- Email: [email protected]
Relationship between Blood Lead Concentrations and Learning Achievement among Primary School Children in Taiwan
Environmental Research Section A 89, 12}18 (2002) doi:10.1006/enrs.2002.4342, available online at http://www.idealibrary.com on
Relationship between Blood Lead Concentrations and Learning Achievement among Primary School Children in Taiwan1 Chao-Ling Wang,* - Hung-Yi Chuang,? A Chi-Kung Ho,- ? Chun-Yuh Yang,A Jin-Lian Tsai,Ting-Shan Wu, B and Trong-Neng Wu- *Department of Family Medicine, Yuan’s General Hospital; -Graduate Institute of Occupational Safety and Health, Kaohsiung Medical University; ?Department of Occupational Medicine, Koahsiung Medical University Hospital; ASchool of Public Health, College of Health Science, Kaohsiung Medical University; Institute of Medicine, Kaohsiung Medical University; and BThe Koahsiung City Government’s Health Department Received July 6, 2001
ger inBuence on childern’s language ability (Chinese) than on their ability to calculate (Mathematics). Our results suggest that environmental lead exposure adversely affects a child’s academic achievement, making a direct link between exposure to lead and academic attainment. 2002 Elsevier
Over the past 20 years lead has been proven to exert an inBuence on the intelligence of children. Especially for children exposed to environmental lead, average blood lead was often lower than the ofAcially recognized intoxication level. Because Kaohsiung is an industrial area in Taiwan and lead exposure is an important environmental issue, we attempted to ascertain the extent to which environmental lead inBuences the achievement of primary school children. We randomly selected 934 children from 32 primary schools in 11 districts of Kaohsiung City. Blood lead levels of the children were checked, and they were administered a questionnaire about their family information. Scores of several courses were used in this study on the relationship between a child’s blood lead and his or her academic performance (Ranking with his or her classmates), including Chinese (reading and writing short Chinese articles), Mathematics, History and Society, and Natural Science. Multiple regression models were done with adjustments for the confounding effects of their parents’ socioeconomic levels. The mean (SD) of 934 blood lead level was 5.50 (1.86) g/dL. Spearman’s coefAcient showed that class rankings in Chinese, Mathematics, Natural Science, and History and Society were all strongly associated with blood lead levels (P0.01). The multiple regression models revealed that blood lead level exerts a stron-
Science (USA)
Key Words: children; blood lead; learning achievement; school attainment.
INTRODUCTION
The rapid industrialization of the past three decades has resulted in prosperity and democratization in Taiwan; however, it has also led to environmental pollution. Lead is one of the leading sources of pollution in Taiwan. As mentioned in previous studies, in Taiwan about 10% of lead workers have high blood lead levels (Wu et al., 1998a,b). In addition, two previous studies found that children residing near a battery factory had lower intelligence than those residing in a control area (Rabinowitz et al., 1992; Wang et al., 1992). In those studies, the authors suggest that lead concentrations in the soil and air might have had a signi7cant effect on the children’s intelligence. Kaohsiung City, a southern port city and industrial center in Taiwan, is well known for its air and water pollution problems. Exhaust gas from motor vehicles is an area of great public concern. Lead from leaded gasoline and various industries has been the main source of lead pollution in this area. Although the phase out of leaded gasoline started before 1988 and by 2000 it was totally phased out in Taiwan, there are still concentrations of lead in waste and dust deposits.
1 This study was supported by the Department of Health, the Executive Yuan, Taiwan (Grant DOH89-TD-2106, NSC 892320-B-037-045), and Kaohsiung City Government Health Department. 2 To whom correspondence should be addressed at Graduate Institute of Occupational Safety and Health, Kaohsiung Medical University, 100 Shih-Chuang First Road, Kaohsiung City, Taiwan, 807. Fax: #886-7-3115948. E-mail: kmco6849@ms14. hinet.net.
12 0013-9351/02 $35.00 2002 Elsevier Science (USA) All rights reserved.
CHILDREN BLOOD LEAD AND LEARNING ACHIEVEMENT
The potential neurotoxicity of lead found in children all over the world has been explored in many recent epidemiological studies. These studies have gradually shifted from describing the general neurotoxicity observed in a relatively few individuals to studying more subtle impairments caused by this form of pollution in large numbers of children. Changes in the functional abilities of a signi7cant proportion of a population have potentially serious consequences for the affected individuals as well as for society (Banks et al., 1997). Because several previous studies have documented obvious neurotoxicity of lead and its in8uence on childhood development of intelligence using the intelligence quotient (IQ) as a measure, our aim was to focus on environmental lead in relation to academic achievement of children. The purpose of this study was to better understand the relationship between exposure to environmental lead and attainment in speci7c school subjects. MATERIALS AND METHODS
All 11 Kaohsiung City districts were covered in this study. Children were randomly chosen from 32 primary schools. At least one school per district was represented. At each school, one class of thirdgraders (out of a total of six grades) was randomly chosen for the study. Parental consent was obtained for the collection of blood from each child. Blood lead levels were determined and questionnaires were administered from December 1998 to March 1999. The blood samples were analyzed for whole blood lead levels in Kaohsiung Medical University Hospital. After pretreatment with cell lysis and digestion in a Class 100 hood to which air was supplied and cleaned by a high-ef7ciency particulate 7lter, the blood samples were analyzed by Zeeman effect graphite furnace atomic absorption spectrometry (GF-AAS, Perkin-Elmer 5100 PC with AS 71 autosampler), with intralaboratory quality controls. With use of commercial standard materials (Betherning Institute, Bio-Rad), all coef7cients of variation (CVs) were less than 3% for measurements at high levels (40.5}72.7 g/dL) and medium levels (22.4}45.3 g/dL), and were less than 5% for those at low levels (5.6}8.9 g/dL) during the 6 months of the study. Since 1993, the laboratory in Kaohsiung Medical University Hospital has been participating in the interlaboratory blood lead pro7ciency testing program of the Taiwan Department of Health. The results of our measurements are all within the reference ranges, indicating that our blood lead measurements are relatively accurate. At the same time,
13
children were required to answer the questionnaires regarding personal information such as gender, birth date, area of residence, health status, and the occupations, ages, educational levels of the parents under the supervision of teachers or school nurses. In the mean time, we collected the children’s semester grades for Chinese (reading and writing short Chinese articles), Mathematics, History and Society, and Natural Science. In order to prevent interference from different teacher’s grading criteria, we converted the children’s credit scores into class rankings. For example, a class rank of 1 in mathematics would mean that the child had the highest score in his or her math class. Therefore, a larger rank number means the achievement was poor. A small amount of score data was missing (0.6%) due to changes in residence and schools among the children. Parents were classi7ed into seven educational levels; illiterate, primary school, junior high school, senior high school, junior college, university, and higher, with coding numbers ranging from 0 to 6, 0 being illiterate and 6 the highest. The father was classi7ed into 7ve social}economic levels adjusted by Hollingshead’s Index of Social Position according to the father’s occupation and education. In brief, the Hollingshead’s index includes 7ve social}economic classes, the lowest in degree of professionalism and lowest educational level being classi7ed as a Class 5 and the highest in professionalism and the highest in educational level being classi7ed as Class 1. In Taiwan, because women usually resign from work after getting married or after giving birth, the classi7cation of maternal occupation was dropped in this study and the maternal education level was considered a confounding factor in the analysis. We used the SAS system to analyze the data and to study the relationship between blood lead and academic achievement in class. The distribution of children’s blood lead levels had a bell-shaped distribution, so the data needed no transformation. Pearson correlation and Spearman correlation coef7cients were used to determine the bivariate correlation. Continuous variables such as age were performed by Pearson correlation, and ordinal variables, such as ranks, were performed by Spearman correlation. In addition, multiple regression models with adjustments to control for the confounding factors of gender, father’s social economic status, and mother’s education level were also used to investigate the association of blood lead levels and these children’s academic achievement. The standardized residuals of these regression models were analyzed to test the assumptions.
14
WANG ET AL.
TABLE 2
RESULTS
In total, 934 children (454 girls and 480 boys) were involved in this study, which took place between December 1998 and March 1999. Their ages ranged from 8 to 12 years old (mean 8.9 years, SD"0.41). Most of the children in this study were around 8}9 years old. Some children had delayed their schedule due to moving from another country or to a health condition. Four children were older than 11 years old and six were over 10 years old. However, there were no mental or learning problems among these older
TABLE 1 Characteristics of All 934 Children and Their Parents’ Occupation and Education Levels Total (%) Number: Blood lead* (g/dL): Children’s age* (years): Father’s occupation Level 1 Level 2 Level 3 Level 4 Level 5
Girls (%)
Boys (%)
934 5.50 (1.86)*
454 (48.6) 5.30 (1.74)*
480 (51.4) 5.68 (1.96)*
8.85 (0.41)*
8.81 (0.34)*
8.90 (0.46)*
70 280 310 200 15
33 138 149 97 8
37 142 161 103 7
Father’s education (years) Primary school (1}6) 47 Junior high school (7}9) 204 Senior high school (9}12) 344 Junior college (10}15) 172 University (13}16) 108 Master ('16) 17
(8) (32) (35.4) (22.9) (1.7)
(5.3)
25 (5.7)
(8.2) (31.6) (35.8) (22.9) (1.6)
22 (4.9)
(22.9)
101 (23)
103 (22.8)
(36.8)
161 (36.6)
183 (40.5)
(18.4) (11.6) (1.8)
89 (20.2) 56 (12.7) 8 (1.8)
83 (18.4) 52 (11.5) 9 (2)
Father social economic status Class 1 83 (9.8) Class 2 152 (17.9) Class 3 305 (36.0) Class 4 219 (25.9) Class 5 88 (10.4) Mother’s education (years) Primary school (1}6) 70 Junior high school (7}9) 221 Senior high school (9}12) 423 Junior college (10}15) 106 University (13}16) 67 Master ('16) 2
(7.3) (30.4) (32.8) (21.4) (1.8)
42 73 149 108 46
(10.0) (17.5) (35.6) (25.8) (11)
41 79 156 111 42
(9.6) (16.5) (32.5) (23.1) (8.8)
(7.9)
38 (8.7)
32 (7.1)
(24.9)
101 (23.1)
120 (26.5)
(47.6)
202 (46.20)
221 (48.9)
(11.9) (7.5) (0.2)
56 (12.8) 40 (9.2) 0 (0)
50 (11.1) 27 (6.0) 2 (0.4)
Father’s age (years) 40.68 (4.19)* 40.45 (3.90)* 40.91 (4.45)* Mother’s age* (years) 37.79 (4.24)* 37.56 (4.06)* 38.01 (4.41)* Note. Asterisk indicates mean (SD).
Correlation CoefAcients of Blood Lead with Dependent Factors Factors Age of children Father’s age Father’s occupation Father’s education Father’s SES Mother’s age Mother’s occupation Mother’s education Class ranking Chinese Mathematics History Science
Number 925 870 871 888 843 875 894 885 925 925 924 924
Coef7cient 0.048* 0.017* !0.112 !0.236 0.166 0.009* !0.020 !0.179 0.128 0.103 0.126 0.099
P value 0.143 0.611 0.001 (0.001 (0.001 0.795 0.548 (0.001 (0.001 0.002 (0.001 0.003
Note. SES, social economic status. *Pearson coef7cients; others are Spearman coef7cients.
children. The mean ages (SD) of the parents were 40.7 (4.19) years for the fathers and 37.8 (4.24) years for the mothers. All study subjects lived within the limits of Kaohsiung City, and most of them lived near their school. The distributions found in socioeconomic status (SES) of the fathers and education levels of mothers can be seen in Table 1. The SES of the children’s fathers mostly fell within Class 3 and 4 in the Hollingshead’s index classi7cation, with no difference between the parents of boy and girl students. The mean (SD) blood}lead level of the 934 study subjects was 5.50 (1.89) g/dL. The levels ranged form 0.2 to 25.5 g/dL. By correlating blood lead levels and dependent factors, we found that class rankings in Chinese, Mathematics, Natural Science, and History and Society were all strongly associated with blood}lead levels (P0.01) (Table 2). The socioeconomic status of the father and the educational levels of the mother were found related to not only their children’s blood lead levels but also their academic achievement. Therefore, such family conditions were important confounders in the association of blood lead levels and academic achievement in children. In order to clarify this complicated relationship, we used multiple regression models to analyze the relation between blood lead levels and class rankings. The class rankings were found to have signi7cant correlations to blood lead after adjusting for father’s SES and maternal education levels (Table 3). The models revealed that children with higher blood lead levels had signi7cant lower academic achievement in Chinese, History and Society, Mathematics, and
15
CHILDREN BLOOD LEAD AND LEARNING ACHIEVEMENT
TABLE 3 Multiple Regression Models of Class Achievement and Predictors in Grade 3 Children, Kaohsiung City
Rank of Chinese
Lead Gender (girls vs boys) Father SES Mother education Intercept Adjusted R-square
Rank of History and Society
Rank of Mathematics
Rank of Natural Science
estimate
SE
estimate
SE
estimate
SE
estimate
SE
0.37 4.53
0.15* 0.57**
0.43 2.78
0.15* 0.56**
0.38 0.69
0.16* 0.58
0.32 1.24
0.16* 0.58*
0.93 !1.35 11.94
0.30** 0.34** 1.89
1.18 !0.97 10.53
0.29** 0.32** 1.84
0.89 !0.80 12.35
0.14
0.10
0.31** 0.34* 1.93 0.04
0.85 !1.34 13.96
0.31** 0.34** 1.92 0.07
Note. SE, standard error. *P(0.05. **P(0.01.
Natural Science, even after adjusting for other risk factors. The adjusted R-square values for the Chinese model was 0.14, which means that about 14% of the variance found in academic achievement in Chinese could be explained by our model. The girls had achievement scores in Chinese, History and Society, and Natural Science higher than those of the boys. Father’s SES and maternal education levels were still signi7cantly associated with children’s class achievement in all four subjects after adjusting for blood lead levels. Academic achievement in Mathematics was weakly associated with children’s blood lead levels. Blood lead level seemed to play a more important role in the children’s ability to learn language (Chinese) than in their ability to calculate (Mathematics). The plot of residuals for Chinese and Mathematics models both were randomly scattered about zero. The pattern of residuals suggests the presence of an outlier (blood lead"25.5 g/dL). But the outlier did not twist the regression line very much. Rewrite helps to ensure that the model’s predictions about the distribution of the independent variable were not violated. DISCUSSION
Our study compared blood lead levels with course score data on 934 school children. The male-to-female ratio was 1.05, a 7gure quite similar to the ratio found in a primary school population in Taiwan. The proportions of the parents’ educational levels were also similar, but higher than those found national-wide. For example, 19.1% of the parents have junior high school educations nation-wide, compared to the 22.9% found in our study. The
nation-wide percentages and the percentages found in our study continued to be similar with 35.4 vs 36.8% for senior high school, 14.9 vs 18.4% for junior college, and 11.4 vs 13.4% for university and higher (Statistics, 1999). Therefore, sampling of this study was similar to the normal distribution of Taiwan. Also, this study may not have selection bias because data on the dependent variables had not been collected at the time the sample was assembled. Variation in teacher grading was also controlled by random sampling of the classes and the course scores were transformed into class rankings in each subject. The results of our study support the hypothesis that children’s school achievements might be inversely related to their absorption of environmental lead. A signi7cant relationship between lead and achievement was found not only in our analyses using Pearson and Spearman correlation coef7cients, but also in multiple regression models after adjustment for gender, father’s social economic status, and mother’s education. Chinese language skills were more strongly related to blood lead level than was Mathematics. This could mean that lower level lead exposure may have more in8uence on the ability to memorize than on the ability calculate. The outlier was found in the data of one boy. His blood lead was 25.5 g/dL in the time course of the study. We visited with his parents, and they told us that he had taken herbal medicine for his health since childhood. In Taiwan, child lead poisoning resulting from Chinese herbal medicines is not uncommon (Cheng et al., 1998; Chu et al., 1998; Wu et al., 1994, 1997; Yu and Yeung, 1987). But the outlier did not twist the regression line very much. After the outlier was detected from the database, the
16
WANG ET AL.
regression models were similar to the previous one. The regression coef7cient was 0.33 for the Chinese language model and 0.36 for the Mathematics model. Due to the slight change and similar regression line, we did not delete the outlier in our presentation. But we did want to emphasize the important source of childhood lead exposure. Since the 1980s many epidemiological studies have con7rmed that low-level, subclinical lead exposure in early life can be associated with decrements in childern’s intelligence. One meta-analysis study reported that an increase of 1 g/dL in blood lead would cause a loss of 0.26 IQ point. The same study found a slope of 0.323 in studies on children with mean blood lead levels below 15 g/dL, and 0.232 in those with mean blood lead above 15 g/dL (Schwartz, 1994). The slopes in our models are similar to those in the meta-analysis, since most of our blood lead levels were lower than 15 g/dL. Nevin has also concluded that the IQ-to-blood lead slope may increase at lower blood lead levels (Nevin, 2000). In our study only 12 children (1.28%) had blood lead levels over 10 g/dL and only one boy had blood lead levels exceeding 15 g/dL. Most of the children in our study had levels below 10 g/dL. Therefore, much like the studies that demonstrated a relationship between lead and intelligence, our study found a signi7cant inverse relationship between lead and academic achievement. Moreover, our evidence associating blood lead levels with children’s class achievement was more direct than the evidence associating blood lead and IQ, and our 7ndings remained signi7cant even at blood lead levels lower than 10 g/dL. This means that the lead exposure can affect not only children’s intelligence, but also their personal achievement. The mean age of the children in our study was 8.85 years. Several other studies have reported the in8uence of lead to be obvious in children under the age of 10. Schwartz states that during the 7rst 3 years of life, when basic cognitive abilities develop, young children face the greatest risk of IQ losses due to lead exposure (Schwartz, 1994). In addition, exposure to environmental lead during the 7rst 7 years of life has been associated with cognitive de7cits that seem to persist into later childhood (Tong et al., 1996). Fergusson et al., conducted a study to measure dentine lead levels in children 6 to 8 years old and compared the 7gures with educational outcomes of the same children when they were 18 years old. They found that early mildly elevated dentine lead levels had a modest but detectable effect on measures of individual achievement such as reading ability and school examinations. They found that
these effects extended to late adolescence (Fergusson et al., 1997). However, this study was a cross-sectional study of blood lead and classroom attainment. We did not have the earlier blood lead data for the children in our study, but we showed that blood lead had a similar in8uence on learning achievement simply by studying the class rankings in four main subjects. In a comparison of industralized areas, one study showed that dentine lead levels of shed incisors for children around Taipei City were slightly higher at that time than levels reported in Boston, and the intelligence scores from Raven’s Colored Progressive Matrices Test were negatively correlated with dentine lead levels, especially among girls and children whose parents had lower education levels (Rabinowitz et al., 1991). Dentine lead is not easy to measure, and therefore blood lead levels have been widely used in Taiwan and world-wide. In 1995, 319 preschool children living in the Taipei metropolitan area including Taipei City, Taipei County, and Tao-Yuan County were tested. Their mean (SD) blood lead was 4.4 (2.4) g/dL (Cheng et al., 1998). In another study of children in a Tao-Yuan kindergarten (n"84), children’s mean blood lead level (SD) was 3.9 (2.1) g/dL (Chuang et al., 1996). Our study of children in Kaohsiung City, an industrial port area in southern Taiwan, reports a higher blood lead level, 5.5 (1.9) g/dL than those found in northern Taiwan. These 7ndings indicate that the local Kaohsiung government, elected of7cials, and private citizens should act to reduce lead exposure among the children in this city. Intelligence tests for children in primary schools were often based on three dimensions: verbal, mathematical, and performance IQ. However, there has been no consensus on whether lead exposure can be more strongly associated with any of these three aspects separately. In 1986, Lansdown reported a relationship between lead and children’s reading and spelling (Lansdown et al., 1986). The 7ndings of a study by Wasserman in 1997 demonstrated that perceptual}motor skills are signi7cantly more sensitive to lead exposure than the language-related aspects of intelligence (Wasserman et al., 1997). Other studies have revealed that environmental exposure to lead exerts a negative impact upon psychological functions not only in the intelligence of children but also in hand}eye coordination, perception, memory, reaction time and accuracy, language function, attention, classroom behavior, and behavioral disorders (Dudek and Merecz, 1997; Needleman, 1990b). Therefore, lead in8uences not only the general intelligence of children but also visual}
CHILDREN BLOOD LEAD AND LEARNING ACHIEVEMENT
spatial-dependent neuropsychological behavior and performance. Furthermore, the Chinese language has visual}spatial pathway’s that are different from those of nonideographic languages. There had been no previous study on this aspect of lead in8uences. Our study found that blood lead levels were related more strongly to language skills (Chinese) than to calculating skills (Mathematics). Furthermore, our study provides more direct evidence that lead exposure can affect children’s learning achievement, especially in terms of their language-learning abilities. The main 7nding of our report was that a clear relationship exists between blood lead and a child’s learning achievement. The class rankings have an important social impact because a large number of children are affected. We also found that SES of parents conveys information about a child’s potential lead exposure as well as other independent determinants of achievement. Table 2, for example, shows that certain variables, including father’s SES, mother’s educational level, and class rankings were all signi7cantly related to blood lead. Therefore, in our study of blood lead and children achievement, SES was an important confounding factor. Multiple regression model (Tables 3) show parameter estimates of lead to range from 0.32 to 0.43, meaning that when the blood lead increases to 1 g/dL, the children’s class rankings, may fall by around 0.3 point, after the confounders are held constant. Gender was also an important factor in the relationship between lead and performance. In most epidemiological studies of lead, the means of the female blood lead were often lower than those of the male. Our study revealed the same condition (Table 1). The girls’ performances in Chinese and in History and Society were signi7cantly better than those of the boys (Table 3). There was no difference between girls and boys in Mathematics performance. Therefore, the regression model result of better performance by girls was due not only to gender difference but also to the lead in8uence. For the past decade defects in intelligence and changes in behavior caused by low-level exposure to lead have been a world-wide issue. Lead exposure has been associated with hyperactivity and has been considered a risk factor for learning problems (Needleman, 1990a). In Taiwan, since the use of alkyl lead in gasoline additives was prohibited in 2000 government statistics have shown a decrease in air lead. Therefore, follow-up studies on the behavior of children with higher blood lead levels in Taiwan may be needed to clarify the extent to which lead in8uences their lives.
17
ACKNOWLEDGMENTS The authors are grateful for the cooperation of the school nurses, the teachers, and the lovely children themselves. In addition, we thank Shao-Duan Liu, Chen-Yang Chang, and Hung-Pin Tu for assistance in connecting with the schools and Chin-Wen Liu and Shiu-Chu Lee for help with data collection.
REFERENCES Banks, E. C., Ferretti, L. E., and Shucard, D. W. (1997). Effects of low level lead exposure on cognitive function in children: A review of behavioral, neuropsychological and biological evidence. Neurotoxicology 18, 237}281. Cheng, T. J., Wong, R. H., Lin, Y. P., Hwang, Y. H., Horng, J. J., and Wang, J. D. (1998). Chinese herbal medicine, sibship, and blood lead in children. Occup. Environ. Med. 55, 573}576. Chu, N. F., Liou, S. H., Wu, T. N., Ko, K. N., and Chang, P. Y. (1998). Risk factors for high blood lead levels among the general population in Taiwan. Eur. J. Epidemiol. 14, 775}781. Chuang, H. Y., Chao, K. Y., and Wu, T. N. (1996). The knowledge of parents about lead hazard and their attitude towards children blood lead screening in a kindergarten in Tao-Yuan area. Chin. J. Fam. Med. 6, 101}108. Dudek, B., and Merecz, D. (1997). Impairment of psychological functions in children environmentally exposed to lead. Int. J. Occup. Med. Environ. Health 10, 37}46. Fergusson, D. M., Horwood, L. J., and Lynskey, M. T. (1997). Early dentine lead levels and educational outcomes at 18 years. J. Child Psychol. Psychiatry 38, 471}478. Lansdown, R., Yule, W., Urbanowicz, M. A., and Hunter, J. (1986). The relationship between blood}lead concentrations, intelligence, attainment and behaviour in a school population: The second London study. Int. Arch. Occup. Environ. Health 57, 225}235. Needleman, H. L. (1990a). The future challenge of lead toxicity. Environ. Health. Perspect. 89, 85}89. Needleman, H. L. (1990b). What can the study of lead teach us about other toxicants? Environ. Health Perspect. 86, 183}189. Nevin, R. (2000). How lead exposure relates to temporal changes in IQ, violent crime, and unwed pregnancy. Environ. Res. 83, 1}22. Rabinowitz, M. B., Wang, J. D., and Soong, W. T. (1991). Dentine lead and child intelligence in Taiwan. Arch. Environ. Health 46, 351}360. Rabinowitz, M. B., Wang, J. D., and Soong, W. T. (1992). Children’s classroom behavior and lead in Taiwan. Bull. Environ. Contam. Toxicol. 48, 282}288. Schwartz, J. (1994). Low-level and exposure and children’s IQ: A meta-analysis and search for a threshold. Environ. Res. 65, 42}55. Statistics, Burean of (1999). Social Indicators of Republic of China. Directorate-General of Budget, Accounting and Statistics, Executive Yuan, Taipei. Tong, S., Baghurst, P., McMichael, A., Sawyer, M., and Mudge, J. (1996). Lifetime exposure to environmental lead and children’s intelligence at 11}13 years: The Port Pirie cohort study BMJ 312, 1569}1575. Wang, J. D., Jang, C. S., Hwang, Y. H., and Chen, Z. S. (1992). Lead contamination around a kindergarten near a battery recycling plant. Bull. Environ. Contam. Toxicol. 49, 23}30.
18
WANG ET AL.
Wasserman, G. A., Liu, X., Lolacono, N. J., Factor-Litvak, P., Kline, J. K., Popovac, D., Morina, N., Musabegovic, A., Vrenezi, N., Capuni-Paracka, S., Lekic, V., Preteni-Redjepi, E., Hadzialjevic, S., Slavkovich, V., and Graziano, J. H. (1997). Lead exposure and intelligence in 7-year-old children: The Yugoslavia Prospective Study. Environ. Health Perspect. 105, 956}962. Wu, T. N., Liou, S. H., Shen, C. Y., Hsu, C. C., Chao, S. L., Wang, J. H., Chang, S. F., Ko, K. N., Chiang, H. C., and Chang, P. Y. (1998a). Surveillance of noise-induced hearing loss in Taiwan, ROC: A report of the PRESS-NHL results. Prev. Med. 27, 65}69. Wu, T. N., Shen, C. Y., Liou, S. H., Chao, S. L., Hsu, C. C., Lin, F. T., Ko, K. N., and Chang, P. Y. (1998b). Reducing lead
exposure by surveillance system: The Taiwan experience. Arch. Environ. Health 53, 75}78. Wu, T. N., Shen, C. Y., Liou, S. H., Yang, G. Y., Ko, K. N., Chao, S. L., Hsu, C. C., and Chang, P. Y. (1997). The epidemiology and surveillance of blood lead in Taiwan (ROC): A report on the PRESS-BLL project. Int. Arch. Occup. Environ. Health 69, 386}391. Wu, T. N., Yang, K. C., Wang, C. M., Lai, J. S., Ko, K. N., Chang, P. Y., and Liou, S. H. (1996). Lead poisoning caused by contaminated Cordyceps, a Chinese herbal medicine: Two case reports. Sci. Total Environ. 182, 193}195. Yu, E. C., and Yeung, C. Y. (1987). Lead encephalopathy due to herbal medicine. Chin. Med. J. 100, 915}917.
Relationship between Blood Lead Concentrations and Learning Achievement among Primary School Children in Taiwan1 Chao-Ling Wang,* - Hung-Yi Chuang,? A Chi-Kung Ho,- ? Chun-Yuh Yang,A Jin-Lian Tsai,Ting-Shan Wu, B and Trong-Neng Wu- *Department of Family Medicine, Yuan’s General Hospital; -Graduate Institute of Occupational Safety and Health, Kaohsiung Medical University; ?Department of Occupational Medicine, Koahsiung Medical University Hospital; ASchool of Public Health, College of Health Science, Kaohsiung Medical University; Institute of Medicine, Kaohsiung Medical University; and BThe Koahsiung City Government’s Health Department Received July 6, 2001
ger inBuence on childern’s language ability (Chinese) than on their ability to calculate (Mathematics). Our results suggest that environmental lead exposure adversely affects a child’s academic achievement, making a direct link between exposure to lead and academic attainment. 2002 Elsevier
Over the past 20 years lead has been proven to exert an inBuence on the intelligence of children. Especially for children exposed to environmental lead, average blood lead was often lower than the ofAcially recognized intoxication level. Because Kaohsiung is an industrial area in Taiwan and lead exposure is an important environmental issue, we attempted to ascertain the extent to which environmental lead inBuences the achievement of primary school children. We randomly selected 934 children from 32 primary schools in 11 districts of Kaohsiung City. Blood lead levels of the children were checked, and they were administered a questionnaire about their family information. Scores of several courses were used in this study on the relationship between a child’s blood lead and his or her academic performance (Ranking with his or her classmates), including Chinese (reading and writing short Chinese articles), Mathematics, History and Society, and Natural Science. Multiple regression models were done with adjustments for the confounding effects of their parents’ socioeconomic levels. The mean (SD) of 934 blood lead level was 5.50 (1.86) g/dL. Spearman’s coefAcient showed that class rankings in Chinese, Mathematics, Natural Science, and History and Society were all strongly associated with blood lead levels (P0.01). The multiple regression models revealed that blood lead level exerts a stron-
Science (USA)
Key Words: children; blood lead; learning achievement; school attainment.
INTRODUCTION
The rapid industrialization of the past three decades has resulted in prosperity and democratization in Taiwan; however, it has also led to environmental pollution. Lead is one of the leading sources of pollution in Taiwan. As mentioned in previous studies, in Taiwan about 10% of lead workers have high blood lead levels (Wu et al., 1998a,b). In addition, two previous studies found that children residing near a battery factory had lower intelligence than those residing in a control area (Rabinowitz et al., 1992; Wang et al., 1992). In those studies, the authors suggest that lead concentrations in the soil and air might have had a signi7cant effect on the children’s intelligence. Kaohsiung City, a southern port city and industrial center in Taiwan, is well known for its air and water pollution problems. Exhaust gas from motor vehicles is an area of great public concern. Lead from leaded gasoline and various industries has been the main source of lead pollution in this area. Although the phase out of leaded gasoline started before 1988 and by 2000 it was totally phased out in Taiwan, there are still concentrations of lead in waste and dust deposits.
1 This study was supported by the Department of Health, the Executive Yuan, Taiwan (Grant DOH89-TD-2106, NSC 892320-B-037-045), and Kaohsiung City Government Health Department. 2 To whom correspondence should be addressed at Graduate Institute of Occupational Safety and Health, Kaohsiung Medical University, 100 Shih-Chuang First Road, Kaohsiung City, Taiwan, 807. Fax: #886-7-3115948. E-mail: kmco6849@ms14. hinet.net.
12 0013-9351/02 $35.00 2002 Elsevier Science (USA) All rights reserved.
CHILDREN BLOOD LEAD AND LEARNING ACHIEVEMENT
The potential neurotoxicity of lead found in children all over the world has been explored in many recent epidemiological studies. These studies have gradually shifted from describing the general neurotoxicity observed in a relatively few individuals to studying more subtle impairments caused by this form of pollution in large numbers of children. Changes in the functional abilities of a signi7cant proportion of a population have potentially serious consequences for the affected individuals as well as for society (Banks et al., 1997). Because several previous studies have documented obvious neurotoxicity of lead and its in8uence on childhood development of intelligence using the intelligence quotient (IQ) as a measure, our aim was to focus on environmental lead in relation to academic achievement of children. The purpose of this study was to better understand the relationship between exposure to environmental lead and attainment in speci7c school subjects. MATERIALS AND METHODS
All 11 Kaohsiung City districts were covered in this study. Children were randomly chosen from 32 primary schools. At least one school per district was represented. At each school, one class of thirdgraders (out of a total of six grades) was randomly chosen for the study. Parental consent was obtained for the collection of blood from each child. Blood lead levels were determined and questionnaires were administered from December 1998 to March 1999. The blood samples were analyzed for whole blood lead levels in Kaohsiung Medical University Hospital. After pretreatment with cell lysis and digestion in a Class 100 hood to which air was supplied and cleaned by a high-ef7ciency particulate 7lter, the blood samples were analyzed by Zeeman effect graphite furnace atomic absorption spectrometry (GF-AAS, Perkin-Elmer 5100 PC with AS 71 autosampler), with intralaboratory quality controls. With use of commercial standard materials (Betherning Institute, Bio-Rad), all coef7cients of variation (CVs) were less than 3% for measurements at high levels (40.5}72.7 g/dL) and medium levels (22.4}45.3 g/dL), and were less than 5% for those at low levels (5.6}8.9 g/dL) during the 6 months of the study. Since 1993, the laboratory in Kaohsiung Medical University Hospital has been participating in the interlaboratory blood lead pro7ciency testing program of the Taiwan Department of Health. The results of our measurements are all within the reference ranges, indicating that our blood lead measurements are relatively accurate. At the same time,
13
children were required to answer the questionnaires regarding personal information such as gender, birth date, area of residence, health status, and the occupations, ages, educational levels of the parents under the supervision of teachers or school nurses. In the mean time, we collected the children’s semester grades for Chinese (reading and writing short Chinese articles), Mathematics, History and Society, and Natural Science. In order to prevent interference from different teacher’s grading criteria, we converted the children’s credit scores into class rankings. For example, a class rank of 1 in mathematics would mean that the child had the highest score in his or her math class. Therefore, a larger rank number means the achievement was poor. A small amount of score data was missing (0.6%) due to changes in residence and schools among the children. Parents were classi7ed into seven educational levels; illiterate, primary school, junior high school, senior high school, junior college, university, and higher, with coding numbers ranging from 0 to 6, 0 being illiterate and 6 the highest. The father was classi7ed into 7ve social}economic levels adjusted by Hollingshead’s Index of Social Position according to the father’s occupation and education. In brief, the Hollingshead’s index includes 7ve social}economic classes, the lowest in degree of professionalism and lowest educational level being classi7ed as a Class 5 and the highest in professionalism and the highest in educational level being classi7ed as Class 1. In Taiwan, because women usually resign from work after getting married or after giving birth, the classi7cation of maternal occupation was dropped in this study and the maternal education level was considered a confounding factor in the analysis. We used the SAS system to analyze the data and to study the relationship between blood lead and academic achievement in class. The distribution of children’s blood lead levels had a bell-shaped distribution, so the data needed no transformation. Pearson correlation and Spearman correlation coef7cients were used to determine the bivariate correlation. Continuous variables such as age were performed by Pearson correlation, and ordinal variables, such as ranks, were performed by Spearman correlation. In addition, multiple regression models with adjustments to control for the confounding factors of gender, father’s social economic status, and mother’s education level were also used to investigate the association of blood lead levels and these children’s academic achievement. The standardized residuals of these regression models were analyzed to test the assumptions.
14
WANG ET AL.
TABLE 2
RESULTS
In total, 934 children (454 girls and 480 boys) were involved in this study, which took place between December 1998 and March 1999. Their ages ranged from 8 to 12 years old (mean 8.9 years, SD"0.41). Most of the children in this study were around 8}9 years old. Some children had delayed their schedule due to moving from another country or to a health condition. Four children were older than 11 years old and six were over 10 years old. However, there were no mental or learning problems among these older
TABLE 1 Characteristics of All 934 Children and Their Parents’ Occupation and Education Levels Total (%) Number: Blood lead* (g/dL): Children’s age* (years): Father’s occupation Level 1 Level 2 Level 3 Level 4 Level 5
Girls (%)
Boys (%)
934 5.50 (1.86)*
454 (48.6) 5.30 (1.74)*
480 (51.4) 5.68 (1.96)*
8.85 (0.41)*
8.81 (0.34)*
8.90 (0.46)*
70 280 310 200 15
33 138 149 97 8
37 142 161 103 7
Father’s education (years) Primary school (1}6) 47 Junior high school (7}9) 204 Senior high school (9}12) 344 Junior college (10}15) 172 University (13}16) 108 Master ('16) 17
(8) (32) (35.4) (22.9) (1.7)
(5.3)
25 (5.7)
(8.2) (31.6) (35.8) (22.9) (1.6)
22 (4.9)
(22.9)
101 (23)
103 (22.8)
(36.8)
161 (36.6)
183 (40.5)
(18.4) (11.6) (1.8)
89 (20.2) 56 (12.7) 8 (1.8)
83 (18.4) 52 (11.5) 9 (2)
Father social economic status Class 1 83 (9.8) Class 2 152 (17.9) Class 3 305 (36.0) Class 4 219 (25.9) Class 5 88 (10.4) Mother’s education (years) Primary school (1}6) 70 Junior high school (7}9) 221 Senior high school (9}12) 423 Junior college (10}15) 106 University (13}16) 67 Master ('16) 2
(7.3) (30.4) (32.8) (21.4) (1.8)
42 73 149 108 46
(10.0) (17.5) (35.6) (25.8) (11)
41 79 156 111 42
(9.6) (16.5) (32.5) (23.1) (8.8)
(7.9)
38 (8.7)
32 (7.1)
(24.9)
101 (23.1)
120 (26.5)
(47.6)
202 (46.20)
221 (48.9)
(11.9) (7.5) (0.2)
56 (12.8) 40 (9.2) 0 (0)
50 (11.1) 27 (6.0) 2 (0.4)
Father’s age (years) 40.68 (4.19)* 40.45 (3.90)* 40.91 (4.45)* Mother’s age* (years) 37.79 (4.24)* 37.56 (4.06)* 38.01 (4.41)* Note. Asterisk indicates mean (SD).
Correlation CoefAcients of Blood Lead with Dependent Factors Factors Age of children Father’s age Father’s occupation Father’s education Father’s SES Mother’s age Mother’s occupation Mother’s education Class ranking Chinese Mathematics History Science
Number 925 870 871 888 843 875 894 885 925 925 924 924
Coef7cient 0.048* 0.017* !0.112 !0.236 0.166 0.009* !0.020 !0.179 0.128 0.103 0.126 0.099
P value 0.143 0.611 0.001 (0.001 (0.001 0.795 0.548 (0.001 (0.001 0.002 (0.001 0.003
Note. SES, social economic status. *Pearson coef7cients; others are Spearman coef7cients.
children. The mean ages (SD) of the parents were 40.7 (4.19) years for the fathers and 37.8 (4.24) years for the mothers. All study subjects lived within the limits of Kaohsiung City, and most of them lived near their school. The distributions found in socioeconomic status (SES) of the fathers and education levels of mothers can be seen in Table 1. The SES of the children’s fathers mostly fell within Class 3 and 4 in the Hollingshead’s index classi7cation, with no difference between the parents of boy and girl students. The mean (SD) blood}lead level of the 934 study subjects was 5.50 (1.89) g/dL. The levels ranged form 0.2 to 25.5 g/dL. By correlating blood lead levels and dependent factors, we found that class rankings in Chinese, Mathematics, Natural Science, and History and Society were all strongly associated with blood}lead levels (P0.01) (Table 2). The socioeconomic status of the father and the educational levels of the mother were found related to not only their children’s blood lead levels but also their academic achievement. Therefore, such family conditions were important confounders in the association of blood lead levels and academic achievement in children. In order to clarify this complicated relationship, we used multiple regression models to analyze the relation between blood lead levels and class rankings. The class rankings were found to have signi7cant correlations to blood lead after adjusting for father’s SES and maternal education levels (Table 3). The models revealed that children with higher blood lead levels had signi7cant lower academic achievement in Chinese, History and Society, Mathematics, and
15
CHILDREN BLOOD LEAD AND LEARNING ACHIEVEMENT
TABLE 3 Multiple Regression Models of Class Achievement and Predictors in Grade 3 Children, Kaohsiung City
Rank of Chinese
Lead Gender (girls vs boys) Father SES Mother education Intercept Adjusted R-square
Rank of History and Society
Rank of Mathematics
Rank of Natural Science
estimate
SE
estimate
SE
estimate
SE
estimate
SE
0.37 4.53
0.15* 0.57**
0.43 2.78
0.15* 0.56**
0.38 0.69
0.16* 0.58
0.32 1.24
0.16* 0.58*
0.93 !1.35 11.94
0.30** 0.34** 1.89
1.18 !0.97 10.53
0.29** 0.32** 1.84
0.89 !0.80 12.35
0.14
0.10
0.31** 0.34* 1.93 0.04
0.85 !1.34 13.96
0.31** 0.34** 1.92 0.07
Note. SE, standard error. *P(0.05. **P(0.01.
Natural Science, even after adjusting for other risk factors. The adjusted R-square values for the Chinese model was 0.14, which means that about 14% of the variance found in academic achievement in Chinese could be explained by our model. The girls had achievement scores in Chinese, History and Society, and Natural Science higher than those of the boys. Father’s SES and maternal education levels were still signi7cantly associated with children’s class achievement in all four subjects after adjusting for blood lead levels. Academic achievement in Mathematics was weakly associated with children’s blood lead levels. Blood lead level seemed to play a more important role in the children’s ability to learn language (Chinese) than in their ability to calculate (Mathematics). The plot of residuals for Chinese and Mathematics models both were randomly scattered about zero. The pattern of residuals suggests the presence of an outlier (blood lead"25.5 g/dL). But the outlier did not twist the regression line very much. Rewrite helps to ensure that the model’s predictions about the distribution of the independent variable were not violated. DISCUSSION
Our study compared blood lead levels with course score data on 934 school children. The male-to-female ratio was 1.05, a 7gure quite similar to the ratio found in a primary school population in Taiwan. The proportions of the parents’ educational levels were also similar, but higher than those found national-wide. For example, 19.1% of the parents have junior high school educations nation-wide, compared to the 22.9% found in our study. The
nation-wide percentages and the percentages found in our study continued to be similar with 35.4 vs 36.8% for senior high school, 14.9 vs 18.4% for junior college, and 11.4 vs 13.4% for university and higher (Statistics, 1999). Therefore, sampling of this study was similar to the normal distribution of Taiwan. Also, this study may not have selection bias because data on the dependent variables had not been collected at the time the sample was assembled. Variation in teacher grading was also controlled by random sampling of the classes and the course scores were transformed into class rankings in each subject. The results of our study support the hypothesis that children’s school achievements might be inversely related to their absorption of environmental lead. A signi7cant relationship between lead and achievement was found not only in our analyses using Pearson and Spearman correlation coef7cients, but also in multiple regression models after adjustment for gender, father’s social economic status, and mother’s education. Chinese language skills were more strongly related to blood lead level than was Mathematics. This could mean that lower level lead exposure may have more in8uence on the ability to memorize than on the ability calculate. The outlier was found in the data of one boy. His blood lead was 25.5 g/dL in the time course of the study. We visited with his parents, and they told us that he had taken herbal medicine for his health since childhood. In Taiwan, child lead poisoning resulting from Chinese herbal medicines is not uncommon (Cheng et al., 1998; Chu et al., 1998; Wu et al., 1994, 1997; Yu and Yeung, 1987). But the outlier did not twist the regression line very much. After the outlier was detected from the database, the
16
WANG ET AL.
regression models were similar to the previous one. The regression coef7cient was 0.33 for the Chinese language model and 0.36 for the Mathematics model. Due to the slight change and similar regression line, we did not delete the outlier in our presentation. But we did want to emphasize the important source of childhood lead exposure. Since the 1980s many epidemiological studies have con7rmed that low-level, subclinical lead exposure in early life can be associated with decrements in childern’s intelligence. One meta-analysis study reported that an increase of 1 g/dL in blood lead would cause a loss of 0.26 IQ point. The same study found a slope of 0.323 in studies on children with mean blood lead levels below 15 g/dL, and 0.232 in those with mean blood lead above 15 g/dL (Schwartz, 1994). The slopes in our models are similar to those in the meta-analysis, since most of our blood lead levels were lower than 15 g/dL. Nevin has also concluded that the IQ-to-blood lead slope may increase at lower blood lead levels (Nevin, 2000). In our study only 12 children (1.28%) had blood lead levels over 10 g/dL and only one boy had blood lead levels exceeding 15 g/dL. Most of the children in our study had levels below 10 g/dL. Therefore, much like the studies that demonstrated a relationship between lead and intelligence, our study found a signi7cant inverse relationship between lead and academic achievement. Moreover, our evidence associating blood lead levels with children’s class achievement was more direct than the evidence associating blood lead and IQ, and our 7ndings remained signi7cant even at blood lead levels lower than 10 g/dL. This means that the lead exposure can affect not only children’s intelligence, but also their personal achievement. The mean age of the children in our study was 8.85 years. Several other studies have reported the in8uence of lead to be obvious in children under the age of 10. Schwartz states that during the 7rst 3 years of life, when basic cognitive abilities develop, young children face the greatest risk of IQ losses due to lead exposure (Schwartz, 1994). In addition, exposure to environmental lead during the 7rst 7 years of life has been associated with cognitive de7cits that seem to persist into later childhood (Tong et al., 1996). Fergusson et al., conducted a study to measure dentine lead levels in children 6 to 8 years old and compared the 7gures with educational outcomes of the same children when they were 18 years old. They found that early mildly elevated dentine lead levels had a modest but detectable effect on measures of individual achievement such as reading ability and school examinations. They found that
these effects extended to late adolescence (Fergusson et al., 1997). However, this study was a cross-sectional study of blood lead and classroom attainment. We did not have the earlier blood lead data for the children in our study, but we showed that blood lead had a similar in8uence on learning achievement simply by studying the class rankings in four main subjects. In a comparison of industralized areas, one study showed that dentine lead levels of shed incisors for children around Taipei City were slightly higher at that time than levels reported in Boston, and the intelligence scores from Raven’s Colored Progressive Matrices Test were negatively correlated with dentine lead levels, especially among girls and children whose parents had lower education levels (Rabinowitz et al., 1991). Dentine lead is not easy to measure, and therefore blood lead levels have been widely used in Taiwan and world-wide. In 1995, 319 preschool children living in the Taipei metropolitan area including Taipei City, Taipei County, and Tao-Yuan County were tested. Their mean (SD) blood lead was 4.4 (2.4) g/dL (Cheng et al., 1998). In another study of children in a Tao-Yuan kindergarten (n"84), children’s mean blood lead level (SD) was 3.9 (2.1) g/dL (Chuang et al., 1996). Our study of children in Kaohsiung City, an industrial port area in southern Taiwan, reports a higher blood lead level, 5.5 (1.9) g/dL than those found in northern Taiwan. These 7ndings indicate that the local Kaohsiung government, elected of7cials, and private citizens should act to reduce lead exposure among the children in this city. Intelligence tests for children in primary schools were often based on three dimensions: verbal, mathematical, and performance IQ. However, there has been no consensus on whether lead exposure can be more strongly associated with any of these three aspects separately. In 1986, Lansdown reported a relationship between lead and children’s reading and spelling (Lansdown et al., 1986). The 7ndings of a study by Wasserman in 1997 demonstrated that perceptual}motor skills are signi7cantly more sensitive to lead exposure than the language-related aspects of intelligence (Wasserman et al., 1997). Other studies have revealed that environmental exposure to lead exerts a negative impact upon psychological functions not only in the intelligence of children but also in hand}eye coordination, perception, memory, reaction time and accuracy, language function, attention, classroom behavior, and behavioral disorders (Dudek and Merecz, 1997; Needleman, 1990b). Therefore, lead in8uences not only the general intelligence of children but also visual}
CHILDREN BLOOD LEAD AND LEARNING ACHIEVEMENT
spatial-dependent neuropsychological behavior and performance. Furthermore, the Chinese language has visual}spatial pathway’s that are different from those of nonideographic languages. There had been no previous study on this aspect of lead in8uences. Our study found that blood lead levels were related more strongly to language skills (Chinese) than to calculating skills (Mathematics). Furthermore, our study provides more direct evidence that lead exposure can affect children’s learning achievement, especially in terms of their language-learning abilities. The main 7nding of our report was that a clear relationship exists between blood lead and a child’s learning achievement. The class rankings have an important social impact because a large number of children are affected. We also found that SES of parents conveys information about a child’s potential lead exposure as well as other independent determinants of achievement. Table 2, for example, shows that certain variables, including father’s SES, mother’s educational level, and class rankings were all signi7cantly related to blood lead. Therefore, in our study of blood lead and children achievement, SES was an important confounding factor. Multiple regression model (Tables 3) show parameter estimates of lead to range from 0.32 to 0.43, meaning that when the blood lead increases to 1 g/dL, the children’s class rankings, may fall by around 0.3 point, after the confounders are held constant. Gender was also an important factor in the relationship between lead and performance. In most epidemiological studies of lead, the means of the female blood lead were often lower than those of the male. Our study revealed the same condition (Table 1). The girls’ performances in Chinese and in History and Society were signi7cantly better than those of the boys (Table 3). There was no difference between girls and boys in Mathematics performance. Therefore, the regression model result of better performance by girls was due not only to gender difference but also to the lead in8uence. For the past decade defects in intelligence and changes in behavior caused by low-level exposure to lead have been a world-wide issue. Lead exposure has been associated with hyperactivity and has been considered a risk factor for learning problems (Needleman, 1990a). In Taiwan, since the use of alkyl lead in gasoline additives was prohibited in 2000 government statistics have shown a decrease in air lead. Therefore, follow-up studies on the behavior of children with higher blood lead levels in Taiwan may be needed to clarify the extent to which lead in8uences their lives.
17
ACKNOWLEDGMENTS The authors are grateful for the cooperation of the school nurses, the teachers, and the lovely children themselves. In addition, we thank Shao-Duan Liu, Chen-Yang Chang, and Hung-Pin Tu for assistance in connecting with the schools and Chin-Wen Liu and Shiu-Chu Lee for help with data collection.
REFERENCES Banks, E. C., Ferretti, L. E., and Shucard, D. W. (1997). Effects of low level lead exposure on cognitive function in children: A review of behavioral, neuropsychological and biological evidence. Neurotoxicology 18, 237}281. Cheng, T. J., Wong, R. H., Lin, Y. P., Hwang, Y. H., Horng, J. J., and Wang, J. D. (1998). Chinese herbal medicine, sibship, and blood lead in children. Occup. Environ. Med. 55, 573}576. Chu, N. F., Liou, S. H., Wu, T. N., Ko, K. N., and Chang, P. Y. (1998). Risk factors for high blood lead levels among the general population in Taiwan. Eur. J. Epidemiol. 14, 775}781. Chuang, H. Y., Chao, K. Y., and Wu, T. N. (1996). The knowledge of parents about lead hazard and their attitude towards children blood lead screening in a kindergarten in Tao-Yuan area. Chin. J. Fam. Med. 6, 101}108. Dudek, B., and Merecz, D. (1997). Impairment of psychological functions in children environmentally exposed to lead. Int. J. Occup. Med. Environ. Health 10, 37}46. Fergusson, D. M., Horwood, L. J., and Lynskey, M. T. (1997). Early dentine lead levels and educational outcomes at 18 years. J. Child Psychol. Psychiatry 38, 471}478. Lansdown, R., Yule, W., Urbanowicz, M. A., and Hunter, J. (1986). The relationship between blood}lead concentrations, intelligence, attainment and behaviour in a school population: The second London study. Int. Arch. Occup. Environ. Health 57, 225}235. Needleman, H. L. (1990a). The future challenge of lead toxicity. Environ. Health. Perspect. 89, 85}89. Needleman, H. L. (1990b). What can the study of lead teach us about other toxicants? Environ. Health Perspect. 86, 183}189. Nevin, R. (2000). How lead exposure relates to temporal changes in IQ, violent crime, and unwed pregnancy. Environ. Res. 83, 1}22. Rabinowitz, M. B., Wang, J. D., and Soong, W. T. (1991). Dentine lead and child intelligence in Taiwan. Arch. Environ. Health 46, 351}360. Rabinowitz, M. B., Wang, J. D., and Soong, W. T. (1992). Children’s classroom behavior and lead in Taiwan. Bull. Environ. Contam. Toxicol. 48, 282}288. Schwartz, J. (1994). Low-level and exposure and children’s IQ: A meta-analysis and search for a threshold. Environ. Res. 65, 42}55. Statistics, Burean of (1999). Social Indicators of Republic of China. Directorate-General of Budget, Accounting and Statistics, Executive Yuan, Taipei. Tong, S., Baghurst, P., McMichael, A., Sawyer, M., and Mudge, J. (1996). Lifetime exposure to environmental lead and children’s intelligence at 11}13 years: The Port Pirie cohort study BMJ 312, 1569}1575. Wang, J. D., Jang, C. S., Hwang, Y. H., and Chen, Z. S. (1992). Lead contamination around a kindergarten near a battery recycling plant. Bull. Environ. Contam. Toxicol. 49, 23}30.
18
WANG ET AL.
Wasserman, G. A., Liu, X., Lolacono, N. J., Factor-Litvak, P., Kline, J. K., Popovac, D., Morina, N., Musabegovic, A., Vrenezi, N., Capuni-Paracka, S., Lekic, V., Preteni-Redjepi, E., Hadzialjevic, S., Slavkovich, V., and Graziano, J. H. (1997). Lead exposure and intelligence in 7-year-old children: The Yugoslavia Prospective Study. Environ. Health Perspect. 105, 956}962. Wu, T. N., Liou, S. H., Shen, C. Y., Hsu, C. C., Chao, S. L., Wang, J. H., Chang, S. F., Ko, K. N., Chiang, H. C., and Chang, P. Y. (1998a). Surveillance of noise-induced hearing loss in Taiwan, ROC: A report of the PRESS-NHL results. Prev. Med. 27, 65}69. Wu, T. N., Shen, C. Y., Liou, S. H., Chao, S. L., Hsu, C. C., Lin, F. T., Ko, K. N., and Chang, P. Y. (1998b). Reducing lead
exposure by surveillance system: The Taiwan experience. Arch. Environ. Health 53, 75}78. Wu, T. N., Shen, C. Y., Liou, S. H., Yang, G. Y., Ko, K. N., Chao, S. L., Hsu, C. C., and Chang, P. Y. (1997). The epidemiology and surveillance of blood lead in Taiwan (ROC): A report on the PRESS-BLL project. Int. Arch. Occup. Environ. Health 69, 386}391. Wu, T. N., Yang, K. C., Wang, C. M., Lai, J. S., Ko, K. N., Chang, P. Y., and Liou, S. H. (1996). Lead poisoning caused by contaminated Cordyceps, a Chinese herbal medicine: Two case reports. Sci. Total Environ. 182, 193}195. Yu, E. C., and Yeung, C. Y. (1987). Lead encephalopathy due to herbal medicine. Chin. Med. J. 100, 915}917.