NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Cantor AG, Hendrickson R, Blazina I, et al. Screening for Elevated Blood Lead Levels in Children: A Systematic Review for the U.S. Preventive Services Task Force [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2019 Apr. (Evidence Synthesis, No. 174.)

Cover of Screening for Elevated Blood Lead Levels in Children: A Systematic Review for the U.S. Preventive Services Task Force

Screening for Elevated Blood Lead Levels in Children: A Systematic Review for the U.S. Preventive Services Task Force [Internet].

Show details

Chapter 3Results

The search and selection of articles are summarized in the literature flow diagram. Two reviewers independently identified 3,147 unique citations and 233 full-text articles based on predefined criteria (Appendix A2). A total of 21 studies met inclusion criteria for this review (N=10,449). Appendix A3 shows the results of the literature search and selection process, Appendix A4 lists the included studies, and Appendix A5 lists the excluded full-text papers.

Key Question 1. Is There Direct Evidence That Screening for Elevated BLLs in Asymptomatic Children Age 5 Years and Younger Improves Health Outcomes?

As in the prior USPSTF review, no studies directly compared the effectiveness of screening versus no screening for elevated BLLs in children age 5 years and younger on health outcomes.

Key Question 2a. What Is the Accuracy of Questionnaires or Clinical Prediction Tools That Identify Children Who Have Elevated BLLs?

Summary

Nine fair-quality studies (six included in the prior USPSTF report) reported the diagnostic accuracy of questionnaires or clinical prediction tools for identifying asymptomatic children with elevated BLLs, defined as a BLL greater than 10 µg/dL.4351 All studies used a BLL greater than 10 µg/dL as the reference standard. Five fair-quality studies that used the threshold of one or more positive answers on the five-item 1991 CDC screening questionnaire reported a pooled sensitivity of 48 percent (95% CI, 31.4% to 65.6%) and specificity of 58 percent (95% CI, 39.9% to 74.0%) for identifying children with a venous BLL of 10 µg/dL or greater.

Four fair-quality studies that used versions of the CDC questionnaire modified for specific populations or settings did not demonstrate improved accuracy (sensitivity range, 25% to 68%; specificity range, 49% to 58%).

Evidence

The prior USPSTF review1 found fair evidence that a validated questionnaire can correctly identify 64 to 87 percent of children at high risk in urban and suburban populations with BLLs of 10 µg/dL or greater. However, eight of the studies in the prior review did not meet criteria for this update and were excluded due to having the wrong comparison or reference standard.5259 The prior report also found fair evidence that a validated questionnaire had not been adequately evaluated as a screening tool to detect higher BLLs (e.g., ≥20 to 25 µg/dL) or lead exposure in specific populations (e.g., migrant workers, rural communities). Five studies from the prior review on accuracy of screening instruments met inclusion criteria for this update.45,47,48,50,51 Four additional studies were identified for this update.43,44,46,49

Nine studies reported on the diagnostic accuracy of questionnaires or clinical prediction tools for identification of children with elevated BLLs (Appendixes B1 and C1).4351, Five studies evaluated the accuracy of the 1991 CDC questionnaire and four evaluated versions of the CDC questionnaires modified for specific populations and settings.4351 The CDC questionnaire is a five-question survey developed in 1991 that aims to assess residential, household, and personal risk factors for lead exposure in children. Specific items include the age of the child’s housing and the condition of the paint; siblings or playmates with BLLs of 15 µg/dL or greater; parental exposure through work or hobbies; and a home in close proximity to lead industry. Sample sizes ranged from 167 to 2,978 (total N=6,873). Mean age was not reported in six studies, was reported as 9 months in one study,43 and reported as 28 and 31 months in two other studies.47,49 Females comprised 46 to 51 percent of participants in five studies and sex was not reported in the other five. Seven studies were conducted in urban or suburban communities and three studies were conducted in rural communities. Two of the studies identified their population as high risk44,46 and others did not characterize study populations by risk level; however, many of the populations surveyed were from public programs such as Medicaid or public health clinics. In all studies, children were reported as asymptomatic. The prevalence of children with a BLL of 10 µg/dL or greater ranged from 2.2 percent47 to 29 percent.43 In study populations characterized as higher risk, the prevalence of an elevated BLL of 10 µg/dL or greater ranged from 7.7 to 22 percent.44,46 Nine studies were rated as fair quality. One poor-quality, retrospective study was excluded from this analysis.60 Methodologic shortcomings included unclear enrollment methods and exclusion of some patients from analysis (Table 3). The poor-quality study performed retrospective surveys of exposures after BLL was known.

Table 3. Characteristics and Results for Studies of Screening Questionnaires.

Table 3

Characteristics and Results for Studies of Screening Questionnaires.

Five fair-quality, cross-sectional studies (total N=2,265) conducted in mostly urban4346 and one rural U.S. community (n=368)47 evaluated the diagnostic accuracy of the 1991 CDC questionnaire3 for identification of children with venous BLLs of 10 µg/dL or greater. The studies used a threshold of one or more positive answers from the five-question survey to indicate a positive screen. Across studies, sensitivity ranged from 32 to 83 percent and specificity ranged from 32 to 80 percent, with a pooled sensitivity of 48 percent (95% CI, 31.4% to 65.6%) and pooled specificity of 58 percent (95% CI, 39.9% to 74.0%) (Figure 2).4347 The positive likelihood ratio was 1.15 and the negative likelihood ratio was 0.89, indicating that either a positive or negative screen had little effect on informing the likelihood of elevated BLLs.

Figure 2 is a forest plot reflecting meta-analysis of 5 studies. The combined sensitivity of the five studies was 0.48 (95% CI, 0.31 to 0.66) and combined specificity was 0.58 (95% CI, 0.39 to 0.74).

Figure 2

Sensitivity and Specificity of CDC Screening Questionnaire (>1 Positive Answers and >10 μg/dL Venous BLL). Abbreviations: BLL=blood lead level; CDC=Centers for Disease Control and Prevention.

Four diagnostic accuracy studies4851 evaluated a modified 1991 CDC questionnaire by changing some of the language in the CDC questions3 or expanding the CDC questionnaire by adding additional questions to address local risk factors to adapt the questionnaire for use in specific study populations. One study conducted in a low-income, inner city population (n=2,978) found that the adapted questionnaire had low accuracy for identifying children with elevated BLLs (sensitivity, 57%; specificity, 51%).48 Another study (n=705) conducted in a rural setting51 used two items from the CDC questionnaire and two additional items for rural community risk factors and found limited benefit in detecting rural children at higher risk. Compared with the CDC questionnaire, there was a 12-percent increase in sensitivity for identifying children with BLLs of 10 µg/dL or greater (75% vs. 88%) and a 5-percent increase in negative predictive values (93% vs. 98%) using the modified questionnaire. A smaller study (n=171) conducted in rural New York50 that added six items to the CDC questionnaire found no difference compared with the standard CDC questionnaire for predicting elevated BLLs (sensitivity, 50% vs. 50%). Another study conducted in an urban population (n=754)49 with a 3.1 percent prevalence of a BLL of 10 µg/dL or greater found that adding two items to the CDC questionnaire did not increase accuracy for detection of children with elevated BLLs.

Key Question 2b. What Is the Accuracy of Capillary Blood Lead Testing in Children?

Summary

Four fair-quality studies conducted in the urban United States27,6163 found that capillary blood lead testing was associated with sensitivity of 87 to 91 percent and specificity greater than 90 percent (92% to 99%) for identification of elevated BLL compared with venous sampling; two of the studies were included in the prior USPSTF review.

Evidence

The prior USPSTF report included two studies that compared the accuracy of capillary versus venous blood lead testing.27,63 We identified four fair-quality cohort studies assessing the diagnostic accuracy of capillary testing compared with venous sampling for elevated BLLs,27,6163 including the two studies in the prior report (Appendixes B2 and C1).27,63 All four studies were conducted in the urban United States and were published between 1994 and 1998. Sample sizes ranged from 124 to 513 participants (total N=1,431). The mean age was 3 years in one study63 and was not reported in the other studies. Females comprised 41 to 47 percent of the sample in three studies; the fourth study did not report sex. Two studies predominately enrolled black children,61,63 and one study evaluated a more diverse study population (38% white, 28% black, 21% Hispanic, and 6% Asian27); the fourth study did not report race/ethnicity.62 Among the three studies that reported baseline BLLs, the proportion of children with a BLL of 10 µg/dL or greater ranged from 21 to 31 percent.27,61,62 Methodologic shortcomings of the studies included unclear methods of patient enrollment and exclusion of some patients from analysis.

At a BLL cutoff of 10 µg/dL or greater in capillary sampling, three studies reported sensitivities ranging from 87 to 94 percent, and specificities ranging from 92 to 99 percent (N=1,136).27,61,62 For a BLL cutoff of 15 µg/dL or greater, three studies reported sensitivities ranging from 36 to 83 percent and specificities from 95 to 98 percent.27,61,62 For a BLL cutoff of 20 µg/dL or greater, three studies reported sensitivities ranging from 78 to 96 percent and specificities from 91 to 100 percent (N=918).27,61,63

One study evaluated different preparation methods for capillary blood sampling (N=295)63 (alcohol wipe; alcohol wipe and silicone barrier; soap and water followed by alcohol wipe; or soap and water, alcohol wipe, and 1% nitric acid solution). Using a capillary sampling threshold of greater than 20 µg/dL, the most commonly employed sampling method (i.e., soap and water plus alcohol) had the highest specificity (100%) and similar sensitivity (88%) compared with the other methods (sensitivity, 86% to 96%; specificity, 91% to 96%).

Key Question 3. What Are the Harms of Screening for Elevated BLLs (With or Without Screening Questionnaires) in Children?

As in the prior USPTF report, no studies evaluated the harms of screening versus not screening for elevated BLLs in children.

Key Question 4. Do Counseling and Nutritional Interventions, Residential Lead Hazard Control Techniques, or Chelation Therapy Reduce BLLs in Asymptomatic Children With Elevated BLLs?

Summary

One large, good-quality RCT included in the prior USPSTF review found that chelation therapy with DMSA in children with a mean blood lead concentration of 20 to 45 µg/dL was associated with decreased blood lead concentrations versus placebo at 1 week, 6 months, and 1 year, but there were no effects at longer-term followup at 4.5 to 6 years.6467 One fair-quality RCT included in the prior USPSTF review found no differences between chelation therapy versus placebo in blood lead concentration at 1 or 6 months.68

There was insufficient evidence from two poor-quality studies to determine effects of nutritional supplementation on BLLs. Three fair-quality RCTs from the United States and Australia (all included in the prior USPSTF review) found no clear effects of home lead remediation in lowering blood lead concentrations.

Evidence

The prior USPSTF review found that chelating agents may result in short-term reductions in blood lead concentrations in children but that reductions may not be sustained over longer periods in the absence of repeated or continuing chelation therapy or environmental interventions. Effects of cleaning, abatement, and education on blood lead concentrations were mixed, based on a descriptive summary of 11 studies. The prior USPSTF review also found conflicting evidence on the effects of nutritional interventional on elevated BLLs, based on a descriptive summary of 16 studies.

Seven RCTs6473 (reported in 10 publications) evaluated the effects of interventions to reduce blood lead concentrations in asymptomatic children with elevated BLLs (Appendixes B3 and C2); four of the studies were included in the prior USPSTF review.6467,71,68 Two studies evaluated chelation therapy,6468 two studies evaluated counseling and nutritional interventions,71,72 and three studies evaluated residential lead hazard control techniques.69,70,73 Sample sizes ranged from 39 to 780 (total N=1,419). Five studies were conducted in the United States and one study each in Australia and Costa Rica. The mean age of study participants was 1.6 to 3.6 years and sex distribution was balanced in studies that provided this information (44% to 58% female). One study was rated good quality, four fair quality, and two poor quality. The poor-quality studies lacked descriptions of randomization methodology, allocation concealment, and masking, and one study had poor followup; the poor-quality studies were included because no fair- or good-quality studies were available.

Chelation

One fair-68 and one good-quality6467 trial found inconsistent effects of DMSA chelation therapy on blood lead concentrations in asymptomatic children with BLLs of 20 to 45 µg/dL at baseline.6468 Although the good-quality trial found that chelation therapy was associated with lower blood lead concentrations versus placebo at 1 week, 6 months, and 1 year, it found no differences at 4 to 5.6 years. The fair-quality trial found no effect of chelation therapy on BLLs at 1 or 6 months. Both trials were included in the prior report.

The Treatment of Lead-Exposed Children (TLC) study, a good-quality RCT (n=780), evaluated 12- to 33-month-old children with blood lead concentration between 20 and 44 µg/dL.6467 All children received vitamin and mineral supplements and had home inspections with lead abatement. Children were randomized to treatment with DMSA (1,050 mg/m2 per day for 7 days, then 700 mg/m2 for 19 days) or placebo. Children could be treated with DMSA up to three times, with a goal blood lead concentration of less than 15 µg/dL. DMSA was associated with a mean difference in blood lead concentration at 1 week that was 11 µg/dL lower than placebo. However, blood lead concentrations increased once treatment was completed, and at 52 weeks the mean difference had decreased to 2.7 µg/dL in favor of DMSA (95% CI, 1.9 to 3.5 µg/dL).67 In a followup study of 7-year-old participants (approximately 4.5 to 6 years after treatment), mean blood lead concentrations were identical in both groups (8.0 µg/dL).65

A small, fair-quality study (n=39)68 randomized children ages 2.5 to 5 years with blood lead concentrations between 30 and 45 µg/dL to one course of DMSA or control. DMSA was dosed according to weight (≤15 kg, 100 mg dose; >15 kg, 200 mg dose), and each dose was administered three times a day for 5 days followed by twice a day for 14 days. There were no significant differences in mean blood lead concentrations at 1 month (27.4 µg/dL [standard deviation (SD), 7.5] vs. 33.2 µg/dL [SD, 10.3]; p=0.16) or at 6 months (28.8 µg/dL [SD, 6.4] vs. 25.1 µg/dL [SD, 6.8]; p=0.06).

Nutritional Interventions

Two poor-quality studies provided insufficient evidence to determine the effects of nutritional interventions on blood lead concentrations.71,72 One double-blind, placebo-controlled trial conducted in New York City (n=88) that was included in the prior review evaluated the effects of calcium supplementation on blood lead concentrations but had high attrition (34%) and inadequate descriptions of randomization, allocation concealment, and masking techniques.71 The other study evaluated effects of iron supplementation in Costa Rican children72 with elevated blood lead concentrations (mean, 10.98 µg/dL) at baseline. Results were difficult to interpret because iron supplementation was given to children who were iron depleted and placebo was given to children who were iron replete, with no matching on blood lead concentrations. Children were randomized to either intramuscular iron or oral iron. Iron was associated with a decrease in blood lead concentration in iron-deplete children and placebo was associated with slightly increased BLLs in iron-replete children, but it is unclear how baseline iron levels may have affected blood lead concentrations independent of iron supplementation. Another limitation of the trial is that results were reported for the subgroup of patients in the iron-deplete group who received oral iron, but not for those who received intramuscular iron.

Residential Lead Hazard Control Techniques

Three fair-quality RCTs found no clear effects of home lead abatement in lowering blood lead concentrations in asymptomatic children with elevated BLLs at baseline.69,70,73 None of the studies were included in the prior review and home lead abatement interventions differed in each trial.

One trial (n=175) randomized children younger than age 28 months in Rhode Island with blood lead concentrations of 15 to 19 µg/dL70 to a home intervention (five home visits that included testing samples, tailored education, and assessment of nutrition and parent-child interaction plus lead remediation strategies) or control intervention (one to two standard educational visits from an outreach worker). Blood lead concentrations in both groups decreased, but there was no significant difference between the intervention and control groups at 3, 6, or 12 months after baseline.

Another fair-quality trial (n=90)69 conducted in Australia randomized pairs of 12- to 60-month-old children with mean blood lead concentrations between 15 and 30 µg/dL matched by age and BLL to home remediation and lead abatement versus delayed intervention for 1 year. Despite reductions in home lead concentrations after intervention, the effects of remediation on mean BLL were small (17.5 vs. 17.9 µg/dL; mean change, 1% [95% CI, -11% to 11%]), with no significant difference between groups.

A fair-quality trial (n=84)73 conducted in Florida enrolled asymptomatic children from the WIC and Head Start programs and the local health department with blood lead concentrations of 3 to 10 µg/dL (mean, 5.29 µg/dL [range, 3.0 to 9.3 µg/dL]). Participants were randomized to receive an educational brochure, a home cleaning kit, a formal home inspection and remediation, or passive control. The educational brochure included information about diet, cleaning, and habits to reduce lead exposure. The home cleaning kit included a HEPA (high-efficiency particulate air) vacuum, trisodium phosphate detergent, gloves, rags, and buckets. The formal inspection/remediation group received a home risk assessment by a professional company that included dust wipe samples that were evaluated with on-site X-ray fluorescence spectrometry and laboratory testing. The inspection was followed by a second home visit and a written report with a range of optional steps on how to decrease lead exposure. The passive control group received no intervention or information. All groups experienced a decrease in blood lead concentration of 2.26 to 2.99 µg/dL over 6 to 12 months, with no significant difference between groups.

Key Question 5. Do Counseling And Nutritional Interventions, Residential Lead Hazard Control Techniques, or Chelation Therapy Improve Health Outcomes in Asymptomatic Children With Elevated BLLs?

Summary

One good-quality randomized study included in the prior USPSTF review found no differences between chelation therapy versus placebo on neuropsychological outcomes despite a decrease in blood lead concentrations following chelation therapy.6567

There was no evidence on effects of counseling and nutritional interventions or residential lead hazard control techniques on health outcomes in asymptomatic children with elevated blood lead concentrations at baseline.

Evidence

The prior USPSTF review found no clear evidence to support a clinical benefit from chelation therapy in children with elevated blood lead concentrations at baseline, based on one trial,6567 and found no studies on effects of environmental or nutritional interventions on health outcomes.

The TLC6567 trial (N=780) of DMSA chelation therapy versus placebo (see Key Question 4 for study details), included in the prior USPSTF review, was the only study to evaluate the effect of interventions for lowering elevated blood lead concentrations on health outcomes in children by measuring neuropsychological outcomes. At 36 months, there were no significant differences between chelation therapy and placebo in the Wechsler Preschool and Primary Scale of Intelligence-Revised, the Developmental Neuropsychological Assessment, or the Conners’ Parent Rating Scale-Revised. In a followup study65 of the same children at age 7 years (4.5 to 6 years after treatment), chelation therapy was associated with lower (worse) scores on the adjusted Attention and Executive Functions subscore of the Developmental Neuropsychological Assessment (unadjusted difference, −1.8 [95% CI, −4.5 to 1.0]; adjusted p=0.045). There were no statistically significant effects on any other cognitive, neuropsychiatric, or behavioral outcome.

We identified no new study on effects of chelation therapy, environmental interventions, or nutritional interventions on health outcomes. Evidence on the effects of interventions for lowering blood lead concentrations on health outcomes remains very limited.

Key Question 6. What Are the Harms of Interventions in Asymptomatic Children With Elevated BLLs?

Summary

One good-quality RCT64,67 and one poor-quality observational study74 reported adverse effects of chelation therapy. The good-quality RCT found that children treated with DMSA had a small but statistically significant decrease in height growth over 34 months and slightly poorer scores on attention and executive function tests at age 7 years (Appendixes B3 and C2).65

The poor-quality study reported adverse events associated with the less commonly used chelator d-penicillamine, including leukopenia, thrombocytopenia, urticarial and maculopapular rashes, urinary incontinence, abdominal pain, and diarrhea.74

No study evaluated harms of counseling, nutritional interventions, or residential lead hazard control techniques.

Evidence

The prior USPSTF report found adverse effects of environmental interventions including transient elevation in blood lead concentrations, inconvenience associated with abatement work or relocation, and cost-benefit considerations, but the number of studies on which these narrative findings was based was unclear. It also identified adverse effects after DMSA chelation therapy that included mild gastrointestinal (vomiting and diarrhea) and systemic symptoms, rashes, transient hyperphosphatasemia, neutropenia, eosinophilia, and elevations in serum aminotransferases. Most evidence from the prior report did not meet our inclusion criteria due to study design, lack of comparison group, wrong outcomes, or lack of a reference standard. The prior USPSTF review included data on harms from one good-quality RCT, which was also included in this update.

The TLC trial compared DMSA chelation therapy with placebo in children ages 12 to 33 months with blood lead concentrations between 20 and 44 µg/dL (N=780).67 DMSA was associated with a small but statistically significant decrease in height growth over 34 months (difference of 0.35 cm [95% CI, 0.05 to 0.72 cm]) and slightly poorer scores on attention and executive function tests (unadjusted difference of −1.8; adjusted effect P=0.045) at age 7 years. There were no significant differences in laboratory values, including neutrophil count, platelet count, aminotransferase concentrations, and alkaline phosphatase concentration.64,67 Children treated with DMSA were more likely to have evidence of minor traumatic injuries on physical examination (14.9% vs. 9.9%).64 However, a mechanism for this association is not known or theorized.

A poor-quality retrospective cohort study (n=75) evaluated d-penicillamine in children with blood lead concentration of 25 to 40 µg/dL.74 Twenty-nine adverse events were reported in 37 percent of study participants, including leukopenia (11%; white blood cell count <4,000/mm3), rash (9%), low platelet count (9%; <300/mm3), enuresis (4%), abdominal pain (3%), and hematuria (1%) (Appendix C3). No study identified harms of counseling, nutritional interventions, or residential lead hazard control techniques.

Contextual Question 1. What Is the Reliability of Capillary and Venous BLL Testing at Various Lead Levels in Children?

Understanding whether current methods for testing for elevated BLLs is reliable would be helpful for confirming that a standard, predictable measure of blood lead exists and for informing testing strategies. We sought evidence to determine whether children are consistently classified as having elevated BLL at standard thresholds and whether tests perform reliably between laboratories and between patients across the minimum or standard threshold of BLLs. However, we found no studies on these aspects of reliability of BLL testing in children.

Contextual Question 2. What Is the Association Between Reduced BLLs and Improved Health Outcomes in Asymptomatic Children With Elevated BLLs?

One good-quality randomized study (in four publications) addressed the association between reduced BLLs and improved health outcomes in children with elevated BLLs. The previously described TLC study of chelation therapy with DMSA6567 (n=780) found an inverse relationship between cognitive test scores and changes in blood level concentration, with a decrease in cognitive test scores of 3.2 to 3.3 points for every 10-µg/dL increase in BLL. However, there was no correlation between short-term decreases in blood lead concentration and long-term cognitive test scores in the DMSA group compared with placebo.66

Contextual Question 3. Are There Valid Risk Prediction Tools Available That Identify Communities at Highest Risk for Lead Exposure That Could Be Used in Primary Care Practices to Target Screening Efforts in Children?

We identified no studies on the accuracy of community-level risk prediction tools for use in primary care screening to identify children at highest risk for lead exposure. Risk assessment tools for individuals are addressed in Key Question 1.

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.6M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...