Document peGLbdk1jdy2wgpmNp9kmygVa
From:
James L. Pirkle, M.D., Pb.D.
Subject: Sampling of the Urban Population and the Downward Trend in NHANES II Bipod Lead Levels
Questions have arisen as to whether the downward trend in the NHANES II blood lead levels could be attributed to a trend in the percentage of urban dwellers (i.e., living in areas with greater than one million population) sampled during the survey* DuPont replotted NHANES II lead data overlayed with the percent urban dwellers in the sample (attached Figure 1). Figure 2 shows the same overlay with the percent urban dwellers which were actually used in the weighted analysis at the National Center for Health Statistics* Figure 1 and Figure 2 are qualitatively similar although quantitative values for percent urban dwellers is markedly different*
The somewhat similar downward trend in percent urban dwellers sampled and mean lead levels is due to the plotting intervals of January-June and July-December. When the data are plotted with April-September and October-March intervals, the similarity disappears (Figure 3). A plot of mean lead level vs. percent urban dwellers sampled for Jarjuary-June and July-December intervals is shown in Figure 4* The Pearson correlation coefficient is 0.76. A similar plot for April-September and October-March intervals is shown in Figure 5. The Pearson correlation coefficient is only 0.36*
By contrast, the strong linear relationship between mean lead levels and lead used in gasoline is independent of the plotting intervals. Figures 6 and 7 show the mean blood lead level overlayed with the amount of lead used in gasoline for both sets of intervals (January-June and July-December; April-September and October-March). Figures 8 and 9 show the corresponding correlation plots. Note that the Pearson correlation coefficient is 0.95 independent of the type of interval chosen.
Additional evidence that urban sampling cannot account for the downward trend in lead levels can be seen in plotting subsets of the lead data according to degree of urbanization. Figures 10-15 show three subsets of the lead data plotted vs. time: 1) urban greater than one million population, 2) urban less than one million, and 3) rural. Each subset is plotted using both types of the 6-month intervals (January-June and July-December; April-September and October-March). Distinct downward trends are present in the urban less than one million population and the rural subsets, neither of which includes any of the urban dwellers of subset one. This clearly demonstrates that trends in the percent of urban dwellers (greater than one million population) cannot account for the downward trend in the NHANES II lead data*
TEH 0533237
rigraie
Original Houk Figure
LEAD USED IN GASOLINE PRODUCTION AND AVERAGE NHANES It BLOOD LEAD LEVELS
(FEB* 1976 - FE8* 1980)
Fifure I
Another Plot of NHANES Data.
PERCENT URBAN DWELLERS IN SAMPLE AND AVERAGE NHANES II BLOOD LEAD LEVELS
Sc
TEH 0533238
DUP050034505
X U J C E Z -IIU C C O _ im > U 4_4
NHANES II LEAD MEANS AND PCT URBAN POPULATION SAMPLED
JANUARY - JUNE AND JULY - DECEMBER INTERVALS
TEH 0533239
DUP050034506
Dbz.
NHANES II LEAD MEANS AND PCT URBAN POPULATION SAMPLED
APRIL - SEPT AND OCT - MARCH INTERVALS
TEH 0533240
DUP050034507
ISS
` ~.... .....
Fit- V
rp i< m r o s n r s s n z
NHANES II MEAN LEAD VS PCT URBAN POPULATION SAMPLED
JANUARY - JUNE AND JULY - DECEMBER INTERVALS
16-
15-
11312-
i*;
11-
10-
a CoaRBLAnotJ Co<=-(=i-i<iietur ~ oj}(g
T1 I' ftT I' t l*t "{' 1
5 10
rnrTf
15
20
/\r CO
per c en t u r ban po pu l at io n
WT'j'Trt T*7*
m
TEH 0533241
DUP050034508
ILL Fi6-. S'
NHANES II LEAD MEANS VS. PCT URBAN POPULATION SAMPLED
APRIL - SEPT AND OCT - MRRCH INTERVALS
1?H
16-
r-rn<nr- o s n r z x t iz
15m-
13-
12 ih
o
Q
C*Retmw>\J Co<=^Fii <=-tuT ~ CL3&
9{mnT,TTrnrTTinrj-rri`.-i'j rt'i'Ti , r.ny t ' i|' 0 3 6 9 12 15 18 21 29 2?
|-ri. p.. rtfrm . 11 . . p . ) .1. |r. i. T
30 33 36 39 92 45 46 51 $9
PERCENT URBAN POPULATION
TEH 0533242
DUP050034509
u ua
pi & 6
NHANES II LEAD MEANS AND LEAD USED IN GASOLINE
JANUARY - JUNE AND JULT - DECEMBER INTERVALS
TEH 0533243
DUP050034510
r - T O < P if- o n n z a jn t a s"
f\&. 7
LEAD USED IN GASOLINE PRODUCTION AND AVERAGE NHANES II BLOOD LEAD
APRIL - SEPT RND OCT - MARCH INTERVALS
TEH 0533244
DUP050034511
Ft G- ?
X U iC C Z -JIU C E O _ltU > lU _J
NHANES II MEAN LEAD VS LEAD USED IN GASOLINE
JifjNUfiRT - JUNE RND JULT - DECEMBER INTERVALS 17-
16
O
15
D
13-
1211- D
10.
p
CO(lREkfrno(0 CbtFFiCleWT ~ 0,9S'
...II i I ">'1 i i I I I - I
i-i- f-i-yr I Ti-j II I II I-TT-I"! 1 I I !
115 51 57 63 69 75 81 87 93
LEAD USED IN GASOLINE I1D00 TONS)
'1 1 1 ' ...I 1 !
99
105
TEH 0533245
DUP050034512
"S9 L
h (& ?
f-n)<m r- oaw nf zrom z
NHANES II LEAD MEANS VS. LEAD USED IN GASOLINE
APRIL - SEPT AND OCT - MARCH INTERVALS 174
IB-
15-
D D DO
13-
V2-
U-
10- CORftEt- frTioM Co f r c tmJT O * 9f
I l ` 1 1 ` ' T-i t . .-T-i | . r-T-, | . .
- .
| . , , . | ..................
45 51 57 63 69 75 81 87 93 99 105
LEAD USED IN GASOLINE 11000 TONS)
TEH 0533246
DUP050034513
69S FIS. 10 NBANES II MEAN LEAD VS YEAR
JANUARY - JUNE AND JULY - DECEMBER INTERVALS GROUP=URBRN GREATER THAN L MILLION POP,
DUP050034514
6SL FI 6. II NHANES II LEAD MEANS VS. YEAR
APRIL - SEPT AND OCT - MARCH INTERVALS GROUP=URBAN GREATER THAN 1 MILLION POP.
TEH 0533248
DUP050034515
NHANES II MEAN LEAD VS YEAR
JANUARY - JUNE AND JULY - DECEMBER INTERVALS GROUP=URBAN LESS THAN 1 MILLION POP,
TEH 0533249
DUP050034516
i/S
F/. !3
NHANES II LEAD MEANS VS. YEAR
APRIL - SEPT AND OCT - MARCH INTERVALS GROUP-URBAN LESS THAN 1 MILLION POP.
TEH 0533250
DUP050034517
pit?- t<f
NHANES II MEAN LEAD VS YEAH
JANUARY - JUNE AND JULY - DECEMBER INTERVALS GRQUP=RURAL
TEH 0533251
DUP050034518
3C-ii
fl&.
LEAD 17 -
NHANES II LEAD MEANS VS. YEAR
RPRIL - SEPT AND OCT - MARCH INTERVALS GR0UP=3
I ''' r"' ' 1 1 ' ' ' I ' `
0 1976
'1 1*I'
1977
" I....................... ' L L 1 ,............................................. ,
1978
1979
1980
TEAR
TEH 0533252
DUP050034519