Document wqGDzEoxwovpnN5jnrNYde6jV

eT UJ TEH 0532438 PO-598 csrnj$M& >802 E. I. d u Po n t d e Ne mo u r s & Co mp a n y INCORPORATED Pe t r o l e u m La b o r a t o r y Wil min g t o n , De l a w a r e 19bs b cc: D. R, Diggs W. E. B.ettoney A. J. Pahnke Snee May 31, 1979 To: D. H. From: E. s. "LONGITUDINAL CHANGES IN BLOOD LEAD LEVEL IN CHILDREN . AND THEIR RELATIONSHIP TO SEASON, AGE, AND EXPOSURE TO PAINT OR PLASTER" The attached article appeared in the April, 1979 issue of the American Journal of Public Health, I call it to your attention, because the article indicates there is little, if any, seasonal change in blood lead levels of children in New York City. In addition, the authors found the major variable in predicting blood lead level change were the child's age at screening and a history of exposure to paint or plaster. This may be helpful in responding to draft articles by Billick. ESJ/dmg Attach. BETTER THINGS FOR BETTER LIVING . . .THROUGH CHEMISTRY TEH 0532439 DUP050033705 Longitudinal Changes in Blood Lead Level in Children And Their Relationship to Season, Age, And Exposure to Paint or Plaster j a s k Mc Cu s k k k . MD. Dk PH Abstract: Children screened for lead poisoning in the Brownsville district of New York City in either summer or w inter were followed with blood lead tests for approximately six months to one year from screen ing to measure longitudinal changes in blood lead level and to identify some determinants of the changes. On ly minimal evidence was found of the hypothesized summer rise in blood lead level, while the predominant trend seemed to be for blood lead levels to display sta tistical regression to the mean. In children found to have low to intermediate blood lead levels 55 ng/ lUOnjl.) at screening, variables which were found to predict a rise in blood lead level of 10 fig/100ml or greater from winter to summer were under age three and/or exposure to paint or plaster. (Am. J. Public Health 69:34X-352. 1979.) introduction I.cad poisoning screening programs aim to prevent symptomatic and asymptomatic lead poisoning and its se quelae through the early detection of children with excessive lead absorption, followed by medical andior environmental intervention. The blood lead test has been in w idespread use as : screening test, but Its reliability and validity hate been questioned.1 More recently the free erythrocyte pro toporphyrin t FKP) lest, a measure of the toxic effects of lead on the synthesis of heme, has been advocated as ah alterna tive or supplement to the Wood lead test,1' - Reigart and Whitlock5 evaluated both tests longitudinally in a'group of Black children in Charleston, South Carolina and found greater .fluctuations in the blood lead than in the FKP level. In general, in children in whom the test results were dis cordant. ii,e.. high FKP,'.uid low blood lead level or vice Ver sa t, subsequent changes in Wood lead level tended to follow the direction suggested by the initial FKP level. Guidelines for screening programs have gradually re duced the blood lead level cut-oil which separates positive from negative test results.- 4 Increasing emphasis is placed on the periodic resereening of " high-risk" children, gcnc.ra.l- Address reprint requests to Or. Jane MgCusker. Clinical Assist ant Professor of.Preventive Medicine and Community Health. L'ni- sersity of Rochester School of Medicine and Dentistry. Rochester, NY IJ642. At the time of the-Muds. Dr. McCuskcr was with the Department of Kpidentioltigy. Columbia I'tmersity School of Public Health ansi Administrative Medicine. This paper, submitted to the Journal January ft. PTtt, was revised and accepted for publication August .21. I97K, ly agreed to be children under six years of age who live in or frequently visit deteriorating housing constructed prior to the IVfiOs. Children with pica, defined as the ingestion of non-food substances, are also considered to be at high risk.1 in addition to the sociodemographic risk factors men tioned above, more frequent screening has been advocated during the summer months when lead poisoning cases are said to be ntore likely to be diagnosed.-' Evidence supporting summer screening comes largely from cross-sectional data gathered from large-scale screening programs in Chicago and New York City.** * Smaller longitudinal studies have report ed conflicting results. An early study by Chisolm and Harri son of 32 children with onset of acute lead encephalopathy before two years of age showed that the vast majority of them experienced onset of symptoms during the summer months.7 Chisolm later reported on 15 children treated for acute lead-encephalopathy during one summer whom he fol lowed wjlh blood lead and urinary coproporphyrin vests over the subsequent year." Both tests -showed a steady decline, without any evidence of a seasonal effect, and Chisolm con cluded that season has little effect upon children not actively ingesting lead. Klein and Kchlageter' followed 31 children initially screened in the summer with persistently elevated blood lead levels (greater than 40 fig'100ml on three succes sive determinations J for at least 12 months. They found an overall tendency for blood lead levels to decline from a mean screening level of 51 fig/100 ml to a mean level of about 40 /up IIX) ml the following winter, followed by a slight rise to about 44 fig 100 ml the following summer. The Charleston children studied by Reigart and Whitlock5 were followed over a six-month period from winter to summer and no rise in Wood lead level wax detected. The authors suggest that 348 AJPH April 1979, Vol. 69. No. 4 TEH 0532440 DUP050033706 < LONGITUDINAL CHANGES IN BLOOD LEAD LEVEL IN CHILDREN this may have been due to the smaller average seasonal fluc tuations in temperature in Charleston in comparison with the northern cities where the previous studies had been con ducted. The objectives of this study were: 1. To determine whether seasonal fluctuations in blood iead level could be detected in cohorts of children in New York City followed longitudinally. The demonstration of such fluctuations might suggest that a given blood lead level detected during the winter months has a different signifi cance from the same lead level delected during the summer months. 2. To investigate the value of using certain sociodemographic variables and a history of pica to identify those children whose blood lead levels would rise over the ensuing six to 12 months. Subjects and Methods tests to obtain information on social and demographic vari ables. Mothers were asked whether their child ever put pa per, ashes, or dirt in his mouth, and whether the child had ever handled, mouthed, or ingested pieces of paint and plas ter. The mothers were reassured that various mouthing prac tices were quite normal in young children. Even with this precaution, this part of the history proved one of the most difficult and it was often hard to pin the mother down on details of the history. Comments such as "I've never caught him at it." or "1 don't date leave him alone in case he does it," were not uncommon, and neither was the mother who 'flatly denied that her child had ever mouthed foreign objects implying that she was top good a mother to let this happen. Because of the lack of reliability of this history it was de cided to define the variable "exposure to paint and piaster" to include a history of handling in addition to mouthing or ingesting one or both of these substances. A positive history indicates that exposure to paint or plaster was occurring at the time of the initial screening test. This study was carried out in the Brownsville -arca of Brooklyn. New York, where a lead poisoning screening pro gram is conducted by the Brownsville District of the New York City Health Department (BDHD) as part of the citywide screening program.10 At the time the study was con ducted, children whose blood lead levels were 55 f<gMOO mi and over were designated "cases" and followed with envi ronmental and/or medical intervention. Initial screening blood lead levels were divided Into three groups: low (0 to 34 jt&'lOO ml), intermediate (35-54 Mg/100 mil. and high (55 Mg'' 100 ml and overt. Five cohorts of children comprised the study population. Three winter co horts (A, B. and C with low, intermediate, and high blood lead levels respectively) who were first screened between December I, 1970 and February 28, 1171 were followed to the summer of 1971, Two summer cohorts (D and E with low and intermediate blood lead levels respectively) who were initially screened in the summer of 1970 were followed to the summer of 1971.' Screening and follow-up blood lead levels were mea sured on a 5 ml senous blood sample at the New York City Health Department Laboratories by atomic absorption spectophotometry.** Interviews with the child's parent or guardian, using a standard pre-coded questionnaire., took place at BDHD at the time of the first and second follow-up Cohort A was a 10 percent random sample (n = 6f>) while co horts Bln -- UOlandCln * 38) were 100 percent samples. Of the 38 members of cohort C, 13 received chelation therapy during the study period, and only the remaining 25 are included in the results. Cohorts l> and I: were sampled-systematically from BDHD files of UuMicn initially Screened between July I and September.3IJ, 1970 lo give a total or 30 in each cohort. Study specimens were not specially marked but were sent along with routine screening and follow-up blood specimens. The temporal stability of the blood lead level (which includes laboratory variance t over the range of lead levels encountered in this study is expressed as an intraindividual yai lance of approximately 25 Mg' KXI ml (standard deviation of 5 MgTbO mb. furtherdetails on methods of estimation of laboratory and total intraindividua) variance may be obtained from the author. Results The ethnic composition of the study population was al most entirely Black (70 per cent) or Spanisb-American (29 per cent). Seventy-three per cent of the children came from families on welfare, One-ftfih of all mothers had received on ly an elementary education: only 36 per cent were high school graduates. The age range of the children was from 10 months to 12 years with approximately one-third of the sample 0-2 years of age, one-third 3-4 years of age. and onethird 5 years of age or older. The age distribut ions of the five cohorts did not differ significantly. Repeal blood lead test results 6 months after the screen were obtained on 116/211 members of cohorts A, B, and C (55 per cent) and 30/60 members of cohorts D and E (50 per cent). Failure to follow-up members of the cohorts was most marked in the low blood lead cohorts (A and D). This is likely to be due to lack of motivation to return as the parents were informed that the screening result was in the normal range. In addition, the latter families were more likely to have moved away from their address at screening. Further investigation into characteristics of the children not fol lowed, which included contacting a sample of the non-re spondents who had not moved, indicated that non-respond ents were similar to respondents with regard to age, sext race, and other demographic characteristics, but tended to be less mobile, having lived .in New York City for longer than respondents. Seasonal Pattern of Blood Lead levels Table 1 and Figure I contain data on seasonal changes in blood lead levels. The percentage of children in each cohort whose blood lead levels rose over the half-year following screening is greatest in the winter cohorts A and B. A com parison of the pe rcentages with a Wood lead level rise in the two low blood lead level cohorts A and D yields a X? of 13.656 (p < .01). A comparison of the same percentages for cohorts B and E yields a X; of 3.080 <p > .05). Analysis of ' AJPH April 1979. Vol. 69. No. 4 349 TEH 0532441 DUP050033707 MCCUSKER TABLE 1--Blood Lead Level Changes in Children Tested Both in Summer and Winter Cohort n Percentage with Increase m Blood Lead Level Mean Winter Blood lead Level Ug/lOOml) Mean Summer Blood Lead Level (mQ/100ml) Mean Change in Blood Lead Level (fignoomt) 95 Per Cent Confidence Interval (jig/IOOml) ' Winter Cohorts (Winter to Summer Transition) i A 3(2 75 24,5 29.9 h 5.4 2.0 < n < 8.8 I: B 70 C' 14 39 0 41.5 58.7 39.3 46.1 - 2.2 - 12.6 -- 4.7 < n < 0.3 - 17 6 < fi < -- 7.0 I | Summer Cohorts (Summer lo Winter Transition) D 10 10 27.1 23.5 - 3.6 - 7.8 < < 0.6 I E 20 15 41,2 33.9 - 7.3 11.4 < n < - 3.2 i 'Untreated members ol cohort C only I 1 it 8 mean blood lead level changes by cohort indicates a statisti Although the numbers are small, and the changes not statisti cally significant rise for cohort A and a fail for cohorts C and cally significant."cohorts Dand Edid show small fluctuations <r E Ip < .051. Exclusion from the analysis of children aged five and in the expected directions. When children with a history of paint or plaster ex over (approximately one-third of the sample) did not affect posure were excluded from the analysis there was no signifi the mean direction of blood lead level change by cohort hut cant change in the blood lead level patterns already de did affect the -size of the change by increasing a mean rise or scribed. It therefore seemed unlikely Unit a seasonal pattern lessening a mean fall in blood lead level. was being obscured by children currently ingesting lead or Subsets of cohorts D and E were retested the foliowing whose lead levels had been elevated previously. The pos summer: their mean winter and summer blood lead levels sible effects of various interventions including Public Health were as follows: Nurse home visits. Sanitarian home visits and inspections, n wean u hiter wean mmmer environmental intervention (home repair!, and study inter views were examined in a similar way. but the exclusion of .bloodle,ul level tag/100 ml) hi uni lead level tMg'TOOm!) children in whom these interventions had been applied did not significantly alter the direction of previously described blood lead level changes. cohort D cohort E 8 14 24.0 33.9 24.8 Statistical regression to thq mean seemed to be the pre 35.1 dominant determinant of blood lead level change. Analysis by 10 m&'IOO ml groupings demonstrated regression to the 35-44 /xg'lOO ml range: children with screening levels below this range showed a mean increase and those with screening levels above this range showed a mean decrease, regardless 4j of season of testing. I 5f6t Prediction oflilood Lead Level Change At the time of screening. 70 per cent of the children t were living in or frequently visited housing described by the parent or guardian as poor, i.e., old with peeling paint arid/or broken plaster (no home visits were made to confirm this verbal report). A further 16 per cent had previously lived in poor housing. A significantly larger percentage of cohort C (high blood lead levels) currently or previously lived in poor housing (94 per cent) in comparison with cohort 0 (88 per cent) and cohort A (72 per cent). The prevalence of mouthing of foreign substances in the total sample tin whom the history was known) was 34 per cent for paper. 12 per vent for ashes, and 4 per cent for dirt only C- i;c i.*-) ! FIGl'KK I--Mean Blood Level Changes in fig'IDD ml in Five Cohorts (95*1 confidence intervals are shun n in parentheses). while 17 per cent were exposed to paint or plaster. More than one of the above characteristics was reported in 39 per cent of the children. Not surprisingly, exposure to paint or 350 AJPH April 1979, Vol. 69, No. 4 TEH 0532442 DUP050033708 LONGITUDINAL CHANGES IN BLOOD LEAD LEVEL IN CHILOREN plaster was found only in children living in poor housing. It was interesting that mouthing of other substances ipaper. dirt, ashes) was also more common in this group, being re ported in 39 per cent of those liv ing in poor housing as com puted to 17 pet cent of children living in good housing. litis relationship held even after agc-adjustmcnt. To examine the role of various sociodemographic vari ables in predicting Wood lead level change, those children first screened in the winter but not designated eases were examined (cohorts A and B). (Three members of cohort B were excluded because they were not Interviewed, leaving 67 children in cohort B and 32 children in cohort A.) Of these 99 children. 14 had blood lead levels which rose 10 teg' 100 nil or more by the following summer, seven from cohort A and seven from cohort B,* Only two of the sociodemographic and pica variables used in the study showed any clear and consistent relationship to .blood lead level change: age tinder three at screening, and a history of exposure to paint or plas ter. Table 2 shows the sensitivity and specificity of using these variables alone or in combination to identify those chil dren whose blood lead levels rose 10 p.gM00 ml or more by the following summer. Discussion In discussing the results of this study, it :is first appropri ate to consider the extent to which the results may be gener alized. The reference population tor this study was the popu lation of children screened for lead poisoning in the Browns ville district of New York City, This population was in many ways typical of one at "high-risk" for lead .poisoning, being an urban ghetto population of Black and Spanish-Amcrican children of predominantly welfare families, living in deterio rating housing with all its attendant problems. Although the non-response rate was high, non-respondents appeared to differ from respondents only in mobility which wits not found to be related to Wood lead level changes. It therefore seems unlikely that selection Was could account for the results of this study. The sample, however, is a small one. Although subtle evidence was found of a seasonal pat tern of Moot! lead levels, season did not exert as great an effect as had appeared front a review of the literature. Data from the Chicago' and New York City1' lead poisoning screening programs indicated a clear seasonal pattern in the percentage of high blood lead levels detected, with a mini mum in the winter and a maximum in the summer. The dif ference between mean winter and summer blood lead levels in the Chicago study was 15 /agrlOO ml. These data were cross-sectional rather than longitudinal and it is possible that factors such as increased screening of high-risk arcus during the summer months might account in part for the results. The usual explanation for the summer increase in blood lead level is increased lead absorption due to vitamin D as a response to increased exposure to sunlight."' i; More recent work by Sorrell and associates" suggests that, regardless of blood lead level and season, serum 25-OHD (the most re liable indicator of an individual's vitamin D status) is primar ily a function of vitamin D intake and that high blood levels tend to he.associated with a relative vitamin D deficiency, rather than an excess. The relationship between lead and vi tamin 13 is clearly complex and may also involve interference with the biogenesis of the hormonal form of the vitamin (I, 25-(OHL) from 25-0HD ** Another explanation for a ,sea sonal increase in blood lead levelsis increased ingestion and/ or inhalation of dust-borne lead during the summer months," In this study, the hypothesized seasonal effects, if in deed they exist, appear to be of lesser importance than the phenomenon of regression to the mean.15 The roles of vari ous types ofintervention (medical, environmental, and inter view effect) in causing this decline in blood lead level were evaluated, and no evidence for a relationship was found (al though sample sizes were small). There was also no sugges tion that the seasonal pattern might have been obscured by TABLE 2--Age* and Pleat as Predictors of Blood Lead Level Change** Predictor Rise :10;ignv 100ml \f\ - 14) with Predictor (a) without Pie.diclor (ti) No Rise (n = 85) wth Predictor to without Predictor tdl Per Cent Sensitivity ithri Age < 3 6 6 12 73 57 Pica 10 4 19 66 71 Age 3 or pica"* 13 1 20 59 93 'Age at screening $Hstoiy of handling, mouthing, of ingesting paint or plaster at time ol screening 'Data from all members of cohorts A and 0 as tsied n Table 1, except three children whose pica history was unknown "4? 19SZ0. p < 001 Per Cent Specificity !cd- > 86 70 69 ** The cut-off of Itlj/gi 100 ml rise in blood lead level was chosen as being twice the standard deviation for intraindividual variation. **Rosen j; Personal communication to the author. AdPH Aonl 1979, Vol. 69, No. 4 351 TEH 0532443 DUP050033709 MCCUSKER the presence of children who were either currently ingesting lead or who had previously suffered high blood lead levels. Two other explanations should he considered: the over all decline in blood lead level could have been caused by the educational effects of the screening process itself, or there could base been an inconsistency overtime in the laboratory analyses. A change in the laboratory analytic results over the period of the study seems an unlikely explanation as stan dard specimens were used daily as reliability checks. The educational effect was a factor shared .by all study members and it could have produced the observed effects by causing tin acceleration in the normal aging process whereby a child stops mouthing foreign objects. ' The vanables which were found to be of major impor tance in predicting subsequent blood lead level change were the child's age at screening and a history of exposure to paint or plaster. Other authors hate described the important rela tionship of age to increased lead absorption*'- and so it was not surprising that age turned out to be an important variable in predicting subsequent blood lead level change: children under age three were at greatest risk of showing an increase in blood lead level over the sis months following screening. It was surprising, however, that there was not a greater cor relation of age with actual blood lead level: the mean and range of age at screening were very similar in each of the cohorts. Furthermore, a recent history of exposure-to paint or -plaster was present to the same extent in both the older and younger members of cohort C. Although the numbers involved are rather small, it appears that children over three y ears old are more likely to have high blood lead levels than had previously been thought. A history of pica for paint for plaster has been used as a screening device in New York City in the past1' but has largely been abandoned as too unreliable and of low sensitiv ity. The results front the present study suggest that about one-third of the children with high blood lead levels (Cohort C) would have been missed if a history of exposure to paint for plaster (recent or past) had been used alone as a screen ing test. The results of this study suggest that selective follow-up of children whose screening blood load levels arc in the low to intermediate range <<55 Mg 100 ml) and who at screening are cither under age three or whose parents respond affirm atively to the question. "Does your child ever handle pieces of paint or plaster or put them in his mouth?", may be a sensitive and fairly specific means of identifying those ehil- dren whose blood lead levels will rise over the ensuing months. This approach could be tested in a variety of screen ing programs to determine its value. REFERENCES 1. Chisolm D: Is lead poisoning still a problem? Clin Chcm23:252- 255, 1977, 2. Increased Lead Absorption and Lead Poisoning in Young Chil dren, A statement by the Center for Disease Control. U.S. De partment of HEW. Public Health Service. 1975. 5. Rcigart JR and Whitlock NH: Longitudinal observations on the relationship between free erythrocyte prophyrius and whole blood lead. Pediatries 57:54-<9, 1976. 4. Sleinfeld JL: Medical aspects of childhood lead poisoning. HSMHA Health Rep 86: 140-143. 1971. 5. Hlanksma L. Sachs HK. Murray HF and O Connell MJ: In cidence of high blood load levels in Chicago children. Pediatrics 44:661-667, 1969. 6. Guinee VF: Epidemiologic studies of lead .exposure in New York City. Int. Symposium on Environmental Health Aspects of Lend. Amsterdam. Oct. 2-6. 1972. 7. Chisolm JJ and Harrison HE; The exposure of children to lead. Pediatrics IX:943_958. 1956. 8. Chisolm JJ-, Discussion of paper by Baetjcr. A. Ind Med Surg 28:140-142. 1959, 9. Klein MC anJ Schlageter M: Non-treatment of screened chil dren.with intermediate blood lead levels. Pediatrics 56:298-302, 1975. 1(1. Guinee VF: Lead poisoning. Am J Med 52:283-288. 1972. II. Sobel A. Gawron <) and Kramer B; Influence of vitamin D in experimental lead poisoning. Proe Soc Exp ffiul Med 38:433434, 193:8. 12. Rapaport M and Rubin M: Lead poisoning: A clinical and exper imental sluily of the factors influencing the seasonal incidence in children. Am J Diseases of Children 61:245-255, 1941. 13. Sorrell M, Rosen 3F and Roginsky 6). Interactions of lead, cal cium. vitamin D. and nutrition in lead-burdened children. Arch Environ Health 32:160-164. 1977. 14. Sayre JW, Charney E, Vostal J and Pless 18: House and hand dust as a potential source of childhood lead exposure. Am J Dis Child 127:167-170. |974. 15. Dav is CK: The .effect of regression to the mean in epidemiologic and clinical studies. AmcrJ Epidemiol 104:493--198, 1976. 16. Jacotvincr H: Leadj'oisoning mchildhood; epidemiology, man ifestations and prevention. Clin Pcdiat 5:277-286, 1966. 17. Greenberg M, Jaeob/incr H. McLaughlin M, ct al: A study of pica in relation to lead poisoning. Pediatrics 22:756-760. 1958. ACKNOWLEDGMENT This research was supported by The Mental Retardation Train ing Program Grant No, 5-'101-MO-00322. Part of this paper was pre sented at the Society for Epidemiology Research meeting June 1977 in Seattle. Washington. a*?- 352 AJPH April 1979. Vol. 69. No. 4 TEH 0532444 DUP050033710