Document peZ6BDQGv8D0gqbKJzQjXQNbB

3Dt. a.b OCT 2 3 1985 Sw^fesis Lead, Blood Pressure, and Cardiovascular Disease To: Lester Grant Paul Mushak From: Joel Schwartz This memo has two purposes. First, to call to your attention several additional articles on lead and blood pressure that probably bear mentioning in chapter 12, and second to discuss the question of drawing inferences to cardiovascular disease, which you will address in your addenda. X am enclosing two additional human studies on lead and blood pressure. One, from England,(Khera et al. Int J. Environmental Studies 1980 vol 14:309-312) was a case control study that found significantly higher lead levels in the blood of hypertensive and cardiovascular patients than in the blood of hospital controls. Interestingly, the lead levels were highest in the hypertensive patients, and elevated, but not as much, in the patients with ischaemic heart disease. This is what one would expect, since ischaemic heart disease is caused by other factors besides blood pressure. The second human study is of a small sample of men in 2utphen, Again, a significant correlation between blood lead and blood pressure was found. Two recent animal studies by Meridth et al. have found evidence of increased plasma renin in animals and humans exposed to moderate levels of lead, and they suggest it may be part of a mechanism whereby lead promotes hypertension, I have also enclosed them. In addition, there were a number of papers that I discussed in my memo to CASAC last spring which report the results of animal studies on lead and blood pressure, which I had not -located when you were drafting the Corrigenda. The criteria document should at least refer to them, and several, such as Ficinnini et al. and Carmignani et al. merit at least some discussion. Revis et al. probably merits mention just because it is in a different species. I have copies of most of the papers, so please let me know if you have any difficulty locating them. LEAD, BLOOD PRESSURE, AND CARDIOVASCULAR DISEASE Fortunately, the question of whether higher blood pressure means higher risk of cardiovascular and cerebrovascular disease is not one that EFA needs to settle. Every prospective cardiovascular epidemiological study has found current blood pressure to be a risk factor in future cardiovascular and cerebrovascular events. There is no debate in the medical profession over this question. Table 1 shows the striking nature of these results for the largest study, the Framingham study. Notice in particular that it shows no discontinuity or threshold at the level currently defined as hypertension. Rather* the risk of myocardial infarction is a continuous function of blood pressure TEH 0412493 This issue of whether lowering a persons blood pressure will result in a reduced rate of cardiovascular and cerebrovascular disease is also one that EPA does not have to take the lead on. The National Institutes of Health periodically convenes "consensus panels" of experts on certain diseases to reach conclusions related to those diseases. The 1384 report of the Joint Committee on Detection, Evaluation, and Treatment of High Blood Pressure (Archives of Internal Medicine May 1984, vol 144) concluded that: "Risk related to hypertension increases continuously as systolic and diastolic blood pressure rise." 'The goal of treating patients with hypertension is to prevent the morbidity and mortality attributable to high BP. This means the reduction of elevated blood pressure to the extent that excess cardiovascular risk is eliminated." "Reducing blood pressure with drugs decreases cardiovascular mortality and morbidity in patients with moderate and severe hypertension. Trials of antihypertensive drugs in patients with mild hypertension have uniformly shown protection against stroke, left ventricular hypertrophy, congestive heart failure and progression to more severe levels of hypertension." . "It is apparent from prospective epidemiologic studies as well as from pooled actuarial data that the lower the systolic and diastolic blood pressure, even within the normal range (emphasis added), the better the prognosis." This conclusion is well accepted in the medical community; they have voted with their prescription pads. Over 2u million Americans are no receiving anti-hypertensive medication, at a cost of about $5 billion dollars per year, plus the incidence or undesirable side affects. This medication is being given to reduce the risk of heart disease and strokes. For ETA to .conclude that lowering blood pressure would not reduce tne risx of heart disease and stroke would be for it to assert that there are 30 million continuing cases of medical malpractice, as well as for it to reject the conclusions of the N1H expert panel. The NIH committee's recommendation is supported by the epidemiological evidence. The hypertension, Detection, and Followup program found that reducing blood pressure even in people with only slightly elevated blood pressure recuceo tne occurrence of coronary events. (NEJM 19b3;3u/:s/b-youi. me Australian national trials showed a substantial decrease in cardiovascular morbidity and mortality tor patients witn diastolic blood pressure from 95-109 mm Hg, who receiveo active treatment compared to a group receiving a placebo, u-ancet 1980,1:1261-1267). Recently Levy examined the 27% decline m coronary near t TEH 0412494 DUP050453546 i' disease deaths between 1972 and 1982 to elucidate the causes. (Am. J. Cardiology 1984 v.ol 54 7-130. Noting that the survival rates for 3 years after acute myocardial infarction have not changed in the last 20 years, based on an almost complete ascertainment of all such cases in Baltimore by the Johns Hopkins epidemiology department (Goldberg et al. Johns Hopkins Med. J. 1979 144:73-80). and that the coronary artery surgery study (circulation 1981; 63 supp; Is 11-181) also failed to find improved survival from bypass operations, he examined the changes in risk factors. He found that inserting the observed changes in smoking, serum cholesterol, and blood pressure in Framingham regression results predicted a 22% drop in mortality over the period, compared to the observed 27%, and concluded that risk factor modification was the major cause. This again supports the belief that reducing blood pressure will reduce the chance of heart disease. Moreover, the magnitude of the effect of lead on blood pressure, as indicated by the NHANES regressions, is similar to the size of the reduction achieved by hypertensive medication. The Hypertension detection and Followup Program (NEJM 1983;307:976-980) found a 20% decrease in mortality from an intervention program that reduced blood pressure by about 5 mmHg. The NHANES regressions (Pirkle et al. AJ Epid. 1985) shows that a reduction in blood lead levels from 17ug/dl to 6ug/dl would produce a similar drop in blood pressure. There aren't very many direct studies of lead and cardiovascular mortality, and they are usually occupational, and hence confounded by the healthy worker effect. However, the most recent analysis of Cooper found an SMB for hypertensive heart disease of 128 and 208 in battery workers and smelter workers respectively. More interestingly, a general population autopsy survey (Voors et al. Arch. Env. Health 37;98--102 1982) looked at lead concentrations in aorta and whether the death was due to heart disease. Lead was highly significant as a predictor of cardiovascular death, and explained 12% of the variation. The finding of Khera et al. of an association between blood lead and heart disease in a general population is also confirmatory. I believe that these results indicate that accepted medical judgment finds that decreases in blood pressure are expected to result in decreases in cardiovascular and cerebrovascular disease, and that the limited data available directly relating lead and cardiovascular disease in the general population supports this expectation. TEH 0412495 DUP050453547 An h u c a n Jo u r n a l or Et id e mio l o o v Copyright01985 by Tbs Johns Hopkins UnhvnitySchool ofHygisiw and Public Htaltb All ti(btiMciv*d TRACE METALS AND CORONARY HEART DISEASE RISK INDICATORS IN 152 ELDERLY MEN (THE ZUTPHEN STUDY) D. KROMHOUT,* AAR WIBOWO* R. F. M. HERBER* L. M. DALDERUP,* H. HEERDINK.* C. d e LEZENNE COULANDER,1 a n d R. L. ZIELHUIS* Kromhout, D. (Institute of Social Medicine, State University of Leiden, 2300 RC Leiden, The Netherlands), AAR Wibowo, R. F. M. Herber, L M. Daiderup, H. Heerdink, C. de Lezenne Couiander, and R. L Zielhuis. Trace metals and coronary heart disease risk indicators in 1S2 elderly men (The Zutphen Study). Am J Epidemiol 1985;122:378-85. Information about trace metals and coronary heart disease risk indicators was collected in 1977 among 152 men aged 57-75 years in the town of Zutphen, the Netherlands. Serum zinc, serum copper, blood cadmium, and blood lead were determined by atomic absorption spectrometry and serum lithium by flame emission spectrometry. After uni* and multivariate regression analysis, the follow ing statistically significant relations were found: serum zinc was inversely related to resting heart rate; serum copper was positively related to cigarette smoking and inversely to high density lipoprotein cholesterol; blood cadmium was strongly positively related to cigarette smoking and Inversely to Quetelet Index; the positive relation between blood lead and cigarette smoking was of borderline ignlficance; and blood lead was related to btood pressure, with the relation being stronger for systolic than for diastolic blood pressure. blood pressure; cadmium; cholesterol; copper; coronary disease; lead; smok ing; zinc Descriptive data on serum and blood trace metal distributions in elderly popu lations ate scarce. This type of information Received for publication July 27,1934 and in final form February 6,1985. 1 Institute of Social Medicine, Faculty of Medicine, State University of Leiden, P.O. Box 9605, 2300 RC Leiden, The Netherlands. (Reprint requests to Or. O. Kromhout at this address.) ' Cotonel Laboratory for Occupational and Envi ronmental Health, Faculty of Medicine, University of Amsterdam, Amsterdam, The Netherlands. * Factory inspectorate, Ministry of Social Affairs, Amsterdam, The Netherlands. This study was supported by grants from the Neth erlands Organization for Food and Nutrition Re search, the Netherlands Nutrition Council, and the Netherlands Prevention Fund. The authors thank H. Pieters and P. del Castilho, Corone! Laboratory, Amsterdam, and B. Nijhof, Lipid Laboratory, Department of Human Nutrition, Agri cultural University Wageningen (Head: Prof. Dr. J. G. A. J. Hautvast) for their contributions to the trace metal and cholesterol analyses. 1 is of interest because cadmium and lead have been suggested to play a role in the etiology of hypertension (1-4). Animal ex perimental evidence (5) suggests that the dietary zinc/copper ratio may be a deter minant of coronary heart disease, and, based on descriptive epidemiologic data, a protective effect of lithium on coronary heart disease has been postulated (6). To our knowledge, with the exception of the notion that Mood cadmium and lead levels are higher in smokers than in non-smokers (7), the relationships betweeu serum and blood trace metals and established coro nary heart disease risk indicators have not been studied in elderly populations. It was therefore decided to collect infor mation about lithium, copper, and zinc in serum and cadmium and lead in blood in a 378 subgroup of 153 m participated in the phea Study. Also heart disease risk ured: total and fc cholesterol, systol pressure, smoking dex (weight/heigh the results of the trace metals and risk indicators am between them. Ma t e r ia l s Since 1960, a lo of the relations h indicators, and coi been carried out t from the town ofZ: The Zutphen Stud tribution to the Se 10). In 1960, a ram bom between 1900 for at least five yt iected for the Bto men, 919 (84.5 p< examined. Between 1960 as went a medical exa and 1978, the 15tl of all the men who per cent) were re-e: men examined in Iected of 152 men had lived since 19f this sample, relatic m m tween the type of copper) and some blood. Informatioi w disease risk indies Statistically signif onary heart disea m were not found bel ticipated in the tr other men examini v**r Information abc obtained by a que; -'5* the men at home, determined accorc TEH 0412496 N 30996.01 DUP050453548 Vgi.i2i.No.ai Prmttd in VJSA.-i 2 RISK STUDY) M. HEERDINK,* , 2300 RC derup, H. coronary y). Am J itofswas 5hen, the tad were vf flame e followy related smoking strongly lex; the irderline relation ; smok- am and lead a role in the ). Animal ex* ests that the y be a deter* disease, and, ilogic data, a on coronary fated (6). To ption of the d lead levels aon-smokers i serum and Iished coro* its have not ions. ollect infor* and zinc in n blood in a TRACE METALS AND CORONARY HEART DISEASE RISK 379 subgroup of 152 men aged 57-76 years who participated in the 15th round of the Zutphen Study. Also the following coronary heart disease risk indicators were meas ured: total and high density lipoprotein cholesterol, systolic and diastolic blood pressure, smoking habits, and Quetelet in* dex (weight/height2). This paper reports the results of the descriptive analyses of trace metals and coronary heart disease risk indicators and discusses the relations between them. Ma t e r ia l s a n d me t h o d s Since 1960, a longitudinal investigation of the relations between diet, other risk indicators, and coronary heart disease has been carried out among middle-aged men from the town ofZutphen, the Netherlands. The Zutphen Study forms the Dutch con tribution to the Seven Countries Study (&10). In 1960, a random sample of 1,088 men born between 1900 and 1919 who had lived for at least five years in Zutphen was se lected for the study, and, of these 1,088 men, 919 (84.5 per cent) were medically examined. Between 1960 and 1973, the men under went a medical examination yearly. In 1977 and 1978, the ILth round took place, and, of all the men who were still alive, 611 (92 per cent) were re-examined. Among the 473 men examined in 1977, a sample was se lected of 152 men aged 57-76 years who had lived since 1960 in the same house. In tiiis sample, relations were investigated be tween the type of waterpipes (e.g., lead or copper) and some trace metal levels in blood. Information about coronary heart disease risk indicators was also collected. Statistically significant differences in cor onary heart disease risk indicator levels were not found between the men who par ticipated in the trace metal study and all other men examined in 1977. Information about smoking habits was obtained by a questionnaire completed by the men at home. Height and weight were determined according to a standardized protocol by a trained staff (8). Blood pres sure was measured according to a standard ized protocol by one internist using a mer cury sphygmomanometer (8). Blood pres sures were taken at the right arm with the men in a supine position. The first record ing was done at the beginning and the second and third at the end of the medical examination. Only the systolic and dia stolic (fifth phase) values ofthe third meas urement were recorded. Resting heart rate was calculated from the electrocardiogram. Non-fasting serum total and high density lipoprotein cholesterol determinations were carried out at the Lipid Laboratory of the Department of Human Nutrition, Ag ricultural University Wageningen, the Netherlands. Serum total cholesterol was determined automatically according to Huang et al. (39) using serum standards in which total cholesterol was determined ac cording to Abell-Kendall (11, 12). High density lipoproteins were isolated by pre cipitation of the apo-B containing lipopro teins with heparin-manganese (13). The cholesterol content of the high density li poprotein fraction was determined in the same way as described for total cholesterol (14). During the time that the cholesterol determinations were carried out for the present study, the Lipid Laboratory was standardized according to the criteria ofthe Lipid Standardization Laboratory of the Centers for Disease Control, Atlanta, GA. To prevent contamination, cadmiumand lead-free disposable syringes and polyethylene tubes were used for blood col lection. Lead-free heparin was used to pre vent blood clotting and the blood samples were deep-frozen till analysis took place. Blood lead, blood cadmium, serum copper, and serum zinc were determined by elec trothermal atomization atomic absorption spectrometry with improvements in preci sion by using a peakshape monitoring de vice (15). Serum lithium was determined by flame emission spectrometry. The ac curacy of blood lead and blood cadmium was checked by participation in interna- TEH 0412497 iSEfifc m ill to rss P*g sgjjaf iM m to DUP050453549 380 KROMHOUT ET AL. tional round robin studies. The inaccuracy was 50 jig/liter for blood lead and 0.7 jig/liter for blood cadmium. The accuracy of serum copper and serum zinc was checked in a comparison between the in ductively coupled plasma spectrometry and the electrothermal atomization atomic ab sorption spectrometry method (16). The inaccuracy for zinc and copper was for both 0.1 mg/liter. For lithium, no comparison could be made. For the statistical analyses, SPSS pack age programs were used (17). Besides de scriptive statistics, means and standard de viations, correlation coefficients were cal culated for continuous variables. Analysis of variance was used if the dependent vari able was continuous and the independent variable categorical. Multiple comparisons were tested by the Scheffe method (17). Multiple regression analyses were carried out using coronary heart disease risk indi cators as dependent variables and blood trace metals and other determinants of these risk indicators as independent vari ables. For variables with skewed distribu tions, log transformations were used in uniand multivariate regression analyses. No difference was observed whether trans formed or untransformed variables were used. Therefore, only the results of un transformed variables will be reported. - Re s u l t s The levels of the coronary heart disease risk indicators were generally high in this elderly population (table 1). If a cut-off point of 160/95 mmHg was taken, the prev alence of systolic hypertension amounted to 34.4 per cent and that of diastolic hyper tension to 37.1 per cent. The prevalence of obesity (Quetelet index >27 kg/ms) was 26.5 per cent and the prevalence of hyper cholesterolemia (serum cholesterol >260 mg/100 ml) was 17.9 per cent The per centage of current smokers amounted to 64.9 per cent The serum concentrations of the essen tial trace metals, lithium, copper, and zinc were normally distributed (table 2). The means for the nonessential blood metals, cadmium and lead, were higher than the medians, indicating that the distributions were skewed to the right Blood cadmium levels above 5 #ig/liter were not found. Blood lead levels above 300 pg/liter were present among 8.6 per cent and levels above 400 Mg/liter among 1.3 per cent of the el derly men in Zutphen. Serum copper was significantly related to cigarette smoking (table 3). A dose-re sponse relation was present but only the men smoking 10 or more cigarettes/day had significantly higher serum copper levels than n^nsmokers. Serum copper was in versely related to high density lipoprotein cholesterol (r -0.22, p < 0.01) (table 4). This relation became somewhat stronger (standardized regression coefficient = --0.27, p < 0.01) after multiple repession analyses taking other determinants of high density lipoprotein cholesterol, e.g., Que telet index, age, cigarette smoking, and resting heart rate, into account. Ta b l e 1 Coronary heart disease risk indicator distributions of 152 men aged 57-76 years in Zutphen, the Netherlands. 1977 Risk indicator Mean Standard deviation P10 P50 P90 Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Total cholesterol (mg/100 ml) High density lipoprotein cholesterol (mg/100 ml) Quetelet index (kg/m1) Besting heart rate (b/min) 154.3 22.0 126 149 184 91.8 14.0 75 91 107 226.9 38.2 180 223 276 46.3 11.2 25.4 2.8 75.4 12.3 33 45 59 22 26 29 59 74 93 Trace mete Trace melsl Serum lithium <*ig/lit y. Serum copper {mg/lh ! Serum zinc (mg/liter) Blood cadmium 0g/li " V. Blood lead Witter) , Analyses ofvariant No. Of C. . dcinttCA . mokd/dy ~.v * . Nona I"' <1 :T" 10 No. or men 53 58 40 * Contrast, p< 0.0 Correlations betwee. Trace metal Seram lithium Serum copper Seram zinc Blood cadmium Blood lead *0.01 p< 0.05. "0.001 sp<0.0: "*/>< 0.001. A dose-respon tween blood cadi ing (table 3). T1 cigarettes that wi blood cadmium 1 also significant!; and Quetelet ind regression mode! die dependent va ing, age, and Que variables, only ci telet index remss to blood cadmh between blood ca TEH 0412498 DUP050453550 as taken, the tnt tension amount ofdiastolic hyp4 The prevalence! >27 kg/m5) wf rvalence of hype; cholesterol >26ff r cent. The pe%.J ers amounted * 3ns of the essen-g copper, and zinc 1 (table 2). Th| al blood metals! higher than the! he distributions* Blood cadmiuml ere not founds X) ag/liter were^ and levels above = cent of the el- ficantly related 3). A dose-re* t but only the irettes/day had copper levels upper was insty lipoprotein 101) (table 4). what stronger ^efficient = pie regression inants of high ol, e.g., Quesmoking, and it. Wen P90 TRACE METALS AND CORONARY HEART DISEASE RISK 381 Ta b l e 2 Trace metal distributions of 152 men aged 57-76 years in Zutphen, the Netherlands. 1977 Truce met*! Mean Standard deviation P10 PSO P90 s.T.iin lithium (iwg/lit*r) srnini copper (mg/Iiter) serum line (mg/liter) l({,nl cadmium (#g/liter) ulnoci lead (jig/liter) 18.3 2.1 16.3 18.5 20.8 1.21 0.22 0.91 1.18 1.52 0.79 0.12 0.61 0.78 0.94 1.55 0.89 0.68 1.26 2S1 183 74 108 167 280 Ta b l e 3 Analyses of variance of xerum and blood trace metals on cigarette smoking in 151 men aged 57-76 years in Zutphen, the Netherlands, 1977 So. of cigarettes moked/diy No. of sen None <10 210 53 58 40 ?ncc element (mean standard deviation) Serum copper Img/lher) Blood cadmium ta/Uter) Blood teed (xg/Iiteri 1.14 0.22 -I 1.21 0.20 * lAl 0.23-1 F 7.3, p- 0.001 LOS 0.41"j] 1.48 0.794]* 2.32 1.00J] F 33.5, p< 0.001 172 63 176 64 209 93 F 3.5, p 0.032 ' Contrast, p < 0.05. 1 Ta b l e 4 t nrrelations between trace metals and coronary tear? disease risk indicators in 151 men aged 57-76years in Zutphen, the Netherlands. 1977 Trace metal Ruk indicator Age Systolic blood pressure Diastolic blood pressure Total cholesterol High dentity lipoprotein cholesterol Quetelet index Resting bean rate Serum lithium Serum copper Serum zinc Blood cadmium Blood lead -0.02 0.08 -0.05 -0.17* 0.03 0.09 0.04 -0.02 -0.12 0.24" 0.06 0.07 -0.00 -0.10 0.16* -0.10 0.09 0.00 0.07 0.07 -0.09 -0.22" -0.06 -0.02 0.08 -0.06 -0.07 0.09 -0.24" 0.11 0.12 0.15* -0.16* 0.13 -0.02 0.01 p< 0.05. "0.001 p< 0.01. "p<0.(K>l. A dose-response relation was found be tween blood cadmium and cigarette smok ing (table 3). The greater the number of cigarettes that were smoked, the higher the blood cadmium level. Blood cadmium was also significantly inversely related to age and Quetelet index {table 4). In a multiple regression mode! using blood cadmium as the dependent variable and cigarette smok ing, age, and Quetelet index as independent variables, only cigarette smoking and Que telet index remained significantly related to blood cadmium. The inverse relation between blood cadmium and age (r-- -0.17, p < 0.05) observed in univariate analysis disappeared after multiple regression anal yses due to the inverse relation between cigarette smoking and age (r = --0.27, p < 0.001). Blood cadmium and serum copper were significantly related (r = 0.29, p < 0.001). These trace metals were both significantly related to cigarette smoking (table 3). Therefore a partial correlation was calcu lated controlling for cigarette smoking. This partial correlation coefficient amounted to 0.15 (p < 0.05). Blood lead was borderline significantly TEH 0412499 DUP050453551 382 KROMHOUT ET AL. related to cigarette smoking (table 3). A dose-response relation was not present and the differences between the different cate gories of smokers, tested by Scheffe, were not statistically significant. Blood lead was significantly positively related to systolic and diastolic blood pressure (table 4). Mul tiple regression analyses with systolic blood pressure as the dependent variable and blood lead, age and Quetelet index as in dependent variables showed that the stan dardized regression coefficient for blood lead was reduced from 0.24 (p < 0.01) to 0.21 (p < 0.01). When a similar model was used for diastolic blood pressure, the stan dardized regression coefficient for blood lead was reduced from 0.18 (p < 0.05) to 0.15 (p = 0.05). Besting heart rate was significantly pos itively related to serum copper and signifi cantly inversely to serum zinc (table 4). The positive relations between serum lith ium and resting heart rate and between blood cadmium and resting heart rate were borderline statistically significant (p = 0.07 and p * 0.05, respectively). In a multiple regression model using resting heart rate as the dependent variable and the four trace metals as independent variables, ony serum zinc was significantly (p = 0.04) inversely related to resting heart rate. Dis c u s s io n The serum lithium levels seen in this 'study could not be compared with those observed in other studies due to differences in analytic methods. The average serum zinc level of the elderly men in Zutphen was similar to that of men aged 65-94 years in Belfast, Northern Ireland (18). In both studies, no relation was found between se rum zinc and age. That may be due to the small age range, because in a study includ ing men aged 20-84 years, a significant inverse relation with age was observed (19). In studies mainly including men aged 60 years or less (19-21)< average serum zinc levels of 1.0 mg/liter were found, compared to 0.8 mg/liter in the present study and in the Belfast study (18). This difference may be of physiologic importance hut methodologic differences between these studies in the determination of zinc cannot be ruled out as a major explanation. The average serum copper level of the Zutphen men was 0.1 mg/liter higher than in the elderly men from Belfast (18). No difference was found between the average serum copper level of the elderly men in Zutphen and middle-aged men from Hei delberg, West Germany (21). In the Zut phen men, serum copper was unrelated to age, probably due to the small age range. In men aged 20-69 years from Omaha, Ne braska (22), an increase in serum copper levels was found with age. A study on blood cadmium levels among men in a comparable age range, 56-72 years, was carried out in Sweden (7). The median cadmium level of these men was considerably lower than the median level of the Zutphen men, 0.8 versus 1.3 /tg/liter. This difference may partly be due to the different percentages of smokers, 46 versus 65 per cent. In a carefully standardized international cooperative study among men of an unspecified age range (23,24), median values^ of 0.4-0.9 pg/liter were found in Sweden, Yugoslavia, Israel, and the United States, while median values of 1.0-1.8 j*g/ liter were found in men from Belgium, Peru, Mexico, Japan, and China. These comparisons may also be confounded by differences in the cadmium concentration of the cigarettes and the smoking habits of the men in these countries. The percentage of smokers varied from 20 per cent for Peru to 64 per cent for Japan. It may be con cluded that the blood cadmium levels of the elderly men in Zutphen were comparable with those of men in countries with the higher levels, but values above 5 pg/liter were not found. The median blood lead level of men aged 56-72 years in Sweden (7) amounted to 76 Mg/liter, compared to 167 pg/liter among the men in Zutphen. In the above-men tioned cooperative international study (23, -* `7 ' : ff O % % J " ' 24), median le found among r Japan, while and 200 pg/lit< the United St Peru. The hig liter was four aged men from had average vt 25). Comparis difficult due t alcohol, smoki 23-25) on bloo difficulties, it blood lead levc phen ranked 1 median of 167 levels above 3 with levels abt Serum zinc related to rest, lated to total < cholesterol. In positive relatic zinc and high terol. A study men (28), shov high density 1 the men had i per day for fiv that the relati lipoproteins is In the prese was found bet' density lipopr and multivarif eral epidemio shown that lc cholesterol Ie% increased risk Serum copper coronary heart corroborated 1 control studies dle-aged perso coronary hear higher serum < Also, a statist! lation was four TEH 0412500 DUP050453552 Hu's difference these studies g ic cannot be rule on, JPper level of i}M t/liter higher thajf Belfast (18). jji ^ntheave^p e elderly men*^ !2m1)e- nIfrothme 2Hei|.l was unrelated tof smell age range;! Omaha, Nef| n serum copper-f m levels among s range, 56-72 `veden (7). The: these men was e median level ius 1.3 ug/litet. he due to the kers, 46 versus standardized iy amongmen 3> 24), median ere found in id the United >f 1.0-l.s ng/ Jm Belgium, hina. These 1founded by mcentration ng habits of i percentage nt for Peru y he conevels of the comparable s with the 5 ^S/liter men aged nted to 76 cr among ove-menstudy (23, TRACE METALS AND CORONARY HEART DISEASE RISK 383 it. median levels below 100 ^g/liter were total cholesterol (21). A similar, but not iimnd among men from Sweden, China, and statistically significant relation was ob .Lipan, while median levels between 100 served in the present study (table 4). Per .snd 200 >ig/liter were observed in men from sons with elevated total cholesterol or low ilu* llnited States, Yugoslavia, India, and density lipoprotein cholesterol levels are at IVru. The highest median value, 271 fig/ high risk for coronary heart disease (10, 'tier was found in Mexico City. Middle* 33). From these results, it may be hypoth aged men from England and West Germany esized that high serum copper levels are had average values around 150 pg/iiter (21, related to low levels of the anti-atherogenic Comparisons between countries are high density lipoprotein fraction and high difficult due to the confounding effects of levels of the atherogenic low density lipo alcohol, smoking, and traffic intensity (7, protein fraction. <-25) on blood lead levels. In spite of these Based mainly on the results of animal difficulties, it may be concluded that the experimental studies (5,34), the hypothesis hlood lead levels of the elderly men in Zut- has been formulated that a dietary copper l>hen ranked high internationally, with a deficiency either alone or in combination median of 167 jig/liter and 8.6 per cent with with an elevated intake of zinc is associated levels above 300 /ig/liter and 1.3 per cent with an increased total cholesterol level. with levels above 400 pg/liter. ( This relation was probably due to ex Serum zinc was significantly inversely tremely low levels of copper in the diet. related to resting heart rate but was jmre- When copper levels were used comparable laied to total and high density lipoprotein to the North American diet, no relation was cholesterol. In rats (26, 27), a significant found (35). It is questionable whether these positive relation was found between serum results are relevant for the relations be zinc and high density lipoprotein choles tween serum copper and serum lipids, be terol. A study in 12 healthy young adult cause serum copper may not be indicative men (28), showed a significant reduction in for copper intake (36, 37). Serum copper high density lipoprotein cholesterol after levels are, however, significantly positively the men had ingested 440 mg zinc sulfate related to cigarette smoking and coronary per day for five weeks. It may be concluded heart disease (21). Studies on the relations that the relation between serum zinc and between copper intake and exposure, serum lipoproteins is unclear. copper and coronary heart disease are ur In the present study, an inverse relation gently needed. was found between serum copper and high In the present study, an inverse relation density lipoprotein cholesterol alter uni- was found between blood cadmium and age. and multivariate regression analyses. Sev This relation disappeared after controlling eral epidemiologic studies (29-31) have for cigarette smoking due to the inverse shown that low high density lipoprotein relation between cigarette smoking and age. cholesterol levels are associated with an It may therefore be concluded that at least increased risk of coronary heart disease. in this study the decrease in blood cadmium Serum copper may therefore be related to with age is due to the decrease in the per coronary heart disease. This hypothesis is centage of cigarette smokers with increas corroborated by the results of two case- ing age. control studies (21, 32) carried out in mid The strong positive relation seen be dle-aged persons, in which persons with tween blood cadmium and cigarette smok coronary heart disease had significantly ing in the elderly men in Zutphen was also higher serum copper levels than controls. found among men of the same age in Swe Also, a statistically significant positive re den (7). This relation was also present in lation was found between serum copper and several studies carried out in younger men TEH 041250l"ipi DUP050453553 384 KROMHOUT ET AL. (23,24). It may therefore be concluded that lation, the individual with the highest blood Liebermann- cigarette smokers are at high risk for ele lead level (525 pg/liter) who also had hy vated blood cadmium levels. pertension (218/138 mmHg) was excluded " 1982:28:683- 13. Burstein M, cholesterol 1 iij i In the present study, an inverse relation from the analyses. Thereafter, a borderline serum. Clin < was noted between blood cadmium and significant correlation was found between 14. Van deT Has et al. Metho Quetelet index. That may be due to the blood lead and systolic blood pressure, and cholesterol, c confounding effect of smoking because the correlation between blood lead and di Clin Chem A smoking is related to blood cadmium (table astolic blood pressure became insignificant. 4) and Quetelet index (r = --0.15, p < 0.05). The relation between blood lead and sys 15. Del Castilho 6 nation ofcad - blood by at Therefore, a multiple regression analysis tolic blood pressure became insignificant electrotherm was carried out using blood cadmium as the after multiple regression analysis with dependent variable and Quetelet index, cig other determinants, e.g., age and Quetelet > cision with a Chim Acta 1 16. Herber RFM arette smoking, and age as independent index, included in the model. The correla variables. After this analysis, blood cad tions between blood pressure and Quetelet mium remained significantly inversely re index and age, respectively, did not change parison of atomic emisi atomization ruination oft lated to Quetelet index. This means that after exclusion of the hypertensive man body leanness is associated with high blood with the elevated blood lead level It may cadmium levels. An explanation for this therefore be concluded that blood lead is Chem 1982;J 17. Hie NH, Hu package for Graw-Hill, 1 relation is not available. probably a less important determinant of 18. Vir SC, Lovt Among the elderly men in Zutphen, no blood pressure than age and Quetelet index. elderly. Am < 19. LindemanK relation was found between blood cadmium and blood pressure. In both animals and Re f e r e n c e s ofage and se tration. J Ge humans, consistent relations between blood 1. Party HM Jr, Kopp SJ. Doei cadmium contribute to human hypertension. Sci Total Environ 20. Halsted JA, and disease. : cadmium and blood pressure have not been 1983;26:223-32. . 21. Manthey J, : i observed (1, 2, 4). That may be due to the 2. Beevers DG, Campbell BC, Goldberg A, et al. Magnesium - inverted U-shaped relation between cad Blood-cadmium in hypertensives and normotensives. Lancet 1976;2:1222-4. onary heart 22. Harman D.r mium intake and blood pressure found in 3. Beevers DG, Erskine E, Robertson M, et al. Blood- of age on i rats (38). The maximum increase in blood lead and hypertension. Lancet 1976;2:1-3. pressure was found at a cadmium intake of 4. Beevers DG, Cmickshank JK, Yeoman WB, et eL Blood-lead and cadmium in human hypertension. 10 Mg/day. This is considerably lower than J Environ Pathol Toxicol 1980;4:251-60. 1965:20:15123. Valuer M. e lead and cad Stockholm: the average intake among Americans and Europeans of 50-70 Mg/day (38). It may therefore be possible that the cadmium in 5. Klevay LM. Coronary heart disease: the zinc/ copper hypothesis. Am J Clin Nutr 1975:28:76474. 6. Voors AW. Does lithium depletion cause athero ronmental K mental Hygi 24. Friberg L, \ lead and cad take among humans is too high for showing a relation with blood pressure. Cigarette smoking was borderline signifi sclerotic heart disease? Lancet 1969;2:1337-9. 7. Elinder CG, Friberg L, Lind B, et al. Lead and cadmium levels in blood samples from the general population of Sweden. Environ Res 1983:30:233- Results of a Res 1983;30 25. Shaper AG. of alcohol a cantly related to blood lead. No dose-re 53. aged British sponse relation was present. Also, in other studies (7, 23, 24), a generally weaker rela 8. Keys A, Aravanis C, Blackburn H, et al. Epide miological studies related to coronary heart dis ease: characteristics of men aged 40-59 in seven tion was found between cigarette smoking countries. Acta Med Scand 1967;460(Suppl). and blood lead than between cigarette 9. Keys A. Coronary heart disease in seven countries. Circulation 1970;41(Suppl 1). smoking and blood cadmium. 10. Keys A. Seven countries: a multivariate analysis In the present study and in studies from of death and coronary heart disease. Cambridge. Scotland (3, 4), a positive relation was MA: Harvard University Press, 1980. 11. Abell LL, Levy BB, Brody BB, et al. A simplified found between blood lead and blood pres method for the estimation of total cholesterol in sure. Among the elderly men in Zutphen, a serum and a demonstration of its specificity. J stronger relation Was found for systolic Biol Chem 1952;195:357-66. than for diastolic blood pressure. In order 12. Katan MB, Van der Haar F. Kromhout D, et al. Standardization of serum cholesterol essays by to get insight into the stability of this re use of serum calibrators and direct addition of TEH 0412502 DUP050453554 SJbL9flU*.y~~'*___--r'XS1Z'~\z5!~' s*+**^* ssaFrgg^Rsaaaa.jasuae m ith the highestMj ) who also hadj mHg) was el4 rafter, a borderi| was found betw blood pressure, as blood lead andl came insignificaffe. lood lead and*||| same insignifi$ on analysis wit) age and Queteler odel. The correlgl sure and QueteiS' ly, did not chanp iypertensive mac ead leveL It may hat blood lead & t determinant ofl d Quetefet ind S} cadmium contribute Sci Total Environ j Goldberg A. et all uvea and nonnoten-* tson M.etal. Blood- 11976;2:1-3. $ Yeoman WB, et at) Jmati hypertension. : 9:4:251-60. disease: the line/ i Nutr 1975;28:76+- ? ttion cause athero1969;2:1337-9. 3. et al. lead and a front the general Res 1983:30:233- n H, et el. Epide. ironary heart diasd 40-59 in seven ';460(Suppl}. a seven countries. tivariate analysis ease. Cambridge. '.980. t al. A simplified al cholesterol in its specificity. J imhout D, et al. terol assays by ect addition of TRACE METALS AND CORONARY HEART DISEASE RISK 385 l.irlierniann-Burchard reagent. Clin Cheat l!)S2;2H:689-6. liursu-in M, Samaille J. Sur un dosage rapide du . linio-iierol lie aux a- et aux 0-lipoproteines du Clin Chem Acta 1960;5:609. Viin Her Haar F, Van Gent CM, Schouten FJM, ill. Methods for the estimation of high density . Jmlfiiierol, comparison between two laboratories, i |.n Chem Acta 1978;88:469-81. I i.-l Castilho P, Herber RFM. The rapid determin.i< um ofcadmium, lead, copper and zinc in whole 1,1,h k I hv atomic absorption spectrometry with 'U-cmithermal atomization. Improvements in pre, m with a peak-shape monitoring device. Anal Cliiro Acta 1977;94:269-74. Herber RFM, Pieters HJ, Elgersma JW. A com parison of inductively coupled argon plasma atomic emission spectrometry and electrothermal atomization absorption spectrometry in the deter mination of copper and zinc in serum. Free Z Anal Chem 1982;313:103-7. Vie N'H, Hull CH, Jenkins JG, et al. Statistical package for the social sciences. New York, Mc- C.raw-HiU. 1975. Vir SC, Lave AHG. Zinc and copper status of the elderly. Am J Clin Nutr 1979;32:1472-6. Undeman RD, Mervin LC, Colmore JP. Influence of age and sex on plasma and red cell zinc conceni ration. J Gerontol 1971;26:358-63. Hoisted JA, Smith JC Jr. Plasma zinc in health and disease. Lancet 1970;1:322-4. Manthey J, Steepler M. Morgenstem W, et al. Magnesium end trace metals: risk factors for cor onary heart disease? Circulation 1981;64:722-9. Harman D. The free radical theory ofaging: effect of age on serum copper levels. J Gerontol 1965;20:151-3. SI. Vahter M, ed. Assessment of human exposure to lead and cadmium through biological monitoring. Stockholm: National Swedish Institute of Envi ronmental Medicine and Department ofEnviron mental Hygiene, Karolinska Institute, 1982. -4. Friberg L, Vahter M. Assessment of exposure to lead and cadmium through biological monitoring. Results of a UNEP/WHO global study. Environ Res 1983;30:95-128. 25. Shaper AG, Pocock SJ, Walker M, et al. Effects of alcohol and smoking on blood lead in middleaged British men. Br Med J 1982;284:299-302. 26. Koo SI, Williams DA. Relationship between the nutritional status of zinc and cholesterol concen tration of serum lipoproteins in adult male rats. Am J Clin Nutr 1981;34:2376-81. 27. Koo SI, Ramlet JS. Dietary cholesterol decreases the serum level of zinc: further evidence for the positive relationship between serum zinc and high density lipoproteins. Am J Clin Nutr 1983;37:918- 23. 28. Hooper PL, Visconti L, Garry J, et al. Zinc lowers high-density lipoprotein-cholesterol levels. JAMA 1980;244:1960-1. 29. Rhoads GG, Gulbrandsen CL, Kagan A. Serum lipoproteins and coronary heart disease in a pop ulation study of Hawaii Japanese men. N Engl J Med 1976;294:293-8. 30. Cordon T, Castelli WP, Hjortland MC, et al. High density lipoprotein as a protective factor against coronary heart disease. Am J Med 1977;62:70714. 31. Miller NE, Thelle DS, Fnrde OH, et al. The Tromse Heart Study. High-density lipoprotein and coronary heart disease: a prospective casecontrol study. Lancet 1977;1:965-8. 32. Misra R, Bhambal SA, Misra NP, et al. Serum copper, ceruloplasmin and iron in ischaemic heart disease. Indian Heart J 1978;30:339-44. 33. Pooling Project Research Group. Relationship of blood pressure, serum cholesterol, smoking habit, relative weight and ECG abnormalities to inci dence of major coronary events: final report on the Pooling Project. J Chronic Dis 1978:31:201- 306. 34. Allen KGD, Klevay LM. Cholesterolemia and car diovascular abnormalities in rats caused by copper deficiency. Atherosclerosis 1978;29:81-93. 35. Fischer PWF, Giroux A, Belonje B, et al. The effect of dietary copper and zinc on cholesterol metabolism. Am J Clin Nutr 1980;33:1019-25. 36. Solomons NW. On the assessment of zinc and copper nutriture in men. Am J Clin Nutr 1979;32:856-71. 37. Klevay LM. Reply to letter by Fisher and Collins. Am J Clin Nutr 1981;34:597-8. 38. Kopp SJ, Gionek T, Petty HM Jr, et al. Cardio vascular actions of cadmium at environmental exposure levels. Science 1982;217:837-9. 39. Huang TC, Chen CP, Wefler V, et al. A stable reagent for the Liebermann Burchard reaction. Anal Chem 1961;33:1045-7. TEH 0412603 DUP050453555 .<LAD-&H-S1-X, jfilA* (L. (e i j * i 7 Additive Statistical Effects of Cadmium and Lead on Heart-Related Disease in a North Carolina Autopsy Series ANTONIE W. VOORS, M.O., Dr. P.H. Department of Preventive Medicine WILLIAM D. JOHNSON, M.S. Department of Biometry Louisiana Sate University Medical Center New Orleans, Louisiana 70112-1394 MARKS. SHUMAN, Ph.D. Department of Environmental Sciences and Engineering School of Public Health University of North Carolina Chape) Hill, North Carolina 27S14 ABSTRACT. The association of heart-related morality with tissue cadmium and lead in a study of autopsies per formed on persons who resided in a soft-water, leached-soil area of North Carolina was examined. Liver cadmium con centrations and aortic lead level were indices of these ele ments. Both cadmium and lead levels had sutistically signif icant correlations with cause of death (heart-related disease vs. non-heart-related disease, excluding cancer). Although cause of death was significantly associated with age, the association with cadmium and lead persisted after statistical adjustment for the effect of age. The combined effects of cadmium and lead provided sufficient information in an additive model to predict cause of death correcdy for 8094 of the cases, with age contributing insignificantly. These findings indicate the intimate relation of these two trace metals with increased risk of heart-related mortality, even in light of known conventional causes of such deaths. EXCESS CARDIOVASCULAR MORTALITY on the east coast of the United States (particularly the rural Coast al Plains area}: and around Glasgow in Great Britain has not yet been explained definitely.* Although the water is soft and the soil is acidic and leached in both areas, a biological link with cardiovascular disease is tenuous. Excess cadmium (Cd) and lead (Pb) have been implicated individually in car diovascular pathology,*"* but their combined effects have not been demonstrated by their accumulation In local autopsy tissues. Previously, Cd level in the liver and Pb level in the aorta have been shown to be positively associated with death resulting from cardiac disease in an autopsy pop ulation of North Carolina.1, * Presumably, liver and aorta tissues are indicators of the respective body burdens of the two trace metals as they relate to the circulatory system. It has been suggested by several investigators7*9 that the additive effects of Cd and Pb result in damage to the car diovascular system. This question is of practical importance because it is likely that these two heavy trace metals often have common sources in contemporary environments.>0*u In the present analysis these two effects are studied simul taneously to assess the magnitude of their respective roles in an additive statistical model. MATERIALS AND METHODS The population studied and methods used have been described previously.13 Briefly, 106 autopsies were done in 7 hospitals in North Carolina during an uninterrupted period. Patients who died of cancer were excluded because cancer was found to be associated with an increased varia bility of Cd and Pb levels.14 Choice of index organs. Since the heart and aorta share the arterial blood as an immediate environment, the level of Pb in the aorta was used as an index of the heart's exposure to Pb. Cadmium accumulates in the iiver and kidney; its concentration in the kidney rises until the age TEH 0412504 98 0003-9896/82/3702-0098$ 1.00 Archives of Environmental Health DUP050453556 N 30996.02 Table 1.--Heart-Related Death Causes, by Liver Cadmium and Aortic Lead Lewd, Excluding Cancer Patients (North Carolina Autopsy Study, 1971) ( Pathologic-Anatomical Diagnosis Age Cd in Liver Pb in Aoru (yr) Sex (ppm ash) (ppm ash) Myocardial infarction.(Ml) Atherosclerosis, renal infarction. Ml Congestive heart failure, obstructive lung disease Cardiac arrest, surgery for mitral valve replacement Cardiac arrest, hypertension, arteriolar nephrosis Acute coronary insufficiency Ml, fatty liver Renal failure due to arterionephroscierosis. cardiomyopathy Ml Bacteremia and myocarditis, cardiorenal atherosclerosis Atherosclerotic heart disease, cardiac arrest, aortic aneurysma Mi Coronary thrombosis, aortic aneurysma Congestive heart failure due to severe kyphoscoliosis Arteriosclerotic occlusion of the iliac arteries, diabetes msllitus, congestive heart failure Ml Pulmonary thromboembolic disease, cor pulmonale Ml, diabetes meilitvs Thrombi, myocardial insufficiency, tung tuberculosis Acute alcohol poisoning with focal fatty infiltration of myocard, fatty liver Acute coronary insufficiency Ml, pulmonary congestion, fatty liver Post-operative cholecystitis, congestive heart failure Rheumatic heart disease with valvular and coronary involvement Atherosclerotic coronary insufficiency Bronchopneumonia with focal interstitial fibrosis, recent Ml, fatty liver 76 37 76 63 63 44 54 46 59 60 67 85 61 34 74 58 67 47 70 42 58 48 71 24 47 65 F F M M F M M M M M M M M M M M M M M M F M F F M M 1720 684 60S 460 449 391 390 375 368 353 350 282 281 279 273 273 265 236 2!4 207 181 147 111 097 092 unknown 0S2 084 076 CIS 015 015 ots 091 052 107 152 130 128 104 015 223 067 212 478 465 064 315 161 540 unknown 226 of SO yr and then decreases, whereas its concentration in the liver remains more constant after 30 yr of age. Liver concentration of Cd was, therefore, chosen as a more con sistent index of relatively recent exposure.1* Tissueprocessing. The entire thoracic and abdominal aorta and the leftmost peripheral portion of the liver from each subject were cleaned, freeze-dried, and dry-ashed under low temperature to concentrate the trace metals without loss. The material was then examined by atomic absorption spectrometry for Cd and Pb content. Statistical analysis. A stepwise logistic regression analy sis was performed with the cause of death as the depend ent variable. Age and the log-transformed trace metal levels (expressed in ppm ash weight) were entered as independ ent variables. Trace metal levels less than the sensitivity threshold of the chemical analytic method were arbitrarily assigned the value of 3/S of the threshold level. The authors controlled for the amount of mineralization of the tissues by expressing trace metal concentration per ash weight rather than per dry or wet weight. RESULTS No cancer was found in S3 of 106 autopsies. Of those 83,8 did not yield aoru or liver tissues. The present analy- March/April 1982 [Vol. 37 (No. 2)1 sis, therefore, is based on the remaining 7S autopsies. The deaths of a 23-yr-old patient who died of acute bacterial endocarditis and a 64-yr-old patient who died during sur gery for an aortic valve replacement were considered nonheart-related disease deaths for this study. A list of causes of death related to heart disease is given in Table 1. Liver Cd was placed into the stepwise logistic regression equation first, followed by aorta Pb and age. With all three variables in the equation, both liver Cd (P = .0033) and aorta Pb {P * .0021) demonstrated statistically significant effects. The independent effect of age did not attain sta tistical significance at the 5% level [P * .220, Table 2).** A plot of aortic Pb by liver Cd for the heart disease deaths and other deaths (Fig. 1) suggests an additive effect on heart disease-related deaths. It should be noted that there is a concentration of aortic Pb values around 3/S of the threshold level for the chemical analysis, which is an artifact generated by the arbitrary evaluation of belowthreshold observations. The five heart disease deaths with the lowest aortic Pb values stand out in Fig. 1. Descriptions additional to the diagnoses of Table 1 are, in order of list ing for these S cases: (1) post-operative death after mitral valve replacement; (2) hypertension and adrenal cortex TEH 0412505 99 1VlI 1 l DUP050453557 TEH 0412506 rn ia m n i w ii^ u im m m ill Fig. 1. Proportion of deaths resulting from heart disease, related to hepatic Cd and aortic Pb levels; cancer deaths are excluded. (North Carolina Autopsy Study, 1971) cases, mostly with multiple diseases, are exceptions to our finding of the predictive power of an additive model pre dicting heart-related disease. Table 3 shows the data of Figure 1 in contingency table format, and Figure 2 presents the same results as a three-dimensional bargram. In general, the proportion of deaths caused by heart-related disease is the lowest when levels of both minerals are low and increases as the com bined levels of the two metals increase in a manner com patible with an additive effect for the two metals. DISCUSSION The possibility cannot be excluded that the relation ships described herein may result, in part, from consump tion of moonshine alcohol (aortic Pb) and cigarette smok ing (hepatic Cd). The presence of fatty liver, however, did not account for the relation between Pb and heart disease death ratio, nor could amount of cigarette smoking account for the relation between Cd and heart disease death ratio.1 It is more probable that in the absence of sufficient environmentally determined nutrient protection (calcium, selenium, etc), that Cd and Pb act additively in their effects on atherogenic conditions or on events that cause arteriosclerosis and cardiovascular complications. Recent animal experiments and clinical observation have demonstrated low-dose effects, both in vivo and in vitro, of Cd and Pb on the cardiac conduction system.15'15 These low-dose effects are probably not limited to the cardiac conduction tissue; Cd may also have a pressor effect1* and Pb may also have a central nervous system effect 8oth may act on cardiac tis sue.* In vivo. Perry et al.' have observed additive effects of Cd and Pb on systolic blood pressure in small rodents. Revis et al.33 reported that an additive effect of Cd and Pb occurred in pigeons, and was manifested as athero sclerosis and hypertension which were counteracted by calcium. Walker and Moses* observed a Pb mobilizing Fig. 2. Three-dimensional bargain of the proportion of deaths result ing from heart diseaje, related to hepatic Cd and aortic Pb levels. The height of the bars represents this proportional mortality and increases with higher tissue levels for either metal. If a dose-effect relationship for each of these meals exists, then some additive effect seems prob able. Cancer deaths are excluded. N represents sample site. (North Carolina Autopsy Study, 1971) adenoma; (3) severe atheromatosis, obesity, acute coronary insufficiency with old coronary thrombus; (4) severe ath eromatosis, moderately old and organizing myocardial infarcts, fatty liver; and {5} diabetes meilitus. These 5 100 Table l.-Logistic Regression of the Proportion* of Deaths Resulting from Heart Disease on Trace Metal Levels and on Age (/V * 75) (North Carolina Autopsy Study, 1971} Variable BETA P Dt intercept Cadmium in liver, ppm ash weighty Lead in aorta, ppm ash weightg Age (yr) 26.01 5.40 2.12 0.03 .0003 .0033 .0021 .2204 0.108 0.118 0.021 NOTES: This model classifies 80% of cases correctly. Cancer deaths have been excluded. *Probit transformation. Statistical Analysis System (SAS), PROC LOG 1ST.1* fO is equivalent to partial A* if the distribution would have been Gaussian. t Logjg transformation. Bivariate correlation coefficients aret age vs.Cd 0.24 [A .034);age vs. Pb, 0.05; Cd vs. Pb.0.11. j i ' i > | Archives of Environmental Health DUP050453558 Tabic 3.-Proportion of Duthi Resulting from Mean Oisease and Total Number of Otaths (Nm 75> in Categories of Cd and Pb Levels, Cancer Oeaths Excluded (North > Carolina Autopsy Study, 1971) Cd in Liver (ppm ash wt.) < 30 Pb in Aorta, onm ash weight 30-99 100-299 300* Total 100 100-199 200* Total 0.00 (8)* 0.00 (13) 0.36 (14) 0.14 (3S) 0.00 (4) 0.25 (4) 0.60 (10) 0.39 (18) 0.00 0) 0.25 (4) 0.70 (10) 0J3 (15) 1.00 0) 0.25 (4) 1.00 (2) 0.50 (> 0.07 (14) 0.12 <2S) 0.56 (36) 0.32 (7S) Sample size appears in parentheses. action of Cd in rats, and Mahaffey et al.1* observed inter actions altering the levels and toxicity of Pb and Cd. Cadmium is quite dissipated in the western industrial ized countries.11 A main source of entry into the human body is food.11 Cadmium in the soil is more available to food and fodder plants if the soil is acidic and leached,11 and if the water is soft.1* A second source of Cd is cigar ettes.1*"31 Crawford and Clayton31 have demonstrated an association between softness of drinking water and Pb levels in human bones. They attributed this association to the plumbosolvency of soft water in lead-containing pipes. Sauer3* found a partial correlation between Pb level in drinking water and mortality resulting from car diovascular-renal disease in 92 Metropolitan State Eco nomic Areas in the United States after controlling for a number of socioeconomic and environmental variables. The observed relationship between softness of drinking water and sudden heart death33,3* and the observed rela tive concentration of cardiovascular deaths in white males in the southeastern United States* point to intimate rela tions of Cd and Pb uptake with cardiovascular disease, per haps involving the cardiac neuromuscular impulse conduc tion system, the arterial tissue, the blood pressure, or a com bination of these.31 Although Cifand Pb each have separate tissue predilections, both Cd and Pb share an affinity to cardiac tissue.3*"*0 CONCLUSION The data suggest additive statistical effects of tissue Cd and tissue Pb on cause of death (heart-related disease vs. non-heart-related disease) and support the hypothesis that the excessive presence of these heavy metals in crucial tis sue components of persons autopsied in the eastern United States has a close relationship with cardiovascular mortality in this autopsy population, even in light of known conven tional causes of such deaths. Cardiovascular death in softwater, leached-soil areas needs further study. Mareh/April 1983 (Vet. 37 (No. 2)1 The authors are grateful to Georfe P. Woodward, B.5., who assisted in the atomic absorption spectrometry; to Philip N. Galla gher, Or. P. H., who prepared the tissues; to all pathologists who cooperated in tissue collection; and to John F. Meyer, III, who assisted in computing for data analysis. This research was sponsored by the Environmental Protection Agency under contract No. CPA 70-1 OS. This paper docs not necessarily represent the views or stated policy of that agency. Submitted for publication July 24,1980; revised; accepted for publication January 24,1982. Requests for reprints should be sent to: A. W. Voors, M.O.. Department of Preventive Medicine, LSU Medical Center, 1542 Tulane Avenue, New Orleans, LA 70112-1394. REFERENCES 1. Fabsitt, R., and Feinleib, M. 1980. Geographic patterns in county mortality rates from cardiovascular diseases. Am / Epidemiol 111: 315-28. 2. Amavis, R.; Hunter, W. J.; Smuts, J.G.P.M., (eds.l. 197S. Hardness of Orinhinj Water and Public Health. New York: Pergamon Press. 3. Beavers, D.G.; Campbell, S.C.; Goidbert, A.; Moore, M.R.; Hawthorne, V.M. 1976. Blood-cadmium in hypertensives and normotensives. Lancet 2: 1222-24. 4. Beavers, D.G.; Ersking, E.; Robertson, M.; Beattie, A.D.; Campbell, 8.C.; Goldberg, A.; Moore, M.R.; Hawthorne, V.M. 1976. Blood-lead and hypertension. Lancet 2: 1-3. 5. Voors, A.W., and Shuman, M.S. 1977. Liver cadmium levels in North Carolina residents who died of heart disease. Bull Environ Cant Toxicol 17: 692-96. fi. Voors, A.W.;$human, M.5.; Woodward, G.P.;Gallagher, P.N. 1973. Arterial lead levels and cardiac death: A hypothesis. Environ Health Perspect Experimental issue No. 4: 97. 7. Perry, H. M., and Erlangen, M. 1978. Pressor effect of chron ically feeding cadmium and lead together. Trace Substances Environ Health 12: 268-75. 8. Walker, H. L., and Moses, H.A. 1979. Cadmium: Hypertension induction and lead mobilization. / Nett Med Assoc 71: 118789. 9. Kopp, S.J.; Barany, M.; Erlangen, M.; Perry, E.F.; Perry, H. M. 1980. The influence of chronic low-level cadmium and/or lead TEH 0412507 101 DUP050453559 feeding on myocardial contractility related to phosphorylation of cardiac myofibrillar proteins. Toxicol Appl Pharmacol 54: 48-56. 10. StenstromT., and Vahter, M. 1975. Cadmium and lead in brewed coffee. Ifar Foeda (Stockholm) 27: 150-56. 11. Lvon, T.DJB., and Lenihan, J.M.A. 1977. Kitchen boilers as source of lead and cadmium. Lancet 1:423. 12. Lewis, S.C, and Forney, R.8. 1976. Correlation of lead and cadmium In human urine. J Forensic Set 21: ISO-53. 13. Voors, A.W.; Shuman, M.S.; Gallgher, P.N. 1975. Atheroscler osis and hypertension ip relation to some trace elements in tis sues. World Rev Hutr Diet 20: 299-326. 14. Hammer, D.i.;Coiucei, A.V.; Hasselblad, V.; Williams, W.E.; Pinkerton, C. 1973. Cadmium and lead in autopsy tissues. / Occup Mid 15: 956-63. 15. Friberg, L.;Piscator, M.; Nordberg, G. 1971. Cadmium in the Environment. Cleveland. OH: Chemical Rubber Company. 16. Statistical Analysis System Institute. 1980.545 Supplemental Library User's Guide. Cary, NC: Statistical Analysis System Institute, Inc., Box 800. 17. Kopp. S.J.; Perry, H.M., |r.; Clonek, T.; Erianger, M.;Perry, E.P.: Barany, M.; d'Agrosa, L.S. 1980. Cardiac physiologic-metabolic changes after chronic low-level heavy metal feeding. Am J Physiol 239: H22-H30. 18. Slofen, 0.1974. Environmental lead and the heart. / Mol Cell Cardiol 6:285-90. 19. Perry, H.M., |r.; Erianger, M.; Perr, E.F. 1977. Hypertension following chronic, very low dose cadmium feeding. Prac Soc Exp Biol Med 156: 173-76. 20. Valciukas, J.A.; Lilis, R.; Fischbein, A.; Selikoff, I.).; Eisinger, j.; Blumberg, W.E. 1978. Central nervous system dysfunction due to lead exposure. Science 201: 465-67. 21. Haenninen, H.; Hernberg, S.; Mantere, P.i Vesanto, R.; |alkanen, M. 1978. Physiological performance of subjects with low exposure to lead. J Occup Health 28: 683-89. 22. UliS, R.: Fischbein, A.: Diamond, S.; Anderson, H.A.; Selikoff, J.5.; Blumberg, W.E.; Eisinger. ). 1977. Lead effects among sec ondary lead smelter workers with blood lead levels below 30 micrograms/100 ml. Arch Environ Health 32: 256-66. 23. Revis, N.W.; Zinsmeister, A.R.; Bull, R. 1981. Atherosclerosis and hypertension induction by lead and cadmium ions: An effect prevented by calcium ion. Proc Had Acad Sci 78: 6494- 98. 24. Mahaffey, K.R.; Capar, S.G.; Gladen, 8.C.; Fowler, B.A. 1981. Concurrent exposure to lead, cadmium, and arsenic--effects on toxicity and metal concentrations in the rat. / Lab Clin Med 90: 463-81. 25. Fulkerson, W., and Goeller, H.E. 1973. Cadmium, the Dissi pated Element. Oak Ridge, TN: Oak Ridge National Laboratory. 26. Kiellstroem, T.; Borg, K.; Lind. B. 1978. Cadmium in feces as an estimator of daily cadmium intake in Sweden. Environ Res 15: 242-51. 27. Lagerwerff, J.V. 1972. Lead, mercury, and cadmium as envir onmental contaminants. In Micronutrienl in Agriculture, J. J. Mortvedt, P. M. Giordano, and W. L. Lindsay, eds., pp. 593-36. Madison, Wl: Soil Science Society of America, Inc. 28. Elinder, C.-G.; Stenstroem, T.; Piscator, M.; Linnman, L; lohnson, L. 1980. Water hardness in relation-to cadmium accumulation and microscopic signs of cardiovascular disease in horses. Arch Environ Health 35: 81-84. 29. Lewis, C.P.; iusco, W. |.; Coughlin. L.L.; Hartz, S. 1972. Con tribution of cigarette smoking to cadmium accumulation in man. Lancet 1:291-92. 30. Shuman, M.S.; Voors, A.W.; Gallagher, P.N. 1974. Contribu tion of cigarette smoking to cadmium accumulation in man. Butt Environ Com Toxicol 12: 570-76. 31. Oestergaard, K. 1977. Cadmium and hypertension. Lancet 1:677-78. 32. Syversen, T.L.M.; Stray, T.K.;Syversen, G.B.;Ofstad, |. 1976. Cadmium and zinc in human liver and kidney. Stand ) Clin Lab Invest 36: 251-56. 33. Crawford, M.O., and Clayton, D.G. 1973. Lead in bones and drinking water in towns with hard and soft water. Br Med / 1: 21-23. 34. Sauer, H.L 1974. Relationship between Trace Element Con tent oi the Drinking Water and Chronic Diseases, Observed Effects of Trace Metals in Drinking Water on Human Health. Presented at 16th Water Quality Conference, University of Illinois at Urbana. 35. Anderson, T.W.; leRiche, W.H.; Mackay, |.S. 1969. Sudden death and ischemic heart disease. N Engl j Med 280: 805-67. 36. Neri, L.C.; Hewitt, 0.; Mandei. |.S. 1971. Risk of sudden death in soft water areas. / Epidemiol 94:101 -04. 37. Voors, A.W. 1979. The association of trace eiements and car diovascular diseases: A selected review of positive findings in the literature. In Geochemistry of Water in Relation to Car diovascular Disease, pp. 82-90. Washington, D.C: National Academy of Sciences. 38. Berlin, M., and Uilberg, S. 1963.The fate of Cd109 in the mouse: An autoradiographic study after a single intravenous injection of Cd10,Clj. Arch Environ Health 7:686-93. 39. Amacher, D.E., and Ewing, K.L. 1975. Cadmium deposition in canine heart and major arteries following intravascular administration of cadmium chloride. Bull Environ Contam Toxicol 14: 457-64. 40. Thind, G.S., and Fischer, G.M. 1975. Cadmium and zinc dis tribution in cardiovascular and other tissues of normal and cadmium-treated dogs. Exp Mol Pathol 22: 326-34. 41. Moore, M.R.; Meredith, P.A.: Goldberg, A.; Carr, K.E.; Toner, P.G.; Lawrie, T.D.V. 1975. Cardiac effects of lead in drinking water of rats. Clin Sci MoI Med 49: 337-41. 1EM 102 Archives of Environmental Health DUP050453560 fntm / Fnvimnmentaf Studies 1980. Vol. 14. pp. 309-312 00;0-7233iSi>14U4-03U9 S04.50-'0 Jlfi-Crt (o Gordon and Breach Science Publishers Inc.. l^SO Printed in the United Kingdom CADMIUM AND LEAD LEVELS IN BLOOD AND URINE IN A SERIES OF CARDIOVASCULAR AND NORMOTENSIVE PATIENTS A. K. KHERA+, D. G. WIBBERLEYt, K. W. EDWARDS* and H. A. WALDRON! (Received February 28. 1979) The mean blood-lead, blood-cadmium, urine-l:ad and urine-cadmium levels of a series of male Caucasian patients suffering from cardiovascular disease were significantly higher at every age group than the corresponding levels in a control series of normotensive patients. INTRODUCTION Despite intensive studies, particularly over the past six years, the question as to whether the levels of lead and cadmium in the blood of hypertensive patients or those with cardiovascular disease, are higher than the corresponding levels in normal subjects remains unsolved. Even more intractable problems concerned with whether such conditions are actually caused by the presence of either of these heavy metals must await this resolution. The higher levels of cardiovascular disease in soft water districts has been cited' as a possible reason for implicating lead as a factor and the induction of hypertension in laboratory animals with additional cadmium in their diet as a reason for suggesting that cadmium is a factor.2 In the case of blood-lead levels the most important contribution has come from the Glasgow team of D.G. Beeverser a/.*who have shown a significant excess of cases with high blood-lead levels among hypertensives in the West of Scotland where tap-water lead levels are noioriously high. The same authors,*4 hSo*wever, did not detect an appreciable difference in average blood-cadmium levels between two groups of hypertensive and control patients matched for age and sex. In contrast, S. C. Glauserex a/.s{pund more than a three-fold difference between average bloodcadmium levels in 17 untreated hypertensive t Department of Pharmacy, University of Aston. Birming ham B4 7ET (U.K.). S Technology Policy Unit. University of Aston. Birmingham B4 7ET (U K.). London School of Hygiene and Tropical Medicine. Keppel Street. London EC1E 7HT(U.K.). patients and a group of ten somewhat younger normotensive controls. One of the problems is undoubtedly the consid erable difficulty of determining with accuracy and consistency the low levels of the two metais in blood. In the case of cadmium for example, in the last five years, nine different authors have claimed blood-cadmium levels in the normal population of between 0.003/s mole/1 (0.4 ng/ml)4 and 0.33/t mole/1 (37 ng'ml).7 Lauwerys* has claimed that the mean bloodcadmium levels for the general population is0.04ju mole/1 (4.1 ng/ml) and that it is now generally recognised* that the normal concentration is below 0.09/s mole/1 (10 ng/ml). PATIENTS For a period of 15 months from October 1976. 50 patients who attended the General Hospital Bir mingham because of a moderate to severe cardiac condition and'or hypertension gave, on the same day. one sample of blood and one of urine w hich were analysed for cadmium and lead. During the same time similar determinations were made for 75 patients attending the hospital with no known cardiovascular symptoms. METAL DETERMINATIONS Blood was drawn from each patient into a heparinised metal-free container and urine collected in a glass metal-free tube, lg samples were dissolved in soluene (5 mi) and diluted with toluene to 10 ml 309 lfcH 041215Ua DUP050453561 N 30996.03 310 A. K. KHERA a tl. and aliquots (25m1) pipetted into a H.G.A. 74 graphite furnace of a Perkin Elmer 360 A.A.S. modie. In the case of cadmium the mixture was dried at 60C for 30 seconds, charred at 550C for 30 sec* ends and atomised at 1900C for 10 seconds. The samples were then burnt off at 2700*C for five seconds. Controls were prepared from blood spiked with cadmium chloride to give blood* cadmium concentrations in the range of 0.0090.04ft mole/1 (or lng/ml-4 ng/ml). For lead, dry ing, charring and burn-off temperatures and tim ings were the same as for cadmium. The samples were atomised at 2100C for ten seconds. The lead standards were in the range of 0.24-2.4m mole/1. these matched groups are recorded in Table II. The urine metal levels were extremely variable and other workers1" have shown that 24 hour sam ples are necessaty to overcome diurnal variations. The blood-cadmium and blood-lead levels are recorded in Table III for four matched age groups. There was some evidence of variations in heavy metal levels with the nature of the cardiovascular disease (Figure 1) but the small numbers in certain categories make definite correlations uncertain. RESULTS For all four analytical determinations there was an aooreciable difference between the mean heavy metal value* in the cardiovascular pationrs-anrf in th< (Table I). The car diovascular patients contained small numbers of women (7) Negroid and Asian patients (7) and the normotensive patients included 14 patients less than 30 years of age. When these three groups were removed from consideration because of difficulties in matching, there remained 38 male cardiovascular patients aged over 30 and 48 matched normotensive patients. The results from HT Hypertensive (13 patients) of whom three atso had ischaemic heart disease. FIGURE 1 Disease pattern in cardiovascular patients and normotensives in relation to blood-cadmium and blood lead. TABLE 1 Cadmium and lead levels (p motc/t) in all patients Patients Cardiovascular Normotensive Blood-Cd 0.10 0.07 Metal level (ji mold) Urine-Cd Blood-Pb 0.07 0.06 2.12 1.35 Urine-Pb 0.37 0.29 V t TABLEU Cadmium and lead levels in blood and urine in Caucasian malesageti-ovet-SO' Patients Cadiovasiular Normotensive Blood-Cd mean range 0.10 0.07 0.02-0.35 0.01-0.18 Metal level (u mole.'I) Urine Blood-Pb mean range 0.07 0.05 2.17 0.43-4.00 1.4 0.58-2.2 Urine 0.34 0.27 IfcH 041Z&10 DUP050453562 CADMIUM AND LEAD Cardiovascular Normotensive Cardiovascular Normotensive Cardiovascular Normotensive Cardiovascular Normctensive TABLE III Cadmium and lead levels lor four matched age groups Age Average No. of group age patients 30-39 30-39 40-49 40-49 50-59 50-59 >60 > 60 34 35 46 45 54 51 65 67 5 9 8 9 16 16 8 17 Blood-Cd (p moie'l) mean range 0.14 0.06 0.10 0.07 0.09 0.07 0.11 0.07 0.04-0.27 0.09-0.10 0.02-0.25 0.01-0.18 0.02-0.35 0.02-0.13 0.07-0.21 0.03-0.15 Btood-Pb l{t mole.l) ' mean range 2.65 1.35 2.12 1.45 1.9S 1.30 2.17 1.40 1.16-3.96 0.42-1.74 0.68-3.23 0.58-1.95 L5O-3.04 0.5$-1.98 0.43-3.96 0.87-2.03 311 Similarly, smoking habits were not determined with sufficient accuracy for firm conclusions, although overall results showed little change with smoking and lead levels but cadmium levels ca 20 percent higher than normal in normotensive smokers compared with non-smokers. For the cardiovascular patients the group of ex-smokers had the highest blood-cadmium levels but for all three groups the disease state appeared to mask any smoking effect. DISCUSSION The results of this pilot study have proved to our satisfaction that average blood-lead and bloodcadmium levels are consistently higher in patients with a wide range of cardiovascular-disease than in matched normal subjects with no such disease symptoms. We have found that the analyses are extremely sensitive to a whole range of factors and are near the limits of sensitivity and reliability of the instrument. The results demonstrate yet again the difficulty of inter-laboratory comparisons which must await validated control work11 on the same blood samples. The patients were also far from being a cohesive group having a wide range of heart disease varying from severe myocardial infection and ischaemic heart disease to moderate hypertention. Most were also receiving a range of drug therapies and in view of the results found with varying disease states these should be standardised in future trials. Beevers et al.* have suggested that chronic exposure to unsatisfactory levels of lead in drink ing water can lead to the development of hyperten sion. Voors and Shuman1* have similarly concluded, on the basis of high liver-cadmium levels in North Carolina residents who died of heart disease, that cadmium has a toxic effect on the cardiac conduc tion system. Our work on placental lead levels did not enable us to clearly implicate lead as a causa tive factor in stillbirths and neonatal deaths13 and in our opinion, a similar dilemma exists in the case of cardiovascular disease. Environmental levels of cadmium and lead could be inducing cardiovascu lar disease in some patients or such disease may result in abnormal metabolism and consequential higher blood-cadmium and blood-lead levels. REFERENCES 1. M. R. Moore era/., `'Environmental lead pollution in an urban soft water area'* Br. Med. J. a. 491 (1972). 2. HA. Schraederef aL." Hypertension in rats from injection of cadmium" Arch. Environ. Health 13. 788 (1966). 3. D. C. Beevers of., "Blood-lead and hypertension" Lan cet II. 1 (1976). 4. D. C. Beeverseto/., "Blood-leadcadmium in hypertensives normotensives" ibid, ii, 1222 (19*6). 5. S. C. Glauxer era/.. "Blood-cadmium levels in normoten sive and untreated hypertensive humans" ibid I. 717 (1976). 6. A Ulander and 0 Axelson, "Measurement of bloodcadmium levels" ibid i, 682 (1974). 7. A. V. Coiucci l of.. "Pollution burdens and biological response" Arch. Environ. Health 27. 152 (1973). 8. R. R. Lauwerys a/.. "The Relationship between cad mium exposure or body burden and the concentration of cadmium in blood and' urine in man" Ini. Arch. Occup. Environ. Health 36. 275 (1976). 9. R. R. Lauwerys */ a/.. "Placental transfer of lead, mer cury, cadmium and carbon monoxide in women" part 1. Environ. Ret. IS. 278 (1978). 10. S. C. Lewis et al.. "Correlation of lead and cadmium in human urine" /. Forensic ScL 21. 150 (1976). 11. R. R. Lauwerys et al.. " Imercomparison programme on the analysis of lead, cadmium and mercury in biological I fcli U41ZD11 DUP050453563 312 A. K. KHERA tt at. fluids." In Pw. Intern. Symp. Recent advancer in the assessment ofthe health effects ofenvironmental pollution. Paris 1974 Luxembourg Comm. Europ. Comm, p. 2185 (197?). 12. A. W. Voors and M. S. Shuman." Liver-cadmium levels in North Carolina residents who died of heart disease" Bull Envir. Comem and Toxicology 17. 692 (1977). 13. O. C. Wibberley et al.. "Lead levels in human placentae from normal and malformed births" J. Med. Genetics 14. 339 (1977). Intern. I o o io -t : '1 i i. f INTI The reprt lutio publ In n lion.* lead leve repc neai dati alor con alot leai bot the also EX1 1. 5 * Sa: att dii in; of K: f sit TEH 0412512 DUP050453564