EXPOSURE INVESTIGATION REPORT
SOLUTIA INCORPORATED/MONSANTO COMPANY
ANNISTON, CALHOUN COUNTY, ALABAMA
In the United States, no study of PCB blood levels in a statistically-based
sample of the population has been conducted. Therefore, there is no national
reference range that can be used as a comparison population for this EI. However,
several studies have measured PCB levels in populations that had no known exposures
to PCBs other than typical background levels. The results from these studies
are listed in Tables 1 and 2.
As indicated in Table 1, mean background PCB levels in adults range from 3.7
to 6.8 ppb. The geometric mean and median PCB values calculated from these studies
are somewhat lower, since these statistical measures of central tendency reduce
the influence of individuals with unusually high PCB levels in the test population.
In comparing the results of these studies, several caveats must be considered.
First, different analytical methodologies were used, which could account for
some of the variability between studies. Earlier studies used packed column
gas chromatography with electron capture detectors; more recent studies have
used glass capillary columns with mass spectroscopy to detect individual PCB
congeners. In addition, different populations were studied, and background PCB
exposures of these populations could vary. In the United States, dietary intake
of PCBs is thought to be the major source of exposure .
Therefore, regional variations in food consumption patterns could explain some
of the variability in the studies. The age of the participants in these studies
also varied. Since blood PCB levels increase with age [6,7],
some of the variability between studies might be accounted for by the different
age distributions of the study populations. Finally, these studies were conducted
over the time period, 1982-1995. Since production of PCBs in the United States
was halted in 1977, PCB releases to the environment have decreased. This could
cause PCB blood concentrations to decrease over time, although this speculation
has not been adequately documented.
In spite of these limitations and differences between studies, the results
are fairly consistent. These studies suggest that mean PCB levels in blood serum
from adults without unusual PCB exposures are 3 to 7 ppb, with lower levels
being reported in more recent studies.
The upper end distribution of PCB levels in the general population has not
been well characterized. In a review paper, Kreiss estimated that the 95th
percentile PCB blood concentration in adults is 20 ppb . Since this estimate
was based on studies conducted in the late 1970s and early 1980s, it is likely
that the 95th percentile level would be lower today. However, in
the absence of a more recent estimate, ATSDR will assume that a blood PCB level
in excess of 20 ppb is significantly elevated.
In this EI, five adults had a blood PCB level (22, 54, 93, 97, and 210 ppb)
that exceeded 20 ppb. It was noted that the blood serum sample from the individual
with 22 ppb PCBs was hyperlipidemic (total lipid = 1,417 mg/dl). Following a
heavy fat meal, blood PCB levels can be temporarily elevated by the transient
lipidemia . Therefore, the elevated blood PCB level in this individual may
have been spuriously elevated by the presence of hyperlipidemia.
If the five individuals with elevated PCB levels are excluded, the mean PCB
level in the rest of the adult population is 3.5 ppb, which is within the normal
background range. Therefore, these results indicate that PCB exposures for most
of the adult EI population were within the normal background range.
Blood concentrations of PCBs were also calculated as a blood lipid concentration.
For adults, the mean blood concentration was 2,537 ng/g lipid, and the median
concentration was 392 ng/g lipid. The large difference in these values indicates
that the mean value is skewed by the inclusion of the elevated individuals.
If the five highest concentrations are not included, the mean PCB concentration
in the rest of the adult population is 544 ng/g lipid.
There are few studies in the published scientific literature that have reported
blood concentrations of PCBs on a lipid-adjusted basis, so only limited comparisons
can be made between the EI participants and other populations with background
exposures. In a study of non-fasting adults, a mean blood serum PCB concentration
of 666 ng/g lipid was reported . In another study of Canadian adults, the
median blood plasma PCB concentration (50th percentile) was reported
to be 263 ng/g lipid . However this value is likely to be less than the actual
PCB total, since only 15 PCB congeners were included in the total, and the blood
samples had been diluted 10-15 percent by a citrate anticoagulant. Therefore,
based on these limited studies, the lipid-based PCB concentrations in most of
the EI participants were comparable to those reported in other normal populations.
Relatively few studies have measured PCB levels in children. However, based
on two published studies (Table 2), PCB levels in blood serum from children
are about 2 to 4 ppb, which is less than for adults. In the EI population ,
the average blood PCB level in children was 0.37 ppb, and the median value was
0. These levels are less than the values cited in the reference studies.
Only 10 of 37 children tested in the EI had a detectable PCB level in their
blood serum. The highest blood level detected in a child was 4.6 ppb, which
was detected in an older child (a 13-year old girl). Reference ranges for PCB
concentrations in teenagers are not available, but they would likely be between
those for young children and adults. Nevertheless, all of the PCB concentrations
detected in children from the EI were within the ranges detected in the reference
studies for children (Table 2). ATSDR concludes that none of the children tested
in the EI had a significantly elevated blood PCB level.
Because PCBs are resistant to metabolism
in the body, they bioaccumulate as a person ages. Therefore, blood PCB levels
tend to increase with age. In Figure 1, the blood PCB levels in the EI participants
are plotted as a function of age. As indicated, all of the people with elevated
blood PCB levels (> 20 ppb) are aged 45 or older. In people below the age
of 40, none of the PCB concentrations exceeded 8.2 ppb. The absence of elevated
PCB levels in people below 45 years old suggests that PCB exposures in the past
may have exceeded more recent exposures.
To assess the relationships between blood PCB levels, age, and length of residency
in the vicinity of the manufacturing plant, Spearman rank correlation coefficients
were calculated. Age was strongly correlated with the blood PCB level (rs
= 0.729, p < 0.001). Length of residency was also correlated with the blood
PCB level (rs = 0.645, p < 0.001). By calculating the Spearman
partial correlation coefficient, it was determined that if age were controlled
for, there was still a significant correlation between the blood PCB level and
length of residency in the vicinity of the Solutia facility (rs =
0.310, p = 0.0054). This suggests that people were exposed to PCBs while living
at their current residences, although it is not known when the exposures occurred
or what the source of PCBs was.
PCB Congener Analyses
There are 209 possible congeners of PCBs, which differ in the number and position
of the chlorine atoms on the biphenyl ring structure. The congeners that contributed
most to the total blood concentration of PCBs in most of the EI participants
were PCB congeners 153, 138/158, 180, 187, and 118 (IUPAC designation). In general,
these congeners accounted for more than 50 percent of the total PCBs in individuals with elevated or normal PCB levels.
Several studies have reported that these same congeners are the major ones
detected in blood, breast milk, and adipose tissue samples from other human
populations [10, 11]. The presence of relatively high concentrations of these
congeners in humans is likely related to several factors: (1) they are major
constituents of commercial PCB mixtures (e.g., Aroclor 1260), (2) they bioconcentrate
in animal food chains, and (3) they resist metabolic degradation.
To interpret the PCB congener profiles in the EI participants, ATSDR employed
principal components analyses (PCA) [see description in Methods section].
As shown in Figure 2, most of the children were in Cluster A, and most of the
adults were in Cluster B. In general, the blood samples from children in Cluster
A contained the major PCB congeners discussed above, but they lacked other,
minor PCB congeners that were detected in individuals in Cluster B. In addition,
the children generally had lower total PCB levels.
All of the adults with elevated blood PCB levels grouped in Cluster B, along
with other adults with normal PCB levels. This indicates that the PCB profiles
in adults with normal or elevated PCB levels were similar; they differed only
in the total concentration of PCBs.
Two of the EI participants (A* and C*) had distinctly different PCB profiles.
These two individuals were teenagers who lived in different homes. Blood samples
from both of these individuals contained several lower chlorinated PCB congeners.
Lower chlorinated congeners tend to be more susceptible to metabolic degradation;
hence, they have shorter biological half-lives. In particular, PCB congeners
52 and 44, which lack chlorine atoms on the meta and para
position of the biphenyl ring, are especially susceptible to metabolic degradation
. Therefore, the presence of these congeners in human blood serum suggest
relatively recent exposure. Nevertheless, the total concentrations of PCBs in
blood samples from these individuals (0.66 and 4.6 ppb) were within the normal
The blood levels of PCBs that cause adverse health effects have not been well
characterized. In occupational studies, exposures to PCBs have been associated
with chloracne and subtle evidence of liver damage, although it is possible
that polychlorinated dibenzofuran impurities in the PCBs might have contributed
to the toxicity [13, 14]. Slight elevations in serum enzymes of hepatic origin,
such as gamma-glutamyl transferase, have been reported in workers with blood serum
PCB levels of several hundreds of parts per billion . However, in occupational
studies, it is difficult to unequivocally attribute the observed health effects
to PCBs because of the possible confounding effect of concurrent exposure to
other chemicals. Nevertheless, based on one such study, a blood PCB level of
200 ppb was suggested as a no effect level .
Based on this criterion, only one of the EI participants had a blood PCB level
that exceeded this level of concern. However, it should be recognized that occupational
studies have examined only a limited number of health endpoints and have limited
statistical power to detect an effect. Therefore, no firm conclusions can be
drawn regarding a safe blood level for PCBs in adults.
Several studies have reported that low level PCB exposure during fetal or neonatal
development can effect the infant's neurobehavioral development [15, 16]. However,
several limitations of these studies have been noted: (1) possible exposure
to other neurotoxic chemicals besides PCBs (e.g., dioxins, mercury, lead, or
organochlorine pesticides) that may have contributed to the effects, (2) inadequate
control for confounding socioeconomic variables such as maternal smoking, alcohol,
and other drug use, and (3) inadequate control for maternal birth weight and
nonspontaneous deliveries [17, 18]. In addition to these methodological limitations,
different studies have measured different neurobehavioral endpoints, which impedes
comparisons between studies.
Therefore, these studies suggest, but do not conclusively prove, an association
between prenatal or neonatal exposures to PCBs and neurobehavioral and developmental
effects in young children. Furthermore, these purported effects were reported
to occur in populations with background exposures to PCBs, so a threshold effect
level has not been defined. Because of these limitations, it cannot be determined
whether the low level PCB exposures detected in the EI participants could affect
neonatal development. Nevertheless, to minimize any potential health risks,
it is prudent public health policy to reduce exposure to environmental PCB contamination.
The concentration of PCBs detected in composite surface soil samples from 19
homes ranged from non-detected to 11.7 ppm. Soil samples from four homes contained
PCB concentrations in excess of 1 ppm (11.7, 5.14, 1.31, and 1.20 ppm).
At Superfund sites, the U. S. Environmental Protection Agency (EPA) has set
a Recommended Soil Action Level - Analytical Starting Point of 1 ppm PCBs for
residential properties . Higher action levels may be set if warranted by
The concentration of PCBs detected in house dust samples from 18 homes ranged
from non-detected to 10.3 ppm. House dust samples from two homes contained PCB
concentrations in excess of 1 ppm (10.3 and 1.71 ppm). The EPA has not established
an action level for PCBs in house dust.
By calculating the Spearman rank correlation coefficient, ATSDR was able to
show that there was a significant correlation between the concentrations of
PCBs in soil and house dust samples from individual homes (rs = 0.628,
p < 0.0052). This finding was expected since soil constitutes about 50 percent
of the mass of house dust .
Statistical analyses failed to demonstrate a significant correlation between
blood PCB levels and the concentration of PCBs in either soil (rs
= -0.128, p = 0.26), house dust (rs = 0.078, p = 0.51), or the house
dust loading (ng/ft2) concentration (rs = 0.046, p = 0.70).
Further analyses in which the EI population was divided into adults and children
also failed to demonstrate any significant correlations between blood and environmental
PCB levels. The absence of a correlation between blood PCB levels and soil or
house dust PCB levels suggests that residents have been exposed to other sources
of PCBs besides those in soil and house dust.
Characterization of PCB Exposures
In the general population, the major background source of PCB exposure is from
food . Among foodstuffs, the major contributors to the body burden of PCBs
are fish, meat, and poultry. It is likely that trace levels of PCBs in comercially-available
foods have contributed to the body burden of PCBs in the EI participants. However,
for those EI participants with elevated PCB levels, additional sources of PCB
exposure are likely.
The EI participants with elevated blood PCB levels share several characteristics:
(1) They are older adults, aged 45 and above. (2) They reported no known occupational
exposure to PCBs. (3) They grew up in neighborhoods near the facility and have
lived there most of their lives. (4) They ate locally-grown fruits and vegetables
and locally-raised chickens and eggs. In addition, three of the five individuals
with elevated blood PCB levels reported eating clay that they collected in the
In recent years, the potentially responsible party has purchased, remediated,
and restricted access to some off-site, PCB-contaminated properties. However,
it is likely that prior to remediation, long-term residents of the area had
exposure to environmental contamination by direct contact with contaminated
soil, sediment, water, and air, and by indirect contact from eating locally-raised
animal or plant foodstuffs.
Once PCBs get inside the body, they are resistant to metabolic degradation
and are stored in adipose tissue for long periods of time. The biological half-life
of PCB congeners vary, but for PCB mixtures, a collective half-life of 2 to
6 years has been estimated . Therefore, the
high body burdens of PCBs seen in some long-term residents may be the result
of exposures that occurred years ago. This conclusion is supported by the absence
of elevated blood PCB levels in younger residents of the community (Figure
1), as well as by the strong correlation between blood PCB level and length
Therefore, the available evidence suggests that past exposures to environmental PCB contamination exceeded current exposures. Nevertheless, this study and others conducted by the EPA have documented that elevated levels of PCBs remain in off-site soils and sediments. Since the future use of these areas cannot be predicted, they should be remediated to prevent further human exposure to environmental contamination.