Updated Summary of Evidence
Gary Rischitelli, M.D., J.D., M.P.H.a, b; Peggy Nygren, M.A.a; Christina Bougatsos, B.S.a; Michele Freeman, M.P.H.a; Mark Helfand, M.D., M.P.H.a
The authors of this article are responsible for its contents, including any clinical or treatment recommendations. No statement in this article should be construed as an official position of the U.S. Agency for Healthcare Research and Quality (AHRQ) or the U.S. Department of Health and Human Services.
Address correspondence to: Gary Rischitelli, M.D., J.D., M.P.H., Oregon Health and Science University, Mail Code L606, 3181 SW Sam Jackson Park Road, Portland, Oregon 97239. E-mail: rischite@ohsu.edu
This evidence summary was first published in Pediatrics. Select for
copyright, source, and reprint information.
Contents
Abstract
Introduction
Methods
Literature Review and Synthesis
Results
Evidence Synthesis and Conclusions
References
Acknowledgments
Notes
Abstract
Background: In 1996, the United States Preventive Services Task Force (USPSTF) provided recommendations for routine screening of asymptomatic children and pregnant women for elevated blood lead levels. This review updates the evidence for the benefits and harms of screening and intervention for elevated blood lead in asymptomatic children and pregnant women.
Methods: We searched MEDLINE®, reference lists of review articles, and tables of contents of leading pediatric journals for studies published 1995 or later that contained new information about the prevalence, diagnosis, natural course, or treatment of elevated lead levels in asymptomatic children aged one-five years and pregnant women.
Results: The prevalence of elevated blood lead levels among children and women in the United States, like that in the general population, continues to decline sharply, due primarily to marked reductions in environmental exposure, but still varies substantially among different communities and populations.
Similar to the findings in 1996, our searches did not identify direct evidence from controlled studies that screening children for elevated blood lead levels results in improved health outcomes and there was no direct evidence identified from controlled studies that screening improves pregnancy or perinatal outcomes.
No new relevant information regarding the accuracy of screening for lead toxicity was identified during the update and we did not identify evidence that demonstrates that universal screening for blood lead results in better clinical outcomes than targeted screening. Substantial new relevant information regarding the adverse effects of screening and interventions was not identified.
Conclusions: There is no persuasive evidence that screening for elevated lead levels in asymptomatic children will improve clinical outcomes. For those children who are screened and found to have elevated levels, there is conflicting evidence demonstrating the clinical effectiveness of early detection and intervention.
Similarly, there are no controlled trials evaluating screening for elevated lead levels in pregnant women, nor are there sufficient data to construct an adequate chain of evidence demonstrating benefit.
Community-based interventions are likely to be more effective than office-based screening, treatment, and counseling.
Return to Contents
Introduction
In 1996, the USPSTF recommended screening for elevated blood lead levels at age 12 months in all children with identifiable risk factors, and in all children living in communities in which the prevalence of elevated blood lead levels was high or unknown. There was insufficient evidence, however, to recommend a specific community prevalence below which targeted screening could be substituted for universal screening. The USPSTF found insufficient evidence to recommend for or against routine screening for lead exposure in asymptomatic pregnant women. The USPSTF also found insufficient evidence to recommend for or against trying to prevent lead exposure by counseling families to control lead dust by repeated household cleaning, or to optimize caloric, iron, and calcium intake specifically to reduce lead absorption.1
Return to Contents
Methods
Problem Formulation
USPSTF members defined the scope of this update, in cooperation with AHRQ and the Oregon Evidence Based Practice Center (EPC) personnel. The USPSTF's goal for this update was to review the literature published since its 1996 recommendation to identify new evidence addressing the previously-identified gaps in the literature, including the accuracy of risk assessment questionnaires in children with varying blood lead levels, the population prevalence at which to change from targeted screening to universal screening, the effectiveness of interventions to lower lead levels, and cost-effectiveness analyses of lead screening programs. (Go to Appendix 1 and Figure 1 (20 KB) for key questions and analytic framework.)
Return to Contents
Literature Review and Synthesis
We developed literature search strategies and terms for each key question (KQ) and then searched MEDLINE®, CINAHL, and the Cochrane Library, assisted by a EPC reference librarian, to comprehensively update the literature from 1995 to August 2005 that contained new information about the prevalence, diagnosis, natural course, or treatment of elevated lead levels in asymptomatic children ages 1-5 years and pregnant women. The search was supplemented with reference lists of review articles, references from experts in the field, and reports, guidelines, and recommendations from government, non-government, and medical professional organizations.
Inclusion criteria included:
- The study was an original meta-analysis, prospective cohort study, controlled trial, quasi-experimental study with concurrent controls, or case-control study.
- The study was not included in the 1996 review.
- The study was rated at least "fair-quality" using USPSTF criteria (Appendix 2).
Consistent with the scope of USPSTF recommendations, interventions needed to be relevant to primary care and feasible for delivery in primary care or by referral. Interventions were classified as pharmaceutical (chelation), environmental (residential lead paint, dust, or soil abatement), or nutritional. A primary reviewer abstracted relevant information from included studies for each of the intervention categories in KQ 5.
Return to Contents
Results
Key Question 1: Screening in Asymptomatic Children and Pregnant Women
Similar to the 1996 findings, our searches did not identify direct evidence that screening children for elevated blood lead levels improved health outcomes. There was also no direct evidence that screening improves pregnancy or perinatal outcomes.
Key Question 2: Prevalence and Risk
The prevalence of elevated blood lead levels among children and women in the United States, like that in the general population, continues to decline sharply, primarily due to marked reductions in environmental exposure to lead (e.g., gasoline, air, dietary sources, and residential paint). These reductions are largely the result of regulatory interventions at the federal, state, and local levels of government. The prevalence of elevated blood lead levels, however, varies substantially among different communities and populations, and children and pregnant women share many of the same risk factors for lead exposure. Correlates of higher blood lead levels at all ages include minority race/ethnicity, urban residence, low income, low educational attainment, older (pre-1950) housing, home renovation or remodeling, pica, use of ethnic remedies, cosmetics, lead glazed pottery, occupational exposures, and recent immigration. Alcohol use and smoking are known risk factors among pregnant women. (Go to Appendix 3 for a complete discussion.)
Recent observational studies have demonstrated an inverse relationship between historical blood lead levels in children and subsequent measures of behavioral and cognitive performance at blood lead levels of <10 µg/dL. Observational studies of infants provide preliminary data that prenatal blood lead levels <10 µg/dL may be associated with neurodevelopmental delay or impairment. Study design and measurement issues, however, limit interpretation of these studies. Studies also suggest that levels of maternal exposure in this range may be associated with increased risk for spontaneous abortion, hypertension in pregnancy, and adverse effects on fetal growth3 (Appendix 3).
Key Question 3: Accuracy of Screening Tests
Can screening tests accurately detect elevated blood lead levels?
We identified no new relevant information regarding the accuracy of screening for lead toxicity. Readers are referred to the 1996 USPSTF Statement.1 Blood lead testing has largely supplanted protoporphyrin levels as a screening tool because of poor performance of the latter at blood lead levels (BLL's) <25 µg/dL.19
What is the accuracy of using questionnaires (or other tools) for risk factor assessment at various blood lead levels?
In communities where there is a low prevalence of elevated blood lead levels, screening will identify few cases and yield a significant proportion of false-positive tests. Older cross-sectional studies in urban and suburban populations showed that one or more positive responses to five questions (about exposures to deteriorated paint from older or renovated housing, to other lead-poisoned children, or to lead-related hobbies or industry) detected 64-87% of children with blood lead levels >10 µg/dL.1 Higher sensitivities (81-100%) for blood lead levels >15-20 µg/dL were reported,1 but none of these studies evaluated the ability of questionnaires to detect levels above 20 µg/dL, in part because so few patients had levels so high. Specificity among the studies ranged from 32% to 75%. False negative results were predictably low (0.2-3.5%) in low-prevalence (2-7%) samples, but increased to 19% when the population prevalence of elevated lead levels was higher (17-28%). Questionnaires, therefore, may have greater utility in identifying children at low risk of elevated blood lead (negative predictive value) where the population prevalence is low and local risk factors are known. Negative predictive values of 96–100% have been reported in these settings.1,51
More recent studies of questionnaires in urban and rural settings, however, demonstrated a low prevalence of elevated blood lead levels and poor sensitivity and specificity.52-55 Studies of questionnaires modified for local use provide some evidence of improved clinical utility for identifying children with elevated blood lead levels,55-57 when compared to the panel of screening questions recommended by the Centers for Disease Control and Prevention (CDC) in 1991.108
Other studies have reported high false-positive rates for questionnaires53,55 and that resource considerations52 are important when formulating a screening program. A population-based followup study (n=31904) showed that raising the action level for screening to 15 µg/dL in this sample would have eliminated the unnecessary followup of 5162 children, 3360 of whom were falsely identified as having elevated lead levels.58
A recent study identified housing risk factors associated with elevated blood lead levels (>10 mcg/dL) among 481 children residing in Rochester, New York. Housing characteristics including rental status, lead-contaminated floor dust, and poor housing condition were all associated with elevated blood lead levels (EBLL) (sensitivity 47-92%, specificity 28-76%, positive predictive value 25-34%, negative predictive value 85-93%), suggesting that housing characteristics and floor dust lead levels can be used to identify homes where a lead hazard may exist before or during occupancy.59
Prenatal screening with questionnaires
A maternal survey using four questions recommended by the CDC was evaluated in a study of 314 new prenatal patients. The prevalence of elevated maternal lead levels (at or greater than 10 µg/dL or 0.483 µmol/L) was 13%. Subjects with a positive response to at least one question were more likely to have elevated blood lead than those who answered negatively to all four questions (relative risk = 2.39, 95% confidence interval 1.17-4.89; P = .01). The CDC questionnaire had a sensitivity of 75.7%. Among women who answered "No" to all 4 questions, the probability of having an elevated lead level was reduced from 13% to 6.9% (negative predictive value of 93.1%). The most predictive single item was "home built before 1960." The study also identified a high prevalence of elevated blood lead among children living with women with elevated blood lead levels.6
Key Questions 5: Effectiveness of Early Detection
Detecting elevated blood lead levels before the development of clinical manifestations allows a clinician to recommend interventions to limit further exposure and, when necessary, begin medical treatment with chelating agents. Early detection may also result in interventions that prevent lead exposure in other children (the child with elevated blood lead level acting as a sentinel for a hazardous environment). There is relatively little convincing evidence, however, that these interventions effectively improve health outcomes. First, most available studies in asymptomatic children evaluate the effects of various interventions on blood lead levels rather than on clinical outcomes. Second, blood lead levels in childhood, after peaking at about two years of age, decrease without intervention 1,5 a result attributable in part to regression to the mean, random variation, laboratory error, and redistribution of lead from blood to other tissue compartments. Studies must account for these changes over time, preferably by using controls who do not receive the intervention, to adequately evaluate the interventions' effects on blood lead levels or health outcomes.
Effect of screening on clinical outcomes
EPC staff did not identify evidence demonstrating that universal screening for blood lead results in better clinical outcomes. The 1996 USPSTF recommendation cited several older studies that reported intensive screening programs targeting children in high-risk neighborhoods reduced case fatality rates, mortality rates, and proportions of children detected with very high blood lead levels or who developed symptomatic lead poisoning.1 Lacking concurrent controls, however, it was possible that the reported reductions in mortality and case fatality rates were due to other factors, such as advances in medical care, rather than the effect of screening. The reduction in mean blood lead levels in the US population is primarily the result of diminishing exposure in the environment through regulatory interventions. The available evidence regarding the efficacy of screening programs, therefore, is weak.
Do interventions for elevated lead levels result in improved health outcomes?
While chelating agents benefit children with symptomatic lead poisoning, no studies have demonstrated clinical benefits of chelation therapy in asymptomatic children. The Treatment of Lead-Exposed Children (TLC) Trial, a large multicenter randomized controlled trial sponsored by the U.S. National Institute for Environmental Health Science (NIEHS), enrolled children from 1994-97 to assess the effect of oral chelation therapy with succimer on IQ in young children with venous blood lead concentrations of 20-45 µg/dL.60 followup testing at 36 months demonstrated a mean IQ one point lower, and poorer parental ratings of behavior, among the succimer group, compared to placebo. Although succimer treated children did slightly better on a test of learning ability, none of the differences between groups were statistically significant.61 Reanalysis of the same data using the change in blood lead level as the independent variable demonstrated a 4.0 point improvement in cognitive scores for every 10 µg/dL reduction in blood lead level, but only in the placebo group, suggesting that factors other than declining blood lead contributed to cognitive improvement, or that treatment had an adverse effect on cognitive performance.62 Assessment of neurobehavioral outcomes at seven years of age revealed no statistically significant differences on a battery of neurobehavioral tests except that the succimer group had worse attention-executive function scores.63 Treatment also appeared to have an adverse effect on mean height.64 The TLC Group concluded that chelation therapy was not indicated for children with blood lead levels <45 µg/dL.61,63
Despite evidence of efficacy in lowering blood lead on a short-term basis, there is little evidence confirming a clinical benefit from chelation therapy for children with lead levels <45 µg/dL.
We found no studies evaluating clinical outcomes after environmental or nutritional interventions.
Effects of chelation therapy on blood lead levels
In the previously cited NIEHS-sponsored randomized controlled trial (RCT) of oral chelation in young children with venous blood lead concentrations of 20-45 µg/dL (TLC Study) reporting no effects of chelation on IQ60-63,65 (Tables 1 and 2), blood lead levels fell steeply in the treatment group in the first week (mean 11 µg/dL lower), but rebounded after. Blood lead levels also dropped in the placebo group but more slowly. Blood lead levels were 77% of baseline in the succimer group (88% of baseline among placebo) at seven weeks after initiation of therapy. Mean blood lead levels among the treatment group were 4.5 µg/dL and 2.7 µg/dL, at six and twelve months respectively, but the difference between treatment and placebo groups at 24 months was not significant.65
Chelating agents have demonstrated short-term reductions in blood lead levels in children whose pretreatment values ranged from 20 to 70 µg/dL in studies where chelation therapy was often combined with environmental interventions, but these reductions were not sustained over longer periods in the absence of repeated or continuing chelation therapy or environmental interventions.1,66-68
These data provide good evidence that chelating agents may result in short-term reductions in blood lead levels in children, but suggest that these reductions may not be sustained over longer periods in the absence of repeated or continuing chelation therapy or environmental interventions. Further, there is no evidence that these reductions result in improved neurobehavioral or health outcomes.
Effect of residential lead hazard control on blood lead levels
Recent studies of household dust and paint hazard control through cleaning, abatement, and education have mixed results. Of the eight controlled studies published since 1995, one has shown a modest, but significant, decline, five have shown non-significant declines, and two have shown non-significant elevations in blood lead levels among children. Reduced blood lead levels were seen among children with higher baseline lead levels (15+ or 20+ µg/dL) in two studies (one meta-analysis, one retrospective chart review with no comparison group), but not in children with lower baseline levels. Recent studies differ from older studies in that newer paint hazard control techniques result in lower lead-dust levels. Population venous lead levels have decreased over time, and lead-poisoned children in older studies had higher mean blood lead levels than in recent studies. (Go to Tables 3 and 4 and Appendix 4 for a detailed assessment.)
Effect of counseling and education interventions on blood lead levels
Overall, the evidence to determine whether education and counseling improve outcomes among children with moderately elevated blood lead levels is weak and conflicting (go to Appendix 5 for a detailed assessment).
Effect of soil abatement on blood lead levels
Recent studies of soil remediation in residential areas have shown only modest or non-significant effects.80,85,86 Soil remediation in communities near lead mining, milling, or smelting operations may have a beneficial effect, but was not considered within the scope of review (Go to Appendix 6 for a detailed assessment).
Effect of nutritional interventions on blood lead levels
There is conflicting evidence whether nutritional interventions are an efficacious way to lower children's blood lead levels. Depending on the nutritional intervention under investigation, findings are limited, preliminary, and somewhat contradictory (Go to Tables 5 and 6 and Appendix 7 for a detailed assessment).
Key Questions 4 and 6: Adverse Effects of Screening and Intervention
We identified no substantial new relevant information regarding the adverse effects of screening and interventions for lead toxicity. The most common adverse effects of screening for elevated lead levels remain those identified in the 1996 USPSTF Statement1 (i.e., false-positive results and the associated anxiety, inconvenience, work or school absenteeism, and financial costs of return visits and repeat tests). Adverse effects of environmental interventions may include transient elevation in blood levels, inconvenience associated with abatement work or relocation, and cost-benefit considerations.
Reported adverse effects of treatment with succimer (meso-2, 3-dimercaptosuccinic acid, or DMSA) include mild gastrointestinal (vomiting and diarrhea) and systemic symptoms, rashes, transient hyperphosphatasemia, neutropenia, eosinophilia, and elevations in serum transaminases. These effects occurred in up to 10% of cases.1,60-63, 65
Return to Contents
Evidence Synthesis and Conclusions
There is no direct evidence that screening for elevated lead levels in asymptomatic children at increased risk for lead exposure will improve clinical outcomes (Table 7). Because there have been no controlled trials directly evaluating screening for elevated lead levels, this conclusion is based on a chain of evidence constructed from studies of weaker design. First, in young asymptomatic children, blood lead levels as low as 10 µg/dL, and perhaps lower, are associated with measurable neurodevelopmental dysfunction. Therefore, a relevant threshold level for screening and subsequent intervention cannot be specified based on clinical evidence. Second, the national prevalence of elevated lead levels has declined dramatically in the past two decades, although high prevalence persists in some communities, particularly poor urban communities in the Northeast and Midwest. Third, although current interventions (e.g., residential lead hazard control and chelation therapy) can reduce blood lead levels in children identified with levels >25 µg/dL, the quality of evidence supporting their effectiveness is weak and a beneficial effect on IQ or other clinical outcomes has not yet been demonstrated. Further, well-designed, randomized controlled trials do not support beneficial effects and suggest adverse effects of chelation therapy for asymptomatic children with levels <45 µg/dL.
For those children who are screened and found to have initial blood lead levels <25 µg/dL, there is no evidence regarding the effectiveness of early detection and intervention, or of repeated screening to detect further increases in blood lead. Longitudinal and cross-sectional studies suggest that in children >2 years, such levels will decline naturally with time, but elevated levels may persist in children who are chronically exposed.
There is no direct evidence comparing the outcomes of universal screening with the outcomes from targeted screening for elevated lead levels. Recent studies indicate that the prevalence of elevated blood levels in the U.S. has declined dramatically in the past two decades, but local prevalence is highly variable, with more than 10-fold differences between communities. In a community with a low prevalence of elevated blood lead levels, universal screening may result in disproportionate risks and costs relative to benefits. The prevalence level at which targeted screening can replace universal screening is a public health policy decision requiring consideration beyond the scientific evidence for effectiveness of early detection, such as available resources, competing public health needs, and costs and availability of alternative approaches to reducing lead exposure. Clinicians can consult their local or state health departments regarding appropriate screening policy for their populations. (Go to Appendix 8 for recommendations of other groups.)
In communities where data suggest that universal screening is not indicated, there may be some children who are at increased risk of blood lead levels in the range for which individual intervention by chelation therapy or residential lead hazard control has been demonstrated to be effective. In addition to risks from housing, these children may have had exposure to other lead sources such as lead-based hobbies or industries, traditional ethnic remedies, or lead-based pottery. Selective blood lead screening of such high-risk children is appropriate even in low-prevalence communities.
Questionnaires that have been locally validated and are of known and acceptable sensitivity and specificity can assist in identifying those at high risk. In several studies, the CDC108 and similar questionnaires correctly identified 64% to 87% of urban and suburban children who had blood lead levels >10 µg/dL. Because of frequent false positives in low prevalence communities, questionnaires may have greater utility in identifying children at low risk of elevated blood lead (negative predictive value) where the population prevalence is low, and local risk factors are known. Locale-specific questionnaires inquiring about likely local sources of lead exposure may lead to improved prediction.
There are no controlled trials evaluating screening for elevated lead levels in pregnant women, and there are insufficient data to construct an adequate chain of evidence demonstrating benefit. The prevalence of levels >15 µg/dL appears to be quite low in pregnant women. There is some evidence that mildly-elevated lead levels during pregnancy are associated with small increases in antepartum blood pressure, but only limited evidence that these levels have important adverse effects on reproductive outcomes. An extensive literature search failed to identify studies evaluating screening or intervention for lead exposure in pregnant women. There are potentially important adverse effects of chelation therapy on the fetus and of residential lead hazard control on both the pregnant woman and fetus if they are not performed according to established standards. While removal to a lead-free environment would theoretically be effective in reducing lead exposure, it has not been specifically evaluated in pregnancy.
Community-based interventions for the primary prevention of lead exposure are likely to be more effective, and may be more cost-effective, than office-based screening, treatment, and counseling.21 Evaluating the effectiveness of community-based interventions, and recommendations regarding their use, are important areas of future research.
Return to Contents
Proceed to Next Section
540 Gaither Road Rockville, MD 20850