Vulnerability of Children and the Developing Brain to Neurotoxic HazardsBernard WeissDepartment of Environmental Medicine, University of Rochester School of Medicine and Dentistry, Rochester, New York, USA

For much of the history of toxicology, the sensitivity of the developing organism to chemicalperturbation attracted limited attention. Several tragic episodes and new insights finally taught usthat the course of early brain development incurs unique risks. Although the process is exquisitelycontrolled, its lability renders it highly susceptible to damage from environmental chemicals. Suchdisturbances, as recognized by current testing protocols and legislation such as the Food QualityProtection Act, can result in outcomes ranging from death to malformations to functionalimpairment. The latter are the most difficult to determine. First, they require a variety of measuresto assay their extent. Second, adult responses may prove an inadequate guide to the response ofthe developing brain, which is part of the reason for proposing additional safety factors for children.Third, neuropsychological tests are deployed in complex circumstances in which many factors,including economic status, combine to produce a particular effect such as lowered intelligencequotient score. Fourth, the magnitude of the effect, for most environmental exposure levels, maybe relatively small but extremely significant for public health. Fifth, changes in brain function occurthroughout life, and some consequences of early damage may not even emerge until advancedage. Such factors need to be addressed in estimating the influence of a particular agent or group ofagents on brain development and its functional expression. It is especially important to considerways of dealing with multiple risks and their combinations in addition to the prevailing practice ofestimating risks in isolation. Key words: behavioral teratology, developmental toxicity, earnings,Food Quality Protection Act, IQ, multiple risks, Parkinson's disease, pesticides, social environment.- Environ Health Perspect 1 08(suppl 3:375-381 (2000).http.//ehpnetl.niehs.nih.gov/docs/2000/suppl-3/375-381weiss/abstract.html

Near the beginning of our just-expired cen-tury, the toxicology of that time was goadedby pioneers like Harvey Wiley, whose persis-tence and stubbornness helped create theU.S. Food and Drug Administration (FDA).Wiley became the first FDA Commissionerin 1907, but he had earlier become gravelytroubled by the adulteration of foods byunscrupulous or ignorant producers. Whilechief chemist of the U.S. Department ofAgriculture, he recruited what became knownas the "Poison Squad," a group ofyoung menwho consumed food treated with variouschemicals to determine if they might experi-ence adverse reactions including depression,headaches, and even more severe symptoms.Photographs taken around 1905 show themseated at dining tables, wearing coats and ties,as they valiantly consumed their tainted diets.Informed consent was not an issue at the timebut probably would not have dissuadedWiley's acolytes. They considered themselvespublic servants, an intriguing distinction withthe current practice and ethical quandaries ofpaying volunteers to consume pesticides so asto change acceptable exposure standards.

Healthy young men would have been ajudicious choice for poison squad subjectsbecause, if they proved sensitive, Wiley couldargue that less robust individuals might moreeasily succumb. There are no reports to indi-cate that he did so. In fact, the question ofchildren as special targets of food additives,environmental poisons, and even drugs

apparently arose much later. A distinctivelabel for prenatal ethanol toxicity, FetalAlcohol Syndrome, did not appear until the1970s (1) despite centuries of anecdotalobservations. The contemporary discipline ofteratology is largely a product of the shoddybehavior of the company that concealed thehorrors of thalidomide, which had been mar-keted as a safe effective sedative during preg-nancy. A worldwide epidemic of birth defectsbegan to emerge during the 1950s and early1960s in the babies of women who had beenprescribed thalidomide. The German firm,Chemie Griunenthal that had manufacturedand marketed thalidomide kept such reportssecret until the evidence became overwhelm-ing, but, by that time, the drug had claimedthousands of victims.

The United States was spared this tragedybecause of the stubbornness of Frances Kelsey,who received the President's Award forDistinguished Federal Civilian Service in1962 for denying FDA approval of thalido-mide. The Kefauver-Harris amendments thatsame year to the Food, Drug and CosmeticAct (2), which had last been amended in1938, were enacted in response to the thalido-mide crisis. The amendments added a require-ment for proof of effectiveness for drugs, andalso required manufacturers to send reports ofadverse drug reactions to the FDA.

Even after the thalidomide incident, thefull dimensions of developmental toxicityremained unappreciated. The ancient metals,

mercury and lead, exemplify how little recog-nition, until quite recently, was accorded thespecial vulnerabilities of children. Needleman(3) has related the lead story on many occa-sions, noting the recognition of lead poison-ing in workers at the same time that theeffects on child development escaped notice.The mercury story is equally baffling in itsexclusion of children. Mercury was recog-nized as a poison even in antiquity. Its neuro-toxic properties, in the form of mercuryvapor, received labels such as "hatter'sshakes," which described its propensity toinduce excessive tremor, and "erethism," acollection of psychological disorders manifest-ing features such as being hyperirritable;blushing easily; having a labile temperament;avoiding friends and/or public places; beingtimid and/or shy; being depressed and/ordespondent; suffering from insomnia; andsuffering from fatigue.

Mercury as a potential developmentaltoxicant received no attention. The childhoodaffliction of Pink Disease, or acrodynia,which had been discussed in the medical liter-ature earlier in the century, was typicallyascribed to infectious disease. Its symptomsdid not coincide at all with those expressed byadults who had been poisoned by mercury.Pink disease symptoms included: apathy andirritability; rashes and sloughing of skin;reddening of cheeks and nose; cold blue fin-gertips and toes; profuse sweating and hyper-tension; itching and/or burning sensation;photophobia and anorexia; hypotonicity andtremor; pain in extremities and paresthesias;and muscle twitches.

Not until 1947 was Pink Disease linkedto mercury. Mercury appeared mostly in theform of calomel (mercurous chloride), which

This article is based on presentations made at the 20thRochester Conference on Environmental Toxicity titled"The Role of Environmental Neurotoxicants inDevelopmental Disabilities" held 23-25 September1998 in Rochester, New York, and at the conference"Environmental Influences on Children: Brain,Development, and Behavior" held 24-25 May 1999 inNew York, New York.

Address correspondence to B. Weiss, Departmentof Environmental Medicine, Room G-6820, Universityof Rochester Medical Center, 575 Elmwood Ave.,Rochester NY 14642 USA. Telephone: (716) 275-1736.Fax: (716) 256-2591. E-mail: bernard_weiss@urmc.rochester.edu

Preparation supported in part by grants ES01247,ES08109, and ES08958 from the National Institute ofEnvironmental Health Sciences.

Received 5 November 1999; accepted 20 January2000.

Environmental Health Perspectives * Vol 108, Supplement 3 * June 2000 375

B. WEISS

was widely used in teething powders (4).Once calomel was removed from teethingpowders, the incidence of Pink Disease plum-meted. Now, Pink Disease is largely a prod-uct of exposure to mercury vapor, which isstill pervasive in the environment. Older latexpaints, which contained a phenylmercury as afungicide, emitted mercury vapor after appli-cation; jewelry factories used elemental mer-cury to produce gold amalgams and literallyleft pools of it behind in sites they aban-doned; fluorescent bulbs contain mercury inthe ballast assemblies; some religious groupsuse elemental mercury in their ceremonies;and certain ethnic groups use mercury com-pounds in cosmetics such as skin lighteners.

The Current SceneTesting for developmental toxicity hasbecome a fundamental component of safetyassessment for environmental chemicals. Itencompasses the end points of fetal death,malformations, altered growth, and the muchbroader question of functional abnormalities.This last criterion is the most difficult todetermine. It embraces a vast array of possibil-ities, most of which are themselves comprisedof multiple indices; learning deficits, for exam-ple, describe a class of problems rather than asingle problem. They may emerge clearly longbeyond gestation and only in specific settingssuch as the classroom. Functional measuresare typically affected by lower exposure levelsthan other end points relied on as indices ofdevelopmental neurotoxicity.

Fetal Alcohol Syndrome offers an exampleof the progression from an emphasis on struc-ture to an emphasis on function (1). In itsoriginal formulation, it featured craniofacialdysmorphologies. With further investigationfocused on the effects of lower levels ofmaternal alcohol intake than those evokingstructural defects, a new syndrome, labeledFetal Alcohol Effects, emerged. Now knownas Alcohol-Related NeurodevelopmentalDisorder, it is applied to patients who displaybehavioral or cognitive abnormalities such aslearning difficulties.

The special status of functional disordersinspired the term and discipline of behavioralteratology. Behavioral teratology describesfunctional abnormalities induced by prenatalexposures rather than the structural deficitsassayed by teratologists. That identificationhas taken on a quasi-official status. Regulatorybodies around the world have come to recog-nize the importance of testing for functionalabnormalities produced by pharmaceuticals,spurred by data that indicate the remarkablescope of drugs capable of leaving a residue ofbehavioral problems (5). These types ofdrugs include: vitamin analogs, cytostatics,gastrointestinals, hyperlipemics, bronchiolyt-ics, spasmolytics, analgesics, antiallergens,

antibiotics, cardiovasculars, anticoagulants,and anticonvulsants.

The FDA lags behind its counterparts inGreat Britain, France, and Japan in recogniz-ing neurobehavioral assays as an essentialaspect of drug testing. One example of howthe FDA sometimes discounts neurobehav-ioral toxicity in children is seen in its require-ments for the safety testing of food additives.Despite a string of publications since the late1970s showing that some children respondadversely to additives, especially food dyes, atlevels prevailing in the diet, the agency persistsin its claim that "... well-controlled studies ...have produced no evidence that food coloradditives cause hyperactivity or learning dis-abilities in children" (6). Such a statement dis-regards reports based on controlled clinicaltrials (7-9) and lacks familiarity with the kindof careful statistical analysis required toidentify susceptible subpopulations (10).

For most potentially toxic chemicals, wepossess only animal, typically rodent, data.To provide an adequate margin of safety forhumans, current practice first examines thedose-response function to select a dose statis-tically indistinguishable from the controlpreparation [the no-observed-effect level(NOEL)]. If the NOEL is derived from achronic exposure study, it then divides thisdose by a safety or uncertainty factor of 100to calculate a reference dose or allowable dailyintake: a factor of 10 to account for speciesdifferences and a factor of 10 to account forvariations in sensitivity among individuals.Currently, for food additive toxicity testing,the FDA requires manufacturers to fulfill anextensive set of protocols, including teratolog-ical evaluations, in rodents. It then applies the100-fold safety factor. Neurobehavioral test-ing is not required. The difference betweenwhat the literature shows in children andallowable daily intakes based on traditionalprotocols is highlighted in Table 1. The levelsof food colors eliciting behavioral responses inchildren are 50-60 times less than the con-ventional exposure standard.

Table 1. Comparison of food dye doses (in milligrams).Weiss Nutrition

Color et al.a Foundationb ADIcBluel 0.80 0.85 300Blue 2 0.15 0.46 37Red 3 0.57 1.66 150Green 3 0.11 0.04 150Red 40 13.80 10.45 420Yellow 5 9.07 7.34 300Yellow 6 10.70 6.20 300ADI, allowable daily intake.aln a challenge experiment in which they elicited acute behav-ioral responses in young children 17). bEstimated intakes in chil-dren calculated by the Nutrition Foundation and used in severalchallenge studies (11). CADI derived from animal testing andbased on NOELs divided by a safety factor of 100 (12. The doseseliciting behavioral responses in children are many times lowerthan the ADI, which is not based on neurobehavioral testing.

Pesticides and the FoodQuality Protection ActOne major outcome of our recognition thatdeveloping organisms may be uniquely vul-nerable to chemical challenge is the FoodQuality Protection Act (FQPA) of 1996 (13).It may be viewed as the second stage of theNational Academy of Sciences report (14)that noted the enhanced health risks arisingfrom the combination of developmental sus-ceptibility and the higher relative intakes ofpesticides by children. The FQPA embodiedthe following risk assessment issues: inclusionof an extra 10-fold safety factor for children,in utero exposures, cumulative effects of pesti-cides and substances with a common mode ofaction, aggregate exposure (e.g., food andwater), and endocrine disruption.

The most controversial feature of theFQPA specifies, at the discretion of the U.S.Environmental Protection Agency (EPA), anadditional 10-fold safety factor for calculatingacceptable intakes for children. This attributeof the FQPA has prompted objections amongmany elements of the pesticide industrybecause it can sharply restrict sales of manyproducts whose residues appear in foods.

The other entries in the list arise from thesame apprehensions that prompted the addedsafety factor. In utero exposures are requiredbecause, for so many agents, as recognized bycurrent pesticide regulations, fetal develop-ment is an exceedingly sensitive stage of thelife cycle, offering the prospect of latent dam-age whose consequences will surface later inlife. The reference to cumulative effectsacknowledges that different products may actin similar ways, especially if they come fromthe same chemical class; individual chemicalexposure levels may fall within regulatory lim-its but prove toxic in the aggregate.Moreover, even chemicals from differentclasses may still share the same mechanisms oftoxicity or be measured by equivalent endpoints; for example, both lead and polychlori-nated biphenyls (PCBs) can reduce intelli-gence quotient (IQ) scores. Similarly,exposure may come from different sourcessuch as produce residues, pesticide deposits inareas such as schoolyards where children play,contaminated drinking water, and residentialsurfaces. It makes little sense to treat stan-dards for each source separately.

The FQPA (13) also included languagedirected at the question of endocrine disrup-tion. We have come to recognize that manychemicals in the environment, including thosefound in nature, can exert powerful effects onhormone action, particularly on the role hor-mones play in early development. In responseto FQPA concerns, the EPA created theEndocrine Disruptor Screening and TestingAdvisory Committee (EDSTAC) to design a

Environmental Health Perspectives * Vol 108, Supplement 3 * June 2000376

VULNERABILITY OF CHILDREN TO NEUROTOXIC HAZARDS

screening and testing strategy to determinehow to select chemicals for screening, how tochoose the assays included in a screening bat-tery, and the circumstances under whichchemicals should be subjected to further test-ing. EDSTAC's targets are primarily estro-genic chemicals, however. A number ofpesticides act as androgen antagonists and, inrats exposed in utero, induce signs of feminiza-tion, such as prominent nipples, in males(15). Testing for antithyroid properties is alsonot projected, although such agents may posean even greater developmental hazard thanestrogenic chemicals because of the highprevalence of iodine deficiency.

Safety Factors for ChildrenTwo opposing viewpoints have clashed overthe added 10-fold safety factor. Industrygroups contend that current testing guidelinesalready include provisions aimed at develop-mental exposures such as two-generationstudies. Their challengers argue that the pro-visions are insufficient to protect children.These latter groups base their argument partlyon the lack of information about variations indietary practices, inadequate monitoring ofresidues, and other gaps in knowledge. Butthe strongest argument for the additional pro-tection is what might be called the ecology,particularly the spatial ecology, of childhood.These include such factors as activity altitude(young children spend much time on floors;they stir up and breathe dust and residues;and exposure to dust may be 10 times greaterthan adults); proportionality (children con-sume relatively more juice, fruit, and waterthan adults); exploratory behavior (youngchildren literally lead a hand-to-mouth exis-tence); and breast milk (lipid-soluble agentscan be transmitted in quantity).

Children occupy a distinctive ecologicalniche in a world tuned to adults. Size isimportant because of what might be termedthe altitude factor. Children's activities takeplace close to ground level. Dense vapors suchas mercury collect at much higher concentra-tions near the floor than at adult waist level.The same would be true for any volatile mate-rials that are denser than air. With their activi-ties, children also stir up floor dust that theythen inhale. One reason that farm familiesand agricultural communities typically experi-ence greater pesticide exposures than the bulkof urban dwellers (16-18) is inadvertent depo-sition in the home through air currents, foottraffic, and contaminated clothing. In somecities, as in the upper west side of Manhattan,New York, pesticide use also can be heavybecause of attempts to eliminate vermin (19).Lead in household dust is an example of a pre-viously neglected exposure source that is sig-nificantly correlated with blood levels in thechildren (20). One frequent concomitant of

child exploration, hand-to-mouth sampling, isanother exposure source that is poorly quanti-fied. A recent survey and analysis (21) indi-cates that a small percentage of children mayingest much greater quantities of soil than theEPA standard of 200 mg/day.

Finally, breast milk, despite its great advan-tages, is also an efficient medium for transfer-ring lipid-soluble chemicals from mothers toinfants. All of these exposure sources demon-strate the futility of trying to characterize riskby reliance on a deductive sequence stretchingfrom NOELs, to reference doses, to residuelevels. The stipulations of the FQPA representonly the prelude to the more global, compre-hensive analysis of exposure required for adetailed accurate assessment of risk.

Even an additional safety factor for chil-dren fails to take account of exposure spikesthat may be experienced by many children.Developmental data from humans consistalmost entirely of poisoning cases, probably aresult of the fact that most calls to poison con-trol centers about pesticides involve children.In Minnesota in 1988, pesticide poisoningsaccounted for 1,428 case files (22). The meancase age was 5 years, and 50% of all cases wereyounger than 3 years of age. Insecticides inresidences were identified as the source inmost of the case files. In Toronto, Ontario,Canada, the Hospital for Sick Children saw1,026 cases of pesticide poisoning in a singleyear, 597 of which occurred in childrenyounger than 6 years of age. These rates aretypical of poison control centers (23). Follow-up observations are rare, unfortunately, evenfor those cases that require medical interven-tion. Children treated for poisoning by pesti-cides such as lindane and diazinon wereobserved by Angle et al. (24) to suffer appar-ently persistent neurobehavioral problems.Angle et al. (24) noted that "... there are nolong-term studies of the effects of poisoningson children," and "... a provocative associa-tion ... with later intellectual deficit . butthe intervening 30 years have seen no system-atic attempts to explore these questions (25).One exception to the dearth of developmentaldata is the novel report (26) comparing twoagricultural communities in Mexico.Residents came from the same ethnic stock,but one community had adopted chemical-based agriculture, whereas the other hadrejected it. Various indices of neuropsycholog-ical development suggested that children fromthe traditional community proved superior onseveral measures.

One other factor not explicitly acknowl-edged by the FQPA is what we have learnedfrom agricultural workers. Evidence of subtleneurobehavioral toxicity in adults, even atclinically silent exposure levels (27-29) hasbeen accumulating steadily. Data verifyinglingering neurobehavioral deficits in adult

farmworkers after a clinically toxic exposure(30-33) have also been accruing. Given thepronounced vulnerability of the developingbrain to most neurotoxicants, we shouldanticipate that children will display evenmore pronounced responsiveness to the neu-rotoxic properties of pesticides than adults.The question might be framed in the form ofa pyramid, as depicted in Figure 1. Clinicalpoisoning severe enough to necessitate med-ical treatment occurs in a relatively smallnumber of children. Subclinical poisoningdetectable with neuropsychological testingwould reveal an even larger affected popula-tion. Latent toxicity, also conceived of assilent toxicity (34), may not become apparentuntil additional challenges to function, suchas the demands of the classroom, supervene.Another is aging.

Many investigators have suggested thatthe roots of some neurodegenerative diseasesmay be found in events that occurred duringearly development (35). These authorsproposed that

... Alzheimer's disease, Parkinson's disease(PD), and motoneurone disease are due toenvironmental damage to specific regionsof the central nervous system and that thedamage remains subclinical for severaldecades but makes those affected especiallyprone to the consequences of age-relatedneuronal attrition.

One source of support for such suggestions isthe relationship between a history of pesticideexposure and PD. In some parts of the brain,cell numbers dwindle with age. The change isespecially noticeable in the substantia nigra(SN), an area in the subcortical collection ofnuclei known as the basal ganglia. The cells inone region of SN secrete the neurotransmitterdopamine. Clinical signs of PD emerge whenmost of these nerve cells die and can no longersynthesize adequate supplies of dopamine.Latent toxicity is one way of viewing the

Clinical pesticidepoisoning

Subclinical pesticideK poisonlng

Latent pesticidei toxicity

Figure 1. The toxicity pyramid for pesticides is designedto show that, although only a small proportion of chil-dren will manifest clear signs of excessive exposure,many more will show effects detectable by neurobehav-ioral testing, and even more will endure latent or silenttoxicity that only emerges with additional challengessuch as other pesticides, other environmental chemicals,or other health problems.

Environmental Health Perspectives * Vol 108, Supplement 3 * June 2000 377

B. WEISS

association between a history of pesticideexposure and an elevated risk ofPD (36). Thepossibility of such an environmental sourcefor PD is buttressed by the evidence (37)showing no difference in concordance ratesfor PD in monozygotic and dizogotic twins60 years of age and older.

One possible explanation of the associa-tion is an acceleration of natural cell loss inSN induced by pesticide exposure, as shownin Figure 2. It compares the rate modeled onpublished data (38) with rates calculated onthe basis of additional accelerations of 0.1and 0.3% annually. Even a rather smallacceleration of 0.1% per year means that the40% reduction in cell density that wouldhave occurred by the early 1970s is advancedby about 10 years. It is hardly a negligibleshift and entails significant health and med-ical consequences.

These consequences are depicted in Figure3. The rate of onset of PD rises steeply withage. The baseline curve plots representativedata from a U.S. urban community (39). Theother two curves show equivalent rates of risebut are displaced by 5 and 10 years, respec-tively. At age 70 and above, the empirical

- 1.01-O 1 0 Xc_ * McGeer et al. (38)0.9 ~~~~~~~~~AExtra 0.1%

' ~ 0.8 - Extra 0.3%cE 0.7 -

0.6 -0.5-

2 0.40.3

O 10 20 30 40 50 60 70 80 90Age (years)

Figure 2. Cell numbers decline with age in severalareas of the human brain. In the SN, the decline seemsto progress throughout life (38). Slight accelerations ofthe process (0.1 and 0.3% per year) can foster the ear-lier clinical emergence of PD. The arrows correspond toa 40% loss, which some data suggest is a precursor ofclinical PD.

°. 1,400 -

° 1,200 -CD 1,000-o 800-~ 600-

400-0w 200-Oc> 0CL

A Baselineo Advanced 5 years* Advanced 10 years / /

~~1.11-l..,-.,.~~'i/---- /

IT T T

30 35 40 45 50 55 60 65Age (years)

70

Figure 3. Incidence of PD with age. Baseline data (39)are compared to earlier hypothetical displacements of 5and 10 years. Assume a baseline prevalence of about0.012 at age 70 and above (39). In the year 2000, thenumber of PD patients older than 70 years of age wouldcome to 309,672. Assuming the same prevalence at age60, that is, displaced by 10 years, would bring the totalto 555,696. If the prevalence function were to be shifted5 years earlier, that number would be 424,728.

curve shows a prevalence of about 0.012.Demographic projections based on that preva-lence figure indicate that, in the year 2000,the number of PD patients above 70 years ofage would become 309,672. If the prevalencefunction were to be shifted by 10 years, thetotal would come to 555,696.

The Social Environment as aDimension of RiskExposures to environmental chemicals repre-sent only one dass of risks during child devel-opment. Investigators have learned to behighly sensitive to this problem, and whenthey undertake to study risk factors in humanpopulations, neurotoxicologists typically obeythe conventions of traditional epidemiology.If an investigation targets a particular agent, itattempts to strip away those influences thatmight confound the effects of that agent inisolation. This practice is essential to theanalysis of dose-response relationships, to theformulation of experimental hypotheses, tothe planning of laboratory research, and tothe pursuit of biological mechanisms. Thevirtues of such a focused approach begin todissolve, however, when they have to betranslated into risk characterizations and pol-icy implications. Neurotoxicant exposure andits consequences for development are moldedinto a larger environment from which theycannot easily be extricated. The literature onlead offers one illuminating example.

Lead research during the past 30 years hasmostly converged on behavioral developmentin children. The predominant measure,adopted because it is the one most easilycommunicated, because it is the beneficiary ofan immense research effort extending over acentury, and because it summarizes a varietyof functions, is IQ score. Often confused bythe public, and even by some scientists, withan abstract quality termed "intelligence," itremains useful because of its correlations withacademic success, earnings, and even broadercriteria such as welfare dependency.

What is now a vast literature documentsan inverse relationship between indices of leadexposure in children, such as blood level, andIQ score. The relationship itself, as a simplestatistical measure, arouses relatively muteddebate. What does excite vigorous argument isinterpretation. The basis for argument inneurobehavioral toxicology is the contributionof other variables. The contenders agree thatlead is only one component of a multitude offactors acting on IQ score. Maternal intelli-gence, family income, parental intellectualstimulation and care, maternal and paternaleducation, race, marital status, maternal age,extent of prenatal care, maternal drug use(alcohol, tobacco, and illicit drugs), maternalpsychopathology, and other aspects of thechild's history and environment also influence

IQ score. The strategy universally adopted byserious investigators is to compensate statisti-cally for these other influences, typically bymultiple regression procedures. Disagreementsarise over which confounders or control vari-ables to include, and how to do so.

The practice of isolating the contributionof a selected neurotoxicant is scientifically log-ical but possibly misleading for public policy.Stripping away potential confounders maydistort that contribution in a larger context,partly because it can be carried to a point atwhich it effectively eliminates lead exposure asa variable (40). Given all we know about lead,might it be time to proceed to larger ques-tions? Should the old debates be put to rest?Should we frame another kind of questionand adopt another perspective? Should neuro-toxicity be viewed as one element in a combi-nation of factors acting jointly as a form ofrisk vector? These questions are not novel.They have been posed by others and in otherforms, as in the dilemma described byBellinger and Stiles (41). They discussed theutility of different study populations in thecontext of their own investigations as follows:

Other investigators chose to recruit theirsamples from inner-city populations at farhigher risk of developmental handicap dueto a variety of poverty-related risk factors,of which lead is only one. The trade-offsinvolved in this decision are that, on theone hand, it may be difficult to reject thepossibility that any apparent lead-relateddecrement in performance is not the resultof residual confounding (a potential TypeI error). On the other hand, it may be dif-ficult to estimate, with adequate accuracyand precision, any true lead-related decre-ment amidst the "statistical noise" con-tributed by correlated risk factors (apotential Type II error) ...

Community attributes, beyond the formof covariates, are one of the ways in whichcustomary measures of neurobehavioral func-tion may distort the total effect of toxic expo-sure. How the influence of lead on IQ scoresis evaluated provides one example. In hismeta-analysis of lead and IQ, Schwartz (42)calculated that a 10-g/dL rise in blood leadwould produce a mean estimated loss of 1.85IQ points in a socially disadvantaged popula-tion and a mean loss of 2.89 IQ points in asocially advantaged population. Such a dis-crepancy might be used to argue that reduc-tions in lead exposure might not be ascost-effective in disadvantaged populations.IQ Loss in Advantaged andDisadvantaged CommunitiesIf the two populations were to be compared bybroader criteria, the message might be differ-ent. Figure 4 compares the effects of shifts inmean IQ in two different communities, one

Environmental Health Perspectives * Vol 108, Supplement 3 * June 2000

I378

VULNERABILITY OF CHILDREN TO NEUROTOXIC HAZARDS

that would be ranked as advantaged and one asdisadvantaged. In many surveys, the differ-ences in mean IQ scores of such populationsapproximate about 15 points (43). Assume,then, for modeling purposes, initial IQ distrib-utions with respective means of 100 and 85,both with standard deviations of 15. As animpact index, calculate the number of scoresbelow 70. Conventional education standardstend to assume (despite the rise in scores overseveral decades known as the Flynn effect) thata score below 70 indicates the need for reme-dial measures, and for some school systems,signifies a classification of retarded.

With population sizes of 100,000 each, asshown in Figure 4, a loss of 1 IQ point in theadvantaged population increases the numberof individuals below 70 from 2,280 to 2,660.In the disadvantaged population, the lossassigns 17,530 rather than 15,870 individualsto the below-70 category. Although the pro-portional shift is greater in the advantagedpopulation (16.7%) than in the disadvantagedpopulation (10.5%), the number of individu-als added to the developmentally disabled cate-gory is much larger in the disadvantagedpopulation (1,660) than in the advantagedpopulation (380). The discrepancies enlargewith greater IQ losses, which could result fromhigher neurotoxicant exposures. One result isthat expenditures arising from increaseddemands for remedial education in the disad-vantaged community would also greatly exceedthose in the advantaged community.

Figure 5 shows one perspective on themultitude of stressors acting on disadvan-taged communities. It depicts how an array ofchallenges to optimal development, none ofwhich by itself exerts a large effect, mightcombine to create a pronounced depressionof functional potential as measured by IQscore. For this conceptual model, the sizes ofthe entries and their overlaps do not representactual data. The chart is designed only tounderscore the point that individual risk fac-tors, none of which alone might exert con-spicuous influence, can jointly effect markedchanges in developmental potential. Jointactions might take many forms; additivity isone possibility, synergism is another.

Degraded intellectual potential is only onefacet ofhow environmental stressors can inter-fere with neurobehavioral development. TheRochester Longitudinal Study (43) followed acohort of children, beginning at birth, toascertain the influence of maternal psy-chopathology, primarily schizophrenia, ondevelopmental outcome. They included otherpotential risk factors, as well, in their analyses(Figure 6). The unexpected message fromtheir data led the authors to condude that thenumber of risk factors, rather than any specificfactor, determined outcomes such as IQ andsocial competence. Figure 6 depicts effects on

measures of IQ and social-emotional compe-tence at 4 years of age. Individually, each riskfactor contributes a small amount of the totalvariance averaging about 4%, about equiva-lent to a 10-pig/dL increment on blood leadlevel. Their cumulative effect, as shown inFigure 6, can be dramatic.

Sameroff et al. (43) did not include neu-rotoxicant exposure as a risk factor, althoughsome portion of their study populationinhabited housing with high levels of lead inpaint and dust (20). The trends depicted inFigure 6 indicate that neurotoxicant exposuremight well exert a greater effect in combina-tion with other risk factors of the type listedin Figures 5 and 6 than in isolation. Figure 4shows that the home and community envi-ronment exert an enormous influence on theoutcomes of neurotoxic exposures.A second method to illustrate the way in

which different stressors, including toxicexposures, might influence development isshown in Figure 7. The uppermost curvedepicts the normal trajectory of development.Add exposure to a neurotoxic substance anddevelopment proceeds at a lowered rate and

0

cv 25,000 -

~2,00 Moon of 8520000

15,000

3:iomw0CD

500.0

= 0 1 2 3Z IQ loss (points)Figure 4. Advantaged communities typically show abouta 15-point or 15% higher mean IQ score as compared todisadvantaged communities. If both populations, as aresult of neurotoxic exposure, suffer an equivalentdecrease in mean IQ, the effect will be greater in thedisadvantaged community, as gauged by the number ofIQ scores below 70. A score below 70 is taken as evi-dence of retardation and, in many school districts,requires remedial education.

Pescide exposureOther neurooic agents

Povrt incomInadequate schools

Poor nufflhnTeen pregnancy

Poor prenatel careMetmaltobacco, drugsLowmatemal educaton

Violent neighborhood

in%._

Figure 5. Schematic model to show how individual com-ponents of a stressful environment might cumulate toreduce performance on IQ and other tests. The individualstressors are shown as overlapping to suggest a lack ofindependence, and their length is meant to indicate thatno single component is overwhelming in isolation.

reaches a lower asymptote. Add a secondstressor, low income (which, in practice,would embody a collection of stresses), andrate and asymptote may fall even further.

Economic Implications ofDevelopmental NeurotoxicityImpaired child development imposeseconomic burdens on the entire society, notsolely on the individual. Behavioral difficultiesexpressed in higher rates of crime and delin-quency comprise one set of outcomes (44) butare difficult to quantify in economic terms.

AcO 1.9-, 121

2.3-

o 2.5-E

2.7-

2.9-

0 1 2 3 4 5 6 7

B120 -115-110105-

100-95-90_85 -80-

0 1 2 3 4 5 6 7

Number of risk factors

Figure 6. Charts modified from a report on theRochester Longitudinal Study (43). Two end points areshown for children tested at 4 years of age: (A) a ratingof social-emotional competence and (B) verbal IQ on theWechsler Preschool and Primary Scale of Intelligence(WPPSI). Each individual risk factor accounted for about4% (4 IQ points). The difference between the lowest andhighest risk children was approximately 30 IQ points or 2SDs. Risk factors are maternal mental illness, highmaternal anxiety, rigid parental perspectives, few inter-actions with infant, minimal maternal education,unskilled occupation, disadvantaged minority, singleparenthood, stressful life events, and large family size.

0 1 2 3 4Developmental stage

Figure 7. Schematic to show how the normal course ofdevelopment (intellectual attainment) might be slowedby the addition of toxic exposures, poverty, and otherrisk factors. SES, socioeconomic status.

Environmental Health Perspectives * Vol 108, Supplement 3 * June 2000 379

B. WEISS

Lowered IQ, as measured by standardizedtests, provides the best documented relation-ship with economic criteria such as earnings.

Much of the public, and even some scien-tists, are confused about what IQ scores mea-sure. IQ tests are not used to gauge someabstract quality called intelligence, but,instead, constitute a sample of performancewith established predictive power. For exam-ple, IQ scores are correlated with social class,educational attainment, and income. The lat-ter relationship facilitates one form of eco-nomic analysis of the costs arising fromenvironmental neurotoxicants. Such an analy-sis should not be interpreted as a claim thatIQ determines earnings rather than as adescription of an empirical relationship.

The relationship ofIQ to income providesa gauge by which to determine how muchsociety loses by neglecting environmental con-tamination. Salkever (45) has assembled themost convincing calculations. They includerelationships among IQ and years of schoolingand also take account of the cost of lostincome while in school and the discountedvalue of future earnings. By Salkever's calcula-tions, in terms of 1995 dollars, each IQ pointis worth $8,346 in lifetime earnings. Althoughthis sum is equivalent to only $200/year overa working lifetime, it has to be viewed from apopulation perspective. If environmental con-tamination diminishes IQ in the U.S. popula-tion by an average of 1%, the annual costwould come to $50 billion and the lifetimecosts to trillions. This sum is an underestimatebecause a shift of the population mean to alower IQ means more individuals thrust into ascore category below 70, which typicallymandates remedial education at a cost ofapproximately $6,000 annually. Salkever's cal-culation, if it were to indude this expenditure,might raise the loss to the community of oneIQ point to as high as $10,000. Restrictionson environmental pollution cost far less. In itsreport to Congress on the Costs and Benefitsof the Clean Air Act (46), EPA investigatorsestimated the direct costs of implementing theact at $523 billion in 1990 dollars. Its esti-mate of benefits during the same period,mostly from the avoidance of adverse healtheffects, ranged from $5.6 to $49.4 trillion,with a mean of $22.2 trillion. Based on thesecalculations, the ratio of benefits to costsranges from 10.7 to 42. Those who opposerestrictions on the environmental discharge ofchemical contaminants typically cite only theestimated costs.

The tobacco settlement offers a com-pelling example of how ignoring IQ lossesseriously underestimated the ultimate healthcosts of smoking. The economic implicationsof developmental impairment due to mater-nal smoking are based on a comprehensivereview of the relationship between IQ and

earnings (45) and the effects on IQ of maternalsmoking during pregnancy (47). The impli-cations are as follows:* four million children born in 1994* 20% of their mothers smoked during

pregnancy (800,000 children)* offspring of smoking mothers average a

3% (3-point) drop in IQ* assume 1 IQ point = $ 10,000 over a

working lifetime ($250/year)* multiplying 800,000 children by 3 =

2,400,000 lost IQ points* for the 1994 cohort, total loss = $24

million* look back over 30 years (1964-1994)

with the same smoking rates* for 30 years, total earnings loss would be

$720 million (for a 30-year period, thecumulative loss of earnings greatly exceedsthe $200 billion proposed tobacco settle-ment negotiated in November 1998 by28 of the State Attorneys General.These calculations are based only on

projected income. Distributed across the off-spring of smoking mothers, their sum is morethan twice the size of the tobacco settlement.It does not include the costs of other behav-ioral disturbances associated with maternalsmoking such as conduct disorders (47). Itdisregards cancer, emphysema, heart disease,asthma, premature rupture of membranes inpregnancy, low birthweight, and other costlyconsequences of tobacco use.

Earnings are far from the sole index ofhow reductions in population IQ impinge onsociety. Herrnstein and Murray (48) pro-posed the thesis that psychometric intelli-gence, as measured by IQ, determines anindividual's position and success in society. Itfurther stoked a fierce debate that continuesunabated about the heritability of IQ and therole of race. The empirical data upon whichthe Herrnstein and Murray (48) analyses relycome from the National Longitudinal Surveyof Youth (49), a federal program operated bythe Department of Labor to gather informa-tion at multiple points in time on the labormarket experiences of five selected groups.These data, independent of any particularstance on genetic contributions to IQ, aredepicted in Figure 8. They show that a rela-tively small shift in IQ can produce substan-tial effects on a number of social indices.

Herrnstein and Murray (48) nevermentioned the contribution of environmentalcontaminants such as lead to IQ scores. Forthe environmental health sciences, the ironiccounterpoint is its implied message that deter-mined efforts to reduce exposures to lead,PCBs, and other developmental neurotoxi-cants have inadvertently promoted social ben-efits beyond our original expectations basedon improved health. We can also inspect themirror image of those data to calculate the

massive costs of failure to act on behalf ofenvironmental protection for children.

We have only begun to grasp the breadthof challenges that chemical contamination ofthe environment poses to the developingbrain. Although we finally grasped the extentof hazards arising from lead exposure, webecame complacent about modern chemicals,such as pesticides, because their contributionsto our lives were striking and transparent.Their shortcomings proved elusive becausethey were subtle enough not to be immedi-ately apparent; moreover, we simply did notunderstand how to ask the proper questionsabout their toxicity. Recall the struggle todocument the dangers of smoking. Or theastonishment provoked by Carson's SilentSpring (50) and the attacks on hypothesesadvanced in Our Stolen Future (51). We cer-tainly will discover new threats unforeseen byour current knowledge that likely will takethe form of an unraveling puzzle.A provocative statement about lead from

the report by the Centers for Disease Control(52) could serve as a guide. If the referencesto lead are left blank, the paragraph belowcan be viewed as a terse description of howour understanding of threats to neuro-behavioral development tends to materialize.

[...] is ubiquitous in the human environ-ment as a result of industrialization. It hasno known physiologic value. Children areparticularly susceptible to [...] toxiceffects. [...] poisoning, for the most part,is silent: most poisoned children have nosymptoms. The vast majority of cases,therefore, go undiagnosed and untreated.[...] poisoning is widespread. No socio-economic group, geographic area, or racialor ethnic population is spared.

Neurobehavioral toxicology is one of thosedisciplines inextricably entangled with largerissues and public policy. A superficially directindex of developmental neurotoxicity such asIQ score evokes broader questions such as howit positions an individual in society, forinstance, and in the career and educational

Poverty during first 3 years of lifeBottom decile Home scores

Out-of-wedlock birthsLow-weight births

Wesfare recipiencyChildren wifthout parents 25

High school dropoutsMales interviewed in jail

Poverty rate

0 5 10 15 20 25 30Reduction (%)

Figure 8. Societal benefits (i.e., reduction in certainundesirable social factors) (48) of a 3% rise in 10, basedon National Longitudinal Survey of Youth data (49). TheHome score was based on the Home Observation forMeasurement of the Environment Inventory (53).

Environmental Health Perspectives * Vol 108, Supplement 3 * June 2000380

VULNERABILITY OF CHILDREN TO NEUROTOXIC HAZARDS

opportunities it affords and in how it deter-mines social class. Even more broadly, it awak-ens questions about the structure of society,about its economic potential, and even aboutthe processes of aging. The title of this paperencompasses only the first step in our explo-ration of children's health and its implications.

REFERENCES AND NOTES

1. Streissguth A. Fetal Alcohol Syndrome. Baltimore:Brookes, 1997.2. U.S. Food and Drug Administration Center for Drug Evaluation

and Research. Time line: Chronology of Drug Regulation in theUnited States. Available: http://www.fda.gov/cder/about/his-tory/timel.htm [cited 10 April 2000].

3. Needleman Hl. Childhood lead poisoning. Curr Opin Neurol7:187-190(1994).

4. Weiss B, Clarkson TW. Mercury toxicity in children. In:Chemical and Radiation Hazards to Children. Columbus,OH:Ross laboratories, 1982;52-58.

5. Ulbrich B, Palmer AK. Neurobehavioral aspects of developmen-tal toxicity testing. Environ Health Perspect 104(suppl2):407-412 (1996).

6. Center for Food Safety and Applied Nutrition. Food Color Facts.Rockville, MD:Food and Drug Administration, 1993.

7. Weiss B, Williams JH, Margen S, Abrams B, Caan B, Citron lJ,Cox C, McKibben J, Ogar D, Schultz S. Behavioral responses toartificial food colors. Science 207:1487-1489 (1980).

8. Swanson JM, Kinsbourne M. Food dyes impair performance ofhyperactive children on a laboratory learning test. Science207:1485-1487 (1980).

9. Kaplan BJ, McNicol J, Conte RA, Moghadam HK. Dietaryreplacement in preschool-aged hyperactive boys. Pediatrics83:7-17 (1989).

10. Weiss B. Food additive safety evaluation: the link to behavioraldisorders in children. In: Advances in Clinical Child Psychology(Lahey BB, Kazdin AE, eds). New York:Plenum, 1984;221-251.

11. Lipton MA, Nemeroff CB, Mailman RB. Hyperkinesis and foodadditives. In: Nutrition and the Brain (Wurtman RJ, WurtmanJJ, eds). New York:Raven Press, 1974;1-27.

12. Code of Federal Regulations. 21 CFR 70.40, 1 April 1999.13. Congressional Research Service. Report for Congress. Pesticide

Legislation: Food Quality Protection Act of 1996. Washington,DC:Congressional Research Service, 1998.

14. National Research Council. Pesticides in the Diets of Infantsand Children. Washington, DC:National Academy Press, 1993.

15. Gray LE Jr, Wolf C, Lambright C, Mann P, Price M, Cooper RL,Ostby J. Administration of potentially antiandrogenic pesticides(procymidone, linuron, iprodione, chlozolinate, pp'-DDE, andketoconazole) and toxic substances (dibutyl- and diethylhexylphthalate, PCB 169, and ethane dimethane sulphonate) duringsexual differentiation produces diverse profiles of reproductivemalformations in the male rat. Toxicol Ind Health 15:94-118(1999).

16. Eskenazi B, Bradman A, Castorina R. Exposures of children to

organophosphate pesticides and their potential adversehealth effects. Environ Health Perspect 107(suppl 3):409-419(1999).

17. Loewenherz C, Fenske RA, Simcox NJ, Bellamy G, Kalman D.Biological monitoring of organophosphorus pesticide exposureamong children of agricultural workers in central WashingtonState. Environ Health Perspect 105:1344-1353 (19971.

18. Gladen BC, Sandier DP, Zahm SH, Kamel F, Rowland AS,Alavanja MC. Exposure opportunities of families of farmer pes-ticide applicators. Am J Ind Med 34:581-587 (1998).

19. Landrigan PJ, Claudio L, Markowitz SB, Berkowitz GS, BrennerBL, Romero H, Wetmur JG, Matte TD, Gore AC, Godbold JH, etal. Pesticides and inner-city children: exposures, risks, and pre-vention. Environ Health Perspect 107(suppl 3):431-437 (1999).

20. Lanphear BP, Matte TD, Rogers J, Clickner RP, Dietz B,Bornschein RL, Succop P, Mahaffey KR, Dixon S, Galke W, et al.The contribution of lead-contaminated house dust and residen-tial soil to children's blood lead levels. A pooled analysis of 12epidemiologic studies. Environ Res 79:51-68 (1998).

21. Calabrese EJ, Stanek EJ, James RC, Roberts SM. Soil inges-tion: a concern for acute toxicity in children. Environ HealthPerspect 105:1354-1358 (1997).

22. Olson DK, Sax L, Gunderson P, Sioris L. Pesticide poisoning sur-veillance through regional poison control centers. Am J PublicHealth 81:750-753 (1991).

23. Ferguson JA, Sellar C, McGuigan MA. Predictors of pesticidepoisoning. Can J Public Health 82:157-161 (1991).

24. Angle CR, Mcintyre MS, Meile RL. Neurologic sequelae of poi-soning in children. J Pediatr 73:531-539 (1968).

25. Weiss B. Pesticides as a source of developmental disabilities.Ment Retard Develop Disabil 3:246-256(1997).

26. Guillette EA, Meza MM, Aquilar MG, Soto AD, Garcia IE. Ananthropological approach to the evaluation of preschool chil-dren exposed to pesticides in Mexico. Environ Health Perspect106:347-353 (1998).

27. Beach JR, Spurgeon A, Stephens R, Heafield T, Calvert IA, LevyLS, Harrington JM. Abnormalities on neurological examinationamong sheep farmers exposed to organophosphorous pesti-cides. Occup Environ Med 3:520-525 (1996).

28. Stephens R, Spurgeon A, Calvert IA, Beach J, Levy LS, Berry H,Harrington JM. Neuropsychological effects of long-term expo-sure to organophosphates in sheep dip. Lancet 345:1135-1139(1995).

29. Stokes L, Stark A, Marshall E, Narang A. Neurotoxicity amongpesticide applicators exposed to organophosphates. OccupEnviron Med 52:648-653 (1995)

30. Savage EP, Keefe TJ, Mounce LM, Heaton RK, Lewis JA, BurcarPJ. Chronic neurological sequelae of acute organophosphatepesticide poisoning. Arch Environ Health 43:38-45 (1988).

31. Rosenstock L, Keifer M, Daniell WE, McConnell R, Claypoole K.Chronic central nervous system effects of acute organophos-phate pesticide intoxication. The Pesticide Health Effects StudyGroup. Lancet 338:223-227 (1991).

32. McConnell R, Keifer M, Rosenstock L. Elevated quantitativevibrotactile threshold among workers previously poisoned withmethamidophos and other organophosphate pesticides. Am JInd Med 25:325-334 (1994).

33. Muller-Mohnssen H. Chronic sequelae and irreversible injuries

following acute pyrethroid intoxication. Toxicol Lett107:161-176 (1999).

34. Weiss B, Reuhi K. Delayed neurotoxicity: a silent toxicity. In:Principles of Neurotoxicology (Chang LW, ed). NewYork:Dekker, 1994; 765-784.

35. Calne DB, Eisen A, McGeer E, Spencer P. Alzheimer's disease,Parkinson's disease, and motoneurone disease: abiotrophicinteraction between ageing and environment? Lancet2(8515):1067-1070 (1986).

36. Gorell JM, Johnson CC, Rybicki BA, Peterson EL, Richardson RJ.The risk of Parkinson's disease with exposure to pesticides, farm-ing, well water, and rural living. Neurology 50:1346-1350(1998).

37. Tanner CM, Ottman R, Goldman SM, Ellenberg J, Chan P,Mayeux R, Langston JW. Parkinson disease in twins: an etio-logic study. JAMA 281:341-6 (1999).

38. McGeer PL, Itagaki S, Akiyama H, McGeer EG. Rate of celldeath in Parkinsonism indicates active neuropathologicalprocess. Ann Neurol 24:574-576 (1988).

39. Kessler 11. Epidemiologic studies of Parkinson's disease. 3: Acommunity-based survey. Am J Epidemiol 96:242-254(1972).

40. Needleman HL. The future challenge of lead toxicity. EnvironHealth Perspect 89:85-89 (1990).

41. Bellinger DC, Stiles KM. Epidemiologic approaches to assessingthe developmental toxicity of lead. Neurotoxicology 14:151-160(1993).

42. Schwartz J. Low-level lead exposure and children's ID: a meta-analysis and search for a threshold. Environ Res 65:42-55(1994).

43. Sameroff AJ, Seifer R, Bartko WT. Environmental perspectives onadaptation during childhood and adolescence. In: DevelopmentalPsychopathology. Perspectives on Adjustment, Risk, and Disorder(Luthar SS; Burack JA; Cicchetti D, Weisz JR, eds). Cambridge,UK:Cambridge University Press; 1997; 507-526.

44. Needleman HL, Riess JA, Tobin MJ, Biesecker GE, GreenhouseJB. Bone lead levels and delinquent behavior. JAMA275:363-369 (1996).

45. Salkever D. Updated estimates of earnings benefits fromreduced exposure of children to environmental lead. EnvironRes 70:1-6 (1995).

46. U.S. EPA. Final Report to Congress on Benefits and Costs of theClean Air Act, 1970 to 1990. EAP 410-R-97-002. Washington,DC:U.S. Environmental Protection Agency, 1997.

47. Olds D. Tobacco exposure and impaired development: a reviewof the evidence. Ment Retard Develop Disab Res Rev3:257-269 (1997).

48. Herrnstein RJ, Murray C. The Bell Curve. New York:Free Press,1994.

49. Center for Human Resource Research. National LongitudinalSurveys. NLS Bibliography. Available: http://www.chrr.ohio-state.edu/nis-bib/bib-home.htm (cited 10 April 2000).

50. Carson R. Silent Spring. Boston:Houghton Mifflin, 1962.51. Colborn T, Dumanoski D, Myers JP. Our Stolen Future. New

York:Dutton, 1996.52. CDC. Preventing Lead Poisoning in Young Children. Atlanta,

GA:Centers for Disease Control, 1991.53. Caldwell B, Bradley R. The Home Manual. Center for Child

Development and Education. Little Rock, AR:University ofArkansas at Little Rock, 1979.

Environmental Health Perspectives * Vol 108, Supplement 3 * June 2000 381