Research Grants/Fellowships/SBIR

An Epidemiologic Study of Time Trends and Health Effects of Persistent Organic Pollutants, Mercury and Micronutrients

EPA Grant Number: R833705
Title: An Epidemiologic Study of Time Trends and Health Effects of Persistent Organic Pollutants, Mercury and Micronutrients
Investigators: Berner, James E.
Institution: Alaska Native Tribal Health Consortium
EPA Project Officer: McOliver, Cynthia
Project Period: June 1, 2009 through May 31, 2013 (Extended to May 21, 2014)
Project Amount: $948,121
RFA: Issues in Tribal Environmental Research and Health Promotion: Novel Approaches for Assessing and Managing Cumulative Risks and Impacts of Global Climate Change (2007) RFA Text |  Recipients Lists
Research Category: Environmental Justice , Global Climate Change , Tribal Environmental Health Research , Health , Climate Change


This research will address cumulative exposure to multiple environmental stressors in rural Yupik Alaska Natives (AN). The stressors to be investigated are anthropogenic persistent organic pollutants (POPs) and mercury (Hg), accumulated in pregnant Yupik residents through ingestion of traditional subsistence wildlife species, particularly salmon and seals.  This set of stressors may be associated with adverse health outcomes, but the threat of illness caused by the traditional foods that have sustained the Yupik people for thousands of years, and which form a central part of the culture is very threatening to the Yupik worldview.  This has caused some Yupik residents of the Yukon-Kuskokwim River Delta (YKD) to decrease the use of traditional marine subsistence species, and increase western foods where possible, and particularly high sugar and fat containing choices, such as soft drinks, chips, and solid shortening agents.

The concern with safety of traditional foods caused the Alaska Native Tribal Health Consortium (ANTHC) and the Yukon-Kuskokwim Health Corporation (YKHC), the tribal health organizations responsible for health care in the YKD, to initiate a pilot study of maternal and newborn blood levels of POPs, metals, and micronutrients (CDC protocol 2320) starting in 1999. The U.S. Fish and Wildlife Service (FWS) carried out a study in 2001, to determine tissue levels of POPs, heavy metals, omega-3 fatty acids, and selenium in the two most commonly consumed salmon species in the Kuskokwim and Yukon Rivers, which are the drainage for the entire YKD. 


This research will extend and fully analyze a previously obtained data base of human tisse and salmon tissue for POPs, metals, and micronutrients. The expasion of the data base will accomplish the following objectives:

(1) Assure an adequate sample size of Yupik women and infants to document any associations between prenatal exposure to POPs and/or Hg, with adverse health outcomes and document any association between micronutrients and positive outcomes.

(2) Provide a time series in human and subsistence species tissue levels of POPs and Hg to evaluate the impact of climate regime change on ocean and atmospheric current delivery of POPs and Hg from lower latitudes into the Bering Sea, addressing possible climate change impact on subsistence resources and human health.

(3) Provide and communicate risk and benefit data to Yupik residents to enable the development of region-wide strategies and policies to reduce the risk, and inrease the benefits of the traditional diet, which will strengthen that critical component of Yupik culture.

The research hypotheses will specifically address the following required topics:  (a) subsistence resources, (b) subsistence practices, (c) chemical exposures, (d) the differential impact on sensitive sub-populations, and (e) potential impact of climate change on subsistence resources and human health.

The first objective has two parts:  the collection of a representative sample (~40 salmon) of the two most commonly consumed salmon (Chinook salmon [Oncorhynchus tshawytscha] and chum salmon[O. keta]) species in the YKD, and the recruitment of 200 pregnant Yupik women for the human cohort study.



Objective 1 has two parts, collection of a representative sample of the two most commonly consumed salmon species in the YK.D, and recruitment of 200 pregnant Yupik women. The salmon study will be described, followed by the human cohort study.

The Salmon Study:
The FWS collected baseline data on contaminants in Chinook salmon (Oncorhynchus tshawytscha) and chum salmon (0. keta) from the Yukon and Kuskokwim rivers in 2001(10). We propose to analyze potential increases in contaminant concentrations by repeating salmon data collections in 2008. Fish are the largest single subsistence resource harvested by rural residents in Alaska, representing approximately 60% of the total harvest (11). The mix of subsistence foods harvested differs across communities. Salmon, halibut, herring, whitefish, sheefish, blackfish and cod are among the primary fish species harvested for subsistence in Alaskan communities(12). Chinook salmon and chum salmon are the two most heavily used salmon species in the Yukon River(13) and the Kuskokwim River(14). In 1989, chinook salmon represented 44% (by weight) of all salmon harvested for human consumption on the Yukon River, and 64% in 1998(15). The relative importance of chinook and chum in the local diet varies between communities and may change in relation to total fish abundance(15). For example in 1989, a year of high salmon abundance, approximately 61% of the fish harvested for human consumption in Tanana were Fall chums. Subsistence use differs not only among commtmities, but among families within a community, and among individuals within a family.

Fall chum are an important subsistence food resource for people living in communities along the Yukon River, and this run is needed to provide a complete picture of contaminants in Yukon River salmon consumed by humans. Subsistence harvest in the Kuskokwim River in 1996 was estimated at 78,729 and 89,430 chinook and chum salmon, respectively (14) and the total subsistence salmon harvest for the region exceeded 240,000 fish. Maximum size and trophic position were important factors in salmon species choice. In arctic and subarctic ecosystems, tissue concentrations of persistent contaminants are related to trophic position (animals higher in the food web typically have higher levels of these pollutants). Among Pacific salmon species, chinook and chum salmon are the largest in size, spend a longer time at sea (potentially accumulating contaminants for a longer period), and at least for chinook salmon, likely feed at a higher trophic level than coho, pink, and sockeye salmon. Chinook and chum differ considerably in size, life history, food preferences, and geographic range, all of which may influence exposure to and accumulation of environmental contaminants. Chinook salmon are the largest of all Pacific salmon, with weights of individual fish commonly exceeding 14 kilograms in the Yukon River. Fish dominate the diets of ocean-phase chinook salmon (16). They grow quickly and reach sexual maturity after two to seven years (17). Chum salmon are considerably smaller, with a mean weight of 3.17 kg in the Yukon and Kuskokwim rivers ( 16), and reach sexual maturity after three to six years (18). Food preferences for chum salmon in the Bering sea include amphipods, euphausiids, pteropods, and copepods; fish; and squid larvae (16). These differences in fish size, time spent at sea, selection of prey items and geographic range all may potentially influence exposure to contaminants. Lipid concentration also affects retention of persistent organic pollutants, since these chemicals partition preferentially to lipids. Pacific salmon typically do not feed after they enter fresh water and fat stores are progressively used as they travel upriver to spawning areas. The Yukon River is the third longest river in North America and is the longest river in Alaska. To support an extended migration to upriver spawning sites, many fish stocks within the Yukon have a high lipid content relative to some other salmon stocks in the State. Within the Yukon River system, salmon that spawn in the upper reaches have greater lipid concentrations than those spawning further downstream, thus up-river spawners presumably may have higher burdens of some fat-soluble persistent organic chemicals. Due to their ecological and physiological characteristics, up-river spawners may be at increased risk from persistent bioaccumulative contaminants, and are of particular interest for initial data collection, since many are harvested long before they reach their spawning site. Partitioning and use of fat stores in migrating and gametogenic fish can affect contaminant concentrations in those fish, and potentially affect both availability of contaminants to human consumers and the release of contaminants to ecosystems. Salmon entering rivers contain fat stores which are progressively used as they travel up river to spawning areas (18, 19). Yukon River salmon that spawn in the upper portions ofthe Yukon River drainage have greater lipid concentrations when they enter the river than those spawning further downstream (18), and therefore potentially greater concentrations of lipophilic contaminants. Since many persistent organochlorine contaminants are lipophilic, they are liberated as fat is metabolized and may concentrate in the remaining fat (20). Ewald et al. (19) observed that as sockeye salmon migrated upstream on the Copper River, Alaska, lipids were reallocated from muscle to female gonads concurrent with a decrease in muscle lipids from 5.5% to 2.2%. DDT and PCB concentrations in muscle lipid increased ~550% and ~370%, respectively, and organochlorine concentrations in roe approximately doubled (19). By measuring tissues from male and female fish, and eggs, we will determine if there are gender-specific differences in elimination or reallocation of contaminants, especially due to elimination into eggs by females. This will allow species, and sex-specific stratification of potential hwnan exposme to POP's, as many AN residents consume salmon eggs.

Sample Collection
Chinook and chum salmon will be collected at Rampart Rapids (65°20'00"N, 151°04'30"W), a productive subsistence fishing spot on the Yukon River, and at Bethel (60°47'30"N, 161°45'00"W) on the Kuskokwim River. At both sites, fish will be collected in cooperation with subsistence users. Fish will be collected using gill nets or fish wheels. Fish will be dissected using clean techniques; samples will be collected and frozen as soon as possible, and the body of the fish will be returned to subsistence users. Twenty fish (10 male, l0 female) of each species will be collected at each sample site. The sample size of twenty fish per species per sample site should provide sufficient data to characterize the sample variance for comparison to published effect concentrations of organochlorine, metal, and metalloid contaminants. Fatty acids will be measured in muscle tissue. Samples and analyses proposed for chinook and chum salmon on the Yukon River and Kuskokwim River, Alaska, 2008. Sampling for Chinook and chum salmon will begin in mid-June at Bethel, and Chinook sampling will begin in mid-June at The Rapids. Sampling for chum salmon on the Yukon River will begin during mid-July, as it is a fall run. Fish will be weighed (to nearest g), measw-ed (mid-eye to fork length and mid-eye to total length, to nearest mm), and receive a careful external examination for parasites and abnormalities. Muscle and egg samples for mercury and other metals will be collected using acid-cleaned stainless steel instruments, then transferred to new acid-precleaned polyethylene bottles. Liver, muscle, and egg samples for organochlorine analysis will be collected using acetone/hexane-cleaned stainless steel instruments and transferred to new precleaned glass jars (I-Chem Series 200 or equivalent). A thorough internal examination, for Ichthyophonus infection and organ condition, will occur concurrently with the tissue dissections. All excised tissues will be weighed prior to subsampling for the various analyses. Samples to be frozen will be stored in a freezer ( ~-10°C) in the field and an ultralow (-40°C) freezer in Fairbanks until shipped (frozen, ovemight) to analytical laboratories. The exterior and interior of each fish will be photographed. Vertebrae or scale samples will be collected to determine fish age.

Analytic Methods for Salmon Tissue
All chemical analyses will be conducted at a U.S. Fish and Wildlife Service contract laboratory. All analytical methods, quality assurance, and quality control procedures will meet U.S. Fish and Wildlife Service contractual standards (; Appendix A).

Organochlorine analysis includes percent lipid, pp'-DDE, alpha BHC, gamma chlordane, pp'DDD, beta BHC, cis-nonachlor, pp'-DDT, gamma BHC, trans-nonachlor, op'-DDE, dieldrin, endrin, op'-DDD, heptachlor epoxide, mirex, op'-DDT, oxychlordane, toxaphene, HCB, alpha chlordane, PCB-TOTAL, and PCB congeners, and brominated flame retardants, and PFOS. Generally methods involve solvent extraction (Soxhlet for tissues and sediments/soils, and mixing for water), cleanup (Gel Permeation Chromatography and/or Florisil column chromatography, plus carbon column for planar PCBs), separation of pesticides from PCBs (Silica gel column chromatography), and quantification by electron capture capillary gas chromatography and confirmation by mass spectrometry. The Limit of Detection is the "Method Limit of Detection" as described by the Environmental Protection Agency in 40 CFR Part 136, Appendix B. In summary, the limit of detection is defined as the Student's t for 99% confidence times the standard deviation of seven replicate measurements of the same low level sample.

Metal analysis includes percent moisture, arsenic, selenium, mercury. Methods include Graphite Furnace Atomic Absorption (CLINS lO5A-E, 205A-E, 305A-E, etc.) (arsenic and selenium), Cold Vapor Mercury Atomic Absorption (CLINS 106A-E, 206A-E, 306A-E, etc.) (mercury), Inductively Coupled Plasma Emission Spectroscopy (CLINS 107 A-E, 207 A-E; 307A-E, etc.) (all other metals), and ICP-Mass Spectrometry (CLINS 108A-E, 208A-E, 308A-E, etc.) (for confirmation). Metal detection limits vary by analyte; the most current can be found at: inorg sow.btml.

Data analysis will be multivariate whenever possible. Methods include Analysis of Variance (ANOVA), Analysis of Covariance (ANCOVA), multiple regression, Piincipal Components Analysis (PCA), and other univariate and multivariate techniques. Because specific statistical analyses must account for variable percent detections among groups with appropriate techniques (non-parametric or data reduction methods), actual analyses will be based on the percent detections in the data set. Analyses will utilize sample percent lipid, age or size of fish, and stable isotope ratios as covariates when these are significantly correlated with contaminant concentrations. Specific analyses include: 1) tests for differences between species, sexes, and rivers, and 2) time series analyses comparing contaminants concentrations between 2001 and 2008. Contaminant concentrations will also be compared to published values to determine if concentrations exceed effect thresholds.

The Human Cohort Study: In 1998, rising concern around contaminants in the traditional marine subsistence wildlife prompted the YKHC to request that the ANTHC Division of Community Health Services (DCHS) conduct an analysis of the exposure of the Yupik residents to POPs and heavy metals, and to determine, if possible whether adverse health effects were associated with current exposure levels, and to determine if the micronutrients, particularly 03 FA and selenium (Se) were associated with positive health effects. Since pregnancy and infancy were felt to be the most sensitive life stages, it was elected to focus exclusively on pregnant women and the frrst 12 months of life (the infant period.) The basic sampling protocol of the Arctic Council Arctic Monitoring and Assessment Program (AMAP) was selected as a template. The AMAP Human Health Advisory Group (HHAG) had completed a preliminary assessment of POPs and metals exposure in maternal blood of residents of the Arctic nations, (22) but the US had not fully participated. As Alaska is the only Arctic territory in the US and because AN rural residents are highly subsistence dependent, ANTHC bec~"tle the US participant in the AMAP HHAG, and the comparison levels of AN YKHC pregnant women, compared to representative maternal cohorts from the other Arctic nations, are presented in Table 1. The essential elements of the CDC protocol 2320 have been in continuous use since 1999. Protocol elements are: recruitment - pregnant Yupik women age 18 years of age or greater are offered an opportunity, by a Yupik RN, employed by ANTHC Division of Community Health, in the YKHC prenatal clinic at the first prenatal visit. If interested, the woman is given information on the program, and if she chooses to enroll, signs a consent to have an additional three tubes (15cc) of blood drawn as her routine prenatal labs are obtained, two 5cc clot tubes and one 5cc anticoagulated tube and medical record reviews for her infant and herself. She fills out a brief dietary history focused on types and frequency of consumption of the subsistence species in the YKD, cunently, in the past year, and past 10 years. This survey has been in use, and approved as part of the protocol since 1999. At delivery, 15cc of blood similar to the maternal sample are obtained from the umbilical cord, after it is clamped and cut. The blood is analyzed for POPs and metals, 03 FA, and Se as presented in Table 1.

Table 1: POPs in Maternal Blood (micrograms/Kg lipid)1 Geometric Mean Values

GR InuitCA InuitNorwaySwedenIcelandFinlandRussia
Tox P501.932
Tox P261.382

1 - All values not footnoted are from Van Oostdam, J.C., et. al. Science of the Total Environment 330:2004, 55-70.
2 - Values from Alaska Native Traditional Food Safety Monitoring Program; 3 10-woman pooled specimens from 2000-2002 cohort. J. Berner, personal communication.
3 - Values from Guvenius, D.M.; Aronsson, A; et. al. Environmental Health Perspectives, 111:9, July 2003.

With the information in the FWS 1999 salmon study, toxaphene was added to cohort 2 analyses, as was brominated flame retardants (BFRs) and perflorinated octane sulfonate compounds (PFOS). There was insufficient volume of blood left for each woman in cohort 1 to have the additional analyses, so two pooled specimens, representing river village residents, and coastal residents were created, with results presented in Table 1. The flrst cohort was enrolled between 312000 - 5/2003 with 144 women. A second cohort of 204 women was enrolled 7/2004- 5/2006, with the last birth in that cohort in 1/07. The samples from this cohort will be analyzed by CDC DLS by 4/07. Throughout the 1999-2007 time period the CDC DLS, and the CDC HSB of National Center for Environmental Health (NCEH) have been collaborators and in-kind funders with funding support from the EPA office of International Activities, through IAGDW75-93818701 from 1999-2005. The funding level did permit systematic record review of the first cohon, but the agreement ended before the second cohort could undergo record review and data abstraction. The sample size used by other investigators looking for POPs and Hg health effects in prenatally exposed indigenous Inuit mothers and infants in Arctic Canada has varied from 350-600, which has usually provided adequate statistical power to document adverse effects. Health effects so far associated with prenatal POPs exposure in the highest quartile include a small increase in risk for acute otitis media (AOM), and lower respiratory tract infection as well as neurobehavioral effects detected in infants exposed in the highest quartile of matemal blood in some studies (23). Blood levels in Inuit maternal specimens, for Aroch1orl260, (Arochlor 1260 is 5.2 (PCB138+ 153), mean levels varied from 2.4 micrograms/liter plasma in 31 women in Inuvik, to 8.0 micrograms/liter of plasma in 30 women in Baffin, Canada. For comparison, the mean Arochlor 1260 level in 153 Yupik women was 1.5 micrograms/liter of plasma. Exposure levels are low enough that neurocognitive deficits, based on prenatal POPs exposure would not be expected in Yupik infants (24,25). Prenatal Hg levels in the YKD show a mean of 6. l micrograms/liter with a maximum of 14.0 micrograms/liter which is far below the EPA BMLD for prenatal blood exposure, of 44 microgramslliter, it is well within the 1-10 micrograms/liter range shown to increase childhood blood pressure in the Faroe Island and Japanese cohort (8,9). For these reasons, as well as the huge cost of flying psychologists frequently to the 55 YKD villages to test cohort infants and children, to duplicate studies done in the Faroe Islands, detailed developmental and neurophysiologic tests will not be possible. Standard tests of growth and development, perfomed by field physicians from YKHC, and State of Alaska Public Health Nurses will be used to follow these parameters. Blood pressure, beginning at age 4 years, will be taken by the State of Alaska Public Health Nurse (PHN), trained by the PI. During the PHN's regularly scheduled village visits, they will use pediatric blood pressure device (General Electric Dynaprocare Monitor) to assure reliable, BP measurement. The AN infant population of the YKD bas one of the highest rates ever reported of hospitalization in the first year of life for severe lower respiratory infection (7). Most of these infections are viral, 50% being due to Respiratory Syncytial Virus (RSV). The frequency of hospitalization most years is between 120-150/1000 infants. The reasons for this high rate are not known. Factors such as crowded housing and smoking rates could all be factors. The rate of this event makes it possible to use risk for hospitalization with lower respiratory infection as an adverse outcome to be investigated for association with prenatal POPs exposure. In addition, AK children in Southwest Alaska (YKD) have 5% incidence of congenital heart disease, 2.5 times greater rate then non-Native infants (6), even when adjusted for alcohol and tobacco use. Among Yupik residents of YKD, the prevalence of type 2 Diabetes is 20/1000 population (26). The risk of type 2 DM is higher in US NHANES III participants with higher POPs exposure (27), and the possible association between YK maternal POPs levels and risk for DM will be examined. In addition, all complications of pregnancy including preterm delivery will be examined for association with POPs and Hg, as the rate of preterm delivery is 14.7/1000 births. Each maternal participant under goes a complete standard prenatal history and physical examination and a two page subsistence-focused dietary survey as previously described. The participants have blood for the POPs, Hg, and 03FA levels drawn with the same venipuncture as the initial prenatal lab studies.

After delivery a maternal record review (described and approved in CDC protoco1 2320) is carried out, and a data abstraction form is completed. After the infant is 12 months old a medical record review and data abstraction form is completed. The two prior maternal and infant cohorts will be combined with the 200 additional women and infants recruited in this proposal. The total of approximately 550 women and infants will be examined by a within-cohort comparison for any increased risk for adverse outcome with POPs and/or Hg exposure in the highest quartile of exposure, compared with the lowest quartile. Confounders such as smoking, alcohol use, obesity, family history will be statistically analyzed. The approximately 350 infants who will between 4-8 years old during the grant period will be evaluated as a cohort to look at BP in relation to prenatal Hg exposure, adjusting for family history or other risk factors, such as extreme prematurity.

Statistical Analvsis of Human Data: The POPs and Hg levels from cohort 1 have log normal distributions. It is expected that this will be the case for cohort 2, now being analyzed, and cohort 3, which will be recruited if this proposal is funded. The concentrations will be presented as geometric means with values below the limits of detection (LOD) given a value of 0.5 LOD, as is the standard in the AMAP HHAG. Poisson regression will be used to analyze the incidence rate ratio (IRR), with the number of infections requiring hospitalization in the first 12 months of life used as the dependent variable and the three POPs with the highest geometric mean (GM) concentration, as well as Hg used in individual analyses as the independent variable. The infants in the lowest quartile will be examined against the infants in each of the upper three quartiles. Confounding factors will be tested by multiple regression, and those influencing IRR by greater than 5% will be included. The same approach will be used for the other infant and maternal outcomes to be analyzed. Statistical analysis of the 4- 7 year old children to evaluate the effect of prenatal Hg exposure on BP will utilize the approach of Sorensen, et al. (7) in the Faroe Islands cohort.

Expected Results:

The primary results of the proposal will be: 1.) New salmon contaminant and nutrient data. 2.) New human tissue contaminant, Hg, and nutrient data. 3.) New pediatric blood pressure data. 4.) New subsistence dietary data for YK mothers. The benefits will be: 1.) The first risk assessment of exposure to POPs and Hg from dietary sources for any AN population. 2.) The first time trend data on human tissue levels of POPs and Hg for any ethnically specific US population. 3.) The first region species-specific salmon POPs and Hg time trend data in the Bering Sea. 4.) The first risk assessment in any Native American children for chronic disease related to prenatal dietary exposure to Hg, or DM in AN women related to POPs exposure. Outcomes: The benefits listed will allow the State of Alaska, ANTHC and the YKHC to formulate subsistence diet risk reduction strategies, and strategies to take advantage of any health benefits associated with micronutrients. This would allow specific strategies to reduce exposure for the sensitive subpopulation of AN, including infants, female children, and women who are planning pregnancy, already pregnant, or at risk for pregnancy. Such strategies could include reduction in certain foods which might include items such as salmon eggs, eatjng larger and older marine mammals. It might include increasing land mammal intake, increasing salmon intake, and limiting use of seal oil. Findings of risk of higher blood pressure in children exposed to higher levels of prenatal Hg would warrant a BP surveillance protocol for all Y-K children, something not now done routinely. This finding if present might be related to the known risk of stroke mortality in AN, twice the US all races rate (28) and thus has potential to address this dispariJy in mortality.

Finally, salmon tissue POPs and Hg reflects the levels in the trophic levels at which they feed, and at the lowest levels, these lower level organisms reflect the seawater POPs and Hg levels. Seawater receives river and atmospheric deposition, and there is concern that the recent increase in global mean temperature has increased volitilization of POPs, and air current transport to northern latitudes. Increased river discharge in mid and lower latitudes has increased from increased precipitation secondary to more frequent E1 Nino events. Thus, it is quite possible that the time trend in salmon may show increases in toxaphene, Hg and PBDEs, all of which are rising in the western Arctic. Findings like these, if present, would furnish both human health and wildlife agencies in the US, Canada, and internationally to advocate more forcefully for more stringent, effective controls on production and release of these compounds and for institution of measures to reduce global warming. The data gained in the proposal will be shared with the YKHC, villages, State and Federal wildlife and hwnan health agencies, and the AMAP HHAG.

Publications and Presentations:

Publications have been submitted on this project: View all 7 publications for this project

Progress and Final Reports:
2009 Progress Report
2010 Progress Report
2011 Progress Report
2012 Progress Report