TABLE ~A.--COTO~M;GT~ heart disease (men). Age+tandardized o!eath ,'a&8 for t?X-Ci@Z?Vtte SmOkeT8 With hi&n-y Of C&re#e SmOkiTbg &y, by former number of cigarettes smoked per day and years since last cigarette smoking. Death rate8 for current cigarette smoker8 uith history of cigarette smoking only ano? men who new smoked regukzriy are 8h0m fqr comparison. Men aged 60-69 Smoked l-19 cigarettes a day Smoked !B+ dtpretted a day Ep&gIretta SmOh~ (YearS since last cigarette smokW z% Nlllllber Death %z NUCULXE "Sh ofdesths rate of de8tba -- ----- -- -- Under 1 year--- _-____ 745 27 11,005 2,244 77 1 1,070 lt04years ____--__-- 1,844 51 718 5,435 195 1,003 5togyeal-s __-------- 1,770 48 725 5,803 152 732 lo+ years ___________ 4,209 84 498 8,142 205 679 Total ex- smokers----- 8,569 210 635 21,624 630 813 Current cigarette smokers- _ _ _ _- -_ _-_ 22,808 781 947 56,886 1,895 1,029 Never smoked regu- larly- _ _ _ ___ _-- _--- 55,728 1,114 502 55,728 1,114 502 1 Four or mom but less than 10 deaths expected in 6ome of the component &yaar age groups. SOURCE: Hammond, E. C. WI, p. 1481. nolZ and EiJZ In $5 prospective study by Doll and Hill (n) of mortality among British physicians whose smoking habits had been previously recorded, there were 1,369 deaths in the course of 10 years in which the underly- ing cause was coronary heart disease ($7, table 1). The physician popu- lation under observation totaled 320,185 person-years. The CHD deaths were clas.si&d into three major subcategories: Group 1, com- prising 35 CHD deaths in which an associated condition related to smoking, e.g., lung cancer, was recorded on the death c&&ate; Group i), comprising 721 CHD deaths in which no other significant contribu- tory cause of death was recorded on the death certificate; and Group 3, c&omprising 613 CHD deaths which were associated with some other contributory cause, including conditions known to predispose t.o core- llary heart disease, e.g., hypertension, diabetes, and obesity. The CHD death rates for smokers and nonsmokers based only on Group 1 deaths, while subject to large variatiofi, show the largest differentials (data not, shown). Among smokers of 25 or more cigarettes daily, the age-ad- justed CHD death rate was nearly eight times that in nonsmokers. Based on Group 2 coronary heart disease deaths, presumably uncom- l>licated by any other sign&ant disease, the mortality ratio of age- :tdjusted death rates among continuing cigarette smokers to non- smokers is found to be 1.6, and for heavy smokers to nonsmokers the ratio is 2.0. However, as shown in table 5, the mortality differentials between smokers and nonsmokers are much larger at the younger ages. ThstE: FL-iUorta&y ralioe for di$krent typea of cormv h.mrt disease by mnoking h&$8 I Group 2 ' CHD I Oroup 8 CED 35 to44------------- 45to54------------- 55to64 _-___________ 65to74------------- 75to84------------- ssplua -----------___ - Age adjusted-All agsa- 1.0 4.7 9.7 1.0 3.8 3.5 1.0 1.4 1.6 1.0 1.4 1.8 1.0 1.1 2.0 1.0 1.0 Q 1.0 1.6 2.0 __ 1.0 -_ Cl (`1 1.1 1.7 1.4 1.9 1.4 . .8 .a .2 1.4 (9 1.1 .9 ' Not calahbk no rata for mmamok~ because of so few deaths. *Very few men in this category. Souacr: Data in above table based on valups from Study of British l'by&Jans. Table 8 (27J. The mortality ratios shown for Group 3 deaths, i.e., CHD deaths accompanied by some other complicating disease, suggest that, for all age groups combined, smokers do not have any special risk to this type of coronary death. However, smokers below the age of 65 appear to be at a somewhat greater risk, while no consistent differen- tials are observed among persons in the older age groups. In summary, the study substantiates other mortality studies' find- ings that CHD mortality ratios (current cigarette smokers vs. non- smokers) increase with the number of cigarettes smoked daily, that the ratios are highest in the age group 45-54, and that they decrease as age advances. Moreover, smoking apparently is assooiated with deaths from coronary heart disease among persons free of other serious disease states. In a prospective study of California longshoremen, Borhani (17) reported on the mortality experience of more than 3,700 men observed for 10 years. Table 6, derived from his data, provides some additional insights on both the independent and the inter&ion effects of ciga- rette smoking. Men 45-64 years of age who were heavy smokers experienced higher death rates from coronary heart disease than did nonsmokers independ- ent of whether they were hypertensive or nonhypertensive. 52 TABLE 6.-Mortulitg ratio from corvnury h&W disea&?e amo?zg m&e hypertensives and nonhypertensives by smoking history and age Age group Blood ptpssarestatns 1 NOll- r -v smokus~ Emokus' -- 45 to ,rx,- _ ___ _ ____- Nonhypertenaive ______________ LO 2.2 Hypertensive, _ _ _ _ ___ ---i ____ _ 2.5 9. 6 55 to 64 ____________- Nonhypertensive -___-___------ LO 5.8 Hypertensive, _ _ _______ __ _____ 5.9 9. 4 1 Hyp&?l,S,V~ are de&& B(J tboSe b8m E$`StOlk blood m of MD mm. as. OX OV= M dbtCUc ,,,d presure of 9smm. H&T. or ov=. Nonh &mslvfa hava ayatdlc blood praaau~ lem than 18 mm. m cr dtastok? blood P-m @JE than 96 mm. r~o,,smokaalnthiSpsrtiCUlSC8tUdyaFe 6fhed~SthOaen0tSmokh@8UYdW0tt=Orl~th1m~ Y! cigarettes per day. Smokem are those who smoke !ixl or mere pu day. SOURCE: Borhsni, N. o., et 8l. (II). ,.n analysis was made by Schor (80) of 181 adult males who died from coronary heart disease generally less than 2 years after receiving ;I periodic health examination. The results of this study and those of Doll and Hill suggest that sudden death from previously undetected coronary heart disease frequently occurs among cigarette smokers. If tIlis is true, it may, in part, account for the small differentials in the prevalence of coronary heart disease between smokers and nonsmokers observed in some morbidity prevalence surveys. As will be described in the following section, longitudinal, prospective morbidity studies nlso show that smokers are more likely to die from sudden attacks of coronary heart disease. CORONARY I&ART DISEASE MORBIDFIT* In chapter 11 of the 1964 Surgeon General's Report, several prospec- tive studies on the incidence of coronary heart disease (!34,31,78,88) established that smokers were subject to higher rates than nonsmokers. The relationship was reported to be more marked under 60 years of age than among older persons and appeared to be associated with myocardial infarction but not with angina pectoris. Since the 1964 report, recent findings from large-scale, on-going prospective studies have been reported, providing additional insight on the interaction between smoking and other important coronary heart disease risk fac- tors. Current tidings are summarized in the following pages includ- ing tables 7 to 13. Whenever possible, data are shown separately for findings related to angina pectoris and those pertaining to myocardial infarction, including sudden death attributed to coronary heart dis- ease. Higgins (60) has drawn attention to the fact that "many factors lnay influence or be affected by smoking habits, and obscure those differences between smokers and nonsmokers which are directly re- lated to the use of tobacco." In her review of the literature, Higgins identified differences between smokers and nonsmokers in genetically o 8lrso may include mortality data in this presentation. determined qualities (M), in physique (77,98), in pemonality (87, 47, @, 65, 68, 69), and in social, cultural, religious, and economic characteristics ($6, &,68,84). Age The elect of age on the incidence of coronary heart disease with regard to cigarette smoking is shown in table `7 based on recent data from the Framingham Study as yet unpublished. TABLE 7.-ha+&imm raks and morbidity ratios for coronay be& disease by age a& emoting status of men l%yew experience, Fm- minghum, Maas. AP 35to44 -_-___-_______________ 1.4 4.1 2.7 T 1.0 2.9 45to54 -_-__--_-________-____ 4.6 11.1 0.5 1.0 2.4 55to64-----_-_---__-_____,__ 16.2 25.4 9.2 1.0 1.6 SOURCE: U.8. Pnblk He&h Semke~, Fiamhgbam study (es). (Updnted 1967) When the incidence rate of coronary heart diseaseOamong male non- smokers.between 35-44 yeers of age is compared with that among older nonsmokers, the rate is seen to triple every 10 years. This marked increase in incidence among nonsmokers reflects the effect of other important risk factors and perhaps accounts for the decrease in mor- bidity ratio as age advances. The independent effect of smoking on the incidence of coronary heart disease is believed to be more appro- priately represented by the excess morbidity rates, which increase from 2.7 per 1,000 smokers in the age group 35-44 to 9.2 per 1,000 smokers 55-64 years of age. &$b &!OOd ht?88WW? Although the inhalation of cigarette smoke is frequently accom- panied by acute transient elevations in blood pressure, habitual smok- ers tend to have lower blood pressures than do nonsmokers (48). But, given the presence of high blood pressure in an individual, smoking acts as an additional risk factor for the development of coronary heart disease (l7,$?8, m, 90,53,55,95, 96). Both the independent and the combined effect of cigarette smoking is clearly shown in table 8 de- rived from the experience of the Framingham and Albany studies (W* 54 TABLE 8.-Ageadjualed morbidity ratios for coronary heart disease among smokers and nonsmokers according to level of systolic blood pressure Systolic blood p- Noos"p;z of "2ZJC - --- Under 130mm.Hg___--__------___-___--__--_-- 1.0 2. 1 130 mm. Hg and over- _ _ _______________________ 1.8 3.8 SOURCE: l&year Framingham and &yesr Albany experienci (3% Eigh sem choze-stt?roz It is not now conclusively known if cigarette smoking by itself can cause increases in serum cholesterol. Dietary influences as well as en- dogenous production and elimination of cholesterol must be evaluated in greater detail with simultaneous analysis of the roles of other risk factors, including smoking. One study of a small population of twins in Sweden, as reported by Lundman (67)) suggests that smoking monozygotic twins tend to have lower cholesterol levels than their nonsmoking control twins, although the differences are not statistically significant. Other studies suggest that smokers generally have higher serum cholesterol than nonsmokers (13, 37, 88). However, given the presence of high serum cholesterol, smoking increases the risk of cor- onary heart disease (96,96'). The independent any synergistic effect of cigarette smoking is dem- onstrated by the data in table 9 derived from the combined experience of the Framingham and Albany studies (30). TABLE O.-Age-adjusted morbtiity ratios for coronary heart disease among smokers and nonsmokers according to level of serum cholesterol Serum cholest~ol level Low'-_-___--____________________________----- 1. 0 1. 8 Highs-___________-___________________------ 2. 0 4.5 1 "Low" is below m&an. %igh" is above median value of sermn chloeaterol. SOURCE: l&year Fmmiughcm and &year Albany esperfence (Jo). Puzmonar/y Fzunction The acute effects of cigarette smoking upon pulmonary function are expressed mainly through increase in airway resistance. The dif- ferences in pulmonary function between smokers and nonsmokers ap- pear to be greater than can be accounted for by acute effects from a recently smoked cigaretta (50,91). The relationship of coronary heart disease to lowered pulmonary function as reflected by low vital capacity and cigarette smoking is observed in the data published by the Na- z7&SQ4 o-67--5 55 UUU~~ msmutes oi Health based on the 12-year experience in Framing- ham (96). Morbidity ratios derived from this publication are shown in table 10. TABLE lO.-Age+dju&ed mo&idiiy ratios fw coronary he& disease among smokers and nonsmokers according to IePel of erital cupa&y wal capacity Under 3 liters _-_---_ - ___ ___ _ ______ _ _________ ___ 1.0 I I 3 liters or more --------------------------------- 1.7 1.7 2.4 Sonaclr: The FraxnIngham heart study. (6W). Here again, the independent and combined effects of cigarette smok- ing are observed. Physical znaotivity A physically inactive or sedentary individusl seems to run a higher risk of developing coronary heart disesse (89, .&I, 0,76). Spain (88) reported that, in his prospective study of 3,000 men "* * * the rela- tionship of occupational physical activity to smoking habits revealed that one of six sedentary workers were heavy smokers and one of five strenuous workers were heavy smokers." Weinblatt, in reporting the experience of the Health Insurance Plan of Greater New York (100) also found that a higher proportion (41.9 versus 36.0 percent) of cig- arette smokers were classified in the "most active" physical activity category. The independent and the combined effects between cigarette smok- ing and physical activity are shown in table 11. The morbidity ratios for myocardial infarctions are derived from published data. TABLE ll.-Age-cs&justed morbidity ratios for myocardti infarctions among smokers ano! nowmokers according to phgsi.eal act&i@ &we1 Physical actitity Most active------------_------_---_____-____-__ 1. 0 I I Lenstactive---------_-_-----------~----------- 24 2. 6 3. 4 SOUBCE: Weinblatt, E. (fOrI). Socioenvironmenta.2 Stress Since 1955, research on socioenvironmental stress in relation to cor- onary heart disease has increased greatly (#, 99). Among the factors studied that indicate a strong association with coronary heart disease incidence and prevalence is sociocultural mobility, that is, moving from one social setting to another. The interaction of this factor and 56 cigarette smoking has been reported by Syme (90,91) in both an ur- ban and rural setting. Apparently in both areas cigarette smokers \vere more culturally mobile than nonsmokers. The independent effect of cigarette smoking on the incidence of coronary heart disease is shown in the morbidity ratios in table 12 derived from the North Dakota study (91). TABLE 12.-Age-adjusted morbidity ratios for coronary heart d&ease among smokers and nonsmokers according to sociocu&uraJ mobility status socioouItumI status Ne;wwT$ed Current and fonn~ok~grette stable __________________ - ____ - _______________ - 1. 0 1 1 Ijighly mobile--- __ _- ____ -__ ___ ___ _ __ _ ___ _ ___ ___ 2. 3 1. 5 3. 2 SOURCE: North Dakota study. (81) Personality Type Various investigators have long `suspected a possible pathogenetic role of the central nervous system in coronary heart disease (35). In a series of reports, Rosenman (82,8@ and Jenkins (51) have described a personality pattern or overt emotional complex which, while asso- ciated wit.h other known risk factors, appears to predict coronary heart disease more effectively than do other risk factors. This emotional com- plex, `&which they have termed Behavior Pattern Type A, is composed of an enhanced competitiveness, drive, aggressiveness and hostility, nud an excessive sense of time urgency." Recent unpublished data based upon prospective observation of more than 3,000 men for a 4+`&-year period (51) discloses that smokers have a higher percentage (54 versus 4'7 percent) of type A persons among them. Moreover, the incidence of coronary heart disease is shown to be related independently to both smoking status and personality type. Morbidity ratios, derived from the incidence data, are shown in table 13 which clearly demonstrates the independent effects of cigarette smoking and its interaction with personality characteristics. TABLE 13.-Morbidity ratios of cigarette smokers as compared to non.- smokers by personality type SOUBCE: Unpublished data from We&em Collaborative Group Study, San kandseo, C&if. (eb). Mdtiple-R&k Factor8 The method of analysis traditionally employed by epidemiologists, that of the comparison of rates for. multiple cross-classifications of the data, generally requires a large study population at relatively high incidence of significant events. Since coronary heart disease incidence rates are low and study populations are necessarily small because of practical and practicable limitations, definitive analysis of the inde- pendence and interaction between risk factors have generally been re- stricted to two factors at a time. Truett (96) applied a multiple logistic function proposed by Cornfield to investigate the independent effect on the incidence of coronary heart disease of seven risk factors: Age, serum cholesterol, systolic blood pressure, relative weight, hemoglobin, cigarettes per day, and ECG abnormalities. The method was used in the analysis of data compiled in the Framingham study during a 12-year period. A discriminant function coefficient was computed for each risk factor. These-coefficients represent the relative importance of each fac- tor with respect. to the other six factors in the development of coronary heart disease. While theoretical considerations underlying the logistic risk function are not fully satisfied by the actual data, the approxima- tion given by the function to observed rates is very good. Consequently, Truett and Cornfield believe that the present compu- tations permit the conclusion that "the most important risk factors, aside from age itself, are cholesterol, cigarette smoking, ECG abnor- mality, and blood pressure" (96). MIANIFE~TATION OF CORONARY HIURT DIEIEASE Coronary heart disease is essentially comprised of three major mani- festations or subcategories : 1. Fatal myocardial infarctions, including sudden deaths attrib- uted to coronary heart disease ; 2. Nonfatal myocardial infarction; and 3. Angina pectoris. Generally, investigators in their analysis of the relationship of risk factors to the incidence of coronary heart disease have not subdivided the observed coronary events into the three major subcategories pri- marily because the paucity of events in each category did not permit definitive conclusions on any differences observed. However, the pool- ing of data from some of the larger prospective studies holds promise of a more complete analysis of the independent and synergistic effects of each risk factor on each of the subcategories of coronary disease. Findings from these analyses might provide some insights into the underlying pathophysiological mechanisms t.hrough which a risk factor operates. The pooled data from the Albany and Framingham studies a.nd data from the HIP study include the observed associations of 58 cigarette smoking with each of the three major manifestations. Mor- bidity ratios have been derived from these studies and are presented in table 14. TABLE 14.-Agemfjusted nwrbidity mtios for subcategories of coronary heart disease among smokers and nonsmokers Framingham-Albany Ewaltll fneurance plan Disease category Non- "&Tg NOU- smokem of smokea of cIgarettea clgncettes "m% - --- ~-- --- ~_ Fatal myocardid infarction _________ 1. 0 2. 4 1. 0 2. 1 Non-f&d myocardial infarction- _ _ _ _ 1. 0 2. 3 1. 0 1. 8 Angina pcctorie __________ ___ ___ _ ___ 1. 0 1. 1 1. 0 1. 7 Socmcr: Second Report of the Comblned Erperlence from Albany and Framingbarn Studies (JO). VII- &dished Data from Health Innuance Plan Study (loo). The association of cigarette smoking to angina pectoris is not a con- sistent one. A clear-cut association was found in the Health Insurance Plan Study (ratio of 1.7) ; a similar association is also found in un- published data from Framingham considered separately. However, no association between cigarette smoking and the incidence of angina pectoris was found in the Albany experience. Cederlof (19)) in his analysis of prevalence data on angina pectoris obtained by question- naire, found no statistically significant difference in prevalence rates between 453 monozygotic twin pairs with dissimilar smoking habits. In a larger study of about 9,000 persons from the twin register where genetic factors were uncontrolled, Cederlof (19, m) did find a sig- nificantly higher prevalence of angina pectoris among smokers than nonsmokers, particularly in men (ratio of 1.6) (fl) . Friedman (&!) and Epstein (96) have clearly described the in- herent biases in prevalence studies which may lead to findings of risk gradients that are different from those obtained in prospective inci- dence studies. One of these limitations is that fatal cases are under- represented in a prevalence survey. Thus, since it appears that cigar- ette smoking is more closely related to the incidence of fatal myocar- dial infarctions than to other forms of coronary heart disease, it is ex- pected that morbidity ratios derived from prevalence surveys would be lower than those computed from incidence data. With these restrictions in mind, Russek (83) in a survey of 12,000 men in 14 occupational groups found that the morbidity ratio of coronary heart disease preva- lence among cigarette smokers was as high as 1.8. In contrast, in a study of 77 identical and 89 fraternal twins in Sweden, comparing smokers with their respective nonsmoking twins, Lundman (67) reported no excess prevalence of overt or silent coronary heart disease. However, the prevalence of these conditions, as Lundman concluded, "was too low to permit of definitive conclusions." ~ARDI~~A~~~LAR RJWPONSE. m SIKOKING AND/OR NIOOT~VE As noted in the Surgeon General's 1964 Report, nicotine has definite physiologic effects on the cardiovascular system of experimental ani- mals and of man. These include increases in heart rate, systemic arte- rial pressure, cardiac output, stroke volume, and velocity of myocardial contraction, all resulting in an increased myocardial tissue oxygen demand (16). Coronary blood flow studies will be reported in the next section under a separate subheading. These effects parallel those ob- served with catecholamines (epinephrine and norepinephrine) . The effects can be blocked by the injection of tetraethylammonium chloride and markedly reduced by adrenalectomy (28). Nicotine has been re- peatedly shown to release endogenous catecholamines (67,68,69, M, Y&Z?). However, the mechanism by which nicotine affects the cardio- vascular system is more complex than the release of catecholamines from the adrenal medulla. Direct and indirect (via the carotid body and other chemoreceptors) stimulation of the vasomotor center, stimu- lation of sympathetic ganglia, release of norepinephrine from local stores, and release of antidiuretic hormone are included among other postulated mechanisms of action involved in nicotine's effect on the cardiovascular system (IS, &`,85). Coronary Blood Fho in Nod Subjects The action of smoking and/or nicotine on the coronary blood flow of normal human subjects has not yetbeen definitively established, but apparently normal subjects can increase their coronary blood flows sufhciently to maintain a compensatory blood supply to the myocar- dium despite the increased myocardial tissue demand for oxygen caused by cigarette smoking. Earlier findings of increased coronary blood flow in normal men, in response to cigarette smoking (11)) were not reproduced in a more recent study (16). In this latter study, al- though a trend towards a slight increase in coronary blood flow was observed in the particular normal persons studied, it was not significant. Direct injection of nicotine into the left coronary artery of dogs in- der conditions of constant flow rate resulted in increased coronary VW+ cular resistance (38,S4). This response could be reduced by vagal nerve stimulation or prior adminstration of acetylcholine; an immediate in- crease in cardiac contractile force was also observed that could be similarly reduced. It was concluded that these responses to nicotine resulted from sympathetic nervous system activity or from the release of catecholamines by myocardial chromaffin tissue (64). Blood from the smoke-exposed lung tissue of dogs, directly perfused 60 into the coronary artery, failed to increase coronary resistance (38). This effect was thought to be secondary to that of histamine, known to act as a coronary vasodilator, which apparently is released from the lung tissue of dogs during their exposure to smoke (8). Inen blood from the smoke-exposed lung was perfused through the systemic circulation of dogs while the coronaries were being perfused with non-smoke-exposed blood, the typical release of cate- cholamines occurred with many of the usual effects on cardiovascular parameters except that the coronary vascular resistance increased under these experimental conditions, apparently due to the increased activity of the sympathetic nervous system (98). Since it is well kuown that exposing dogs to cigarette smoke without isolating and separately perfusing the coronary circulation normally results in an increase of the coronary blood flow (38)) the manipulation of experimental conditions as described suggests that there is a masking effect by the catecholamines on nicotine's direct action on the coronary circulation (38). These studies may relate, by analogy, to humans and indicate that smoking, in "normal" individuals, may produce at least two actions that can affect coronary blood flow : (1) a decrease in coronary blood flow by a possible direct action of nicotine on the coronary circulation (demonstrated in dogs), and (2) an increase in coronary blood flow as the usual resultant of varying responses to the intermediating action of catecholamines and other physiologic processes (demonstrated in both animals and humans). Cormmy Blood Pkm in Subjects with Coronay Heart Disease The effect of cigarette smoking on coronary blood flow was studied in patients with coronary heart, disease (79). As was seen in normal subjects, significant increases in heart rate,' arterial pressure, and cardiac output were noted. In contrast to the normal individuals studied, patients with coronary heart disease distinctly showed a much less significant compensatory increase in their coronary blood flows. These results were confirmed by a later study (16)) using the Rubidium 84 method