Most of the available data on the neurologic effects of hormonally active agents (HAAs), particularly in humans, are limited to polychlorinated biphenyls (PCBs) and dioxins. A connection has been hypothesized between prenatal and postnatal exposure to HAAs and disturbances in neurological and behavioral development. Possible mechanisms underlying such interactions have been suggested. In humans particularly, a number of difficulties are associated with the investigation of potential neurological and behavioral effects of exposure to HAAs in utero, including the recognition of such effects that may be quite subtle, and the possible long delay between exposure and outcome. The objective of this chapter is to review and analyze the available information in this case.
Studies with laboratory animals indicate that PCBs can cause impaired locomotor ability in rodents, impaired learning in rodents and monkeys, and impaired cognition in monkeys. For example, offspring of female mice treated with 32 mg/kg body weight (BW) 3,4,3',4'-tetrachlorobiphenyl on d 10-16 of gestation showed signs of neurotoxicity (intermittent stereotypic circling, head bobbing, and hyperactivity) in adulthood (Tilson et al. 1979). In a prenatal study with rats, the offspring of female rats fed 30 mg/kg Clophen A30 (a mixture of PCBs) in their diet 60 d prior to mating until postnatal d 21 had significant changes in active avoidance learning, operant conditioning, and open-field ambulations (Lilienthal et al. 1990; Lilienthal and Winneke 1991). Impaired learning and altered activity levels, such as increases in the reactivity to aversive events, were observed in rats fed a 30% diet of Lake Ontario salmon, which are contaminatedcontinue
with PCBs, DDT, DDE, mercury, and dioxin (Daly et al. 1989; Daly 1991). The offspring of monkeys fed 1.0 ppm Aroclor 1016 (a mixture of PCB congeners) during gestation and lactation were impaired in their ability to perform spatial discrimination-reversal learning tasks, but this was not observed in offspring of monkeys exposed to 0.25 ppm Aroclor 1016 or 1.0 ppm of Aroclor 1248 (Schantz et al. 1989). These varied effects with the amount of exposure and PCB congener type, suggest the importance of PCB congener identification in the measures of human exposure. In addition, significant deficits in cognitive ability were observed in infant and juvenile monkeys born 3 yr after maternal exposure to 2.5 ppm Aroclor 1248 had ended (Schantz et al. 1989). Possible mechanisms that have been suggested include alterations in neurotransmitter function (Rosin and Martin 1981), intracellular signaling mechanisms (Kodavanti et al. 1993, 1994, 1996), gonadal hormones (Tilson et al. 1979), and thyroid hormone function (Bastomsky and Murthy 1976; van den Berg et al. 1988). The structure of the PCB congener and the age of the animals at the time of exposure were shown to influence the effects (Seegal 1996). Furthermore, it is possible that some congeners influence the development of neurochemicals indirectly by altering the concentrations of thyroid hormones and neuroactive steroids (Seegal 1996). Schantz et al. (1995) reported on spatial learning behavior that was altered in adult rats by exposure to three PCB congeners, but thyroid hormone function was altered by only two of those. This highlights the complexity of the problem of teasing apart neurobehavioral and neurodevelopmental effects and mechanisms for cognitive functioning.
Neurobehavioral effects of exposure to PCBs and other chemicals that might act as HAAs have been studied extensively in four human populations, and studies are continuing in three other populations (Table 6-1). All but one of the studies focus on neurodevelopment, and most of the studies involve PCBs, polychlorinated dibenzofurans (PCDFs), DDE, and their metabolites.
In 1968, an industrial accident in Yusho, Japan, caused PCBs that were being used to clarify rice cooking oil to leak into the product. The leak went undetected for 9 mo, and it is estimated that 1,700 people were exposed (Kuratsune et al. 1972). Data on the developmental neurotoxicity from this incident are sparse, but reports described prenatally exposed children as dull, apathetic, and hypotonic and having subnormal intelligence (Harada 1976; Urabe et al. 1979). Little data on exposure concentrations were available, so no dose-response relationship could be demonstrated. However, this incident did serve to warn epidemiologists of the possible neurologic effects of PCB exposure.break
In May 1979, an unusual cluster of symptoms, characterized by chloracne and hyperpigmentation of the skin and nails, was reported in Taiwan (Hsu et al. 1985). By October 1979, PCBs and their heat-degraded products in samples of rice cooking oil were identified as the cause of the outbreak. In addition to PCBs, the contaminated oil contained an unusually high concentration of PCDFs, which can be more toxic than PCBs and could explain some of the reported effects (Schantz 1996). The resulting symptom complex, Yu-Cheng disease, was similar to that described in Yusho, Japan, ascribed to PCB-contaminated rice cooking oil. By the end of 1980, 1,843 cases had been reported in four central-Taiwan counties. Most were classified as mild (40.0%) or moderate (26.2%). Most of the affected population were between the ages of 11 and 20, and most were students or factory workers. It was estimated that PCB intake was 0.7-1.84 g, and the onset of clinical signs occurred 3-4 mo after exposure (Hsu et al. 1985).
Between 1979 and 1983, 39 babies with hyperpigmentation were born to exposed mothers. Their mortality rate was high; eight infants died of a variety of causes. The registry of Yu-Cheng, maintained by the Taiwanese Bureau of Disease Control, identified 128 surviving children born to 74 exposed women between June 1978 and March 1985. Over the next 2 decades, 1 8 of the children (born to 69 mothers) were followed to identify the cognitive-development effects of prenatal (transplacental) and postnatal (breast milk) exposure to PCBs (Hsu et al. 1985; Rogan et al. 1988; Yu et al. 1991). Exposure data were incomplete: Serum PCB concentrations were available from 61 of 69 exposed mothers (mean: 49.3 ppb; maximum: 456 ppb), serum concentrations were listed for 21 of 118 children (mean: 1 ppb; maximum: 77.8 ppb), and only fragmentary dietary data were available for the mothers. All of the exposed mothers in the study, however, had exposure-related symptoms at the time of enrollment. Control children were matched for neighborhood, age, sex, mother's age, educational level, and occupation of both parents. Testers were unaware of the child's exposure status (unless revealed by the parents' chloracne). Between the ages of 4 and 7 yr, there was a consistent, statistically significant, 5-point difference in the Chinese versions of the Stanford Binet test and Wechsler Intelligence Scale for Children, Revised (WISC-R) between exposed and control children. However, there was no relationship between either the mother's or the child's PCB serum concentration and developmental outcome. When the children's scores were arrayed by year of birth, there was some moderation in the effect among the 4- and 5-yr-olds born the longest after exposure. However, data from the 6- and 7-yr-olds indicated that children born up to 6 yr after the mothers' exposure and those born within 1 yr of exposure were equally affected (Chen et al. 1992).
The Japanese and Taiwanese studies demonstrate developmental delays and cognitive defects in young children prenatally exposed to PCBs and other products of heat-degraded PCBs. Because the studies were designed to assess thecontinue
(table continued on next page)break
(table continued from previous page)break
(table continued on next page)break
(table continued from previous page)break
effects of accidental exposures, there was little opportunity for consistent assessment of PCBs in cord blood or in maternal and infant serum. Thus, lack of good exposure measures makes assessment of a dose-response relationship difficult (Schantz 1996).
A series of studies on the cognitive development in offspring of women exposed to low concentrations of PCBs in their diet was conducted by Jacobson and Jacobson (1996) and Jacobson et al. (1985). A sample group of white, urban. predominantly middle-class women was recruited between 1980 and 1981 from four western Michigan hospitals for a longitudinal study (Michigan/Maternal Infant Cohort Study) of intrauterine and postnatal PCB exposure. The initial sample included 313 subjects: 242 women who reported moderate consumption of Lake Michigan sport fish (at least 1 1.8 kg during a 6-yr period) and 71 women (chosen as a 4.6% random sample) who reported no consumption of fish. Umbilical cord blood, maternal blood samples taken shortly after delivery, and breast milk samples taken 0.2-4.5 mo (median 1 mo) after delivery were analyzed for PCBs. PCBs were not detected in 70% of cord serum samples or in 22% of maternal serum samples. The mean concentrations of PCBs were 6 ng/mL in maternal serum, 3 ng/mL in cord serum, and 841 ng/g in breast milk.
The children of the exposed women were tested at 7 mo old with Fagan's test of visual recognition. The investigators controlled for infants whose gestational age was less than 38 wk by excluding them from the analysis, and they controlled for confounding variables. Preference for novelty decreased in a dose-dependent fashion with increasing prenatal PCB exposure, although postnatal exposure was not related. The effect of PCB exposure on recognition memory in newborns was not mediated by smaller body size or by behavioral deficits associated with PCB exposure: infants performing poorly at 7 mo had not necessarily exhibited deficits at birth (Jacobson et al. 1985).
When the same cohort of infants was tested at 4 yr (n = 236), prenatal exposure to PCBs was a good predictor of poorer short-term memory function on verbal and quantitative tests in a dose-dependent fashion; those effects persisted after controlling for a broad range of confounding variables (Jacobson et al. 1992). When the same children were tested on tasks designed to evaluate cognitive processing, prenatal exposure to PCBs was associated with less-efficient visual discrimination and with more errors in short-term memory, but not with a lack of sustained attention. Although much greater quantities of PCBs are transferred postnatally through breast-feeding, postnatal exposure was unrelated to cognitive and visual performance, suggesting that the critical period of exposure is during prenatal development (Jacobson et al. 1992).
Jacobson and Jacobson (1996) published a follow-up of this cohort assessed at 11 yr old. They examined 212 children, 68% of the 313 newborns studied Incontinue
1980-1981. Blood samples were analyzed for seven organochlorine pesticides and for lead. Hair samples also were taken for mercury analysis. Of the 212 children, 176 had been classified as fish-eaters. Children were tested at home at a mean age of 11 ± 0.2 yr for reading comprehension and with the WISC IQ test by testers blind to the exposure history of the child. Prenatal exposure to PCBs was associated with significantly lower full-scale and verbal IQ scores (p = .02). The IQ score of the most highly exposed group (1.25 µg/g breast milk, 4.7 ng in cord serum, or 9.7 ng in maternal serum) averaged 6.2 points lower than the other four groups after adjusting for potential confounders, including socioeconomic status, maternal age, age at testing, delivery complications, maternal drinking and smoking, parity, and number of children in household (p = .007). The deficit in IQ in the highly exposed children is similar to the effect of low-concentration exposure to lead (Bellinger et al. 1992). This highly exposed group was also three times more likely to have low average IQ scores (p < 0.001) and to have a twofold increased risk of being 2 yr behind in reading comprehension (p = 0.03). Intellectual impairments occurred only in relation to transplacental exposure, consistent with other studies. The investigators concluded that ''in utero exposure to PCBs in concentrations slightly higher than those in the general public can have a long-term impact on intellectual function" (Jacobson and Jacobson 1996). Although further corroboration from other studies is needed, this finding is consistent with the Jacobsons' earlier work (Jacobson et al. 1985, 1990, 1992) and with that of Lonky et al. (1996), reported below.
Several criticisms of the Jacobson studies must be considered (Paneth 1991: Schantz 1996; Seegal 1996). It has been charged that the work suffers from sample selection bias, differential loss to follow-up of the study cohort, and incomplete or inconsistent exposure assessment. Paneth (1991) questions the initial sample selection and comparability between exposed and control groups. The control group was slightly less than one-third the size of the exposed group (n = 71) (Schantz 1996).
Taken together, those criticisms suggest that the studies should be interpreted cautiously (Schantz 1996; Seegal 1996). However, Seegal (1996) points out that the Jacobson studies are the most intensive and detailed work to date on cognitive and developmental effects of human in utero exposure to complex mixtures.
Recently, the Agency for Toxic Substances and Disease Registry (ATSDR) evaluated the public-health implications of persistent toxic substances in the Great Lakes and St. Lawrence basins (Johnson et al. 1998). With regard to neurologic effects, ATSDR noted that "The findings of elevated PCB levels in persons who consume large amounts of Great Lakes-St. Lawrence fish, together with findings of developmental deficits and neurologic problems in children of fish-consuming mothers, are compelling." It was concluded that "neurobehavioral and developmental deficits occur in newborns and continue through school-age children from in utero exposure to persistent toxic substances, e.g., PCBs."break
Rogan et al. (1986a) conducted the North Carolina Breast Milk and Formula Project, a prospective birth-cohort study of 930 children born between 1978 and 1982. These children were followed to study the possible neurologic consequences of prenatal and postnatal exposure to environmental concentrations of PCBs and DDE. Women were recruited at or near term from hospitals, childbirth classes, and private and public prenatal clinics. No attempt was made to assemble a random sample. Of the women studied, 92% were white and well-educated, and most worked outside the home. A questionnaire was administered during hospitalization, and samples of placenta, maternal and cord blood, and breast milk and colostrum were assayed for PCB and DDE concentrations. Subsequently, additional serum and breast milk samples were obtained. The results from all available milk samples were combined, after scaling to account for differences between matrices and over time, to estimate the concentration of each chemical in milk fat at birth. No associations were seen for PCBs or DDE and either birth weight, head circumference, or hyperbilirubinemia, end points that had been associated with exposure to PCBs in the Taiwan and Yusho episodes. Neonatal testing was conducted during the first month of life using the Brazelton Neonatal Behavioral Assessment Scales (BNBAS). Because of attrition, results for this testing were obtained for 867 infants (93%) (Rogan et al. 1986a). The BNBAS analysis controlled for mother's age, education, occupation, smoking, drinking, and consumption of sport fish, as well as infant's sex, race, birth weight, and presence of jaundice. Tonicity and reflex cluster scores were significantly related to exposure to PCB or DDE. Exposure to high concentrations of PCBs in milk fat (=3.5 ppm) was associated with hypotonicity and hyporeflexia; and DDE exposure through milk fat (=4.0 ppm) was related to hyporeflexia. There was no evidence of deficit in memory function (Rogan et al. 1986b).
At the ages of 6 and 12 mo, 802 of the children were retested, using two instruments, the Bayley Mental Development Index (MDI) and the Bayley Psychomotor Development Index (PDI). Neither postnatal PCB exposure nor postnatal DDE exposure was related to either score at 6 or 12 mo, nor was the PDI related to prenatal DDE concentrations. However, the Bayley PDI decreased significantly with increasing prenatal PCB exposure at both points; 0.96/ppm (2.1 SE) and 1.34/ppm (2.2 SE) at 6 and 12 mo, respectively. A significant association between estimated prenatal DDE concentrations and Bayley MDI score was seen at 6 mo, but not at 12 mo (Gladen et al. 1988).
The children were again examined using the Bayley scales at 18 and 24 mo. At both ages, adjusted scores on the psychomotor scales were 4 to 9 points lower among children in the top quintile of exposure to transplacental PCB, and scores differed significantly at 24 mo. As with earlier test results from this cohort, no significant effects were seen in association with exposure through breast milk (Rogan and Gladen 1991). When tested at older ages (5.5-10.5 yr), using thecontinue
McCarthy Scales of Children's Abilities, the developmental changes seen at younger ages in relation to transplacental exposure to PCBs were no longer detectable (Gladen and Rogan 1991). This series of studies of neurologic function in children exposed to concentrations of PCBs and DDE that were ambient to the southeastern United States in 1978-1982 suggests small, but significant, neuromotor deficits that are no longer detected after early childhood.
Oswego, New York
An ongoing prospective longitudinal study has been designed to examine the behavioral effects of maternal consumption of Lake Ontario fish that contained a range of persistent chemicals, including PCBs, hexachlorobenzene (HCB), polychlorinated dibenzo-p-dioxin (PCDD), dieldrin, lindane, chlordane, cadmium, mercury, and mirex (Daly et al. 1996; Lonky et al. 1996). The study involves pregnant women recruited at 20 wk gestation and their offspring born between 1991 and 1994. Women were classified into exposure groups of high (consumed >40 equivalent pounds of fish; n = 152), low (consumed <40 equivalent pounds of fish; n = 243) and no exposure (n = 164). The parameter "equivalent pounds of fish" is a score based on fish species, number of years eating fish, number of meals per year, and serving size. Behavioral development was tested by the Neonatal Behavioral Assessment Scale. Trained observers, who were blind to the exposure status of the children, conducted the tests. Children were assessed 12-24 hr after birth and again at 25-48 hr. Principal component analyses controlled for three sets of variables: demographics, substances consumed during pregnancy, and labor-delivery-birth characteristics. The study sample was similar to the hospital population with respect to demographics and characteristics of labor and delivery, as were the three study groups. The study controlled for more than 100 confounders.
No significant associations were found between exposure and either weight or head circumference. Newborns in the high-exposure group scored more poorly than did the low or unexposed groups on three of the seven clusters: reflex, autonomic maturity, and habituation. The findings of poorer reflex functioning and greater autonomic immaturity in highly exposed infants are similar to the findings of Jacobson et al. (1985) and Rogan et al. (1986b) and thus strengthen the findings from those studies, which have been criticized on methodologic grounds. Further follow-up will be done to determine the persistence of effects as the children mature.
Efforts are under way to study the effect of recency of maternal fish consumption on neonatal coping behavior and infant temperament in this cohort at 24 mo (Darvill et al. 1997, as cited in Johnson et al. 1998).break
An ongoing study of perinatal exposure to PCBs is being carried out in the Netherlands (Huisman et al. 1995a). The protocol involves 418 children of normal birth (37-42 wk gestation), half of whom were breast fed. PCB exposure was determined with umbilical cord blood, maternal serum, and breast-milk samples. Neurologic optimality of the newborns was evaluated with the Prechtl neurologic examination between postnatal d 10 and 21. Children were classified as normal. suspect, or abnormal, and two clusters were scored: postural tone and reflex. PCB concentrations in cord and maternal sera were not related to neurologic dysfunction. However, higher concentrations of PCB, PCDD, and PCDF congeners in breast milk were associated with reduced neonatal neurologic optimality. Higher concentrations of planar PCBs in breast milk were related to an increased incidence of hypotonia (Huisman et al. 1995a).
A subset of these children (105 breast-fed and 102 formula-fed infants) were tested at 3, 7, and 18 mo with the Bayley Scales of Infant Development (Koopman-Esseboom et al. 1996). Prenatal exposure to PCBs had a small negative effect on psychomotor scores at 3 mo, but had no effect on mental scores. PCB and dioxin exposure through breast feeding had adverse effects on psychomotor outcome at 7 mo, while mental outcome was positively influenced in children exposed pre-and postnatally compared with children who were exposed prenatally only. At 18 mo. no adverse effects on psychomotor or mental development were found.
The entire cohort of 418 children were also tested with the Neurologic Examination for Toddler-Age at 18 mo. After adjustment for covariates (father's education, smoking of father, and parity), it was determined that transplacental PCB exposure is significantly and negatively related to motor function (grasping. sitting, crawling, standing, and walking) at 18 mo (ß coefficient -0.149: p = .003). Father's smoking modified this association, which was seen predominantly in children whose fathers did not smoke. On the other hand, little association was seen with additive lactational exposure, consistent with other studies (Huisman et al. 1995b).
Recently, 395 children from the original 418 mother-infant pairs in this study were tested with the Dutch version of the Kaufman Assessment Battery for Children (sequential and simultaneous processing) at 3.5 yr. Prenatal exposure to PCBs was measured using maternal and umbilical cord plasma samples, postnatal exposure to PCBs and dioxins was measured using breast-milk samples, and current exposure to PCBs was determined in the child's plasma. After adjustments for covariates, prenatal exposure to PCBs was significantly associated with lower cognitive scores, while postnatal exposure to PCBs and dioxin and current exposure to PCBs were not (Patandin et al. 1999). The children were also evaluated for motor function using the Touwen/Hempel method (motor functions prehension, sitting, crawling, standing, and walking) at 3.5 yr, and no effect on neurologic condition was found (Lanting et al. 1998).break
In collaboration with the studies in the Netherlands, Winneke et al. ( 1998) investigated the neurodevelopmental effects of perinatal exposure to PCBs in 171 newborns in Düsseldorf, Germany. Children were evaluated using the Bayley scales of infant development (BSID), psychomotor development tests, and the Fagan visual recognition memory test at 7 mo. Mean concentrations of PCBs were 0.55 ng/mL in cord blood and 427 ng/g in the fat of breast milk. After adjusting for confounders, the only significant association was a negative association between PCB concentrations in breast milk and mental development index.
A neurologic assessment of an aging population of Great Lakes fisheaters is currently being conducted by Schantz et al. (1996). In all, 104 fisheaters and 84 nonfisheaters, age 50 or older, were enrolled in the study. Blood samples from 180 of the 188 study participants were analyzed for PCBs and ten other contaminants. The blood concentrations of PCBs and DDE in the fish-eaters were found to be elevated in comparison to age- and sex-matched nonfisheaters. A battery of neuropsychologic tests, including tests of motor function, memory and learning, executive functions, visual-spatial function, and abstract reasoning, have been conducted, but the data are still under analysis.
A study has also been conducted that compares nervous system function in adult consumers of fish from the upper St. Lawrence River lakes to nonfisheaters (Mergler et al. 1998). No significant differences were found between the groups in tests of sensory function, visual memory and recognition, fine motor performance, and some motor tests. However, the fisheaters performed more poorly on tests requiring cognitive flexibility, word naming, auditory recall, and more complex motor task compared with individuals who do not eat fish. It should be noted that the observations were made within a larger study of early neurotoxic effects of environmental exposure to manganese, and was not designed to examine the effects of fish eating. Furthermore, paired analyses showed that consumers of fish had higher levels of organic mercury and lead in their blood.
New Bedford, Massachusetts
Mother-infant pairs residing near the New Bedford Superfund site, which is known to be significantly contaminated with PCBs, are being evaluated for adverse neurologic effects (Korrick 1998 as cited in ATSDR 1999). Preliminary data suggest that cognitive development, as measured by the Fagan Test of Infant Intelligence, is impaired by in utero exposure to PCBs.break
Summary and Conclusions
Most of the human data on neurologic effects of HAAs involve studies of newborns and children because the developing nervous system is thought to be the most sensitive to toxic agents. Long-term epidemiologic studies of cognitive and neurobehavioral development have been conducted in Michigan, New York, North Carolina, and the Netherlands on children exposed pre- and postnatally to PCBs from maternal consumption of contaminated fish or other food products. Studies of cognitive development (i.e., short-term memory, visual discrimination, and IQ scores) in Michigan show consistent correlations between prenatal exposure to PCBs and deficits at up to 11 yr (Jacobson and Jacobson 1996). Similarly, in the Netherlands, lower cognitive scores were associated with prenatal exposure when tested in 3.5-yr-old children (Patandin et al. 1999). In contrast, studies in North Carolina (Gladen et al. 1988; Gladen and Rogan 1991; Rogan and Gladen 1991) and Germany (Winneke et al. 1998) found no association between prenatal exposure and cognitive development.
In studies of neurobehavioral development (i.e., sitting, crawling, and walking), deficits were observed in the North Carolina cohort of children until 2 yr but not thereafter. Neurobehavioral deficits have also been reported in the Michigan and New York cohorts. Reports regarding neurobehavioral development from the Netherlands are inconsistent. For example, one study measuring psychomotor development at 18 mo with the Neurologic Examination for Toddler-Age found a significant reduction in neurologic function in prenatally exposed children (Huisman et al. 1995b), while another study of psychomotor development using the Bayley Scales of Infant Development found slight effects at 3 mo but none at 7 or 18 mo (Koopman-Esseboom et al. 1996). A study conducted with children at 3.5 yr also found no effects (Lanting et al. 1998). Some adverse neurobehavioral effects have been reported from postnatal exposure in studies of 10-21-d-old children (Huisman et al. 1995a) and 7-mo-old children (Koopman-Esseboom et al. 1996) in the Netherlands. However, no significant effect on neurobehavioral development was found in perinatally exposed children in Germany (Winneke et al. 1998).
Thus, data from the United States and some of the data from the Netherlands suggest a correlation between prenatal exposure to PCBs (due to maternal consumption of contaminated fish or other food products) and effects on cognitive and behavioral development in children. For the most part, postnatal exposure via breast milk does not appear to contribute significantly to these outcomes. Studies of PCBs in laboratory animals support these findings. In particular, studies of monkeys exposed prenatally to PCBs have shown impaired learning ability and prenatal exposure of rats and mice to PCBs have also shown impaired locomotor ability and learning. Taken together, the results of animal and human studies indicate that prenatal exposure to PCBs can affect neurologic development. The mechanisms underlying these effects have yet to be determined.break
The majority of the data reviewed by the committee on the neurologic effects of HAAs focused on only a few chemicals, namely PCBs, PCDF, dioxin, and DDT. Based on the committee's evaluation of those data, the following are recommended:
Human populations suspected of being affected by HAAs should continue to be monitored for adverse neurologic effects. Longitudinal tests should be conducted on developmental landmarks or milestones from conception through adulthood. A standardized set of criteria should be established to measure functional and social development and physical and clinical parameters. Burdens of HAAs should be measured in relation to the effects.
Further attempts should be made to identify the specific agents and the mechanisms of action underlying the human neurologic effects that have been associated with exposure to PCBs, PCDFs, dioxins, and DDT.break