TMLE 3.Pudden death from coronary Project. Amcriua aart hwciation. 1970. U.SA. (88). TABLE 4.-Coro7uzry heart disease (Risk rstlos-actus1 number of CHD [S?d = Smokers NS = Nonsmokcn Dorh 2.282 nula Dctdkd 10 24a mro- NS . . . . . . . . . . . . 1.00 (62) cc al.. F~minghlm. mcdkd CSKdM AU smokers . _. .2.36(191) 1964. 3W2 yean usminn- infrrc- <20 . . . . . . . . . .I.98 (44) . USA. of Lge. tian snd tions and 20 . . . . . . ..__.._ 2.06 (64) (84). 1,913 m&s folka-up. CHD >20 . . . . . .._.. 3.04 (83) An=w. dathx. X9-66 ran of sgc. 8Lunlcr 1.329 CHD- Int.e?vicv 4 46 CHD NS . . . . . . . . . . . . 1.00 (2) ct& free mrk mnd uamio- 1666. anDlo7rn of stion 4th =`; ;ig;;t.ty: 2.92 (6) 4 USA. PcoulQ Gas clinic < 6 vbca..... (177) _ COUlD=V folbr-up. l&19 cigarette.3.67 (8) 40-59 ,an of *se. > 6 PIPCS..... 4 96 mak. Male* 92 femak 40-58 CED in- NS . . . . . . . . . . ..l.OO (1) duding EX . . . . . -. . . ...6.63 (10) deaths, Clg.rrttea . . . . .6.20 (36) msirm. and Femakm ,nyocudIII NS . . . _. . . . . . . . 1.0 (21) idsrctlon.. EX . . . . . . . ._ . . , 0.80 (1) Cfp.RtrcS . . ...1.02 (X0 26 heart disease related to smoking of death shorn fn Darentbesca) Ci~rcttesldru Clean, Dive Comment Never smoke3 ........... 1.00 (16) 1.00 (16) See t&k 1 for der.crlDthn of 210 ............... z.. .. .t.90 (23) 1.86 (13) Pooling Project. 20 . .:.: ...... . ....... .x.90 (60) z-20 .................... 5.36 (40 morbidity US related to smoking msoifstrtions shown in psrcothesea)' EX = Exxsmokcnl PROSPECTIVE STUDIES-Continued Pipes, clears Age variation cornmenb Dats include CHD deaths. only on m&s 4049 remn of we md f rrr of CHD on entry. NS includa, PIPeE. c&w% and cx-smoken. NS include cr-smokera. Includn all CHD. Mole&ntinued M.Zk. Recxrmination 80 and WdT N-59 Of pdenb 1.00 (7) SM . . ..l.SO(Z) WOJ spread l.z?(ll) 60 and over cwer 1y,.a-yar 1.96(23) SY . ...0.86(6) Femah-contlnued period. but dsta are re- 1.00(47) DOdZd in 1.11 (6) tcrma of 0.42 (2) 4-year in& dence r.tes. Actual number of CHD inci- dents derived from dsta on incidmce and tot4 In smok- ha CLsl. 27 TABLE 4.-Coronary heart disease (Rhk rrtIm--sctrul number of CHD (SM = Smokcn NS = Nonamatcm PROSPECTIVE STUDIES eAUthOr. *-. Number knd Lkt, Polkm- Number ot camtrr. t7De Of eolkctlcm UP incidenti Clgl3rtttM/&T rdcrarcc pOpllhti0n la- Jenkins. 3,lfJZ mda Initid 4% 104 m,o- NS __ . . . . ..- . ..I.00 (21) et 8L. 19-69 ,cIn m&cl1 UldLl EX ._.___ . . . . . . 1.47 (16) 1968. of sge at usmin8- hhrrtbnm. Cumnt _ ., ._. .2.78 (68) U.S.A. mtl7. uon rnd o-lS/daT ._-._. t1.as (46) (901. fol!er-up >I6 _ -.*-.. . . . . . WI6 (69) br -t aunlnr- uon8- Kmoel. 6.127 mslea Y&d l2 226 myc- kr,c,a rnforc:ifm et al.. and females crur.fnstbn UIdW M&8 1968. to-59 yt*ra end follow- Snflrc- NS . . .,* . . . . . . . 1.00 (21) U.S.A. of *pz. UP- tiorl~. AU slid . . . . . . . . 1.61(16a) (94). a80 CHD. aesvy SM . . ..I.86 (69) Ri.k of CHD (owrdI) Mdu NS .._____ _ . .._ 1.00 (61) l-10 . . . . . . . . . . . I.24 (25) II-20 . . . . . . ..__ 1.80 (90) >20 . . . . .._..._ 2.41 (76) Shapiro 110.000 male Bwelioc mcd- a Tot&l Mok. et IL. sod femak kd ioter- un*pecl- NS . . . _. . . . __ .I.OO 1969. cnr0lk3 ricr md tied. At1 cul~cnt . . . .2.14 U.S.A. of Ecmhh rxamin*ticm cigsrettea (p20 . . , . . . . . . .2.33 Greater I >40 . . . . . . . . . ..B.JC New York (HIP) a-4 le.= of .gc. K,. 9,186 m&s Iotervie~ 5 65 dcatha. NS. EX 1970 In 6 coun- and r-- 80 nwocIIr- Ytlp+ (SM <20) I. .1.00(306) tria lo-59 Isr foibw- dial in- AU current ah& Icall of np examins- f.rction% PinLnd (>20) . . . . . ..1.31(103) .*e at entrY. lion br 128 8ngir-m ItAb Lxrl NdhU- DCCti?iU. Dhysici*nr 156 other lands - C- (III). t428 totaL 28 mwbidity as related to smoking (cont.) mmnifahtioos show in p~renthmcs)' EX = Ex-emokersl PROSPECTIVE STUDIES--Continued (~o.ol) fD20 . . . . . . . . . . . 1.17 (13) myocardial infarctions. DtUIZl 13,145 male Dsta only up to 14 Total un- et rl., p*tienb in on *cur ape&&d. 1970 periodic health incidents U.S.A. examination extracted (55). clinics. from clinic recorda. POOtiIlp 7.427 while Medical 10 638 Prujtct. mrla 30-59 examination Includea Never Fmokcd .l.OO (63) American ,can of xnd follow- fatal and 20 . . . . . . . . . ..3.28(164) 1970. infarction USA. nnd audden car). death. Pmxl CL IL. 1.939 Wntcrn Screening lP63, Ekctric co. crsmination COTOMW TJmsa rmk workera and CaBCS 167) (li.4). prticipatinp history. NS .__ .__.... . 23 In. prmpcc- l- 7 _ . . . . . . . 2 tire atw3Y 8-12 . . . . . . . . _ 9 for 4% ,C.rn. la-17 . . . . . . . . - 6 18-22 .----- 41 .._..... - 23-27 , . . . . . . . a >28 ___ . . . . . . . 9 30 ,/. ,,... y. -I .`J. ~- ,J ., -1 . . , : . . . ,, , ,, -.. :' ,. . ._ .:,, .-; _ , :s. ! ,`.' .' ; 1. i, :i:;.-, . . :* ;-,2 . I,,.i: , ,.Z' `: ,_ ,>' f `., ; -' ' .; .,. , : ,I ..- ` tios JO-19 .( ; bo-bf 60-59 t Includea ; SM 1.00(26) 1.00(126) 1.00(167, NS. EX. and ." _._ I,. tHlgb <20 cigarettes/ . --: . , day. .I SM 2.17(10) 0.90 (31) 1.41 (63) :>ZOei`~a- . . . .`.d ..I )*,-J.. .iz- ._, ; , r&es/dry. ,' I :.:.* ;`.. ,_I.' -: ,`., : Include. all ' . CHD blrt __ I .._. . . .,_. _.. :.I :: :,:-* I ;-car. ,., -; `.. : - . . r `- .' czcludr. ._. 1.26(60 ' = . i . . -.. 1. . . -* . . . ..1..--- :. . .-, `. 1; ,.,., -; , =...; ,. I : ,,, +i : 1;. -*. __ .:. studies have shown an increased risk of this manifestation among smokers, others have not (see table 5). From these longitudinal studies, it has become increasingly clear that cigarette smoking is one of several risk factors for CHD and that it exerts both an independent effect and an effect in conjunc- tion with the other risk factors. The basic concept may be ex- pressed as follows: The more risk factors a given individual has, the greater the chance of his developing CHD. The importance of the constellation of coronary risk factors which include cigarette smoking, high blood pressure, and high serum cholesterol in pre- dicting the risk for CHD is illustrated in figures 1 through 3. Other risk factors are included in certain of these figures and are dis- cussed below. Knowledge of the effects of cigarette smoke on the cardiovascu- lar system has developed concurrently with the knowledge derived from the epidemiological studies. Nicotine, as well as cigarette smoke, has been shown to increase heart rate, stroke volume, and blood pressure, all most probably secondary to the promotion of catecholamine reIease from the adrenal gland and other chromaffin tissue. This release of cateeholamines is also considered to be the cause of the rise in serum free fatty acids observed upon the in- halation of cigarette smoke. Studies concerning the effect of nico- tine on cardiac rhythm have also suggested that smoking might contribute to sudden death from ventricular fibrillation. In addition, research efforts have also been directed toward the effects of smoking on blood clotting and thrombosis; since many _ cases of sudden death and myocardial infarction are associated with thrombosis in a diseased coronary artery branch. Cigarette smoking may be associated with increased plateIet aggregation in Vitro and thus might play a role in the development of such throm- bi or platelet plugs in viva. , Other mechanisms have been investigated. Because cigarette smoking has been shown in some studies to be related to the prev- alence of angina pectoris as well as to the incidence of myocardial infarction, it has been suggested that smoking enhances the de- velopment of atherosclerotic lesions. Autopsy and experimental studies have shown that cigarette smoking plays a role in athero- genesis. The administration of nicotine has been observed to in- crease the severity of cholesterol-induced atherosclerotic lesions in exper+nental animals. Attention is presently being given to carbon monoxide, which is present in cigarette smoke in such concentra- tions as to cause carboxyhemoglobin concentrations in the blood of smokers as high as 10 percent. Based on research in animals, it is reasonable to conclude that the atherosclerotic process may be enhanced, in part, by the relative arterial hypoxemia in cigarette 32 ;f .' ,;p, - -. $,,:" : 1(' -: : TABLE S.-Coronary heart disease morbidity aa related to smoking -angina pectoris+woapective atudiee .+, . (Rl#k +.l~-utu~l number of CHD mrnlfntatlon# shown In parenthesn)~ `,,. ;, ,;, -. ! [SY = Smokera NS z Nonsmokers] Cll?#M -. Clmrcttedday wd DiDC. :.-, : .I `., : AgevarIatIon ,. `t `.I " Commcntr ?`. . . NS Include ez- Malu Fmalsc M&b Male* t (PCO.01) >4o ..*,..a . . . . . . . - 4.12 - ' Unb c4h-d~ ~e~ltlcd. dlwultla between the total number of ,A , -ntf-Ulom *ad the bum OK the lndlrldurl wnoklns caegx,ries .re due to the acluslon of citbw aculonal, mLcelLne0ur. mlred. or ex-rmokrtn. smokers caused by the increased carboxyhemoglobin level. With respect to the acute event of myocardial infarction, atten- tion has been focused on the role of nicotine. Nicotine stimulates the myocardium, increasing its oxygen demand. Other experiments have demonstrated that in the face of diminished coronary flow `(due to partial occlusion from severe atherosclerosis in man or to partial mechanical obstruction in the animal), nicotine does not lead to an increase in coronary blood flow as seen in the normal individual. These effects exaggerate the oxygen deficit when the supply of oxygen has already been decreased by the presence of carboxyhemoglobin. Thus, a marked imbalance between oxygen demand (which has been increased) and oxygen supply (which has been decreased) is created by the inhalation of CO and nico- tine. This imbalance may contribute to acute coronary insufficiency and myocardial infarction. EPIDEMIOLOGICAL STUDIES Numerous epidemiological studies, bath retrospective. and pros- pective, have been carried out in various countries in order to iden- tify the risk factors associated with the development of coronary heart disease (CHD) . Many of these studies have included smok- ing as one of the variables investigated. Tables 2 to 4 present the major findings. CORONARY HEART DISEASE MORTALITY Table 2 lists the various prospective studies concerning the rela- tion of CHD mortality and smoking. These studies demonstrate the dose-related effect of cigarette smoking on the risk of deveIoping CHD. For example, the Dorn Study of U.S. Veterans as reported by Kahn (93) reveals progressively increasing mortality ratios, from 1.39 for those smoking 1 to 9 cigarettes per day to 2.00 for those smoking more than 39 cigarettes per day: Although the data are not detailed in the accompanying tables, several of these stud- ies have also shown that increased rates of CHD mortality are associated with increased cigarette dosage, as measured by the degree of inhalation and the age at which smoking began. Although not as striking, the data for females reveal the same trends. In most studies, the smokers' increased risk of dying from CHD appears to be limited mainly to those who smoke cigarettes. Some studies that have investigated other forms of smoking have shown much smaller increases in risk for pipe and cigar smokers when compared to nonsmokers. However, the recent study by Shapiro, et al. (172) of a large population enrolIed in the Health Insurance Plan (HIP) of New York City showed a significantly increased risk for the development of myocardial infarction and rapidly fatal myocardial infarction for a group consisting of both pipe and cigar smokers. Table 3 details the findings of the American Heart Association Pooling Project on sudden death. The Pooling Project, a national cooperative project of the AHA Council on Epidemiology, is de- scribed in table 1 (88). Cigarette smokers in the 30 to 59 Year age group incurred a risk of sudden death from CHD substantially greater than that of nonsmokers. Pipe and cigar smokers were observed to show a risk slightly greater than that of nonsmokers -- (table 3). The relative risk of CHD mortality is greatest among cigarette smokers (as well as among those with other risk factors) in the younger age groups and decreases among the elderly. In table 2, Hammond and Horn found that for those smoking more than one pack per day, the risk is 2.51 in the 50 to 54 year age group and 1.56 in the 65 to 69 year age group. Although the relative risk for CHD among smokers decreases in the older age groups, the actual number of excess deaths among smokers continues to climb since the differences in death rates between smokers and nonsmok- -ers continue to rise. CORONARY HEART DISEASE MORBLLIITY Tables 4 and 5 list the prospective studies carried on in a num- ber of countries to identify the risk of CHD morbidity incurred by smoking. Here, CHD morbidity includes myocardial infarction as well as angina pectoris. Certain studies, notably those of Doyle, et al. (541, Keys, et al. (III), and Taylor, et al. (185) include a number of CHD deaths in their data that could not be separated out .using the information provided in their respective reports. As noted in the discussion on CHD mortality, the CHD risk ratio increases significantly as the number of cigarettes smoked perday increases. Similarly, the HIP data of Shapiro, et al. (17.2) sha$ that the elevated morbidity ratios declined with increasing age as has been shown for mortality ratios. A recent monograph edited by Keys (111) dealt with the &year CHD incidence in males age 40 to 59 from seven countries. As summarized in table 4, cigarette smoking &as found to be associ- ated with an increased incidence of CHD in the U.S. railroad worker population, 2,571 individuals (183). None of the differences in ratio between smokers and nonsmokers was statistically sjmifi- e-ant for the 13 other population samples which varied in size from 505 m 962 individuals, from the five other countries. (Smoking was not considered in the two Japanese populations.) men more cases . . 35 become available to provide greater statistical stability to the rates, this intercultural comparison should prove illuminating. The results of those studies which have separated out angina pectoris as a manifestation of CHD are presented in table 5. Doyle, et al. (54) found no relationship between this manifestation of CHD and cigarette smoking. Both Jenkins, et al. (90) and Kannel, et al. (94) observed increased risk ratios among male cigarette smokers although these. differences were not statistically signifi- cant. More recently, Shapiro; et al. (172) found a significantly increased risk for angina among their male cigarette smokers as well as increasing risk ratios with increasing dosage among both males and females, particularly in the .younger age groups. A variety of hypothetical explanations have been advanced to account for this seeming contradiction- Among these are the relatively small number of cases, the difficulties associated with the definitive diagnosis of the syndrome, and differences in the methods of clas- sifying those cases of angina pectoris which are followed bv mvo- cardial infarction. RETROSPECTIVE STUDIES Table A6 presents data from the various retrospective studies of CHD prevalence. Most of these are case-control studies and show an increased percentage of smokers among those with clinical CHD when compared with a selected control population, usually without apparent CHD. Two of these studies include data on mortality. THE INTERACTION OF CIGARETTE SMOKING AND OTHER CHD RISK FACTORS The preceding section has reviewed the epidemiologic evidence which supports the judgment that' cigarette smoking is a signifi- cant risk factor in the development of CHD. Many of the studies discussed above have identified a number of biochemical, physio- logical, and environmental factors; other than cigarette smoking, which also increase the risk of developing CHD. These risk factors include elevated serum lipids (particularly serum cholesterol) and hypertension, which, with cigarette smoking, are considered to be of greatest importance. .Other facto&-are obesity, physical inac- tivity, elevated resting heart rate, diabetes (as we11 as asympto- matic hyperglycemia), electrocardiographi~~abnormalities, and a positive family history of premature CHD (88). A number of these studies have also found that these factors, when present in the same individual, exert a combined effect on the risk of developing CHD. Figures 1 through 3 depict this inter- action of risk factors. As may be noted in Figures 1 and 2, the 36 additional factor of smoking greatly increases the risk of develop- ing CHD among those people already at high risk because of other factors. Furthermore, these studies have shown that the effect of smok- ing on the risk of deveIoping CHD is statistically independent of the other risk factors. That is, when the effect of the other factors is statistically controlled, smoking continues to exert a significant effect on increasing the risk of developing and dying from CHD. Smoking and Serum Lipids The interaction of smoking and serum lipid Ievels in the develop- ment of CHD should be considered in the light of information con- cerning the relationship of smoking to serum lipid levels. Table A7 presents studies which deal, with the association between smoking and lipids, notably cholesterol, triglycerides, and lipoproteins (con- cerned with lipid transport) _ While some of the studies have indi- &,ed that smokers show increased serum IeveIs of these lipid con- stituents, others have not. The populations investigated and the methods of the various studies show significant variation. This lack of comparability makes interpretation of the tidings diflicult It is clear, however, that in the presence of high serum choles- terol, cigarette smoking increases the risk of. CHD. Figure 4 de- picts the data from the Chicago Peoples Gas, Light and Coke Com- pany study which show that smoking greatly `&eases the risk of CHD in each of the cholesterol groups. Smoking and Hypwtemion Some epidemiological studies have indicated that smokers tend to have lower mean systolic and/or diastolic blood pressures than nonsmokers, while other studies have not found this to be the case (table A 8). Reid, *et al. (155)) in a study of 1,300 British and American postal workers, found that the blood pressure difference between the smoking and nonsmoking groups was eliminated after controlling for body weight. Tables 9 through 11, derived from the study by Borhani, et aL (27) , demonstrate the following associations : That for both amok- ers and nonsmokers, the risk of dying from CHD increases with increasing diastolic or systolic pressure, and that the risk of mor- tality from CHD is higher among smokers than among nonsmokers in each blood pressure group. Cigarette smoking, therefore, ham been shown to elevate CHD mortality independently botb'of its effect on blood pressure and of the effect of hypertension on CL-ID. Smoking and Phyticul inactivity The recent study by Shapiro, et al. (172) of more fhan 110,00fr TABLE Q.-Dsath rate6 from coronas hem9 dineuse, b sy&& b&d $wuMc: 4664 . . . . . . . . . . . . . . . . `.:. - lO Never cig.rruc./day smoked r~~ulrb _ ._ . . .1.00l1.841) l.OO(1.841) Yale da. only CUlTCUt clemrettc amaken . _. . _ ._ l.DO(1.063) 2.66 (2.822 ) stoD,Xd 20 . . . . . . . . . ..-....... 1.08 (70) 1.06 (80) AU cx-dnardte smokers . -1.16 (263) 1.28 (6641 N Total &finite myacardial infarction ever smoked .._...._._... _ . . . . . . .._...... . . . . . . ..I.00 Current cigarcite smokers ......................... 1.87 StDPM S6 Yeara ................................. .0.78 AR CHD &aLhs Never smoked . . . ._. . . . . . . .1.00(27) >`j4 wck/dsy _ _.._.___. ._ .1.66(34) 1 wck/dq . . . _ . . . .._. . ..1.70(.66) >1 Drck/dry . . . . . . . . . .._.. 3.00(68) Exsmokcrx . . _ . . . . . . . . . . . . 0.80(19) Pirmf major CormION event 1.00 (63) See table 4 1.66 (72) for description 2.08(205) of Pmlinp 3.28 (154) PrOiKt. 1.25 (61) TAEKE 16.-Annual probability of death front coronary heart da%ease, in current and discontinued enwkers, by age, maximum amount smoked, and age started Bmoking .-e-v._ _ ..___. 0 601 - 601 - lo-20 198 608 811 661 7.149 S60 766 612 698 &`I(' -.__ . . . . . . .._ 0 1.016 - 1.016 - lo-20 1.601 1.169 1.478 1.21t 21-M l.710 l.ur 1.ma 1.0911 * For - group 66-74, ~r.,b.blUUa for dkeantlnrrsd -ken o R for 10 Or mOrr - d dh-- adinn.~cc dacc d.t. for the 64 ,err dk.zantLntunu craw .R not riven. Sxmcx: thmtldd, J, Yft&c~I. 8. (44). Bud cm da@ derived Irma IhIm. 1. A. (#I). 42 Smoking and Electrocardiographic Abnormalltzes Electrocardiographic (ECG) abnormalities such as T-wave and ST-segment changes as well as a number of arrhythmias are use- ful indicators of CHD and may, therefore, be predictive of the development of clinically overt CHD manifestations. The results summarized`in table 13, from the prospective study by Borhani, et al. (279, reflect the joint predictive value of smoking and ECG abnormalities on the death rate from CHD. Smokzng and Heart Rate Recent analysis by Berkson, et al. (25) of the data derived from the Chicago Peoples Gas, Light `and Coke Company study of middle-aged men revealed that resting heart rates of 80 or greater were associated with an increase in the risk of death from CHD. These authors found that this association was independent of the other major coronary risk factors. Table 14 presents the interaction between smoking, blood pres- sure, and elevated heart rate in increasing the risk of CHD mor- tality. This study shows that cigarette smoking increases CHD risk in the presence of elevated heart rate as well as in its absence. THE EFFECT OF CESSATION OF CIGARETTE SMOKING ON CORONARY HEART DISEASE A number of epidemiological studies have been concerned with the CHD incidence and mortality among ex-cigarette smokers as compared with current smokers (51, 76, 88, 90, 93, 172). These studies are listed in table 15. Table 16 presents the data derived by Cornfield and Mitchell (45) from the Dorn Study of U.S. Veterans (93). Ex-cigarette smokers show a reduced risk of both myocardial infarction and death from CHD relative to that of continuing ciga- rette smokers. The Pooling Project (88) and the Western CollaB orative Study Group (192) which adjusted for the other risk fac- tors of elevated serum cholesterol and blood pressure observed this relationship. Hammond and Garfinkel (76) noted that cessation of smoking is accompanied by a relative decrease in risk of death from CHD within 1 year after stopping. This decreased risk of CHD among ex-smokers further strength- ens the relationship between smoking and CHD. It must be noted, however, that the group of ex-smokers is composed of individuals who have stopped smoking for a variety of reasons. Those who stop because of ill health and the presence of symptoms are gen- erally at high risk and can bias the group results in one direction; 43 those healthy persons who stop as part of a general concern about their health and may adopt a number of self-protective health prac- tices are generally at low risk and can bias the group results in the other direction. Therefore, ex-smokers as a group are not fully representative of the entire population of smokers and may have limited value in predicting what would happen if large numbers of cigarette smokers stopped smoking purely for self-protection. Cer- tain incidence studies, such as the Pooling Project, (88)) were initi- ated with only clinically healthy individuals. The data from such studies, as well as those from the British physicians study, contain ex-smoker data less influenced by these biases. Fletcher and Horn (63) have recently presented data derived from the British physicians study of Doll and Hill. Over the past lo-16 years, cigarette smoking rates among British physicians have declined significantly in comparison with those of the general British population. The information presented by these authors concerning all cardiovascular diseases showed that for individuals between the ages of 36 and 64, the age-adjusted death rate for CHD declined by 6 percent among physicians and rose by 10 percent among the male population of England and Wales during the period from 1953-57 to 1961-65. THE CONSTITUTIONAL HYPOTHESIS The effect of smoking on the incidence of CHD has been found to be independent of the influence of the other CHD risk factors. When such risk factors as high serum cholesterol (1771, increased blood pressure (27)) elevated resting heart rate (251, physical in- activity (172), obesity (27), and electrocardiographic abnormali- ties (27) have been controlled, cigarette smokers still show higher rates of CHD than nonsmokers. It haa been suggested by some (39, f7U) that the relationship between cigarette smoking and CHD has a constitutional basis. That is peopIe with certain constitutional make-ups are more likely to develop CHD, and the same people are more likely to smoke cigarettes. This hypothesis maintains that the relationship between cigarette smoking and CHD is thus largely fortuitous and that the significant relationships are between the genetic make-up of the individual and CHD and between the genetic make-up of the indi- vidual and his becoming a cigarette smoker. Two sets of epidemic logic data bear on this hypotheeis. It has been maintained that people with a certain temperament are more likely to smoke and also more likely to develop CHD. These characteristics have been demonstrated for those with the 44 Type A behavior pattern of Rosenmann, et al. (159) which is characterized by competitiveness, excessive drive, and an enhanced sense of time urgency. The prospective study organized by the western Collaborative Group indicates that individuals who ex- hibit this type of personality are more likely to have or develop CHD than those without it (Type B), whether or not they smoke. When the incidence rates of CHD are analyzed with respect to smoking and personality types (tables A 17, A 18)) it is noted that in both Type A and Type B individuals the incidence of CUD is greater among cigarette smokers than among nonsmokers. This research indicates that both personality type, as measured in these studies, and cigarette smoking contribute independently as risk factors to the development of CHD. To what extent such behavior patterns are determined constitutionally or represent acquired characteristics is still open to question. The other type of research designed to study the genetic hypoth- esis has made use of data from registries of twins. Cederlof, et al. (37, 38, 39, 40) have utilized the Twin Registries of Sweden and the Veterans Follow-Up Agency of the U.S. National Academy of Sciences-National Research Council to investigate the relative contributions of heredity and smoking to cardiovascular and bron- chopulmonary symptom prevalence. Data obtained by mailed ques- tionnaires were analyzed for the following characteristics: zy- gosity of the same-sex twin pair, urban-rural residence differences, smoking concordance, and history of various symptoms. Compari- sons were made between smoking discordant monozygotic (iden- tical) pairs and smoking discordant dizygotic (fraternal) pairs, and between unmatched twin pairs and matched twin pairs. Smok- ing discordance has been defined somewhat differently in various reports but, in general, describes twin pairs in which the smoking habits differ between the two members of the same twin pai: Analyzing the data obtained from 9,319 Swedish twin pairs (`72.3 percent of the possible respondents), Cederlof, et al. (39) found that respiratory symptoms were more common among smok- era in both the unmatched and matched smoking discordant twin pair groups. The authors analyzed the data in two distinct man- ners. Group A analysis, which did not control for genetic factors ntihzed two groups; the first composed of all the firstborn, and the second of those listed second on the birth certificates. Group B analysis utilized MZ and DZ t&n pairs which were discordant for smoking, thereby controlling genetic factors. "Angina pedoris," a8 defined by a certain pattern of responses to the questionnaire, was found to be more prevalent among smokers in Group A, but this difference was not present when the data from Group B were an- alyzed. Males in the first group exhibited a "hypermorbidityratio" 45 of 1.6, whiIe those in the second group were found to have one of approximately 1.1. The authors concluded that this difference be- tween the two groups provides better support for the importance of constitutional factors as against the importance of cigarette smoking in the development of angina pectoris. A similar study was done using the responses of 4,379 U.S. Vet- eran twin pairs (approximately 60 percent of estimated available total) who completed the mailed questionnaires (38). Cederlof, et al. found a significantly increased prevalence of chest pain and "angina pectoris" among smokers when Group A. ~8s analyzed. Analysis of the smoking-discordant matched twin pairs (Group B) revealed no association between smoking and cardiovascular symp- tams among the monozygotic pairs. The dizygotic pair data did show a slight association. The authors concIuded that this Iack of association among the monozygotes and its presence among the dizygotes and unmatched pairs strengthens the case for a constitu- tional hypothesis. A major problem in these studies is the small number of cases available and, therefore, the statistical instability of the results. In the Swedish study, among the 274 monozygotes, only 19 smokers and 16 nonsmokers were classified as having angina pectoris while among the 733 dizygotes, 25 smokers and 25 nonsmokers were so classified. In neither group was the difference between the prev- alence ratios found in the Group A analysis and that in the Group B analysis of statistical significance. Analysis of the data on women shows a similar lack of significance. Similar criticisms may be made of the study which utiliyed the U.S. Veteran Twin Registry. In that study, the authors observed that the difference in the prevalence of angina pectoris Mween the low-cigarette-exposure and high-cigarette-exposure dizygotic groups was not present among the monozygotes. The authors ques- tioned whether the excess morbidity associated with cigarette smoking found in the dizygotic group was causal as it was not pos- sible to reproduce the association when studying monozygotic smoking-discordant twin pairs. As noted above, the numbers in this study are also small so that the differences in rates do not approach statistical significance. ~ Tibblin (188) has questioned the value of a mailed questionnaire to diagnose heart disease. The questionnaire as originalIy con- structed was used and validated by interview technique alone (157, 158). Cederlof, et al. (40) conducted a study to determine the validity of this questionnaire as a mailed instrument by personally interviewing and examining 170 of the twin pairs who had replied. Of the eight males who were diagnosed as having "angina pectoris" by the questionnaire. four were found to be free of symptoms on 46 clinical examination, while among 204 responding negatively, two were found to have angina by clinical criteria. None of the 11 women who were diagnosed as positive by questionnaire was found to be cIinicaIIy affected, and of the 136 reporting as negative, three had symptoms of angina pectoris. Other major difificulties associated with these studies include the problems of using prevalence data in the investigation of a disease (CHD) from which a significant number of those affected die shortly after the onset of symptoms, the inclusion of ex-smokers in the smoking population, and the,low numbers of heavy cigarette smokers in the Swedish population. In general, the problems of using twin registries to study the etioIogy of cardiovascular disease with mortality and morbidity ratios in the neighborhood of 2 to 1 are much more difficult than in studying the etiology of bronchopulmonary disease in which the relationships are of the order of magnitude of 4 to 1. More recently, Friberg, et al. (69) reported on mortality data from the Swedish Twin Registry. The authors suggested that part of the increased mortality observed among smokers when com- pared with nonsmokers was not due to smoking per se but to fac- tors associated with smoking. The very small numbers of total deaths presently available (47 deaths among 706 dizygotic pairs and 13 deaths among 246 monozygotic pairs) do not provide a sta- tically stable base for deriving any conclusions at the present time. huge, et al. (81) have recently reported on the influence of smoking on the morbidity and mortality observed in the Danish Twin Register. Among 762 monozygotic and same-sexed dizygotic twin pairs, angina pectoris was found to be significantly more fre- Went in those cotwins with a higher consumption of tobacco than h those with a lower or no consumption. A similar tendency was observed for myocardial infarctions but was not of statistical significance. Seltzer, who has been a proponent of the constitutional hypothe- sis, in a recent review of some of the experimental, clinical, and PothO~Ogica~ data relating smoking and CJID, concluded that the evidence from these areas has not "reasonably substantiated" the "hypothesis" of the acute effect of cigarette smoking on the coro- narY circulation, nor has the chronic effect of cigarette smoking on ihe cardiovascular system been shown to be a "clear" and eon- gist& one (170). His views are contrary to those of most re- achers in this fieid. mhough the data from the twin studies are inconclusive with regard to a role for genetic factors in heart disease, it w&Id be surprising if genetic factors did not play such a role. It is open to 47 question whether findings from twin studies can be used to distin- guish between the hypothesis that genetic factors govern the level of host susceptibility or resistance to the effects of an exogenous influence such as cigarette smoking and the hypothesis that genetic factors "cause" both heart disease and smoking. AUTOPSY STUDIES RELATING SMOKING, ATHEROSCLEROSIS, AND SUDDEN CHD DEATH A number of researchers have investigated the ci&rette smoking habits and the cardiovascular pathology of th,ose.individuaIs dying suddenly from CHD and of large populations of individuals with and without histories of overt CHD. Spain and Bradess (175) recently analyzed the smoking habits of X89 individuals who died suddenly and unexpectedly, apparently from the first acute clinical episodes of CHD. The authors nofed a close correlation of a history of cigarette smoking with this type of sudden death and also with shorter survival times following the acute episode. This association was strongest in those persons under 50 years of age. The Authors also observed that those survlvmg very short pe- riods of time showed a notable lack of intracoronary artery throm- bi at autopsy and that the frequency of thrombi present increased with increasing survival time. They suggested that thrombi found at autopsy may be the result rather than the cause of certain instances of myocardial infarction, particularly of lesions showing subendocardial necrosis. This finding is of significance in the study of the effect of smoking on myocardial metabolism and oxygen supply and demand rather than on thrombus or platelet plug formation. While the autipsy study of Spain and Bradess (175) concerned sudden death among smokers, other autopsy studies from various countries have been directed towards the relationship of cigarette amoking to the presence of atherosclerotic disease in the aorta-and .coronary arteries. These are concerned with the long-term effects .which smoking has on the cardiovascular system and are `sum- marized in table 19. The studies of Auerbach, et al. (IZ), Avtan- dilov, 8 al. (IS), Sackett, et al. (165), and Strong, et al. (~82) round that aortic and coronary atherosclerosis were more common and more severe among smokers than amongnonsmokers. Auerbach, et al. (12) .Gted that this relationship persisted when the cases were matched for both age and cause of death or when the follow- ing cases were excluded ; men.with a history of diab&; men who had died of any type of heart disease ;. and men whose hearts weighed 400 grams, or more. Sackett. et al. (165) found that the 40 TAEILE lO~Autopet( studiea o/ athsroaclaroaia (Flguta la pwmthnn an number of Indivldualc In that amoklnc ertrn0r7)~ [SM = anoken NS = nonmokcrs] AdhOt, 7*rr, Aubpr7 D&tA COUl-li~. DODUktlOC, collcctlon Comment4 odjwkd rrrulk) No othcro- NkTO6b Slight NS .;.., ,.,. 6.6 (69) --.`-. 67.8 Cumnt elgarctte <20 *.,.*1 l.6(139) 30.9 20-39 .,.,0.8(299l lg.7 >49 ,,.`.;.0.6(144) 18.1 the percentage of men with .?I o dv*need deorn of M0hub. Advonccd coronary d,herolckm~k 21.8 16.8 wu hlpbcr .mong elsn- - rette rmokcn thnn among nonrmokem and that the 873 29.2 ' pcrecntagc hcreued 42.1 87.4 with amount of elgrretts 96.4 46.9 I 8moklng. This relation- lblp DM'dlted CvCIl nhm cad" were mhtehed for age and UUI~ of death. TABIX 19.-Autopsy studies oj athumsch'osia (cm:.) (?`I- In DuSrrthC.4. w number of IndlrlduLL in that fimOkiU ukmr7)' [9X = mmken NB = noo~mekenl Ln Anthor, 0 Y-t AntaDtoprl Dab OoOMtn, DODUbtk"3 ' wnudm Ckamtta DCr dw conctuliona cornlnmta rdmu Avtandlh, 219 mab and Nat BDrChdd, fh,SDWdW #(I# of maan a,~ of athbtcrobrutb b&BU The rutbar eoncluda that Cawa of dath 9l-&mro- 1966, 141 f*mmb but there rarer in inn47 mat of covmary a.rkriar. the wont ehanga rm Acrotle, lOZ-wcldrntA, Flunrlr ~~~D~l~. 180 9X uid Rlpht COTOMY ortsru h/1 mrona.ry orby found In the left and 202varloru dlrcua. (Ia). 220 NS. srd Nf mf NS rlsklt ??????? o ?????? tT-tent for +~.lltuncr a049 ,*.tla.s(ao, Lscaz) ma 2.2 nlth h mvcra changer of hlffcrenca klwcen 40-49 . , tzs.ecao 11.6(27) tu.a 4.4 hi clreu.m.8a artery means II llgnldunt so-t.9 ..tae.a(so, 14.8(89) t27.9 9.9 &rid wti. at P