Document VxqokgwzRgnd8Z5YzVZaB1Lw

DownloadSearchable PDFRandom document
SOCIETY OF TOXICOLOGY ABSTRACT FORM See format for heading (other side). Leave double space between heading and abstract. Abstract must be single spaced. There will be a charge of *5 for any abstract which needs to be retyped. Leave no margins -- type must be within rectangle. Single space, elite type, except for double space between body of text and title and authors. Deadline for receipt in Executive Secretary's office is October 1, 19B4. Follow instructions accompanying this form. Submit original and two copies. Original must be on this form. Proof carefully. EFFECTS OF LEAD 0N COGNITIVE FUNCTION OF BLACK CHILDREN FROM LOW SOCIOECONOMIC STATUS (SES) FAMILIES: A REPLICATION. B.A. Hawk, S.R. Schroeder, S. Robinson, D.A. Otto, P. Mushak and C.N. Barton. University of North Carol ina,Chapel Hill, NC and U.S. Environmental Protection Agency. Sponsor: R Dyer. Children from low SES minority families are con- sidered to be at risk for lead poisoning. Deere-; ments in cognitive function ^as indexed by tradi- j ItionaVTQ tests-have been reported ^luseveraT^^ i studies of children with low to moderate lead ex-; ; posure,although'these' results -areiver^-contro-' | versial. An independent replication of a "pre- ] vious study (Schroeder et al. Environ, Res., in \ press) of the interaction of lead and social ! , factors on Stanford-Binet IQ.was performed on 80, low SES Black children screened 'by-county health : departments in North Carolina. Children's mean ! ! blood lead (PbB) level was 20.85 yg/dl (range, { 6.3-47.4). Multivariate regression analyses * showed no interactions between PbB and age, sex, maternal IQ, Caldwell home environment score or SES (Hollingshead Two-Factor Index). There was a highly significant negative relationship of both mean and maximum PbB levels on IQ (P<.002). That is, IQ decreased linearly as PbB increased. Tests for quadratic and cubic components were not significant. These results replicate our , previous findings and are consistent with recent studies of Pb effects on cognitive function in children from low-income families. Name and address of author to contact: Stephen R. Schroeder 136 DDDL/BSRD 220-H University of North Carolina Chapel Hill, NC 27514 Telephone --iiijl-966-5171 ___________ Indicate whether you prefer a platform or poster session. Platform ...... Poster -- Either Select one topic category from EACH of three lists belowwhich you think best describes your work. Type of Study -Environmental Type of Toxicant -Metal Type ofTarget -Central Nervous System Other--------------------------- ------------ Please check box if you would be willing to serve as session chairman D Signature of Sponsor TYPE OF STUDY Analytical Aquatic Clinical Developmental/Reproduction Environmental General Toxicology Evaluation Hypersensitivity inhalation In Vitro Mechanism Metaboiism/Disposition Mutagenesis Oncogenesis Reactive Intermediates Regulatory /?ev. 6/64 TYPE OF TOXICANT Air Poliutants/Gases Drug/Cosmetic Food Additives Halogenaied Hydrocarbon Metal Natural Products Pesticides Solvents/Vapors Other TYPE OF TARGET Cardiovascular Cellular Central Nervous System Dermal Hematological Hepatic or Gastrointestinal Immune Molecular/Genetic Ocular Peripheral Nervous System Pulmonary Renal Reproductive TEH 0350271 N36713 STANFORD-BINET IQ SCORE <oD o10 0350272 *3er 13.01 DUP0502989U*? / Separating the Effects of Lead and Social Factors on IQ by Stephen R. Schroeder2, Barbara Hawfc2t David A* Otto-*, Paul Mushak2 and Robert E. Hicks^ University of North Carolina2 and U.S. Environmental Protection Agency^ ^Paper-presented at the Second International Conference on Prospective Lead Studies. Cincinnati, 0. April, 1984. ORIGINAL FILE COPY DO NOT REMOVE ECAO/TIC TEH 0350273 N36713.02 DUP050298910 Separating Lead and Social Factors - Schroeder 2 ABSTRACT Initial evaluations of 104 lower socioeconomic class Black children screened fay the local community health departments In North Carolina showed significant effects of lead in the 6-59 ug/dl range on IQ after controlling for concomitant social factors, such as socioeconomic status, hone environment and maternal IQ. The main concomitant variable was socioeconomic status, which was multicollinear with other social factors. Five years later, when all blood lead levels were 30 ug/dl or less, lead effects on IQ were no longer significant. The correlation between maternal and child IQ which had been suppressed initially in children with higher lead levels, returned to expected levels when decreases in blood lead level occurred, while concomitant variables remained stable over the five-year peroid. Suggested Running Mead: Separating Lead and Social Factors Address Editorial Correspondence to: Stephen R. Schroeder 136 DDDL/BSRC 220H University of North Carolina School of Medicine Chapel Hill, N.C. 27514 3S02?4 DUP05029891 1 . Separating Lead and Social Factors - Schroeder 3 Children, with incompletely developed nervous systems, are considered a special risk group-for the neurobehavioral effects of lead. Extreme levels (blood lead level 100 ug/dl) are associated with lead encephalopathy, which begins with dullness, restlessness. Irritability, headaches, and muscular tremor and proceeds to convulsions, paralysis, coma, and frequently, death. Even when chelation therapy is appropriately and quickly employed, death still occurs in 20-40% of encephalopathy cases (Ennis and Harrison, 1950; Agerty, 1952; National Academy of Sciences, 1972). In fact, chelation therapy may even hasten death when the severity of symptoms and blood lead level are underestimated. High lead levels that remain asymptomatic in terms of acute encephalopathy have been clearly associated with intellectual deficits which are frequently long-lasting (Byers and Lord, 1943; Chisolm, 1968; Perlstein and Attala, 1966). Given that blood lead levels in the 80 ug/dl and greater range are generally proven to be extremely harmful, the question arises as to what effect lower levels such as 3Q - 80 ug/dl may have. Answers to this question are inconsistent and contradictory. Many studies have shown that moderate lead exposure has significant effects on neurobehavioral (including cognitive) functioning (EPA, in press). Other studies report no' such relation. Tremendous controversy surrounds these studies regarding selection bias, sensitivity of instruments, indices of exposure, control of confounding covariates, blind evaluation, and appropriate statistics. As one might expect, this research is difficult to do from the standpoint of sample selection and experimentally rigorous control. Many studies fail to take into account such factors as socioeconomic status (SES), maternal intelligence and quality of the care-giving environment. Others have examined older children several years after lead exposure, increasing the probability that intervening events have either exacerbated or minimized the effects of lead exposure. 03S02?5 DU PO 50298912 Separating Lead and Social Factors - Schroeder 4 Many studies fail on a crucial statistical control procedure:--i .e., to test interaction effects of lead and various control variables upon the various dependent measures. For example, consider the interaction of parental SES, lead, and IQ. Most studies have statistically "controlled" SES and tested the lead effect on IQ. Suppose, in actuality, that the relationship of lead to IQ differs as a function of SES. This is'not an unreasonable question, since SES has been shown to interact with other environmental insults such as low birth weight. Such an effect would be missed if appropriate interactions were not tested. Related to the first point, analysis of covariance is probably not appropriate for these studies because it is likely that regression coefficients are not equal across groups. The various interaction vectors provide a powerful test for homogeneity of regression. Another general criticism deals with a cormon failure to assess possible dissociative aspects of lead intoxication. Such an assessment is easily accomplished by intercorrelating all control and dependent measures separately in the several lead groups: e.g., consider Parental SES, Parental IQ, and Child IQ. Suppose that in the low lead group Parental SES and Parental IQ are highly related to Proband IQ. These relationships should be somewhat weaker in the high lead group. If this does not occur, it would leaa one to suspect either (a) parent-chita covariance on lead intoxication, or (b) mean difference (on IQ) as a function of lead are due to confounding with the various familiar variables, or (c) both possibilities. Restriction of subjects to high risk groups, such as blacks, severely restricts generalizability of a study. Consider the relation between child IQ and parental SES in blacks and whites. For example, the relationship between child IQ and parental SES is much stronger among whites than among blacks. DUP050298913 , 'Separating Lead and Social Factors - Schroeder 5 Therefore, IQ may be less correlated with lead intoxication in blacks than in whites. If both variablesare included, an interaction vector will test for the above possibility. Order of entry of predictor variables In a regression analysis may be crucial in determining the result with nonorthogonal predictors* For example, consider three interrelated variables: early CNS dysfunction, age, and handedness. When age (at time of testing) is partialed out of handedness, CNS dysfunction continues to predict left-handedness. However, when CNS dysfunction is partialed out of handedness, age at time of testing has zero relationship to handedness. Thus, obtaining an age effect depends on the ordering of variables in the equation. This ordering should always be guided by theory, so as not to capitalize on chance; but different orderings on the same data are generally useful when attempting causal modeling. I. Initial Screening Study Our initial research project was designed to investigate possible subtle cognitive deficits resulting from undue lead exposure and the relationship between lead and hyperactivity. The present study is concerned with the relationship among various factors and IQ. Specifically, the covariates considered important in children include parental socioeconomic status {Needleman, et a'l., 1979), maternal IQ (Perino and Ernhart, 1974) pica (Barltrop, 1966), home environment (Milar, Schroeder, Mushak, and Boone, 1981), and age at exposure. These covariates are believed to be particularly important when considering "threshold effects" of lead. Thus, covariates may interact to pose a cumulative risk which may have differential weight given different lead levels (EPA, in press). This study represents only one small TEH 0350277 DU PO 502 98914 ' Separating Lead and Social Factors - Schroeder 6 aspect of a larger research project and of the comprehensive evaluation which the children received, but It is one of the main topics of current research controversy* METHOD Subjects 104 children in Uake County, M.C. aged ten months to six and one half years at the time of initial testings served as subjects. They were mostly black (94%) from lower social classes--i,e., x=4.5 on the Hollingshead Two-Factor Index (Uollingshead and Redlich, 1953). Children previously evidencing CHS disease or insult and children with prediagnosed language delay or mental retardation were excluded from data analysis. Approximately half of the children were less than 30 months of age and half were over 30 months of age. Half of the subjects were EPSDT children and half were children of battery factory workers referred from a local county health department to the Lead Screening Program at North Carolina Memorial Hospital as part of a statewide screening program. Procedure The order of evaluation was the same for all children. At the time of testing, the examiner was blind as to the child's lead exposure. The first evaluation consisted of an intellectual assessment using the Bayley Scales of Mental Development (Bayley, 1969) for children less than 30 months of age and the Stanford-Binet Intelligence Scale (Terman and Merrill, 1960) for children 30 months of age or older. The second evaluation was a measure of free field activity (Routh, Schroeder, and D'Tuama, 1974). We have reported these results elsewhere and will not discuss hyperactivity here (Milar, Schroeder, Mushalc, and Boone, 1981). TElH 03s0278 DUP050298915 Separating Lead and Social Factors - Schroeder 7 While the childl$ intellectual evaluation was taking place, the primary caregiver (in most cases the mother) was interviewed by the patient coordinator. The following areas were covered: 1) family composition, 2) previous and present residence of child, 3) symptoms relating to lead poisoning, 4) socioeconomic background of parents including education and employment (Hoi'1 ingshead and Redlich, 1958), 5) Werry-Weiss-Peters Activity Scale (Kerry, l%a), t) Conner's rating scale (Conners, 1969), 7) estimate of parental intelligence using the Quick Test (Ammons and Ammons; 1962). Each child was also seen by a pediatrician for a complete medical evaluation. The purpose of this evaluation was to screen for medical problems not related to lead that could contribute to delayed development. The last event of the day was a venous blood sample for lead level and other biochemical determinations. Following the evaluation, a home visit was made to the primary residence of each child. During the visit, the mother was interviewed and the Home Observation for Measurement of the Environment (HOME) Inventory developed by Caldwell was completed. This inventory is an overall indicator of the quality of the caregiving environment. Bradley and Caldwell (1976) have demonstrated a significant relationship between HOME score and intellectual performance and language delay (Vlulbert, Inglis, Kriegsmann and Mills, 19-75), In particular, the subscales dealing with emotional and verbal responsivity of the mother and maternal Involvement with the child showed the highest relationship. Two different forms of the HOME inventory were used depending on the age of the child. For children under 30 months of age the 45 Item scale designed for younger children was used. For children over 30 months of age the 80 item HOME inventory designed for children 3 to 6 years of age was used. All data were collected at a time when both mother and child were present. At the time TEH 0350279 DUP050298916 Separating Lead and Social Factors - Schroeder 8 of assessment the interviewer was blind as to the possible lead exposure of the children. During the hone visit X-ray flourimeter readings were taken on painted surfaces to evaluate possible lead hazard from paint. Dust samples were also obtained from the floor to evaluate the possible contribution of lead in the house dust to lead hazard. Results have been'reported eslewhere (Milar & Mushak, 1982). The analytical^involved exploratory analysis of the data as part of a larger project in which several other hypotheses were also tested (Otto et al. 1989). The plan employed a form of hierarchical regression accomplished with backward stepwise regression (Kleinbaum, Kupper, & Morqanstern, 1982). The initial full model included the following sets of predictors: (1) linear, quadratic, and cubic components of lead; (2) concomitant variables comprised of the Caldwell H.O.H.E. score, maternal I(], the child's chronological aoe, sex, socioeconomic status of parents, test (Bayley vs. Stanford-Ginet), presence of the father in the hone, number of siblings, and (3) the omnibus interaction tern between the first three components of lead (linear, quadratic and cubic) and all concomitant variables. Only linear components of the concomitant variables were used because many studies have reported the lack of curvilinear relation ships with IQ (e.g.a Willerman, 1979). The three sources were backstepped out of the full model in reverse order, i.e., (3) first and (1) last. A significant lead x concomitant variable interaction would lead to decomposition of the onnibus component. INSERT TABLE 1 ABOUT HERE The source summary table is presented in Table 1. Only the socioeconomic status and the linear component of lead are TEH 0350280 DUP050298917 . Separating Leaid and Social Factors - Schroeder. g significant. The concomitant variables had been removed from the equation simultaneously, and the only concomitant variable to reach statistical significance Is socioeconomic status: regression coefficient * 0.375, F (1,63) * 7,496, p < 0.01. The regression coefficient for lead is -0.199. The failure of the omnibus interaction term to reach significance obviated the necessity for analyzing this source any further. The zero-order (linear) relationships between IQ and lead and IQ and socioeconomic status are presented in Figures 1 and 2 respectively. INSERT FIGURES 1&2 ABOUT HERE The failure of concomitant variables other than socioeconomic status to reach significance is probably due to multi col linearity. The multiple correlation (Rj of this variable with all other concomitant variables = 0.677, The intercorrelation matrix of concomitant variables, lead, and IQ is presented in Table 2. INSERT TABLE 2 ABOUT HERE II. Five-Year Follow-Up Study Approximately five and a half years later we retested these children to see whether there were any residual effects of early exposure that might show up at school age. Because all the children have grown out of the high exposure risk toddler stage, we expected all blood lead levels to be low. Method Subjects. Of the 104 children originally seen in 1977-78, 80 were located TEH 0350281 DUP050298918 Separating Lead and Social Factors - Schroeder 10 through the County Health Department and school records* Each family received a letter inviting them to participate in the study and a telephone call from the nurse coordinator of the Lead Screening Clinic explaining the study. The parents of 50 children agreed to participate. Procedures. Children were seen at the local county health department in a mobile laboratory where they received a physical examination, an extensive family history using the Fantus Clinic Lead Poisoning Questionnaire, and a Stanford-Binet IQ test. Socioeconomic Status (Hollingshead Two-Factor Index) and maternal IQ (Ammons & Ammons Quick Test) were repeated, but a home visit and the Caldwell H.O.M.E. Inventory were not. Electrophysiological tests were also made and are treated in a separated paper (Otto et al., 1935). Results and Discussion Fig 7. shows that all of the children's blood lead levels five years later were equal to or below 30 ug/dl. A similar regression model was used for these data as that used in the initial study, except that "Test" was no longer a variable since all children took the Stanford-Binet IQ test this time. "Age" was also dropped due tc its ubiquitous lack of relationship with other variables. The original lead level was used as the independent variable of interest (tested last in the model) with five-year lead level serving as a concomitant variable* The two lead scores are highly correlated (Fig. 3), as are the original and follow-up IQ scores (Fig. 4), Interestingly, the simple correlation between follow-up lead and follow-up IQ (Fig. 5) is of the same magnitude as the correlation between original lead and follow-up IQ (Fig. 6). INSERT FIGS 3,4,5,6 ABOUT HERE 0350282 DUP050298919 Separating Lead and Social Factors - Schroeder n In the regression analysis, however, neither original lead nor follow-up lead (when concomitant variables are included in the regression model) predicted follow-up IQs (F<1 in both cases). The only significant effect was for the concomitant variable term; F (7,35) = 3.614, p <.01. none of the individual predictors was significant. Apparently, the decline in lead levels, while not appreciably disturbing the structure of individual differences of lead on IQ, was sufficient to negate its relationship to IQ when confounding factors were controlled. Nor were INSERT TABLE 3 ABOUT HERE these results due to sampling bias at follow-up. Table 3 shows that the demographic characteristics of those who returned for follow-up and those who did not were very similar. Maternal IQ was also stable from the 1978-1983 samples (83 vs. 82) as was socioeconomic status (63.0 vs 64.8). Another indirect test of an effect of lead on child IQ might be to examine its effect on the relationship between maternal and child IQ (Perino and Ernhart, 1974). According to the polygenetic model of hereditability intelligence (Bouchard and McGue, 1981) one would expect the correlation betwet?n maternal IQ and child IQ to be about 0.50. Lead might disrupt this relationship, as has been found by Bellinger and Needleman (1983). A similar result was found in the present studies. In the initial study the correlation between maternal and child IQ for the children with blood lead levels of 6 to 30 ug/dl it was 0.528; for children with blood lead levels of 31-56 yg/dl it was 0.058. At five-year follow-up, when all blood lead levels were at or below 30 pg/dl, the correlation was back up to 0.45. From Fig. 7 it can be seen that about half of these children had had blood lead levels above 30 ug/dl and half below that level in 1977-78. TEH 0350283 DU P050298920 .Separating Lead and Social Factors - Schroeder 12 Summary It appears that, in a high risk population which had shown effects of lead on IQ, a reduction in lead levels resulted in a reduction of this effect five years later. This disrupted relationship between maternal and child IQ returned to near the expected level once decreases in blood lead level occurred. Furthermore, it appeared that the effect of the major covariates was relatively homogeneous over time. Finally, controlling for SES appears to capture much of the variance related to other covariates due to nulticollinearity--e.g., care giver practices, maternal IQ, number of siblings, and exposure history. This would suggest that, for a given population, if one controlled for the major covariates--i.e., SFS, caregiver environment, maternal IQ, age and exposure history,--then other related covariates would be considerably less important in terms of the degree to which failure to control for then would distort the results. TEH 0350284 DU P050298921 jc^aiatmy ceaa ana iociai factors - Schroeder 13 A9erty, H.W. (1955). Lead poisoning in children. Hed. Clin, of N. Amer. 36, 1587-1597. Ammons, R.9. A Ammons* C.H. (1962). The Quick Test: Provisional Manual. Psych. Rep. 11, 111-161. Barltrop, D. (1966). The prevalence of pica. Amer. J. Pis. Child. 112, 116-123. Bayley, N. (1969). Bayley Scales of Infant Development: Birth to 2 Years. Psychological Corporation, New York. Bellinger, D.C* and Needleman, H.L. (1983). Lead and the relationship between maternal and child intelligence. J. Pediatr. 102, 523-527. Bouchard, T.O. and McGue, M. (1981). Familiar studies of intelligence: A review. Science, 212, 1055-1059. Bradley, R. and Caldwell, B. (1976). Early home environment and changes in mental test performance from 6 to 36 months. Dev. Psychol. 12, 93-97, (a) * Bradley, R. and Caldwell* B. (1976). The relation of infants' home environments to mental test performance at fifty-four months: a follow up study. Child Dev. 47, 1172-1174, (b). Byers, R.K, and Lord, E.E. (1943). Late effects of lead poisoning on mental development. Amer. J^. Pis. Child. 66, 471-494. Chisolm, J.J. (1968). The use of chelating agents in the treatment of acute and chronic lead poisoning in childhood. J. Pediatr. 73, 1-38. Conners, C.K. (1969). A teacher rating scale for use In drug studies with children. Amer. 0. Psychiatr. 126, 152-156. Ennis, J.M., and Harrison, H.E. Treatment of lead encephalopathy with BAL (2,3 diraercaptopropanol). Pediatrics, 1951), j>, 853-868. TEH 0350285 DUP050298922 Separating Lead and Social Factors - Schroeder 14 Environmental Protection Agency (1985). Air Quality for Lead. Chapter 12: Biological Effects of Lead. Washington, D.C. In press.' Hollingshead, A.d.B., and Redlich, F.C. (1958). Social Class arid Mental Ilness. Wiley, New York. Kleinbaum, D., Kupper, L., & Morgenstern, H. (1982). Epidemiologic Research: Principles and Quantitative Methods. London: Lifetime Learning Publications. Milar, C., S Mushak, P. (1982). Lead-contaminated housedust: Hazard, measurement, and decontamination. In 0. Chisholm and O'Hara 0. (Eds.). Paper presented at the conference o:i Management of Increased Lead Absorption in Children: Cl inical, Social, arid Environmental Aspects (Baltimore, Maryland: Urban Schwarzenburq). Milar, C.R., Schroeder, S.R., Mushak, P., Oolcourt, J.I., and Grant, L.O. {1980). Contributions of the care-giving environment to increased lead burden of children. Amer. J* Ment. Defic. 84, 339-344, Milar, C.R., Schroeder, S.R,, Mushak, P., and Boone, L. (1981). Failure to find hyperactivity in preschool children with moderately elevated lead burden. J. Med. Psychol. 6, 85-95, National Academy of Sciences (1972). Lead: Airborne Lead in Perspective. Washington, D.C, Needleman, H.L., Gunnoe, C., Leviton, A., Reed, R., Maher, C., and Barrett, P. (1979). Deficits in psychologic and classroom performance of children with elevated dentine lead levels. New Eng. 0_. Med. 300, 689-695. Otto, 0., Robinson, 6., Baumann, S., Schroeder, S., Kleinbaum, D., Barton, C., Mushak, P., and Boone, L. (1985). Five-Year follow-up study of low-to-moderate lead absorption: Electrophysiologlcal evaluation. Env. Res. In press. Perino, J., and Ernhart, C.B. (1974). The relation of subclinical lead level to cognitive and sensorimotor impairment in black preschoolers. 0. of Lrng. Pi sab, ]_* 26-30. TEH 0350286 DUP050298923 Separating Lead and Social Factors - Schroeder 15 Routh, D.K., Schroeder, C.S., and O'Tuama, L. (1974). Development of activity level in children. Dev. Psychol. 10, 163-168. Terman, L.M. ft Merrill, M.A. (1960). Stanford-Binet Intelligence Scale. Houghton Mifflin, Boston. Werry, O.S. (1968). Developmental hyperactivity. Red. Clin. H* Amer. 15,581599. Wulbert, M., Inglis, S., Kriegsmann, E., and Mills, R, (1975). Language delay and associated mother-child interactions. Dev. Psychol. 11, 61-70. WiHerman, L. (1979). Effects of families on intellectual development. Amer. - Psychol . 34, 923-929. TEH 0350287 DUP050298924 Separating Lead and Social Factors - Schroeder 16 ACKNOWLEDGEMENTS We wish to acknowledge NIEHS Grant #ES-Q1I04; USPKS Grant #HD-03110 to the Child Development Research Institute; MCH Project 916 to the Division for Disorders of Development and Learning, where the children were evaluated; Wake County Public Health Department* Helen Cannon, M.D., Director, where children were also screened. Thanks are also due to the Department of Pediatrics, North Carolina Memorial Hospital and the Health Effects Research Laboratories of the Environmental Protection Agency for their countless examples of support and collaboration in this project. TEH 0350288 DU PO50298925 Separating Lead and Social Factors - Schroeder 17 TABLE 1 Source Table of the Effects of Lead on Children's IQ (1977-1978 Initial Study) Lead (Linear) Socioeconomic Status Residual Sourcedf_ ' ss - 1 1530.549 1 4012.906 96 199.055 F 7.689 20.159 P <0.01 <0.001 TEH 0350289 DUP050298926 Separating lead and Social Factors - Schroeder 18 Table 2 Correlation Matrix of Blood Lead Level and Child IQ <19771 and Concomitant Variables ------------------------------------------------------------------ ------------------------- Variables Pb.B ia H.O.M.E. Mat IQ CA SEX SES FaPr Blood Lead Level (PbB) Child IQ Caldwell H.O.M.E. Maternal IQ (MatIQ) Chronological Age (CA) Sex Socioeconomic Status (S1ES) Father Present (FaPr) Number of Siblings (Sibs) -.276 -.269 .451 -.193 .379 -.068 -.254 -.072 .130 .183 -.449 -.010 .165 .314 -.284 .522 .089 .123 -.624 .433 -.202 .029 -.028 -.475 .178 -.177 .062 -.143 -.064 .161 .054 -.495 .116 -.003 .161 -.064 TEH 0350290 DU PO 502 98927 Separating Lead and Social Factors - Sctiroeder 19 SUMMARY TABLE 3 Initial (1978) Scores of Children on Key Variables Who Were or Were Not Followed Up Five Years Later Key Variables Mean Blood Lead Level Mean Child IQ Mean Caldwell HOME Mean Maternal IQ Mean Chronological Age (Mos.) Mean Socioeconomic Class Followed-Up Yes" (S.D.) No (S.D.) 28.7 (14.0) 30.5 (11.6) 88.0 (17.3) 87.8 (16.7) 47.5 (20.6) 43.9 (26.1) 80.5 (9.8) 80.3 (10.1) 37.0 (18.0) 35.1 (23.2) 4.6 (0.6) 4.5 (0.7) TEH 0350291 DU P0502 98928 .Separating Lead and Social Factors - Schroeder 2D FIGURE CAPTIONS Fig 1. Initial! IQ levels as a function of Initial blood lead levels: Data were collected 1977-1978* Fig. 2. Initial IQ levels as a function of initial socioeconomic status (197778) usiing the Hollingshead Two-Factor Index. Fig. 3. Five-year follow-up blood lead levels (1983) as a function of initial blood lead levels (1977-1978) for 50 of the original subjects. Fig 4. Five-year follow-up IQ levels (1983) as a function of initial IQ levels (1977-1978). Fig. 5. Five-year follow-up IQ levels (1983) as a function of five-year follow- up blood lead levels (1983). Fig. 6. Five-year follow-up IQ levels (1983) as a function of initil blood lead levels (1977-1978). Fig. 7. Frequency distribution comparing initial and five-year follow-up lead levels after lead abatement had occurred for 50 subjects. TEH 0350292 DU P050298929 13A31 01 IVIliN! =104 =Q077,F (1,102) =8.491 p< IN IT IA L BLOOD LEAD LEVEL ( /ig /d l) TEH 0350293 DU P050298930 N =104 r2=.l9l5,F( 1,102) =24.15 IQ. =121.39-0.52015 (S SOCIOECONOMIC STATUS DU P 0502 98 931 o <D IN ITIAL BLOOD LEAD LEVEL (ju g /d l) O O O bv CL rrrooo ci LO it CO u_ ro oto itno o If ii <Mw CD JO 0^ CO roo o + ro <7> a CD -Q 0- X oro OP/BrO S13A31 QV31 00019 dn-M01103 9V3A-3AI3 TEH 0350295 DU P05 029 8932 N=50 r2=0.4362, F (1,48) =37.054, p<0.0000l INITIAL IQ LEVEL o OrO O OJ O o ocn O co o rO co o m 13A31 01 dn-M0110d d V3A-3AI3 TEH 0350296 DUP050298933 FIVE-YEAR FOLLOW-UP BLOOD LEAD L E V E L to O d V in Cl 00 f -O CL 00 rcoo ii 03 OroJ co u_ cm C7) o in o 6 II ii 04 ro (T) ii m a oro in C\J _o CM _ in -O o CM o o o CD o CO R 33A31 01 dfl-MOTlOJ dV3A--3AI3 o CO TEH 0350297 DU P050298934 INITIAL BLOOD LEAD LEVEL (ju g /d l) 13A31 01 dn-M0T10J 8V3A-3A13 TEEhH ,,0350298 DU P050298935 s03rans 30 A0N3no333 TEH 0350299 DUP050298936 5-10 11-20 21-30 31-40 41-50 51-60 61-70 BLOOD LEAD LEVELS (yjg/dl)