Blood Lead Levels Among Resettled Refugee Children in Select US States, 2010–2014

State and local refugee health programs provided the CDC with BLL test results and demographic information routinely collected during the domestic refugee health examinations of children aged 6 months to 16 years resettled to the United States from January 1, 2010, to September 30, 2014. Some participating partners received funding through the CDC’s CK12-1205 Strengthening Surveillance for Diseases Among Newly-Arrived Immigrants and Refugees cooperative agreement. We preferentially included venous BLL testing results (because capillary BLLs may be subject to contamination from lead traces on fingers¹⁷) but accepted capillary or undocumented specimen results if venous results were unavailable. We defined a valid initial BLL test as a quantitative blood lead value from testing conducted ≤90 days after arrival. If a child had both capillary and venous BLL testing during the first 90 days, we preferentially selected the venous BLL if the date of collection was either before or ≤45 days after the capillary test (Fig 1). Tests administered 91 to 183 days (3–6 months) after the initial selected test were designated valid follow-up tests. Qualitative test results (ie, reported without a numeric value) and results of specimens collected outside the defined time frames were excluded. We were unable to collect data on potential lead exposures.

Refugee health program partners provided demographic data, including sex, age at time of BLL testing, US arrival date, and overseas examination country. When available, height, weight, and hemoglobin results from the initial domestic medical examination were also provided. If sites could not provide complete demographic information but provided a unique identification number, we used this number to extract relevant data from the CDC’s Electronic Disease Notification system.¹⁸ We categorized participants by examination country (location of the overseas immigration medical screening examination and typically where a refugee lived during the 3–6 months before US arrival), because our data on nationality were incomplete and may not reflect where a refugee was exposed to lead. Our reference country was Malaysia because it had the lowest EBLL prevalence among the 5 examination countries with the largest arrival volumes.

We defined EBLL as a BLL of ≥5 µg/dL.⁶ The comparison group for all analyses was children with BLL values <5 µg/dL. We used the month of testing, grouped by quarter (eg, January to March), to evaluate the potential association between the time of year and BLL. We defined moderate to severe anemia as hemoglobin <10 g/dL, regardless of age or sex.¹⁹ We used height and weight measurements to calculate anthropometric z scores using the World Health Organization’s Child Growth Standards SAS igrowup macro package (SAS version 9.4; SAS Institute, Inc, Cary, NC). Stunting was defined as <−2 SD from median height-for-age z score for reference population and wasting as <−2 SD from median weight-for-height or BMI z score for reference population.²⁰ 

We assessed the relationship between EBLL on the initial screening test and demographic characteristics and other covariates in our data set by calculating prevalence ratios (PRs) and 95% confidence intervals (CIs) using generalized estimating equations to account for state-level clustering and both with and without age stratification (comparing children aged <7 years with children aged 7–16 years). We included variables in the adjusted model if they were significantly associated with EBLL in the bivariate model at the .05 α level. In our age-stratified analysis, sex was the only variable with significantly different BLLs, so we included an interaction term for age and sex in our model. We also analyzed the association between nutritional status (ie, stunting, wasting, or severe anemia) and EBLL using the same modeling approach on the subset of our population with available nutritional indicator data. We used the Cochran-Armitage test to assess for trends in EBLL prevalence over the time period covered by our data set. To evaluate changes in BLL after arrival, we calculated the prevalence of a ≥2 µg/dL increase in BLL using a generalized estimating equation and used the sign test to compare the change in median BLL from the initial and follow-up tests among children who had EBLL on both tests. We evaluated the relationship between EBLL on the follow-up test and multiple covariates using the same modeling approach described for the initial testing. All data analysis was performed using SAS 9.4 software. This analysis was determined to be nonresearch surveillance by a CDC human subjects advisor; institutional review board review was not required.

We received valid BLL results for 27 284 resettled refugee children aged 6 months to 16 years old at the time of initial domestic medical examination from 12 sites: 11 states (CO, ID, IL, KY, MA, MN, NC, NY [excluding New York City], TX, UT, and WA) and 1 county health department (Marion County Public Health Department of IN); these data represent nearly a quarter (24.0%) of all refugee arrivals <17 years old to the United States over the time period.²⁰ Girls comprised 49% (n = 13 355) of our data set, and the mean age was 95.6 months (8 years; median: 92 months; interquartile range: 46–142 months) (Table 1). The top 5 overseas examination countries by arrivals were Thailand (n = 5574; 20.7%), Nepal (n = 4117; 15.2%), Malaysia (n = 3431; 12.8%), Iraq (n = 2360; 8.8%), and Kenya (n = 1962; 7.3%). Demographics of children in our data set did not differ from those of the overall population of refugee child arrivals to the United States during the same period by age, sex, arrival year, or country of examination.²⁰ 

Overall, 5275 (19.3%) children had EBLL on initial testing; children <7 years old had the highest prevalence (n = 2836; 22.8%) (Supplemental Table 5). Children ≥7 years old had an EBLL prevalence of 16.5%. There were 579 (2.1%) children with BLL over 10 µg/dL and 6 (0.02%) children with BLL ≥45 µg/dL. The mean time elapsed between arrival in the United States and the initial BLL test was 29.8 days (SD = 20.2 days), and 72.1% of initial tests were venous (Table 1).

Of the 5 examination countries with the largest volume of arrivals, children from Nepal had the highest EBLL prevalence (n = 1128; 27.5%) followed by Thailand (n = 1181; 21.0%) and Iraq (n = 489, 20.7%); Kenya had the highest prevalence of BLL ≥10 µg/dL (3.4%) (Table 1). Independent of arrival volume, EBLL prevalence was highest among refugees from India (57.9%) and Afghanistan (55.1%) (Table 2); the examination country with the highest prevalence of BLL >10 µg/dL was Afghanistan (16.7%). Boys had a higher prevalence of EBLL (22.8%) than that of girls (15.7%; P < .001) (Table 1), but the geometric mean EBLL did not differ between sexes (6.5 µg/dL [95% CI = 6.4–6.6] for both). EBLL prevalence did not differ between sexes for children aged <2 years; in all other age groups, girls had a significantly lower prevalence of EBLL than did boys, and the sex disparity increased with age. EBLL was significantly associated with the time of year during initial testing (highest during July through September; 21.2%; P < .001). EBLL prevalence declined by arrival year, dropping from 24.4% of arrivals in 2010 to 14.4% of arrivals in 2014 (P < .001; Fig 2) (Supplemental Table 5).

The nutritional analysis subset included 8951 children from 8 sites with complete hemoglobin results and height (or length) and weight information; 1697 (19.0%) of these children had EBLL. In unadjusted analyses, EBLL was associated with moderate to severe anemia (PR = 1.4; 95% CI 1.2–1.7; P = .001) and stunting (PR = 1.2; 95% CI 1.1–1.3; P = .001) but not wasting (PR = 1.2; 95% CI 1.0–1.4; P = .09). However, none of these variables remained associated with EBLL after adjustment for age, sex, and time of year.

In the adjusted model (Table 3), prevalence of EBLL among boys and girls was not significantly different in children <2 years of age. In children ages 2 years and older, boys were more likely to have EBLL than girls of the same age, and the disparity increased with age. No other covariates differed by age in age-stratified analysis. Children examined in Nepal, Iraq, Kenya, and Thailand were more likely to have EBLL on follow-up testing than children examined in Malaysia.

Five sites (CO, IL, IN, MN, and NY) provided follow-up testing results on 3532 (13.0%) children. Among children with multiple BLL results, 1121 (31.7%) had valid follow-up tests (Fig 2) of which 76.3% were venous tests. The proportion of children with a valid follow-up test was 5.0% among children <7 years old (22.8% EBLL prevalence) and 3.4% among children ≥7 years old (16.5% EBLL prevalence).

Among 1121 children with valid follow-up BLLs, 183 (16.3%) had EBLL on both initial and follow-up tests, and 71 (6.3%) did not have EBLL initially but had EBLL at follow-up (Fig 3). In total, 117 (10.4%) children experienced a ≥2 µg/dL increase in BLL on the follow-up test. Increases in BLL were most common among children <2 years (20.8% of retested children <2 years), but 5.2% of children 7 years and older experienced a ≥2 µg/dL increase in BLL. Overall, the median BLL declined significantly between initial and follow-up testing (8.0 µg/dL [95% CI = 8.0–8.7] and 7.0 µg/dL; [95% CI = 6.2–7.1], respectively; sign test P < .0001) for children with EBLL on both tests.

A ≥2 µg/dL rise in BLL on the follow-up test was associated with younger age, particularly <2 years old, testing time of year (April through June), and examination country (adjusted model, Table 4). An EBLL result on the follow-up test was associated with younger age (<7 years), earlier arrival years (2010–2011), country of examination (Iraq and Kenya), and time of year (July through September) (adjusted model; data not shown) .

This report expands our knowledge of the prevalence of EBLL among resettled refugee children, using data from a representative sample of children resettled to multiple states. The EBLL prevalence in this population of recently arrived refugee children (23.7% among 1–5-year-olds) was 10-fold higher than the NHANES-estimated prevalence of 2.3% among all 1- to 5-year-old children in the United States from 1999 to 2010.²¹ In addition, up to 10% of children had increases in BLL 3 to 6 months after the initial test, including 5% of children who were >6 years old.

Initial EBLL results likely indicate overseas lead exposures, because testing was conducted within 3 months of arrival. Among 5 overseas sites with the largest arrival volume, EBLL prevalence was highest among children with overseas medical examinations in Nepal (generally children of Bhutanese origin or nationality), Thailand (Burma origin), and Iraq (Iraq origin). Children examined in India, Afghanistan, Burma, and Nepal had the highest overall prevalence of EBLL, and children examined in Afghanistan had the highest prevalence of BLL ≥10 µg/dL, although the relatively small numbers of refugee arrivals from these countries limited our ability to draw definitive conclusions, and EBLL risk may vary by subnational factors such as city or refugee camp of residence, which we were unable to analyze. Our findings may warrant further investigation of EBLL as a common health concern for children from these countries.

Younger age and male sex (regardless of age) were also associated with EBLL at arrival. Although the oldest children (age 12–16 years) had a lower EBLL prevalence (14%) than children aged 2 to 4 (24%), they had a higher prevalence of EBLL than has been reported among the general US adolescent population,^22,23 supporting the CDC’s current recommendations for testing refugee children through age 16. As previously noted,^9,12 initial EBLL prevalence was highest among children tested between July and September.

EBLL prevalence among refugee arrivals declined over the period of our analysis, which could have resulted from many factors, including improved conditions overseas, such as the introduction of electricity in some refugee camps²⁴; phasing out of leaded gasoline in multiple countries in the years leading up to this analysis²⁵; and changes in the countries of origin of resettlement populations. For example, refugees resettling from Thailand may have been exposed to lead poisoning prevention messaging as part of an overseas educational campaign on lead early in the analysis period, although our analysis was unable to evaluate the effect of this education. Although EBLL declined overall, we noted upward trends in Thailand and Kenya in 2014; further investigation would be needed to identify contributors.

We identified an overall decline in median BLL among children who received a follow-up test, although a minority experienced postresettlement BLL increases, which may indicate domestic lead exposures. We had no information on specific lead exposures, but other investigations of EBLL among refugee children have identified associations between postresettlement increases in BLL and anemia, parasitic infections, pre-1950s housing,⁹ and imported infant remedies and cosmetics,⁷ for example.

There were limitations to this analysis. First, reporting differences between sites lead to the exclusion of some data from analysis (eg, qualitative results). Some sites did not provide height, weight, and hemoglobin data. Most sites were unable to provide the laboratory limit of detection data, so we could not calculate mean BLLs for children with results <5 µg/dL. Most sites were unable to provide follow-up BLL data, usually because refugee health programs only have access to initial postarrival screening data, and follow-up testing is performed outside the refugee health program. Accordingly, children in our follow-up data set had a higher initial prevalence of EBLL (30.8%) than children in our overall data set (19.3%), which may explain why some children >6 years were retested. A majority (89.4%) of our follow-up test results were collected from 2 sites; therefore, related findings cannot be generalized to the broader data set. We were unable to ascertain whether follow-up tests had taken place after settlement in permanent housing; therefore, we chose to use a narrow definition for follow-up testing, which excluded a number (2411) of follow-up tests not meeting the retesting time period we specified. Finally, we were not always able to determine which BLL testing method was used in each site. On the basis of a May 2017 Food and Drug Administration advisory, use of LeadCare analyzers for venous BLL testing may produce falsely low BLL results.²⁶ If participating sites used this method, it could have underestimated EBLL prevalence.

Efforts to reduce EBLL among foreign-born children, including refugee children, are an important part of the Healthy People 2020 strategy of achieving BLLs <5.2 µg/dL for ≥97.5% of the population aged 1 to 5 years.²⁷ Although EBLL prevalence in resettled refugee children declined over the period of this analysis, it remains far higher than the general US prevalence. EBLL prevalence varied by examination country, and refugee children remain susceptible to both overseas and domestic exposures, as illustrated by postarrival increases in BLL for some children. Refugee children with potential lead exposure can arrive in any state, so it remains important to improve linkages between federal, state, and local lead and refugee health programs to facilitate collaboration, ensure appropriate screening and follow-up for refugee children, and share best practices. We encourage states to notify the CDC of specific lead exposures or refugee populations with unusually high EBLL prevalence rates, which can help facilitate overseas and domestic interventions. Finally, we suggest that states develop and share translated resources for lead education and encourage lead poisoning prevention education for refugee families early and often. Because refugee children are a subset of all immigrant children who may have similar exposures, such resources may also benefit other children migrating to the United States.

Refugee children remain at risk for EBLL, despite declines in prevalence over time. With our findings, we underscore the importance of screening all arriving refugee children aged 6 months to 16 years for EBLL, followed by rescreening of refugee children aged 6 months to 6 years 3- to 6-months postresettlement. Although domestic BLL increases were more likely to occur in younger children, our data suggest that older children and adolescents also experience increases in BLL after arrival. Because our follow-up data were limited to 5 sites, further work is necessary to determine if this pattern is consistent across sites and its implications for the recommended age limits for follow-up testing.

We thank Collin Elias (ID), Kenneth Mulanya (Marion County, IN), Shandy Dearth (Marion County, IN), P. Joseph Gibson (Marion County, IN), and Amelia Self (UT) for the use of their data and their assistance. We also thank Adrienne Ettinger (CDC Division of Environmental Health Science and Practice) for providing lead subject matter expertise and Yecai Liu, Christina Phares, and Emily Jentes (CDC Division of Global Migration and Quarantine) for their guidance and support during data analysis and article preparation.

BLL

blood lead level

CDC

Centers for Disease Control and Prevention

CI

confidence interval

EBLL

elevated blood lead level

PR

prevalence ratio