- Email: [email protected]
Blood Lead Levels in Young Children: US, 2009-2015
Blood Lead Levels in Young Children: US, 2009-2015 Leland F. McClure, PhD, Justin K. Niles, MA, and Harvey W. Kaufman, MD Objectives To evaluate trends in blood lead levels in children <6 years of age, this Quest Diagnostics Health Trends report builds on previously reported National Health and Nutrition Examination Survey data with a much larger national group and adds more detail and novel assessments. Study design This report describes the results from a 6-year retrospective study (May 2009-April 2015) based on >5 million blood lead level results (including >3.8 million venous results) from children <6 years old living in all 50 states and the District of Columbia. We evaluated yearly changes and examined demographic categories including sex, pre-1950s housing construction, poverty income ratios (PIRs), Medicaid enrollment status, and geographic regions. Results Among children <6 years old, 3.0% exhibited blood lead levels $5.0 mg/dL (high). There were significant differences in high blood lead levels based on sex, pre-1950s housing construction quintiles, and PIR <1.25 and PIR >5 (all P < .01). Health and Human Services regions, states, and 3-digit ZIP code areas exhibited drastically different frequencies of high blood lead levels and blood lead levels $10.0 mg/dL (very high). Generally, levels declined over time for all groups. Conclusion Examination of more than 5 million venous blood lead level results in children younger than 6 years old allowed for a robust, detailed analysis of blood lead level group results by geography and other criteria that are prohibited with the narrower National Health and Nutrition Examination Survey database. Progress in reducing the burden of lead toxicity is a public health success story that is incomplete with some identified factors posing larger, ongoing challenges. (J Pediatr 2016;-:---).
C
hildhood lead toxicity is a preventable environmental disease with long-lasting adverse health and behavioral effects.1 Public health services and other health professionals throughout the US have dedicated more than 4 decades of efforts to screen children, especially those at high risk, for lead exposure and to identify primary sources of lead.2 Federal and local environmental policies have included the removal of lead from gasoline, reduction of lead in paints, and testing of homes for lead-based paint. These efforts along with laboratory testing and case management efforts have been instrumental in significantly reducing blood lead levels in the US. The 2007-2010 National Health and Nutrition Examination Survey (NHANES) estimate of the geometric mean blood lead level was 1.3 mg/dL,3 which is a 90% decrease compared with the 1976-1980 NHANES II 12.8 mg/dL estimate.4 In 1991, the Centers for Disease Control and Prevention (CDC) recommended changes for preventing childhood lead poisoning, which included a reduction for the blood lead level deemed safe (from 25 mg/dL to 10 mg/dL).5 In May 2012, the CDC Advisory Committee on Childhood Lead Poisoning Prevention (ACCLPP) identified that there is no safe blood lead level and the CDC accepted ACCLPP recommendations to remove all CDC blood lead level references to “blood lead level of concern.”6 The CDC position of “no safe blood lead level” is based on an absence of blood lead levels without effects and low blood lead levels that are associated with intellectual deficits, attention deficit behaviors, and poor academic achievement.7,8 That these effects appear to be irreversible9-11 emphasizes a public health care shift to primary prevention rather than secondary and tertiary prevention efforts, which are based on responses after detecting lead exposure. In May 2012, the CDC also adopted the ACCLPP committee recommendations to use the NHANES 97.5th blood lead level percentile (5.0 mg/dL) as an upper reference interval threshold to identify children with elevated blood lead levels. The 5.0 mg/dL value is based on 2 consecutive cycles of the NHANES blood lead level data distribution among study children 1-5 years of age. Based on the 5.0 mg/dL threshold, the 2012 ACCLPP committee report estimated 450 000 children in the US as having blood lead levels greater than the new reference limit.12 The NHANES analysis includes demographic categories with long-standing disparities of risk for elevated blood From the Quest Diagnostics, Madison, NJ
ACCLPP CDC HHS NHANES PIR
Advisory Committee on Childhood Lead Poisoning Prevention Centers for Disease Control and Prevention US Department of Health and Human Services National Health and Nutrition Examination Survey Poverty income ratio
Funded by Quest Diagnostics, which provided support in the form of salaries for all authors but did not have any additional role in the study design, collection, analysis and interpretation of data, writing of the manuscript, or decision to publish. The authors declare no conflicts of interest. 0022-3476/$ - see front matter. ª 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2016.05.005
1
THE JOURNAL OF PEDIATRICS
www.jpeds.com
lead levels, including age, sex, race/ethnicity, age of housing, poverty income ratio (PIR), and Medicaid enrollment status. Despite the insights provided by the NHANES analysis, the study has several limitations. One such limitation is that the low numbers of NHANES-enrolled children with blood lead levels $10 mg/dL (only 9 children in 2007-2008; 6 children in 2009-2010) make interpretation of population estimates of very high blood lead levels unreliable. In addition, the NHANES was not designed to produce estimates at the state and local level and may not detect statistically significant disparities with important public health implications. This Quest Diagnostics Health Trends report describes the results of a 6-year retrospective study based on a large national clinical laboratory database with more than 5 million results from children younger than 6 years of age. Our analysis builds upon previously reported NHANES data and includes insights into yearly trends and the distributions of blood lead levels by specimen type (venous and capillary), sex, payer type, US Department of Health and Human Services (HHS) region, residence state, PIR, and pre-1950s housing construction.
Methods The specimen requirement for venous blood lead level analysis is whole blood collected into an evacuated collection tube certified for lead testing, such as tan-top and royal blue-top tubes containing the anticoagulant EDTA. For the capillary collection method, the specimen collection container is the lavender-top capillary tube. The blood lead level analyses were performed by the use of either inductively coupled plasma/mass spectrometry or the Zeeman graphite furnace atomic absorption spectroscopy. Instrument calibrations are performed with standards traceable to the National Institutes of Standards and Technology. Performance for all methods is in compliance with the 4 mg/dL (or 10%, whichever is greater) CDC accuracy standards.13 The blood lead level results were evaluated with a 3.0 mg/dL lower reporting threshold. The laboratory analysis of venous specimens is consistent with the CDC definition for “confirmed elevated blood lead level” when indicated.14 The blood lead level data set includes deidentified results of testing performed for children <6 years of age, from May 2009 through April 2015 (3 years before and after the 2012 CDC change from the 10 mg/dL “level of concern” to the 5.0 mg/dL reference interval threshold). Instances of blood lead level results reported as a specimen submitted in a tube/container not certified for lead testing were excluded from the study. This study was deemed exempt by the Western Institutional Review Board. To avoid duplication of patient data, when 2 or more tests were associated with the same individual, only the first venous result (or the first capillary result if there were no venous results) within the data set was included in this study. The 3.0 mg/dL reporting threshold precluded our ability to estimate the mean blood lead level for the study with sufficient precision. Instead, analyses focused on the proportions 2
Volume of the population falling into each of 4 blood lead level groups: #3.0 mg/dL (below the reporting limit); 3.1-4.9 mg/dL (above the reporting limit and below the CDC 2012 reference interval threshold); 5.0-9.9 mg/dL (between the 2012 reference interval threshold and the previous 1991 CDC “level of concern”); and $10.0 mg/dL. Patient data were limited to patients <6 years of age, corresponding to the CDC age definition for high risk. Blood lead levels results missing patient sex were excluded from sex analysis. Data from the US Census Bureau’s 2009-2013 American Community Survey 5-Year Estimates15 were used to determine the proportion of housing constructed before 1950 by ZIP code. According to the CDC, “houses built before 1950 pose the greatest hazard to children because they are much more likely to contain lead-based paint than newer houses.”16 Quintiles were defined as the percentage of the housing category by ZIP code. Quintile thresholds for pre-1950s housing were defined as <3.6%, 3.6%-12.9%, 13.0%-29.9%, 30.0%50.9%, and $51.0%. All quintile thresholds were chosen to provide approximately equal numbers in each quintile group. Demographics were divided into quintiles to demonstrate trends in blood lead level proportions. ZIP codes are based on patient residence, not the site of the blood collection. Data from the United States Census Bureau’s 2008-2012 American Community Survey 5-Year Estimates17 were used to determine PIR of children’s area of residence by ZIP code. Quintiles were defined as the percentage of PIR <1.25 (low income) and PIR >5 (high income) by ZIP code. Quintile ranges were defined as <16.0%, 16.0%-27.9%, 28.0%38.9%, 39.0%-51.9%, and $52.0% for PIR <1.25, and <2.8%, 2.8%-6.9%, 7.0%-13.9%, 14.0%-27.9%, and $28.0% for PIR >5. This study included specimens submitted for blood lead level testing from all 50 states and the District of Columbia. Data were grouped for analysis by HHS region, state, and 3-digit ZIP code region. We limited our state analyses to those with at least 2000 children and our 3-digit ZIP code analysis to areas with at least 1000 children. The proportion of housing that was constructed before 1950 in various geographical regions also was analyzed. These data were weighted by the number of patients with specimens from individual ZIP codes. Statistical Analyses The Cochran-Armitage test was used to analyze trends in proportions of children with blood lead levels $5.0 mg/dL (high blood lead level) and $10.0 mg/dL (very high blood lead level) for various groups. Testing for statistical significance between the 2 groups was conducted with the c2 test. Multivariable logistic regression models to determine characteristics associated with high blood lead level and very high blood lead level also are reported. Variables in both models were chosen based on plausibility and/or statistical significance in previous studies.3 Living in ZIP codes associated with the greatest quintile of pre-1950s housing, low income, and high income were included as binary variables. Living in McClure, Niles, and Kaufman
- 2016
ORIGINAL ARTICLES
HHS regions 1, 3, or 7 also was included as a binary variable. Despite the small measures of association determined by unadjusted ORs in logistic regression models, year of service, patient age, sex, and payer status were included in the multivariable models because of their potential as confounding factors. Results of adjusted and unadjusted models were stated as ORs and 95% CIs. Data were analyzed with SAS, version 9.4 (SAS Institute, Cary, North Carolina).18
Results This study included 5 266 408 blood lead levels from children <6 years of age. Venous blood draws represent 72.2% of all specimen submissions (n = 3 803 070). Blood lead levels were below the reporting threshold of #3.0 mg/dL in 94.9% of venous specimens; 3.1-4.9 mg/dL in 2.2%; and $5.0 mg/ dL in 3.0%. Results from the 1 178 000 capillary specimens were slightly, but statistically significantly (P < .01), more likely to fall into greater blood lead level categories: capillary blood lead levels were #3.0 mg/dL in 93.6%, 3.1-4.9 mg/dL in 3.3%, and $5.0 mg/dL in 3.1%. There also was a group of 285 338 specimens with unknown draw type with the following blood lead level distribution: 95.8% #3.0 mg/dL, 2.3% between 3.0 and 4.9 mg/dL, and 1.9% $5.0 mg/dL. The remainder of the results section will focus on the 3 803 070 children with venous blood results. Sex Of the venous specimens tested, 1 947 693 (51.2%) were from boys and 1 842 881 (48.5%) were from girls; the other 12 496 (0.3%) specimens had no sex information. High blood lead level was only slightly greater for boys (3.1%) than girls (2.8%), although the difference was statistically significant (P < .01; Table I). Pre-1950s Housing Construction Living in an area with a high proportion of pre-1950 housing construction was strongly associated with having a high blood lead level (Table I). Living in a ZIP code where $51.0% of housing units were constructed before 1950 (the highest quintile) was associated with a significantly larger proportion of high blood lead levels (OR 5.86, 95% CI 5.716.01) or very high blood lead levels (OR 6.34, 95% CI 5.976.74) than living in an area with the lowest quintile. There was a statistically significant trend for both high blood lead levels and very high blood lead levels among the pre-1950s housing quintiles (both P < .01, Table I). Payer Type and PIRs Approximately 20% of patients (n = 768 879) with venous blood draws had Medicaid as the payer and 687 had Medicare as the payer. Patients with Medicaid or Medicare had a slightly, but statistically significantly, lower frequency of high blood lead level than did those with private payers (2.6 vs 2.9%; P < .01). Comparisons of other lead groups between payer types were also similar (Table I). The data also showed a Blood Lead Levels in Young Children: US, 2009-2015
positive association between percentage of PIR <1.25 in patient ZIP codes and high blood lead level (Table I). Children living in ZIP codes in which $52.0% have PIR <1.25 (the highest quintile) had a greater proportion of high blood lead level (OR 2.85, 95% CI 2.79-2.91) and very high blood lead level (OR 2.67, 95% CI 2.54-2.80) compared with those living in the lowest quintile. There was a statistically significant trend in both high blood lead level and very high blood lead level among the poverty quintiles (Table I). There was an inverse association between the greater-income ZIP codes and high blood lead level. Those living in ZIP codes with the greatest percentage of PIR >5 (the highest quintile) were much less likely to exhibit high blood lead levels (OR 0.43, 95% CI 0.42-0.44) and very high blood lead level (OR 0.43, 95% CI 0.41-0.46) than were those in the lowest quintile. There was a statistically significant trend in both high blood lead level and very high blood lead level among the PIR >5 quintiles (Table I). Geographic Regions Blood lead levels group results were analyzed by HHS region (Table I). Region 7 (Iowa, Kansas, Missouri, and Nebraska) had the largest proportion of both high blood lead levels (5.7%) and very high blood lead levels (1.1%). Region 1 (Connecticut, Massachusetts, Maine, New Hampshire, Rhode Island, and Vermont) and Region 3 (Delaware, District of Columbia, Maryland, Pennsylvania, Virginia, and West Virginia) also had notably large proportions of high blood lead levels (5.4% and 5.1%, respectively), and very high blood lead levels (1.0% and 1.1%, respectively). The states with the largest proportions of high blood lead levels were Minnesota (10.3%), Pennsylvania (7.8%), Kentucky (7.1%), Ohio (7.0%), and Connecticut (6.7%) (Table II). All these states and 20 others (25/37, 68%) exhibited a decline in the proportion of high blood lead levels between the first year and the final year of the study. New Hampshire had the largest absolute decline (from 9.7% to 2.6% high blood lead levels), and Mississippi had the largest absolute increase (from 3.1% to 6.3% high blood lead levels) between the first year and the final year of the study. Florida and California had the lowest proportions of high blood lead levels (1.1% and 1.4%, respectively) and very high blood lead levels (0.1% and 0.2%, respectively). Data from 325 3-digit ZIP code regions with more than 1000 specimens (36.0% of all 3-digit ZIP codes with specimens in the study) also were analyzed to assess their impact on state data. In 6 regions, >14.0% of specimens had high blood lead levels: Syracuse, NY (40.1%), Buffalo, NY (18.8%), Cincinnati, OH (16.0%), Poughkeepsie, NY (14.9%), York, PA (14.4%), and Oil City, PA (14.0%). The 11 regions with the largest proportions of specimens with very high blood lead levels also were all in New York, Pennsylvania, or Ohio. Syracuse, NY (16.0%), Buffalo, NY (6.0%), York, PA (5.5%), Poughkeepsie, NY (4.9%), and Oil City, PA (4.3%) had the greatest rates in the study. 3
THE JOURNAL OF PEDIATRICS
Volume -
www.jpeds.com
Table I. Blood lead levels in children <6 years old: 2009-2015
Total Sex1 Male Female Pre-1950s housing quintile2 <3.6% 3.6-12.9% 13.0-29.9% 30.0-50.9% $51.0% Payer type3 Private Medicaid/Medicare Low income (PIR <1.25) quintile4 <16.0% 16.0-27.9% 28.0-38.9% 39.0-51.9% $52.0% High income (PIR >5) quintile4 <2.8% 2.8-6.9% 7.0-13.9% 14.0-27.9% $28.0% HHS region5 1: CT, MA, ME, NH, RI, VT 2: NJ, NY 3: DE, DC, MD, PA, VA, WV 4: AL, FL, GA, KY, MS, NC, SC, TN 5: IL, IN, MI, MN, OH, WI 6: AR, LA, NM, OK, TX 7: IA, KS, MO, NE 8: CO, MT, ND, SD, UT, WY 9: AZ, CA, HI, NV 10: AK, OR, ID, WA Time period (Year) May 2009-April 2010 May 2010-April 2011 May 2011-April 2012 May 2012-April 2013 May 2013-April 2014 May 2014-April 2015
‡10.0 mg/dL (Very high BLL)
‡5.0 mg/dL (High BLL)
Number (Percent) 90 092 (2.37)
22 237 (0.58)
112 329 (2.95)
44 032 (2.26) 38 384 (2.08)
48 312 (2.48) 41 384 (2.25)
11 907 (0.61) 10 238 (0.56)
60 219 (3.09) 51 622 (2.80)
647 545 (98.06) 643 939 (97.22) 637 074 (96.24) 617 464 (93.65) 564 372 (89.96)
5777 (0.87) 8522 (1.29) 11 099 (1.68) 17 581 (2.67) 25 622 (4.08)
5817 (0.88) 7994 (1.21) 11 116 (1.68) 19 334 (2.93) 29 936 (4.77)
1248 (0.19) 1872 (0.28) 2647 (0.40) 4935 (0.75) 7446 (1.19)
7065 (1.07) 9866 (1.49) 13 763 (2.08) 24 269 (3.68) 37 382 (5.96)
2 665 330 769 566
2 531 010 (94.96) 734 185 (95.40)
57 982 (2.18) 15 157 (1.97)
61 863 (2.32) 16 288 (2.12)
14 475 (0.54) 3936 (0.51)
76 338 (2.86) 20 224 (2.63)
703 782 624 981 643 452 698 627 597 828
682 503 (96.98) 602 837 (96.46) 614 280 (95.47) 658 817 (94.30) 549 471 (91.91)
9066 (1.29) 9874 (1.58) 12 798 (1.99) 17 067 (2.44) 19 730 (3.30)
9779 (1.39) 9816 (1.57) 13 141 (2.04) 18 214 (2.61) 23 144 (3.87)
2434 (0.35) 2454 (0.39) 3233 (0.50) 4529 (0.65) 5483 (0.92)
12 213 (1.74) 12 270 (1.96) 16 374 (2.54) 22 743 (3.26) 28 627 (4.79)
657 317 666 236 692 538 614 839 637 740
613 075 (93.27) 627 939 (94.25) 658 850 (95.14) 590 014 (95.96) 618 030 (96.91)
18 333 (2.79) 16 417 (2.46) 14 352 (2.07) 10 800 (1.76) 8633 (1.35)
21 018 (3.20) 17 607 (2.64) 15 261 (2.20) 11 198 (1.82) 9010 (1.41)
4891 (0.74) 4273 (0.64) 4075 (0.59) 2827 (0.46) 2067 (0.32)
25 909 (3.94) 21 880 (3.28) 19 336 (2.79) 14 025 (2.28) 11 077 (1.74)
165 624 722 385 406 681 509 539 262 144 264 833 96 232 21 372 874 401 11 700
151 422 (91.43) 681 603 (94.35) 374 655 (92.13) 494 543 (97.06) 244 002 (93.08) 255 900 (96.63) 87 446 (90.87) 20 346 (95.20) 848 895 (97.08) 11 255 (96.20)
5184 (3.13) 19 431 (2.69) 11 407 (2.80) 6637 (1.30) 6711 (2.56) 3549 (1.34) 3292 (3.42) 436 (2.04) 13 291 (1.52) 212 (1.81)
7306 (4.41) 16 950 (2.35) 16 313 (4.01) 7085 (1.39) 8849 (3.38) 4240 (1.60) 4424 (4.60) 503 (2.35) 10 268 (1.17) 172 (1.47)
1712 (1.03) 4401 (0.61) 4306 (1.06) 1274 (0.25) 2582 (0.98) 1144 (0.43) 1070 (1.11) 87 (0.41) 1947 (0.22) 61 (0.52)
9018 (5.44) 21 351 (2.96) 20 619 (5.07) 8359 (1.64) 11 431 (4.36) 5384 (2.03) 5494 (5.71) 590 (2.76) 12 215 (1.40) 233 (1.99)
823 198 737 826 633 374 560 906 540 016 507 750
770 737 (93.63) 698 385 (94.65) 602 166 (95.07) 534 990 (95.38) 516 139 (95.58) 485 457 (95.61)
22 217 (2.70) 16 847 (2.28) 13 522 (2.13) 10 955 (1.95) 10 184 (1.89) 9142 (1.80)
24 185 (2.94) 18 449 (2.50) 14 171 (2.24) 11 993 (2.14) 10 941 (2.03) 10 353 (2.04)
6059 (0.74) 4145 (0.56) 3515 (0.55) 2968 (0.53) 2752 (0.51) 2798 (0.55)
30 244 (3.67) 22 594 (3.06) 17 686 (2.79) 14 961 (2.67) 13 693 (2.54) 13 151 (2.59)
Total
£3.0 mg/dL
>3.0 and <5.0 mg/dL
3 803 070
3 607 874 (94.87)
82 867 (2.18)
1 947 693 1 842 881
1 843 442 (94.65) 1 752 875 (95.12)
660 387 662 327 661 936 659 314 627 376
‡5.0 and <10.0 mg/dL
PIR, povery income ratio. Notes: 1 - 12 496 results did not have gender data. 2 - 531 730 results did not have pre-1950 house construction data available. 3 - 368 174 results did not have payer data available. 4 - 534 400 results did not have poverty/wealth data. 5 - 468 159 results did not have state data available.
In 49 3-digit ZIP codes, <1.0% of specimens had high blood lead levels. Eighteen of these 3-digit ZIP codes were found in California, including San Jose (0.61%), which had the lowest proportion in the study. Eight were found in Florida, with South Florida (0.65%) having the lowest proportion in Florida. Yearly Changes in Blood Lead Level Distribution The distributions of patient blood lead levels showed yearover-year percentage increases in blood lead level groups #3.0 mg/dL (Table I). This increase outpaced the decrease in the 3.1-4.9 mg/dL blood lead group, resulting in a net decrease of those below the 2012 CDC 5 mg/dL threshold: 4
from 3.67% for May 2009-April 2010 to 2.59% for May 2014-April 2015. For the 5.0-9.9 mg/dL blood lead level group and the very high blood lead level group, the patient distributions showed year-over-year decreases during the same period (Table I). May 2014-April 2015 showed a slight reversal of these trends. The top 2.5% blood lead level threshold (97.5th percentile) was 5.1 mg/dL every year of the study. Yearly Changes in High Blood Lead Level for Various Risk Factors Table III illustrates the trends in high blood lead levels for all the risk factors mentioned previously across the 6 years of the McClure, Niles, and Kaufman
- 2016
Blood Lead Levels in Young Children: US, 2009-2015
Table II. Blood lead levels in children < 6 years old by state May 2009-April 2015 Total
£3.0 mg/dL
>3.0 and <5.0 mg/dL
‡5.0 and <10.0 mg/dL
State
‡10.0 mg/dL
‡5.0 mg/dL
Total
‡5.0 mg/dL
May 2014-April 2015 Total
‡5.0 mg/dL
Number (Percent) 2345 190 843 8530 35 703 76 520 5216 82 200 10 571 30 132 17 891 9869 7114 20 645 78 016 11 315 168 103 23 605 15 204 598 685 15 731 9484 2375 157 736 7865 11 859 56 864 18 528 2621 6562 123 700 26 911 8647 212 851 2474 11 229 873 944 386 743
2030 (86.57) 168 214 (88.14) 7598 (89.07) 31 715 (88.83) 68 697 (89.78) 4700 (90.11) 74 375 (90.48) 9616 (90.97) 27 427 (91.02) 16 491 (92.17) 9044 (91.64) 6611 (92.93) 19 295 (93.46) 72 651 (93.12) 10 561 (93.34) 158 835 (94.49) 22 356 (94.71) 14 372 (94.53) 562 660 (93.98) 14 995 (95.32) 8990 (94.79) 2243 (94.44) 150 431 (95.37) 7518 (95.59) 11 332 (95.56) 54 194 (95.30) 17 674 (95.39) 2520 (96.15) 6302 (96.04) 118 943 (96.15) 26 088 (96.94) 8415 (97.32) 207 015 (97.26) 2391 (96.65) 10 946 (97.48) 848 450 (97.08) 378 739 (97.93)
74 (3.16) 7728 (4.05) 324 (3.80) 1501 (4.20) 2705 (3.54) 170 (3.26) 2976 (3.62) 344 (3.25) 967 (3.21) 524 (2.93) 362 (3.67) 179 (2.52) 479 (2.32) 2123 (2.72) 318 (2.81) 3520 (2.09) 467 (1.98) 331 (2.18) 17 291 (2.89) 270 (1.72) 222 (2.34) 64 (2.69) 2849 (1.81) 129 (1.64) 205 (1.73) 1147 (2.02) 359 (1.94) 35 (1.34) 114 (1.74) 2140 (1.73) 309 (1.15) 80 (0.93) 2394 (1.12) 45 (1.82) 123 (1.10) 13 285 (1.52) 3802 (0.98)
184 (7.85) 11 633 (6.10) 461 (5.40) 1889 (5.29) 4099 (5.36) 255 (4.89) 3931 (4.78) 474 (4.48) 1411 (4.68) 647 (3.62) 388 (3.93) 261 (3.67) 663 (3.21) 2698 (3.46) 338 (2.99) 4447 (2.65) 616 (2.61) 417 (2.74) 14 880 (2.49) 371 (2.36) 225 (2.37) 56 (2.36) 3713 (2.35) 171 (2.17) 265 (2.23) 1309 (2.30) 423 (2.28) 53 (2.02) 105 (1.60) 2070 (1.67) 381 (1.42) 117 (1.35) 2753 (1.29) 30 (1.21) 131 (1.17) 10 262 (1.17) 3653 (0.94)
Absolute change reflects the difference in >5.0 mg/dL from the first to the second time periods.
57 (2.43) 3268 (1.71) 147 (1.72) 598 (1.67) 1019 (1.33) 91 (1.74) 918 (1.12) 137 (1.30) 327 (1.09) 229 (1.28) 75 (0.76) 63 (0.89) 208 (1.01) 544 (0.70) 98 (0.87) 1301 (0.77) 166 (0.70) 84 (0.55) 3854 (0.64) 95 (0.60) 47 (0.50) 12 (0.51) 743 (0.47) 47 (0.60) 57 (0.48) 214 (0.38) 72 (0.39) 13 (0.50) 41 (0.62) 547 (0.44) 133 (0.49) 35 (0.40) 689 (0.32) 8 (0.32) 29 (0.26) 1947 (0.22) 549 (0.14)
241 (10.28) 14 901 (7.81) 608 (7.13) 2487 (6.97) 5118 (6.69) 346 (6.63) 4849 (5.90) 611 (5.78) 1738 (5.77) 876 (4.90) 463 (4.69) 324 (4.55) 871 (4.22) 3242 (4.16) 436 (3.85) 5748 (3.42) 782 (3.31) 501 (3.30) 18 734 (3.13) 466 (2.96) 272 (2.87) 68 (2.86) 4456 (2.82) 218 (2.77) 322 (2.72) 1523 (2.68) 495 (2.67) 66 (2.52) 146 (2.22) 2617 (2.12) 514 (1.91) 152 (1.76) 3442 (1.62) 38 (1.54) 160 (1.42) 12 209 (1.40) 4202 (1.09)
638 41 568 1291 6420 16 095 2275 20 109 1236 7650 4180 1140 1011 2277 7564 1611 37 675 3468 1727 140 558 1818 1670 458 30 589 1593 2376 10 445 3395 296 1242 29 570 7343 1969 38 871 636 2518 200 413 72 951
60 (9.40) 4444 (10.69) 125 (9.68) 575 (8.96) 1324 (8.23) 165 (7.25) 1450 (7.21) 120 (9.71) 502 (6.56) 249 (5.96) 35 (3.07) 36 (3.56) 115 (5.05) 374 (4.94) 95 (5.90) 1815 (4.82) 96 (2.77) 49 (2.84) 5448 (3.88) 43 (2.37) 58 (3.47) 20 (4.37) 1179 (3.85) 33 (2.07) 84 (3.54) 257 (2.46) 126 (3.71) 4 (1.35) 32 (2.58) 810 (2.74) 125 (1.70) 31 (1.57) 699 (1.80) 15 (2.36) 41 (1.63) 3538 (1.77) 1070 (1.47)
345 21 782 1727 3668 10 250 284 9918 1792 4163 2470 572 801 3282 20 969 1965 23 036 3909 3368 72 265 7517 1616 352 23 340 805 1772 7952 3750 964 956 13 706 3031 1101 31 439 193 1749 107 443 59 019
26 (7.54) 1644 (7.55) 70 (4.05) 209 (5.70) 639 (6.23) 19 (6.69) 474 (4.78) 47 (2.62) 208 (5.00) 89 (3.60) 36 (6.29) 54 (6.74) 106 (3.23) 777 (3.71) 55 (2.80) 644 (2.80) 116 (2.97) 95 (2.82) 1799 (2.49) 264 (3.51) 40 (2.48) 19 (5.40) 512 (2.19) 23 (2.86) 34 (1.92) 226 (2.84) 85 (2.27) 26 (2.70) 30 (3.14) 251 (1.83) 70 (2.31) 22 (2.00) 519 (1.65) 5 (2.59) 17 (0.97) 1448 (1.35) 513 (0.87)
Absolute % change 1.86 3.14 5.63 3.26 2.00 0.56 2.43 7.09 1.56 2.36 3.22 3.18 1.82 1.23 3.10 2.02 0.20 0.02 1.39 1.14 0.99 1.03 1.66 0.79 1.62 0.38 1.44 1.35 0.56 0.91 0.61 0.43 0.15 0.23 0.66 0.42 0.60
5
ORIGINAL ARTICLES
MN PA KY OH CT WI MO NH MI LA MS SC IN MA WV IL OK TN NY NC AL UT MD AR KS GA CO NM OR NJ VA DE TX WA DC CA FL
May 2009-April 2010
May 2009April 2015
May 2009April 2010
May 2010April 2011
May 2011April 2012
May 2012April 2013
May 2013April 2014
May 2014April 2015
3 803 070
823 198
737 826
633 374
560 906
540 016
507 750
Significant trend
Number (Percent) 60 219 (3.09) 51 622 (2.80)
16 216 (3.86) 13 856 (3.47)
12 079 (3.20) 10 417 (2.91)
9357 (2.89) 8245 (2.68)
8063 (2.81) 6828 (2.51)
7379 (2.66) 6277 (2.40)
7125 (2.72) 5999 (2.44)
* *
660 387 662 327 661 936 659 314 627 376
7065 (1.07) 9866 (1.49) 13 763 (2.08) 24 269 (3.68) 37 382 (5.96)
1555 (1.21) 2124 (1.50) 3587 (2.50) 6706 (4.74) 10 766 (7.59)
1295 (1.05) 1993 (1.52) 2725 (2.13) 4743 (3.77) 7573 (6.27)
1097 (0.99) 1785 (1.59) 2288 (2.05) 3817 (3.51) 5615 (5.54)
1015 (1.01) 1411 (1.42) 1923 (1.95) 3310 (3.34) 4669 (5.25)
855 (0.89) 1262 (1.37) 1642 (1.78) 2966 (3.09) 4550 (4.94)
1248 (1.23) 1291 (1.51) 1598 (1.82) 2727 (3.10) 4209 (5.13)
P = .0446 P = .0290 * * *
2 665 330 769 566
76 338 (2.86) 20 224 (2.63)
22 060 (3.53) 4371 (3.35)
15 933 (2.94) 3492 (2.80)
11 504 (2.74) 3583 (2.50)
9531 (2.53) 3092 (2.44)
8899 (2.49) 2869 (2.11)
8411 (2.45) 2817 (2.60)
* *
703 782 624 981 643 452 698 627 597 828
12 213 (1.74) 12 270 (1.96) 16 374 (2.54) 22 743 (3.26) 28 627 (4.79)
2993 (2.02) 3116 (2.34) 4327 (3.19) 6424 (4.19) 7842 (6.25)
2266 (1.70) 2586 (2.10) 3334 (2.65) 4433 (3.35) 5684 (4.98)
2005 (1.73) 1947 (1.88) 2531 (2.35) 3476 (2.99) 4632 (4.58)
1610 (1.54) 1633 (1.75) 2100 (2.19) 3067 (2.97) 3899 (4.38)
1528 (1.54) 1566 (1.75) 2034 (2.20) 2752 (2.74) 3383 (3.91)
1811 (1.78) 1422 (1.73) 2048 (2.37) 2591 (2.78) 3187 (3.92)
* * * * *
657 317 666 236 692 538 614 839 637 740
25 909 (3.94) 21 880 (3.28) 19 336 (2.79) 14 025 (2.28) 11 077 (1.74)
7241 (5.06) 6152 (4.31) 4894 (3.37) 3605 (2.77) 2810 (2.08)
5045 (4.03) 4346 (3.46) 3883 (2.87) 2863 (2.38) 2166 (1.76)
4156 (3.78) 3359 (3.05) 3046 (2.62) 2183 (2.12) 1847 (1.76)
3487 (3.58) 2886 (2.91) 2672 (2.56) 1843 (2.05) 1421 (1.49)
3063 (3.28) 2631 (2.69) 2513 (2.53) 1673 (1.95) 1383 (1.51)
2917 (3.31) 2506 (2.77) 2328 (2.52) 1858 (2.17) 1450 (1.65)
* * * * *
165 624 722 385 406 681 509 539 262 144 264 833 96 232 21 372 874 401 11 700
9018 (5.44) 21 351 (2.96) 20 619 (5.07) 8359 (1.64) 11 431 (4.36) 5384 (2.03) 5494 (5.71) 590 (2.76) 12 215 (1.40) 233 (1.99)
1824 (7.30) 6258 (3.68) 5915 (6.91) 1673 (1.82) 3232 (5.68) 1081 (2.23) 1610 (7.02) 156 (3.92) 3540 (1.77) 58 (2.51)
1636 (7.39) 4997 (3.25) 4128 (5.23) 1598 (1.75) 1974 (4.41) 1012 (1.96) 1010 (5.55) 138 (3.69) 2328 (1.32) 51 (1.87)
1330 (5.06) 3342 (2.84) 2941 (4.35) 1368 (1.57) 1942 (4.15) 1041 (2.18) 866 (5.55) 63 (1.98) 2004 (1.41) 24 (1.07)
1147 (4.45) 2393 (2.42) 2822 (4.52) 1376 (1.68) 1740 (4.20) 816 (2.08) 774 (5.56) 57 (1.95) 1574 (1.24) 34 (1.87)
1607 (4.82) 2311 (2.41) 2439 (4.22) 1046 (1.40) 1331 (3.55) 661 (1.73) 654 (4.93) 69 (2.03) 1320 (1.09) 28 (2.11)
1474 (4.46) 2050 (2.38) 2320 (4.38) 1298 (1.57) 1212 (3.48) 773 (1.95) 580 (4.72) 107 (2.58) 1449 (1.35) 38 (2.96)
* * * * * * * * * P = .2356
Volume -
All values Number (Percent) reflect results of $5.0 mg/dL; CDC changed threshold to 5.0 mg/dL in May 2012. Notes: 1 - chi2 P < .01 for all years; 12 496 results did not have gender data. 2 - Cochran Armitage test for trend P < .01 for all years; 531 730 results did not have pre-1950 house construction data available. 3 - 368 174 results did not have payer data available. 4 - Cochran Armitage test for trend P < .01 for all years; 534 400 results did not have poverty data. 5 - 468 159 results did not have state data available.
www.jpeds.com
1 947 693 1 842 881
McClure, Niles, and Kaufman
Gender1 Male Female Pre-1950 housing quintile2 <3.6% 3.6-12.9% 13.0-29.9% 30.0-50.9% 51.0% + Payer type3 Private Medicaid/Medicare Low income (PIR <1.25) quintile4 <16.0% 16.0-28.9% 29.0-39.9% 40.0-51.9% 52.0% + High income (PIR >5) quintile4 <2.8% 2.8-6.8% 6.9-13.9% 14.0-27.9% 28.0% + HHS region5 1: CT, MA, ME, NH, RI, VT 2: NJ, NY 3: DE, DC, MD, PA, VA, WV 4: AL, FL, GA, KY, MS, NC, SC, TN 5: IL, IN, MI, MN, OH, WI 6: AR, LA, NM, OK, TX 7: IA, KS, MO, NE 8: CO, MT, ND, SD, UT, WY 9: AZ, CA, HI, NV 10: AK, OR, ID, WA
THE JOURNAL OF PEDIATRICS
6
Table III. Trends in high blood lead levels of children <6 years old
- 2016
ORIGINAL ARTICLES
Table IV. Logistic models: Factors associated with high blood lead levels and very high blood lead levels Factors associated with high BLL (‡5.0 mg/dL)
Study year Private payer Male Top housing quintile Top 3 HHS regions (1,3,7) Top poverty quintile Top wealth quintile
Factors associated with very high BLL (‡10.0 mg/dL)
Unadjusted OR
95% CI
aOR
95% CI
Unadjusted OR
95% CI
aOR
95% CI
0.93 1.09 1.11 2.99 2.43 2.06 0.56
(0.92-0.93) (1.08-1.11) (1.09-1.12) (2.94-3.02) (2.40-2.46) (2.03-2.09) (0.54-0.57)
0.92 0.84 1.11 2.50 2.23 1.64 0.65
(0.91-0.92) (0.82-0.85) (1.09-1.13) (2.46-2.54) (2.20-2.26) (1.61-1.67) (0.64-0.67)
0.94 1.06 1.10 2.96 2.48 1.95 0.53
(0.93-0.95) (1.02-1.10) (1.07-1.13) (2.87-3.04) (2.40-2.55) (1.88-2.01) (0.51-0.55)
0.93 0.80 1.08 2.58 2.21 1.55 0.62
(0.92-0.94) (0.77-0.83) (1.04-1.11) (2.50-2.68) (2.13-2.29) (1.50-1.61) (0.59-0.65)
BLL, blood lead level; CI, confidence interval; OR, odds ratio.
study. Male subjects maintained a greater proportion of high blood lead level than female subjects during each year of the study, and this difference remained similar over the years (0.21%-0.39%). The statistically significant trend among ZIP codes with increasing proportions of pre-1950s housing quintiles and PIR <1.25 remained when we examined data from each year of the study individually. Patients with private payer insurance had larger rates of high blood lead levels than patients with Centers for Medicare and Medicaid Services payers each year from May 2009-April 2014 but exhibited a lower rate of high blood lead levels than those with Centers for Medicare and Medicaid Services payers during the final year of the study. HHS Regions 1, 3, and 7 all exhibited a substantial decline in high blood lead level proportion during the study period but were the only regions with >4.0% high blood lead levels during the final year of the study. HHS Region 10 was the only region that had an increase in high blood lead levels during the study period and was the only factor examined that did not demonstrate a statistically significant (P < .05) downward trend in high blood lead level over the study period. High Blood Lead Level and Very High Blood Lead Level Models Logistic regression models were used to analyze the impact of the factors examined in Table I on high blood lead levels and very high blood lead levels. In unadjusted models, the strongest measures of association with high blood lead levels were being in the greatest pre-1950s housing quintile (OR 2.99, 95% CI 2.95-3.03) and living in HHS Region 1, 3, or 7 (OR 2.43, 95% CI 2.40-2.46). Living in a region in the lowest income quintile also was associated with high blood lead level (OR 2.06, 95% CI 2.03-2.09), whereas living in a region in the top income quintile had a significant protective effect (OR for high blood lead level = 0.56, 95% CI 0.54-0.57). The measures of association for these factors were similar in the adjusted multivariable model (Table IV). Private payer status had a small but significant association with high blood lead levels in the unadjusted model (OR 1.09, 95% CI 1.08-1.11) but had a significant protective effect in the adjusted model (aOR 0.84, 95% CI 0.82-0.85). Measures of association in the very high blood lead level models (both adjusted and unadjusted) were similar to those in the high blood lead level models (Table IV). Blood Lead Levels in Young Children: US, 2009-2015
Discussion Reducing blood lead levels in children has been and continues to be a major public health success. The declining blood lead levels are the result of public health initiatives that include the removal of leaded gasoline, the banning of lead paint, effective treatment of potable water supplies, and remediation of homes found to be contaminated with lead. Studies conducted by the CDC3,19 have provided valuable insight into the changing lead levels in children and associated risk factors. The millions of test results reported by individual states have allowed the CDC to track the changes in lead levels over time.19 The demographic factors associated with high blood lead levels, however, come from a much smaller set: n = 793 from 2007-2010 for ages 1-2 years19 and n = 1653 from 2007-2010 for ages 1-5 years.13 The study presented here examines similar factors in a much larger data set of 3.8 million children. The CDC studies also were unable to look at factors associated with blood lead level $10.0 mg/dL because of small numbers (9 children in 2007-2008; 6 children in 2009- 2010).3 In contrast, our study was able to examine factors associated with very high venous blood lead level results (n = 22 237). Our study provides insights for the time period through mid-2015, 4.5 years after the end of data collection in the most recent NHANES report, which included data through the end of 2010. This timeframe extension enabled the examination of the continuing decline in high blood lead level and very high blood lead level levels. This was true for the population as whole (from 3.67% to 2.59% high blood lead levels and from 0.74% to 0.55% very high blood lead levels) and for most demographic groups and geographic regions. The results also demonstrate a slight increase in the incidence of high blood lead levels and very high blood lead levels in the final year of the study. After years of consecutive decline, the rise was unexpected, especially given that there was no major observable change in demographic proportions during the final year. Future studies are needed to evaluate the evolution of high blood lead levels and very high blood lead levels. To a large extent, the results presented here confirm the significance of factors examined by the CDC3,19 and exhibit similar measures of association. Our study, using a different methodology, confirms the findings in NHANES associating 7
THE JOURNAL OF PEDIATRICS
www.jpeds.com
pre-1950s housing data and low PIR with high blood lead level. Examining PIR and housing construction data by ZIP code instead of direct patient data is a limiting factor but also highlights the ability of physicians to identify geographic regions with increased risk, particularly in the event that individual patient risk factors are unavailable or difficult to obtain. Using data available on the US Census Bureau website, health care experts can determine the proportion of housing constructed before 1950 and PIR of the ZIP codes they serve to help assess the level of risk in their patients, even if individual factors are not available. All previously reported NHANES cycles found increased levels of high blood lead level in Medicaid patients (the differences in 1999-2002 and in 2003-2006 were statistically significant). The current Quest Diagnostics Health Trends study found that children with private payers had a significantly larger rate of high blood lead level than those with Medicaid/Medicare payers. In the adjusted model, however, private payer status had a significant protective effect. Thus, with all factors explored in the study considered together, the private payer results seem to indicate the same relationship found in the NHANES studies. There were also some differences between our findings and those of the NHANES studies. The 2003-2006 and 2007-2010 NHANES cycles both found female participants to have slightly greater proportions of high blood lead level than male participants (although these differences were not found to be statistically significant). In the present study, male participants exhibited a small but statistically significantly larger proportion of high blood lead levels than female participants. Our study examined results from patients living in all 50 states and the District of Columbia; however, analyses by state were limited to the 36 states and the District of Columbia with more than 2000 specimens during the study period. Our results for high blood lead level at the state level are often similar (within 3%) to the results reported by the CDC in their national surveillance data (2010-2014),20 but there are also interesting differences. The rates of high blood lead level in our study were considerably lower than those reported by the CDC for New Hampshire (6.2% vs 12.0%), Mississippi (4.5% vs 9.2%), and Illinois (4.5% vs 7.9%). Only Minnesota showed a considerably larger rate of high blood lead level in our study than in the CDC report (10.3% vs 2.9%). At the national level, we found a virtually identical rate of very high blood lead level (0.58%) as the CDC (0.57%) over a similar 5-year period, but the rate of high blood lead level was lower in our study (2.95% vs 5.48%). The reasons for these differences are unclear but could reflect the inclusion of capillary blood results in the CDC data. According to the CDC, “houses built before 1950 pose the greatest hazard to children because they are much more likely to contain lead-based paint than newer houses.”16 Leadcontaminated dust on floors, windowsills, and window wells is associated with elevated blood lead levels in children.21 Our findings showed that, in general, the regions with the greatest adjusted level of pre-1950s housing construction had the 8
Volume greatest proportions of patients with high blood lead levels. HHS Region 2 (New York and New Jersey) was an interesting deviation from this trend. Despite having the greatest level of housing built before 1950 (based on study participant ZIP codes), the overall levels of high blood lead level were below average for the study. It is also interesting that New York State (3.1%) was near the national average (3.0%) for high blood lead level while having several 3-digit ZIP codes with the largest proportions of high blood lead levels and very high blood lead levels found in the study. The impact of abatement and active lead surveillance in these states is also unclear. In general, southern regions had lower proportions of high blood lead levels. This includes Florida, with more than 386 000 specimens and 1.1% with high blood lead levels, and California, with more than 873 000 specimens and 1.4% with high blood lead levels. We have no easy explanation why 12 of the 36 states and the District of Columbia, with sufficient number of results, had an increase in the proportion of high blood lead levels between the first year and the final year of the study. We found that living in ZIP codes with greater percentages of residents <1.25 PIR exhibited a stronger association with high blood lead levels than individual payer type. The reasons for this perplexing difference are unclear. One possibility may be that ZIP codes with greater levels of low income also have greater proportions of housing built before 1950 or other associated factors. Although this relationship may exist, the multivariable models indicate that it is not reason for the association, because both variables maintain significant associations with high blood lead levels and very high blood lead levels in the adjusted multivariable models. All risk factors associated with high blood lead levels also were associated with very high blood lead levels. In both unadjusted and adjusted models, the measures of association were remarkably similar. This finding is reassuring because the risk factors identified in previous studies were not only confirmed for the most part in this larger study but found to be associated with very high blood lead levels as well. The major strengths of this study are its large size, national representation with data from 50 states and the District of Columbia, and the inclusion of data through mid-2015. Analyses were conducted on more than 5 million results for children <6 years of age. The analyses focused on more than 3.8 million results of venous blood draws from children <6 years of age, consistent with the CDC definition of confirmed elevated blood lead level. This study also examined demographics for the vast majority of participants. Having demographics available for so many participants enabled analysis of factors associated with very high blood lead levels. This study also had limitations. The 3.0 mg/dL reporting threshold precluded our ability to estimate the mean blood lead level for the study population. Some blood lead level groups are narrower than the range of instrument variability, but we assume there is no bias in the data set. It is also possible that some patients were tested as a follow-up to prior blood lead level results at another laboratory or because healthcare providers suspected a high probability of elevated McClure, Niles, and Kaufman
- 2016
ORIGINAL ARTICLES
blood lead level. It also may be possible that population segments or regional populations deemed to be at risk are being tested more frequently. Sociological, familial, and environmental factors also may play a role in determining who is tested for lead; however, with our 97.5th percentile being the same as found in NHANES, we feel the selection bias is minimal. Quest Diagnostics does not perform all lead testing in the country, and these data should only be seen as a large sample of national data. These data do not necessarily reflect the population as a whole but reflect those tested in medical practices in the US. In summary, progress in reducing the burden of lead toxicity is a public health success story that is incomplete. This analysis of more than 5 million blood lead level results over a 6-year period, extending through April 2015, includes 3.8 million venous blood lead level results for infants and children <6 years of age. This allowed for a robust, detailed analysis of results by geography and other criteria that are impossible with the narrower NHANES approach. Unique observations include the correlation of results for patients with blood lead level results of $10.0 mg/dL to those with results of 5.1-10.0 mg/dL. Many of the 3-digit ZIP code regions with high blood lead levels and very high blood lead levels are in New York, Pennsylvania, and Ohio. Additional study will be necessary to assure that the improvements reported will continue and that specific state and local efforts achieve the desired goals. n Submitted for publication Jan 26, 2016; last revision received Mar 29, 2016; accepted May 3, 2016. Reprint requests: Leland F. McClure, PhD, Director, Medical Science Liaison, Medical Affairs - Quest Diagnostics, 11636 Administration Drive, St Louis, MO 63146. E-mail: [email protected]
References 1. Agency for Toxic Substances and Disease Registry. Toxicological profile for lead. Atlanta, GA: US Department of Health and Human Services, CDC, Agency for Toxic Substances and Disease Registry; 2007. 2. Parsons PJ, Reilly AA, Esernio-Jenssen D. Screening children exposed to lead: an assessment of the capillary blood lead fingerstick test. Clin Chem 1997;43:302-11. 3. Centers for Disease Control and Prevention. Blood lead levels in children aged 1-5 years – United States, 1999-2010. MMWR Morb Mortal Wkly Rep 2013;62:245-8. 4. Pirkle JL, Brody DJ, Gunter EW, Kramer RA, Paschal DC, Flegal KM, et al. The decline of blood lead levels in the United States: the National Health and Nutrition Examination Surveys (NHANES). JAMA 1994; 272:284-91.
Blood Lead Levels in Young Children: US, 2009-2015
5. Centers for Disease Control and Prevention. Preventing lead poisoning in young children. Atlanta: CDC; 1991. 6. CDC Advisory Committee on Childhood Lead Poisoning Prevention. Low level lead exposure harms children: A renewed call for primary prevention. Atlanta: CDC; 2012. 7. Chandramouli K, Steer CD, Ellis M, Emond AM. Effects of early childhood lead exposure on academic performance and behavior of school age children. Arch Dis Child 2009;94:844-8. 8. Nigg JT, Nikolas M, Mark Knottnerus G, Cavanagh K, Friderici K. Confirmation and extension of association of blood lead level with attention-deficit/hyperactivity disorder (ADHD) and ADHD symptom domains at population-typical exposure levels. J Child Psychol Psychiatry 2010;51:58-65. 9. Needleman HL, Schell A, Bellinger D, Leviton A, Allred EN. The longterm effects of exposure to low doses of lead in childhood. An 11-year follow-up report. N Engl J Med 1990;332:83-8. 10. Bellinger DC, Stiles KM, Needleman HL. Low-level lead exposure, intelligence and academic achievement: a long-term follow-up study. Pediatrics 1992;90:855-61. 11. Rogan WJ, Dietrich KN, Ware JH, Dockery DW, Salganik M, Radcliffe J, et al. The effect of chelation therapy with succimer on neuropsychological development in children exposed to lead. N Engl J Med 2001;344: 1421-6. 12. Centers for Disease Control and Prevention. Response to Advisory Committee on Childhood Lead Poisoning Prevention Recommendations in “Low Level Exposure Harms Children: A Renewal Call of Primary Prevention” CDC, http://www.cdc.gov/nceh/lead/acclpp/cdc_response_ lead_exposure_recs.pdf. Accessed May 10, 2016. 13. Centers for Disease Control and Prevention. Screening young children for lead poisoning: Guidance for state and local public health officials; Appendix C1: The Lead Laboratory, www.cdc.gov/nceh/lead/ publications/screening.htm. Accessed December 12, 2015. 14. CDC Standard Surveillance Definitions and Classifications, http://www. cdc.gov/nceh/lead/data/definitions.htm. Accessed December 12, 2015. 15. US Census Bureau, 2009-2013 5-Year American Community Survey, http://factfinder.census.gov/faces/tableservices/jsf/pages-/productview. xhtml?pid=ACS_13_5YR_B25034&prodType=table. Accessed December 12, 2015. 16. CDC. Facts on Lead, http://www.cdc.gov/nceh/lead/publications/1997/ factlead.htm. Accessed December 12, 2015. 17. US Census Bureau, 2008-2012 5-Year American Community Survey, http://factfinder.census.gov/faces/tableservices/jsf/pages/-productview. xhtml?pid=ACS_12_5YR_B17024&prodType=table. Accessed December 12, 2015. 18. SAS, version 9.4. Cary, NC: SAS Institute. 19. Raymond J, Wheeler W, Brown MJ. Lead screening and prevalence of blood lead levels in children aged 1–2 years—Child Blood Lead Surveillance System, United States, 2002–2010 and National Health and Nutrition Examination Survey, United States, 1999–2010. MMWR Suppl 2014;63:36-42. 20. Centers for Disease Control and Prevention. CDC’s National Surveillance Data (1997-2014), http://www.cdc.gov/nceh/lead/data/national. htm. Accessed March 10, 2016. 21. Lanphear BP, Weitzman M, Winter NL, Eberly S, Yakir B, Tanner M, et al. Lead-contaminated house dust and urban children’s blood lead levels. Am J Public Health 1996;86:1416-21.
9
C
hildhood lead toxicity is a preventable environmental disease with long-lasting adverse health and behavioral effects.1 Public health services and other health professionals throughout the US have dedicated more than 4 decades of efforts to screen children, especially those at high risk, for lead exposure and to identify primary sources of lead.2 Federal and local environmental policies have included the removal of lead from gasoline, reduction of lead in paints, and testing of homes for lead-based paint. These efforts along with laboratory testing and case management efforts have been instrumental in significantly reducing blood lead levels in the US. The 2007-2010 National Health and Nutrition Examination Survey (NHANES) estimate of the geometric mean blood lead level was 1.3 mg/dL,3 which is a 90% decrease compared with the 1976-1980 NHANES II 12.8 mg/dL estimate.4 In 1991, the Centers for Disease Control and Prevention (CDC) recommended changes for preventing childhood lead poisoning, which included a reduction for the blood lead level deemed safe (from 25 mg/dL to 10 mg/dL).5 In May 2012, the CDC Advisory Committee on Childhood Lead Poisoning Prevention (ACCLPP) identified that there is no safe blood lead level and the CDC accepted ACCLPP recommendations to remove all CDC blood lead level references to “blood lead level of concern.”6 The CDC position of “no safe blood lead level” is based on an absence of blood lead levels without effects and low blood lead levels that are associated with intellectual deficits, attention deficit behaviors, and poor academic achievement.7,8 That these effects appear to be irreversible9-11 emphasizes a public health care shift to primary prevention rather than secondary and tertiary prevention efforts, which are based on responses after detecting lead exposure. In May 2012, the CDC also adopted the ACCLPP committee recommendations to use the NHANES 97.5th blood lead level percentile (5.0 mg/dL) as an upper reference interval threshold to identify children with elevated blood lead levels. The 5.0 mg/dL value is based on 2 consecutive cycles of the NHANES blood lead level data distribution among study children 1-5 years of age. Based on the 5.0 mg/dL threshold, the 2012 ACCLPP committee report estimated 450 000 children in the US as having blood lead levels greater than the new reference limit.12 The NHANES analysis includes demographic categories with long-standing disparities of risk for elevated blood From the Quest Diagnostics, Madison, NJ
ACCLPP CDC HHS NHANES PIR
Advisory Committee on Childhood Lead Poisoning Prevention Centers for Disease Control and Prevention US Department of Health and Human Services National Health and Nutrition Examination Survey Poverty income ratio
Funded by Quest Diagnostics, which provided support in the form of salaries for all authors but did not have any additional role in the study design, collection, analysis and interpretation of data, writing of the manuscript, or decision to publish. The authors declare no conflicts of interest. 0022-3476/$ - see front matter. ª 2016 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2016.05.005
1
THE JOURNAL OF PEDIATRICS
www.jpeds.com
lead levels, including age, sex, race/ethnicity, age of housing, poverty income ratio (PIR), and Medicaid enrollment status. Despite the insights provided by the NHANES analysis, the study has several limitations. One such limitation is that the low numbers of NHANES-enrolled children with blood lead levels $10 mg/dL (only 9 children in 2007-2008; 6 children in 2009-2010) make interpretation of population estimates of very high blood lead levels unreliable. In addition, the NHANES was not designed to produce estimates at the state and local level and may not detect statistically significant disparities with important public health implications. This Quest Diagnostics Health Trends report describes the results of a 6-year retrospective study based on a large national clinical laboratory database with more than 5 million results from children younger than 6 years of age. Our analysis builds upon previously reported NHANES data and includes insights into yearly trends and the distributions of blood lead levels by specimen type (venous and capillary), sex, payer type, US Department of Health and Human Services (HHS) region, residence state, PIR, and pre-1950s housing construction.
Methods The specimen requirement for venous blood lead level analysis is whole blood collected into an evacuated collection tube certified for lead testing, such as tan-top and royal blue-top tubes containing the anticoagulant EDTA. For the capillary collection method, the specimen collection container is the lavender-top capillary tube. The blood lead level analyses were performed by the use of either inductively coupled plasma/mass spectrometry or the Zeeman graphite furnace atomic absorption spectroscopy. Instrument calibrations are performed with standards traceable to the National Institutes of Standards and Technology. Performance for all methods is in compliance with the 4 mg/dL (or 10%, whichever is greater) CDC accuracy standards.13 The blood lead level results were evaluated with a 3.0 mg/dL lower reporting threshold. The laboratory analysis of venous specimens is consistent with the CDC definition for “confirmed elevated blood lead level” when indicated.14 The blood lead level data set includes deidentified results of testing performed for children <6 years of age, from May 2009 through April 2015 (3 years before and after the 2012 CDC change from the 10 mg/dL “level of concern” to the 5.0 mg/dL reference interval threshold). Instances of blood lead level results reported as a specimen submitted in a tube/container not certified for lead testing were excluded from the study. This study was deemed exempt by the Western Institutional Review Board. To avoid duplication of patient data, when 2 or more tests were associated with the same individual, only the first venous result (or the first capillary result if there were no venous results) within the data set was included in this study. The 3.0 mg/dL reporting threshold precluded our ability to estimate the mean blood lead level for the study with sufficient precision. Instead, analyses focused on the proportions 2
Volume of the population falling into each of 4 blood lead level groups: #3.0 mg/dL (below the reporting limit); 3.1-4.9 mg/dL (above the reporting limit and below the CDC 2012 reference interval threshold); 5.0-9.9 mg/dL (between the 2012 reference interval threshold and the previous 1991 CDC “level of concern”); and $10.0 mg/dL. Patient data were limited to patients <6 years of age, corresponding to the CDC age definition for high risk. Blood lead levels results missing patient sex were excluded from sex analysis. Data from the US Census Bureau’s 2009-2013 American Community Survey 5-Year Estimates15 were used to determine the proportion of housing constructed before 1950 by ZIP code. According to the CDC, “houses built before 1950 pose the greatest hazard to children because they are much more likely to contain lead-based paint than newer houses.”16 Quintiles were defined as the percentage of the housing category by ZIP code. Quintile thresholds for pre-1950s housing were defined as <3.6%, 3.6%-12.9%, 13.0%-29.9%, 30.0%50.9%, and $51.0%. All quintile thresholds were chosen to provide approximately equal numbers in each quintile group. Demographics were divided into quintiles to demonstrate trends in blood lead level proportions. ZIP codes are based on patient residence, not the site of the blood collection. Data from the United States Census Bureau’s 2008-2012 American Community Survey 5-Year Estimates17 were used to determine PIR of children’s area of residence by ZIP code. Quintiles were defined as the percentage of PIR <1.25 (low income) and PIR >5 (high income) by ZIP code. Quintile ranges were defined as <16.0%, 16.0%-27.9%, 28.0%38.9%, 39.0%-51.9%, and $52.0% for PIR <1.25, and <2.8%, 2.8%-6.9%, 7.0%-13.9%, 14.0%-27.9%, and $28.0% for PIR >5. This study included specimens submitted for blood lead level testing from all 50 states and the District of Columbia. Data were grouped for analysis by HHS region, state, and 3-digit ZIP code region. We limited our state analyses to those with at least 2000 children and our 3-digit ZIP code analysis to areas with at least 1000 children. The proportion of housing that was constructed before 1950 in various geographical regions also was analyzed. These data were weighted by the number of patients with specimens from individual ZIP codes. Statistical Analyses The Cochran-Armitage test was used to analyze trends in proportions of children with blood lead levels $5.0 mg/dL (high blood lead level) and $10.0 mg/dL (very high blood lead level) for various groups. Testing for statistical significance between the 2 groups was conducted with the c2 test. Multivariable logistic regression models to determine characteristics associated with high blood lead level and very high blood lead level also are reported. Variables in both models were chosen based on plausibility and/or statistical significance in previous studies.3 Living in ZIP codes associated with the greatest quintile of pre-1950s housing, low income, and high income were included as binary variables. Living in McClure, Niles, and Kaufman
- 2016
ORIGINAL ARTICLES
HHS regions 1, 3, or 7 also was included as a binary variable. Despite the small measures of association determined by unadjusted ORs in logistic regression models, year of service, patient age, sex, and payer status were included in the multivariable models because of their potential as confounding factors. Results of adjusted and unadjusted models were stated as ORs and 95% CIs. Data were analyzed with SAS, version 9.4 (SAS Institute, Cary, North Carolina).18
Results This study included 5 266 408 blood lead levels from children <6 years of age. Venous blood draws represent 72.2% of all specimen submissions (n = 3 803 070). Blood lead levels were below the reporting threshold of #3.0 mg/dL in 94.9% of venous specimens; 3.1-4.9 mg/dL in 2.2%; and $5.0 mg/ dL in 3.0%. Results from the 1 178 000 capillary specimens were slightly, but statistically significantly (P < .01), more likely to fall into greater blood lead level categories: capillary blood lead levels were #3.0 mg/dL in 93.6%, 3.1-4.9 mg/dL in 3.3%, and $5.0 mg/dL in 3.1%. There also was a group of 285 338 specimens with unknown draw type with the following blood lead level distribution: 95.8% #3.0 mg/dL, 2.3% between 3.0 and 4.9 mg/dL, and 1.9% $5.0 mg/dL. The remainder of the results section will focus on the 3 803 070 children with venous blood results. Sex Of the venous specimens tested, 1 947 693 (51.2%) were from boys and 1 842 881 (48.5%) were from girls; the other 12 496 (0.3%) specimens had no sex information. High blood lead level was only slightly greater for boys (3.1%) than girls (2.8%), although the difference was statistically significant (P < .01; Table I). Pre-1950s Housing Construction Living in an area with a high proportion of pre-1950 housing construction was strongly associated with having a high blood lead level (Table I). Living in a ZIP code where $51.0% of housing units were constructed before 1950 (the highest quintile) was associated with a significantly larger proportion of high blood lead levels (OR 5.86, 95% CI 5.716.01) or very high blood lead levels (OR 6.34, 95% CI 5.976.74) than living in an area with the lowest quintile. There was a statistically significant trend for both high blood lead levels and very high blood lead levels among the pre-1950s housing quintiles (both P < .01, Table I). Payer Type and PIRs Approximately 20% of patients (n = 768 879) with venous blood draws had Medicaid as the payer and 687 had Medicare as the payer. Patients with Medicaid or Medicare had a slightly, but statistically significantly, lower frequency of high blood lead level than did those with private payers (2.6 vs 2.9%; P < .01). Comparisons of other lead groups between payer types were also similar (Table I). The data also showed a Blood Lead Levels in Young Children: US, 2009-2015
positive association between percentage of PIR <1.25 in patient ZIP codes and high blood lead level (Table I). Children living in ZIP codes in which $52.0% have PIR <1.25 (the highest quintile) had a greater proportion of high blood lead level (OR 2.85, 95% CI 2.79-2.91) and very high blood lead level (OR 2.67, 95% CI 2.54-2.80) compared with those living in the lowest quintile. There was a statistically significant trend in both high blood lead level and very high blood lead level among the poverty quintiles (Table I). There was an inverse association between the greater-income ZIP codes and high blood lead level. Those living in ZIP codes with the greatest percentage of PIR >5 (the highest quintile) were much less likely to exhibit high blood lead levels (OR 0.43, 95% CI 0.42-0.44) and very high blood lead level (OR 0.43, 95% CI 0.41-0.46) than were those in the lowest quintile. There was a statistically significant trend in both high blood lead level and very high blood lead level among the PIR >5 quintiles (Table I). Geographic Regions Blood lead levels group results were analyzed by HHS region (Table I). Region 7 (Iowa, Kansas, Missouri, and Nebraska) had the largest proportion of both high blood lead levels (5.7%) and very high blood lead levels (1.1%). Region 1 (Connecticut, Massachusetts, Maine, New Hampshire, Rhode Island, and Vermont) and Region 3 (Delaware, District of Columbia, Maryland, Pennsylvania, Virginia, and West Virginia) also had notably large proportions of high blood lead levels (5.4% and 5.1%, respectively), and very high blood lead levels (1.0% and 1.1%, respectively). The states with the largest proportions of high blood lead levels were Minnesota (10.3%), Pennsylvania (7.8%), Kentucky (7.1%), Ohio (7.0%), and Connecticut (6.7%) (Table II). All these states and 20 others (25/37, 68%) exhibited a decline in the proportion of high blood lead levels between the first year and the final year of the study. New Hampshire had the largest absolute decline (from 9.7% to 2.6% high blood lead levels), and Mississippi had the largest absolute increase (from 3.1% to 6.3% high blood lead levels) between the first year and the final year of the study. Florida and California had the lowest proportions of high blood lead levels (1.1% and 1.4%, respectively) and very high blood lead levels (0.1% and 0.2%, respectively). Data from 325 3-digit ZIP code regions with more than 1000 specimens (36.0% of all 3-digit ZIP codes with specimens in the study) also were analyzed to assess their impact on state data. In 6 regions, >14.0% of specimens had high blood lead levels: Syracuse, NY (40.1%), Buffalo, NY (18.8%), Cincinnati, OH (16.0%), Poughkeepsie, NY (14.9%), York, PA (14.4%), and Oil City, PA (14.0%). The 11 regions with the largest proportions of specimens with very high blood lead levels also were all in New York, Pennsylvania, or Ohio. Syracuse, NY (16.0%), Buffalo, NY (6.0%), York, PA (5.5%), Poughkeepsie, NY (4.9%), and Oil City, PA (4.3%) had the greatest rates in the study. 3
THE JOURNAL OF PEDIATRICS
Volume -
www.jpeds.com
Table I. Blood lead levels in children <6 years old: 2009-2015
Total Sex1 Male Female Pre-1950s housing quintile2 <3.6% 3.6-12.9% 13.0-29.9% 30.0-50.9% $51.0% Payer type3 Private Medicaid/Medicare Low income (PIR <1.25) quintile4 <16.0% 16.0-27.9% 28.0-38.9% 39.0-51.9% $52.0% High income (PIR >5) quintile4 <2.8% 2.8-6.9% 7.0-13.9% 14.0-27.9% $28.0% HHS region5 1: CT, MA, ME, NH, RI, VT 2: NJ, NY 3: DE, DC, MD, PA, VA, WV 4: AL, FL, GA, KY, MS, NC, SC, TN 5: IL, IN, MI, MN, OH, WI 6: AR, LA, NM, OK, TX 7: IA, KS, MO, NE 8: CO, MT, ND, SD, UT, WY 9: AZ, CA, HI, NV 10: AK, OR, ID, WA Time period (Year) May 2009-April 2010 May 2010-April 2011 May 2011-April 2012 May 2012-April 2013 May 2013-April 2014 May 2014-April 2015
‡10.0 mg/dL (Very high BLL)
‡5.0 mg/dL (High BLL)
Number (Percent) 90 092 (2.37)
22 237 (0.58)
112 329 (2.95)
44 032 (2.26) 38 384 (2.08)
48 312 (2.48) 41 384 (2.25)
11 907 (0.61) 10 238 (0.56)
60 219 (3.09) 51 622 (2.80)
647 545 (98.06) 643 939 (97.22) 637 074 (96.24) 617 464 (93.65) 564 372 (89.96)
5777 (0.87) 8522 (1.29) 11 099 (1.68) 17 581 (2.67) 25 622 (4.08)
5817 (0.88) 7994 (1.21) 11 116 (1.68) 19 334 (2.93) 29 936 (4.77)
1248 (0.19) 1872 (0.28) 2647 (0.40) 4935 (0.75) 7446 (1.19)
7065 (1.07) 9866 (1.49) 13 763 (2.08) 24 269 (3.68) 37 382 (5.96)
2 665 330 769 566
2 531 010 (94.96) 734 185 (95.40)
57 982 (2.18) 15 157 (1.97)
61 863 (2.32) 16 288 (2.12)
14 475 (0.54) 3936 (0.51)
76 338 (2.86) 20 224 (2.63)
703 782 624 981 643 452 698 627 597 828
682 503 (96.98) 602 837 (96.46) 614 280 (95.47) 658 817 (94.30) 549 471 (91.91)
9066 (1.29) 9874 (1.58) 12 798 (1.99) 17 067 (2.44) 19 730 (3.30)
9779 (1.39) 9816 (1.57) 13 141 (2.04) 18 214 (2.61) 23 144 (3.87)
2434 (0.35) 2454 (0.39) 3233 (0.50) 4529 (0.65) 5483 (0.92)
12 213 (1.74) 12 270 (1.96) 16 374 (2.54) 22 743 (3.26) 28 627 (4.79)
657 317 666 236 692 538 614 839 637 740
613 075 (93.27) 627 939 (94.25) 658 850 (95.14) 590 014 (95.96) 618 030 (96.91)
18 333 (2.79) 16 417 (2.46) 14 352 (2.07) 10 800 (1.76) 8633 (1.35)
21 018 (3.20) 17 607 (2.64) 15 261 (2.20) 11 198 (1.82) 9010 (1.41)
4891 (0.74) 4273 (0.64) 4075 (0.59) 2827 (0.46) 2067 (0.32)
25 909 (3.94) 21 880 (3.28) 19 336 (2.79) 14 025 (2.28) 11 077 (1.74)
165 624 722 385 406 681 509 539 262 144 264 833 96 232 21 372 874 401 11 700
151 422 (91.43) 681 603 (94.35) 374 655 (92.13) 494 543 (97.06) 244 002 (93.08) 255 900 (96.63) 87 446 (90.87) 20 346 (95.20) 848 895 (97.08) 11 255 (96.20)
5184 (3.13) 19 431 (2.69) 11 407 (2.80) 6637 (1.30) 6711 (2.56) 3549 (1.34) 3292 (3.42) 436 (2.04) 13 291 (1.52) 212 (1.81)
7306 (4.41) 16 950 (2.35) 16 313 (4.01) 7085 (1.39) 8849 (3.38) 4240 (1.60) 4424 (4.60) 503 (2.35) 10 268 (1.17) 172 (1.47)
1712 (1.03) 4401 (0.61) 4306 (1.06) 1274 (0.25) 2582 (0.98) 1144 (0.43) 1070 (1.11) 87 (0.41) 1947 (0.22) 61 (0.52)
9018 (5.44) 21 351 (2.96) 20 619 (5.07) 8359 (1.64) 11 431 (4.36) 5384 (2.03) 5494 (5.71) 590 (2.76) 12 215 (1.40) 233 (1.99)
823 198 737 826 633 374 560 906 540 016 507 750
770 737 (93.63) 698 385 (94.65) 602 166 (95.07) 534 990 (95.38) 516 139 (95.58) 485 457 (95.61)
22 217 (2.70) 16 847 (2.28) 13 522 (2.13) 10 955 (1.95) 10 184 (1.89) 9142 (1.80)
24 185 (2.94) 18 449 (2.50) 14 171 (2.24) 11 993 (2.14) 10 941 (2.03) 10 353 (2.04)
6059 (0.74) 4145 (0.56) 3515 (0.55) 2968 (0.53) 2752 (0.51) 2798 (0.55)
30 244 (3.67) 22 594 (3.06) 17 686 (2.79) 14 961 (2.67) 13 693 (2.54) 13 151 (2.59)
Total
£3.0 mg/dL
>3.0 and <5.0 mg/dL
3 803 070
3 607 874 (94.87)
82 867 (2.18)
1 947 693 1 842 881
1 843 442 (94.65) 1 752 875 (95.12)
660 387 662 327 661 936 659 314 627 376
‡5.0 and <10.0 mg/dL
PIR, povery income ratio. Notes: 1 - 12 496 results did not have gender data. 2 - 531 730 results did not have pre-1950 house construction data available. 3 - 368 174 results did not have payer data available. 4 - 534 400 results did not have poverty/wealth data. 5 - 468 159 results did not have state data available.
In 49 3-digit ZIP codes, <1.0% of specimens had high blood lead levels. Eighteen of these 3-digit ZIP codes were found in California, including San Jose (0.61%), which had the lowest proportion in the study. Eight were found in Florida, with South Florida (0.65%) having the lowest proportion in Florida. Yearly Changes in Blood Lead Level Distribution The distributions of patient blood lead levels showed yearover-year percentage increases in blood lead level groups #3.0 mg/dL (Table I). This increase outpaced the decrease in the 3.1-4.9 mg/dL blood lead group, resulting in a net decrease of those below the 2012 CDC 5 mg/dL threshold: 4
from 3.67% for May 2009-April 2010 to 2.59% for May 2014-April 2015. For the 5.0-9.9 mg/dL blood lead level group and the very high blood lead level group, the patient distributions showed year-over-year decreases during the same period (Table I). May 2014-April 2015 showed a slight reversal of these trends. The top 2.5% blood lead level threshold (97.5th percentile) was 5.1 mg/dL every year of the study. Yearly Changes in High Blood Lead Level for Various Risk Factors Table III illustrates the trends in high blood lead levels for all the risk factors mentioned previously across the 6 years of the McClure, Niles, and Kaufman
- 2016
Blood Lead Levels in Young Children: US, 2009-2015
Table II. Blood lead levels in children < 6 years old by state May 2009-April 2015 Total
£3.0 mg/dL
>3.0 and <5.0 mg/dL
‡5.0 and <10.0 mg/dL
State
‡10.0 mg/dL
‡5.0 mg/dL
Total
‡5.0 mg/dL
May 2014-April 2015 Total
‡5.0 mg/dL
Number (Percent) 2345 190 843 8530 35 703 76 520 5216 82 200 10 571 30 132 17 891 9869 7114 20 645 78 016 11 315 168 103 23 605 15 204 598 685 15 731 9484 2375 157 736 7865 11 859 56 864 18 528 2621 6562 123 700 26 911 8647 212 851 2474 11 229 873 944 386 743
2030 (86.57) 168 214 (88.14) 7598 (89.07) 31 715 (88.83) 68 697 (89.78) 4700 (90.11) 74 375 (90.48) 9616 (90.97) 27 427 (91.02) 16 491 (92.17) 9044 (91.64) 6611 (92.93) 19 295 (93.46) 72 651 (93.12) 10 561 (93.34) 158 835 (94.49) 22 356 (94.71) 14 372 (94.53) 562 660 (93.98) 14 995 (95.32) 8990 (94.79) 2243 (94.44) 150 431 (95.37) 7518 (95.59) 11 332 (95.56) 54 194 (95.30) 17 674 (95.39) 2520 (96.15) 6302 (96.04) 118 943 (96.15) 26 088 (96.94) 8415 (97.32) 207 015 (97.26) 2391 (96.65) 10 946 (97.48) 848 450 (97.08) 378 739 (97.93)
74 (3.16) 7728 (4.05) 324 (3.80) 1501 (4.20) 2705 (3.54) 170 (3.26) 2976 (3.62) 344 (3.25) 967 (3.21) 524 (2.93) 362 (3.67) 179 (2.52) 479 (2.32) 2123 (2.72) 318 (2.81) 3520 (2.09) 467 (1.98) 331 (2.18) 17 291 (2.89) 270 (1.72) 222 (2.34) 64 (2.69) 2849 (1.81) 129 (1.64) 205 (1.73) 1147 (2.02) 359 (1.94) 35 (1.34) 114 (1.74) 2140 (1.73) 309 (1.15) 80 (0.93) 2394 (1.12) 45 (1.82) 123 (1.10) 13 285 (1.52) 3802 (0.98)
184 (7.85) 11 633 (6.10) 461 (5.40) 1889 (5.29) 4099 (5.36) 255 (4.89) 3931 (4.78) 474 (4.48) 1411 (4.68) 647 (3.62) 388 (3.93) 261 (3.67) 663 (3.21) 2698 (3.46) 338 (2.99) 4447 (2.65) 616 (2.61) 417 (2.74) 14 880 (2.49) 371 (2.36) 225 (2.37) 56 (2.36) 3713 (2.35) 171 (2.17) 265 (2.23) 1309 (2.30) 423 (2.28) 53 (2.02) 105 (1.60) 2070 (1.67) 381 (1.42) 117 (1.35) 2753 (1.29) 30 (1.21) 131 (1.17) 10 262 (1.17) 3653 (0.94)
Absolute change reflects the difference in >5.0 mg/dL from the first to the second time periods.
57 (2.43) 3268 (1.71) 147 (1.72) 598 (1.67) 1019 (1.33) 91 (1.74) 918 (1.12) 137 (1.30) 327 (1.09) 229 (1.28) 75 (0.76) 63 (0.89) 208 (1.01) 544 (0.70) 98 (0.87) 1301 (0.77) 166 (0.70) 84 (0.55) 3854 (0.64) 95 (0.60) 47 (0.50) 12 (0.51) 743 (0.47) 47 (0.60) 57 (0.48) 214 (0.38) 72 (0.39) 13 (0.50) 41 (0.62) 547 (0.44) 133 (0.49) 35 (0.40) 689 (0.32) 8 (0.32) 29 (0.26) 1947 (0.22) 549 (0.14)
241 (10.28) 14 901 (7.81) 608 (7.13) 2487 (6.97) 5118 (6.69) 346 (6.63) 4849 (5.90) 611 (5.78) 1738 (5.77) 876 (4.90) 463 (4.69) 324 (4.55) 871 (4.22) 3242 (4.16) 436 (3.85) 5748 (3.42) 782 (3.31) 501 (3.30) 18 734 (3.13) 466 (2.96) 272 (2.87) 68 (2.86) 4456 (2.82) 218 (2.77) 322 (2.72) 1523 (2.68) 495 (2.67) 66 (2.52) 146 (2.22) 2617 (2.12) 514 (1.91) 152 (1.76) 3442 (1.62) 38 (1.54) 160 (1.42) 12 209 (1.40) 4202 (1.09)
638 41 568 1291 6420 16 095 2275 20 109 1236 7650 4180 1140 1011 2277 7564 1611 37 675 3468 1727 140 558 1818 1670 458 30 589 1593 2376 10 445 3395 296 1242 29 570 7343 1969 38 871 636 2518 200 413 72 951
60 (9.40) 4444 (10.69) 125 (9.68) 575 (8.96) 1324 (8.23) 165 (7.25) 1450 (7.21) 120 (9.71) 502 (6.56) 249 (5.96) 35 (3.07) 36 (3.56) 115 (5.05) 374 (4.94) 95 (5.90) 1815 (4.82) 96 (2.77) 49 (2.84) 5448 (3.88) 43 (2.37) 58 (3.47) 20 (4.37) 1179 (3.85) 33 (2.07) 84 (3.54) 257 (2.46) 126 (3.71) 4 (1.35) 32 (2.58) 810 (2.74) 125 (1.70) 31 (1.57) 699 (1.80) 15 (2.36) 41 (1.63) 3538 (1.77) 1070 (1.47)
345 21 782 1727 3668 10 250 284 9918 1792 4163 2470 572 801 3282 20 969 1965 23 036 3909 3368 72 265 7517 1616 352 23 340 805 1772 7952 3750 964 956 13 706 3031 1101 31 439 193 1749 107 443 59 019
26 (7.54) 1644 (7.55) 70 (4.05) 209 (5.70) 639 (6.23) 19 (6.69) 474 (4.78) 47 (2.62) 208 (5.00) 89 (3.60) 36 (6.29) 54 (6.74) 106 (3.23) 777 (3.71) 55 (2.80) 644 (2.80) 116 (2.97) 95 (2.82) 1799 (2.49) 264 (3.51) 40 (2.48) 19 (5.40) 512 (2.19) 23 (2.86) 34 (1.92) 226 (2.84) 85 (2.27) 26 (2.70) 30 (3.14) 251 (1.83) 70 (2.31) 22 (2.00) 519 (1.65) 5 (2.59) 17 (0.97) 1448 (1.35) 513 (0.87)
Absolute % change 1.86 3.14 5.63 3.26 2.00 0.56 2.43 7.09 1.56 2.36 3.22 3.18 1.82 1.23 3.10 2.02 0.20 0.02 1.39 1.14 0.99 1.03 1.66 0.79 1.62 0.38 1.44 1.35 0.56 0.91 0.61 0.43 0.15 0.23 0.66 0.42 0.60
5
ORIGINAL ARTICLES
MN PA KY OH CT WI MO NH MI LA MS SC IN MA WV IL OK TN NY NC AL UT MD AR KS GA CO NM OR NJ VA DE TX WA DC CA FL
May 2009-April 2010
May 2009April 2015
May 2009April 2010
May 2010April 2011
May 2011April 2012
May 2012April 2013
May 2013April 2014
May 2014April 2015
3 803 070
823 198
737 826
633 374
560 906
540 016
507 750
Significant trend
Number (Percent) 60 219 (3.09) 51 622 (2.80)
16 216 (3.86) 13 856 (3.47)
12 079 (3.20) 10 417 (2.91)
9357 (2.89) 8245 (2.68)
8063 (2.81) 6828 (2.51)
7379 (2.66) 6277 (2.40)
7125 (2.72) 5999 (2.44)
* *
660 387 662 327 661 936 659 314 627 376
7065 (1.07) 9866 (1.49) 13 763 (2.08) 24 269 (3.68) 37 382 (5.96)
1555 (1.21) 2124 (1.50) 3587 (2.50) 6706 (4.74) 10 766 (7.59)
1295 (1.05) 1993 (1.52) 2725 (2.13) 4743 (3.77) 7573 (6.27)
1097 (0.99) 1785 (1.59) 2288 (2.05) 3817 (3.51) 5615 (5.54)
1015 (1.01) 1411 (1.42) 1923 (1.95) 3310 (3.34) 4669 (5.25)
855 (0.89) 1262 (1.37) 1642 (1.78) 2966 (3.09) 4550 (4.94)
1248 (1.23) 1291 (1.51) 1598 (1.82) 2727 (3.10) 4209 (5.13)
P = .0446 P = .0290 * * *
2 665 330 769 566
76 338 (2.86) 20 224 (2.63)
22 060 (3.53) 4371 (3.35)
15 933 (2.94) 3492 (2.80)
11 504 (2.74) 3583 (2.50)
9531 (2.53) 3092 (2.44)
8899 (2.49) 2869 (2.11)
8411 (2.45) 2817 (2.60)
* *
703 782 624 981 643 452 698 627 597 828
12 213 (1.74) 12 270 (1.96) 16 374 (2.54) 22 743 (3.26) 28 627 (4.79)
2993 (2.02) 3116 (2.34) 4327 (3.19) 6424 (4.19) 7842 (6.25)
2266 (1.70) 2586 (2.10) 3334 (2.65) 4433 (3.35) 5684 (4.98)
2005 (1.73) 1947 (1.88) 2531 (2.35) 3476 (2.99) 4632 (4.58)
1610 (1.54) 1633 (1.75) 2100 (2.19) 3067 (2.97) 3899 (4.38)
1528 (1.54) 1566 (1.75) 2034 (2.20) 2752 (2.74) 3383 (3.91)
1811 (1.78) 1422 (1.73) 2048 (2.37) 2591 (2.78) 3187 (3.92)
* * * * *
657 317 666 236 692 538 614 839 637 740
25 909 (3.94) 21 880 (3.28) 19 336 (2.79) 14 025 (2.28) 11 077 (1.74)
7241 (5.06) 6152 (4.31) 4894 (3.37) 3605 (2.77) 2810 (2.08)
5045 (4.03) 4346 (3.46) 3883 (2.87) 2863 (2.38) 2166 (1.76)
4156 (3.78) 3359 (3.05) 3046 (2.62) 2183 (2.12) 1847 (1.76)
3487 (3.58) 2886 (2.91) 2672 (2.56) 1843 (2.05) 1421 (1.49)
3063 (3.28) 2631 (2.69) 2513 (2.53) 1673 (1.95) 1383 (1.51)
2917 (3.31) 2506 (2.77) 2328 (2.52) 1858 (2.17) 1450 (1.65)
* * * * *
165 624 722 385 406 681 509 539 262 144 264 833 96 232 21 372 874 401 11 700
9018 (5.44) 21 351 (2.96) 20 619 (5.07) 8359 (1.64) 11 431 (4.36) 5384 (2.03) 5494 (5.71) 590 (2.76) 12 215 (1.40) 233 (1.99)
1824 (7.30) 6258 (3.68) 5915 (6.91) 1673 (1.82) 3232 (5.68) 1081 (2.23) 1610 (7.02) 156 (3.92) 3540 (1.77) 58 (2.51)
1636 (7.39) 4997 (3.25) 4128 (5.23) 1598 (1.75) 1974 (4.41) 1012 (1.96) 1010 (5.55) 138 (3.69) 2328 (1.32) 51 (1.87)
1330 (5.06) 3342 (2.84) 2941 (4.35) 1368 (1.57) 1942 (4.15) 1041 (2.18) 866 (5.55) 63 (1.98) 2004 (1.41) 24 (1.07)
1147 (4.45) 2393 (2.42) 2822 (4.52) 1376 (1.68) 1740 (4.20) 816 (2.08) 774 (5.56) 57 (1.95) 1574 (1.24) 34 (1.87)
1607 (4.82) 2311 (2.41) 2439 (4.22) 1046 (1.40) 1331 (3.55) 661 (1.73) 654 (4.93) 69 (2.03) 1320 (1.09) 28 (2.11)
1474 (4.46) 2050 (2.38) 2320 (4.38) 1298 (1.57) 1212 (3.48) 773 (1.95) 580 (4.72) 107 (2.58) 1449 (1.35) 38 (2.96)
* * * * * * * * * P = .2356
Volume -
All values Number (Percent) reflect results of $5.0 mg/dL; CDC changed threshold to 5.0 mg/dL in May 2012. Notes: 1 - chi2 P < .01 for all years; 12 496 results did not have gender data. 2 - Cochran Armitage test for trend P < .01 for all years; 531 730 results did not have pre-1950 house construction data available. 3 - 368 174 results did not have payer data available. 4 - Cochran Armitage test for trend P < .01 for all years; 534 400 results did not have poverty data. 5 - 468 159 results did not have state data available.
www.jpeds.com
1 947 693 1 842 881
McClure, Niles, and Kaufman
Gender1 Male Female Pre-1950 housing quintile2 <3.6% 3.6-12.9% 13.0-29.9% 30.0-50.9% 51.0% + Payer type3 Private Medicaid/Medicare Low income (PIR <1.25) quintile4 <16.0% 16.0-28.9% 29.0-39.9% 40.0-51.9% 52.0% + High income (PIR >5) quintile4 <2.8% 2.8-6.8% 6.9-13.9% 14.0-27.9% 28.0% + HHS region5 1: CT, MA, ME, NH, RI, VT 2: NJ, NY 3: DE, DC, MD, PA, VA, WV 4: AL, FL, GA, KY, MS, NC, SC, TN 5: IL, IN, MI, MN, OH, WI 6: AR, LA, NM, OK, TX 7: IA, KS, MO, NE 8: CO, MT, ND, SD, UT, WY 9: AZ, CA, HI, NV 10: AK, OR, ID, WA
THE JOURNAL OF PEDIATRICS
6
Table III. Trends in high blood lead levels of children <6 years old
- 2016
ORIGINAL ARTICLES
Table IV. Logistic models: Factors associated with high blood lead levels and very high blood lead levels Factors associated with high BLL (‡5.0 mg/dL)
Study year Private payer Male Top housing quintile Top 3 HHS regions (1,3,7) Top poverty quintile Top wealth quintile
Factors associated with very high BLL (‡10.0 mg/dL)
Unadjusted OR
95% CI
aOR
95% CI
Unadjusted OR
95% CI
aOR
95% CI
0.93 1.09 1.11 2.99 2.43 2.06 0.56
(0.92-0.93) (1.08-1.11) (1.09-1.12) (2.94-3.02) (2.40-2.46) (2.03-2.09) (0.54-0.57)
0.92 0.84 1.11 2.50 2.23 1.64 0.65
(0.91-0.92) (0.82-0.85) (1.09-1.13) (2.46-2.54) (2.20-2.26) (1.61-1.67) (0.64-0.67)
0.94 1.06 1.10 2.96 2.48 1.95 0.53
(0.93-0.95) (1.02-1.10) (1.07-1.13) (2.87-3.04) (2.40-2.55) (1.88-2.01) (0.51-0.55)
0.93 0.80 1.08 2.58 2.21 1.55 0.62
(0.92-0.94) (0.77-0.83) (1.04-1.11) (2.50-2.68) (2.13-2.29) (1.50-1.61) (0.59-0.65)
BLL, blood lead level; CI, confidence interval; OR, odds ratio.
study. Male subjects maintained a greater proportion of high blood lead level than female subjects during each year of the study, and this difference remained similar over the years (0.21%-0.39%). The statistically significant trend among ZIP codes with increasing proportions of pre-1950s housing quintiles and PIR <1.25 remained when we examined data from each year of the study individually. Patients with private payer insurance had larger rates of high blood lead levels than patients with Centers for Medicare and Medicaid Services payers each year from May 2009-April 2014 but exhibited a lower rate of high blood lead levels than those with Centers for Medicare and Medicaid Services payers during the final year of the study. HHS Regions 1, 3, and 7 all exhibited a substantial decline in high blood lead level proportion during the study period but were the only regions with >4.0% high blood lead levels during the final year of the study. HHS Region 10 was the only region that had an increase in high blood lead levels during the study period and was the only factor examined that did not demonstrate a statistically significant (P < .05) downward trend in high blood lead level over the study period. High Blood Lead Level and Very High Blood Lead Level Models Logistic regression models were used to analyze the impact of the factors examined in Table I on high blood lead levels and very high blood lead levels. In unadjusted models, the strongest measures of association with high blood lead levels were being in the greatest pre-1950s housing quintile (OR 2.99, 95% CI 2.95-3.03) and living in HHS Region 1, 3, or 7 (OR 2.43, 95% CI 2.40-2.46). Living in a region in the lowest income quintile also was associated with high blood lead level (OR 2.06, 95% CI 2.03-2.09), whereas living in a region in the top income quintile had a significant protective effect (OR for high blood lead level = 0.56, 95% CI 0.54-0.57). The measures of association for these factors were similar in the adjusted multivariable model (Table IV). Private payer status had a small but significant association with high blood lead levels in the unadjusted model (OR 1.09, 95% CI 1.08-1.11) but had a significant protective effect in the adjusted model (aOR 0.84, 95% CI 0.82-0.85). Measures of association in the very high blood lead level models (both adjusted and unadjusted) were similar to those in the high blood lead level models (Table IV). Blood Lead Levels in Young Children: US, 2009-2015
Discussion Reducing blood lead levels in children has been and continues to be a major public health success. The declining blood lead levels are the result of public health initiatives that include the removal of leaded gasoline, the banning of lead paint, effective treatment of potable water supplies, and remediation of homes found to be contaminated with lead. Studies conducted by the CDC3,19 have provided valuable insight into the changing lead levels in children and associated risk factors. The millions of test results reported by individual states have allowed the CDC to track the changes in lead levels over time.19 The demographic factors associated with high blood lead levels, however, come from a much smaller set: n = 793 from 2007-2010 for ages 1-2 years19 and n = 1653 from 2007-2010 for ages 1-5 years.13 The study presented here examines similar factors in a much larger data set of 3.8 million children. The CDC studies also were unable to look at factors associated with blood lead level $10.0 mg/dL because of small numbers (9 children in 2007-2008; 6 children in 2009- 2010).3 In contrast, our study was able to examine factors associated with very high venous blood lead level results (n = 22 237). Our study provides insights for the time period through mid-2015, 4.5 years after the end of data collection in the most recent NHANES report, which included data through the end of 2010. This timeframe extension enabled the examination of the continuing decline in high blood lead level and very high blood lead level levels. This was true for the population as whole (from 3.67% to 2.59% high blood lead levels and from 0.74% to 0.55% very high blood lead levels) and for most demographic groups and geographic regions. The results also demonstrate a slight increase in the incidence of high blood lead levels and very high blood lead levels in the final year of the study. After years of consecutive decline, the rise was unexpected, especially given that there was no major observable change in demographic proportions during the final year. Future studies are needed to evaluate the evolution of high blood lead levels and very high blood lead levels. To a large extent, the results presented here confirm the significance of factors examined by the CDC3,19 and exhibit similar measures of association. Our study, using a different methodology, confirms the findings in NHANES associating 7
THE JOURNAL OF PEDIATRICS
www.jpeds.com
pre-1950s housing data and low PIR with high blood lead level. Examining PIR and housing construction data by ZIP code instead of direct patient data is a limiting factor but also highlights the ability of physicians to identify geographic regions with increased risk, particularly in the event that individual patient risk factors are unavailable or difficult to obtain. Using data available on the US Census Bureau website, health care experts can determine the proportion of housing constructed before 1950 and PIR of the ZIP codes they serve to help assess the level of risk in their patients, even if individual factors are not available. All previously reported NHANES cycles found increased levels of high blood lead level in Medicaid patients (the differences in 1999-2002 and in 2003-2006 were statistically significant). The current Quest Diagnostics Health Trends study found that children with private payers had a significantly larger rate of high blood lead level than those with Medicaid/Medicare payers. In the adjusted model, however, private payer status had a significant protective effect. Thus, with all factors explored in the study considered together, the private payer results seem to indicate the same relationship found in the NHANES studies. There were also some differences between our findings and those of the NHANES studies. The 2003-2006 and 2007-2010 NHANES cycles both found female participants to have slightly greater proportions of high blood lead level than male participants (although these differences were not found to be statistically significant). In the present study, male participants exhibited a small but statistically significantly larger proportion of high blood lead levels than female participants. Our study examined results from patients living in all 50 states and the District of Columbia; however, analyses by state were limited to the 36 states and the District of Columbia with more than 2000 specimens during the study period. Our results for high blood lead level at the state level are often similar (within 3%) to the results reported by the CDC in their national surveillance data (2010-2014),20 but there are also interesting differences. The rates of high blood lead level in our study were considerably lower than those reported by the CDC for New Hampshire (6.2% vs 12.0%), Mississippi (4.5% vs 9.2%), and Illinois (4.5% vs 7.9%). Only Minnesota showed a considerably larger rate of high blood lead level in our study than in the CDC report (10.3% vs 2.9%). At the national level, we found a virtually identical rate of very high blood lead level (0.58%) as the CDC (0.57%) over a similar 5-year period, but the rate of high blood lead level was lower in our study (2.95% vs 5.48%). The reasons for these differences are unclear but could reflect the inclusion of capillary blood results in the CDC data. According to the CDC, “houses built before 1950 pose the greatest hazard to children because they are much more likely to contain lead-based paint than newer houses.”16 Leadcontaminated dust on floors, windowsills, and window wells is associated with elevated blood lead levels in children.21 Our findings showed that, in general, the regions with the greatest adjusted level of pre-1950s housing construction had the 8
Volume greatest proportions of patients with high blood lead levels. HHS Region 2 (New York and New Jersey) was an interesting deviation from this trend. Despite having the greatest level of housing built before 1950 (based on study participant ZIP codes), the overall levels of high blood lead level were below average for the study. It is also interesting that New York State (3.1%) was near the national average (3.0%) for high blood lead level while having several 3-digit ZIP codes with the largest proportions of high blood lead levels and very high blood lead levels found in the study. The impact of abatement and active lead surveillance in these states is also unclear. In general, southern regions had lower proportions of high blood lead levels. This includes Florida, with more than 386 000 specimens and 1.1% with high blood lead levels, and California, with more than 873 000 specimens and 1.4% with high blood lead levels. We have no easy explanation why 12 of the 36 states and the District of Columbia, with sufficient number of results, had an increase in the proportion of high blood lead levels between the first year and the final year of the study. We found that living in ZIP codes with greater percentages of residents <1.25 PIR exhibited a stronger association with high blood lead levels than individual payer type. The reasons for this perplexing difference are unclear. One possibility may be that ZIP codes with greater levels of low income also have greater proportions of housing built before 1950 or other associated factors. Although this relationship may exist, the multivariable models indicate that it is not reason for the association, because both variables maintain significant associations with high blood lead levels and very high blood lead levels in the adjusted multivariable models. All risk factors associated with high blood lead levels also were associated with very high blood lead levels. In both unadjusted and adjusted models, the measures of association were remarkably similar. This finding is reassuring because the risk factors identified in previous studies were not only confirmed for the most part in this larger study but found to be associated with very high blood lead levels as well. The major strengths of this study are its large size, national representation with data from 50 states and the District of Columbia, and the inclusion of data through mid-2015. Analyses were conducted on more than 5 million results for children <6 years of age. The analyses focused on more than 3.8 million results of venous blood draws from children <6 years of age, consistent with the CDC definition of confirmed elevated blood lead level. This study also examined demographics for the vast majority of participants. Having demographics available for so many participants enabled analysis of factors associated with very high blood lead levels. This study also had limitations. The 3.0 mg/dL reporting threshold precluded our ability to estimate the mean blood lead level for the study population. Some blood lead level groups are narrower than the range of instrument variability, but we assume there is no bias in the data set. It is also possible that some patients were tested as a follow-up to prior blood lead level results at another laboratory or because healthcare providers suspected a high probability of elevated McClure, Niles, and Kaufman
- 2016
ORIGINAL ARTICLES
blood lead level. It also may be possible that population segments or regional populations deemed to be at risk are being tested more frequently. Sociological, familial, and environmental factors also may play a role in determining who is tested for lead; however, with our 97.5th percentile being the same as found in NHANES, we feel the selection bias is minimal. Quest Diagnostics does not perform all lead testing in the country, and these data should only be seen as a large sample of national data. These data do not necessarily reflect the population as a whole but reflect those tested in medical practices in the US. In summary, progress in reducing the burden of lead toxicity is a public health success story that is incomplete. This analysis of more than 5 million blood lead level results over a 6-year period, extending through April 2015, includes 3.8 million venous blood lead level results for infants and children <6 years of age. This allowed for a robust, detailed analysis of results by geography and other criteria that are impossible with the narrower NHANES approach. Unique observations include the correlation of results for patients with blood lead level results of $10.0 mg/dL to those with results of 5.1-10.0 mg/dL. Many of the 3-digit ZIP code regions with high blood lead levels and very high blood lead levels are in New York, Pennsylvania, and Ohio. Additional study will be necessary to assure that the improvements reported will continue and that specific state and local efforts achieve the desired goals. n Submitted for publication Jan 26, 2016; last revision received Mar 29, 2016; accepted May 3, 2016. Reprint requests: Leland F. McClure, PhD, Director, Medical Science Liaison, Medical Affairs - Quest Diagnostics, 11636 Administration Drive, St Louis, MO 63146. E-mail: [email protected]
References 1. Agency for Toxic Substances and Disease Registry. Toxicological profile for lead. Atlanta, GA: US Department of Health and Human Services, CDC, Agency for Toxic Substances and Disease Registry; 2007. 2. Parsons PJ, Reilly AA, Esernio-Jenssen D. Screening children exposed to lead: an assessment of the capillary blood lead fingerstick test. Clin Chem 1997;43:302-11. 3. Centers for Disease Control and Prevention. Blood lead levels in children aged 1-5 years – United States, 1999-2010. MMWR Morb Mortal Wkly Rep 2013;62:245-8. 4. Pirkle JL, Brody DJ, Gunter EW, Kramer RA, Paschal DC, Flegal KM, et al. The decline of blood lead levels in the United States: the National Health and Nutrition Examination Surveys (NHANES). JAMA 1994; 272:284-91.
Blood Lead Levels in Young Children: US, 2009-2015
5. Centers for Disease Control and Prevention. Preventing lead poisoning in young children. Atlanta: CDC; 1991. 6. CDC Advisory Committee on Childhood Lead Poisoning Prevention. Low level lead exposure harms children: A renewed call for primary prevention. Atlanta: CDC; 2012. 7. Chandramouli K, Steer CD, Ellis M, Emond AM. Effects of early childhood lead exposure on academic performance and behavior of school age children. Arch Dis Child 2009;94:844-8. 8. Nigg JT, Nikolas M, Mark Knottnerus G, Cavanagh K, Friderici K. Confirmation and extension of association of blood lead level with attention-deficit/hyperactivity disorder (ADHD) and ADHD symptom domains at population-typical exposure levels. J Child Psychol Psychiatry 2010;51:58-65. 9. Needleman HL, Schell A, Bellinger D, Leviton A, Allred EN. The longterm effects of exposure to low doses of lead in childhood. An 11-year follow-up report. N Engl J Med 1990;332:83-8. 10. Bellinger DC, Stiles KM, Needleman HL. Low-level lead exposure, intelligence and academic achievement: a long-term follow-up study. Pediatrics 1992;90:855-61. 11. Rogan WJ, Dietrich KN, Ware JH, Dockery DW, Salganik M, Radcliffe J, et al. The effect of chelation therapy with succimer on neuropsychological development in children exposed to lead. N Engl J Med 2001;344: 1421-6. 12. Centers for Disease Control and Prevention. Response to Advisory Committee on Childhood Lead Poisoning Prevention Recommendations in “Low Level Exposure Harms Children: A Renewal Call of Primary Prevention” CDC, http://www.cdc.gov/nceh/lead/acclpp/cdc_response_ lead_exposure_recs.pdf. Accessed May 10, 2016. 13. Centers for Disease Control and Prevention. Screening young children for lead poisoning: Guidance for state and local public health officials; Appendix C1: The Lead Laboratory, www.cdc.gov/nceh/lead/ publications/screening.htm. Accessed December 12, 2015. 14. CDC Standard Surveillance Definitions and Classifications, http://www. cdc.gov/nceh/lead/data/definitions.htm. Accessed December 12, 2015. 15. US Census Bureau, 2009-2013 5-Year American Community Survey, http://factfinder.census.gov/faces/tableservices/jsf/pages-/productview. xhtml?pid=ACS_13_5YR_B25034&prodType=table. Accessed December 12, 2015. 16. CDC. Facts on Lead, http://www.cdc.gov/nceh/lead/publications/1997/ factlead.htm. Accessed December 12, 2015. 17. US Census Bureau, 2008-2012 5-Year American Community Survey, http://factfinder.census.gov/faces/tableservices/jsf/pages/-productview. xhtml?pid=ACS_12_5YR_B17024&prodType=table. Accessed December 12, 2015. 18. SAS, version 9.4. Cary, NC: SAS Institute. 19. Raymond J, Wheeler W, Brown MJ. Lead screening and prevalence of blood lead levels in children aged 1–2 years—Child Blood Lead Surveillance System, United States, 2002–2010 and National Health and Nutrition Examination Survey, United States, 1999–2010. MMWR Suppl 2014;63:36-42. 20. Centers for Disease Control and Prevention. CDC’s National Surveillance Data (1997-2014), http://www.cdc.gov/nceh/lead/data/national. htm. Accessed March 10, 2016. 21. Lanphear BP, Weitzman M, Winter NL, Eberly S, Yakir B, Tanner M, et al. Lead-contaminated house dust and urban children’s blood lead levels. Am J Public Health 1996;86:1416-21.
9