Tobacco Documents Online

672 I am Dr. Carl C. Seltzer, Honcrary Researd: Associate at the Peabcdv Museum, Harvard Dniversity and Professor of Nutrition, Tufts G.^.ivers°tv . I was formerly Senior Research Associate at the Harvard University Schoo'_ of Public Health. My work in the smoking and health area is ey.ter.si•ie and I::_^e cublished over thirt7-five (35) articles since 1964 on the s*~`_'ec:, :nany deali.^.g with the relationship between smoking and coror.ay heart disease (C-D). I was a consultant to the surgeon General's Advisory Coaaitt=_e o.. _-ckinc and Health (ccr.tributed a section to the 1964 Report) and aa presentl•; a:e11ow of the Council onEpideni- ology of the American Heart Association. My investigative work in the heart disease field includes participation in studies involving the Fraaingham Heart Study, The American Medical Association, the •ieterans :,d-ir.istraticrr and the 1Caiser-Permanente Foundation. . - H.R. 4957 proposes that the following la'--el statenent aooear cn cigarette packages: "Warning: Cigarette smokir.g is a major cause of Heart Disease."This proposal is apparently based on the clain, put _brth as aCongressional "finding" in the first part of the bill, that one third o° deaths from CHD are "attributable to smoking." , These bald statements are not based on consistent, valid, de^.cns`rable scientific evidence, and are without established ;roof. ?.pparent:y, *_=e,v are based on the views of a succession of Surgeon Generals subsequent to the 1964 Report, on the statements of the American Heart Associat'-en, the Framincham Heart Study and others. ir Page 2 The primary basis : statements from the 1981 Sur Study. The 1981 Report statc many epidemiological criteria dose related, and reversible. "Causal inferences are suppor consistent, demonstrated pros and can be explained by the cardiovascular apparatus. F: s-,oking have only half the ri data have . . . tended to inc reversible, triggering effect Accordingly, the mai between cigarette smoking and that the association is (1) independent of associated risk (6) In addition, the associat cigarette smoking on the cardi turn. .. .... . Claim 1. Is the relationship s that in the D.S. the But, this is , not Professor Hutchison

~iversity Sc^col cf Public .ter.s.ve ar.d I '-a,-e cub:ished _.,-, many deali.^.g with the -.se (CcD). .1 was a consultant c ar.d Health (ccr.tributed a ~ of the Council on Enidemi- :gative work in the heart 1•ring the c'rasi-cham Heart -.s Administraticr.-and the t, ~. -. . . - . - - a1 statement appear on .s a major cause of Heart : clain, put forth as a' _, that one third of deaths i.. :stent, valid, demc^strable :cf. Apparently, tF.ey are aLs subsequent to the 1964 ;ociat:on, the rramir.cham Page 2 673 The primary basis for these views can best be characterized by statements from the 1981 Surgeon General's Report and by the Frani^aham Heart Study. The 1981 Report stated: "The effect of smoking on CvC risk fulfills many epidemiological criteria for a causal asseciation: pcwerful, ir.de_=_r.de.^.t, dose related, and reversible." Kennel of the :raninc2ar.. Heart St.dy stated: "Causal inferences are supported by the fact that the relatior.s:i: is strcnc, consistent, demonstrated prospectively, ir.de=enden`_ of associated risk factcrs, and can be explained by the kncwn effects of -cigarette smoking or, the cardiovascular apparatus. Finally, it has been shown that those *:ic am:it smoking have only half the risk of those who continue to smc:<e. 'pide:^'_clcqic i data have . . . tended to indicate an independent, trar.sient, .*.on^:mulat::e, reversible, triggering effect of cigarette smoking." (J?.1A, 1981) Accordingly, the main evidence advanced for a ca_sa'_ relaticnship between cigarette smoking and Ch'D is essentially eoidemiciogic, tc the effect that the association is (1) strong, (2) consistent, (3) _c=e-related, (4) independent of associated risk factors, and (5) reversible b;% stcppinc smcking. (6) In addition, the association is said to be explained b;• knewz effects of cigarette smoking on the cardiovascular system. Let us exar..ine these __a_r.s ..n turn. Claim 1. Is the relationship strong? The 1964 Surgeon General's °eport stated that in the U.S. the mortality ratio of smokers to nonsmokers is 1.. Hut, this is not a strong relatior.ship, according to Professor Hutchison of the Harvard School of ?ublic vea'-ti. (?ull xY

674 • Page 3 Acad Med, 1968). This independent cbserver feels that a mortality ratio of two or less is actually indicative of a weak relationship which may possibly be explained by confounding factors. Can the relationship be characterized as "strong" when the 1979 Sur;ecn General's Report, although discussi::: smo'<ir.g an d health studies throughout the world, limits its cor.cl,_sion to "r..e n and wcc,ern i n the United States," and Kanael of the Frami ng:^.am Study talks about the excess C.'3D death rates of smokers "in rost of Western societies"? (AHJ, 1981) If the relationship between smokers and CF:^ is "strong," why is it not present in all Western societies oriin non-Western societies? - Claim 2. Is the relationship consistent? Since ;c significant statistical associations between smoking and CFD have •nee.^n reported in Finland, the Netherlands, Italy, Greece, Yugeslavia, ,Tazan and ?•;erto Rico, it is clearly incorrect to claim that the relationship is ccnsistent. The relationship is not even consistent irn the 'raninoham Study. _.7 appeared to be consistent in initial analyses, but when star.dard risk factors and such psychological factors as :ype A personality were controlled, it was found that "cigarettes smoked per day" were not predictive of CHD and myocardial infarction incidence in men, or of CFID and angina pectoris incidence in •.rosen. (:_n J_oidemiol, 1980) Can the relati Surgeon Genera- association of clear"? The Fr to angina pectc Is the relatic declines with that the benef 65 as regards '. Furthermore, t: excess incider. data made avai Can the relati Heart Study re women? C1aim 3. Is the relatic rising gradie smoking. (JAt reversed. (J: not show a cc cigarette smo_ category, a c

rver feels that a mortality ratio of a weak relationship which may actors. . as "strcng" -whe- the 1979 Sur_ecn ? smokir.g an d heal:h st_dies lusion to "cern and wcmern i n the smiagham Study talks about the -n most of Western societies"? een smokers ar.d CH* is "strong," z societies or in non-aestern :ce nc significant .,tatistica_ :avs been reported in :'-r.la--d, the a. Japan and ?uerto Rico, it is alationship is ccnsister.t. int in the 'ranincham Study. :,Z analyses, but when standard risk - :zs as Type A persona'atj were ettes smoked per day" were not rction incidence in men, or of CED .1. (:.m J Eaidemiol, 1980) Page 4 675 Can the relationship be characterized as co-:sistent when the 1979 Surgeon General's Report concluded that "the predictive ris)c factor association of smoking with the incidence of angir.a pectcris is not clear"? The Framingham Heart Study also reports that "the relations::i: to angina pectoris is modest, if it ex:s=s at all." (..` 1981) Is the relationship consistent when t•r-e relative risk progress`_vely declines with increasing age? Seitzer de:a^•strated and Kannel noted that the benefits of quitting smoking do not extend beyond the age of 65 as regards heart attacks. (Am J Med Sci, 1978; =, 1981) ~ Furthermore, the 20 year follow-up data of the Frar.,i-:gaam study show no excess incidence of CHD in men or women 55 yearsan.d over. (Framingham data made available to Seltzer as consultant.) - Car. the relationship be characterized as ccr-siste-^-t when the Framinc:;am Heart Study reports an absence of a relationshi; c= s-oking and c:9 in women? Claim 3. Is the relationship dose-related? -C:iD rates do not consistently shcw a rising gradient in relation to an increased ancunt of cigarette smoking. (JAMA, 1968) In some studies, the gradient is actually reversed. (JAMA, 1968) I found that data from the Franingham Study do not show a consistent rise in CHD risk with increasing amounts of cigarette smoking because of the "heterogenous" character of its "none" category, a deceptive procedure. The "none" category :n Framingham

676 Page 5 consists of a combination of never smckers, ex-smokers, pipe ar:d cigar smokers. A Framingham Study report confirmed that "there is no distinct dose-response relation [of CHO and death rates] with increasing amounts of cigarette snoki:g" over ar. 18-year surveillance period. (Lancet, 1974) , . _ .. , There is no significant dose-response rela=ions4ip betweez smoking and CHD in studies of Finland, the Net:er'-a1ds, Greece, Italy, Yugoslavia and Japan. (Circulation, 1970) , . No claims have been made lately about the alleged a=_sociation of C3D with duration of cigarette smoking. In the ccr,%!:i-.ed albaay and Framingham Heart Studies•(NEJMl, 1962), no association ::as fo•:-d between ~ duration of heavy cigarette smoking and risk of myocardial in`arction. -_ In a study of Canadian veterans, no consistent gradiect occ_rred between CaD mortality and years of cigarette smokir.g. 1968) In Kahn's analysis of Dorn's U.S. veterans data, there was also no significant relationship between duration of cigarette smoking and risk of death from C'(D. (JAMA, 1968) ,_ . Claim 4. Is the relationship independent of associated risk factcrs? Smoking has been found to be associated with C3D independently of associated risk factors in some studies when only cholesterol and blood pressure were considered. But this generally has -ct ~zeen the case when ot:er Page 6 risk fact psycholog Claim 5. Does stop studies f ex-smoker article c claiming to one-ha CHD rates article), were far smokers" letter tc Additionz smoking : first is ex-smoke, characte. is false would be others h represen- data sho

-smokers, pipe and cigar : that 'there is r.o death rates] with ar. 18-year surveillance ahip betxeea smoking and reece,.Italy, Y~goslavia -eged asseciaticn cf C3D ccm:i-ed Albany and ciation was found between 3f myocardial infarction. er.t gradier.t occ'_rred xkir.g. - (J2~a, 1363) In a, there was also no :igarette smoking and risk ? risk factcrs? Smoking apendently of associated ~terol and blood pressure teen the case when ot.`.er Page 6 677 risk factors were investigated, such as ?`yce A personality, a::d other psychological factors (as seen above in the Franingham Stud;•). Claim 5. Does stopping smoking reduce the risk of C:'.D? This claim is .'•.ased or. studies from some Western countries vni h retort that C:.., rates of ex-smokers are lower than those of cor.t-r.~ir.y smokers. The 1?-4 La.^.cet article on the Framinghaa study is the r.cst widely quoted source for claiming that the rate of CYD accng those who stop smoking is reduced to one-half that of continuing smokers. When I was provided with the CHD rates of Framingham "never smokers" (which had :een omitted in the article), however, I was able tc, show that the C3: rates of ex-smokers were far below those of "never" smokers," while the rates of "never smokers" were not far below those of centinuinc smokers. (Seltzer letter to Lancet, 1977) Additional evidence about the fallacy of the clain that stcnping smoking reduces the rate of CHD comes from two other sources. :he first is a study challenging the assumption by e=idemiologists that ex-smokers and- continuing smokers are alike in all pertinent characteristics, except for their smoking habits. if this ass::mptiorn is false, which I believe it to be, then comparisons of the two groups would be biased as to their eventual CHD rates. Friedman, Seltzer, and others have shown that ex-smokers, when they still smoked, were not a representative sample of the population of continuing smokers. The data showed, in fact, that ex-smokers were at lcwer heart disease risk

Y._ 680 Page 9 discusses those statements of the 1979 Surgeon Ger.eral's Re_rcrt which illustrate the absence of firm evidence of mechanisms by wh+cz src.cing affects CHD, such as: (1) little is known about the mecha^isr.s by which smoking enhances athercgenesis, (2, further research is r.:eded to show the mechanisms of sudden death and i__ :rec•.usor states, i3) the . data on the effect of smoking on blood iipids are not very uni'crn, (4) the association between cholesterol and smoking is .mir.imized, (5) the acute and transient effects of smckinc are to increase heart rate and blood pressure to a minor dearee, (6) smoking is not a risk factor for hypertension, (7) the association of smoking with necr^:sy manifestations of atherosclerosis, shown orincipallry by Strong's grc.:p in New Orleans, now has been contradicted by a seud;• by Ho:ne, Strong and others who found that smoking did r.ot show a signi_°icar.t association with coronary artery lesions. (Artericsclerosis, 1931) More evidence comes from other sources. An editorial in the August 1980 British Medical Journal states that the mec^anism by which smokir.g _. affects coronary heart disease is unknown. The American Heart Association Heartbook states: "The mechanisms by which cigarette smoking is associated with higher rates of coronary heart disease are ..not yet fully understood." The 1981 Surgeon General's Report states: "Estimation of the impact of varying cigarettes on coronarl heart disease risk is difficult, because the exact etiologic agent(s) have not been identified." Thus, the bi11's proposal to label c'_garette packages with levels of tar, nicotine, and carbon monoxide is not Page 10 justified has not b etiologic This is al radical ch years (fil oa cardio• consistent carbon morn rates, and etiologic . Since most important : smoking ern evidence. the athero_ the heart smoking is lack of de: It is claa: the Congressional f: in this country arc

on General's Re`ort which echanisms by which smoking about the Wechanisr.s bv :ther research is r.e=ded to crecirsor states, (3) the 3 are ^ot very unifcrn, (4) ing is.minimized, (5) the .,, increase heart rate and ; is not a risk factor for king with necropsy icipally.b_v Strong's group a study_ by Ho;ne, Stror.g _ show a significant rter•: sclerosis, 1931) editorial in the August mechanism by which smo:cir.g . The -american Heart _sms by which cigarette ,oronary heart disease are : Geperal's.Report _states: attes on coronary heart : etiologic agent(s) have oosal to label cigarette carbon mor.oxide is not Page 10 681 justified on scientific grounds. The Surgeon General adaits ttat it has not been clearly demonstrated that these factors are e etiologic agents in the causation of increased CHD in smokers. This is also shown by the conclusizn cf t_e 1:61 Report that despite radical changes in the composition of cigarettes over the last 10 to 15 years (filter, tar, nicotine), there was no clearly demonstrated effect on cardiovascular disease. This surprisir.a result is clearly consistent with the proposition that amounts of tar, niccti::e, and carbon monoxide in cigarette smoke have no demonstrable effects on C: rates, and that these components of tobaccc smcke are r.ot proven s etiologic agents. Since most cardiologists believe that atherosclerosis is the most important factor in coronary heart disease development, they claim that smoking enhances atherosclerosis, despite the lack of definitive evidence. On the other hand, the Framingham Heart Study has disavowed the atherosclerotic argument and claimed that the effect of smoking on the heart is acute (not progressive), and disappears promptly when smoking is stopped. These are contradictory claims and illustrate the lack of definitive knowledge in this field. It is clear from the above that extensive research data do not support the Congressional finding in H.R. 4957 that a major proportion of the Cr2 deaths in this country are attributable to smoking. In turn, the bill's proposed

682 ® Sigaed Date S" I am Shel in pathology, curr Columbia Universit N.Y., and Universi Los Angeles, Calif Lenox Hill Hospita Hygiene Medical Re Stout Society of S New England Pathol Since 193 engaged in medical endocrine and gast: about 10 percent d, on pancreatic canc: Diagnostic Gynecolc boards of five othc publication list a: For the pe Director, Council : agency for biomedic funded by tobacco r. in 1982 is 7 millic Scientific Advisor; 35-M7 0-82-44 ` Page 11 warning, "Cigarette smoking is a major cause of Heart Disease," is not scientifically valid. _ In every instance, an examination of _-_ claims made against cigarette smoking and CHD shows that they are either ar_n;, incnnsistent, selective, unsubstantiated, or, in many respects, contrary te the statements of the Surgeon General, the American Heart Association and t`e ::aninc::&~ Heart St'.idy.