August 12, 1998 FINAL DRAFT
HUMAN VISUAL FUNCTION IN THE NORTH CAROLINA CLINICAL STUDY ON
PFIESTERIA PISCICIDA
H. Kenneth Hudnell
U. S. Environmental Protection Agency, National Health and Environmental Effects Research
Laboratory, Neurotoxicology Division, RTF, NC 27711
Manuscript prepared at the request of the the North Carolina Department of Health and
Human Services for review by the North Carolina Task Force on Pfiesteria (William Roper,
Chairman)
Address for Correspondence: HK Hudnell, PhD
US Environmental Protection Agency
National Health and Environmental Effects Research Laboratory
Neurotoxicology Division
MD-74B
RTP,NC27711
Ph.:919-541-7866
Fx.:919-541-4849
E-mail:hudnell.ken@epamail.epa.gov
This manuscript was reviewed by the National Health and Environmental Effects Research
Laboratory, US Environmental Protection Agency, and approved for publication. Approval does
not signify that the contents necessarily reflect the views and policies of the Agency, nor does
mention of trade names or commercial products constitute endorsement or recommendation.
-------
August 12, 1998 HUDNELL2
ABSTRACT
Human Visual Function in the North Carolina Clinical Study on Pfiesteria piscicida.
Hudnell, H.K., U.S. EPA, RTF, NC. The U.S. Environmental Protection Agency assisted the
North Carolina Department of Health and Human Services in conducting a human-health study to
investigate the potential for an association between fish kills in the NC estuary system and the risk
for adverse human-health effects. Impetus for the study was recent evidence suggesting that the
estuarine dinoflagellate, Pfiesteria piscicida, may release a toxin(s) which kills fish and adversely
affects human health. This report describes one component of the study in which visual system
function was assessed. Study participants worked primarily in estuaries inhabited by P. piscicida
or in off-shore waters thought not to contain P. piscicida. The potentially exposed estuary
(N=22) and unexposed offshore (N=20) cohorts were well matched for age, gender, and
education, but less well matched for occupation. Visual acuity did not differ between the cohorts,
but visual contrast sensitivity (VCS), an indicator of visual pattern-detection ability for stimuli of
various sizes, was significantly reduced in the estuary cohort relative to the offshore cohort. A
further analysis which excluded participants having a history predictive of neuropsychological
impairment also showed significantly reduced VCS in the estuary cohort (N=14) relative to the
offshore cohort (N=10). Additional analyses indicated that differences between the cohorts in age,
education, smoking, alcohol consumption, and total time spent on any water did not account for
the difference in VCS. Finally, an analysis which excluded members of the estuary cohort who
may not have had direct contact with an active fish kill, as well as offshore participants who may
have had direct contact, also indicated that VCS was significantly lower in the estuary (N=17)
than the offshore (N=17) cohort. The profile of VCS deficit across stimulus sizes resembled that
seen in organic-solvent exposed workers, but an assessment of solvent and other neurotoxicant
exposures did not indicate differences between the cohorts. These results suggest that factor(s)
associated with the NC estuaries, including the possibility of exposure to P. piscicida toxin(s) at
active fish kills, may impair visual system function.
-------
August 12, 1998 HUDNELL 3
INTRODUCTION
The North Carolina Department of Health and Human Services, in collaboration with several
universities and the U.S. Environmental Protection Agency, conducted a human-health study in
the Fall of 1997 to investigate the potential association between fish kills in the North Carolina
estuary system and human-health status. Impetus for the study was recent evidence suggesting
that the estuarine dinoflagellate, Pfiesteria piscicida, may release a toxin(s) which kills fish
(Burkholder et al., 1992, 1995) and adversely affects human health, particularly neurological
function, in laboratory (Glasgow et al., 1995) and environmental (Morris et al., 1997; Bever et al.,
1998; Golub et al., 1998; Grattan et al., 1998; Greenberg et al., 1998; Lowitt et al., 1998; Tracy
et al., 1998) settings. A multi-component clinical evaluation was conducted to compare health
status in: 1) two occupational cohorts, one with (the estuary cohort) and one without (the
offshore cohort) potential for exposure to P. piscicida toxins; and 2) a case-control series in
which the cases self-reported to the NC Pfiesteria Hot Line suspicions of having been affected by
exposure to P. piscicida toxin(s). This report describes one component of the evaluation in which
visual system function was assessed.
Visual System Tests
Three tests of visual function were administered to participants in the current study. Two of the
tests, visual contrast sensitivity (VCS) and visual acuity, were previously recommended by a panel
of neurotoxicologists (Anger et al., 1994) for inclusion in a battery of core tests being assembled
by the Agency for Toxic Substances and Disease Registry's (ATSDR) for use in environmental
health field studies. Both tests were included in batteries designed for detecting subtle neurotoxic
effects in adults (ATSDR, 1995) and children (ATSDR, 1996).
VCS is a measure of the ability to detect visual patterns (Ginsberg, 1984; Ginsberg et al., 1984).
Whereas standard tests of visual acuity measure the visual system's resolution limit for high
contrast stimuli, a task critically dependent on the functional integrity of the eye's physiological
optics system, VCS is primarily an indicator of neurological function in the visual pathways from
the retina to the cortex (Bodis-Wollner et al., 1986). The VCS test measures the least amount
(threshold) of luminance difference (contrast) between adjacent areas necessary for an observer
to detect a visual pattern. Contrast (C) is defined as: C = (Lnu* - Lmin) / 0-max+ ^nm) where L^
and L^are the luminances of the brighter and darker areas, respectively. VCS is the inverse of
contrast threshold. A simple card test, the Functional Acuity Contrast Test (F. A.C.T.), measures
contrast sensitivity for five sizes (spatial frequencies) of light and dark bar patterns (sinusoidal
gratings) because spatial vision is mediated by populations of neurons selectively tuned to
different spatial frequency (Bodis-Wollner et al., 1986). If neurons subserving low spatial-
frequency (larger bars) vision are functionally impaired but those underlying high spatial-
frequency (smaller bars) vision are functionally normal, for example, then visual perception also
will be impaired for low frequency patterns but normal for high frequency patterns.
-------
August 12, 1998 HLTDNELL 4
A standard test of visual acuity was administered to participants because VCS deficits at high
spatial frequency could result from either refractive error or neurological dysfunction. A deficit in
VCS at high spatial frequency in the presence of normal visual acuity indicates neurological
dysfunction (Bodis-Wollner et al., 1986). VCS at low-to-mid spatial frequencies is unaffected by
moderate acuity deficits (Bodis-Wollner et al., 1986). It is important to note that normal visual
acuity is often found in neurotoxicological studies which show a significant reduction in VCS
(Mergler et al., 1991; Frenette et al., 1991; Broadwell et al., 1995; Hudnell et al., 1996a,b,c). This
arises from the fact that the acuity test engages only those neurons selectively tuned to small
stimuli, and the stimuli are always of extremely high contrast. The VCS test, on the other hand,
assesses visual function across stimulus sizes and engages neurons sensitive to low contrast in
determining the sensitivity of the system to visual patterns.
A color discrimination test was also administered to screen for congenital color blindness and
color vision deficiencies because severe dyschromatopsia could impair performance on tests of
cognitive function which use chromatic stimuli, such as some neuropsychological tests and tests in
the Neurobehavioral Evaluation System 2 (NES2; Baker et al., 1985), which was used in the
current study. In addition, the color discrimination test included a condition designed to detect the
failure to perform at the level of one's ability due to malingering or low motivation. Together,
these three tests provided a basis for assessing group differences in visual function.
Rationale for Test Selection
Tests of visual function were included in the current evaluation for two primary reasons. Visual
function is a sensitive indicator of neurotoxicity (Boyes, 1994; Mergler, 1995) and an important
determinant of performance on tests designed to assess motoric and cognitive functions (Hudnell
et al., 1996c). Two studies of mixed volatile-organic compound exposure observed VCS deficits
in microelectronics-fabrication workers relative to unexposed control workers matched with
exposed workers for age, gender, ethnicity, and education (Mergler et al., 1991; Bowler et al.,
1991; Frenette et al., 1991; Broadwell et al., 1995; Hudnell et al., 1996a). Both studies observed
a unique VCS profile across spatial frequencies; large VCS deficits were observed at mid-spatial
frequencies with little or no deficit at higher and lower spatial frequencies. Although the
magnitude of the VCS deficits was about 20%, severity was at a sub-clinical level; subjects had
not been diagnosed with visual anomalies and generally attributed the reduction to normal ageing.
Yet the significance of these deficits is striking in that participants of both studies had received
little or no exposure for over a year prior to testing, suggesting that the deficits were permanent
or long lasting. A study of patients previously diagnosed with organic-solvent-induced, chronic-
toxic encephalopathy found a virtually identical pattern of VCS loss in the absence of recent
exposure (Donoghue et al., 1995). Among currently exposed styrene workers, VCS reductions at
the mid-spatial frequencies were significantly and inversely associated with end-of-shift urinary
mandelic acid (a styrene metabolite) concentration (Campagna et al., 1995). Very recent data
collected uy EPA in CuupcittUuu with the New York State Department of Health suggested that
people living in apartments above dry-cleaning facilities were at risk for alterations in VCS
(Schreiber et al., 1998). Concentration of the dry cleaning solvent perchloroethylene (i.e.
-------
August 12, 1998 HUDNELL 5
tetrachloroethene) in apartment air was only about 1 ppm, but VCS was significantly reduced in
17 adults and children relative to matched-control subjects and showed the mid-spatial frequency
deficit characteristic of solvent exposure. These results were consistent with a report of visual
reaction-time and visual memory deficits in a German population with comparable environmental
exposure to perchloroethylene (Altman et al., 1995). Analyses of the influence of VCS on results
from visual reaction-time and visual memory tests (see below) suggested that VCS deficits may
have been at least partially responsible for the reaction-time and visual memory deficits. Similar
effects were observed in cohorts occupationally exposed to perchloroethylene (Seeber, 1989;
Echeverria et al., 1995), and a follow-up study of previously exposed workers reported further
decline in sensory and motoric, but not cognitive, functions after a mean exposure-free duration of
5.9 years (Lindstrom et al., 1982). The consistency of results from these studies, in conjunction
with animal studies (Merigan et al., 1988; Boyes, 1994), suggest that a variety of organic solvents
at relatively low exposure levels may act on a common mechanism to degrade mid-spatial
frequency pattern vision.
Neurotoxicant induced deficits in VCS are not limited to solvent exposures. Visual-pattern
evoked potentials (Hudnell et al., 1990a), a method for assessing pattern vision in animals and
humans (Hudnell et al., 1990b; Hudnell and Boyes, 1991; Benignus et al., 1991), collected from
rats indicated that acute exposures to several classes of pesticides, metals, and other compounds
degrade pattern perception (reviewed in Boyes, 1992). Evidence is also mounting that some
heavy metals may induce VCS deficits in humans. Measurements of VCS in children exposed to
the combustion products of soft-brown coal in the Czech Republic revealed a pattern of low-to-
mid spatial frequency loss. A significant association with methyl mercury body burden was
observed (Hudnell et al., 1996b). Other studies have shown associations between VCS loss and
methyl mercury (Mukuno et al., 1981; Lebel et al., 1996) and inorganic mercury exposures in
adults (Cavalleri et al., 1995) and children (Altman et al., 1998). The VCS spatial-frequency
profiles observed in the mercury exposed populations showed no evidence of the mid-spatial
frequency selective deficit seen in solvent-exposed cohorts. These results are consistent with
observations of methyl mercury-induced reductions of VCS in monkeys (Rice and Gilbert, 1982,
1990; Merigan etal., 1983).
Clinical studies have demonstrated that the VCS test is sensitive to the neurological dysfunction
associated with many diseases affecting the nervous system. Ocular diseases, such as glaucoma,
which manifests a low spatial-frequency deficit (Atkin et al., 1980; Ross et al., 1985; Sample et
al.., 1991), macular disease (Loshin and White, 1984; Greeves et al., 1988), retinitis pigmentosa
(Gawande et al., 1989; Seiple et al., 1993; Alexander et al., 1992, 1995), Type 1 diabetes with
little or no retinopathy (Sokol et al., 1985; Trick et al., 1988 Bangstad et al., 1994), and other
conditions (Bodis-Wollner and Camisa, 1980; Regan and Neima, 1984), produce a variety of
alterations in the VCS spatial-frequency profile. VCS deficits, as well as color discrimination
deficits (Mergler et al., 1987), are commonly present prior to detectible pathology in the retina or
optic nerve head, making this one of the earliest sign of disease (Regan, 1989). With damage
more proximal to the visual cortex, VCS deficits have been observed in cases of optic-nerve
neuropathy (Bodis-Wollner, 1983), optic-nerve compression (Kuppersmith et al., 1982), and
-------
August 12, 1998 HUDNELL 6
cerebral lesions (Bodis-Wollner and Diamond, 1976). Patients that have recovered from optic
neuritis with normal visual acuity retain severe deficits in VCS (Fleishman et al., 1987).
Neurodegenerative diseases that are not well known for their effects on vision also manifest VSC
deficits. Multiple sclerosis patients display VCS deficits which are orientation specific, suggesting
cortical rather than retinal or optic nerve damage (Camisa et al.,1981). A primarily low spatial-
frequency VCS deficit is present in Parkinson's (Regan and Neima, 1984) and Alzheimer's (Sadun
et al., 1987; Cronin-Golomb et al., 1991; Gilmore and Levy, 1991) patients, the latter of whom
show an extent of cognitive impairment predicted by VCS scores (Cronin-Golomb et al., 1995).
AIDS patients display marked color-vision and VCS deficiencies (Quiceno et al., 1992), and
cystic fibrosis patients show a VCS deficit across spatial frequencies which is either secondary to
a vitamin A deficiency (Leguire et al., 1991) or, more likely, a primary manifestation of cystic
fibrosis since the deficit is seen in patients taking vitamin A supplements (Morkeberg et al., 1995).
Micronutirent deficiencies, including the vitamin B complex, are associated with reversible VCS
deficits, as seen in the "Cuban epidemic optic neuropathy" cases of the early 1990s (Sadun et al.,
1994; Roman, 1994). These studies suggest that the perception of visual patterns, as indicated by
VCS, may be an apical endpoint that is commonly altered by a variety of clinical conditions which
affect neurophysiologjcal structures or biochemical processes in the visual pathways from the
retina to the cortex.
VCS deficits thought to be congenital are associated with learning disabilities in children. Earlier
research indicated that low VCS was prevalent among children with reading disabilities
(Lovegrove et al., 1980) and dyslexia (Lennerstrand and Ygge, 1992). Recent evidence suggests
that VCS deficits greatest at mid-to-high spatial frequency may be widespread among children
with various types of learning disabilities (Hudnell et al., 1996b). This same distortion of the
contrast sensitivity function was subsequently observed in Down syndrome children (Courage et
al., 1997). A similar pattern of mid-to-high spatial frequency VCS reduction was seen in monkeys
treated with acrylamide monomer which caused severe degeneration of the parvocellular
retinogeniculate pathway while sparing the magnocellular pathway (Merigan et al., 1985, 1989,
1991). These results suggest that dysfunction in the parvocellular pathway may underlie mid-to-
high spatial frequency VCS deficits.
Therefore, VCS deficits are associated with many abnormal neurological conditions, making the
VCS test well suited as a tool for neurological health screening. Variations in the pattern of VCS
loss across spatial frequencies are often associated with particular diseases and neurotoxicants,
which increases the power of the VCS to assist in differential diagnostics. The above mentioned
studies have demonstrated the potential for the VCS test as an aid in the diagnosis of acquired and
congenital clinical conditions, as well as a tool for detecting neurotoxicant-induced subclinical _
deficits.
VCS appears to be not only a sensitive indicator for the adverse-health effects of a broad range of
neurotoxicants and clinical conditions, but also is an important factor in the interpretation of
computerized-neurobehavioral test data. Computerized tests are designed to assess a number of
specific cognitive functions, for example visual memory and attention. These tests involve small or
-------
August 12, 1998 HUDNELL 7
briefly presented visual stimuli, and test endpoints often incorporate measurements of response
times which may be as short as 150 ms. VCS is known to strongly influence detection of stimuli
and, in turn, response times (Felipe et al., 1993; reviewed in Hudnell et al., 1996c). Yet
historically, measures of visual function have not been used in the assessment of cognitive
functions as indicated by computerized-test performance. The analysis of several data sets from
environmental-epidemiological studies indicated that VCS can account for up to 24% of the
variance hi scores from NES tests (Hudnell et al., 1996c). In our study of solvent-exposed
microelectronics fabricators (Broadwell et al., 1995; Hudnell et al., 1996a), VCS accounted for
18% of the variance in the simple-reaction-time test and with 17% of the variance in a pattern-
memory test (Hudnell et al., 1998). These proportions were based on analyses of data collected
from the unexposed control subjects, rather than the exposed subjects, to avoid the potential for
correlated deficits in visual and cognitive functions induced by neurotoxicant exposure. A model
was developed to remove the influences of vision from the data of both control and exposed
participants in order to better assess the effects of neurotoxicant exposure on cognitive functions.
Along with complementary procedures for calibrating the luminance and contrast of test stimuli
on video screens (Hudnell et al., 1996c), the analytical model helped to both more accurately
attribute performance deficits to the visual or cognitive domains and to reveal group differences in
cognitive performance which were obfuscated by random differences in VCS. This effort was
extended to include assessment and statistical control of motoric influences, as indicated by
finger-tapping performance, on computerized test performance (Hudnell el al., 1998).
The Current Study
Visual system function was assessed in the current study to test the hypothesis that VCS is lower
in populations exposed to P. piscicida neurotoxins that in unexposed populations. The
performance of several groups in the current study on tests of visual acuity, VCS, and color
discrimination is reported. Visual acuity was measured to obtain an indicator of the functional
integrity of the eyes' physiological optics system. VCS was measured as an indicator of
neurological function in the visual system. The Ishihara color discrimination plate test was
included to quickly screen for malingering and motivation to perform at the level of one's ability,
as well as to detect congenital color blindness and color deficiencies because these conditions
could affect performance on tests using colored stimuli. However, the Ishihara test is insensitive
to acquired dyschromatopsia relative to tests designed for this purpose, such as the Lanthony 15
Desaturated Hue test (Geller and Hudnell, 1997). Results are presented for four separate group
analyses: 1) The full occupational estuary and offshore cohorts; 2) the occupational cohorts
restricted to participants without histories which could explain neurological deficits (Savitz,
1998); 3) the occupational cohorts restricted to estuary participants who reported direct contact
with an active fish kill and offshore participants who did not report direct contact; and 4) the
Pfiesteria Hot-Line case-control series participants.
METHODS
-------
August 12, 1998 HUDNELL 8
Study design, study participant selection, methods for assessing socio-demographic factors and
health history, and the criteria used to exclude participants with potentially confounding factors
were described in detail by Savitz (1998), but will be briefly reviewed here. The investigators
were unaware of the study participants' group status during testing, thereby creating a single-
blind clinical investigation.
Study Design & Subject Selection
Occupational Cohorts
The occupational estuary- and offshore-cohort study design called for the selection of people who
worked full'time on the waters in 1997. The potentially exposed cohort worked in the Neuse and
Pamlico river estuaries, whereas the offshore participants worked near their residences on the
Outer Banks between Ocrachoke Village and Hatteras Village. The estuary participants were
primarily licensed gill fishers and crabbers, although four were male state employees, some of
whom had worked on the estuaries for only a few months. Ultimately, 19 males and four females
were included in the estuary group. The offshore participants were licensed, commercial fishers,
who worked in boats thought to be comparable in size to those of the estuary fishers, plus four
male and four female county or state employees. The offshore participants were individually
matched with estuary participants for age, gender, education, and occupation with a few
exceptions, such as matching the four estuary fisherwomen with county or state employees due to
the lack of suitable control fisherwomen. The data from several participants were ultimately
excluded from analysis due to participants' failure to give good effort during testing or being
outside of the targeted age or education range (Savitz, 1998). The current analyses of the full
occupational cohorts are based on data from 22 estuary and 20 offshore participants who met all
qualification criteria (Savitz, 1998). One offshore, but no estuary, participant had diabetes, a
disease which might alter performance on vision or other neurobehavioral tests.
A sub-group from the estuary and offshore cohorts was also identified for analytical purposes
(Savitz, 1998). Restricted cohorts were constructed to eliminate participants who had mild
neurological anomalies that might be accounted for by factors other than exposure. These factors
included a "history of difficulty in school, serious past or current psychiatric symptoms, daily
marijuana use, past or current other drug abuse, and medical conditions with potential cognitive
effects", including diabetes (Savitz, 1998). The restricted cohorts were comprised of 14 estuary
and 10 offshore participants from the full group.
Analyses were also performed on a second sub-group of the full occupational group to exclude
estuary participants who did not report direct contact with an active fish kill and offshore
participants who did report direct contact with an active fish kill. This exclusion criterion resulted
in 17 participants in both the estuary and offshore cohorts.
Case-Control Series
-------
August 12, 1998 HUDNELL 9
Cases were recruited for the study from the approximately 100 callers to the NC toll-free
Pfiesteria Hot Line established after the State upgraded a Fish Kill Precaution to a Health
Advisory in September, 1997 (Savitz, 1998). State telephone operators administered a 6 page
questionnaire on health symptoms and water contact to all callers. NC Division of Epidemiology
staff subsequently administered a more detailed follow-up questionnaire to 65 of the callers. Ten-
point scales were used to quantify both exposure and symptom questionnaire data such that a
maximum total score of 20 points could be attained. The highest scoring individuals were
recruited for the study in descending score order until 11 agreed to participate. Case scores
ranged from 15-20 points. A pool of 11 control subjects individually matched to cases for age,
gender, education, and occupation was also recruited. Two of the controls were identified by
cases as friends willing to participate, whereas the remainder of the controls were county health
department"employees from Eastern North Carolina and the Piedmont (Savitz, 1998).
Vision Tests
All subjects who normally wore corrective lenses for near-point viewing were asked to wear them
during vision testing. The visual acuity and VCS tests were administered monocularly to each eye;
an eye occluder was held over one eye while the other eye was tested. The Ishihara congenital
color-blindness screen was administered binocularly. All vision tests were administered under
illumination from a "daylight" illuminator (fluorescent source with a correlated color temperature
of approximately = 6500° K; color rendering index > 90; intensity = 1150 lux; luminance
approximately 70 foot-lamberts) in a clinical unit at East Carolina University Medical School
which had normal background lighting. A light meter was used to insure that luminance remained
constant throughout the test sessions. A face rest, placed just under the cheek bones and
connected by a calibrated rod to a card holder on the distal end, was used to position the acuity
and VCS test cards at a constant distance from the eyes (acuity -36 cm; contrast sensitivity - 46
cm).
Near Visual Acuity. The acuity test card (Rosenbaum Pocket Vision Screener; Grass Instrument
Co., Quincy, MA) contained 10 rows of numbers in which the size of the numbers progressed
from a larger angular subtense in the top row to a smaller angular subtense in the bottom row.
Participants were asked to first read the numbers in a middle row. Testing proceeded to the next
lower row if all numbers were correctly identified or to the next higher row if an error occurred.
The Snellen distance equivalent of the row with the smallest numbers which were all correctly
identified was recorded as the visual acuity score. Approximately 2 minutes were required to test
both eyes and explain the test results to the participants.
Contrast Sensitivity. The contrast sensitivity test card (Functional Acuity Contrast Test,
F.A.C.T. 101; Stereo Optical Co., Chicago, IL) contained a matrix (5 x 9) of circles filled with
sinusoidal gratings (dark and light bars). Spatial frequency (1.5,3,6,12 and 18 cycles/deg)
increased from top to bottom, and contrast decreased from left to right in steps of approximately
0.15 log units. The grating bars were oriented either vertically, or tilted 15 degrees to the left or
right. As the investigator called out each circle from left to right, row by row, subjects responded
-------
August 12, 1998 HUDNELL 10
by saying either vertical, left, right or blank. Participants were encouraged to name an orientation
if they had any indication that the bars could be seen. Participants were also asked to point in the
direction to which the top of the grating was tilted if they felt any difficulty in verbalizing the
orientation. The contrast of the last test patch correctly identified on each row was recorded as
the contrast sensitivity score for that row (spatial frequency). The procedure was repeated for
each row in descending order. Scores were recorded on a graph showing the normative range
(90th percentile confidence interval). Approximately 6 minutes were required to test each eye
separately and explain the results to the participants.
Color Discrimination. The short version (6 plates) of the Ishihara test (Ishihara's Tests for
Colour-Blindness, 38 plates edition, 1993, K«nehara & Co., Ltd. Tokyo, Japan) was used to
screen individuals for congenital dyschromatopsia. Four plates contained multicolored patches in
which a number could be seen by participants with congenital trichromatic vision. One plate
contained a number that could not be seen by participants with normal, trichromatic color vision;
only participants with a red-green deficiency, either protanomaly or deuteranomaly, could identify
the number. Another plate contained a number which could be seen by dyschromatopics since it
was defined by luminance, rather than color, gradients. This plate served as a control for low
motivation or malingering. All plates were individually displayed for binocular viewing and
participants were informed of the results in about 1 minute.
Vision Test Exclusion Criteria & Statistical Analyses.
The units of analysis for the visual-acuity and contrast-sensitivity tests were the mean scores of
the participant's two eyes for each endpoint, with one exception. The data from an eye was
excluded from analysis if the visual acuity score was poorer than the Snellen Distance Equivalent
of 20:70 in order to avoid confounding of the VCS results by excessive optical-refraction error. In
cases where a participant had only one qualifying eye, the score from that eye was the unit of
analysis. This visual acuity criterion for inclusion in data analysis did not result in the loss of any
data among the occupational cohorts. However, data from four eyes in two Pfiesteria Hot-Line
cases and four eyes in three Hot-Line controls were excluded due to the criterion. All participants
in the occupational and Pfiesteria Hot-Line groups identified the Ishihara plate number defined by
luminance, rather than color, gradients. Therefore, no additional exclusions were required due to
low motivation to perform the tests or malingering. As mentioned above, several participants
were previously removed from the population due to failure to meet the criteria described by
Savitz (1998). Based on these results, the sample sizes remained at 22 estuary and 20 offshore
participants in the unrestricted occupational group, 14 estuary and 10 controls in the group
restricted for confounding factors, and 17 in both cohorts adjusted for fish kill exposure. The
Pfiesteria Hot-Line group was reduced to nine cases and 10 controls.
Visual Acuity
Two-tailed Student t-tests with an a = 0.05 were performed in all analyses of visual acuity, using
each participant's mean visual acuity score, to determine if scores differed significantly between
-------
August 12, 1998 HUDNELL 11
cohorts.
Visual Contrast Sensitivity
The VCS data were analyzed using multivariate analyses of variance (MANOVA, with the Wilks'
lambda statistic) procedures suitable for repeated measures with an a = 0.05. The factors in the
model were group, spatial frequency, and their interaction term. A factor for eye was not required
since the analysis units were the mean scores for the two eyes at each spatial frequency, with the
exception noted above concerning data from excluded eyes. A factor for gender was not included
since no gender differences in susceptibility to P. piscicida-induced effects had been indicated and
the groups were matched for gender. Results which showed a significant group-by-spatial
frequency interaction were further analyzed in step-down, two-tailed Student t- tests (a = 0.05),
the equivalent of a univariate ANOVA, to determine which spatial frequencies accounted for the
overall effect.
Additional analyses used multivariate linear-regression techniques to assess relationships between
VCS, group assignment, and the covariates of age, education, smoking, alcohol consumption, and
total time spent on any water. The model initially included each of these factors. A backward
elimination technique was use to remove covariates and interaction terms which did not appear to
explain any of the variance in VCS. Interaction terms and then covariates which had a p-value
>0.015 were eliminated one at a time from the model if their removal did not alter the ratio of the
group estimate and the standard error of the estimate. This approach allowed an assessment of the
ability of group assignment to predict VCS while taking into account even small between-group
differences in the covariates which might influence VCS. In addition, chi-square tests were used
to assess the significance of group differences in several categorical variables.
Similar multivariate linear-regression analysis techniques were used in exploratory analyses
(Muller et al., 1984) to evaluate the potential for several measures to serve as surrogates of
exposure in dose-response assessments.
Color Blindness
Statistical anlayses were not performed on data from Ishihara's Tests for Colour-Blindness.
Rather, each participant's data were examined using standard methods (Ishihara, 1993) to
determine the presence or absence of congenital color blindness and color-discrimination
deficiency.
RESULTS
Occupational Cohorts: Analyses of Group Differences
According to Savitz (1998), the estuary and offshore occupational cohorts were well matched for
-------
August 12, 1998 HUDNELL 12
age (mean estuary = 41.4; offshore = 42.9 years), education (mean estuary = 13.2; offshore =
13.8 years), and gender (mean estuary male = 78.3%, female = 21.7%; mean offshore male
81.8%, female = 18.2%). The matching of the estuary and offshore cohorts for occupation, based
on job title, only attained a level of 52.2%. Whereas commercial fishers or crabbers comprised
73.9% of the estuary cohort, only 40.9% of the offshore participants shared these job titles. The
lack of better matching for occupation was due in part to the inability to successfully recruit
offshore female fishers or crabbers, which led to the alternative strategy of recruiting four female
government employees for the offshore cohort.
The estuary and offshore participants in the full occupational group were not significantly
different in visual acuity (Table 1). However, as shown in Figure 1, mean VCS was lower in the
estuary than'in the offshore participants at all*five spatial frequencies. Statistical analyses indicated
that the group factor and the group-by-spatial frequency interaction term were significantly
different (Table 1). Significant VCS reductions in the estuary participants at the middle and next
highest spatial frequencies, 6 and 12 cycles/degree (Table 1), were primarily responsible for the
overall difference.
Since about half of the participants in the two occupational cohorts were identified as having
historical factors which could influence neurobehavioral test outcomes (Savitz, 1998), the
analyses of the vision data were repeated on the cohorts restricted to only those participants for
whom potentially confounding medical, life-style, or educational factors were not identified. As in
the full occupational group, the estuary and offshore restricted cohorts showed no statistically
significant difference in, visual acuity (Table 2). Yet, group differences in mean VCS were greater
at each spatial frequency than in the full group due to slightly improved scores in the offshore
participants (Figure 2). Both the group factor and the group-by-spatial frequency interaction term
were significant (Table 2). Further analyses indicated that the significant difference between
cohorts at the middle spatial frequency, 6 cycles per degree, primarily accounted for the overall
difference (Table 2).
Due to the consideration that age, education, smoking, alcohol consumption, or total time spent
on any body of water (an indicator of bright sunlight exposure, Rosenthal et al., 1991) might
affect VCS and, therefore, that even small differences between the cohorts in these factors might
account for the lower VCS scores in the estuary cohort, multivariate linear regression analyses
were performed on the data sets from both the full and restricted occupational cohorts. The
analyses sought to determine the ability of these variables plus group membership (estuary versus
offshore) to predict VCS. Only VCS scores at the mid-spatial frequency, (6 cycles per degree,
VCS-6), were used in the analyses because this variable showed the largest difference between the
cohorts. The results for the full occupational group, shown in Table 3, indicated that group
membership was the most significant predictor of VCS-6, although smoking and total time spent
on any water appeared to account for some of the variance in VCS-6. In addition, there was a
significant interaction between group and age. Muhivariate linear regression analyses conducted
on each cohort separately revealed a trend of decreasing VCS-6 with age in the offshore cohort,
but not in the estuary cohort (Table 3). Similar results were seen in the analyses of the data from
-------
August 12, 1998 HUDNELL 13
the restricted occupational cohort (Table 3).
Since the VCS difference between cohorts was primarily due to reduced sensitivity at the middle
spatial frequencies, as seen in solvent-exposed groups (Mergler et al., 1991; Bowler et al., 1991;
Frenette et al., 1991; Broadwell et al., 1995; Hudnell et al., 1996a), the exposure-questionnaire
data were examined to determine if the cohorts differed in occupational exposures to solvents or
other potential neurotoxicants. The questionnaire design did not allow quantification of non-
occupational exposures to potential neurotoxicants. As shown in Table 4, solvent exposures were
commonly reported by members of both cohorts, but both the frequency of reports and the total
number of years in which these exposures occurred were slightly greater in the offshore than
estuary cohort. In addition, few participants in either cohort reported occupational exposures to
mercury, lead, or pesticides. Although more participants reported exposures to other metals and
fumes, these exposures were more frequent among offshore than estuary participants.
The exposure questionnaire data indicated that five members of the estuary cohort may not have
had direct contact with an active fish kill, although they reported contact with some dead fish, and
that three members of the offshore cohort may have been at the site of an active fish kill.
Therefore, these participants were excluded from the full group in an analysis of VCS differences
between the estuary (N=17) and offshore (N=17) cohorts. Mean VCS at each spatial frequency
was higher in the offshore group than in the estuary group. The significance tests indicated that,
although the group factor no longer showed a significant difference (F(l,32)=2.95, p=0.096), the
difference in the group-by-spatial frequency interaction term remained significant (F(4,29)=3.32,
p=0.023). Analyses of VCS at each spatial frequency again indicated that the overall difference
between cohorts was largely due to a significance difference at the middle spatial frequency
(t=2.34, p=0.026). The difference at other spatial frequencies was not significant.
The results of the color vision screening indicated that the estuary cohort included 1 (4.5%)
congenitally colorblind, 11 normal (50%), and 10 (45.5%) color-deficient participants. The
offshore cohort included no colorblind, 10 normal (50%), and 10 (50%) color-deficient
participants.
Occupational Cohorts: Exploratory Analyses of Potential Pfiesteria-Exposure Indicators
Exploratory analyses (Muller et al., 1984) of dose-response relationships were conducted using
three variables constructed from the questionnaire data to assess the potential for the variables to
serve as quantitative, although surrogate, indicators of exposure to P. piscicida. The analyses
used multivariate linear regression techniques with the dependent variable, VCS-6, and the
covariates of age, education, smoking, alcohol consumption, total time spent on any body of
water, and each of the potential exposure indicators in separate analyses. The first potential
exposure indicator was total time spent in estuary waters (TEW). The data from all members of
the estuary cohort in the full occupational group and the three members of the offshore cohort
who reported spending some time in estuary waters were included in the analysis. The regression
-------
August 12, 1998 HUDNELL 14
coefficient for VCS-6 and TEW, while in the direction of decreasing VCS-6 with increasing
TEW, was not significant (p=0.20). A similar analysis, which omitted the covariate for total time
spent on any body of water, and corresponding analyses on data from the restricted occupational
group also failed to show a significant relationship between VCS-6 and TEW.
The other two potential exposure indicators were participant-reported number of contacts with
fish kills (NFK) and the total number of hours spent at fish kills (HFK). Among the estuary cohort
members, HFK showed a significant interaction with years of education in the multivariate linear
regression with VCS-6 (estimate=-0.28, standard error=0.09, p=0.009), whereas NFK showed a
trend in the same direction (estimate=-6.2, standard error=2.9, p=0.051). Of the 22 members in
the estuary cohort, 10 participants had 12 or fewer years of education and 12 participants had 13
or more ye'ars of education. To address the VCS and education interaction, simple regression
analyses of VCS-6 with HFK were performed separately on the two education groups. HFK and
VCS-6 were significantly related in the more highly educated group (estimate=-0.28, standard
error=0.12, p=0.046) but not in the less educated group (estimate=0.10, standard error=0.19,
p=0.618). Participants in the more highly educated group were more likely to have contacted a
fish kill than those in the less educated group (correlation chi square, p=0.011), and six of the
seven participants who reported contact with 3 or more fish kills were in the more highly
educated group.
Case-Control Series
The mean age of the Pfiesteria Hot-Line cases was 43.7 years and they averaged 12.5 years of
education. Males and females comprised 72.7% and 27.3% of the group, respectively (Savitz,
1998). Although initial screening indicated that all cases had direct contact with fish kills, only six
of the 10 cases confirmed contact during the clinical examination (Savitz, 1998). Demographic
characteristics of the control group were not described (Savitz, 1998).
Visual acuity was slightly better in the Pfiesteria Hot-Line control participants than in the cases,
although the difference was not statistically significant (Table 4). However, the control group
scored lower in VCS than the cases at each spatial frequency (Figure 3), although neither the
group factor nor the interaction term showed a statistically significant difference between the
cases and controls (Table 4).
The results of the color vision screening indicated that the cases included no colorblind, 2 normal
(22.2%), and 7 (77.8%) color-deficient participants. The controls included ! (10%) colorblind, 7
normal (70%), and 2 (20%) color-deficient participants.
DISCUSSION
4___ rf-i_ _,*»„ .i-.™ A 1
i»i VxUiini UiaiAM y '
-------
August 12, 1998 HUDNELL 15
Several analyses indicated that the estuary cohort had a mid-spatial frequency reduction in VCS
relative to the offshore cohort. First, in the full occupational cohort, the significant difference in
the MANOVA group factor indicated that VCS averaged across spatial frequencies was lower in
the estuary cohort than in the demographically-matched offshore cohort. The significance of the
group-by-spatial frequency interaction term indicated that the two cohorts' VCS profiles across
spatial frequencies were not parallel. It can be seen in Figure 1 that peak sensitivity in the offshore
group was at 6 cycles per degree of visual arc, as seen in other unexposed populations (Hudnell et
al., 1996b,c). In the estuary group, however, no clear peak in sensitivity was apparent; sensitivity
was approximately equal at 3 and 6 cycles per degree. Analyses of group differences at each
spatial frequency indicated that VCS was significantly reduced in the estuary cohort at 6 and 12
cycles per degree relative to the offshore cohort.
Second, an assessment of medical history, life-style factors, and educational history indicated that
these factors might explain deficits in neurological function in about half of the participants in
both cohorts (Savitz, 1998). Therefore, group differences in VCS were reassessed after restricting
the cohorts to participants free of these potentially confounding factors. As can be seen by
comparing Figures 1 and 2, the restriction had little effect on the VCS profile of the estuary
cohort, but led to some improvement in the VCS profile of the offshore cohort. The difference in
VCS between cohorts and the group-by-spatial frequency interaction term were significantly
different in the restricted group. Analyses at each spatial frequency indicated that the difference in
VCS at 6 cycles per degree was significant, whereas that at 12 cycles per degree only approached
significance.
Third, multivariate linear-regression analyses with backward elimination of interaction terms and
variables unrelated to VCS were performed to assess the potential for differences between the
cohorts in age, education, smoking, alcohol consumption, and total time spent on any water to
account for the group difference in VCS at the mid-spatial frequency (VCS-6). Age (Green and
Madden, 1987), smoking (West et al., 1989, 1995), alcohol consumption (Roquelaure et al.,
1995), and sunlight exposure (Taylor, 1995; Javitt and Taylor, 1995; Schein et al., 1994; Taylor
et al., 1993; Werner et al., 1990; Taylor et al., 1990; Taylor et al., 1989) have been reported to
affect either the optical properties or neurological function of the visual system. Both smoking
and total time spent on any water appeared to account for some of the variance in VCS-6, and a
relationship between VCS and age was observed in the offshore, but not the estuary, cohort.
Education and alcohol consumption did not appear to influence VCS-6 in the full or restricted
occupational cohorts. However, a strong difference in VCS-6 between the estuary and offshore
groups remained after taking into account the five potential covariates.
Fourth, an analysis of questionnaire data on occupational exposures to potential neurotoxicants
did not indicate greater neurotoxicant exposures in the estuary cohort, suggesting that such
exposures were not likely the cause of VCS differences between the estuary and offshore cohorts.
Fifth, another restricted group was created from the full occupational group because the exposure
questionnaire data indicated that 5 participants in the estuary cohort may not have had direct
contact with an active fish kill, whereas 3 members of the offshore cohort may have had contact.
-------
August 12, 1998 HUDNELL 16
Analyses indicated that, although the group factor for VCS only approached significance, the
interaction term and the group difference at 6 cycles per degree remained significant. However,
this analysis failed to show a larger difference than that seen in the full group as might have been
expected if direct contact with P. piscicida-induced fish kills was the causative agent, although
this qualification is somewhat mitigated by the fact that fish kills contacted by the offshore
participants may have been caused by factors other than P. piscicida..The crude assessment of
color discrimination ability provided by the Ishihara test did not indicate differences in
dyschromatopsia between the full occupational cohorts. However, since the VCS profile
resembled that seen in studies of solvent exposure, and solvent exposures also cause
dyschromatopsia (reviewed in Geller and Hudnell, 1997), future studies of P. piscicida effects in
humans should use color discrimination tests more sensitive to acquired dyschromatcpsia such as
the Lanthohy 15 Desaturated Hue test (Geller and Hudnell, 1997).
Conclusions and Possible Implications. Taken together, analyses of the VCS data confirm the
hypothesis that some factor(s), other than several demographic variables, medical history, life-
style factors, or known neurotoxicant exposures, is associated with lower visual system function
in the estuary cohort than in the offshore cohort. However, the nature of the association remains
unclear. As in all studies involving group comparisons, it is possible that the difference observed
between the cohorts resulted from chance. The selection of cohort members from the entire
populations of offshore and estuary fishers and crabbers could have resulted in the inclusion of
participants with lower VCS in the estuary cohort and participants with higher VCS in the
offshore cohort simply by chance, even if there is no difference in VCS between the entire
populations. However, if the entire populations do differ in VCS, there are at least three primary
hypotheses that may explain the difference. First, since all estuary participants lived on or near the
mainland and all offshore participants lived on the outer banks, the group difference in VCS could
be due to an unknown factor(s) (e.g. neurotoxicant exposure, population genetics) which differs
between these geographical areas. Second, a factor(s) associated with estuary waters other than
P. piscicida (e.g. other neurotoxin or neurotoxicant exposure) could have caused the group
difference in VCS. Third, exposure to P. piscicida neurotoxins in the estuaries could have caused
the deficit in visual function. If the VCS deficit is attributable to contact with P. piscicida-induced
fish kills, the lack of recent exposure (Savitz, 1998) would suggest that the effect is permanent or
long lasting, as previously seen with solvent-induced VCS deficits (Mergler et al., 1991; Bowler
et al., 1991; Frenette et al., 1991; Broadwell et al., 1995; Hudnell et al., 1996). Alternatively, the
VCS deficit could be more readily reversible, but caused by unknown, non-exposure related
factor(s) or by some other, perhaps more continuous, exposure factor(s) associated with the
estuaries or the geographical area. Given the multiple possible explanations and the study
limitations discussed below, each of these hypotheses should be viewed as tentative.
The average contrast of the mid-spatial frequency VCS test stimulus required for the participants
to detect the stimulus was about 30% higher in the estuary cohort than in the offshore cohort. The
functional impact of this difference in perceptual threshold, the inverse of contrast sensitivity, is
difficult to estimate, although a number of studies have assessed the relationship between VCS
and various functional abilities. It is well recognized that the speed of performance on behavioral
-------
August 12, 1998 HUDNELL 17
tasks, as indicated by reaction times, is highly dependent on VCS (Felipe et al., 1993; Hudnell et
al., 1996c). Since delays in the detection of visual stimuli due to VCS deficits cause a delay in the
production.of appropriate responses, VCS scores have been proposed for use as a predictor of
automobile-accident risk in elderly drivers (Shinar and Schieber, 1991). In the elderly, VCS is
strongly associated with postural stability (Lord et al., 1991) and the frequency of falling (Lord et
al., 1994). VCS scores helped to correctly categorize individuals into multiple faller and non-
multiple faller categories with 75% accuracy (Lord et al., 1994). Furthermore, other studies have
associated VCS with reading performance (Carmean and Regeth, 1990) and athletic ability (Kulka
et al., 1989, 1993, 1996; Love and Kulka, 1992; Melcher and Lund, 1992). This evidence
suggests that VCS deficits may result in decreased productivity and a lessening of the quality of
life due to slowed performance and an increased risk for accidents. Although VCS deficits may be
predictive of cognitive impairment in neurodegenerative disease (Cronin-Golomb et al., 1995), no
evidence suggests that the estuary cohort participants are at increased risk for any disease.
Occupational Cohorts: Exploratory Analyses
The exploratory analyses, classified as such to reduce the probability of obtaining false-positive
results due to multiple comparisons (Muller et al., 1984), attempted to identify surrogate
measures of exposures in the estuary waters. No relationship between VCS-6 and total time spent
on estuary waters was apparent. This result suggested that, if VCS was reduced in the estuary
cohort due to exposures encountered while on the estuaries, the exposure was not likely to have
been continuously ongoing, but rather was likely to have been periodic as would be expected if
the causative exposure was P. piscicida neurotoxin(s) released during fish kills. A significant
interaction between total hours spent at fish kills (HFK) and educations was observed in a
multivariate linear regression with VCS-6 in the estuary cohort. Simple linear regressions of VCS-
6 with HFK were performed on roughly even numbers of participants who did and did not have at
least some college education. A significant correlation was observed in the more highly educated
group but not in the less educated group. However, the more highly educated group had
significantly more contacts with fish kills than the less educated group. A stronger association of
VCS with HFK would be expected among more highly exposed than lesser exposed participants if
an above-threshold level of exposure was required to produce the VCS deficit. This result
suggested that future studies may be able to associate health effects with exposure by obtaining
detailed information from study participants on contact with fish kills.
Related Research
The results from the assessment of visual system function were consistent with the outcomes of
the neurological examination (Savitz, 1998) in which a trend towards a difference between the
occupational cohorts in sensory abnormalities was observed. Although the difference was not
significant, overall sensory abnormalities were noted in 41% of the full cohort of estuary
participants versus 20% of the offshore cohort members (risk ratio=2.0, 95% CI=0.8-5.6).
Sensory abnormality ratings in the group restricted for potentially explanatory histories were 43%
in the estuary cohort and 18% in the offshore cohort (risk ratio=2.4, 95% CI=0.3-19.6), but this
-------
August 12, 1998 HUDNELL 18
difference was not significant. The only other category in the neurological examination showing
notable, but nonsignificant, differences between groups was overall peripheral neuropathy. The
frequencies of peripheral neuropathy in the full group were 35% and 18% in the estuary and
offshore cohorts, respectively. Comparable figures in the group restricted for potentially
explanatory histories were 36% in the estuary cohort and 9% in the offshore cohort. Little or no
difference between cohorts in either the full or restricted group was observed for cognitive or
motoric function, behavioral disturbances, neurotoxic complex staging, or affective status (Savitz,
1998).
Scientists exposed to P. piscicida toxins in the laboratory have reported ocular irritation, as well
as blurred vision which persisted for hours to days (Glasgow et al., 1995), although the
neurologicaTsigns and symptoms given greater prominence included narcosis, short-term memory
loss, spatial disorientation, peripheral sensory disturbance, and emotional lability (Glasgow et al.,
1995). Clinical evaluations of humans recently exposured to a Pfiesteria-induced fish kill were
conducted in Maryland during the summer of 1997 (Morris et al., 1997; Bever et al., 1998; Golub
et al., 1998; Grattan et al., 1998; Greenberg et al., 1998; Lowitt et al., 1998; Tracy et al., 1998).
VCS and other neurobehavioral tests of visual system function were not administered in the
Maryland study. Neurological assessment in the Maryland study focused primarily on
neurocognitive functions. Cognitive function was also the focus of the only study reported to date
which investigated the effects of P. piscicida exposure in an animal model (Levin et al., 1997).
Water containing P. piscicida was collected from an aquarium in which a fish kill had been
induced, and single subcutaneous injections were given to rats. The exposed rats and control rats,
which received P. piscicida-free water injections, were repeatedly tested with a radial-arm maze
learning task. Choice accuracy was significantly reduced in the exposed rats relative to controls
for up to 10 weeks. Subsequent tests showed no performance differences between groups. Rats
trained on the task prior to exposure showed no decrement in performance relative to control rats,
suggesting that P. piscicida exposure induced a learning, rather than a memory, deficit.
However, no assessment of visual function in the rats was performed. Since performance in radial-
arm maze tasks is heavily dependent on visual cues (e.g. Zoldek and Roberts, 1978), the learning
delay seen in the exposed rats may have been due to a reduction in visual system function.
Humans or animals with low vision may require a longer period of time to learn a visually oriented
task than those with normal vision, but then perform at a level comparable to those with normal
vision after finally mastering the task.
Case-Control Series
Analyses of data from the Pfiesteria Hot Line case-control series indicated that these groups were
not significantly different in VCS. A notable feature of the data was that the VCS profile of the
control participants (Figure 3), who scored lower at each spatial frequency than the cases, was
well below that of the offshore cohorts in the occupational group (Figures 1, 2). The incidence of
color-vision deficiency, on the other hand, was low in the controls relative to the cases and the
onshore cohorts in the occupaiiuiial giuup. Huwcvcf, due to the relative inscnsitivity of the
Ishihara test to acquired dyschromatopsia and its inability to distinguish congenital from acquired
-------
August 12, 1998 HUDNELL 19
color deficiency (Geller and Hudnell, 1997), the color discrimination data provided only weak
support for the position that the cases and controls differ in acquired dyschromatopsia. Overall,
the VCS and color discrimination data do not indicate that visual function in the Pfiesteria Hot-
Line cases was affected by exposure to NC estuaries.
Study Strengths and Weaknesses
As noted by Savitz (1998), strengths of the current study included: 1) the use of objective criteria
in the subject-selection process; 2) matching of the estuary and offshore cohort participants for
age, gender, and education, with less complete matching for occupation, and; 3) masking of the
investigators with respect to the exposure status of the participants. However, a number of factors
preclude drawing definitive conclusions concerning the cause of the VCS deficits seen in the
estuary cohort. Study limitations noted by Savitz (1998) included: 1) the possibility that the
prevalence of many factors with potential to affect neurological function, and thereby confound
interpretation of the study results, may not have been well balanced between cohorts; 2)
uncertainty about the participants' exposure to P. piscicida toxin(s), and; 3) a lack of any
participants with very recent fish kill contact. Uncertainties concerning potentially confounding
factors are a major source of concern in drawing inferences from the study results. For example,
data obtained on chronic alcohol consumption, which can reduce VCS (Roquelaure et al., 1995),
the use of neuroactive prescription drugs, and avocational or environmental exposures to
neurotoxicants may have been inadequate to insure that these factors did not account for the VCS
differences between cohorts. In addition, no data were collected on other potentially confounding
factors such as dietary habits and the use of sunglasses or brimmed hats. The Cuban epidemic
optic neuropathy episode demonstrated that deficiencies in micronutrients without overt
malnutrition can induce reversible VCS deficits (Sadun et al., 1994, Roman, 1994). Chronic
exposure to intense sunlight, particularly the ultra violet portion of the spectrum, can induce
cataracts and other conditions involving optical aberrations (Taylor, 1995; Javitt and Taylor,
1995; Schein et al., 1994; Taylor et al., 1993; Taylor et al., 1989), which reduce VCS (Lasa et al.,
1992; Drews-Bankiewicz et al., 1992; Burton et al., 1993). Sunlight exposure also has been
proposed as an important determinant of retinal aging (Werner et al., 1990) and age-related
macular degeneration (Taylor et al., 1990), and VCS declines with aging beyond about 40-50
years (Green and Madden, 1987). Less frequent use of brimmed hats and sunglasses by the
estuary workers than the off-shore workers could have induced optical opacities or accelerated
retinal aging which might account for the VCS difference between cohorts. Analyses of the VCS
data only included eyes with a visual acuity of 20:70 or better. However, detection of the
extremely high-contrast stimuli used in the visual acuity test may be less affected by the optical
and neural effects of chronic sunlight exposure than is detection of the low-contrast VCS stimuli.
Future studies could directly address the issue of ocular aberrations by performing slit-lamp
examinations on the participants (Taylor, 1995). It is possible that differences between cohorts in
these or other factors could have caused the difference in VCS.
Uncertainties in human-health risk assessments decrease as more information becomes available to
quantitatively describe relationships between the components of a biologically-based dose-
-------
August 12, 1998 HUDNELL 20
response model (BBDR; e.g. Conolly and Andersen, 1991 Andersen et al., 1992), illustrated in
Figure 4. In order to implicate (or exonerate) P. piscicida as the causative agent of the VCS
difference between cohorts, more definitive data are needed on exposure - the dose of
neurotoxin(s) applied to humans, the absorbed dose, and dose at the target site. In the current
study, time spent on the water and the degree of potential P. piscicida exposure in the estuary
cohort varied greatly. Some government employees in the estuary population had been employed
for only five months at the time of the study and had very limited activity on the water, whereas
some fishers and crabbers had more than 20 years of potential exposure. As a first approximation
of the applied dose, the current study tried to identify the date, location, and time spent in direct
contact (within 6 feet) with active fish kills. Yet uncertainties in the participants' abilities to recall
and report this information, and the lack of evidence that P. piscicida v»us the cause of all
reported fish' kills, permitted only crude quantifications of surrogates for applied dose. These
factors could be better assessed in a prospective study by collecting exposure data during an
active fish kill. Identification of the P. piscicida toxin(s) would enable better quantification of
applied and absorbed doses if methods were developed to measure the toxin(s) in water, air (or
mist above the water), and in biological samples. More complete information on exposure dosage
would permit dose-response assessments and better justify causative inferences between P.
piscicida and effects such as reduced VCS.
Conclusion and Research Needs
The current study sought to provide initial evidence on the human health effects of exposure to P.
piscicida-mduced fish kills, and to identify experimental design factors which might improve
future studies. The current study design allowed associations to be made between VCS reduction
and inland residence with time spent on NC estuaries, relative to offshore residence with time
spent on off-shore waters. Since the time spent working on estuary waters sometimes involved
contact with fish kills thought to be induced by P. piscicida toxin(s), one of the possible causative
agents is P. piscicida. However, due to uncertainties about exposure to P. piscicida toxin(s) and
the confounding of residency area, estuary contact, and fish-kill contact, future research must
verify the tentative association between exposure to P. piscicida neurotoxin(s) and VCS
reduction. Verification of this relationship would indicate that the VCS test is a useful screening
tool for detecting an early sign of P. piscicida exposure-induced effects on neurological function.
Two approaches might be taken in future studies to more strongly implicate (or exonerate) P.
piscicida as the causative agent. First, a strong association between P. piscicida and health status
could be attained using a prospective approach with health measurements taken before and after
exposure to fish kills where environmental monitoring demonstrated the presence of P. piscicida
and, if possible, P. piscicida toxin(s). This approach would also provide an opportunity to detect
any rapidly-reversible effects, an element missing from the current study. However, if the
participants in a prospective study were previously exposed to P. piscicida toxin(s), persistent
exposure-induced effects present during the initial evaluation could lessen the probability that
effects would be detected after a subsequent exposure. Second, a comparison of estuary workers
who were and who were not at a recent fish kill where P. piscicida activity was documented
-------
August 12, 1998 HUDNELL 21
could eliminate any confounding differences between cohorts associated with residence area and
estuary versus off-shore work. However, this approach is also subject to the limitation imposed by
pre-existing, exposure-induced effects. The use of young, previously unexposed workers to
obviate this problem might be inadequate because young workers may be less susceptible to P.
piscicida toxin(s) than older workers due to greater neurological reserve or compensatory
capacities. Susceptibility may increase through an interaction with the functional declines
associated with ageing such that dysfunction is more likely to manifest in older individuals, as
observed with the Parkinson-like motoric dysfunction associated with environmental manganese
exposure (Mergler et al., 1998; Hudnell, 1998).
A more optimal approach might be formed by combining the best aspect of the current study's
design, the'single-blind group comparison, with a prospective, repeated-measuros approach. Two
large cohorts, one with and one without exposure potential, could be established and maintained
long-term. Workers could be selected for the cohort with exposure potential who had not
previously contacted an active fish kill. An extensive assessment could be undertaken during the
selection process to screen potential participants for confounding factors and to individually
match the exposed and control cohort participants for socio-demographic characteristics. Initial
health assessments could evaluate the comparability of neurological function in matched-pairs of
participants. Following subsequent exposure to a confirmed and monitored P. piscicida-induced
fish kill, the matched-pairs could be reassessed to determine the health risk of exposure to P.
piscicida toxin(s). Limitations to this approach are cost and the possibility that a widespread
perception of risk associated with fish kUl contact, or a NC ban on water activities during a fish
kill, could result in none of the participants being exposed. In this case, a study could be designed
to assess the potential for risk in shore-side residents at locations with and without a high
probability for fish kills. In addition, further development of an animal BBDR model for studying
the neurotoxic effects of P. piscicida toxin(s) (Levin et al., 1997) could improve human-health
risk assessments for exposure to P. piscicida toxin(s) by investigating dose-response relationships
and the mechanism(s) of toxicity.
ACKNOWLEDGMENTS
This manuscript benefited greatly from information in the report on the study written by Dr.
David Savitz, UNC School of Public Health, who also supplied materials needed to do the
analyses reported in this manuscript. The assistance of Cristin Cahill-Harrell, US EPA, in
collection of these data, and Dennis House and Judy Schmid, US EPA, in performing the
statistical analyses, is gratefully acknowledged. Dr. Woddhall Stopford, Duke University Medical
School, supplied suggestions for data analyses and data for two variables on time spent in estuary
and other waters which enhanced the analyses. Reviews of earlier versions of the manuscript by
Drs. William Boyes, Hugh Tilson, Robert MacPhail, and Vernon Benignus, US EPA, Dr.
Woodhall Stopford, and anonymous reviewers led to substantial improvements in the manuscript.
The clinical evaluations of study participants were made possible by the organizational skills of
Dr. Marian Swinker, East Carolina University Medical School, and her staff. Finally, Dr. Stanley
Music, NC Department of Health and Human Services, is credited for the insight and leadership
which made the study a reality.
-------
HUDNELL 22
REFERENCES
Alexander KR, Derlacki DJ, and Fishmann GA. Visual acuity vs letter contrast sensitivity in
retinitis pigmentosa. Vision Res. 35(10): 1495-1499, 1995.
Alexander KR, Derlacki DJ, and Fishman GA. Contrast thresholds for letter identification in
retinitis pigmentosa. Invest. Ophthalmol Vis. Sci. 33(6): 1846-1852, 1992.
Altmann L, Neuhann HF, Kramer U, Witten J, Jermann E. Neurobehavioral and
neurophysiological outcome of chronic low-level tetrachloroethene exposure measured in
neighborhoods of dry cleaning shops. Environ. Res., 69:83-89, 1995.
Altmann L, Sveinsson K, Kramer U, WeishofT-Houben M, Turfeld M, Winneke G, and
Wiegadn H. Visual functions in 6-year old children in relation to lead and mercury
levels. Neurotoxicol. Teratol., 20(1):9-17, 1998.
Andersen ME, Krishnan K, Conolly RB, McClellan RO. Mechanistic toxicology research
and biological based modeling: partners for improving risk assessments. CUT Act., 12:1-
7, 1992.
Anger WK, Letz RE, Chrislip DW, Frumpkin H, Hudnell HK, Kilburn KH, Russo JM,
Chappell W, and Hutchinson L. Neurobehavioral test methods for immediate use in
environmental health studies of adults. Neurotoxicol. & Teratol., 16:489-497, 1994.
Atkin A, Wolkstein M, Bodis-Wollner I, Anders M, Kels B, and Podos S. Interocular
comparison of contrast sensitivities in glaucoma patients and suspects. Br. J. Ophthalmol.,
64:858-862, 1980.
ATSDR. Adult environmental neurobehavioral test battery. Amler RW, Anger WK, and
Sizemore OJ (eds). Atlanta, U.S. Department of Health and Human Services, 1995.
ATSDR. Pediatric environmental neurobehavioral test battery. Amler RW, and Gibertini M
(eds). Atlanta, U.S. Department of Health and Human Services, 1996.
Baker EL,. Letz R. and Fiedler AT. A computer-administered neurobehavioral evaluation
system for occupational and environmental epidemiology. J. Occupat. Med., 27:206-212,
1985.
Bangstad HJ, Brinchmann-Hansen O, Hultgren S, Dahl-Jorgensen K, Hanssen KF.
Impaired contrast sensitivity in adolescents and young type 1 (insulin-dependent) diabetic
pstisnts with microslbiiiTiiiiuria. Acts-Onhth£lrno!-Consnh., 72^6^668-673,1994.
-------
HUDNELL 23
Benignus VA, Boyes WK, Hudnell HK, Frey CM, and Svendsgaard DJ. Quantitative
Methods for Cross-species Mapping (CSM). Neurosci. Biobehav. Rev., 15:165-
171, 1991.
Bever, Jr CT, Grattan L, and Morris GJ. Neurologic symptoms following Pfiesteria exposure:
case report and literature review. Maryland Med. J., 47:120-123, 1998.
Bodis-WoIIner I. Methodological aspects of contrast sensitivity measurements in the
diagnosis of optic neuropathy and maculopathy. In Greve EL, Heijl A (eds): Fifth
international visual field symposium. The Hague, Dr. W Junk Publishers, pp225-
237, 1983.
Bodis-Wollner I and Camisa JM. 26. Contrast sensitivity measurement in clinical diagnois.
Neuro-ophthalmology, 1:373-401, 1980.
Bodis-Wollner I, and Diamond S. The measurement of spatial contrast sensitivity in cases of
blurred vision associated with cerebral lesions. Brain, 99:695-710, 1976.
Bodis-Wollner I, Ghilardi MF, and Mylin LH. The importance of stimulus selection in VHP
practice: The clinical relevance of visual physiology. In Cracco RQ, Bodis-Wollner I (eds):
Frontiers of Clinical Neuroscience, Vol 3: Evoked Potentials. New York, Alan R Liss,
Inc., ppl5-27, 1986.
Bowler RM, Mergler D, Huel G, Harrison R, and Cone J. Neuropsychological impairment
among former microelectronics workers. NeuroToxicology, 12:87-104, 1991.
Boyes WK. Testing visual system toxicity using evoked potential technology. In: The vulnerable
brain and environmental risks, Volume 1: Malnutrition and hazard assessment
Isaacson RL and Jensen KF (eds). New York, Plenum Press, 1992.
Boyes WK. Rat and human sensory evoked potentials and the predictability of human
neurotoxicity from rat data. Neurotoxicology, 15(3):569-578, 1994.
Broadwell DK, Darcey DJ, Hudnell HK, Otto DA, and Boyes WK. Work-site clinical and
neurobehavioral assessment of solvent-exposed microelectronics workers. Am. J. Indust.
Med., 27:677-698, 1995.
Burkholder JM, Noga EJ, Hobbs CH, and Glasgow, Jr HB. New 'phantom' dinoflagellate is
the causative agent of major estuarine fish kills. Nature, 358:407-410, 1992.
Burkholder JM, Glasgow, Jr HB, and Hobbs CW. Fish kills linked to a toxic ambush-predator
dinoflagellate: distribution and environmental conditions. Mar. Ecol. Prog. Ser., 124:443-
61, 1995.
-------
HUDNELL 24
Burton CB, Owsley C, and Sloane ME. Aging and Neural Spatial Contrast Sensitivity:
Photopic Vision. Vis. Res., 33:939-946, 1993.
Cameanm SL, Regeth RA. Optimun level of visual contrast optimium level of visual contrast
sensitivity for reading comprehension. Perceptual and Motor Skills, 71:755-762, 1990.
Camisa J, Mylin L, and Bodis-Wollner L The effect of stimulus orientation on the visual
evoked potential in multiple sclerosis. Ann. Neurol, 10:532-539, 1981.
Campagna D, Mergler D, Huel G, Belanger S, Truchon G, Ostiguy C, and Drolet D. Visual
dysfunction among styrene-exposed workers. Scand. J. Work environ Health, 21(5):382-
390, 1995.
-: •<*?.-
Cavalleri A, Belotti L, Gobba F, Luzzana G, Rosa P, and Seghizzi P. Colour vision loss in
workers exposed to elemental mercury vapour. Toxicol. Letts., 77:351-356, 1995.
Conolly RB, Andersen ME. Biologically based pharmacodynamic models: Tools for
lexicological research and risk assessment. Annu. Rev. Pharmacol. Toxicol., 31: 503-
523, 1991.
Courage ML, Adams RJ, and Hall EJ. Contrast sensitivity in infants and children with Down
syndrome. Vis. Res., 37:1545-1555, 1997.
Cronin-Golomb A, Corkin S, Rizzo JF, Cohen J, Growdon JH, Banks KS. Visual
dysfunction Alzheimer's disease: relation to normal aging. Annals of Neurology,
29(l):41-52, 1991.
Cronin-Golomb A, Corkin S, Growdon JH. Visual dysfunction predicts cognitive deficits in
Alzheimer's disease. Opto. Vis. Sci., 72(3): 168-176, 1995.
Donoghue AM, Dryson EW, and Wynn-Williams G. Contrast sensitivity in organic-solvent-
induced chronic toxic encephalopathy. J. Occupat. Environ. Med., 37:1357-1363, 1995.
Drews-Bankiewicz MA, Caruso RC, Datiles MB, and Kaiser-Kupfer ML Contrast
sensitivity in patients with nuclear cataracts. Arch. Ophthalmol., 110, 1992.
Echeverria D, White RF, Sampaio C. A behavioral evaluation of PCE exposure in patients and
dry cleaners: A possible relationship between clinical and preclinical effects. J. Occup.
Env. Med., 37(6): 667-680, 1995.
Felipe A, Buades MJ, Artigas JM. Influence of the contrast sensitivity function on the reaction
tiir.e. VisicnR.es. 33^17^:2461-2456. 1993.
-------
HUDNELL 25
Fleishman JA, Beck RW, Linares OA, and Klein JW. Deficits in visual function after
resolution of optic neuritis. Ophthal, 94(8):1029-111035, 1987.
Frenette B, Mergler D, Bowler R. Contrast-sensitivity loss in a group of former
microelectronics workers with noral visual acuity. Optometry and Vision Science,
68(7):556-560, 1991.
Gawande AA, Donovan WJ, Ginsburg AP, and Marmor MF. Photophobia in retinitis
pigmentosa. Br. J. Ophthalmol, 73(2): 115-120, 1989.
Geller AM, and Hudnell HK. Critical issues in the use and analysis of the Lanthony desaturated
color vision test. Neurotoxicol. Teratol., 19:455-465, 1997.
Gilmore GC and Levy JA. Spatial contrast sensitivity in Alzheimer's disease: A comparison
of two methods. Optometry and Vision Science, 68(10): 790-794, 1991.
Ginsberg AP. A new contrast sensitivity vision test chart. Am. J. Optom. Physiol. Opt., 61:403-
407, 1984.
Ginsberg AP, Evans D, Cannon M, and Mulvanny P. Large sample norms for contrast
sensitivity. Am. J. Optom. Physiol. Opt., 61:80-84, 1984.
Glasgow, Jr HB, Burkholder JM, Schmechel DE, Tester PA, and Rublee PA. Insidious
effects of a toxic estuarine dinoflagellate on fish survival and human health. J. Toxicol.
Environ. Hlth., 46:501-522, 1995.
Golub JE, Haselow DT, Hageman JC, Lopez AS, Oldach DW, Gratan LM, and Perl TM.
Pfiesteria in Maryland: preliminary epidemiological findings. Maryland Med. J., 47:137-
143, 1998.
Grattan LM, Oldach D, Tracy JK, and Greenberg DR. Neurobehavioral complaints of
symptomatic persons exposed to Pfiesteria piscicida or morphologically related
organisms. Maryland Med. J., 47:127-129, 1998.
Green HA and Madden DJ. Adult age differences in visual acuity, stereopsis, and contrast
sensitivity. American Journal of Optometry and Physiological Optics, 64(10): 749-753,
1987.
Greenberg DR, Tracy JK, and Grattan LM. A critical review of fa& Pfiesteria hysteria
hypothesis. Maryland Med. J., 47:133-136, 1998.
Greeves AL, Cole BL, and Jacobs RJ. Assessment of contrast sensitivity of patients with
macular disease using reduced contrast near visual acuity charts. Ophthalmol. Physiol.
-------
HUDNELL 26
Opt, 8:371-377, 1988.
Hudnell HK. Effects from environmental Mn exposure: A review of the evidence from non-
occupational exposure studies. Neurotoxicol., submitted, 1998.
Hudnell HK, Boyes WK, and Otto DA. Stationary pattern adaptation and the early components
in human visual evoked potentials. Electroenceph. Clin. Neurophysiol., 77:190-198,
1990a.
Hudnell HK, Boyes WK, and Otto DA. Rat and Human Visual Evoked Potentials Recorded
Under Comparable Conditions: A Preliminary Analysis to Address the Issue of Predicting
Human Neurotoxic Effects from Rat Data. Neurotoxicol.& Teratol., 12:391-398, 1990b.
Hudnell HK, and Boyes WK. The Comparability of Rat and Human Visual Evoked Potentials.
Neurosci. Biobehav. Rev., 15:159-164, 1991.
Hudnell HK, Boyes WK, Otto DA, House DE, Creason JP, Geller AM, Darcey DJ, and
Broad well DK. Battery of neurobehavioral tests recommended to ATSDR: solvent-
induced deficits in microelectronics workers. Toxicol. Indust. Hlth., 12:235-243. 1996a.
Hudnell HK, Skalik I, Otto D, House D, Subrt P, and Sram R. Visual contrast sensitivity
deficits in Bohemian children. NeuroToxicology, 17(3-3):615-628, 1996b.
Hudnell HK, Otto DA, and House DE. The influence of vision on computerized
neurobehavioral test scores: A proposal for improving test protocols. Neurotoxicol. &
Teratol., 18:391-400, 1996c.
Hudnell HK, Otto DA, House DE, Darcey DJ, Broadwell DK, and Boyes WK. Cognitive
deficits in solvent-exposed microelectronic workers: controlling for visual and motoric
influences on computerized test results. Neurotoxicol., submitted, 1998.
Ishihara S. The series of plates designed as a test for colour-blindness. Kanehara Shuppan Co.,
Ltd. Tokyo, Japan, 1993.
Javitt JC and Taylor Hit Cataract and latitude. Doc. Ophthal., 88:307-325, 1995.
KlukaDA, Love PA, Kuhlman J, HammackG, Wesson MD. The effect of a visual skills
training program on selected female intercollegiate volleyball athletes. International
Journal of Sports Vision, 3(1), 1996.
Kluka DA, and Love PA. The effects of daily-wear contact lenses on contrast sensitivity in
selected professional and collegiate female tennis players. Journal of American
-------
HUDNELL 27
Optomology Association, 64:182-186, 1993.
Kluka DA, Love PA, and Allen S. Contrast sensitivity functions of selected collegiate female
athletes. Sports Vision Journal, 5(1): 18-24, 1989.
Kupersmith MK, Siegel I, and Carr RE. Subtle disturbances of vision with compressive
lesions of the anterior visual pathway measured by contrast sensitivity. Ophthal.
(Rochester), 89(l):68-72, 1982.
Lasa MM, Datiles HI MB, Podgor MJ and Magno DV. Contrast and glare sensitivity:
Assocation with the type and severity of the cataract development. Ophthalmology, 99:7,
1992,
Lebel J, Mergler D, Lucotte M, Amorim M, Dolbec J, Miranda D, Arantes G, Rheault I,
Pichet P. Evidence of early nervous system dysfunction in amazonian populations
exposed to low-levels of methylmercury. NeuroToxicology, 17(1):157-168, 1996.
Leguire LE, Pappa KS, kachmer ML, Rogers GL, Bremer DL. Loss of contrast sensitivity in
cystic fibrosis. American Jouranl of Ophthalmology, 111:427-429, 1991.
Lennerstrand G, and Ygge J. Dyslexia: Ophthalmological aspects 1991. Acta Ophthalmol.
Copenh., 70:3-13, 1992.
Levin ED, Schmechel DE, Burkholder JM, Glasgow HB, Deameer-Melia NJ, Moser VC,
Harry GJ. Persisting learning deficits in rats after exposure to pfiesteria piscicida.
Environmental Health Perspectives, 105(12): 1320-1325, 1997.
Lindstrom K, Antti-Poika M, Rola S, and Hyytiainen A. Psychological prognosis of
diagnosed chronic organic solvent intoxication. Neurobehav. Toxicol. Teratol., 4:581-588,
1982.
Lord SR, Clark RD and Webster IW. Postural stability and associated physiological factors in
a population of aged persons. J. Gerontol. 46(3):M69-76, 1991.
Lord SR, Ward JA, Williams P, Anstey KJ. Physiological factors associated with falls in
older community-dwelling women. J. Am. Geriatr. Soc. 42(10): 1110-1117, 1994.
Loshin DS, and White J. Contrast sensitivity: The visual rehabilitation of the patient with
macular degeneration. Arch. Ophthal., 102:13003-1306, 1984.
Love PA, Kluka DA. Contrast sensitivity function in elite women and men Softball players. Intl
Journal of Sports Vision, 1992.
-------
HUDNELL 28
Lovegrove WJ, Bowling A, Babcock B, and Blackwood M. Specific reading disabilities:
differences in contrast sensitivity as a function of spatial frequency. Science, 210:439-440,
1980.
Lowitt MH, and Kauffman CL. Pfiesteria and the skin: a practical update for the clinician.
Maryland Med. J., 47:124-126, 1998.
Melcher MH and Lund DR. Sports vision and the high school student athlete. J. American
Optom. Assoc. 63(7):466-474, 1992.
Mergler D. Behavioral neurophysiology: Quantitative measures of sensory toxicity.
Neurotoxicology: Approaches and Methods, edited by L.W. Chang and W. Slikker, Jr.,
47:727-736, 1995.
Mergler D, Baldwin M, Belanger S, Larribe F, Beuter A, Bowler R, Fanisset M, Edwards
R, de Geoffrey A, Sassine M-P, and Hudnell K. Manganese neurotoxicity, a continuum
of dysfunction: results from a community based study. Neurotoxicol., submitted, 1998.
Mergler D, Blain L, and Lagace J-P. Solvent related colour vision loss: an indicator of neural
damage. Int. Arch. Occup. Environ. Hth., 59:313-321, 1987.
Mergler D, Huel G, Bowler R, Frenette B, Cone J. Visual dysfunction among former
microelectronics assembly workers. Arch. Environ. Health., 46:326-334, 1991.
Merigan WH. Chromatic and achromatic vision of macaques: role of the P pathway. J.
Neurosci., 9(3):776-783, 1989.
Merigan WH, Barkdoll E, Maurissen JP, Eskin TA, Lapham LW. Acrylamide effects on the
macaque visual system. I. Psychophysics and electrophysiology. Invest.Ophthalmol. Vis.
Science, 26(3):309-316, 1985.
Merigan WH, Wood RW, Zehl D, Eskin TA. Carbon disulfide effects on the visual system I.
Visual thresholds and ophthalmoscopy. Invest. Ophthalmol. Vis. Science, 29(4):512-518,
1988.
Merigan WH. Katz LM. and Maunsell JHR. The effects of parvocellular lateral geniculate
lesions on the acuity and contrast sensitivity of macaque monkeys. The Journal of
Neuroscience, 11(4):994-1001, 1991.
Merigan WH, Maurissen JPJ, Weiss B, Eskin T, and LW Lapham. Neurotoxic actions of
methylmercury on the primate visual system. Neurobehav. Toxicol. Teratol., 5:649-658,
1QR1
-------
HUDNELL 29
Morkeberg JC, Edmund C, Prause JU, Lanng S, Koch C, and Michaelsen KF. Ocular
findings in cystic fibrosis patients receiving vitamin A supplementation. Graefes Arch.
Clin. Exp. Ophthal., 233:709-713, 1995.
Morris JG, Charache P, Grattan LM, et al. Medical evaluation of persons with exposure to
water containing Pfiesteria or Pfiesteria-\k& dinoflagellates. Interim Report to Secretary
Wasserman, Maryland Department of Health and Mental Hygiene, 1997.
Mukuno K, Ishikawa S, and Okamura R. Grating test of contrast sensitivity in patients with
Minamata disease. Br. J. Ophthalmol., 65:284-290, 1981.
Muller, K.E., Barton, C.N. and Benignus, V.A. Rf commendations for appropriate statistical
practice in toxicological experiments. Neurotoxicol, 5:113-126, 1984.
Quiceno Jl, Capparelli E, Pharm D, Sadun AA, Munguia D, Grant I, Listhaus A, Crapotta
J, Lambert B, and Freeman WR. Visual dysfunction without retinitis in patients with
acquired immunodeficiency syndrome. American Journal of Ophthalmology 113:8-13,
1992.
Regan D. Human brain electrophysiology. New York, Elsevier Science Publishing Co., 1989.
Regan D, and Neima. Low-contrast letter charts in early diabetic retinopathy, ocular
hypertension, glaucoma and Parkinson's disease. Br. J. Ophthalmol., 68:885-889, 1984.
Rice, DC, and Gilbert SG. Early chronic low-level methylmercury poisoning in monkeys impairs
spatial vision. Science, 216:759-761, 1982.
Rice, DC, and Gilbert SG. Effects of developmental exposure to methyl mercury on spatial and
temporal visual function in monkeys. Toxicol. Appl. Pharmacol., 102:151-163, 1990.
Roman, GC. An epidemic in Cuba of optic neuropathy, sensorineural deafness, peripheral
sensory neuropathy and dorsolateral myeloneuropathy. J. Neurol. Sci., 127:11-28.
Roquelaure Y, Le Gargasson JF, Kupper S, girre C, Hispard E, Dally S. Alcohol
consumption and visual contrast sensitivity. Alcohol Alcohol, 30(5):681-685, 1995.
Rosenthal FS, West SK, Munoz B, Emmett EA, Strickland PT, and Taylor HR. Ocular and
facial skin exposure to untrlviolet radiation in sunlight: a personal exposure model with
application to a worker poplation. Hlth. Physics, 61:77-86, 1991.
Ross JE, Bron AJ, Reeves BL, and Emerson, PG. Detection of optic nerve damage in ocular
hypertension. Br. J. Ophthal. 69:897-903,1985.
-------
HUDNELL 30
Sadun AA, Borchert M, DeVita E, Hinton DR, and Bassi CJ. Assessment of visual
impairment in patients with Alzheimer's disease. Am. J. Ophthalmol., 104:113-120, 1987.
Sadun AA, Martone JF, Muci-Mendoza R, Reyes L, DuBois L, Suva JC, Roman G, and
Caballero B. Epidemic optic neuropathy in Cuba. Eye findings. Arch. Ophthalmol.,
112:691-699.
Sample PA, Pascal SC, Juang BS, and Weinreb RN. Isolating the effects of primary open-
angle glaucoma on the contrast sensitivity function. Am. J. Opthalmol., 112:308-316,
1991.
Savitz DA._Medical evaluation of estuary-exposed persons in North Carolina, November, 1997.
Report to the NC Pfiesteria Task Force. The University of North Carolina at Chapel Hill,
School of Public Health, Department of Epidemiology, 1998.
Schein OD, West S, Munoz B, Vitale S, Maguire M, Taylor HR, and Bressler NM. Cortical
lenticular opacification: distribution and location in a longitudinal study. Investig. Ophthal.
& Vis. Sci., 35: 363-366, 1994.
Schreiber J, Hudnell K and Parker J. Assessing residential exposure to tetrachloroethene using
biomarkers and visual system testing in populations living above dry cleaning facilities.
Toxicol. Sci. (supplement), 42:132, 1998.
Seeber A. Neurobehavioral toxicity of long-term exposure to tetrachloroethylene.
Neurotoxicol. Teratol., 11:579-583, 1989.
Seiple \VH, Holopigian K, Greenstein VC, and Hood DC. Sites of cone system sensitivity
loss in retinitis pigmentosa. Invest. Ophthalmol. Vis. Sci. 34(9):2638-2645, 1993.
Shinar D, and Schieber F. Visual requirements for safety and mobility of older drivers. Hum.
Factors, 33:507-519, 1991.
Sokol S, Moskowitz A, Skarf B, Evans R, Molitch M, and Senior B. Contrast sensitivity in
diabetics with and without background retinopathy. Arch. Ophthal., 103:51-54, 1985.
Taylor HR. Ocular effects of UV-B exposure. Doc. Ophthal., 88:285-293, 1995.
Taylor HR, Munoz B, West S, Bressler NM, Bressler SB, and Rosenthal FS. Visiblelight and
risk of age-related macular degeneration. Tr. Am. Opthal. Soc., 88:163-173; discussion
173-178, 1990.
Taylor HR, West S, Munoz B, Rosenthal FS, Bressler SB and Bressler NM. The long-term
effects of visible light on the eye. Arch. Ophthal., 110:297-298,1993.
-------
HUDNELL31
Taylor HR, West S, Rosenthal FS, Munoz B, Newland HS, and Emmett EA. Corneal
changes associated with chronic UV irradiation. Arch. Ophthal., 107:1481-1484, 1989.
Tracy JK, Oldach D, Greenberg DR, and Grattan LM. Psychological adjustment of watermen
with exposure to Pfiesteriapiscicida. Maryland Med. J., 47:130-132, 1998.
Trick GL, Burde RM, Gordon MO, Santiago JV, and Kilo C. The relationship between hue
discrimination and contrast sensitivity deficits in patients with diabetes mellitus.
Ophthal., 95(5):693-698, 1988.
Werner JS, Peterzell DH, and Scheetz AJ. Light, vision and againg. Optom. Vis. Sci.,
67(3):214-229, 1990.
West S, Munoz B, Schein OD, Vitale S, Maguire M, Taylor HR, and Bressler NM. Cigarette
smoking and risk for progression of nuclear opacities. Arch. Ophtha., 113:1377-1380,
1995.
West S, Munoz B, Emmett EA, and Taylor HR. Cigarette smoking and risk of nuclear
cataracts. Arch. Ophthal., 107:1166-1169, 1989.
Zoldek L and Roberts WA. The sensory basis of spatial memory in rat. Animal Learning and
Memory, 6:77-81, 1978.
-------
Table 1
Visual Acuity
Full Occupational Group
Group
Offshore
Estuary
N
Snellen Distance
Equivalent (20 :X)
SEM
t-score
p-value
20 29.75 ± 3.0
0.27 0.79
22 30.80 ±2.3
Visual Contrast Sensitivity
Source
.Group
Spatial Frequency (SF)
Group x SF
Degrees of Freedom
1,40
4,37
4,37
F-score
5.401
130.712
5.042
p-value
0.025
<0.001
0.002
1 Univariate ANOVA ' MANOVA (Wilks' lambda statistic)
Visual Contrast Sensitivity by Spatial Frequency
Spatial
Frequency (cpd)
1.5
3
6
12
18
Offshore
Mean VCS SEM
54 ±3.0
89 ±7.1
107 ±9.0
56 ±7.0
26 ±3.3
Estuary
Mean VCS SEM
48 ±3.7
78 ± 5.6
78 ± 6.5
40 ±3.1
21 ±2.5
t-score
1.24
1.19
2.70
2.09
1.38
p-value
0.224
0.242
0.010
0.043
0.174
-------
Table 2
Visual Acuity
Restricted Occupational Group
Group
Offshore
Estuary
N
Snellen Distance
Equivalent (20:X)
SEM
t-score
p-value
10 31.25 ±4.6
0.66 0.52
14 27.9 ±2.7
Visual Contrast Sensitivity
Source
, Group
Spatial Frequency (SF)
Group x SF
Degrees of Freedom | F-score
1, 22 4.591
4, 19 68.S32
4, 19 3.142
p-value
0.044
<0.001
0.038
1 Univariate ANOVA 2 MANOVA (Wilks' lambda statistic)
Visual Contrast Sensitivity by Spatial Frequency
Spatial
Frequency (cpd)
1.5
3
6
12
18
Offshore
Mean VCS SEM
59 ±3.3
98 ±11.9
115 ±15.1
61 ±11.5
29 ± 5.6
Estuary
Mean VCS SEM
50 ±5.2
80 ±6.3
79 ±8.1
40 ± 4.0
22 ±3.4
t-score
1.20
1.41
2.28
1.91
1.14
p-value
0.245
0.173
0.032
0.070
0.265
-------
Table 3 Is the Relationship Between VCS at the Mid-Spatial Frequency and
Group Affected by: Age, Education, Smoking, Alcohol, Total Time on Any
Water or First Order Interactions?
Multiple Linear Regression Analyses With Backward Elimination1
Full Occupational Group
Parameter1
Intercept
Group
Age
Smoking
Total Time on Water
Group x Age
Estimate
175.2
-118.5
-1.80
-0.53
0.008
2.15
Standard Error
33.1
44.1
0.79
0.32
0.004
1.04
p - Value
<0.001
0.011
0.030
0.104
0.039
0.047
1 Parameters were eliminated from the model if p > 0.15 and removal of the parameter did not
change the relationship of VCS at the mid-spatial frequency with group
Regression of VCS-6 with Age By Cohort2
Cohort
Estuary (N=22)
Offshore (N=20)
Slope
Standard Error
0.42 0.67
-1.81 0.87
p - value
0.545
0.055
The model also included smoking and total time on the water as predictors
Restricted Occupational Group
Parameter1
intercept
Group
Age
Smoking
Total Time on Water
Group x Age
Estimate
287.!
-265.3
-4.60
-0.64
0.012
5.75
Standard Error
52.9
66.5
1.30
0.33
0.004
1.62
p - Value
<0.001
0.001
0.003
0.072
0.007
0.002
1 Parameters were eliminated from the model if p > 0.15 and removal of the parameter did not
change the relationship of VCS at the mid-spatial frequency with group
-------
Table 4
Occupational Neurotoxicant Exposures in the Full Group
Estuary
(N=22)
Offshore
(N=20)
Hg
1
2
Pb
1
2
Other
Metals
5
8
Pesticides
2
1
Fumes
6
8
Solvents
10
11
Solvent
Years
130
136
-------
Table 5
Visual Acuity
Hot Line Cases & Controls
Group
Controls
Cases
N
Snellen Distance
Equivalent (20:X)
SEM
t-score
p- value
10 25.80 ±3.4
0.54 0.55
9 28.90 ± 4.7
Visual Contrast Sensitivity
Source
Group
Spatial Frequency (SF)
Group x SF
Degrees of Freedom
1,17
4, 14
4, 14
F-score
1.081
69.812
1.292
p-value
0.314
<0.001
0.319
1 Univariate ANOVA 2 MANOVA (Wilks' lambda statistic)
-------
August 12, 1998 HUDNELL
FIGURE LEGENDS
Figure 1. Visual contrast sensitivity (mean + SEM) functions for the estuary and offshore cohorts
in the full occupational group. MANOVA analyses indicated that the group factor and the group-
by-spatial frequency interaction term were significantly different. Step-down tests indicated that
the VCS scores of the estuary cohort at 6 and 12 cycles per degree of visual arc were significantly
lower than that of the offshore cohort.
Figure 2. Visual contrast sensitivity (mean ± SEM) functions for the estuary and offshore cohorts
in the occupational group restricted to include only participants free of potentially confounding
factors. MANOVA analyses indicated that the group factor and the group-by-spatial frequency
interaction term were significantly different. Step-down tests indicated that the VCS score of the
estuary cohort at 6 cycles per degree of visual arc was significantly lower than that of the offshore
cohort.
Figure 3. Visual contrast sensitivity (mean + SEM) functions for the cases recruited from the NC
Pfiesteria Hot-Line and matched-control participants. MANOVA analyses indicated that neither
the group factor nor the group-by-spatial frequency interaction term were significantly different.
Figure 4. A Biologically-Based Dose-Response Model (BBDR) for the health effects of
environmental toxin exposure in humans. The ultimate goal of the model is to mathematically
describe relationships between each of the stages from sources to health outcome. More complete
data on the dose of Pfiesteria toxin(s) applied to humans, as well as the absorbed dose, are needed
to support the hypothesis of a causal link with the observed difference between cohorts in VCS.
-------
VISUAL CONTRAST SENSITIVITY: FULL GROUP
140
120
_ 100
[Z 80
CO
UJ
CO
CO
o
o
CO
60 -
40 -
20 -
10
OFFSHORE (N=20)
ESTUARY (N=22)
1.5
12 18
SPATIAL FREQUENCY (Cycles / Degree)
-------
VISUAL CONTRAST SENSITIVITY: RESTRICTED GROUP
co
z
UJ
CO
CO
o
o
CO
140
120
100
80
60 -
40 -
20 -
10
OFFSHORE (N=10)
ESTUARY (N=14)
1.5 3 6 12 18
SPATIAL FREQUENCY (Cycles / Degree)
-------
VISUAL CONTRAST SENSITIVITY: HOT LINE
14C
120
_ 100
| 80
CO
g 60
CO
CO
40 -
O
O
< 20
CO
10
CONTROLS (N=10)
CASES (N=9)
1.5 3 6 12
SPATIAL FREQUENCY (Cycles / Degree)
18
-------
A Biologically-Based Dose-Response Model
Research to Improve Human-Health Risk Assessments
Dose
Applied
To
Humans
Human
Health
Outcome
Measure-
able
Change
Target-
Site
Dose
Toxic
Event
Absorbe
Dose
Sources
------- |