Context.— Many studies have found a significant inverse
association between early exposure to environmental lead and cognitive
function in childhood. Whether these effects are reversible when
exposure is reduced is not clear.
Objective.— To assess the reversibility of the apparent
effects of lead on cognitive abilities in early childhood by testing
whether declines in blood lead concentrations beyond the age of 2 years
are associated with improvements in cognition.
Setting.— Urban and rural communities surrounding a large
lead smelter in Port Pirie, South Australia.
Participants.— A total of 375 children followed up from
birth to the age of 11 to 13 years.
Design.— Long-term prospective cohort study.
Main Outcome Measures.— The Bayley Mental Development Index
at age 2 years, the McCarthy General Cognitive Index at age 4 years,
and IQs from the Wechsler Intelligence Scale (revised version) at ages
7 and 11 to 13 years.
Results.— Mean blood lead concentrations in the children
decreased from 1.02 µmol/L (21.2 µg/dL) at age 2 years to 0.38
µmol/L (7.9 µg/dL) at age 11 to 13 years, but cognitive scores in
children whose blood lead concentration declined most were generally
not improved relative to the scores of children whose blood lead levels
declined least. Changes in IQ and declines in blood lead levels that
occurred between the ages of 7 and 11 to 13 years (r = 0.12,
P = .09) suggested slightly better cognition among children
whose blood lead levels declined most.
Conclusion.— The cognitive deficits associated with exposure
to environmental lead in early childhood appear to be only partially
reversed by a subsequent decline in blood lead
level.
DURING THE PAST 25 years there has been
concern about the potential effects of environmental lead on childhood
development. In the late 1970s and early 1980s, cross-sectional studies
examined whether exposure to environmental lead at levels previously
believed to be innocuous affects neuropsychological
development.1-9 Although the findings were not entirely
consistent, most found an inverse association between exposure measures
and neuropsychological performance.1-7
To test this hypothesis and address the limitations of cross-sectional
studies, cohort studies were subsequently conducted in several
countries. These studies collected information on critical features of
lead exposure, such as the timing and extent, together with many other
socioenvironmental factors that may confound the relationship between
lead exposure and neuropsychological development.
Most,10-15 but not all, cohort studies16,17
found a significant inverse association between early exposure to
environmental lead and cognitive functioning in childhood after
adjustment for confounding factors.
A question with important public health ramifications is whether the
apparent deleterious effect of early-life exposure to lead is reversed
when, later in childhood, exposure is reduced. Such exposure reduction
could occur either by environmental management or as a consequence of
the decreased absorption that appears to accompany growth.
Recently, Ruff and colleagues18,19 reported that decreases
in blood lead level were associated with cognitive improvements in
children aged 13 to 87 months. Participants in their study all had
initial blood lead levels between 1.21 and 2.66 µmol/L (25 and 55
µg/dL) and reductions were achieved with the chelating agent calcium
sodium EDTA. However, the question of whether changes in blood lead
levels in individuals are associated with changes in cognition has not
been examined in any long-term cohort study.
The Port Pirie Cohort Study commenced in 1979. The geometric mean blood
lead concentration of the children in this cohort increased from 0.40
µmol/L (8.3 µg/dL) at birth to 1.02 µmol/L (21.2 µg/dL) at age
2 years and decreased to 0.38 µmol/L (7.9 µg/dL) by the age of 11
to 13 years. In previous studies, we reported that exposure to lead in
this cohort of children was inversely associated with cognitive
performance at ages 2, 4, 7, and 11 to 13 years, and that this
association was still apparent after adjustment for a wide range of
confounding factors.15,20-22 While these findings are
strongly suggestive that any causal effect of lead on cognitive
abilities persists throughout the primary school years, it is
nevertheless conceivable that 1 or more (unidentified) confounding
factors affected the relationship between lead exposure and
developmental measures at all ages.
This article pursues an alternative approach to the assessment of
reversibility by examining the relationship between individual changes
in blood lead concentrations and individual changes in measures of
cognitive development during the life of the cohort.
Details of the research design have been reported elsewhere and are
summarized in Figure 1.15,20-22 All
women living in Port Pirie, South Australia (site of one of the largest
lead smelting facilities in the southern hemisphere), or in the
immediately adjoining region who were enrolled for antenatal care
between May 1979 and May 1982 were advised by their physicians of the
purpose and procedures of the study and were encouraged to contact the
study coordinator (a community health nurse) to discuss participation.
The study was approved by the University of Adelaide Research Ethics
Committee, Adelaide, Australia, and written consent was obtained from
the parents or guardians of all participating children.
The initial study population was composed of an estimated 90% of
all pregnancies in and around Port Pirie during a 3-year period from
1979 to 1982. Of the resulting 723 singleton live births, 601 were
assessed at age 2 years, 548 at age 4 years, 494 at age 7 years, and
375 at age 11 to 13 years. The majority of children lost to follow-up
were in families that either left the Port Pirie area or could not be
contacted despite intensive efforts, while a small number of families
simply discontinued their participation. Potential for selective losses
of participants to bias the study findings was evaluated throughout the
study. The results showed that the correlation between lead exposure
and early measures of cognitive development was slightly stronger in
those children lost to follow-up than in those remaining in the cohort.
Therefore, any differential loss to follow-up is likely to have
resulted in an underestimation of the lead-cognition
relationship.22
As many as 3 venous blood samples were collected from each mother
before delivery. A further sample was collected from the umbilical cord
at birth, and capillary blood samples were obtained from each child at
the ages of 6, 15, and 24 months, and annually thereafter until age 7
years. At age 11 to 13 years, a trained community nurse collected a
venous blood sample from 326 of the 375 children still in the cohort.
At the time of each blood sampling, a structured interview was
conducted to obtain information on a wide range of psychosocial,
demographic, environmental, and biomedical factors. The adequacy of
capillary sampling was tested in a prestudy trial in which both venous
and capillary blood samples were collected simultaneously from 47
children. The estimated Spearman rank correlation coefficient for the 2
sampling methods was 0.97.15 Compliance with the capillary
sampling protocol was monitored throughout the study.
The children's developmental status was assessed using the
Bayley Scales of Infant Development at age 2 years,23 the
McCarthy Scales of Children's Abilities at age 4 years,24
and the revised version of the Wechsler Intelligence Scale for Children
at ages 7 and 11 to 13 years.25 The Bayley Scales of Infant
Development are applicable to children aged 30 months or younger and
yield 2 standardized scores, the Mental Development Index and the
Psychomotor Development Index. Only Mental Development Index scores
were used in the analysis presented herein. The McCarthy Scales of
Children's Abilities, which can be used in children 3 to 7 years old,
consist of 5 scales: verbal, perceptual performance, quantitative,
memory, and motor. The first 3 of these scales are combined to form the
General Cognitive Index. The revised version of the Wechsler
Intelligence Scale for Children, a test of general intelligence
developed for use with children aged 6 to 16 years, was used to assess
the cognitive abilities of each child at ages 7 and 11 to 13 years. At
ages 2, 4, and 7 years, all children were assessed by a single research
psychologist who was unaware of each child's lead exposure status. At
age 11 to 13 years, the subjects were evaluated by a single trained
examiner who had not participated in earlier phases of the cohort study
and who was unaware of the children's exposure and developmental
histories.
Blood lead concentration was measured in the Department of
Chemical Pathology at the Adelaide Women's and Children's Hospital,
Adelaide, Australia, by electrothermal atomization atomic absorption
spectrometry.26 Throughout this study, both internal and
external quality-control procedures were used, with consistently
satisfactory results. A certified commercially prepared product was
used to monitor intrabatch accuracy and ensure interbatch
standardization. The coefficient of variation for blood lead
measurements was 5.7% or less. External quality control was maintained
by participation in 3 major programs: the National Quality-Control
Program conducted by the Standards Association of Australia, Sydney,
and the international programs run by the Health Department of
Pennsylvania, Philadelphia, and Wolfson Research
Laboratory, Birmingham, England. Estimates of blood
lead concentration were standardized to a packed-cell volume of 50%
for umbilical cord blood and 35% for all other samples.
Ancillary variables potentially capable of confounding the
relationship between lead exposure and cognitive development
included parents' occupational prestige, which was assessed using the
Daniel Scale of Prestige of Occupations in Australia (lower
scores indicate more prestigious jobs, which are generally associated
with higher socioeconomic status)27; the caregiving
environment, which was measured with the Home Observation for
Measurement of the Environment (HOME) inventory28; and
maternal intelligence, which was assessed using the Wechsler Adult
Intelligence Scale.29 Other information collected
included family size and functioning scores; maternal general health
status and age at delivery; parental smoking habits, marital status,
and education; and child's sex, age, grade in school, birthweight,
birth rank, feeding method during infancy, duration of breast-feeding,
major life events, and prolonged absences from school for any single
school term during the past 5 years.
Statistical analyses were performed on the natural
logarithm of the blood lead concentration, and all reported mean values
are geometric. To calculate lifetime average blood lead concentrations,
a plot of blood lead concentration against age was constructed for each
child. The lifetime average blood lead concentration up to a particular
age for each child was estimated by dividing the area under that
child's curve by the specified age.
Two approaches were used to examine whether the children's cognition
improved as their lead exposure decreased. First, the relationship
between blood lead concentration and cognitive function was assessed
(using multiple regression models to adjust for potential confounding
factors) at various ages. The covariates included in these models were
child's sex, birthweight, birth rank, feeding style during infancy,
duration of breast-feeding, maternal IQ, maternal age at child's
birth, socioeconomic status, HOME score, parental smoking habits, and
whether the parents lived together or apart. Second, analysis of
variance was used to determine whether changes in cognitive scores were
associated with decreases in blood lead concentration across the
interval between any pair of developmental assessments. For each
comparison, 3 groups were formed on the basis of the tertiles of
decrease in blood lead concentration across that interval. In comparing
scores at ages 7 and 11 to 13 years, simple differences in Wechsler IQ
scores obtained at both ages were used. For all other comparisons, the
developmental scores were first transformed to z scores to
take into account variations in the estimated means and SDs of the
various developmental measures used at each age.
Regression analyses for each specific developmental outcome measure
revealed significant inverse associations between blood lead
concentration and cognitive development at all ages, as previously
described.15,21,22 Simple plots of the covariate-adjusted
differences in developmental scores for the 3 exposure groups defined
by the tertiles of lifetime average blood lead concentration at age 2
years (the age at which the children's developmental status was
assessed for the first time) also indicated that the children with
greater exposure continued to perform less well at older ages (Figure 2).
However, neither of these analyses provides definitive
evidence that the effect of early lead exposure persists throughout
childhood. The possibility of residual confounding and/or an
unidentified bias affecting all the measure-specific analyses similarly
cannot be excluded, and the parallel behavior of the 3 groups
identified in Figure 2 may reflect little more than a "tracking" in
lead exposure. In Table 1 it is readily apparent
that many children with high blood lead levels at age 2 years still had
high blood lead level ranking at age 11 to 13 years, even though the
overall mean exposure level decreased as the children grew older.
We therefore performed an additional analysis on the changes in
cognitive development in relation to the changes in blood
lead concentration for each individual in the study. If the alleged
effect of lead is transient, cognitive performance might be expected to
improve most among those children whose blood lead level experienced
the greatest decline in later childhood. On the other hand, if the lead
effect is persistent, children's performance would not be expected to
improve as their exposure declined.
The children's IQ at age 11 to 13 years (mean score, 100; 95%
confidence interval, 98.8-101.2) was generally lower than at age 7
years (mean score, 104.7; 95% confidence interval, 103.5-106.0). There
was a weak (P = .45) trend suggesting that children
experiencing the greatest decline in lifetime average blood lead
concentration from age 7 years to 11 to 13 years exhibited the smallest
decrease in IQ scores (Table 2).
Similar analyses of the change in z scores from age 2 years to
11 to 13 years and from age 4 years to 11 to 13 years provided no
evidence that the cognitive function of the children whose lead
exposure declined most in those age ranges had significantly improved (Table 3).
Analyses on the ungrouped data found that correlations between
changes in blood lead levels and cognitive functioning were generally
weak and not statistically significant. The strongest association was
between the changes in IQ and declines in blood lead levels that
occurred between ages 7 and 11 to 13 years (r = 0.12,
P = .09). This was weakly suggestive of a slight
"recovery" among the children whose blood lead levels declined
most.
Determination of the course of the cognitive effects of low-level lead
exposure is an important issue because the ramifications of potential
effects of lead exposure for regulatory policy depend not solely on the
extent of the initially observable effects per se, but also on their
persistence over time. To our knowledge, this is the first attempt to
systematically relate changes in blood lead concentration throughout
childhood to changes in measures of cognitive functioning in a
long-term cohort study.
Our analysis shows that, compared with children with lower
exposure levels, the cognitive deficit in the group with higher
exposure changed little with age, even though blood lead levels
declined substantially after age 2 years. It might be argued that the
consistency of the association between lead exposure and cognitive
development at all ages is an artifact of some insufficiency in
measuring confounders at all ages or, alternatively, that it reflects a
persistent, largely irreversible impact on cognitive development.
Although it is almost impossible to dispel the concern of residual
confounding in any observational study, several other prospective
studies have also reported an inverse association between an early
measure of lead exposure and subsequent developmental measures made in
later childhood10,13 or early adulthood.12
Moreover, a number of animal studies have reported lasting effects of
low-level exposure to lead in early life.30-32
An alternative approach for testing the persistence of the effects of
exposure to lead is to assess whether individual patterns of change in
lead exposure are associated with individual changes in
cognition. Ruff and colleagues18,19 found an association
between decreases in blood lead levels and cognitive improvements in
"moderately lead-poisoned" children after intervention, although
the results of their study have been criticized for possible problems
of residual confounding by age and educational
factors.33-36
No statistically significant association between declining blood lead
levels and cognitive changes was observed in our data. However, there
are several issues that need to be considered in interpreting our
findings. First, blood lead level may not be the best indicator of
changes in exposure to lead. It is well known that blood lead is
largely determined by variations in the magnitude and direction of
recent exposure and may reflect only exposure that occurred recently
(eg, within about 1 month).37,38 Therefore, some other
measure of exposure (eg, lead in bone) that better reflects the
whole-body burden of lead may have been preferable for an assessment of
the reversibility of the apparent effects of lead. Second, cognitive
development was assessed using different tests at different ages. The
validity of comparing the results of these tests may be questionable if
they focus on different aspects of cognitive functioning. Third, the
intensity, duration, timing, and pattern of exposure to lead may
determine the nature and degree of reversibility. It is conceivable
that the alleged adverse effects of lead are reversible only below a
threshold exposure level and that this threshold was exceeded by the
vast majority of Port Pirie study subjects. Finally, because there is
substantial intraindividual variability in all measures of cognition,
it is also possible that we had insufficient statistical power to
detect an improvement in cognition among children whose lead exposure
decreased the most—but the absence of any trend in the point estimates
in Table 3 indicates that there may be no obvious beneficial effect of
reducing lead exposure in the latter years of childhood.
If the putative effects of lead on the neuropsychological development
of children are genuinely irreversible, efforts to reduce the exposure
of children to environmental lead from a very early age become
increasingly important.
1.Needleman HL, Gunnoe C, Leviton A.
et al. Deficits in
psychologic and classroom performance of children with elevated dentine
lead levels.
N Engl J Med.1979;300:689-695.Google Scholar 2.Hatzakis A, Kokkevi A, Maravelias C.
et al. Psychometric
intelligence deficits in lead-exposured children. In: Smith MA, Grant
LD, Sors AI, eds. Lead Exposure and Child Development: An
International Assessment. Lancaster, England: Kluwer Academic
Publishers; 1989:211-223.
3.Winneke G, Kramer U, Brockhaus U.
et al. Neuropsychological studies in children with elevated tooth-lead
concentrations, II: extended study.
Int Arch Occup Environ
Health.1983;51:231-252.Google Scholar 4.Harvey PG, Hamlin MW, Kumar R, Delves HT. Blood lead,
behaviour and intelligence test performance in preschool children.
Sci Total Environ.1984;40:45-60.Google Scholar 5.Lansdown R, Yule W, Urbanowicz M-A, Hunter J. The
relationship between blood lead concentrations, intelligence,
attainment and behaviour in a school population: the second London
study.
Int Arch Occup Environ Health.1986;57:225-235.Google Scholar 6.Fulton M, Raab GM, Thomson GOB, Laxen DPH, Hunter R, Hepburn W. Influence of blood lead on the ability and attainment of
children in Edinburgh.
Lancet.1987;1:1221-1226.Google Scholar 7.Schroeder SR, Hawk B, Otto DA, Mushak P, Hicks RE. Separating the effect of lead and social factors on IQ.
Environ
Res.1985;38:144-154.Google Scholar 8.Smith M, Delves T, Lansdown R, Clayton B, Graham P. The
effects of lead exposure on urban children: the Institute of Child
Health/Southampton Study.
Dev Med Child Neurol Suppl.1983;47:1-54.Google Scholar 9.Pocock SJ, Ashby D, Smith MA. Lead exposure and
children's intellectual performance.
Int J Epidemiol.1987;16:57-67.Google Scholar 10.Bellinger D, Stiles K, Needleman HL. Low-level lead
exposure intelligence and academic achievement: a long-term follow-up
study.
Pediatrics.1992;90:855-861.Google Scholar 11.Dietrich KN, Succop PA, Berger OG, Hammond PB, Bornschein RL. Lead exposure and the cognitive development of urban
preschool children: the Cincinnati Lead Study cohort at age 4 years.
Neurotoxicol Teratol.1991;13:203-211.Google Scholar 12.Needleman HL, Schell A, Bellinger D, Leviton A, Allred EN. The long-term effects of exposure to low doses of lead in
childhood: an 11-year follow-up report.
N Engl J Med.1990;322:83-88.Google Scholar 13.Fergusson DM, Horwood LJ. The effects of lead levels on
the growth of word recognition in middle childhood.
Int J
Epidemiol.1993;22:891-897.Google Scholar 14.Wasserman G, Graziano JH, Factor-Litvak P.
et al. Independent effects of lead exposure and iron deficiency anemia on
developmental outcome at age 2 years.
J Pediatr.1992;121:695-703.Google Scholar 15.McMichael AJ, Baghurst PA, Wigg NR, Vimpani GV, Robertson EF, Roberts RJ. Port Pirie Cohort Study: environmental
exposure to lead and children's abilities at the age of four years.
N Engl J Med.1988;319:468-475.Google Scholar 16.Ernhart CB, Morrow-Tlucak M, Wolf AW, Super D, Drotar D. Low level lead exposure in the prenatal and early preschool periods:
intelligence prior to school entry.
Neurotoxicol Teratol.1989;11:161-170.Google Scholar 17.Cooney GH, Bell A, McBride W, Carter C. Low-level
exposure to lead: the Sydney Lead Study.
Dev Med Child Neurol.1989;31:640-649.Google Scholar 18.Ruff HA, Bijur PE, Markowitz M, Ma YC, Rosen JF. Declining blood lead levels and cognitive changes in moderately
lead-poisoned children.
JAMA.1993;269:1641-1646.Google Scholar 19.Ruff HA, Markowitz M, Bijur PE, Rosen JF. Relationships among blood lead level, iron deficiency, and cognitive
development in two-year-old children.
Environ Health Perspect.1996;104:180-185.Google Scholar 20.Wigg NR, Vimpani GV, McMichael AJ, Baghurst PA, Robertson EF, Roberts RJ. Port Pirie Cohort Study: childhood blood lead
and neuropsychological development at age two years.
J Epidemiol
Community Health.1988;42:213-219.Google Scholar 21.Baghurst PA, McMichael AJ, Wigg NR.
et al. Environmental exposure to lead and children's intelligence at the age
of seven years: the Port Pirie Cohort Study.
N Engl J Med.1992;327:1279-1284.Google Scholar 22.Tong SL, Baghurst PA, McMichael AJ, Sawyer M, Mudge J. Lifetime exposure to environmental lead and children's intelligence at
11-13 years: the Port Pirie Cohort Study.
BMJ.1996;312:1569-1575.Google Scholar 23.Bayley N. Bayley Scales of Infant Development. New York, NY: Psychological Corp; 1969.
24.McCarthy D. Manual for the McCarthy Scales of
Children's Abilities. New York, NY: Psychological Corp; 1972.
25.Wechsler D. Manual for the Wechsler Intelligence
Scale for Children—Revised. New York, NY: Psychological Corp;
1974.
26.Standards Association of Australia. Whole Blood
Determination of Lead—Electrothermal Atomization Atomic Absorption
Spectrometric Method. Sydney: Standards Association of Australia;
1985. Australian standard 2787.
27.Daniel A. The measurement of social class.
Community Health Studies.1984;3:218-222.Google Scholar 28.Caldwell B, Bradley R. Home Observation for
Measurement of the Environment. New York, NY: Dorsey; 1985.
29.Wechsler D. Wechsler Adult Intelligence
Scale—Revised. New York, NY: Psychological Corp; 1981.
30.Rice DC, Karpinski KF. Lifetime low-level lead exposure
produces deficits in delayed alternation in adult monkeys.
Neurotoxicol Teratol.1988;10:207-214.Google Scholar 31.Gilbert SG, Rice DC. Low-level lifetime lead exposure
produces behavioural toxicity (spatial discrimination reversal) in
adult monkeys.
Toxicol Appl Pharmacol.1987;91:484-490.Google Scholar 32.Rice DC. Lead-induced changes in learning:
evidence for behavioural mechanisms from experimental animal studies.
Neurotoxicology.1993;14:167-178.Google Scholar 33.Schindler AM, Haecker T, Gould J.
et al. Declining
blood lead levels and cognitive change in children.
JAMA.1993;270:828.Google Scholar 34.Ernhart CB. Declining blood lead levels and cognitive
change in children.
JAMA.1993;270:827-828.Google Scholar 35.Cunningham RD. Declining blood lead levels and
cognitive change in children.
JAMA.1993;270:828-829.Google Scholar 36.Binder S, Matte T. Childhood lead poisoning: the impact
of prevention.
JAMA.1993;269:1679-1681.Google Scholar 37.US Agency for Toxic Substances and Disease Registry. The Nature and Extent of Lead Poisoning in Children in the United
States: A Report to Congress. Atlanta, Ga: US Public Health
Service, Dept of Health and Human Services; 1988.
38.US Environmental Protection Agency. Air Quality
Criteria for Lead: Supplement to the 1986 Addendum. Research
Triangle Park, NC: Office of Health and Environmental Assessment, US
Environmental Protection Agency; 1990. EPA document
600/8-89/049F.