Abstract
Objectives. To examine elevated blood lead levels (EBLLs) in refugee children by postrelocation duration with control for several covariates.
Methods. We assessed EBLLs (≥ 5µg/dL) between 2012 and 2016 of children younger than 15 years (n = 1950) by the duration of resettlement to health screening by using logistic regression, with control for potential confounders (gender, region of birth, age of housing, and intestinal infestation) in a cross-sectional study.
Results. Prevalence of EBLLs was 11.2%. Length of time from resettlement to health screening was inversely associated with EBLLs (tertile 2 unadjusted odds ratio [OR] = 0.79; 95% confidence interval [CI] = 0.56, 1.12; tertile 3 OR = 0.62; 95% CI = 0.42, 0.90; tertile 2 adjusted odds ratio [AOR] = 0.62; 95% CI = 0.39, 0.97; tertile 3 AOR = 0.57; 95% CI = 0.34, 0.93). There was a significant interaction between intestinal infestation and age of housing (P < .003), indicating significant risk in the joint exposure of intestinal infestation (a pica proxy) and age of house.
Conclusions. Elevated blood lead levels were reduced with increasing length of time of resettlement in unadjusted and adjusted models. Improved housing, early education, and effective safe-house inspections may be necessary to address EBLLs in refugees.
Lead poisoning presents as a multisystem dysfunction caused by toxic accumulation in nerves and bones. Lead exposure–associated neurobehavioral and cognitive impairment accounts for 13.0% and 14.8% of reading and math failures according to a study of Chicago children.1 Neurocognitive damage caused by lead exposure is long-term, evident in children experiencing learning and developmental difficulties up to early adulthood.2 Children are susceptible to lead exposure because of hand–mouth behaviors, higher gastric absorption rate, increased permeability of the blood–brain barrier, and actively developing brain and bones.3 This is more so in malnourished children, especially those with other metal deficiencies such as iron and calcium deficiencies.4
Children from families of lower socioeconomic status are more vulnerable to the effects of elevated blood lead levels (EBLLs), attributable, in part, to lack of timely diagnosis related to limited access to care,2 poor housing, lack of education about lead toxicity, and malnutrition.4 One vulnerable population is refugee children. These children are likely to come from households with median incomes $8000 less than the median income of US-born households. A thorough review of integration outcomes of US refugees has been addressed elsewhere.5 Instructively, an estimated 35 000 children are admitted as refugees into the United States every year.6 It is documented that US refugee populations have higher prevalence of EBLLs than do US-born children.6–8 In a recent study, a survey of blood levels showed more than 10 times higher prevalence of EBLLs in refugees compared with similar-aged US-born children in Massachusetts (16% vs 1.4%).7 Notably, in 2000, a Sudanese child died shortly after relocation from local lead exposure.9,10 Though refugee children are increasingly becoming an important subgroup of US children, there are limited comprehensive guidelines for addressing health concerns for this demography.6 The persisting effect of lead exposure on neurocognitive development has made it important to examine likely exposure sources germane to this growing demographic group.
Sociocultural practices have been implicated as a likely determinant for higher lead exposures among refugee children.10 However, few epidemiological studies have reported postresettlement exposures, with some identifying quality of housing as an important risk for EBLLs in this subgroup.7,8 Lead exposure is associated with economic disadvantage and therefore it may correlate closely with socioeconomic status11; refugee families are more likely to opt for cheaper housing, putting them at high risk of exposure.10 The limited housing options to newly arriving refugees pose a significant risk. Moreover, refugees are less likely to be aware of the sources of lead exposures in the United States and therefore may unintentionally place their children at higher risk.10
Environmental lead exposures are significantly prevalent in the third world.8 Outside the United States, common exposures to lead occur through gasoline, cooking earthenware, traditional medication, and cosmetics.10 Other studies have cited battery lead recycling and poor delineation of zoning for residential and industrial activities as major contributors. By contrast, residency in houses built before 1950 is the primary risk of exposure to lead in the United States with exposures resulting mainly from consumption of flaking paint and contaminated soils.7
Pica, an eating disorder characterized by ingestion of nonnutritive substances, is commonly associated with anemia, intestinal infestation, and lead poisoning. Theoretically, children with pica who live in older houses are more likely to have a higher risk of lead intoxication. As a consequence, it is important to examine the effectiveness of measures that make old houses safe for refugee resettlement by studying modification of risk of EBLLs with joint exposure of house age and pica activity (using intestinal infestation as a proxy).
Four epidemiological studies have examined EBLLs in resettled refugees.7,8,12,13 These studies have faced some limitations, including use of 10 micrograms per deciliter (µg/dL) cut-off for case ascertainment, working assumption of initial lead screening level being indicative of preadmission exposure, and limited control for potentially confounding covariates. Because of limited amount of epidemiological evidence, we re-examined postresettlement duration and EBLLs while evaluating a range of covariates with the Centers for Disease Control and Prevention’s (CDC’s) new recommended blood lead level (BLL) cut-off of 5 µg/dL.14,15
The primary objective of this research was to examine EBLLs in refugee children by postrelocation duration while controlling for several covariates. In addition, we attempted to quantify the role of housing and, subsequently, evaluated effect modification of the risk of EBLLs by house age in children who may have pica behavior. New information gained from this work may assist to evaluate current interventions aimed at reducing lead exposures in this vulnerable subpopulation.
METHODS
All persons meeting the US Immigration and Naturalization statute definition of a refugee qualify for a free medical screening program that ideally occurs within the first 90 days of admission to the United States.16 The state of Kentucky is ranked 15th in terms of numbers of refugee and other entrants resettled, with approximately 85% of refugees resettling in Louisville. The University of Louisville’s Arriving Refugee Informatics Surveillance and Epidemiology (ARIVE) database, established in August 2012, maintains a repository of health screening and vaccination information for adult and pediatric refugees resettled in the state of Kentucky. A more comprehensive description of ARIVE is available elsewhere.17
ARIVE monitors, among others, panel tests that are recommended by the CDC. As part of the panel, BLLs of refugee children are measured and reported in micrograms per deciliter. Data are entered, validated, and monitored through the secure ARIVE system, which is hosted in Research Electronic Data Capture system (REDCap), a Web-based secure health database.18 Participation in the Refugee Health Screening Program is voluntary.
For the purposes of this study, the analysis was limited to refugees aged 15 years or younger admitted to the United States between September 1, 2012, and June 1, 2016. We included a total of 1950 children meeting our age cut-off criteria from the ARIVE data set (n = 7495). The sample size used in the analysis was 1726 after exclusion of children with missing responses to BLL or address information or both (n = 224).
Measures
We computed time of resettlement to health assessment as the difference between the dates of arrival and screening and grouped into tertiles by the overall length of time. We then analyzed this as a categorical variable in our main analysis. We coded BLLs as a dichotomous variable such that we coded BLLs greater than or equal to 5 µg/dL as “high” and BLLs less than 5 µg/dL as “low.”
We chose covariates that were used in the analysis on the basis of their associations with determinants of lead levels and postresettlement duration in current literature and availability in the ARIVE data set. Region of birth, age of housing, age, gender, body mass index (BMI, defined as weight in kilograms divided by the square of height in meters), intestinal infestation, anemia, and season of lead testing have been identified as potential confounders in previous studies.6–8,19 We abstracted the median year of house structure (from the 2014 American Community Survey) and used it as the average age of housing based on reported address zip code. This was coded as a continuous variable in the analyses. We calculated and reported in years age at entry—the difference between the date of admission into the United States and reported birth date. We coded this as a dichotomous variable using less than or equal to 3 years and greater than 3 years age cutoff similar to categories used by Eisenberg et al.7
We abstracted gender, season, and parasitic infestation from ARIVE data set. We derived anemia as a dichotomous variable following methods described in Eisenberg et al.7 Finally, we computed age- and gender-specific percentiles from reported BMI and coded them as continuous variables while we coded season of testing and region of birth (Asia, Eurasia, Latin America and Caribbean, Middle East, and sub-Saharan Africa) as categorical variables.
Statistical Analysis
We evaluated associations of covariates with the length of time from resettlement to health screening and BLL test to assess for potential confounding. We used the Pearson χ2 test to evaluate categorical variables, the Wilcoxon or Kruskal–Wallis test for continuous variables, and reported P values accordingly. We used the Fisher exact test when we encountered cells less than 5. We calculated unadjusted odds ratios (ORs) and the 95% confidence intervals (CIs) with logistic regression. We determined unadjusted ORs between resettlement time and the likelihood of reporting EBLLs by using logistic regression.
In the multivariable logit model, we examined the relationship of BLL to length of time from resettlement to health screening, while adjusting for covariates. In this analysis, we included age, gender, region of birth, age of housing, and parasitic infestation as categorical variables. Using length of time from resettlement to health screening as the principal predictor, we retained covariates in the model if the likelihood ratio test comparing models with and without the factor was at the P less than .15 level. We did not include child’s age in the final model because it did not meet the P less than .15 cutoff for further analysis.
In a secondary analysis, we estimated the Β in the relationship between house age and log-transformed lead levels in the event that length of time from resettlement to health screening was a significant predictor of lead levels in our full model. To assess whether intestinal parasite infestation, a proxy for pica,20 modifies the association between age of housing and EBLLs, we included an interaction term in the full multivariable logit model corresponding to the joint exposure (intestinal infestation and age of house).20
RESULTS
The population was almost equally distributed between boys and girls (52% boys vs 48% girls) with a population median age of 7.2 years (interquartile range = 3.8–13.3). Birth origin was predominately sub-Saharan Africa (48%), and participating children were more likely to have relocated to the United States in 2015 (34%). In addition, most of the children were not anemic (90%) and recorded a median BMI score of 49th percentile.
Table 1 reports covariate characteristics by postresettlement duration to health screening and also by the status of lead levels. In this analysis, 11.2% (n = 192) of the children had EBLLs of which 41.2% (n = 79), 33.3% (n = 64), and 25.3% (n = 49) were screened within first, second, and third tertiles of the postresettlement period to health screening, respectively.
TABLE 1—
Characteristics of Blood Lead Levels and Length of Resettlement to Health Screening Among Refugee Children: Kentucky, 2012–2016
| Postrelocation Duration, Daysa |
Blood Lead Level |
||||||
| Variable | Tertile 1 (n = 585) | Tertile 2 (n = 582) | Tertile 3 (n = 555) | P | ≥ 5 µg/dL (n = 192) | < 5 µg/dL (n = 1530) | P |
| Age (n = 1722) | .032 | .020 | |||||
| Percentiles (25th/50th/75th) | 4.0/7.6/11.3 | 3.7/6.7/10.6 | 3.9/7.4/11.3 | 3.3/6.3/10.6 | 4.0/7.4/11.1 | ||
| Mean ±SD | 7.7 ±4.2 | 7.1 ±4.0 | 7.6 ±4.1 | 6.8 ±4.1 | 7.6 ±4.1 | ||
| > 3 y, % (no.) | 82 (479) | 80 (467) | 83 (462) | .42 | 77 (147) | 82 (1261) | .048 |
| ≤ 3 y, % (no.) | 18 (106) | 20 (115) | 17 (93) | 23 (45) | 18 (269) | ||
| Gender, % (no.)a | .70 | .002 | |||||
| Male | 50 (294) | 53 (307) | 52 (286) | 62 (119) | 50 (768) | ||
| Female | 50 (291) | 47 (275) | 48 (269) | 38 (73) | 50 (762) | ||
| BMI percentileb | .21 | .23 | |||||
| Percentiles (25th/50th/75th) | 21.8/50.0/80.0 | 22.0/47.5/77.8 | 22.0/51.0/85.0 | 20.0/46.0/75.5 | 22.0/50.0/81.0 | ||
| Mean ±SD | 50.4 ±32.6 | 49.6 ±31.3 | 52.9 ±32.9 | 48.2 ±31.6 | 51.3 ±32.4 | ||
| Region, % (no.)a | <.001 | <.001 | |||||
| Asia (exclusively Near East) | 29 (171) | 27 (155) | 25 (140) | 46 (88) | 25 (378) | ||
| Eurasia | 0 (2) | 0 (0) | 0 (0) | 0 (0) | 0 (2) | ||
| Latin America and Caribbean | 5 (29) | 10 (60) | 20 (113) | 6 (12) | 12 (190) | ||
| Middle East | 20 (115) | 24 (141) | 17 (97) | 14 (27) | 21 (326) | ||
| Sub-Saharan Africa | 46 (268) | 39 (226) | 37 (205) | 34 (65) | 41 (634) | ||
| Season, % (no.)a | .61 | .77 | |||||
| Winter | 51 (300) | 51 (299) | 54 (299) | 53 (102) | 52 (796) | ||
| Summer | 49 (285) | 49 (283) | 46 (256) | 47 (90) | 48 (734) | ||
| Year of arrival, % (no.)a | <.001 | .33 | |||||
| 2012 | 15 (85) | 4 (22) | 3 (17) | 10 (20) | 7 (104) | ||
| 2013 | 23 (135) | 16 (95) | 28 (153) | 21 (41) | 22 (342) | ||
| 2014 | 20 (117) | 38 (220) | 37 (205) | 29 (56) | 32 (486) | ||
| 2015 | 34 (198) | 38 (222) | 31 (170) | 33 (63) | 34 (527) | ||
| 2016 | 9 (50) | 4 (23) | 2 (10) | 6 (12) | 5 (71) | ||
| Anemia, % (no.)c | .31 | .24 | |||||
| No | 92 (530) | 90 (516) | 89 (459) | 88 (165) | 90 (1340) | ||
| Yes | 8 (48) | 10 (60) | 11 (56) | 12 (23) | 10 (141) | ||
| Intestinal infestation, % (no.)d | .52 | .007 | |||||
| No | 59 (202) | 61 (240) | 64 (227) | 50 (63) | 63 (606) | ||
| Yes | 41 (138) | 39 (152) | 36 (130) | 50 (62) | 37 (358) | ||
Note. BMI = body mass index (weight in kilograms divided by the square of height in meters).
n = 1722.
n = 1700. Tertiles 1, 2, and 3 were equivalent to 1 to < 40 days, ≥ 40 to < 62 days, and ≥ 62 to 391 days, respectively.
n = 1669.
n = 1089.
In the examination of covariates and resettlement time before screening, we identified significant associations between region of birth, with respondents of Latin American and Caribbean origin being more likely to report for screening in the third rather than first tertile (20% vs 5%; P < .001). We did not find any significant associations with our remaining covariates: dichotomized age (P = .42), gender (P = .70), BMI (P = .21), season of screening (P = .61), anemia (P = .31), and parasitic infestation (P = .52; Table 1).
In the examination of the association between EBLL and covariates, compared with children with normal levels, children with EBLLs were younger (≤ 3 years; 23% vs 18%; P = .048), and were more likely to be male (50% vs 38%; P = .002), to come from Asia (46% vs 25%; P < .001), and to report intestinal infestation (50% vs 37%; P = .007; Table 1).
In the primary unadjusted logistic model, we observed a significant inverse association between length of time from resettlement to screening and likelihood of reporting EBLLs. Compared with those seen within the first tertile, those seen in the second and third tertiles had reduced odds of EBLLs (second tertile OR = 0.79 [95% CI = 0.56, 1.12]; third tertile OR = 0.62 [95% CI = 0.42, 0.90]). The relationship remained significant in the adjusted model with house age, gender, region of birth, and intestinal infestation as covariates (second tertile OR = 0.62 [95% CI = 0.39, 0.97]; third tertile OR = 0.57 [95% CI = 0.34, 0.93]). This indicates that reporting for screening in the second or third tertiles is protective of EBLL. Thus, a major contributor to the observed lead exposure may be from refugee-source countries, which is interrupted after resettlement. In addition, a 10-year increase in the age of housing was associated with 27% increased odds of recording EBLLs and documentation of previous intestinal infestation was associated with a 63% increased odds in the main adjusted model (Table 2).
TABLE 2—
Unadjusted and Adjusted Odds Ratios by Main Model Covariates Among Refugee Children: Kentucky, 2012–2016
| Variable | No. | Unadjusted OR (95% CI) | No. | Adjusteda OR (95% CI) |
| Postrelocation duration | 1722 | 1083 | ||
| Tertile 1 | 1 (Ref) | 1 (Ref) | ||
| Tertile 2 | 0.79 (0.56, 1.12) | 0.62 (0.39, 0.97) | ||
| Tertile 3 | 0.62 (0.42, 0.90) | 0.57 (0.34, 0.93) | ||
| Age of house | 1715 | 1.02 (1.01, 1.03) | 1083 | 1.27 (1.08, 1.48) |
| Gender | 1722 | 1083 | ||
| Male | 1.00 | 1 (Ref) | ||
| Female | 0.62 (0.45, 0.84) | 0.75 (0.51, 1.10) | ||
| Region | 1722 | 1083 | ||
| Asia (exclusively Near East; Ref) | 1 (Ref) | 1 (Ref) | ||
| Eurasia | 0 | 0 | ||
| Latin America and Caribbean | 0.27 (0.14, 0.49) | 0.35 (0.17, 0.66) | ||
| Middle East | 0.36 (0.14, 0.49) | 0.30 (0.16, 0.54) | ||
| Sub-Saharan Africa | 0.44 (0.22, 0.55) | 0.35 (0.22, 0.56) | ||
| Parasite infestation | 1089 | 1083 | ||
| No | 1 (Ref) | 1 (Ref) | ||
| Yes | 1.67 (1.14, 2.42) | 1.63 (1.09, 2.43) |
Note. CI = confidence interval; OR = odds ratio.
The variables in the fully adjusted model were postrelocation duration, age of house, gender, region of birth, and parasitic infestation.
We observed a statistically significant interaction between age of housing and intestinal infestation on the relationship between the length of time from resettlement to screening and likelihood of EBLLs (P = .003; Table 3). Children with intestinal infestation living in the oldest homes in Louisville were more likely to have EBLLs (adjusted OR = 4.63; 95% CI = 2.11, 11.10). Subsequently, in the secondary analysis, there was a 2.94% (95% CI = 0.95%, 4.97%) increase in the geometric mean of BLL for an increase in a house age from 43 years to 53 years, with adjustment for identified covariates in our main adjusted model.
TABLE 3—
Adjusted Odds Ratios by Strata of House Age (Percentiles) in the Association Between Lead Levels and Intestinal Infestation Among Refugee Children: Kentucky, 2012–2016
| BLL |
||||
| No. | < 5 µg/dL | ≥ 5 µg/dL | Adjusted ORa (95% CI) | |
| ≤ 25th percentile | ||||
| Intestinal infestation | ||||
| No (Ref) | 179 | 162 | 17 | 1 (Ref) |
| Yes | 99 | 92 | 7 | 0.77 (0.28, 1.91) |
| Total | 278 | |||
| > 25th to 50th percentile | ||||
| Intestinal infestation | ||||
| No (Ref) | 186 | 169 | 17 | 1 (Ref) |
| Yes | 86 | 74 | 12 | 2.23 (0.93, 5.33) |
| Total | 272 | |||
| > 50th to 75th percentile | ||||
| Intestinal infestation | ||||
| No (Ref) | 164 | 145 | 19 | 1 (Ref) |
| Yes | 102 | 90 | 12 | 0.93 (0.41, 2.09) |
| Total | 266 | |||
| > 75th percentile | ||||
| Intestinal infestation | ||||
| No (Ref) | 137 | 127 | 10 | 1 (Ref) |
| Yes | 130 | 99 | 31 | 4.63 (2.11, 11.00) |
| Total | 267 | |||
Note. BLL = blood lead level; CI = confidence interval; OR = odds ratio. P value for interaction term between house age and intestinal infestation is .003.
Adjusted by postadmission duration, region of birth, and gender.
DISCUSSION
In this review of the Kentucky ARIVE system, we found that 11% of refugee children had EBLLs. This compares with a statewide estimate of 0.36% from the child BLL surveillance data for Kentucky.21 We observed that refugee children who took longer to complete lead screening were more likely to have lower BLLs than children who were screened earlier. This association persisted after we adjusted for covariates often related to EBLLs.
This study is one of few that have examined the predictors of lead levels in the refugee community and is different in that it investigated factors associated with EBLLs in children who were screened at differing lengths of time after resettlement. To date, 4 studies have examined EBLLs in refugees.7,8,12,13 In a review of refugee data in New Hampshire, Plotinsky et al. observed an increased BLL in about 7% of respondents after admission to the United States with a significant predictor being nutritional wasting (P < .001).8 An earlier study by Zabel et al. identified only 1 child (i.e., 0.7% of respondents) with an EBLL.12 Also, researchers using Massachusetts refugee health data identified 6% of respondents to have EBLLs. Eisenberg et al. showed that the age of housing was an important predictor of BLL increase of more than 2 micrograms in their refugee cohort (hazard ratio = 1.7; 95% CI = 1.2, 2.3).7 Our results are in line with these studies but also confirm a general reduction in risk of reporting EBLLs after resettlement in the United States. However, it also differs focally as we attempted to examine the impact of local factors such as age of housing on the initial lead screening results. This distinction is important because of the widely held view that initial lead screening results are representative of country-of-origin exposures. We have shown that the initial BLL screen result might be significantly influenced by local exposures. Measurement of postresettlement increase in lead levels may be higher in refugees than estimated in some of these studies.
In addition, we observed a statistically significant interaction between intestinal infestation (a potential marker for pica) and age of housing in the association between EBLLs and length of time from resettlement to health screening. This finding suggests that there is a modification of risk of EBLLs in children with pica behavior, with intestinal infestation as a proxy marker, for different ages of houses. The differential odds in reporting EBLLs by house age for children with pica suggest that the effectiveness of measures to reduce lead exposures in older homes may be limited. Current estimates show that about 19% of all houses in Kentucky were built before 1950.21 As a consequence, programs designed to educate refugees who live in older houses about the risks of pica behavior may be an effective approach to reducing EBLLs in this population.
Lead is an important environmental pollutant in developed and developing countries. Improved regulations on environmental lead exposures have reduced morbidity caused by severe EBLLs in US-born children over the past 5 decades.9 On the contrary, lax environmental regulations and lack of awareness of the effect of elevated lead levels have resulted in a persistent elevation of lead levels in refugee children born in foreign countries.6,8 In addition, the potential for postrelocation lead exposures have been recognized, though few studies have examined such risk factors. The effect of an EBLL is multisystemic, leading to hematopoietic, gastrointestinal, urinary, and cardiovascular manifestations. The mechanism of neurotoxicity has been suggested to be mediated through the disruption of calcium homeostasis leading to the inhibition of neurotransmitters.22 Children are more susceptible because of their particular anatomic and physiologic characteristics that cause an increased absorption rate and increased blood–brain barrier permeability of ingested lead.3 A major environmental and child health concern of EBLLs is the impact on the child neurocognitive development.11 Previous work looking at cumulative exposures found that a graded increase in dentine lead concentration was associated with diminishing academic and neurobehavioral performance.11 Furthermore, this insult on the child neurocognitive development has been suggested to be persistent and irreversible.2
This work adds to the body of literature on children’s health in the refugee population. It has shown that, though we observed a general reduction in BLLs after resettlement, the likelihood of exposures from old houses of residence before the initial lead screening measurements cannot be ignored. Also, the significant difference in odds of reporting EBLLs in children with pica living in older houses probes the effectiveness of measures to make old houses safe.
Limitations
A number of potential limitations regarding this study should be noted. First, the half-life of blood lead and bone lead are substantially different.23,24 As we assessed EBLLs among children from resettlement to health screening, which in some cases was longer than 90 days, it is possible that BLLs were affected by lead stored in the bone that was released over time. Second, inconsistency in lead level measurement across different clinics and mismeasurement cannot be ruled out. We are not privy to the methods of measurement across facilities that participate in refugee resettlement programs. These facilities, however, operate under standard state guidelines for providing laboratory data for patient management and reporting to states and federal agencies. It is, therefore, unlikely that these lead measurements varied significantly in a way that affected the true lead levels. Another concern is about exposure from other sources such as water, herbal concoctions, and medicines. Its effect in our models would be nondifferential as we do not anticipate any significant changes in the use of these materials to the likelihood of reporting EBLLs and the time for reporting for lead screening. In the event that the exposures were misclassified, this would bias the estimate to the null. Water quality reports did not indicate that drinking water was a main source of exposure for refugees in Louisville; however, reports may not be 100% accurate and some exposure may have occurred through drinking water.
Participation in the refugee screening program is voluntary. If respondents with higher lead levels were more likely to participate in ARIVE because of exposure symptoms, this population would be overrepresented in our study sample. Also, because symptoms of mild lead intoxication are less dramatic,11 such increase would be nondifferential in the postresettlement duration categories and therefore bias our estimate toward the null. Because ARIVE is voluntary, it is difficult to assess if BLL testing was lower or higher than what should be expected for the refugee population. A 2017 study by Roberts et al. indicated that the reported number of EBLLs in children to the CDC was dramatically lower than the expected number of EBLLs in children throughout the United States.25
Potential misclassifications attributable to lead level measurement and data entry errors are possible, although this is unlikely. Our estimates would be biased to the null in the event that these itemized factors were misclassified. Residual or uncontrolled confounding cannot be ruled out. A potential limitation is geographic clustering effect because of similar housing and location preference by different communities, resulting in similar exposure profile. Furthermore, significant differences in time to screening across these communities because of distance considerations or group behavior may have confounded our estimates. Though we included origin of birth in our models, it is possible that the effect of geographic clustering could not be completely adjusted for because of residual confounding. Mismeasurement of other confounders might have resulted in residual confounding because of inadequate control for those confounders. However, we evaluated a large number of potential confounding factors, including those used in previous studies. Finally, our inability to consider other confounders was because of unavailability in the ARIVE data set. The final limitation of this study is that we had no information on other BLL tests that refugee children may have had before this study and we did not have follow-up BLL tests; therefore, we could not investigate changes in lead levels.
Public Health Implications
In conclusion, we found a reduction in EBLLs in refugees with longer lengths of time from resettlement to health screening. This reduction is significantly affected by the age of housing and also potentially effect modified by the joint exposure of pica behavior and age of the residence. Paints and dust of old houses remain a common way of exposure to lead in the United States and may confound initial lead level screening measurements in refugees. Refugees are more at risk for exposure because of socioeconomic factors and generalized limited awareness of the exposure risks. If our results are replicated in future studies, it may be necessary to institute more aggressive programs to educate the refugee population immediately after admission to the United States. In addition, measures that routinely assess lead and other heavy metal exposures in refugee safe houses may address EBLL risks.
ACKNOWLEDGMENTS
Funding for management of the refugee health database is provided by the Kentucky Office for Refugees.
HUMAN PARTICIPANT PROTECTION
Institutional review board approval was obtained for this study from the University of Louisville.
Footnotes
See also Wells, p. 163.
REFERENCES
- 1.Evens A, Hryhorczuk D, Lanphear BP et al. The impact of low-level lead toxicity on school performance among children in the Chicago Public Schools: a population-based retrospective cohort study. Environ Health. 2015;14:21. doi: 10.1186/s12940-015-0008-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Needleman HL, Schell A, Bellinger D, Leviton A, Allred EN. The long-term effects of exposure to low doses of lead in childhood. N Engl J Med. 1990;322(2):83–88. doi: 10.1056/NEJM199001113220203. [DOI] [PubMed] [Google Scholar]
- 3.Kumar V, Abbas AK, Aster JC. Robbins Basic Pathology: With Student Consult Online Access. Philadelphia, PA: Elsevier Health Sciences; 2012. [Google Scholar]
- 4.Akkus C, Ozdenerol E. Exploring childhood lead exposure through GIS: a review of the recent literature. Int J Environ Res Public Health. 2014;11(6):6314–6334. doi: 10.3390/ijerph110606314. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Capps R, Newland K, Fratzke S The integration outcomes of U.S. refugees. Successes and challenges. Migration Policy Institute. 2015. Available at: http://www.migrationpolicy.org/research/integration-outcomes-us-refugees-successes-and-challenges. Accessed March 14, 2017.
- 6.Yun K, Matheson J, Payton C et al. Health profiles of newly arrived refugee children in the United States, 2006–2012. Am J Public Health. 2016;106(1):128–135. doi: 10.2105/AJPH.2015.302873. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Eisenberg KW, van Wijngaarden E, Fisher SG et al. Blood lead levels of refugee children resettled in Massachusetts, 2000 to 2007. Am J Public Health. 2011;101(1):48–54. doi: 10.2105/AJPH.2009.184408. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Plotinsky RN, Straetemans M, Wong L-Y et al. Risk factors for elevated blood lead levels among African refugee children in New Hampshire, 2004. Environ Res. 2008;108(3):404–412. doi: 10.1016/j.envres.2008.08.002. [DOI] [PubMed] [Google Scholar]
- 9.Centers for Disease Control and Prevention. Fatal pediatric lead poisoning—New Hampshire. 2000. Available at: https://www.cdc.gov/mmwr/preview/mmwrhtml/mm5022a1.htm. Accessed July 5, 2016.
- 10.Schmidt CW. Unsafe harbor? Elevated blood lead levels in refugee children. Environ Health Perspect. 2013;121(6):A190–A195. doi: 10.1289/ehp.121-a190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.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(13):689–695. doi: 10.1056/NEJM197903293001301. [DOI] [PubMed] [Google Scholar]
- 12.Zabel EW, Smith ME, O’Fallon A. Implementation of CDC refugee blood lead testing guidelines in Minnesota. Public Health Rep. 2008;123(2):111–116. doi: 10.1177/003335490812300203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Tehranifar P, Leighton J, Auchincloss AH et al. Immigration and risk of childhood lead poisoning: findings from a case–control study of New York City children. Am J Public Health. 2008;98(1):92–97. doi: 10.2105/AJPH.2006.093229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Centers for Disease Control and Prevention; National Center for Environmental Health. New blood lead level information. Available at: https://www.cdc.gov/nceh/lead/acclpp/blood_lead_levels.htm. Accessed July 5, 2016.
- 15.Raymond J, Brown MJ. Blood lead levels in children aged <5 years—United States, 2007–2013. MMWR Morb Mortal Wkly Rep. 2016;63(55):66–72. doi: 10.15585/mmwr.mm6355a6. [DOI] [PubMed] [Google Scholar]
- 16. 45 CFR §400.107 (1998).
- 17.Carrico R, Bosson R, Koch J et al. Addressing the challenges of refugee health: experiences from the University of Louisville interprofessional refugee immunization clinic. J Nurs Educ Pract. 2015;5(12):88–95. [Google Scholar]
- 18.Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research Electronic Data Capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377–381. doi: 10.1016/j.jbi.2008.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wang N, Chen C, Nie X et al. Blood lead level and its association with body mass index and obesity in China—results from SPECT-China study. Sci Rep. 2015;5:18299. doi: 10.1038/srep18299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Nwaneri DU, Ibadin MO, Ofovwe GE, Sadoh AE. Intestinal helminthiasis in children with chronic neurological disorders in Benin City, Nigeria: intensity and behavioral risk factors. World J Pediatr. 2013;9(2):152–157. doi: 10.1007/s12519-012-0394-9. [DOI] [PubMed] [Google Scholar]
- 21.Centers for Disease Control and Prevention; National Center for Environmental Health. Lead. Kentucky data, statistics, and surveillance. 2016. Available at: https://www.cdc.gov/nceh/lead/data/state/kydata.htm. Accessed March 14, 2017.
- 22.Kumar V, Abbas AK, Fausto N, Aster JC. Robbins & Cotran Pathologic Basis of Disease Elsevieron VitalSource. Philadelphia, PA: Elsevier Health Sciences; 2009. [Google Scholar]
- 23.Rabinowitz MB. Toxicokinetics of bone lead. Environ Health Perspect. 1991;91:33–37. doi: 10.1289/ehp.919133. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Centers for Disease Control and Prevention; National Biomonitoring Program. Biomonitoring summary - lead. 2016. Available at: https://www.cdc.gov/biomonitoring/Lead_BiomonitoringSummary.html. Accessed May 27, 2017.
- 25.Roberts EM, Madrigal D, Valle J, King G, Kite L. Assessing child lead poisoning case ascertainment in the US, 1999–2010. Pediatrics. 2017;139(5):pii:e20164266. doi: 10.1542/peds.2016-4266. [DOI] [PubMed] [Google Scholar]