PUBLIC HEALTH ASSESSMENT
SANGAMO/TWELVE-MILE CREEK/HARTWELL PCB
PICKENS, PICKENS COUNTY, SOUTH CAROLINA
The tables in this section list the contaminants of concern. Weevaluate these contaminants in the subsequent sections of thisassessment and determine whether exposure to them could have aneffect on the public's health. These contaminants are selectedand evaluated according to the following factors:
1. Concentrations of contaminants on and off the site.
2. Field data quality, laboratory data quality, and sampledesign.
3. Comparison of on- and off-site concentrations withbackground concentrations, if available.
4. Comparison of on- and off-site concentrations withpublic health assessment comparison values for (a) noncarcinogenic endpoints and (b) carcinogenic endpoints.
5. Community health concerns.
Contaminants that are listed in the data tables for this sectionshould not be interpreted to mean that they will cause harmfulhealth effects. Rather, the list identifies contaminants thatneed to be evaluated further in later sections of the assessment. When a contaminant of concern is identified in one medium, itwill also be mentioned in all media.
The data tables contain the following acronyms:
CREG = Cancer Risk Evaluation Guide
EMEG = Environmental Media Evaluation Guide
MCLG = EPA Maximum Contaminant Level Goal
MCL = EPA Maximum Contaminant Level
ppm = parts per million
ppb = parts per billion
ppt = parts per trillion
Contaminants requiring further evaluation are selected by usingmedium-specific, comparison values. These values include EMEGs,CREGs, and other relevant guidelines. CREGs are estimatedcontaminant concentrations based on the probability that 1additional cancer case may occur in excess of the number thatwill be expected to occur among 1 million people (assuming theyhave been exposed to the contaminant for a lifetime). CREGs arecalculated from EPA's cancer slope factors. EPA's MCLG is adrinking water health goal. EPA believes that the MCLGrepresents a level at which no known or anticipated harmfuleffects to people should be expected to occur. It allows for anadequate margin of safety. MCLs represent contaminantconcentrations that EPA deems protective of public health;however, in deriving these levels the EPA takes intoconsideration the technical feasibility and economics of watertreatment. While MCLs are regulatory concentrations, MCLGs arenot.
1. Ground Water
In the 1989 Remedial Investigation Study (RI) for OU-1, volatileorganic compounds (VOCs) were found in ground water near thewastewater treatment facility and at two other isolated areas onthe plant site. These findings will not be discussed furtherbecause this public health assessment focuses only on OU-2. Apublic health assessment on OU-1 will be conducted at a latertime.
To date, no groundwater contamination has been associated withOU-2. It is unlikely that local ground water will becontaminated by PCBs from this site because of the demonstratedability of these compounds to bind tightly to sediments and soils(1). This concern was raised by a local resident who used aprivate drinking water well within 500 feet from Twelve MileCreek. There were no PCBs or volatile organic compounds detectedin a sample taken from this well (4).
2. Surface Water
a. Town/Twelve Mile Creek:
Limited ambient surface water sampling has been conducted. PCBshad been detected in EPA's surface water samples collected fromTown Creek in 1976; however, PCBs were below the levels oflaboratory detection in the samples collected from Twelve MileCreek and some of its tributaries. From May 1988 to May 1991, 5surface water samples were collected from Twelve Mile Creek atSCDHEC sampling stations SV-107 (where Twelve Mile Creek joinsthe Seneca River Arm of Lake Hartwell) and SV-282 (in the upperpart of Twelve Mile Creek near the town of Pickens, SC). Allresults were below detectable levels; the level of detection forthe SCDHEC laboratory was 0.5 parts per billion (ppb).
b. Municipal Drinking Water:
Data results from the monitoring program have fluctuated greatlybetween most sampling periods. Since 1983, PCBs have beenperiodically detected in drinking water samples at most samplingpoints; however, PCB concentrations at these same locations oftenreturned to levels below laboratory detection (that is, less than0.1 ppb) by the next sampling period.
Samples were often collected at taps both before allowing waterto run through the distribution lines (pre-flush) and after thewater was allowed to run for several minutes (post-flush). However, this sampling strategy did not effect the samplingresults because PCBs were detected in both kinds of samples. Inaddition, PCBs that were detected in either pre-flush or post-flush samples at a given sampling location could be found toyield the opposite results during later sampling periods, furtherconfounding these findings.
PCB concentrations listed in Table 2 are reported as totalAroclors. Because all sampling results exceeded the comparisonvalue, these data will be evaluated later in the assessment fortheir risk to the public's health. Of note, the comparison valueused here is well below the limits of laboratory detection.
The EPA has established an MCL for PCBs at 0.5 ppb asdecachlorobiphenyl. The data as Aroclors cannot be compared tothis MCL without converting the concentration of each mixtureinto its equivalent decachlorobiphenyl concentration. Currently,there is no accepted methodology to do this.
Samples collected from the fire hydrants had the highestconcentrations of PCBs. Sample results from the 2 commercialsink faucets have been below detectable levels for the last 10sampling periods (See Table 2.I). PCBs were detected in 1 of 3samples collected from residential taps during the June 1991,sampling period (See Table 2.II). During this same period, PCBswere also detected in all the samples collected from the firehydrants.
Since 1984, SCDHEC has collected monthly and quarterly samples ofthe raw and finished water at the 3 Lake Hartwell water treatmentplants. PCBs were detected only once at a concentration of 0.13ppb in a January 1985, raw water sample from the DeFore Millsplant. All other sample results were below the level ofdetection.
The 1991 data results from the drinking water samples collectedat private residences from the 10 water distribution systems arepresented in Table 3. Analytical methods differed between SCDHECand EPA laboratories resulting in different levels of detection(the EPA laboratory used lower standards for quantification). PCBs were detected in 3 of the 21 samples collected by SCDHEC and17 of the 29 samples collected by EPA. Aroclor 1242 was theprincipal mixture identified by both laboratories; Aroclor 1016was identified in one SCDHEC sample and Aroclor 1260 was detectedin one EPA sample. PCBs were detected in samples from 6 of the10 water systems (Easley-Central # I, Southside # I, Highway 88,East Clemson, Town of Central, and Sandy Springs). These watersystems directly or indirectly (through Easley-Central # I) useTwelve Mile Creek as the source of their raw water (See Figure2). The highest concentration (0.29 ppb) was found in a drinkingwater sample from a private residence in Easley-Central # I. ThePCB concentrations detected in these samples exceeded thecomparison value for PCBs in drinking water. However, thiscomparison value is well below the levels of detection for bothEPA and SCDHEC laboratories.
3. Sediment and Soil
SCDHEC has been monitoring Twelve Mile Creek's sediment since1976 (See Table 4). During these years, samples have been takenat different points along this creek. Samples taken from stationSV-107 consistently had the highest PCB concentrations. The datahave fluctuated at each sampling point without evidence ofincreasing or decreasing trends.
There are no data on the amount of sediment that can be consumedduring wading or swimming. Thus, the comparison value used inTable 4 is based on the consumption of surface soil; that is, theconcentration of PCBs in soil corresponding to a 1 in a millioncancer risk from consuming 100 milligrams of soil each day for alifetime (12). Data for SV-107 have consistently exceeded thisvalue. However, this sampling point is located at SC Highway133. The public would be unlikely to encounter these sedimentsthrough swimming because the water is approximately 10 to 15meters deep at this location. The public could accesscontaminated sediments at SV-137, SV-015, and SV-282(approximately 2 miles below Sangamo Weston, Inc.) because theCreek can be waded at these points. The sampling data generatedafter 1984 from these sampling points did not exceed the soilCREG, although there appeared to be an increase in PCBconcentration in 1989.
EPA also collected fish and sediment samples in 1985 and analyzedthem for PCBs, dibenzodioxins (PCDDs), and dibenzofurans (PCDFs). Samples were collected from the top 2 to 3 inches of sedimentwithin a 4 mile stretch of Twelve Mile Creek. Notetrachlorodibenzodioxins (TCDD) were detected (See Table 5). Toassess the combined risk from the various congeners of dioxinsand furans, a set of toxicity equivalency factors (TEFs) wereused to convert the concentration of the various PCDF/PCDDcongeners into an equivalent concentration of 2,3,7,8-TCDD. Forthese calculations the I-TEFs/88 (the TEFs developed in 1988 byan international panel including the EPA) was used (EPA, 1989). The concentrations of PCBs in these sediment samples exceeded thecomparison value for consumption of surface soil. ATSDR does notcurrently have a CREG value to compare to the PCDD/PCDF data fromthese samples because EPA is in the process of revising its riskestimate for 2,3,7,8-TCDD. The data did not exceed ATSDR'schronic soil EMEG of 700 ppt for 2,3,7,8-TCDD (based onnoncarcinogenic effects).
The fish sample results are presented later in this section underFood Chain.
b. Surface soil
No local surface soil samples have been collected from the shoresof these water bodies. Sediment sampling has been limited tomidstream locations in an effort to follow the trend of thehighest sediment concentrations. The depths of these waterbodies have been known to occasionally vary. Shorelines alongTown Creek and the upper portion of Twelve Mile Creek couldpotentially be contaminated because sediments could be depositedon the shores of these shallow waterways during periods when thewater level rises. This would not be likely to occur furtherdownstream along Twelve Mile Creek where it meets the headwatersof the Seneca River arm of Lake Hartwell.
No ambient air samples were collected.
5. Food Chain: Fish
Fifteen years of data from SCDHEC's annual fish monitoringprogram for Twelve Mile Creek/Lake Hartwell are presented inTable 6 (2 pages). Fish were collected primarily at threesampling stations: Twelve Mile Creek (SV-107), Eighteen MileCreek (SV-532), and at Andersonville Island (SV-535). Fishoccasionally were collected at Martin Creek (SV-106), in theTugaloo portion of Lake Hartwell (SV-641) and at the Hartwell Dam(SV-642) (See Figure 3).
Following the EPA's 1976 study plan, SCDHEC collected fishannually in the 2- to 3-year old size class from 1976 to 1986. Since 1986, SCDHEC has actively sought 2- to 3- pound fish fortrend monitoring because creel surveys conducted by the SouthCarolina Wildlife and Marine Resources Department (SCWMRD)indicated that this weight class was most frequently harvested bythe public. The data generated from these samples could best beused to annually update the fishing advisory because theyreflected the amount of PCBs the public would typically consume. Sampling has always been targeted for the spring before thespawning season because fish were found to have higher tissueconcentrations of PCBs during this season than during the fall(8).
PCBs have been detected in all fish samples and at all samplingstations; however, fish caught in Twelve Mile Creek haveconsistently had the highest PCB concentrations. A suitablecomparison value for PCBs in fish cannot be developed for thissite without knowledge of the fish consumption patterns of LakeHartwell fishermen. The data will be evaluated in later sectionsof this Assessment.
Any interpretations and/or conclusions drawn from the data inthis table should be made with caution. Changes in the protocolused to prepare fish for analysis and difficulties collecting thesame species each year have had a direct effect on these samplingresults. Sampling stations SV-107, SV-106, SV-532 are within theSeneca River arm of Lake Hartwell, north of SC highway 24;sampling stations SV-535, SV-641, and SV-642 are south of it (seeFigure 3). Under the current fishing advisory, both SCDHEC andEPA recommend that all fish caught north of this highway shouldbe released and not eaten. For fish caught south of thishighway, the advisory recommends that all fish greater than 3pounds should not be eaten. The fish monitoring program will bediscussed in greater detail in Appendix C.
As stated previously in this section under 3. Sediment and Soil,EPA analyzed fish for the presence of PCDDs, PCDFs, and PCBs. Twenty fish samples of hybrid bass, largemouth bass, and channelcatfish were collected in 1985 from 12 stations on Twelve-mileCreek, 3 stations on Eighteen-Mile Creek, and 4 stations on theTugaloo River. One composite sample of several fish speciescollected in 1976 from Lake Hartwell was also analyzed. 2,3,7,8-TCDD was not detected in any of the fish tissue samples (level ofdetection varied from 10 to 20 ppt). 2,3,7,8-TCDF was theprincipal PCDF congener found in most samples in concentrationsranging from below the detection limit (ranging from 10-20 ppt)to 290 ppt. In 2 of 20 samples, pentachlorodibenzofurans andoctachlorodibenzodioxin were detected (See Table 7).
Applying the International Toxicity Equivalent Factors/'88 forassessing risks associated with dioxins and furans to these data,the amount of 2,3,7,8-TCDD equivalents in fish tissue ranged from0 to 29 ppt. The PCDD/PCDF concentrations in channel catfishcollected from Twelve Mile Creek were below EPA's laboratorydetection limit of 10 ppt even though PCBs were detected in thesesamples up to 68 ppm (the size weights of these fish were notavailable). There is no comparison value for PCDD/PCDFs in fishbecause there is no site-specific, fish consumption data; inaddition, EPA is currently reassessing the toxicity of 2,3,7,8-TCDD. This data, however, will be evaluated in later sections ofthis Assessment.
These fish samples were also assayed for PCBs. Total PCBs infish tissue ranged from 1.8 ppm (in a hybrid bass from EighteenMile Creek) to 147 ppm (in the 1976 fish composite taken from thelake). The hybrid composites collected from the Tugaloo River(outside of the advisory) ranged from 6-13 ppm. However, theweight of these fish reportedly ranged from 5 to 7 pounds;SCDHEC's fish advisory does not recommend consuming fish of thissize.
TRI for 1989 was searched for facilities reporting PCBs withinthe 3 South Carolina counties and the 5 Georgia Counties thatadjoin Twelve Mile Creek/Lake Hartwell. In 1989, only 1 plant inGreenville, SC reported PCBs to TRI. However, this plant did notreport any releases of PCBs directly to the environment; PCBwastes are disposed of in an incinerator in Georgia. Consideringthe location of this plant in relation to OU-2, any potential on-site PCB releases would not be likely to contaminate Lake Hartwell.
SCDHEC personnel collected drinking water, sediment, and fishtissue samples according to set procedures to minimize extraneouscontamination of the samples. Established procedures to maintainchain of custody for these samples were strictly followed.
The majority of the PCB analyses of fish tissue, drinking water,and sediments were conducted by SCDHEC's Environmental QualityControl Laboratories. This laboratory conforms to the QualityAssurance requirements of EPA. All laboratory practices followthe procedures outlined in the Quality Control Manual forChemistry Laboratories, Division of Analytical Services, SCDHEC,January 1987. This manual has been approved by EPA.
The EPA laboratory in Athens, Georgia has periodically done PCBanalyses on fish samples. PCDD/PCDF analyses were conducted byboth an EPA-contract laboratory in Research Triangle Park, NorthCarolina and the EPA laboratory in Duluth, Minnesota (theyanalyzed split samples). We assume that the practices of theselaboratories conform to EPA's QA/QC protocols.
SCDHEC had contracted with a private laboratory to conduct PCBanalyses on fish samples collected in the years 1987 and 1988. Because of concerns about the quality control used to analyzethese samples, the data generated from these samples werediscarded.
No physical hazards other than those normally associated with swimming, boating, and water skiing are evident at this site.
To determine whether nearby residents may be affected bycontaminants at the site, we evaluated the environmental andhuman components that lead to human exposure. This pathwaysanalysis consists of 5 elements: a source of contamination,transport through an environmental medium, a point of exposure, aroute of human exposure, and an exposed population.
We identify exposure pathways as completed, potential, oreliminated. Completed pathways require that the 5 elements existand indicate that exposure to a contaminant has occurred in thepast, is currently occurring, or will occur in the future. Potential pathways, however, indicate that at least 1 of the 5elements is missing, but could exist. Potential pathwaysindicate that exposure to a contaminant could have occurred inthe past could be occurring now, or could occur in the future. Eliminated pathways indicate that at least 1 of the 5 elements ismissing and will never be present. Completed and potentialpathways, however, may be eliminated when they are unlikely toexist. The discussion that follows identifies the completed,potential, and eliminated pathways at this site.
1. Food Chain: Fish
The primary route of human exposure to PCBs from this site is through the consumption of contaminated fish. Fish becomeexposed to PCBs when they consume smaller organisms which havebeen contaminated by PCBs in sediments. To a lesser extent, fishare also exposed to PCBs through the direct ingestion of anddermal contact with contaminated sediments (particularly true forbottom-dwelling fish like catfish). Samples of several speciesof fish popular with fishermen have consistently been found tocontain concentrations of PCBs above FDA's tolerance level of 2.0ppm. Even though SCDHEC has posted advisory signs at recreational areas around the lake, it is not known whether anglers arecomplying with this fishing advisory.
Exposures are believed to have occurred in the past (as far backas the early 1960's), are occurring at the present (as indicatedby sampling data), and will occur in the future (because of thepersistence of PCBs in fish and sediments).
2. Surface Water: Municipal Drinking Water
Local residents may be exposed to PCBs through the consumption ofcontaminated drinking water. PCBs have been detected in watersamples collected from taps supplied by those municipal watersystems that have directly and indirectly (that is, throughinterconnections with other water systems) used Twelve Mile Creekas a raw water source. It is believed that PCBs initiallyentered the Easley-Central # I water system through its raw waterintake on Twelve Mile Creek. From this water system, PCBs couldhave entered the other systems through interconnections. ThePCBs are mostly likely bound to the scale within the waterdistribution lines of these systems, occasionally being releasedthrough both desorption from the scale and through resuspensionof PCB-laden sediments when the water lines are flushed (seeBackground section, A.3. Site History).
Exposures are believed to be presently occurring to localresidents and may continue in the future. Although it ispossible that exposures may have occurred in the past, until1991, only commercial taps and fire hydrants were sampled. Thus,this sampling data are not indicative of the concentrations thatmay have existed at the taps of private residences.
PCBs bind tightly to sediments and are very resistant todegradation. The contaminated lake sediments can bioaccumulatethrough the food chain. Thus, they will continue to serve as asource of contamination into the future.
Past, present, and future exposures to sediment by localresidents are possible through the incidental ingestion ofcontaminated sediments during wading and swimming in Town Creekand Twelve Mile Creek (especially at the ravine described in theBackground section, C.2. Land Use). Sediment data for the TwelveMile Creek recreational area were generated from samplescollected midstream and at the bottom of the Creek. Because ofthe Creek's width and depth at this sampling location, exposureto these sediments would not be likely to occur.
B. Potential Exposure Pathways (See Table 8).
1. Surface Soil
Contaminated sediments could have been deposited along theshoreline because of localized flooding; however, no surface soilsamples have been collected to determine if they have beencontaminated. Without these data, exposure to contaminatedsurface soils can only be categorized as a potential pathway ofexposure. Additional sampling will be necessary to determine ifthese areas have actually been impacted.
2. Surface Water: Ambient water
PCBs will leave the lake water by adsorbing to sediments,reaching higher concentrations than the associated water column. Although adsorption can immobilize PCBs for relatively longperiods of time, they have been found to re-enter the watercolumn (ATSDR Tox. Profile for Selected PCBs). PCB-ladensediments may also become resuspended in the water column becauseof water turbulence created by heavy rainfall.
PCBs were discharged from the plant's wastewater system to nearbyTown Creek. From here PCB-laden sediments were carried with thecurrent to Twelve Mile Creek and eventually to Lake Hartwell. Although PCBs have not been used at the plant since 1977, smallamounts are still being discharged into Town Creek. This isbelieved to be due to the release of PCB-laden residues from thewastewater facility's settling basin.
Visitors to Twelve Mile Creek/Lake Hartwell could potentially beexposed to surface water contaminants during wading and swimming. Exposures to contaminants through this pathway will be discussedfurther in the Public Health Implications section.
1. Ground Water
Ground water is not a pathway of concern for this site. Inwater, PCBs primarily adsorb to sediments or other organicmatter; significant leaching into ground water should not occur. This opinion is supported by the data results from the privatewell sample collected near Twelve Mile Creek (See theEnvironmental Hazards and Other Contamination section, A.1.Ground water).
Under certain conditions, such as through the action of waves,PCBs could be released in aerosols. This would be unlikely tooccur at this site because the lake is a calm water body. PCBscould also volatilize from lake water; however, because theystrongly adsorb to sediments and are very insoluble in water, airreleases would not be likely to occur. We will not address airexposure to PCBs further in this health assessment.
In this section, we will discuss the impact of the site'scontaminants on the public's health, evaluate state and localhealth data, and address the community's health concerns. Unless otherwise indicated, ATSDR's Toxicological Profile forSelected PCBs was used as the primary source of toxicological andhealth effects information.
To determine the health effects that could result from exposuresto site contaminants, we will compare estimates of daily exposureat the site to estimates of exposure considered to be acceptable(that is, without health risk). To evaluate the potentialexposures to cause non-cancer, adverse health effects, they willbe compared to a Minimal Risk Level (MRL). ATSDR has developedMRLs for contaminants commonly found at hazardous waste sites. The MRL is an estimate of daily human exposure to a contaminantbelow which non-cancer, adverse health effects are unlikely tooccur (that is, a threshold dose). MRLs are developed for eachroute of exposure, such as ingestion, inhalation, and dermalabsorption and for the length of exposure, such as acute (lessthan 14 days), intermediate (15 to 364 days), and chronic(greater that 365 days). For compounds that have been classifiedas being able to cause cancer, it is believed that there isessentially no level of exposure that does not pose some finitecancer risk. This belief is currently being reexamined by EPA inlight of the reassessment of dioxin's cancer risk. For thisdiscussion, however, we will assume that carcinogens have nothreshold below which a cancer could not occur. We will compareestimates of daily contaminant exposure to the lifetime averagedaily exposure dose equivalent to a 1 in a million excesslifetime cancer risk (that is, a risk that is generally recognized as acceptable).
1. Polychlorinated biphenyls:
a. Food Chain: Fish
PCBs are clearly the principal contaminants of concern from thissite. People have been exposed to these compounds primarily byconsuming fish from Lake Hartwell and in particular, Twelve MileCreek. We are not exactly sure when exposures could have firstoccurred; fish monitoring only began in 1976. Because the plantbegan operations in the late 1950's, we can assume that fishcould have been contaminated as early as then. We do not know,however, the extent to which they were contaminated. Thus, wewill not be able to determine the daily PCB exposure that anglersmay have faced.
We know that people presently catch and probably consume the fishfrom Lake Hartwell, outside the fishing advisory area. We do notknow if anglers are complying with this advisory or if they areconsuming fish weighing more than 3 pounds. While we have datafrom a 1982 SCWMRD creel survey on the variety of fish harvestedfrom the lake (See Appendix C), we have no information on theamount and species of fish that local residents typicallyconsume. No surveys to assess statewide or regional consumptionpatterns have ever been conducted in South Carolina, a state inwhich fishing is a very popular recreational activity. Consumption rates may have been greater before the fishingadvisory was issued in 1976; thus, people who consumed fish priorto the advisory may have faced greater daily exposures thanpresent day fish consumers. Choosing an appropriate fish tissueconcentration to use in estimating daily dose is quite difficult,considering all the confounding variables in the fish monitoringdata. However, other than for catfish in Twelve Mile Creek,there does not appear to be much change in the tissueconcentrations over the last fifteen years.
Considering this uncertainty, the following assumptions were madeto estimate a reasonable, daily exposure dose to subsistencefishermen:
|Consumption rates would reflect the subsistencefishermen; that is, those fishermen whose diet reliedheavily on Lake Hartwell fish. We assume that this maybe 4, 1/2-pound fish meals per week (130 gramsfish/day).|
|All the fish in their diet came from Lake Hartwell,near Andersonville Island (SV-535).|
|The anglers are following the advisory, keeping andconsuming only those fish outside of the advisory areathat are less than 3 pounds.|
|Only the 1977-85 PCB data for fish fillet would beused; data from the samples using the modified wholefish technique (1986 and 1989 data) would not be used.|
|30% of the fish in the diet would be largemouth bass,30% would be hybrid bass, and 30% would be catfish. Wewould have preferred to use the percentages reported inthe 1982 SCWMRD creel survey, however, we had no BlackCrappie fillet data from SV-535 for 1977-85.|
|A 70 kilogram (150 pound) adult will be consuming fishunder these conditions for a 70-year lifetime.|
Under these conditions, the estimated amount of PCBs that aconsumer could get from eating this fish greatly exceeds ATSDR'schronic MRL for ingestion. At present, ATSDR does not have anacute or intermediate MRL for ingestion.
Aroclors appear to be toxic to the fetus but have not been foundto cause birth defects in various species of laboratory animals(including rats, mice, rabbits, and monkeys); however, theseeffects may possibly be caused by other contaminants. Slightdecreases in birth weight, head circumference, gestational age,and/or neonatal behavioral performance have been reported ininfants born to women who had environmental exposures (particularly women who consumed PCB-contaminated fish) or occupational exposures to PCBs; however, these effects cannot beconclusively attributed to PCBs because these women were alsoexposed to other chemicals. Thus, it is uncertain whether theseeffects will be exhibited in the children of women who subsistedon Lake Hartwell fish during their pregnancies (1).
Several PCB mixtures have been found to be carcinogenic infeeding studies of laboratory animals but it is not clear fromthese studies which of the components of the mixture are actuallycarcinogenic. The liver is the primary target of PCBcarcinogenicity. Because of its ability to cause cancer inanimals, the EPA has classified PCBs as probable carcinogens(that is, a group of chemicals for which there is sufficientevidence of carcinogenicity in animals but inadequate evidence orno data from human epidemiological studies) (1). Using theresults of a feeding study in which female rats were fed Aroclor1260, they were able to extrapolate the potential cancer risk ofPCBs to humans. The EPA cancer potency estimate for Aroclor 1260applies to all PCB mixtures because there is no informationregarding which constituents of any PCB mixture might be carcinogenic. Thus, the effects from Aroclor 1260 is assumed to berepresentative of all PCB mixtures. Using this cancer potencyestimate, subsistence fishermen could have a high increased riskof developing cancer over a lifetime.
This estimate is specific for those individuals who may consumelarge amounts of fish caught in areas not included in the fishingadvisory. Should individuals differ in their consumption rates,in their selection of fish species for consumption, and in theirchoice of fishing locations, the estimates may vary considerably.
The FDA used different assumptions to set the 2 ppm tolerancelimit for PCBs because they were only addressing health risks toconsumers of fish sold in interstate commerce. They used averageconsumption rates for the US population. In addition, theyassume only a portion of the fish in the consumer's diet comesfrom the area of contamination. Thus, their tolerance limitwould not be suitable to use in those situations where the dietof local consumers consists primarily of locally-caught fish.
b. Drinking water
Another important exposure route to local residents is throughthe consumption of PCB-contaminated, municipal drinking water. Based on the 1985 EPA study (See Background section, 3. SiteHistory), PCBs are most likely being released from the scale thatcollects within drinking water distribution lines. Althoughcontamination had been detected in the samples collected at maindistribution points (primarily fire hydrants), there wereessentially no samples taken from residential taps. It wasbelieved that PCBs might be getting through to these taps, butthe sampling data to support this belief has only recently beengathered. The residential taps most likely affected are suppliedby water systems that directly or indirectly use Twelve MileCreek as their raw water source. Using this data, the followingconservative assumptions were used to estimate a reasonable,worst-case exposure dose for the public consuming contaminateddrinking water:
|The average daily concentration of PCBs in the drinking water at all taps is 0.294 ppb (that is, the maximum concentration detected in the sample from a tap within Easley-Central # I).|
|We will assume that a 70 kilogram (that is, 150 pound) adult consumes 2 liters of water every day for 70 years from this tap only.|
|EPA's cancer potency estimate for Aroclor 1260 is applicable to both Aroclors 1242 and 1248.|
In this scenario, exposures to PCBs in drinking water barelyexceeded ATSDR's chronic MRLs for ingestion. We estimated thatthe residents could have a low increased risk of developingcancer from a lifetime of consuming water with this concentrationof PCBs. This risk estimate concerns exposures to contaminateddrinking water that may be occurring at present and in thefuture. We cannot address any excess lifetime cancer risk frompast exposures because there were no prior sampling data fromresidential taps; PCB concentrations could have been higher orlower than what was recently detected.
As previously stated, cancer risk estimates are made using a veryconservative approach. For example, we have assumed that the PCBconcentration in drinking water will remain constant for 70years. Based on the monitoring data, this is highly unlikely. Because PCBs have probably contaminated the scale within thesewater systems, they will probably be intermittently detected infuture drinking water samples. The fluctuations in the dataincreases the level of uncertainty in the estimate of the actualrisk. Thus, the actual cancer risk is unlikely to be higher thanthis estimate and may be lower.
Town Creek/Twelve Mile Creek bathers may also be exposed to PCBsthrough dermal contact with contaminated sediments. The extentto which PCBs can be absorbed through the skin cannot bequantified. ATSDR has not developed MRLs for skin exposure. Exposures could also occur through incidental ingestion ofsediments during wading and swimming. Considering the relativeinfrequency and short duration of these activities, however,exposures via this route are likely to be minimal.
d. Surface soil
The majority of the PCB-laden sediments should deposit along themidstream bottom of these waterways. Periodic changes in theelevation of Town/Twelve Mile Creek could cause some contaminatedsediments to deposit along the shoreline. With rainfall, thesesediments would probably wash back into these Creeks. However,there are no sampling data to confirm this assumption. Thus, wecannot assess the health risk to local residents through thismedium. We need additional surface soil data from areas of theseshorelines that are accessible to the public (particularly fromthe shoreline within the ravine, near Cateechee).
2. Polychlorinated dibenzodioxins and dibenzofurans (PCDDs and PCDFs)
The EPA is currently reassessing the health risk from exposure todioxins. Until this study can be completed, however, thefollowing discussion will reflect the EPA's original,conservative scientific opinion.
a. Food Chain: Fish
PCDFs and PCDDs can be formed by the pyrolysis (that is, achemical change brought about by heat) of PCBs andchlorobenzenes. PCDFs, not PCDDs, can be created as a result ofthe combustion of PCBs. PCDDs have been formed by the combustionof chlorobenzenes. Chlorobenzenes, together with PCBs, have beenused as the dielectric fluid for transformers and capacitors (2). Thus, even though Sangamo did not manufacture or use PCDDs andPCDFs in any of their operations, it was reasonable for EPA toanalyze fish and sediments for these compounds.
Only 4 of the 19 fish samples were taken from areas south of SCHighway 24 (specifically, from the Tugaloo River). These 4 fish,however, weighed between 5 to 7 pounds. Under the currentadvisory, fish weighing more than 3 pounds are not recommended tobe eaten. To determine the public health risk to consumers ofTugaloo River hybrids, fish of the 2- to 3- pound weight classshould have been collected.
Although fish of this weight class are not recommended forconsumption, we evaluated the data to determine the potentialpublic health impact from consuming fish with equivalentconcentrations of PCBs and PCDDs/PCDFs. Assuming subsistencefishermen consume this fish for years, their daily exposure dosewould exceed ATSDR's chronic MRLs for ingestion for both PCBs and2,3,7,8-TCDD (the PCDDs/PCDFs data have been converted to theirequivalent 2,3,7,8-TCDD concentrations). ATSDR has onlydeveloped an MRL for the congener 2,3,7,8-TCDD; it has notdeveloped MRLs for any other PCDD/PCDF congeners. The MRL for2,3,7,8-TCDD is based on this compound's reproductive toxicity tolaboratory rats (2).
These fishermen would also have a very high increased risk ofdeveloping cancer from a lifetime of consuming PCBs at thisconcentration in this fish. However, they would only have a lowincreased risk from consuming the PCDDs/PCDFs in the same fish.
Thus, the potential cancer risks from these fish essentiallycomes from the concentration of PCBs in its tissue.
Because of EPA's current reassessment of the toxicity of 2,3,7,8-TCDD, the preceding discussion could have overstated the riskfrom exposure to this compound. Chloracne, a severe acneiformskin disease, is the only lesion definitively identified inhumans resulting from 2,3,7,8-TCDD exposure. This effect hasbeen primarily found in workers who were exposed to certaincompounds in which 2,3,7,8-TCDD was a contaminant; however, thereare no data available to quantify the dose at which chloracnewill occur. Other effects (hepatotoxicity, immunotoxicity,reproductive toxicity, developmental toxicity, carcinogenicity)have only been reported in studies on various laboratory animals. These effects have not been reported in humans. While 2,3,7,8-TCDD has been actively studied, there has been very limited studyof the other dioxin congeners and of the dibenzofurans (2).
The principal congener detected in the fish tissue was 2,3,7,8-tetrachlorodibenzofuran; no 2,3,7,8-TCDD was detected in thesesamples. Because the I-TEFs were derived as a means to convertthe concentration of any PCDD/PCDF congener into an equivalent2,3,7,8-TCDD concentration, use of these factors increases theuncertainty in the estimates of the actual risk from exposures tothese congeners (12).
Considering the low levels of PCDFs/PCDDs in the sediment and theinfrequent and short duration of contact with sediment duringbathing, these compounds do not pose a significant public healthrisk.
The risk estimates described above were developed to predict areasonably high risk to the most sensitive members of the exposedpopulation. Conservative assumptions are used such that theactual risk would be unlikely to exceed these estimates. Theassumptions chosen for each scenario described above reflect whatwould be considered a reasonably high exposure scenario. However, by decreasing the exposure dose (for example, bychoosing to eat less fish with less contamination for less than alifetime), these estimates would be substantially lower. Thecancer potency factors themselves were developed to be asconservative as possible to ensure that a person will not be atan undue health risk from exposure.
1. County Mortality Rates
Although South Carolina has no population-based, statewide cancerregistry, we decided to review the cancer mortality experiencefor Pickens County from 1985 to 1989. We considered thisactivity to have some merit since exposures to PCBs may haveoccurred for 30 years, approximating the latency period of manycancers (that is, the period of time from first exposure to theonset of disease). Certainly, this review is quite limited. Wehave no way of knowing if any of the individuals who died from acancer had ever been exposed to PCBs from Lake Hartwell. Even ifthey were exposed, we have no way of knowing the duration ofexposure. We have no way of knowing about those individuals wholived the majority of their lives near Lake Hartwell but died atanother residence in another county or state. We have no way ofidentifying any personal risk factors, such as smoking, whichcould contribute to the occurrence of cancers. We also had noway of limiting our review to the mortality experience of thepopulation in closest proximity to the lake; thus, any positivefindings that may have occurred in this population may be undulyinfluenced by the lower cancer experience of other unaffectedcommunities.
With these limitations in mind, we compared the county's five-year crude mortality rates to that of Anderson County, OconeeCounty and the State for the same time period. Pickens County'srate for all cancers was slightly less than the State's rate(162.0 per 100,000 vs. 173.6 per 100,000, respectively). Thisrate was also below the crude rates for Anderson County (175.5per 100,000) and for Oconee County (183.7 per 100,000).
We used crude rates because they were readily available. However, the cancer mortality experience of a population dependson its composition by age, race, and sex. For instance, blackshave a higher cancer death rate than whites. Since PickensCounty's black population is approximately 1/3 the size of theblack population of Anderson County, this difference alone may beresponsible for the difference in their crude rates (the age andsex distribution of each county, however, was approximately thesame). Thus, these rates should be age-race-sex adjusted to makethem more comparable.
2. Local Health Data
The exposure study that was conducted by SCDHEC was quite limitedin scope. It assessed PCB body burdens in a small self-selectedgroup of people. Exposure histories were collected using ashort, self-administered questionnaire. There was no assessmentof the actual amount of fish these individuals consumed; therewas only one question asking whether they had ever consumed LakeHartwell fish (14 individuals reportedly had done so). Therewere no questions about where these fish were collected or whattypes of fish were typically consumed. There were no questionsabout other possible sources of PCB exposures unrelated to thesite (4).
These people were primarily long time residents of the Cateecheecommunity; the mean number of years they reportedly lived in thecommunity was 23 years (ranging from 8 to 56 years).
Only 2 of the 27 individuals were found to have PCB levels inexcess of 20 ppb (the upper level that would be expected for 95%of the people selected at random from the general US populationwithout evidence of PCB exposure). Both individuals werefemales; they were 46 and 73 years old. While they both reportedto have consumed Lake Hartwell fish for approximately 30 years,there is no information as to the amounts they consumed or thelocations where the fish were caught.
At the request of SCDHEC, the results of this survey werereviewed by the Center for Environmental Health, Centers forDisease Control (CDC). In a November 1984 letter from PaulWiesner, MD, Director, Chronic Diseases Division, they determinedthat the observation of 2 out of 27 individuals with serum levelsgreater than 20 ppb would be expected in this situation. Theydid not recommend any further epidemiologic studies since theyconsidered there to be little direct evidence that significantexcess exposures took place. They did recommend that exposurehistories of the 2 individuals be ascertained to identify anyother sources of exposures.
In the same letter, Dr. Wiesner addressed the self-administeredsurvey conducted by local area residents. They determined thatthey could not link exposures to any excess prevalence of self-reported health problems in the community because there was nodocumented evidence of excess exposures to PCBs. They found mostof the symptoms to be very nonspecific and likely attributable toother more common etiologies.
It is quite clear that the community has been and continues to beconcerned about this situation since it was first brought public. Despite attempts by SCDHEC, CDC, ATSDR, EPA to address thepotential public health effects from this site, the publicremains concerned about the health consequences from long termexposures to PCBs. This has been made readily apparent by thepublic's 2 attempts to conduct their own health studies.
The symptoms and diseases that the community has reported to dateare not necessarily related to PCB toxicity. As discussedpreviously, however, subsistence fishermen may be consuming fishat a level to warrant concern for developmental toxicity to theiroffspring. While this concern has sporadically been raised by afew residents, the effects that were self-reported (for example,mental retardation, "liver problems," "lung, heart, and liverproblems" associated with prematurity) cannot be necessarilyattributed to PCB exposure alone.
There have been sporadic anecdotal reports of individuals withinthis community who had developed different types of the morecommon cancers. Individuals have been living and possiblyconsuming contaminated fish for 20 to 30 years, a sufficientperiod of time for cancers to develop. Site-specific, cancerincidence data is needed to determine if the community hasexperienced an excess occurrence of cancers.
To adequately address the validity of these complaints, someeffort must be made to quantify the amount of PCBs the public,particularly subsistence fishermen, could have been exposed to.