Document RjB2qMEjKpENNOd2Q8j6om4R7

w. utmaW0" tSIAIUSKCOiSOl E. I. d u Po n t d e Ne mo u r s & Co mp a n y INCORPORATED Pe t r o l e u m La b o r a t o r y WILMINGTON, DELAWARE 18898 Te l e p h o n e Ar e a Co d e 609-299-5000 March 21, 1983 Dr. David E. Weil Project Manager Environmental Criteria & Assessment Office MD-52 U.S. Environmental Protection Agency Research Triangle Park, NC 27711 Dear Dr. Weil: The attached material has been prepared for distribution to the special statistics panel, assembled by EPA to determine to what extent, if any, NHANES II blood lead data can be used reliably to evaluate within study time trends and to correlate these trends with other environmental lead exposure data. Du Pont has made several submissions to EPA in the matter of the gasoline lead phasedown regulations (Docket No. A-81-36) which contain analyses and viewpoints on many issues bearing on the review of the Air Quality Criteria Document for Lead. As opposed to submitting all these documents, the attached presen tation represents a consolidation of only that material relevant to the chartered task of the special panel. Reference is made on several occasions to Du Pont's submission to EPA of October 8, 1982 which the panel already has received. We would appreciate it if you could distribute this mate rial to the panel in time so they can review it prior to our next session. Seven copies are enclosed for your convenience. We may have additional items for distribution at the next Ses sion. Very truly yours. JMP;mep Attachment BETTER THINGS FOR BETTER LIVING , . . THROUGH CHEMISTRY TEH 0533060 March 18, 1983 DO PONT *S SUBMITTAL TO SPECIAL EPA STATISTICAL PANEL EVALUATING THE NHANES II STUDY FOR THE EPA REVIEW OF THE AIR QUALITY CRITERIA DOCUMENT FOR LEAD The Second National Health and Nutrition Examination Sur vey {NHANES II) was conducted by the National Center for Health Statistics to gather health and nutritional data from a repre sentative cross section of the noninstitutionalized U.S. popu lation. The blood lead phase of the program (1) was intended to: (i) provide information for the first time about the dis tribution of blood lead levels in the general U.S. population, (ii) establish base line estimates for future studies to monitor changes in such exposure over time, (iii) provide normative in formation for use in health policy and regulatory decisions and (iv) correlate levels of exposure with other health and nutri tional parameters. Chronological analyses of NHANES II blood lead data have revealed a general downward trend during the course of the sur vey (February 1976 through February 1980), The blood lead trend and a similar trend in national gasoline lead use (see top figure in Attachment 1) have been widely publicized during 1982 in the proceedings of the gasoline lead phasedown regulations. The drop in blood lead levels from beginning of survey to end of approximately 5,8 /ug/dl has been largely attributed to gasoline lead as a result of the aforementioned coincidental trends and subsequent analyses by ICF (EPA contractor) and CDC which iden tified significant associations between blood lead and gasoline lead. Further, ICF (2) has used empirically derived relation ships between NHANES II blood lead data and gasoline lead use to (i) forecast the incidence and number of children expected to have elevated blood lead levels under various gasoline lead regulation scenarios for the years 1983 to 1990 and (ii) esti mate the proportion of lead in blood that is due to gasoline lead emissions for the period 1976 through 1980, Du Pont continues to be gravely concerned regarding the quantitative and causal implications of these and related analy ses. We do not believe that the NHANES II blood lead data can be used to estimate time trends during the study period, attri bute the portion of the decrease due to gasoline lead use or correlate these data with gasoline lead use to answer questions of the type ICF (2) has attempted to do. A key basis for this concern is a gross inconsistency between the observed blood lead level drop from beginning to end of the NHANES II study and that which might have been expected to occur based on reductions in air lead attributable to reductions only in gasoline lead. One of the best documented and independently assessed relationships used in setting the present air lead standard is the blood lead (jjg/dU-air lead (pg/irr) slope. Most researchers TEH 0533061 DUP050034331 have concluded that he slope is in the 1-2 m /dl range* (EPA used a value of 2 mydl in setting the current air lead stan dard.) The composite quarterly average of ambient air lead values at 10 5 sites across the nation reported by EPA (3) was about 1.25 jig/nr in 1976 and 0.93 pg/nr in 1979, the latest year available. Applying the blood lead/air lead slope J>f 2 m /dl to the estimated national air lead drop of 0.32 pg/nr, provides a rough estimate of 0.64 pg/dl for the expected blood lead drop. This estimate differs by nearly one order of magnitude from the decline observed during NHANES II. Because gasoline lead emis sions contribute approximately 90% to the lead in air, it is implausible that reductions in gasoline lead have resulted in a substantial portion of the observed decrease in blood lead levels. The difficulties of avoidance of contamination and of accurate analysis of lead in biological materials recently have been reviewed by the Karolinska Institute (4). As stated in Section 3.3 of the Institute's report (Attachment 2): "Studies on commercially available blood collection tubes have shown that diluted nitric acid as well as blood may extract lead and cadmium from certain vials and syringes in quantities which would invalidate any measurement of these metals in blood within the normal range of concen tration." As further stated in Section 2,3 of the Institute's report (Attachment 3): "The reviewed studies show that the accuracy and preci sion of trace metal analysis in biological specimens, especially lead and cadmium .in blood and urine, in gen eral appear to be unsatisfactory. This may also hold for laboratories that have gained considerable experience by analyzing a large number of biological samples over many years Because of the factors discussed above, it appears plaus ible that most of the reported decline in blood lead levels may be due to one or more of the following: 1) Biases resulting from sampling itinerary and differ ential stand-to-stand demographic composition, 2) Bias resulting from inadequate quality control. 3) Contributions from reductions in other sources of lead exposure such as water, food and nonfood (e.g. paint lead) items. Consequently, Du Pont continues to critically review the NHANES II data and challenge those who use these data to draw unsUppo r tab1e cone1usions. 2. TEH 0533062 DUP050034332 The following items summarize points that Du Pont has previously raised in submissions to EPA in the matter of the gasoline lead phasedown regulations on the interpretation of the NHANES II blood lead data and its relationship to gasoline lead use. 1. Regional Sampling Variations Although the NHANES II survey was intended to provide a Snapshot of distributional characteristics of all recorded health and nutritional parameters, the study required four years to complete. No controls to insure samples representative of the target population within stands or even within short time periods were imposed. Regarding the scheduled sample itinerary, Annest, et al (1) warned: "A possible logistical factor indirectly influencing the blood lead data is the itinerary of the Mobile Examina tion Centers (MEC's). To minimize the effects of adverse weather conditions on response rates, MEC's were set up in the northern states during the summer and more south ern states during the winter. The potential environmental effects on blood lead levels associated with seasonality and geographic location may be confounded, to some unde termined degree, with those associated with degree of urbanization of place of residence,,, The breakdown by region of numbers of venous blood leads sampled by six-month periods is shown in the table below. Year long periods devoid of data occurred in each region at some point during the study. In 1980, samples were taken only in the south. Sampling Program Number of Persons Sampled Northeast South Midwest West Total 76-1st half 76-2nd half 0 437 41 557 1,035 0 425 821 226 1,427 77-lst half 470 696 0 0 1,166 77-2nd half 845 267 0 0 1,112 78-1st half 78-2nd half 0 424 0 19 1,193 1,212 0 776 0 1,200 79-lst half 0 240 673 517 1,430 79-2nd half 353 330 208 0 991 80-lst' half 0 315 0 0 315 Total 2,092 2,710 2,638 2,493 9,933 3. TEH 0533063 DUP050034333 2. Differential Urban Representation The figure shown at the top of Attachment 1 is the plot that Dr. Vernon Houk, CDC, published (5) and presented in public sessions, which shows a correlation between the weighted arith metic mean NHANES II blood lead levels and gasoline lead use, grouped by six-month (January to June and July to December) in tervals. The figure at the bottom of Attachment 1 is an alterna tive plot showing a similar correlation between the same NHANES Ii blood lead values and actual percent of urban dwellers (urban Categories 1-4 on data tape) represented. Further simple correlation analyses were performed be tween weighted geometric mean blood lead levels and weighted sample proportions associated with various demographic categori zations in six-month intervals. These results are presented in Attachment 4 (see (6), Section III, Appendix A). The urban category is defined as in Attachment 1, which the rural area category reflects urban category 8 on the NHANES II data tape* Consistent significant associations occur over many age, race and sex categories for urban and rural dwellers. 3. Impact on Blood Dead Trend After Accounting for Urbanization and SMSA Category By way of a simple analysis, Du Pont has attempted to underscore the need to take proper and complete account of con founding factors before using NHANES 11 in a manner for which it was not designed. Preliminary graphical time trends by SMSA category (levels are central city, not central city and not SMSA and are a separate and not dependent classification from urbani zation in the NHANES II data base) revealed a decreasing but nonparallel relationship. Regressing loge blood lead on terms reflecting urban area (categories 1-4) , SMSA category, time, time by SMSA category interactions and age, race and. sex main and interaction effect terms confirmed the nonparallelism. (Re fer to (6) , Section HI, Table 1.) Subsequently, another analy sis was performed regressing log blood lead on single degree of freedom terms urban area, central city (versus not SMSA) and not central city (versus not SMSA). The residuals of this analysis on 9701 blood lead observations (only white and black races in cluded) reflect adjustments to the dependent variable for urban area and SMSA category. Weighted (by lead weights) geometric mean comparisons between the observed blood lead data and the adjusted blood lead data (exponential of residual plus overall mean) for the first and last six-month periods of NHANES II are given in Attachment 3 (see (6), Section III, Table 2). The re sults suggest that degree of urbanization and SMSA category may be able to account for a nonincidental portion of the trend. 4. TEH 0533064 DUP050034334 4. Evaluation of Blind Quality Control Data Our analysis of the NHANES II blind quality control data is reported in (6) , Section III, pages 5-7. this analysis was based on data provided last summer by CDC representing averages of duplicate blind quality control data from low (13.5 ug/dl) and high (25.5 /ug/dl) lead bovine pools. In that discussion, it was noted that no blind quality control data were available for the first 334 days or the last 255 days of SHANES II and there were coincidental significant curvilinear effects in both pools. CDC subsequently pointed out that data were available at the end of the study and the testing time associated with the quality control data must be inaccurate. On January 26, 1983, Dr. Pirkle provided Du Pont another listing of these data with correct time associations. The new listing indicates that qual ity control testing began 518 days after the start of NHANES II for the low lead pool and 369 days after the start for the high lead pool. 5. Federal Programs to Reduce Lead Exposure A report of the National Academy of Sciences (7) stated that there are at least eight departments and agencies of the Federal government currently administering programs designed to limit human exposure to lead. Fourteen regulations are identi fied which are aimed at reducing or eliminating human lead expo sure as it occurs through air, water, food and nonfood media (e.g. paint lead). Table Cl of the NAS report (Attachment 6) summarizes these regulations. Unfortunately, very little data of sufficient quality on the impact of these programs are available to relate to the NHANES II blood lead data. Publications by Jelinek (8) and Schaffner (9) indicate major reductions in food lead levels, particularly those of the processed or canned variety. Table 5 from Jelinek is shown below. Lead Content, Food for Infants Product Mean Level, ppm Early 1970's 1976-1977 1979-1980 Infant Formula, Concentrate a. 0.10 0.055 0.02 Infant Juices 0.30 a 0.045 a 0.015 b Infant Foods, Pureed b 0.15 0.05 0.03 Evaporated Milk 0.52 0.10 0.08 a Packed in cans b Packed in glass 5. TEH 0533065 DUP050034335 The paucity of quality data, however, does not allow one to conclude that lead exposure reductions in sources of lead other than via gasoline lead has no, or has an inconsequential, effect on the observed NHANES II blood lead trend; nor does it allow one to conclude that no differential variations in lead exposure from nongasoline sources during NHANES II occurred which is an implicit assumption when trying to forecast blood lead values on the basis of gasoline lead consumption alone. 6, Analyses Relating Blood Lead Levels to Both Gasoline Lead Use and Paint Lead Exposure Variables These analyses are discussed in detail in Section V of (6) . The purpose of these analyses was to demonstrate to the statistical laymen by way of a plausible example the difficul ties associated with trying to unambiguously quantify the por tion of an observed decline in the incidence of the blood lead levels exceeding thresholds (of 30 pg/dl and 49 pg/dl) with simultaneous reductions occurring in more than one lead exposure source. The data used are from the CDC Lead Poisoning Preven tion Program published in the CDC Morbidity and Mortality weekly reports. The results reveal (see (6), Section V, Table 3) that: e Analyses ignoring gasoline lead use lead to the con clusion that lead paint exposure variables are sig nificantly associated with the incidence of elevated blood lead. Analyses ignoring lead paint exposure variables lead to the conclusion that gasoline lead is significantly associated with the incidence of elevated blood lead. e Analyses accounting for both gasoline lead and lead paint exposure variables lead to the conclusion that lead paint exposure variables are better predictors of incidence of elevated blood lead levels than gaso line lead use. In two of the three incidence re sponses analyzed, gasoline lead use was not signifi cant at level p < .05. Subsequent analyses (see (6), Section V, Table 4) demon strated that no significant residual time trend remained after accounting for only lead paint exposure variables, while one did after accounting for only gasoline lead use. These results show that if only data on lead paint expo sure were available, one could conceivably arrive at the conclu sion that lead paint is the primary, or even only, important ex posure factor impacting on the incidence of elevated blood lead. The problem evidenced here and also inherent in all studies relating blood lead to a single lead exposure source such as gasoline lead, is that all lead exposure variables are presum ably highly correlated with one another over time (not to men- 6. TEH 0533066 DUP050034336 tion the inherent problems in the quality of the blood lead data itself). Consequently, the effects and contributions of each can never be unambiguously sorted out. Focusing in on one or a few exposure sources, to the exclusion of others, only exagger ates the confusion. Interpretation of such correlation analyses should always take into account these inherent limitations. 7. TEH 0533067 DUP050034337 REFERENCES Annest, J. L. , Mahaffey, K. R., Cox, D. H. and Roberts, J., "Blood Lead Levels for Persons 6 Months - 74 years of Age: United States, 1976-1980," National Center for Health Statistics, Advance Lata, Number 79, May 12, 1982, "The Relationship Between Gasoline Lead Usage and Blood Lead Levels in Americans: A Statistical Analysis of the NHANES II Data" prepared under Contract (68-01-5845) for u.s. Environmental Protection Agency, Office of Policy and Resource Management, Office of Policy Analysis, by ICF Incorporated, December 1982, "National Trend in the Maximum Quarterly Average Lead Lev els, 1970-1979," EPA Office of Air Quality Standards, Data Monitoring Branch, "Assessment of Human Exposure to Lead and Calcium Through Biological Monitoring," National Swedish Institute of Environmental Medicine and Department of Environmental Hygiene, Karolinska institute, Stockholm, 1982. "Blood Lead Levels in U.S, Population," Morbidity and Mor tality Weekly Report, Centers for Disease Control, 31, 132-134 (1982), Supplementary statement Presented to Environmental Protec tion Agency in the Matter of Regulation of Fuel and Fuel Additives, Lead Phasedown Regulations Proposal Rulemaking, October 8, 1982, Docket No. A-81-36, Petroleum Chemicals Division, E. I. du Pont de Nemours & Co., Inc, Lead in the Human Environment, National Academy of Sci ences, Washington, D.C. (1980). Jelinek, C. F., "Levels of Lead in the U.S. Food Supply ," Food and Drug Administration, Bureau of Foods, Washington, D.C. Schaffner, A. M., "Lead in Canned Foods," Food Technology, December 1981. 8. TEH 0533068 DUP050034338 t o t al Lead USED PER mo nt h PERIOD (1000 t ons) Attachment 1 Original Houk Figure LEAD USED IN GASOLINE PRODUCTION AND AVERAGE NHANES II BLOOD LEAD LEVELS (FEB* 1976 - FES* 1980) YEAR Another Plot of NHANES Data PERCENT URBAN DWELLERS IN SAMPLE ANO AVERAGE NHANES II BLOOD LEAD LEVELS 9. TEH 0533069 DUP050034339 Percent Urban Dwelters in Sample Attachment 2 3J Preanalyt'caf quality control There are many possibilities lo contaminate biological samples through neoTtf. unsuitable Wood collecting vials and contaminated anticoagulants (ZicfA Mitchell. 1976: Naclowtli at al~ 1977; Nisc A Vesterberg. 1978). Furthermore. contamination may originate from the akin if not properly cleaned. or from contaminated cleaning solutions (Braize! A Reed. 1974). Studies on commercially available Mood collection tubes have Shown that diluted nitric acid ai welt at Mood may extract lead and cadmium from certain vials and syringes in quantities which would invalidate any measurement ofthese metals in blood within the normal range of concentration (Nise A Vesterberg. 1971): To avoid contamination as much as possible, evacuated blood collection tubes (Vennjeet. Terumo Corp., Tokyo, lapan) with heparin from the same batch wen provided by the Cl jincr control of the metal content in a suitable number of tubes from the batch. Eight tubes were randomly selected from a boa of IDO heparinized Mood collection tubes and ten grams of 0.01 M nitric acid added to each tube. After 11 days of storage at room temperature the content offend and cadmium was analyzed by AAS (ETA). Some tubes were turned Over so that the acid came into contact with the rubber caps during storage. The results showed a mean cadmium concentration of 006 ag Cd/I m the acid solution, tmd a lead concentration offes* than 0.8 ug Pb/I (both dose to the detection limit of the analysis) in ah lubes tested- No Contamination from the rubber caps was noticed. A aimilar study on 20 Venoject tubes containing EDTA ar anticoagulant showed mean cadmium content of0.14 #g Cd/I and a mean had content of 0.8S | PM. It was derided 10 use tubes with heparin since they teemed to contain less metals compared to the lubes with EDTA. If in the future EDTA tubes with low metal content will be available, the use cif such tubes may have certain advantages (sec Appendix Ol It was recommended that before collecting bleed, the shin should be carefully washed qnd then cleaned with disposable napkins, saturated with 70% isopropyl alcohol fMcdiSwab. Pharma* Limited. Bexley. V.K.). provided by Cl after check for metal content. Written instructions for the sampling of blood were worked out (WHO. 1980a) and a demonstration was made at Cl during the meeting m Stockholm in May 1980- The par ticipating institutions were requested to prepare a protocol with information on the per sonnel collecting the samples, procedures for collection, transport and storage ofsamples and any additional procedure. After collection of the blood samples, the Mood from each tube was transferred to three S ml tubes of polypropylene (washed with diluted nitric acid and deionized water) provided by the Cl- They were deep-frozen as soon as possiMe- One ofthe three tubes was Stored to allow duplicate analysisat the Cl or a reference laboratory. Blood not used for the initial analysis was stored lo make possible reanalysis at the laboratory if necessary. The risk ofa significant contamination was considered leas pronounced for kidney cor tex samples owing to the higher concentrations ofcadmium in the kidneys than in blood. To avoid contamination to the greatest possible extent and to get comparable samples, the laboratories received a film showing procedures for collection of kidney cones samples at autopsies. The film was produced by WHO/tAEA in relation to the project "WHO/IAEA Joint Research Programme on Trace Elements in Cerdkvascutv Disease* (Autopry Studies)" (Mesironi A Parr, 1979). The k kJncys were to be opened longitudinally (with e-gstainless sleet knives) and a slice containing the cortex and medulla was to be isolatedProm this slice pyelic fat should be eliminated and a portion of the cortex carefully hoisted and collected, then put in a suitable container of polypropylene or polyethylene (previously washed with diluted nitric acid and deionized water) and deep-frozen as soon as possible. After collection the samples were kept deep-frozen until analysis. Some ofthe collected materia) was kept in storage to enable reanalysis and/pr duplicate analysis at a reference laboratory, A description of the procedures used for the kidney cortex collection was prepared by each laboratory and sent to Cl. 10 TEH 0533070 DUP050034340 N33868.03 Attachment. 3 2J itntra- anil Inlertaboralory comparison studies on (lie analysts oflead and cadmium and mlerlahoretory comparison studies have been carried owl in the past by different laboratories and organizations. Aa the mulls of sucfi studies may be ofimportance in in* lerprciing the validity of reported data, some of these are summarised below. The American Industrial Hygiene Association sponsored interlaboratory comparisons latest the accuracy of blood lead analysis (Keppler ct at 1910). In this study samples of spiked Mood 1200-2910 n WO were sent to ten different laboratories to be analysed by methods routinely used by these laboratories. The overall variation m results was large Ibr both the high and the very Ipw concentrations. Results of op to 4)00 grg Pb/I wen reported for normal Mood and down to 0 grg Pb/I for Mood spiled with 1050 grg Pb/l Then was also great variation between multi obtained from the tame laboratory with time. The problem was pot associated with specific analytical techniques. Donovan et aMI9TI) have reported on lead analysis at laboratories in Pennsytvania. USA. iln response to an enquiry to 267 different laboratories, 20 laboratories agreed to participate in an intercomparison programme for analysts of lead in urine. Spiked urine 11)6 irg Pb.1 urine) was sent to each laboratory to be analyzed by the method normally used at the laboratory. Reported results ranged from being 10 times lower to 20 limes higher than the "true** value. When notified of their inaccuracy, some ofthe laboratories sought technical assistance. Seven laboratories with relatively good results on lead analysis in urine were later provided with blood samples spiked with 155 Mt Pb/I Mood. The results reported ranged between 10 and 150 ug Pb/I Mood. In 1972. the reference laboratory ofthe European Intercomparison Programme (BcrSii et a).. 197)1 sent an aqueous solution of lead nitrate (100 gig Pb/I) and three samples of blond (two from persons occupationally exposed to lead end one from an unexposed per son) tn 22 different laboratories. Methods used for analyses included atomic absorption spectrophotometry, dithiione extraction and colorimetry, pofarqgraphy and emission spectrophotometry. For the aqueous solution, results varied from 51 to 190 grg Pb/L but 70% or the results did not deviate more than 10% from the True** values. Blood lead levels were reported to range from MO to 960 ug Pb/I (median: 500 Mt Pb/I). 210 to 1170 gig Pb/I (median; 660 grg Pb/I) and 120 to 740gig Pb/l (median; 410gig Pb/I) for the three samples, respectively. ` targe variations were reported also in ,,c*pefie**ced*, laboratories by Browne ft `a). (1974). Samples ofheparinized Mood (45 in 1973 and 49 in 1974) from lead workers were lent to three different laboratories for analysis by atomic absorption spectrophotometry or anodic stripping voltammetry. Each laboratory had been reported to handle more tlpa 2,000 blood samples per year for lead analysis. The mean difference between results reported by the participating laboratories was as high as 290 grg Pb/I blood in 197) and 440 g>g Pb/I blood in 1974. In another study (Lemcr, 1975) one Mood sampk obtained Bom a single person was divided into 35 separate samples and sent to a well recognised laboratory (the Kettering Laboratory) over a period of nine months together with other samples. A considerable variation in lead levels was found with a mean of 191.3 grg Pb/I Mood and a standard deviation of 57.2 grg Pb/I. The values ranged from 120 to 420 g/g Pb/I blood. 11. TEH 0533071 DUP050034341 Liuwfcrjn ft al. (It75) Cvafusted the results from ft* European laboratories par ticipating in ah fotertaboratory comparison programme for the analyst* ofcadmium, lead nd mercury in water, Wood and urine-The analytical metHod* used were atomic absorp tion spectrophotometry (flame and RamelessX anode stripping voltammetry, colorimetry nd neutron activation analysis. There were large variations in results for aB the metals nd all the media analysed. As an example three samples oflead in Wood showed median values, inlerlsboralory coefficients of variation (CV) and ranges as follows: median 128 pg Pb/I. CV 52.2%. range 27-490 pt Pb/I; median 227 pg Pb/t. CV 42.9%. tang* 103--973 pg Pb/I; median 233 pg 7W, CV 773%, range 10--I ISO pg PM. Correspon ding results for cadmium in the same Wood samples were: median 7 pg Cd/L CV !%. range 1-92 pgCd/t; median *pgCd/),CV I IftK. range 0-73 pgCdfc median 10 pg Cd/I. CV (43%, range 0--ItOpg Cd/I. The variability of the results, according to the authors, could not he attributed to either the different analytical methods used or to the difference in eaperieitcc in trace metal analysis at the laboratories. Paufov et at (1978) sent Wood, urine and aqueous solutions spiked with lead and cad mium to five laboratories, three In Scandinavia end one each in Britain and Canada. Recovery of the added lead in the Wood samples (470 pg Pb/1) ranged between 250 and 470 j/g Pb/I. The results for two Wood samples spiked with cadmium (20 and 67 pg Cd/I) tanged for three laboratories between 10 and 21 pg Cd/I and 40 and ft2 pg CdA. respec tively. Maher ct al.(l979) sent pooled Wood from lead workers to 24 laboratories on nine oc casions during 1974 and 1977. They tested different methods but found no significant bias that couM be attributed to the method ofanalysts as |u<h. in sit different rounds, the coef ficient of variation (CV) varied between 1.7% for a sample with a mean value of J7$ pg Pb/I and 10.7% for a sample with a mean value of 250pg Pb.1. The CV did not impro*-c aignificantly with time. It was Concluded thtt with avt'laWc methodologies (atomic absorption ipcctiophotomelry, flameless and Delves Cup. dithtione extraction, colorimetry, anodic stripping voltammetry and spectroscopy) the optimum coefficient of variation between laboratories would be about 7%. Boone ct al. (1979) compared the mutts from 113 laboratories participating in the "Blood Lead Proficiency Testing Program" conducted by the Center for Disease Control. USA, with the results from isotope dilution mass spectroscopy at the U.S. National Bureau of Standards. The Wood for the interlaboratory comparison was obtained from cattle orally fed with lead nitrate, and the concentrations ranged from ) 30 to 1020ng Pb/I blood- Twelve separate samples were dispatched to each laboratory for analysis. It was Concluded that most methods overestimated the lead concentration when the actual con centration was low (less than 400 Pb/I blood) and underestimated it when the actual Concentration was high (more than 500pg PbA blood). The overall coefficient of variation a ranged from 29 to 73% at NB$ values of 124 pg Pb/I Wood and from 9 to 37% at an N8S value of 1020 pg PbA. The reviewed studies show that the accuracy and precision of trace metal analysis fai biological specimens, especially lend and cadmium in Mood and urine, in general appear to be unsatisfactory. This may also hold for laboratories that have gained considerable ex perience by analyzing a large number of biological samples over many years. Of the analytical methods currently available for analyzing trace metals in biological materials, no single method has been found to be distinctly superior to the others. .12 TEH 0533072 DUP050034342 Attachment 4 Correlation Coefficients^1^ Between Weighted Geometric Mean Blood Lead and ___ Sample Proportion by Demographic Groups Demographic Group Race White Black Total Sample .084 -.364 Urban Areas 1-4 .896** -.001 Rural Area -.862** .002 <6 Years 7-17 Years 18-74 years Sex Male Female Race x Age White, <6 Years Black, <6 Years White, 7-17 Years Black, 7-17 Years White, 8-18 Years Black, 8-18 Years Race x Sex White, Male Black, Male White, Female Black, Female Age x Sex <6 Years, Male <6 Years, Female 7-17 Years, Male 7-17 Years, Female 8-74 Years, Male 8-74 Years, Female .506 .734* -.841** .514 -.509 . 495 .505 .702 .308 -.344 -.650 .203 -.438 .010 -.349 .194 .970** -.813** .636 -.253 -.621 .881** .898** .805** .872** ,799** .871** .715* .947** .391 .866** -.190 .915** -.243 .861** .064 .892** .964** .780* .811** .713* .785* -.587 -.853** -.837** -.879** -.799** -.741* .177 -.865** .148 -.839** -.176 -.887** -.177 -.818** .232 -.781* -.686* -.849** -.202 -.804** -.783* Statistically different from zero at .05 test level, ^Statistically different from zero at ,01 test level. '1*BaSed on nine pairs of data corresponding to 6-month period, starting January 1, 1976 through June 30, 1980. 13, TEH 0533073 DUP050034343 Attachment 5 Blood Lead Levels - Observed and Adjusted for Degree of Urbanization and SMSA Category Overall Race White Black Observed Adjusted Observed Adjusted Observed Adjusted Geometric Mean Blood Leadf pg/dl 1st Half 1st Half 1976 1980 A'1' Effect(2> 14.8 13.8 8.8 9.6 6.0 4.2 1.8 14.5 13.7 16.3 14.7 8.6 9.3 10.8 11.7 5.9 4.4 1.5 5.5 3.0 2,5 *2 .5-5 Yrs. Observed Adjusted 6-17 Yrs. Observed Adjusted 18-74 Yrs. Observed Adjusted 18.5 17.5 13.4 12.5 15.1 14.1 9.4 10.2 6.9 . 7.6 9.3 10.1 9.1 7.3 1.8 6.5 4.9 1.6 5.8 4.0 1.8 Race by ,.5<-5 Yrs. Whi te Observed Adjusted Black Observed Adjusted 17.8 17.1 21.6 19.4 8.7 9.5 12.7 13.8 9.1 7.6 8.9 5.6 1.5 V2 a J Difference between first two columns. Effect on trend when accounting for degree of urbanization and SMSA category. 14. TEH 0533074 DUP050034344 TABLE Cl Federal Regulator Actions Governing Human Exposure to Lead Attachment 6 Medium through Which Exposure Occurs Federel Agency Specific Exposure Fuhwiy Being Controlled Federal Regulations Gistion Nonfood Substances Lead paint in housing Housing and Urban Development Lead paint in housing Health, Education, and Welfare Exposure ofchildren to leadbased paint on surfaces of residential structures. Exposure ofchildren toleadbased paint on surfaces of residential structures. Notification to purchasen and tenants ofHUtHttsodated housing constructed prior to 1950 ofthe hazards ofleadbased point poisoning. Prohibition against thauaaof lead-based print m h u d associated housing. Elimination oflead-based print hazards in HUP-aaoriatcd bousing. Elimination oflead-based paint hazards in federally owned properties prior to sale for residential habitation. Prohibition against the use of lead-based paint in federal ' and federally assisted construction orrehabilitation ofresidential structures. Grants todevelop local programs for detection and treatment oflead-based print poisoningand for the Identification and elimination 24 cm 35.5 24cm 35.14 24 cm 35.24 24 cm 35.56 24cm 35.62 42 cm 91 4 Lead in paint, and toys and furniture with printed surfaces Air Ambient air Ambient air (including dust and din) Consumer Product Safety Commission Environmental Protec tion Agency Environmental Protec tion Agency Exposure ofchildren to Idad* bared paint on surfaces of residential structures, toys and furniture. Exposure ofthe general population and especially young children (1-5 yrs.) to airborne lead from motor vehicles combusting leaded gasoline. Exposure ofthe general population and especially young children (1-5 yrs.} to airborne lead from stationary and mobile emissions sources. ofthe hazards oflead print in residential structures. The following consumer products have been declared as banned hazardous products; (1) paint and other similar surface coatings containingmore than 0.06 percent lead. (2) toys and other articles intended for use by children that bear paint containing more than 0.06 percent lead. (3) furniture that bears paint containing more than 0.06 percent lead. No gasoline refiner shall exceed an average lead content of0.5 g Pb/gal after October I, 1979, averaged overt three-month period. The national primary and secondary ambient air quality standards for lead are 1.5 fig/m*. maximum arithmetic mean averaged over a calendar quarter. 16 cm 1303.4 40 cm 80,20 40 cm 50,12 15. TEH 0533075 DUP050034345 TABLE C.1 {continued) Medium through Which Exposure Occurs Federal Agency Wster Drinking water Environmental Proteclion Agency Food and Earing Utensils Apples, apricots. celery, peaches. pears, straw berries, tomatoes, and several other specific fresh fruits and vegetables Gtrus fruits Environmental Protec* tion Agency Pathway Being Controlled Exposure ofthe general population to lead in drinkingwatersupplied by public watersystems. Exposure ofthe general population to lead arsenate pesticide residua in of on raw agricultural commodities. Ceramic dinner* ewe, enamelware and pewter Rood and Drug Administration Exposure ofgeneral population to teadthat is teachable from pottery glams Federal Regulations The interim national primary thinking watermaximum contaminant level for lead is so^n. The tolerance for lead arsenate pesticide residues in or on ten designated raw fruits or vegetables is limited to 7 ppm ofconfined lead. The tolerance for leadarsenate pesticide on atrui fruits is (unitedto | lead. The actionlevel for teachable lead from ceramic work is 7.0 pg/ml, 5-0 Mg/ml andZ5 Mg/ml for flatware, smafi hollow ware, and large hollow ware, respectively. o Citation 40cntl41.lt 400*190.194 PDAAdnrini* native Guideline 7417.00 Silver-plated hollow ware Food and Dtug Administration Occupational Exposures Occupational exposures to lead in all industries coveted by u s h a except construc tion and agriculture Occupational Safety and Health Administration Exposure ofthe general population to lead that is teachable from silver-plated hollow ware. Exposure, to airborne lead in the workplace. The action level for teachable lead fromsilver-plated hollow ware for use by adults is 7.0 Ag/g. The action level for teachable lead from silverl^ated cups intended for use by infants is 0.5 g/mL 50 Ag/m averaged over an 8hour period. FDa Administratjve Guideline 74174)1 29cnt 1910.1025 16 TEH 0533076 DUP050034346