~,~

To:

Richard E. Purdy
12/03/98 11:53 AM

Georjean L. Adams/US-Corporate/3M/US@3M-Corporate

CO:

Subject:

Risk to the environment due to the presence of PFOS

In the attached analysis of possible risk of ecological harm due to environmental levels of PFOS, I
concluded there is a significant risk of harm. In addition to this, my review of the available data on the
properties of PFOS, indicates that the sink for it in the environment is biota. We know of nothing that it
has more affinity for than blood serum and liver components. Other persistent pollutants have a high
affinity for sediment, but PFOS does not. So, PFOS can’t be trapped and buried like they can. It likely
continues to partition into biota from the other compartments in the environment; thus the levels we are
seeing in eagles and other biota is likely to climb each year.

I believe all this taken together constitutes a significant risk that should be reported to EPA under TSCA
8e.

Pioneer Food Chain Risk Assessment of PFO

Made Available by 3M for Inspection and Copying as Confidential Information:
Subject to Protective Order In Palmer v. 3M, No. C2-04-6309

1533.0001

3MA00242688

 Pioneer Risk Assessment of Adverse Effects in Marine

Mammals from PFOS in the Food Chain
Rich Purdy
3 December 1998

Introduction:
PFOS (perfluorooctanesulfonate) was found in two of three populations of naive rats--fl~at is, rats not
knowingly exposed to PFOS. The rats that had PFOS in their blood were fed food fl~at contained fishmeal.
Incomplete studies indicate that PFOS is in fishmeal. This is consistent with the observation that fledgling
eagles that eat predominately fish contain PFOS in their blood.
Most fishineal is made from menhaden (Brevoortia tyrranus)(Drafl Report: Sources ofFish Food
Constituents By Jolm Giesy and Paul MeMe, Septelnber 6, 1998). Menhaden eat planklon and are
considcrcd a second link in the US Atlantic coastal food chain. Thcy arc in turn catcn by other species of
fish, which are in turn eaten by other fish, mammals and birds. The mass of menhaden is so great that their
catch represents 40% of the US commercial finfish fisheIies.

Since these fish appear to contain PFOS, it seemed prudent to calculate the amount of PFOS that might be
transferred up the food chain and compare this value to a concentration that causes adverse effects. These
calculations do not contain precise data so the evaluation is approximate. The purpose is to see whether
predicted enviromnental concentrations are anywhere near concentrations that cause effects.
This assessment was organized and performed in accordance with the guidance given by EPA in Guidelines
for Ecological Risk Assessment (EPA!630/R-95/002F April 1998). The process presented in these
guidelines is a repetitive one where a risk assessment is updated or redone when new information is
available. The new information can be generated by other processes or generation can be driven b~y the risk
assessment.

Problem Formulation:
A study of rats not purposely exposed to PFOS found that they had siguificant levels of it in their livers.
The likely source was the fishmeal in their diet. ff PFOS is in fish then other fish, fish eating mammals and
birds are consuming PFOS. The concentrations found are not likely to cause toxic responses, but what
about after biomagmfication through a food chain?
The hypothesis to test: The concentration of PFOS in food of marine animals causes adverse effects.

It was decided to keep the problem narrow in this assessment in order to simplify the analysis. This is not
to imply that this is the most important or only problem. In addition tlfis is a pioneer and possibly first
iteration of this assessment. Refinements in the analysis and more data will allow other iterations if this
pionccr asscssmcnt indicatcs thcrc may bc a significant risk. Thc purposc of this pionccr asscssmcnt is to
determine the magnitude of the risk and whether other iterations are warranted.

3MA00242689
1533.0002

 Analysis
Characterization of Exposure:
The task of this characterization is to estimate the concentration of PFOS in the food chain organisms that
eat menhaden and the concentration in the animals comprising the ne~ two steps in the food chain. Two
examples of Ihe many food chaius possible are:
Menhaden -> cod -> seal -> killer whale
Menhaden ->carnivores fish 1 -> carnivores fish 2 -> seal
For the purposes of this analysis the biomagnification by fish and mammals is about the same. This is
supported by the metabolic rates, daily consumption and growth rates published (Fugacity-Based Model of
PCB Bioaccumulation in Complex Aquatic Food Webs, Jan Campfcns and Donald Mackay in Environ. Sci.
Technol. 31,577-583(1997) and Wildlife Exposure Factors Handbook, EPA/600/R-93/187 December
1993). Based on the data in these references and recollection from other readings, an average
biomagnification valne is 9.
In an evaluation of PFOS in the hvers of r~’ffve rats an average level found in male and female 10-14 week
rats was 0.09 mg/kg. The food label listed fishmeal as the fifth ingredient. According to labeling
standards, fisluneal was less than 20% of the food. For tiffs analysis it is assumed that fishmeal made up
16% of the food and th~t it was the only source of PFOS. This value could be too high xvhich wonld mean
the ultimate calculated tissuc levels would bc low, but the value cannot bc too low by much because the
most it could be is 19%. If an animal were to eat 100% fish, as many fish and sea mammals do, it should
have about 6 times as much in its livers as these rats do. That level would be 0.56mg/kg (0.09mg/kg/0.16).
This is the estimated level in the liver, which is the organ that contains most of the PFOS. The blood
contains the next lfighest level. Work with rats has shown that blood contains about 1/6 the amom~t of
PFOS to be found in the liver. Assmning that the liver mass is 3% and blood is 8% the mass of an animal
and that the concentration in other organs is insignificant in comparison to the levels in liver and blood,
then the total body burden of PFOS in the first step of the food chain above menhaden is 0.062mg/kg
(0.56mg/kg X 0.03 + (0.56mg/kg/6) X 0.08)
The level in the second food chain link then would be nine times this level or 0.56 mg/kg.
Using data on seals, the accumulated dose at time of whelping is calculated. It is assumed that 100% of
PFOS consumed is retained. Data for ldller whale food consumption sexual maturation was not
immediately available, but they are assumed to be similar.
The mean time for sexual maturi.ty for female harbor seals is 5.5 years and gestation is 11 months (Wildlife
Exposure Factors Handbook, EPA/600/R-93/187 December 1993). So for up to 77months before whelping
a seal was assumed to be eating fish with the above calculated concentration of 0.56 mg~kg PFO$. Seals
eat 6-8% of their weight per ~y in fish. They eat 13% their first year nnd 10% xvhen gestating For the
ease of calculation it was assmned that an 80kg seal eats 8% or 6.4kg/day. This works out to 15 X 103 kg of
fish (77months X 30days/month X 6.4kg/day). Assuming the fish a seal consumed contained 0.56mg/kg,
then a seal would consume about 8.4 X 103 mg PFOS (0.56mg/kg X 15 X 103 kg). This works out to a
cumulative dose before xvhelping ef about 105mg/kg (8.4 X103 mg/80kg)
This cumulative dosc can bc uscd for scals that arc two food chain lcngths above mcnhadcn. This is
probably not the norm. The average seal is probably one food chain length above menhaden. But this
calculated cumulative dose probably represents that seen by populations of killer whales that eat seals.

3MA00242690
1533.0003

 Characterization of Ecological Effects:
In a two generation rat stu@ it was found that 34%or the pups born to ammals dosed by gavage 1.6 mg/kg
per day PFOS were born dead or died within four days of birth. For the purposes of this assessment rats
arc assumed to be an adequate test surrogate species for marine mammals. There is uncertainty both to
whether marine malnlnals sucla as seals, sea lions, and killer whales are less sensitive or lunre sensitive than
rats.
The lowest dose of 1.6mg/kg tlmt caused an adverse effect was given for 6 weeks before mating, the week
of mating and the 22 days of gestation for a total of 71 days. The cumulative dose up to the time of
whelping is 113mg/kg.
It has been reported that natural populations of sea lions are experiencing significant reproductive
dysftmction (Peter Ross, Enviromnental Centaminants and the Risk of Adverse Effects in Marine
Malmnals: An Ovexwiew, 1998 SETAC Almual Meeting). Also 70% of a ringed seal pepulafion was not
rcproducing. This is highcr than the norm. Apparently some field biologists believe killer whale
reproductions has declined, but there has been no scientific study to verify this.

Risk Characterization:
The predicted cumulative dose of PFOS through t~vo food chain links to sea mammals such as seals, sea
lions, killer whales ~nd porpoises was calculated to be abom 105 mg/kg. Tlfis valne is about the same as
the calculated cumulative dose of 113 mg/kg PFOS that causes reproductive impairment to mammals. Thus
there is a significant risk. There is not enough input data to calculate uncertainty in this risk. Such data is
needed to reevaluate the degree of risk.

3MA00242691
1533.0004