Document 06KbXVd4jQXvvJekJro0wNEnR
NEW SOLUTIONS, Vol_ 18(2) 129-143, 2008
AR-22b-3829 ii e# 73
l . r r,(.'t~t~- . nr'. ~-
Scientific Solutions
SELF-REPORTED HEALTH EFFECTS AMONG COMMUNITY RESIDENTS EXPOSED TO PERFLUOROOCTANOATE*
PAMELA ANDERSON-MAHONEY JENNY KOTLERMAN HARPREET TAKHAR DAVID GRAY JAMES DAHLGREN
ABSTRACT
Serious health effects due to perfluorooctanoate (PFOA) exposure are suspected. The aim of this study was to evaluate the health status of nearby residents, with prelonged exposure to PFOA in their drinking water. A population of 566 white residents who were plaintiffs or potential plaintiffs in a lawsuit was evaluated by questionnaire for health history and symptoms_ Standardized Prevalence Ratios were estimated using National Health and Examination Survey (NHANES) data files for comparison rates. The exposed subjects reported statistically significant greater prevalence of angina, myocardial infarction, and stroke (SPR = 8.07, 95% C.I . = 6.54 9.95; SPR = 1 .91,95% C.L = 1 .40 - 2.62, and SPR = 2.17,95% C.l . =1 .47 3 .21, respectively), chronic bronchitis, shortness of breath on stairs, asthma (SPR = 3.60, 95% C.1 . = 2.92 - 4.44; SPR = 2.05, 95% C.I . ~ 1 .70 - 2.46; SPR = 1 .82, 95% C.L = 1 .47 - 2.25, respectively), and other serious health problems . The increased prevalence of adverse health effects may be due to PFOA . Further study is needed .
'This study was supported by The Law Finns of Winter Johnson and [fill PLLC, Law Firnt of "raft, Stettinius and Hollister LLP, Law Firm of Hill, Peterson, Carper, Bee and Deitzler, PLLC. The law firms were not involved with data analysis, data interpretation, or writing of the report. The legal case was settled prior to submission of this manuscript . The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication .
129
2008, Baywood Publishing Co ., Inc. doi 10.2190/NS .182.d http ://baywood .com
130 / ANDERSON-MAHONEY ET AL .
Organic fluorinated chemicals have been the focus of scientific and public health debate for some time. Most of the studies on these compounds have examined perfluorooctyl sulfonate (PFOS) and perfluorooctanoic acid (PFOA) . In addition, PFOS and PFOA are commonly present and persist in the environment [ 1 ] . PFOA is produced as an emulsifier in the manufacturing process for fluoropolymers [2]. Since the 1980s, there has been a steady increase in the use of fluoropolymers due to their many useful properties [3] . Because of their specific chemical and physical properties, fluoropolymers are used in electrical insulation, fire foam, floor polish, insecticides, Teflon, stain-resistant treatments for upholstery, carpet, textiles, and oil-resistant treatments for paper and plastic [4] .
Several sources of human exposure have been suggested, including environmental releases and some consumer products, but due to a dearth of research, specific pathways responsible for human exposure are not well understood (5] . Potential pathways for human exposure include food, water, leaching from commercial products such as food packaging, contact with products such as clothing, and inhalation of dust or vapor phase precursors coming off of household products such as carpet (6, 7-9, 10) .
Perfluoroalkylated substances (PFAS) are eliminated slowly from the human body [111 . The biological half-life of PFOA is approximately 4'/z years [12] . PFOA is persistent in the environment . It does not hydrolyse, photoiyse or biodegrade under environmental conditions [13) . Bioaccumulation [14] and ubiquitous exposure to U.S . citizens heightens concern about adverse health effects [15, 16) .
Although a number of studies have been conducted to determine the toxicity associated with PFAS, the relative toxicities of the different chemical structures like PFOA and PFOS are not clearly understood_
In rodent studies, PFAS cause multiple adverse health effects [17) . Endocrine disruption is observed with elevated estrogen and decreased testosterone in animal studies [18, 19) . Elevated Thyroid Stimulating Hormone (TSH) and elevated prolactin have been observed in humans [20, 21) . Increased liver weight has been documented in rats [22] and in monkeys [23) . Dose-dependent hepatic enzyme elevations have been seen in PFOA-exposed workers [24] and recently a small sample of PFOA workers in Trissino, Italy, was found to have slightly increased levels of non-HDL cholesterol and total cholesterol with increasing blood levels of PFOA [25) . In addition, a study in January 2005 reported statistically significant increases in cholesterol fractions (total and LDL) and increases in triglycerides in the highest serum PFOA exposure group [26]_ Rat studies have shown liver and pancreatic neoplasms [191 . Studies of PFOA-exposed workers have shown increased rates of cancer, including prostate [27), kidney [28), bladder [28, 29), colon, testes, and other cancer types [28) .
This article reports on the results of a cross-sectional comparison study of self-reported health history and symptoms . A separate manuscript reports an increase of cancer prevalence among this PFOA-exposed population [30] .
SELF-REPORTED HEALTH EFFECTS / 131
MATERIALS AND METHODS
Study Subjects
We administered a questionnaire to volunteers from a population of approximately 23,900 households (approximately 70,010 individuals) near a Teflon manufacturing plant located along the Ohio River in Wood County, West Virginia. All of our study participants were mandatorily included in a law suit, regardless of their individual choice to be included, because a non-opt-out class action was certified against the local plant on behalf of all persons whose drinking water was contaminated with PFOA (>0 .05 parts per billion (ppb)} .
PFAS has been used at this plant since the early 1950s . The plant uses PFAS to manufacture Teflon and other materials and has consistently released perfluoroalkylated substances into the surrounding areas . According to the Environmental Working Group (EWG), in 1999 alone, the plant released more than 40 tons (86,806 pounds) of PFOA into the air and the Ohio River. The plant reports it reduced emissions to 10 tons (20,168 pounds) in 2002 (EWG)_ PFOA has been present in the plant's drinking water at 1-5 parts per billion (ppb). PFOA levels above 0 .05 parts per billion have also been detected in public and private drinking water supplies in communities on both the Ohio and West Virginia sides of the Ohio River near the plant, including the water supply of the neighboring manufacturing plant .
Study participants were recruited through a public invitation via TV, radio, and newspaper advertisements inviting all users of the PFOA-contaminated public and private drinking water sources to participate in a study by filling out a questionnaire . Over a period of three days in August 2003, five hundred and ninety-nine (599) volunteers or 2.5% from the targeted population participated in the study. Most were white_ Each person who filled out a questionnaire reported exposure to PFOA-contaminated water for at least a year. They included subjects who used water with PFOA levels as low as 0.05 parts per billion (ppb), such as from Pomeroy, Ohio, and Mason, West Virginia, which is similar to levels measured elsewhere in the greater region. However, the majority (65%) who participated were exposed to water provided by the Lubeck Public Service District in West Virginia where PFOA ranged from 0 .4-3 .9 ppb and the Little Hocking Water Association in Ohio where PFOA ranged from 1 .7-4 .3 ppb . Each water district in the area reported PFOA levels including well testing data, and company records have released information on PFOA in the company drinking water (see Table l)_ Some study subjects worked in the plant where PFAS were used and therefore had both residential and occupational exposures .
Data Collection and Questionnaire
We collected data on 599 adults exposed to PFOA in their drinking water from a nearby manufacturing plant . Trained and experienced proctors administered
132 / ANDERSON-MAHONEY ET AL.
Table 1 . PFOA Levels by Water District/Source and Percent of Study Subjects Reporting Water Consumption
from Each Source
PFOA tevelsa (ppb)
Location
Households
Total populatiorrb (% total pop)
Study sample (% study sample)
1 .7-4 .3 0.4-3 .9 0 .25-0 .37 0 .08-0.13 0 .06-0.1 0.06-0 .07 0.165 1 .0-5 .0
1 -75-1 .87 0 .05-8.6
Little Hocking, Ohio Lubeck, WV Tuppers Plains, Ohio Belpre, Ohio Mason, WV Pomeroy, Ohio Blennerhassett Dupont, Washington Worksc GE Ptasticsc
68 Private Welts WVA & Ohio
4200 3700 4800 6000 4200 1000
71
N/A N/A
68
11,760 (17%) 10,360 (15%) 13,440 (19%) 16,800 (24%) 11,760 (17%) 2,800 (4%)
200 (0%)
2,200 (3%) 700 (1%)
190 (0%) 70,010
161(27%) 219(37%)
19 (3%) 45 (8%) 4 (0.7%)
1 (0.2%) 6 (1.6)
90(15%) 43(7%)
11(2%) 599
'All PFOA levels were reported by their respective water dstricts and are available via the Internet .
bEstimated as 2.8 people per household. Company employment reports. Note : N/A not applicable .
the questionnaire in groups of 30 to 40 persons or individually by telephone interview . In the group sessions, all subjects were given the same instruction and proctors were available to answer questions .
The questionnaire explores demographics, including age, gender, occupational and residential history as well as medical, social, and behavioral history. We asked about health symptoms, diseases, surgeries, medications, family history, income, chemical exposures, and lifestyle measures including smoking and drinking .
One unlikely symptom question is designed to test for the veracity of the responses provided . Questionnaire responses are machine-readable, scanned on-site, and verified before subjects leave . This basic questionnaire has been used in prior studies of exposed and unexposed groups [311 .
SELF-REPORTED HEALTH EFFECTS / 133
Comparison Databases The Centers for Disease Control and Prevention's (CDC) National Center
for Health Statistics (NCHS), Division of Health Examination Statistics (DHES) has conducted health and nutrition surveys since the early 1960s and is designed to obtain a nationally representative sample of the health and nutritional status of the U.S . population . Details for each survey in addition to the public use data files are available online (CDC).
The most recent National Health and Nutrition Examination Survey (NHANES) began in 2001 . Every year, approximately 7,000 civiGan, noninstitutionalized U .S . citizens of all ages are interviewed ; of these citizens, approximately 5,000 complete the health examination component of the survey . NHANES 2001-2002 over-samples low-income persons, adolescents 12-19 years, persons 60+ years of age, African Americans and Mexican Americans_
Statistical Methods
Frequencies and percents for demographic, social, behavioral and physical characteristics are presented . Standardized Prevalence Ratios (SPRs) for certain disease endpoints were calculated using the NHANES 2001-2002 data files for comparison rates . SPRs are simply observed to expected ratios, (number of cases in exposed population/total number exposed) divided by (number of cases in the comparison population/total in the comparison population) . A value of one represents the null or an equal value in both the observed population (survey data) and the expected population (NFIANES) . Values greater than one indicate higher prevalence in the observed population compared to expected and values less than one indicated lower prevalence in the observed compared to expected . Statistical significance is demonstrated with either p-values, confidence limits, or both_ Other studies have used the NHANES data files for similar analyses [33] .
NHANES data was restricted to non-Hispanic whites 18 and older to be comparable to the exposed population . Only whites were included in the analysis because the number of non-white participants was too small to yield robust results . Rates were calculated for each medical condition using sampling weights developed specifically for the NHANES data source, controlling for age and gender. The sampling weights account for the sampling methods as well as the non-response rates, and they render the data representative of the total U .S . population . The weights also provide for accurate assessment of the sampling error of statistics based on these survey data [34j .
Exposed subjects were restricted to those who had lived in the area for a minimum of one year_ The health effects of participants who were diagnosed prior to the time of residence were excluded from the analysis .
134 / ANDERSON-MAHONEY ET AL .
The 76 subjects who had worked at the Teflon plant were compared to the general population of the exposed area . They were subsequently included in the exposed group because no worker-exposed differences were found.
RESULTS The questionnaire was administered to 599 subjects . Ten of the subjects were less than 18 years old and 21 were non-white, and two subjects were missing age and were therefore excluded from the analyses. The remaining group of 566 residentially PFOA-exposed adults are middle aged (49.9 years with a standard deviation (SD) of 14 .43 years), a balanced proportion of males and females, with slightly more than high school attainment of education (12 .76 years of school, SD 2-18) . The average body mass index is 28.36 kg/mz with a large SD (+/- 6.41) . Almost half (48 .25%) of the sample can be classified as over-weight (BMI > 28), one-third (33 .22%) as average (BMI = 23 - 28), and the remaining (18 .53%) as underweight (BMI < 23) . Table 2 shows the Standardized Prevalence Ratios (SPRs) comparing the exposed group and N1fANES 2001-2002 data while controlling for age and gender. The significantly elevated SPR for cardiovascular problems was 4 .29 (95% C.I . = 3 .47 - 5 .29), and included is that of angina (SPR = 8 .07 ; 95% C . 1 . = 6 .54 - 9 .95), myocardial infarction (SPR = 1 .91 ; 95% C .l . = 1 .40 - 2 .62), and stroke (SPR = 2 .17 ; 95% C .I . = 1 .47 - 3 .21) . Respiratory health effects were elevated for chronic bronchitis (SPR = 3 .60 ; 95% C .1 . = 2 .92 - 4 .44) and asthma (SPR = 1 .82 ; 95% C.1 . = 1 .47 - 3 .21) . The prevalence rate for shortness of breath was also elevated (SPR = 2 .05 ; 95% C .l . = 1 .70 - 2 .46). Thyroid problems and diabetes were marginally elevated in this PFOA-exposed group (SPR = 1 .56 ; 95% C. 1 . = 0.22 - 1 .98 and SPR = 1 .54 ; 95% C . 1 . = 1 . 16 - 2 .05, respectively) . The overall rate of liver disease was not significantly elevated . The prevalence of liver problems was 3 .4% (18 persons) with liver disease diagnosed by their doctors and 10 .6% (59 persons) reporting elevated liver enzymes . Table 3 examines selected outcomes by age and gender. The effect of exposure appears to be exaggerated in young adulthood for several of the listed outcomes : high blood pressure, cardiovascular problems, and liver problems . By age 65 and over, the effect measure is either at the null (liver, kidney, and thyroid disease in males) or greatly reduced (cardiovascular problems) . However, the exposed population (EP)/US ratios for the older population remain high for asthma and chronic bronchitis, and are slightly elevated for high blood pressure, shortness of breath, and cardiovascular problems.
DISCUSSION
The U .S . EPA pressured perfluorooctane sulfonic acid (PFOS) off the market in 2000 because of its persistence in the environment and human body, concerns
SELF-REPORTED HEALTH EFFECTS / 135
Table 2 . Standardized Prevalence Ratio (SPR) Comparing Observed Disease Rate Per 100,000 among a Residentially PFOA-(-=xposed Population to the Expected Disease Rate of the General U.S . Population Controlling for Age and Gender
Disease or symptom type
Number
Observed
Expected
diseased
rates
ratesa
in exposed (per 100,oo0) (per 100,000) SPR
Clb
Cardiovascular probk>.msc
170
30,088
7,019
4.29 3.47-5.29"
Chronic bronchitis
113
Kidney disease
21
Shortness of breath
323
on stairs
22,114 3,757 57,270
6,145 1,665 27,994
3.60 2.26 2.05
2 .92-4 .44" 1 .45-3.51' 1 .70-2.46'
Asthma
105
20,669
11,369
1_82 1 .47-2.25'
Thyroid problems
82
15,589
10,019
1 .56 122-1.98"
Diabetes
56
9,947
6,457
1.54 1 .16-2.05"
High blood pressure
186
33,096
28,077
1 .18 0.97-1 .43
Liver problems
19
3,754
3,728
1 .01 0.64-1.59
aExpected rates are from MANES 2001-2002 using sampling weights to unbiased estimate of national rates while adjusting for non-response, survey sampling technique while giving an accurate estimate of sampling error-
be onfidence Interval Includes MI, Stroke, Angina *Statistically significant (p <- 0.05)
calculate an design, and
about its toxicity, and the fact that it is widely found in human populations and wildlife [3, 11, 141. A short time later, one main producer, 3M, discontinued the perfluorooctanesulfonyl fluoride chemistry used to manufacture PFOS-based chemicals. Animal studies have confirmed that PFOA and PFOS are toxic [5, 35, 361 and worker studies indicate human health effects [29, 37). To our knowledge, this is the first study evaluating the health status of residents living near a Teflon production facility .
The reported emissions from the production facility have contaminated the local water supply . Public drinking water delivered to those living near the plant contains more than 0 .05 ppb PFOA as reported by the water districts in the area . Neighbors are exposed to much higher than background levels of PFOA in drinking water and air [37, 38] . A recent study conducted by Edward
136 J ANDERSON-MAHONEY ET AL .
Op0 OpO~ OpOp
Op O p
V
v
'O 00O0 r-CO3pp OOOp
O OO O v V v
C) co to p 0 0 p .-OO O O O p
v
000 0 00 ' 000
O OO v v v
C
U
N
Ul d
C U
O n-
U
C
Nc6
N
OZ)
W
O
m
O o~m
d
~, C7 m
a- J-- o m: S
W O 7
m 7;6 m
U) CC U)
go
to
d) N Co m
CD
en
m cis Wm
Q -
U) m
W
N ffl
Ucl
Z -0 .U`
M ca
O_ O
U d U -U w
Q 'm N
m
io~)cQ~9c~OCOp7 .- .- .-~
O~ONO Op00 Op00
O
v
p
O
Cv!
i- to 00 N o) cu C') M Nr+-M
C(D')cCo 7cNo
ON(D-
O ~f) CD 00
oO_k_ n
O
CO
(') CV N
lo- O O) dCO .[7MC7N
OOOpO OOOp OOOp
O vO vvO p
N d C) O) co N to O) ui 4 (O N
MN C7 (U OC7't0 Oth co O oCort~o at-oloc) rN N N
co N to Ntnc.)O) ~O- O
ONOO 0000 0000
O v
O
O O
vv
V N
(-D
1~-
('7 N
Nr ~ ~-
(")O)tD O) (D - CO
Oatl- N
0)
f- Q
o OfvCla) O o t-
1A 1f)
000 000 ,000
O vO vvO
co fl- fS c") t[3 V
N r- .-
NN ~iCO - N to
OV aD aD Or)nc00nvN _ aD to (h to to 1-
G) "B QL C
N
C7 p _co-C0: a~CL m ~o U ,KcO9
ca -0 a) 4D 0)
moa
U n >, C
Qf N Q
a_ ti-
M N
F- Uv
N
m _U~
U U? Q n vw m~
N N- tO)(')to O U)O)tA0 O1* CD 0)
N ~n n ~ O co fo-
a
C7
00 CD V tO
fl. NNC') co 0)N V
0~r0~ oo O CO 0~0
(0-10 C0001 d OtaD00 U) - V CO
cnl) N ~ V ~ ~
(D (D CD co Cofl
tOf~c7 ' ONto
tA i7 to
~ vv
~_U w_ W a < CD
N
V1 (0
C0700DNCN 0 O)U)t7hOOO-~t Nll O) O fl. V N O) c")- .--
CD on 000 to 0100 NOt- O (7 O Co N NN~ "-
~O
O~~ V
OOCl)O
CD tt Otn
OtA(y Q) t` O) qi (D V
ON(pO -
O14 m u"1
N h Of N N ~t
V
NNC7tfT ~ -It V to tn
01 C
~tUM~0
N 1+ N N~ m CO~~ h V
O(NDN
p~mU?
m N
(7~ACO V'
fD(OOO j OltO~At- t
~ O co (0)U Ocof y -"rCM = iahu O o N ~, On7 cth0 r- ~oA N u>
N 1, O) U) V '! d N to ~ O) CD ('VC') O m
'-
o c
a.o
0
o o
~C7O
~ O GO 1~ 0 m ('M 4n
U
m
Q mU L
c'") V (O .}.
C U
IT (7 d t0 +
-C
~t ~m ~T c'7 ~Y t0 .}-
in
Q a1 Q
a0 i() O IA
~ c) u> co
t CD U) O to
0 - M to (0
2
a~0
tq
v)
O
to
tcno
C V 'R O R U c'7 V tD + L aO u) O u')
u) - cw) in cc>
ooo,n oO00O
0 O O O
V VV
V t- t0 N N ('9 N V O)NV r N r
r r 0000 C3 a 01 C) 0 000 0 0 40 O
VVV V
c~~`c~~
M Op V r m
aD N .41 O ,-N,- N CO ~t0 V
~tf) to O ~ t~ NN('N~)N
P) N r to No~or~ t~0r tflo- (O Oa0 t- ~ t0 O
(,- a0
~ ooaOo~rn `
O OO
.tO tA c0 o ' CM ~O
00~~~ O OOO
vvo~
t()O- O
to
fl-
V'
~_
NO NON'
O O) to ~T(O
Od) w Of coNVN t0 N C1 to Q) V c0 N c0 (O Q) O
NC)C')
SELF-REPORTED HEALTH EFFECTS / 137
to ~ cno
O O O
rtA O ' CO t-aD
KOCV
O ppOm O O 01 O
V
m~n~~
U) N V O
c0 1- r- tc7~Af~~ cU 14 W DO (7f")ODN t0 N 1~ V MT-r~
V to N o~cnU~ t, d) V) I f0 t0 c0 N c7 O
N14
_o
o00
01
ooo~ .cj
46 O O O
V V
v
m
m
r 00 h V 0 3 00)(0(Ni .- .- ~ m
N
N C O
tn (O r O C3 r
O O O
r~0
01 N i0 V daDC00
U1NaDN
N G C2 to
f~
('7
~N('O~(N7
O) Of l') (O fl_*-14: (P (O ON N'I! t<-D 1- V' 'It 1nOCON
O m
E
"N m m
_ C
m Q
~ E
D c
cN 0
v
m
m ao
:3 W U m v m
N O 3
.m.. v U L Q1
C m ~1 3 p) ch C c > ~Q o)
Em mt G' 3
(')(DhO) (D0V: 0 CO to (O O) to ~ _~ ~ NNVOItI
Go U) It N UOC?O
ui c0C00) N U) P M C) to V) d N N LO N
OOO)to OOtQt 000if? OOCOst O t0 O) to N N (O V
t0 O) (O U) a n
I
r: 0 100N0)
O 1- r-
C N L O U O m N '
" pl
O o C Q t
W h
a
= cto w-
Z 6c
Em aN~,~ncnoo~
p 1'~ th V CO V 1- N N tONtr) V 03 a0 t0
cc .oo-o > _.w -rr
v coN ~c~ `!W '-
V' O O
N to t+ 1-
N <t N O
R t0 O) h
It NN~
0) Id,
t0 tI) t~) N
-to N
rnrNocvo
OI V co to tn c") V N
EU N (Q
m Tp N m ~m c
_~
N
~
f0 U)
i0 m O N Q
v
m
m
U. Y
to
Q
o
ma m o
0
>`
Z7
m a ~ v~ d m Q
V O) V t7 V CO
~ Q) V
(D V Of V'
`~ C'7 V ~O + C C7 V (O
O V d) V
a t7 V (U } w
a~
+ CO Ln O to
? (0 t[) O ln
13
+
W QQ U
co to O to L (D t(S O tf)
U r ('M t[) CO J r- c7 to (O Y - M W) (D F- r M to c0
c
138 / ANDERSON-MAHONEY ET AL.
Emmett at the University of Pennsylvania among community residents served by Little Hocking Water Association (LHWA) showed high levels of serum perfluorooctanoate [40] . Dr . Emmett reported median residential blood levels of 327 ng/ml (interquartile range: 187-572) from a random sample of 74 residents served by LHWA . Levels were high for all members of the sampled group including children, adults, and the aged. Occupational exposure was restricted by limiting the study participants to those who had worked one year or less at the nearby chemical plant during the previous 10 years. Even so, the C8 blood levels approached and sometimes overlapped that of reported worker levels . Dr. Emmett contends that the exposure source is the drinking water and not the air. Twenty-seven percent (27/.) of our study sample obtained their water from LHWA ; our remaining subjects were served by other nearby water districts (see Table I ).
This questionnaire study of the PFOA-exposed population in the Ohio and West Virginia communities near the plant suggests an increased prevalence of several serious health problems . The interpretation of these results depends, however, on several unknown factors . The assumption is that the Parkersburg sample is equivalent to the NHANES sample with regard to all factors that influence the health measures in this study, for example: smoking status, alcohol consumption, and genetic predisposition . If this is the case, the prevalence of these problems is significantly increased compared to National rates .
This study showed an increase in the reported prevalence of respiratory and cardiovascular problems among the PFOA-exposed subjects. These findings are consistent with previous occupational studies of exposed workers [25, 28, 41] . In addition, Gilliland found PFOA alteration in high-density lipoproteins (HDL) in a dose-dependent fashion among PFOA-exposed workers [27) . The alteration was observed only in moderate drinkers, and was considered nonsignificant after adjusting for alcohol use, age, BMI, and cigarette use . Olsen reported a positive association between measured serum PFOS and total cholesterol although he claims that the finding is spurious [2]. He also reports a positive association between PFOA and triglycerides, but discounts the finding because it is inconsistent with previous research on rats . Costa reports a slight increase of total cholesterol in workers exposed, which also showed an increasing trend associated with the highest blood PFOA levels [25) . These findings are significant because now both PFOS and PFOA have been associated with an increase in cholesterol . Most recently, another report describes an inverse relationship between PFOA and I{DL cholesterol coupled with a positive relationship between PFOA and non-HDL cholesterol [26] . These studies suggest a biological mechanism for increased cardiovascular disease in PFAS-exposed populations [25).
The increased prevalence of diabetes in the exposed subjects suggests an interference with insulin function . Studies of monkeys and rats exposed to PFOA revealed increased serum glucose levels [42, 43) .
SELF-REPORTED HEALTH EFFECTS l 139
Many studies of PFAS-exposed populations found elevated liver enzymes [6, 20, 24, 27, 44, 451, although most authors discount their findings . The current study was limited in its ability to evaluate the occurrence of liver problems in this population. The reason is two-fold :
1 . small sample size; and 2. the question asked of the exposed residents is similar, but not precisely the
same as that asked in NHANES .
The question asked in NHANES is "Have you ever had liver problems?" and the question asked in the resident's survey was "Have you ever been told by a doctor that you had a liver disease?" Our question was more restrictive than the comparison question, introducing a definite bias toward the null value making it harder to detect an effect. Nevertheless, we still see large prevalence ratios in the youngest age categories for both males and females demonstrating a potential effect of PFAS exposure on liver function in this population .
The limitations of the data from the residential study include small numbers of cases of rare diseases, lack of a true unexposed comparison group, and potential self-selection bias of study participants . We reduce the limitation of the small sample size by analyzing disease outcomes with relatively large numbers in our data set . National statistics estimated using NHANES data address the need for a comparison group, however, it is not possible to control for potential confounding from variables other than age and gender using these data, and some questions like liver problems or diseases were not strictly comparable to one another . Possible selection bias was examined by evaluating the effect of all cancer prevalence in this group on commonly known and well-understood risk factors : age, gender, smoking status, and educational attainment, which are also reported in a companion paper examining cancer incidence and prevalence in this residential population [30] . The all-cancer rate increased with increasing age, increased smoking, and for those in the lowest category of educational attainment. If one or more of these variables behave outside the norm, we would surmise that this data reflected self-selection bias or reporting bias . However, each category we examined behaved as expected.
Furthermore, the diseases and symptoms identified in this analysis parallel those seen in PFAS-exposed workers . Even if there were selection bias with a disproportionate number of patients with various health problems, there is no reason to assume that only patients with certain types of diseases would be more likely to volunteer for the study_ Our study results are consistent with the literature on PFOA and simply support previously reported findings .
The plant workers in this study were presumed to have larger exposures compared to those who did not work at the plant . Unexpectedly, we did not see a difference in health status. Although puzzling, we did not dismiss our findings in residents because of the lack of differential . Subsequently, we learn from Dr . Emmett's timely study that the residential blood levels in LHWA are
140 / ANDERSON-MAHONEY ET AL.
remarkably similar to those of workers, 60-80 times higher than expected, given commonly measured background levels of PFAS. Although these blood levels were not measured among our study participants, the measurements did come from residents in the same water district serving 27/. of our participants . We feel this adequately explains the lack of difference between worker and non-worker health effects in this study.
Some observers have suggested that plaintiffs in a law suit create an inherent selection bias . The basis for this idea is that the subjects who volunteer to be plaintiffs exaggerate or lie about their symptoms. In addition, a study on the question of selection bias reports that plaintiffs are not more likely to enhance their symptoms than is a normal population (46] .
CONCLUSION In conclusion, living near this Teflon production facility appears to be associated with serious adverse health effects. Additional studies are needed to further delineate the extent and type of health problems associated with PFAS exposure. Such studies are also necessary to define safe levels for these environmental exposures so that governmental agencies can make recommendations and rules to ensure the public's health . The subjects of this study are still living in the same conditions . The water is still contaminated and the plant still continues to dump tons of PFOA every year into the environment. Our findings indicate that chemical companies should reduce their emissions and discharges, and governmental agencies should ensure the public's health by limiting this exposure.
AUTHORS' CONTRIBUTIONS PA and HT assisted in writing the manuscript, JK completed the analysis, DG reviewed drafts of the manuscript, and JD conceived of the study, supervised all aspects of its implementation, and reviewed drafts of the manuscript.
ACKNOWLEDGMENTS
We thank Richard Clapp, D .Sc ., for comments on an earlier draft of the manuscript . We wish to also thank Elaine Nitta and Evelyn Alvarez for their help in revising the manuscript .
REFERENCES
1 . F. M. Hekster, R. W. Laane, and P. de Voogt. (2003) . Environmental and toxicity effects of pertluoroalkylated substances . Rev Enriron Contain Toxicol, 179, 99-121 .
SELF-REPORTED HEALTH EFFECTS 1 141
2. G. W. Olsen, J. M. Burris, M. M. Burlew, and 1. H. Mandel. (2003). Epiderniologic assessment of worker scrnun perlluorooetanesulfonate (PFOS) and per(luorooctanoate (PFOA) concentrations and medical surveillance examinations . J Occup Environ t6fed, 45(3), 260-70 .
3 . K . J. Hansen, L . A . Clemen, M. E . Ellefson, and H. O. Johnson. (2001) . CompoundSpecific, Quantitative Characterization of Organic Fluorochemicals in Biological Matrices . Environmental Science and Technology, 35, 766-770.
4. B . D . Key, R_ D . Howell, and C. S . Criddle. (1997). Fluorinated Organics in the Biosphere . Environmental Science and Technology, 31, 2445-2454.
5. M. M. Schultz, D. F. Barofsky, and J_ A. Field. (2003). Fluorinated Alkyl Surfactants . Environmental Engineering Science, 20(5),487-50 1 .
6 . F. A. Ubel, S . D. Sorenson, and D. E. Roach. (1980) . Health status of plant workers exposed to fluorochemicals-a preliminary report. Am !nd Hyg Assoc J, 41(8),584-9-
7 . 3M . (2001) . Sulfonated Perfluorochemicals in the Environment : Sources, Dispersion, Fate and Effects. USEPA public docket .
8. S. Fields. (2003). Another fast-food fear. Environ Health Perspect, I I I(16), A872 . 9. H. Moriwaki, Y. Takatah, and R. Arakawa. (2003). Concentrations of perfluorooctane
sulfonate (PFOS) and perfluorooctanoic acid (PFOA) in vacuum cleaner dust collected in Japanese homes. JEnviron Monit, 5(5), 753-710- S. Taniyasu, K. Kannan, Y. Horii, N. Hanari, and N. Yamashita. (2003). A survey of perfluorooctane sulfonate and related perfluorinated organic compounds in water, fish, birds, and humans from Japan. Environ Sci Technol, 37(12), 2634-9 . 11 . K. Kannan, J. Koistinen, K. Beckmen, T. Evans, J. F. Gorzelany, K. J_ Hansen, P_ D. Jones, E. Helle, M_ Nyman, and J. P. Giesy. (2001)_ Accumulation of Perfluorooctane Sulfonate in Marine Mammals- Environmental Scienceand Technology, 35,1593-1598 . 12 . J . M . Burris, J. L . Lundberg, G. W . Olsen, C. Simpson, and 1 . Mandel . (2002). Determination of serum half lives of several fluorochemicals . Interim report #2, 3M Medical Department_ USEPA Public Docket . 13 . Revised Draft Hazard Assessment of Perfluorooctanoic Acid and Its Salts . (2002). USEPA Office of Pollution Prevention and Toxics Risk Assessment Division, Public Docket.
14- J_ P. Giesy and K. Kannan. (2001). Global Distribution of Perfluorooctane Sulfonate in Wildlife . Environmental Science and Technology, 35,1339-1342-
15- G . W. Olsen, T_ R. Church, J . P . Miller, l . M . Burris, K_ 1 . Hansen, J. K . Lundberg, J. B. Armitage, R M . Hen-on, Z. Medhdi7adehkashi, 1. B . Nobiletti, E. M . O'Neill, J . H. Mandel, and L. R . Zobel . (2003) . Perfluorooctanesulfonate and other fluorochemicals in the serum of American Red Cross adult blood donors. Environ Health Perspect, 111(16), 1892-1901 .
16- K . Kannan, S . Corsolini, J. Falandysz, G. Fillmann, K. S . Kumar, B . G. Loganathan, M. A . Mohd, J. Olivero, N . Van Wouwe, J. H . Yang, and K . M. Aldoust. (2004). Perfluorooctanesulfonate and related 1luorochemicals in human blood from several countries . Environ Sci Technol, 38(17), 4489-95-
IT F . D_ Griffith and 1_ E . Long. (1980). Animal toxicity studies with ammonium perfluorooctanoate . Am Ind HygAssoc J, 41(8), 576-83-
18 . L . B . Biegei, R. C . Liu, M . E . Hurtt, and J . C. Cook . (1995) . Effects of ammonium perfluorooctanoate on Leydig cell fimction : in vitro, in vivo, and ex vivo studies. Toxicol Appl Pharmacol, 134(t) .18-25 .
142 1 ANDERSON-MAHONEY ET AL_
19. L_ B. Biegel, M . E. Hunt, S . R Frame, J . C . O'Connor, and J. C. Cook . (2001). Mechanisms of extrahepatic tumor induction by peroxisome proliferators in male CD rats. Toxicol Sci, 60(1), 44-55 .
20. F. D. Gilliland_ (1992) . Fluorocarbons and Human Health: Studies in an Occupational Cohort. University of Minnesota. USEPA Public Docket.
21 . G. W. Olsen, M . M. Burlew,1. M. Burris, and 1. H. Mandel. (2001)_ A cross-sectional analysis of serum perfluorooctanesulfonate (PFOS) and pertluorooctanate (PFOA) in relation to clinical chemistry, thyroid hormone, hematology, and urinalysis results from male and female employee participants of the 2000 Antwerp and Decatur fluorochemical medical surveillance program. USEPA Public Docket_
22 . L . A. Kinney, N_ C. Chromey, and G . L. Kennedy Jr (1989). Acute inhalation toxicity of ammonium perfluorononanoate_ Food Chem Toxicol, 27(7),465-8 .
23 . J. Butenhoff, G . Costa, C. Elcombe, D . Farrar, K_ Hansen, H. Iwai, R . lung, G . Kennedy Jr, P. Lieder, G . Olsen, and P. Thomford. (2002). Toxicity of ammonium perfluorooctanoate in male cynomolgus monkeys after oral dosing for 6 months. Toxicol Sci, 69(l), 244-57 .
24 . G_ W_ Olsen, 1_ M. Burris, M . M_ Burlew, and 1 . H . Mandel. (2000)_ Plasma cholecystokinin and hepatic enzymes, cholesterol and lipoproteins in ammonium perfluorooctanoate production workers . Drug Ckem Toxicol, 23(4), 603-20.
25 . G_ Costa_ (2004). Ammonium Perflurooctanoate . USEPA Public Docket, Risk Assessment Division.
26 . M. Kaplan. (2005)_ Ammonium Perfluorooctanoate : Cross Sectional Surveillance of Clinical Measures ofGeneral Health Status Related to a Serum Biomarker of Exposure and Retrospective Cohort Mortality Analysis in a Polymer Production Plant . USEPA Public Docket.
27 . F. D . Giltiland and J_ S. Mandel. (1996) . Serum perfluorooctanoic acid and hepatic enzymes, lipoproteins, and cholesterol : a study of occupationally exposed men_ Am J Ind Med, 29(5), 560-8 .
28 . R . C . Leonard. (2003). Epidemiology Surveillance report Cancer Incidence for Washington Works Site 1959-2001_ USEPA Public Docket .
29_ B . H. Alexander, G . W. Olsen,1 . M . Burris, J . H. Mandei, and J_ S . Mandel. (2003)_ Mortality of employees of a perttuorooctanesulphonyl fluoride manufacturing facility_ Occup Environ Med, 60(10), 722-9 .
30_ J. G. Dahlgren, P . J_ Anderson-Mahoney, J . Kotierman, H . Takhar, A . Lee, T. Robinson, M_ Zwass, R Warshaw, and A_ B . Ibrahim . Prevalence and Incidence of Cancer in Populations to PFOA . Archives ofEnvironmental Health . Under review.
31 . B. Boulanger, 1. Vargo, J. L_ Schnoor, and K. C_ Hombuckle. (2004). Detection of perfluorooctane surfactants in Great Lakes water. Environ Sci Teclenol, 38(15), 4064-70-
32 . J. Dahlgren, R_ Warshaw, 1. Thornton, C. P_ Anderson-Mahoney, and H_ Takhar . (2003)_ Health effects on nearby residents of a wood treatment plant. Environ Res, 92(2),92-8 .
33 . S. M_ Pinney, R. W. Freyberg, G. E. Levine, D. E. Brannen, L. S. Mark, J_ M. Nasuta, C_ D. Tebbe, J. M_ Buckholz, and R Wones_ (2003)_ Health effects in community residents near a uranium plant at Femald, Ohio, USA. Int J Occup Med Environ Ilealth, 16(2),139-53-
SELF-REPORTED HEALTH EFFECTS / 143
34 . National Center for Health Statistics, 1992 . (On-line). Available: bttp ://www.cdc .gov/nchslnhanes_htm
35 . J. P. Giesy and K_ Kannan . (2002). Perfluorochemcial surfactants in the environment. Environmental Science and Technologv, 36, 147A-152A.
36 . A . M. Seacat, P . J. Thomford, K . 1 . Hansen, G. W . Olsen, M_ T . Case, and J . L . Butenhoff. (2002). Subchronic toxicity studies on perfluorooctanesulfonate potassium salt in cynomolgus monkeys. Toxicol Sci, 68((), 249-64 .
37 . F . D. Gilliland and J. S . Mandel . (1993)_ Mortality among employees of a perfluorooctanoic acid production plant. J Occup Med, 35(9), 950-4 .
38 . P. M. Hinderliter and G . W . Jepson. A Simple, Conservative Compartmental Model to Relate Ammonium Perfluorooctanate (APFO) Exposure to Estimates of Perftuorooctanoate (PFO) Blood Levels in Humans_ 2001 . USEPA .
39 . A . Hartten. (2D02) . 3Q02 and 4Q02 Public Water Supply results, West Virginia and Ohio_ USEPA Public Docket.
40. E. A. Emmett, F. S. Shofer. H. Zhang, D. Freeman, C. Desai, and L. M. Shaw_ (2006) . Community exposure to perfluorooctanoate: relationships between serum concentrations and exposure sources. J Occup Em"iron Med, 48(8), 759-70.
41 . G_ W. Olsen, M_ M . Burlew, B_ B . Hocking, J . C. Skratt, J . M . Burris, and J. H_ Mandel_ (2001). An Epidemiologic Analysis of Episodes of Care of 3M Decatur Chemical and Film Plant Employees, 1993 - 1998. USEPA .
42 . E. I . Goldenthal . (1978) . Ninety Day Monkey Toxicity Study. Final Report USEPA Public Docket_
43_ E . 1 . Goldenthal . (1978)_ Ninety Day Rat Toxicity Study . Final Report_ USEPA . 44. W. E. Fayerweather_ (1980)_ Liver Study of Washington Works Employees Exposed
to C8 : Results of Blood Biochemistry Testing_ USEPA Public Docket . 45 . G. W . Olsen, D_ C . Madsen, J . M . Burris, et al_ (2001). Descriptive Summary of
Senun Fltrorochemical Levels Among 236 Building Employees . Final Report. USEPA Public Docket. 46 . S. L_ Allred and J . R_ Burg . (1995). Environmental personal injury litigation as one source of response effects: findings from the National Exposure Registry_ Toxicol lnd llealth, 11(2), 217-30 .
Direct reprint requests to :
Pamela Anderson-Mahoney 13627 Hatteras St. Van Nttys, CA 91401 e-mail : pamela.mahoney@epidemiologyresources .com
