured twelve PFAS in serum by high-performance liquid chromatography (HPLC) with electrospray tandem mass spectrometry (MS/MS) after ion-pairing extraction (Kannan et al., 2004; Honda et al., 2018). The following PFAS were measured: PFOS, PFOA, PFDA, perfluorohexane sulfonic acid (PFHxS), perfluorononanoic acid (PFNA), perfluoroundecanoic acid (PFUnDA), perfluorododecanoic acid (PFDoA), perfluoroheptanoic acid (PFHpA), N-methyl perfluorooctane sulfonamide acetic acid (MeFOSAA), perfluorodecane sulfonate (PFDS), perfluorobutane sulfonic acid (PFBS) and perfluorooctane sulfonamide (PFOSA). Laboratory measurements were monitored for quality control and quality assurance, including the Centers for Disease Control and Prevention proficiency testing and quality control review. The laboratory method detection limits (LOD) for PFAS measures ranged from 0.0400 to 0.400 ng/mL (Supplemental Table 1).

2.3. Statistical analysis

Statistical analyses were performed with SAS® software version 9.4 (SAS/STAT software). The study aimed to assess exposure for two angler populations: (1) refugees from Burma and Bhutan; and (2) low-income, primarily African Americans persons. Whereas the refugees were enrolled as a single study population, descriptive analysis of lifestyle factors and PFAS results indicated notable differences between three ethnic groups. We analyzed data separately for four study cohorts henceforth referred to as the name in quotations that follow: Burmese persons of Karen ethnicity (“Karen”; n = 101), Burmese persons of other ethnicity (“other Burmese”; n = 84), Bhutanese persons (“Bhutanese”; n = 126), and urban anglers (“urban anglers”; n = 88). The analyses excluded three Karen, two Bhutanese and seven urban anglers for whom PFAS measurements were missing due to insufficient serum samples.

Percentiles and distribution-free 95% confidence intervals (CI) were calculated for each analyte using SAS PROC UNIVARIATE with CIQUANTDF option. Since our study participants were mostly aged 18–59 years, we used public use data files for the National Health and Nutrition Examination Survey (NHANES) 2015–2016 survey cycle to calculate reference PFAS levels for comparisons for adults 18–59 years [Calafat, 2012; NCHS, 2019a, 2019b]. The NHANES sample design is based on a complex multi-stage strategy, and the public use data file provides sample weights for analysis. We estimated the 50th and 90th percentiles and corresponding 95% CIs using statistical software packages SAS 9.4 and SAS callable SUDAAN® (Research Triangle Institute) and methods which are described in the National Center for Environmental Health’s (NCEH) Fourth Report (Appendix A) [CDC, 2009]. Consistent with NHANES guidelines, non-detects in our study data were assigned a value equal to the limit of detection (LOD) divided by the square root of 2. We also used Spearman’s rank correlation coefficients to examine the association between PFOS and total PFAS (the sum of the PFAS measures, excluding PFOSA and PFBS with 100% < LOD) and individual PFAS measures.

We conducted descriptive analysis of the questionnaire data to describe demographic characteristics, residential history, lifestyle factors, dietary intake, recreational activities, species and frequency of locally caught fish consumed, parts of fish eaten and preparation methods, store-bought fish consumption, and wildlife consumption. Total number of locally caught fish meals was calculated for the refugee cohorts by summing the frequency of reported past year seasonal local fish consumption. For the refugee cohorts, we also characterized locally caught fish consumption by fish meals per week in the past summer (June, July and August, “when it is hot outside”) as a binary variable (0 for 1–2 meals per week; 1 for 3 or more meals per week). Locally caught fish were eaten more frequently in the summer months, and most respondents reported a weekly frequency of meals consumed. Since questionnaire interviews were conducted starting in September, we assumed that reporting of locally caught fish consumption in the summer could be less influenced by recall bias. Urban anglers reported species-specific locally caught past year consumption and a fish consumption rate was calculated by summing the frequency for all species.

Exposure assessment for this article focused on whether eating locally caught fish or wildlife is associated with serum PFOS, PFOA, PFHxS, PFNA, PFDA, and PFUnDA concentrations, adjusting for other potential covariates and sources of exposure. These PFAS had a detection rate of 80% or higher in all cohorts, except for PFUnDA in the Bhutanese and urban angler cohorts. For modeling PFUnDA, random values from a uniform distribution between zero and the LOD were substituted for biomarker concentrations less than the LOD (Rocque and Winker, 2004). Preliminary analyses stratified by gender did not indicate that gender was an effect modifier for the association between a PFAS measure and past year local fish consumption. The preliminary analysis for females included number of months breastfeeding which was not found to be an independent predictor of any of the PFAS measures. Further modeling was performed for males and females combined and included gender as an independent variable. We evaluated demographic and other potential important predictor variables using bivariate analyses including analysis of variance for categorical variables and regression models for continuous variables. Important predictors identified in the bivariate analysis were included in each model to then assess fish and waterfowl consumption. Specifically, we examined the association between PFAS measures and age, gender, years residing in Syracuse, years residing in a refugee camp, eating fish viscera, eating fish from specific local waterbodies, number of times eating shrimp in the past year, chewing betel nut (refugees only), current smoker, eating specific wildlife in the past year (refugees only). Fishing location was examined for each of five specific waterbodies (Lake Ontario, Seneca River, Oswego Rivers, and Onondaga Lake including tributaries) as a bivariate variable (1 for yes vs. 0 for otherwise). We used stepwise selection after the inclusion of age and gender to identify potential confounders. For the stepwise selection, the default entry significance level of p < 0.15 was used. Multiple linear regression was used with age, gender, and potential confounders significant at the p < 0.10 level included in the model to evaluate the associations between each PFAS and local fish consumption. Local fish consumption was characterized as the number of meals in the past year. For the refugee cohorts, models were also run defining local fish consumption as 3 or more meals per week in the summer versus 1–2 meals as the reference. In addition to the total number of fish meals in the past year, eating fish species with a restricted local fish consumption advisory (yes/no for refugees and number of meals for urban anglers) was examined as an independent predictor of PFAS concentrations. A natural logarithm transformation was performed on analyte concentrations and past-year local fish meals to help normalize the distributions and satisfy the normality assumption for regression models. For each model, diagnostic plots were used to examine how the assumptions of linearity, normality, and homoscedasticity were met.

3. Results

3.1. RDS recruitment

Between September 2015 and January 2016, 311 refugees (including 7 seeds) enrolled in the project over 16 sampling events. A total of 711 referral coupons were distributed, of which 327 (46%) were redeemed. Of those redeeming the coupons and completing the eligibility screening survey, 23 were ineligible. The final sample was derived from six active seeds, with 58% of participants derived from one seed (all but one were Burmese participants), and 20% and 18% derived from the next two seeds (all were Bhutanese participants).

Between October 2015 and January 2016, 89 urban anglers (including 4 seeds) enrolled in the project over 12 sampling events. A total of 177 coupons were distributed, of which 100 (56%) were redeemed. Of those redeeming the coupons and completing the eligibility screening survey, 15 were ineligible. The final sample was derived from four active seeds, with 61% of participants derived from one seed and 33% of participants derived from a second seed.

3.2. Characteristics and lifestyle factors

The Burmese participants consisted of various ethnic groups born in Burma (87%) or Thailand (13%), and the majority were of Karen ethnicity (55%). The Bhutanese participants were born in Bhutan (92%) or Nepal (8%). For 90% of all refugee participants, Syracuse, NY was the only place of residency in the United States. The urban anglers were born in the U.S. and longtime residents of Syracuse.

Demographic and select lifestyle characteristics by study cohort are presented in Table 1. The Karen and Burmese of other ethnicities cohorts had similar demographic characteristics, being mostly female (75%), aged 18–39 years (61%), having resided in Syracuse for 5–8 years (60%), and being unemployed (73%). About 33% percent of the Burmese participants did not go to school, 44% had less than 8 years of school, but most were able to read (data not shown). In contrast to Burmese of other ethnicity cohort, a greater percentage of Karen participants spent less than 9 years in a refugee camp (48% vs. 22%), used thanaka cosmetic cream which is a potential source of mercury exposure (68% vs. 57%), chewed the stimulant betel nut (41% vs. 29%), ate fish paste (93% vs. 77%), and ate fish or shellfish in the past week (98% vs. 89%). About half of the Bhutanese refugee participants were female (52%), 60% were aged 18–39 years, 73% had resided in Syracuse for 1–4 years, 55% were unemployed, and 95% spent at least 17 years in a refugee camp. Thirty percent of the Bhutanese participants used chewing tobacco or snuff, and 43% reported using betel nut on most days. Compared to the Burmese cohort, the Bhutanese participants tended to have more education (47% had at least 9 years of school), and none used thanaka. Bhutanese persons culturally do not eat fish paste, and this questionnaire item was not asked. Of the 81 urban angler participants, 51% were female, 53% were aged 40–59 years, 65% had resided in Syracuse for more than 20 years, 38% currently smoked, and 33% were unemployed. Most of these anglers were non-Hispanic Black (80%), had lower incomes (53% annual family income less than $25k and 80% less than $50k), and 36% had at least some post-high school education.

Table 1.

Selected demographic and lifestyle characteristics of study cohorts.

Characteristic (number of missing responses)a	Burmese Karen (n = 99)	Burmese Other Ethnicity (n = 84)	Bhutanese (n = 124)	Urban anglers (n = 81)
	Number (%)b
Male	23 (23%)	24 (29%)	59 (48%)	40 (49%)
Age group
18–39 years	57 (58%)	54 (64%)	74 (60%)	27 (33%)
40–59 years	39 (39%)	26 (31%)	42 (34%)	43 (53%)
60 years or older	3 (3%)	4 (5%)	8 (6%)	11 (14%)
Women of childbearing age, 18–49 years	56 (57%)	50 (60%)	56 (45%)	28 (35%)
Years residing in Syracus
1–4 years	31 (31%)	34 (41%)	90 (73%)	7 (9%)
5–8 years	60 (61%)	50 (60%)	34 (27%)	10 (12%)
9–11 years	8 (8%)	0	0	2 (2%)
12–20 years	0	0	0	9 (11%)
21 years or more	0	0	0	53 (65%)
Years in a refugee camp (7-Karen, 7-Other Burmese, 16-Bhutanese)				Not Applicable
None	18 (23%)	6 (7%)	0
1–8 years	19 (25%)	14 (15%)	2 (3%)
9–16 years	25 (32%)	31 (34%)	2 (2%)
17 years or more	15 (20%)	41 (45%)	103 (95%)
Body Mass Index - 30.0 kg/m2 or more (Obese) (1-Bhutanese)	19 (19%)	17 (20%)	18 (15%)	41 (51%)
Unemployed (1-Karen)	72 (74%)	60 (71%)	68 (55%)	27 (33%)
Current smoker (1-Burmese/Other Ethnicity)	14 (14%)	7 (8%)	17 (14%)	31 (38%)
Use chewing tobacco or snuff (1-Burmese/Other Ethnicity)	8 (8%)	12 (15%)	37 (30%)	Not asked
Use betel nut	41 (41%)	24 (29%)	53 (43%)	Not asked
Use thanakar powder/cream	67 (68%)	48 (57%)	0
Eat fish paste	92 (93%)	65 (77%)	Not asked
Swim, dive or wade in specified local waterbodies (1-Bhutanese)	28 (28%)	22 (26%)	15 (12%)	18 (22%)
Ate fish/shellfish in the past week (1-Burmese/Other Ethnicity)	97 (98%)	75 (89%)	77 (62%)	71 (88%)
No	2 (3%)	8 (9%)	47 (38%)	10 (12%)
within past 3 days	82 (82%)	57 (69%)	33 (27%)	43 (53%)
4–7 days ago	15 (15%)	18 (22%)	44 (35%)	28 (35%)
Fish meals per week in the summer				Not asked
≤2	26 (26%)	33 (39%)	78 (62%)
3–4	44 (44%)	36 (43%)	41 (33%)
≥5	29 (29%)	15 (18%)	5 (4%)

^aA response of “don’t know” or “refused” is counted as missing.

^bPercentages are based on non-missing responses.

3.3. Fish and wildlife consumption

The refugee participants were asked to identify nearby waterbodies they had eaten fish from, and how they got fish from these waters – e.g., caught themselves, received from family or friends, or bought. Both male and female refugee participants received locally caught fish from family members or friends, while males were more likely than females to catch their own fish (74% vs 32% of the Karen cohort, 54% vs 35% of the other Burmese cohort, and 59% vs 29% of the Bhutanese cohort). The nearby waterbodies where the fish came from are presented in Supplemental Table 2. The Burmese participants tended to identify several waterbodies, most frequently citing the Oswego River (47%) or Onondaga Lake (38% of Karen participants and 58% of the other Burmese participants); however, a large percentage (33% of Karen participants and 25% of the other Burmese participants) said that they did not know. The Seneca River was the predominant local waterbody where the Bhutanese participants (84%) obtained their fish, and only 10% reported eating fish caught in Onondaga Lake. The urban angler participants tended to report getting fish from 3 or more of the local waterbodies listed, and unlike the refugee cohorts, 58% ate fish caught from Lake Ontario.

Participants were asked to identify which fish species from these areas they had eaten in the past 12 months, and the results are provided by the four study cohorts in Supplemental Table 3. All study cohorts reported eating channel catfish (30%–61%), common carp (15%–61%) and large (>15 inches) largemouth and smallmouth bass (21%–41%) which are species with a don’t eat consumption advisory for PCBs, mercury, and/or dioxin for Onondaga Lake and some other local waterbodies (NYSDOH, 2019). The Burmese cohorts ate more species that had a restricted consumption advisory compared to the Bhutanese or urban angler cohorts (Supplemental Table 3). Percentiles of locally caught fish meals in the past 12 months by study cohort showed that the Karen participants were the highest consumers with a median of 135 meals in the past 12 months compared to 103 meals for the other Burmese participants, 70 meals for the urban anglers and 44 meals for the Bhutanese participants (Fig. 1). For the refugee cohorts, the frequency of local fish consumption was collected by season. As expected, local fish was most often eaten in the summer months. The Burmese cohorts tended to eat local fish during the summer 3 or more times per week with almost 30% of the Karen participants eating local fish 5 or more times per week, while 62% of the Bhutanese participants ate local fish 1–2 times per week (Table 1). The Burmese participants tended to eat the smaller fish (hand size) whole with the head and the larger fish whole without the head. The Bhutanese participants mostly reported eating fish of all sizes whole without the head, and the urban anglers usually ate their fish filleted with or without the skin (data not shown). Unlike the other study cohorts, about 20% of the Karen participants also ate the offal of both small and large fish.

Fig. 1.

Percentiles of the number of locally caught fish meals in the past 12 months by study cohort.

Both the Karen and other Burmese cohorts were also frequent consumers of store-bought fish/shellfish with at least 90% eating shrimp on average 1 to 2 times per week (data not shown). Wildlife consumed by the Burmese participants in the past year included waterfowl (ducks or geese) (31%), deer (40%), rabbits (30%), and frogs/toads (31%) (Table 2). The only commonly eaten wildlife among the Bhutanese participants was deer (40%), and less than 5 urban anglers reported eating wildlife hunted in the study areas in the past year (Table 2).

Table 2.

Local wildlife eaten in the past year, Burmese and Bhutanese participants^a.

Species consumed	Burmese Karen (n = 99)	Burmese Other Ethnicity (n = 84)	Bhutanese (n = 124)
	Number (%)
Waterfowl (ducks or geese)	30 (30%)	27 (32%)	7 (6%)
Rats or mice	3 (3%)	2 (2%)	0
Crows or other scavenger birds	5 (5%)	5 (6%)	1 (<1%)
Squirrels	19 (19%)	14 (17%)	0
Deer	35 (35%)	37 (44%)	49 (40%)
Rabbits	39 (39%)	16 (19%)	19 (15%)
Frogs or toads	39 (39%)	18 (21%)	2 (2%)
Groundhogs	2 (2%)	4 (5%)	0
None	45 (45%)	41 (49%)	60 (48%)

^a< 5 urban anglers reported eating any wildlife.

3.4. PFAS biomonitoring results

PFOS, PFOA, PFHxS, PFNA and PFDA were detected in 88%–100% of samples within each study group (Supplemental Table 1). For PFAS serum concentrations, we examined the 50th and 90th percentiles by study cohort in comparison to NHANES, excluding PFDS (32% detects in the Karen cohort only), PFOSA (no detects), and PFBS (no detects) (Table 3). The 50th percentile for PFAS measures with the proportion of results below the LOD greater than or equal to 40% are not reported, since the 50th percentile is likely to be biased when the percentage of results below the detection limit is near or above 50% (Caudill et al., 2007). Among the Burmese cohort, serum concentrations of PFOS, PFNA, PFDA, and PFUnDA were at least 2 times higher than the NHANES comparisons. The Karen cohort had markedly elevated PFOS and PFDA levels compared to NHANES with median concentrations 9.5 times greater (41.6 ng/mL vs. 4.4 ng/mL) and 26.9 times greater (2.69 ng/mL vs. 0.10 ng/mL), respectively. PFNA median concentration was about 3 times greater than NHANES (1.54 ng/mL vs. 0.5 ng/mL). The other Burmese cohort had moderately elevated levels of PFOS and PFDA with median concentrations 3.0 times greater (13.3 ng/mL vs. 4.4 ng/mL) and 7.3 times greater (0.73 ng/mL vs. 0.10 ng/mL) than NHANES, respectively. At the 90th percentile, PFUnDA concentrations among the Karen participants were 15 times higher (3.32 ng/mL vs. 0.20 ng/mL) and among the other Burmese participants 5 times higher (1.32 vs. 0.20 ng/mL) than NHANES. PFAS levels were not elevated in the Bhutanese or urban angler cohorts compared to NHANES, except PFDA which at the median was 2.9 times greater (0.29 ng/mL vs. 0.10 ng/mL) in the Bhutanese cohort and 2.2 times greater (0.22 ng/mL vs. 0.10 ng/mL) in the urban anglers.

Table 3.

Serum concentrations of PFAS (ng/mL) for study cohorts, Syracuse, NY, 2015–16 and U.S. NHANES 2015–16.

PFAS (ng/mL)	Percentile	Burmese Karen Ethnicity (n = 99)	Burmese Other Ethnicity (n = 84)	Bhutanese (n = 124)	Urban Anglers (n = 81)	NHANES* (n = 1175)
PFOS	50th	41.6 (32.0–55.0)	13.3 (9.26–16.9)	5.40 (4.68–7.20)	5.12 (3.84–6.88)	4.4 (4.1–4.8)
	90th	119.6 (96.4–152)	46.8 (35.2–71.3)	20.0 (17.56–33.56)	20.8 (13.3–56.6)	11.2 (10.5–12.2)
PFOA	50th	1.21 (1.07–1.41)	1.18 (1.03–1.34)	1.05 (1.00–1.13)	1.42 (1.25–1.72)	1.471 (1.37–1.67)
	90th	2.29 (2.09–2.77)	2.97 (2.39–3.36)	2.27 (1.91–2.89)	3.07 (2.63–4.43)	3.17 (2.77–3.47)
PFHxS	50th	0.59 (0.49–0.66)	0.38 (0.36–0.47)	0.35 (0.32–0.40)	0.83 (0.60–1.04)	1.2 (1.0–1.3)
	90th	1.33 (1.12–1.77)	1.03 (0.75–1.17)	0.77 (0.66–1.23)	2.40 (2.08–3.78)	3.2 (2.6–3.5)
PFNA	50th	1.54 (1.34–1.96)	0.78 (0.64–0.91)	0.44 (0.39–0.47)	0.65 (0.56–0.78)	0.5 (0.50–0.60)
	90th	3.30 (2.84–4.26)	1.61 (1.51–2.78)	0.91 (0.80–1.4)	1.41 (1.12–2.16)	1.2 (1.1–1.4)
PFDA	50th	2.69 (2.12–3.20)	0.73 (0.53–0.91)	0.29 (0.24–0.33)	0.22 (0.17–0.29)	0.10 (0.10–0.20)
	90th	6.26 (4.92–7.36)	2.34 (2.18–3.42)	0.69 (0.62–1.23)	1.00 (0.55–2.40)	0.40 (0.30–0.50)
PFUnDA	50th	1.54 (1.25–1.69)	0.51 (0.40–0.65)	**	0.18 (0.14–0.21)	<LOD
	90th	3.32 (2.92–4.99)	1.32 (1.14–1.84)	0.38 (0.31–0.61)	0.56 (0.41–1.47)	0.20 (0.20–0.40)
PFDoA	50th	0.27 (0.20–0.31)	0.08 (0.70–0.10)	<LOD	<LOD	<LOD
	90th	0.67 (0.53–0.81)	0.34 (0.23–0.72)	0.08 (0.07–0.11)	0.08 (0.06–0.23)	<LOD
PFHpA	50th	0.05 (0.04–0.05)	**	**	**	<LOD***
	90th	0.12 (0.09–0.26)	0.32 (0.14–1.96)	0.13 (0.11–0.20)	0.14 (0.10–0.23)	0.10*** (<LOD-0.10)
MeFOSAA	50th	<LOD	<LOD	**	<LOD	<LOD
	90th	0.43 (0.27–1.28)	0.57 (0.39–1.20)	0.45 (0.36–0.71)	0.37 (0.23–0.80)	0.30 (0.20–0.40)

Bold type indicates estimates greater than twice the upper 95% confidence limit for the NHANES comparison.

^*ages 18–59;

^**Not calculated: the proportion of results below the limit of detection was too high (≥40%) to provide a valid 50th percentile.

^***NHANES 2013–14 (PFHpA was not measured as part of NHANES, 2015–16).

The NYS Wadsworth Laboratory limits of detection for the PFAS are: PFOS 0.2 ng/mL, PFOA 0.1 ng/.mL, PFHxS 0.1 ng/mL, PFNA 0.04 ng/mL, PFDA 0.1 ng/mL, PFUnDA 0.1 ng/mL, PFDoA 0.04 ng/mL, PFHpA 0.04 ng/mL, MeFOSAA 0.04 ng/mL; and for NHANES are 0.1 for all except for PFOS and PFOA (N/A for 2015–16).

Most PFAS were significantly correlated with one another, and the pattern of correlation strengths tended to be similar across study groups (Fig. 2). Total PFAS was highly correlated with PFOS (the predominant PFAS), with Spearman’s rank correlation coefficients ranging from 0.96 for the urban anglers to 1.0 for the Burmese cohorts. In addition, PFOS was consistently

Fig. 2.

Spearman’s rank correlation coefficients for the association between PFOS and other PFAS* by study cohort.

Strongly correlated with PFNA, PFDA and PFUnDA for all study groups, with Spearman’s rank correlation coefficients ranging from 0.73 to 0.92. The correlation between PFOS and PFOA ranged from 0.40 for the Bhutanese cohort to 0.60 for the Karen cohort.

3.5. Exposure assessment

Table 4 shows regression model results by study cohort for the association between PFOS and PFOA (natural logarithm [ln] transformed) and past year local fish consumption as a continuous variable (natural logarithm transformed) including age, gender, and variables selected via preliminary stepwise regression analysis. Since both variables were ln-transformed, the regression coefficient for past year fish consumption is the percent increase in the PFAS level for every 1% increase in past year fish consumption (University of Virginia, 2020). Regression coefficients (effect estimates) for the other variables were exponentiated for interpretation of results and represent the proportional change in the geometric mean of the PFAS concentration for a 1-unit change in continuous variables or relative to the referent group for dichotomous variables (e.g., males versus females). Males had higher levels of PFOS and PFOA in all study cohorts and male gender was consistently the strongest predictor of PFAS. Age did not show a consistent relationship with PFOS or PFOA among cohorts. PFOS increased with age among the Karen and urban angler cohorts, and PFOA increased with age among the Bhutanese cohort. Among the other demographic and lifestyle factors, a variety of relationships were observed. Among the Karen cohort, PFOS increased with an increase in the number of years living in Syracuse. Years residing in a refugee camp was associated with increased PFOS and PFOA among the other Burmese cohort. For the other Burmese participants, chewing betel nut was associated with higher PFOS, and PFOA increased with the frequency of eating shrimp. A positive association between the number of local fish meals in past year was indicated only for PFOS among urban anglers. The association between local fish consumption and either PFOS or PFOA was not statistically significant among the refugee cohorts, and the regression coefficients for PFOS indicated a non-statistically significant negative relationship among both Burmese cohorts. Regression models for total PFAS, PFHxS, PFNA, PFDA and PFUnDA had results similar to those reported in Table 4 for PFOS and/or PFOA (data not shown). For the refugee cohorts, models using a binary variable for local fish consumption (1–2 meals versus 3 or more meals in summer) also showed similar results (data not shown).

Table 4.

Results of multivariate regression models for serum PFOS concentrations and past year local fish consumption including age, gender, and covariates correlated in preliminary analyses (p < 0.1) by study cohort.

Karen
	PFOSa			PFOAa
Parameter	exp(beta)	P-value	Partial R2	exp(beta)	P-value	Partial R2
Age at interview (years)	1.024	0.0005	0.1336	1.007	0.07	0.0503
Male	2.142	<0.0001	0.1343	1.818	<0.0001	0.2402
Years residing in Syracuse	1.099	0.0074	0.0497
Number of local fish meals in past yeara	−0.014b	0.2158	0.0111	0.001b	0.98	0.0000
Total R2			0.3287		Total R2	0.2905
Other Burmese
	PFOSa			PFOAa
Parameter	exp(beta)	P-value	Partial R2	exp(beta)	P-value	Partial R2
Age at interview (years)	1.007	0.34	0.0478	0.995	0.35	0.0042
Male	2.371	<0.0001	0.0974	1.756	0.0002	0.1353
Years residing in a refugee camp	1.047	<0.0001	0.2175	1.016	0.03	0.0418
Chew betel nut	1.647	0.01	0.0434
Eat fish from the Seneca River	2.467	0.0003	0.0811
BMI	1.044	0.03	0.0220
Times eaten shrimp in past year				0.148b	0.01	0.0767
Number of local fish meals in past yeara	−0.147b	0.22	0.0106	−0.231b	0.01	0.0664
Total R2			0.5199		Total R2	0.3244
Bhutanese
	PFOSa			PFOAa
Parameter	exp(beta)	P-value	Partial R2	exp(beta)	P-value	Partial R2
Age at interview (years)	0.100	0.97	0.0068	1.012	0.002	0.0947
Male	1.851	0.0008	0.0602	1.430	0.0005	0.0896
BMI	1.779	0.01	0.0564
Number of local fish meals in past yeara	0.162b	0.17	0.0143	0.005b	0.94	0.0000
Total R2			0.1375		Total R2	0.1843
Urban anglers
	PFOSa			PFOAa
Parameter	exp(beta)	P-value	Partial R2	exp(beta)	P-value	Partial R2
Age at interview (years)	1.020	0.0137	0.1096	.0998	0.71	0.0023
Male	1.813	0.0065	0.0909	1.474	0.02	0.0453
BMI				1.025	0.03	0.0732
Number of local fish meals in past yeara	0.2020b	0.0215	0.0534	0.070b	0.29	0.0130
Total R2			0.2539		Total R2	0.1338

^aNatural logarithm transformed.

^bCoefficient not exponentiated for ln-transform dependent variable.

Scatter plots with linear regression lines for PFOS and past year local fish consumption are shown in Fig. 3 by study cohort and all cohorts combined overlaid. The effect of locally caught fish consumption as the primary exposure variable differed depending on subgroup, and an overall estimate of association is misleading. Overall, the regression line indicates a steep increase in PFOS with the number of local fish meals in the past year. This apparent association was driven mostly by aggregating data which included the Karen participants who had the highest level of both local fish consumption and PFOS concentrations. An inverse association, however, was found between locally caught fish consumption and PFOS within the Karen and other Burmese cohorts. Given the apparent effect modification by ethnicity/study cohort, the overall/combined result may not be meaningful.

Fig. 3.

Scatter plots of PFOS (natural logarithm) and number of past year local fish meals (natural logarithm) by study cohort, including linear regression lines. Ln(PFOS)_Karen = 4.318–0.134 × Ln(local fish meals in past year). Ln(PFOS)_{Other Burmese} = 2.683–0.023 × Ln(local fish meals in past year). Ln (PFOS)_Bhutanese = 1.330 + 0.110 × Ln(local fish meals in past year). Ln(PFOS)_{Urban anglers} = 0.563 + 0.299 × Ln(local fish meals in past year). Black line: Ln(PFOS)_{all cohorts} = 0.471 + 0.456 × Ln(local fish meals in past year).

3.6. Community outreach and education

Since the 1980s, NYS has provided healthy fish consumption advice to New Yorkers. Brochures and materials developed by the NYSDOH Outreach and Education Program have been informed by an understanding of fishing practices of anglers and their families, and an understanding of contaminants detected in fish in NYS waters. Over time, fish consumption outreach messages have also been informed by refugee populations to better represent cultural and traditional fishing practices, and fish consumption behaviors. Advisories are available in multiple languages, are posted on the NYSDOH website, and are available in print. For this biomonitoring program, the NYSDOH staff held community gatherings for all study participants and their friends and family. In addition to sharing personal results for select analytes one-on-one, these sessions were used to present aggregate results from the study. Detailed fish advisory information was presented, including color coded maps of recommended waters to fish from, fish consumption advice, and advice related to supermarket fish. The best methods of fish preparation to reduce exposure were also presented. Educational fact sheets including the NYSDOH fish advisory brochure for the region (NYSDOH, 2019) and the New York City Department of Health and Mental Hygiene’s Eat Fish, Choose Wisely Protect Against Mercury brochure (NYCDHMH, 2014) were made available. As information on emerging contaminants of concern, including PFAS, in fish has become available, NYS has continued to publish relevant advisories (NYSDOH, 2017). Although fish consumption advice available at the time of outreach conducted for this biomonitoring program was not specific to PFAS, NYS advisories are, in general, more restrictive with the aim to mitigate the overall risk of exposure to contaminants. Continued public health actions taken by the NYSDOH Outreach and Education Program have included translating the fish advisory into the Karen language (NYSDOH, 2019).

4. Discussion

PFAS substances have emerged globally as persistent organic pollutants in the environment which bioaccumulate in fish and wildlife (Lau et al., 2007; Buck et al., 2011; Kannan, 2011). PFOS is the most predominant PFAS found in fish samples with higher concentrations typically present in fish blood, followed by liver, brain, and muscle tissue (Becker et al., 2019) and whole fish compared to fillets (Fair et al., 2019). PFAS were detected in fish from all the Great Lakes with PFOS levels higher in fish from Lake Erie and Lake Ontario (Gewurtz et al., 2013). The NYS Department of Environmental Conservation (NYSDEC) also conducts fish sampling programs to evaluate PFAS concentrations in lakes and waterbodies throughout the state. A 2010–2018 NYSDEC report shows that PFAS concentrations vary by location and species, with PFOS the most prevalent PFAS followed by PFUnDA and PFDA and with the highest concentrations found in the viscera and whole fish (Becker et al., 2019). The NYSDEC fish sampling program also helped demonstrate that PFOS is highly bioaccumulative relative to PFOA. NYSDEC unpublished data for PFOS concentrations in fish had a mean of 39.8 ng per gram (ng/g) for Lake Ontario smallmouth bass collected in 2015 and a mean of 54.5 ng/g for largemouth bass collected from Onondaga Lake in 2018 (NYSDEC, 2018, NYSDOH, 2017; Fish contaminants database, Albany, NY). Unlike PCBs and mercury, PFOS concentrations in fish can change relatively quickly over time as PFOS sources to waterbodies change. The median PFOS concentration of Onondaga Lake surface water samples in 2004 was 756 ng/L (Sinclair et al., 2006) which suggests that fish concentrations at that time would have been much higher than the 2018 data. For example, applying a bioaccumulation factor reported in the literature (logBAF = 3.4; Bhavsar et al., 2016), fish in the lake at that time would have had a mean concentration of 1900 ng/g. Similarly, the U.S. Environmental Protection Agency reported that smallmouth bass from the Oswego River had the highest PFOS level in their national 2008–2009 survey with a mean of 127 ng/g (EPA, 2009b), while more recent NYSDEC sampling of other species from this waterbody suggests much lower levels. These apparent historical trends suggest that earlier exposures may have resulted in higher current serum levels in the study cohorts.

Human biomonitoring studies have measured PFAS to help inform on the risk of fish and shellfish consumption in both the general population and in targeted susceptible communities. The current biomonitoring study assessed PFAS concentrations among urban anglers and refugees from Burma and Bhutan who live in proximity to Onondaga Lake in Syracuse, NY and who tend to be high frequency consumers of locally caught fish putting them at potential higher risk of exposure. The Bhutanese and urban angler participants had PFAS concentrations consistent with the general U.S. population. The Burmese study participants had elevated PFOS, PFDA, PFUnDA levels in comparison to the U. S. population, and the Burmese persons of Karen ethnicity had substantially elevated levels of these contaminants and moderately higher levels of PFNA. Furthermore, the Burmese participants were more likely to eat the whole fish which tends to have higher concentrations of PFOS, PFDA and PFUnDA. These comparisons could underestimate exposure since PFOS tends to be lower among women due to its elimination during breast feeding and menstruation (Berg et al., 2014; Colles et al., 2020); 75% of the Burmese women were aged 18–43 years and had breast fed on average 61 months. In our study, male gender was the strongest predictor of PFAS concentrations in all study cohorts, and we did not find an association between past year fish consumption and any PFAS for each of the refugee cohorts. On the other hand, we did find a positive association between PFOS levels and locally caught fish consumption among the urban angler cohort among whom PFAS levels were not elevated compared to NHANES, except perhaps at the 90th percentile.

Whereas diet is accepted as an important source for widespread PFAS exposure, geographic location and race/ethnicity play a role and need to be better characterized (Zhang et al., 2019; Park et al., 2019; Wu and Kannan, 2019). In addition to the current study, several human biomonitoring programs have been conducted in the past decade in the Great Lakes area to assess exposure to both legacy and emerging contaminants of concern, including PFAS, among a diversity of susceptible populations (Wattigney et al., 2019). A population-based study of American Indians who live within the Lake Superior basin in Minnesota found all PFAS are either similar to or lower than both U.S. and First Nations Canada population comparisons (FDL and MDH, 2015). Furthermore, men eating greater amounts of local fish tended to have higher serum levels of PFNA, PFHxS, PFOA and PFOS, while women tended to have only higher PFNA in association with eating more local fish (FDL and MDH, 2015). The NYSDOH conducted a biomonitoring program in Western NY state in 2013–14 similar to our current study that likewise found substantially elevated PFOS concentrations among a community of Burmese refugees and immigrants living in Buffalo, NY who frequently ate fish from local contaminated waterbodies including Lake Erie, the Buffalo River and the Niagara River (Savadatti et al., 2019). In the Buffalo study, an association was found for consumption of more Great Lakes fish and higher PFOS and PFDA among the Burmese of “Other” ethnicity participants, but not among participants of Karen ethnicity (Liu et al., 2018b). Older male sport anglers in Wisconsin had PFAS levels comparable to the U.S. population, and all the PFAS examined, except PFHxS, showed a positive association with eating more locally caught fish (Christensen et al., 2016). Burmese refugees and licensed anglers in Milwaukee, WI were targeted as part of a 2017–2018 Great Lakes biomonitoring program, and both cohorts had moderately elevated PFOS levels similar to the non-Karen Burmese participants in our study (Christensen et al., 2019).

Perhaps a positive association between caught fish consumption and these PFAS was not found among the Burmese cohort due to the lack of ‘lower-end’ exposure or exposure variability (Ozkaynak et al., 2008). Alternatively, PFAS in human sera can be influenced by inter-individual toxicokinetics, elimination pathways, current exposures, historical exposures, several dietary sources, lifestyle, and other factors (Jain, 2014; Worley et al., 2017; ATSDR, 2019; Hu et al., 2018). Hu et al. explored identifying profiles of PFAS in human serum to provide information on major exposure sources and found that in the Faroe Islands, a North Atlantic fishing community, individuals exposed to PFAS from marine food had a higher proportion (relative to total PFAS) of long-chain perfluoroalkyl carboxylates (such as PFOS, PFDA, and PFUnDA) (Hu et al., 2018). These long-chain PFAS were also correlated with hair mercury as a proxy for seafood exposure (Hu et al., 2018). PFOS concentrations in Onondaga Lake surface water samples and fish samples from the Oswego River decreased considerably in the past 10–15 years which suggests that earlier exposures may have resulted in higher current serum levels in the study cohort (Sinclair et al., 2006; EPA, 2009b). Most (65%) of the urban anglers, however, lived in the area for at least 21 years and did not have elevated PFAS levels compared to NHANES. The Burmese participants ate considerably more locally caught fish but lived in the area for eight years or less. Therefore, it is difficult to interpret the positive association indicated between years residing in Syracuse and PFOS concentrations only among the Karen cohort. Given the markedly elevated PFOS concentrations among the Karen, we suspect that the association with years residing the Syracuse area (statistically significant at p = 0.007) may be confounded with an unknown factor.

The dietary importance of fish and seafood among the Burmese refugee study cohorts extends back to years spent in refugee camps before coming to the U.S. Almost all our Burmese study participants had resided in the U.S. less than 8 years, and dietary or environmental exposures while living in refugee camps prior to coming to Syracuse could also account for their elevated PFAS levels. Studies have shown that PFAS compounds are well absorbed orally, but poorly eliminated. Once absorbed, PFAS with long carbon chains such as PFOS can bioaccumulate in the human body with elimination half-lives from years to decades (ATSDR, 2019). The half-lives for these long carbon chain PFAS in humans have shown possible gender differences which are hormonally controlled (EPA, 2009a). Furthermore, PFAS will persist and bioaccumulate in the environment, and relatively low exposures over a long period of time can result in large body burdens. Most refugees from Burma were housed in camps located in Thailand (NCEZID, 2016). Recent reports document that PFAS have been poorly controlled in Thailand, including the Chao Phraya and Bang Pakong Rivers, and industrial wastewater was a major source of PFOS contamination in the water system (Boontanoon et al., 2013; EARTH, 2019). Thailand became a Party to the Stockholm Convention in 2005, and regulatory action to restrict PFOS went into force in Thailand in 2010. However, other PFAS are for the most part unregulated.

There were some limitations to our study. The initial seed participants were selected in a nonrandom fashion per RDS methodology. Although RDS has proved to be an effective methodology for recruitment of refugee populations (Liu et al., 2018a, 2018b), the final sample of refugees was stratified into three cohorts which may not adequately represent the corresponding populations. Urban angler participants were recruited using RDS mostly from only two seeds and study participants may not represent a random sample of the source population. Although survey data were collected by trained interviewers and interpreters, data quality could be limited by the participants’ recall ability, ability to understand questions, and fatigue given the extensive number of dietary questionnaire items. Another limitation is that PFAS biomonitoring indicates exposure from any and all sources over time. Our exposure assessment focused primarily on locally caught fish consumption in the past year which could represent only short-term consumption. Lastly, elevated PFAS concentrations relative to a comparison population are an indication of exposure and cannot be interpreted in the context of adverse health effects.

5. Conclusions

The Burmese subpopulation in our study, particularly those of Karen ethnicity, had elevated levels of PFOS and other PFAS compounds typically associated with consuming fish and seafood. Since 1999–2000, NHANES data have shown an apparent downward trend in PFOS concentrations in the U.S. population (CDC, 2021). More data are needed to characterize current PFAS levels in fish samples from Onondaga Lake and connecting waters where these communities fish. Burmese persons fish and hunt as part of their cultural identity of being self-sufficient (Schantz et al., 2010) and need for an economical resource for food. Given the dietary importance and nutritional benefits of fish consumption (Mozaffarian and Rimm, 2006; Turyk et al., 2012), the NYSDOH fish advisory program developed community outreach materials that supported continued consumption of sport fish. The NYSDOH focused their educational outreach to study communities on reducing exposure to contaminants by providing visual aids to assist with choosing less contaminated fish to eat, and fish species to avoid/limit. People were given color-coded maps for the region which showed local fishing waters to avoid in red and preferable fishing waters in blue. Specific fish consumption advice was also provided for women of childbearing age and children. Community events were held in collaboration with community partners and translators to help ensure effective delivery of health messages. NYSDOH will continue to develop culturally appropriate messages and conduct focused outreach among populations at disproportionate risk in NYS.

Our study adds to existing data on human exposure to PFAS and other contaminants in high-end fish consumers residing in the Great Lakes Basin. The Burmese subpopulation’s relatively recent residence in the area, however, made exposure assessment difficult. This suggests the need for additional research to better understand both ongoing and historic exposures. Future biomonitoring studies like ours could benefit by including measures of contaminants in potential exposure sources such as study participants’ dietary sources (e.g., fish, fish paste and local wildlife), cosmetics, and herbal medicines as well as more information on contaminants in areas of prior residence. Lastly, it is important to continue to develop culturally appropriate messages and conduct focused outreach among populations at increased risk of exposure.

Abbreviations

ATSDR

Agency for Toxic Substances and Disease Registry

CI

Confidence interval

LOD

Limit of detection

NCEH

National Center for Environmental Health

NHANES

National Health and Nutrition Examination Survey

NYS

New York State

NYSDEC

NYS Department of Environmental Conservation

NYSDOH

New York State Department of Health

PFAS

Per- and polyfluoroalkyl substances

PFCs

Perfluorochemicals

PFBS

Perfluorobutane sulfonic acid

PFDA

Perfluorodecanoic acid

PFDoA

Perfluorododecanoic acid

PFHpA

Perfluoroheptanoic acid

PFHxS

Perfluorohexane sulfonic acid

PFNA

Perfluorononanoic acid

PFOA

Perfluorooctanoic acid

PFOS

Perfluorooctane sulfonic acid

MeFOSAA

N-methyl perfluorooctane sulfonamide acetic acid

PFUnDA

Perfluoroundecanoic acid

PFDS

Perfluorodecane sulfonate

PFOSA

Perfluorooctane sulfonamide

RDS

Respondent driven sampling