"The term asthma is not appropriate for the bronchial narrowing which results solely from widespread bronchial infection, e.g., acute or chronic bronchitis; from destructive diseases of the lung, e.g., pulmonary emphysema; or from cardiovascular disorders. Asthma, as here defined may occur in vascular diseases, but in these instances the airway obstruc. tion is not causally related to these diseases." In rare instances, allergy to tobacco products has been ascribed a causa- tive role in asthma (99, 105, 168, 169, 189). Support for this association comes largely from the presence of skin test reactions to tobacco products and passive transfer tests (168, 169). In the "Tokyo-Yokohama Asthma" studies, a severe asthma-like disease, presumed to be caused by air pollution, affected cigarette smokers predomi- nantly ( 155 J . The absence of smoking data on unaffected members of the same population leaves the question of an additive effect of cigarette smoking unanswered. One study suggests that non-smokers may have a slightly greater prevalence of asthma than smokers; the possibility of bias due to self-selection of the base population could not, however, be excluded in this study (84). Apart from the exceptions noted above, it is clear that cigarette smoking is of no importance as a cause of asthma. A hypothetical contraindication to cigarette smoking can be postulated for asthmatics on the basis of the physiologic alterations induced in the tracheobronchial tree by tobacco smoke. Nonetheless, substantiation of worsening from cigarette smoking in asthmatics has not been reported frequently. A cause-and-effect relation- ship between cigarette smoking and asthma, as defined above, is not supported by evidence available. RELATION OF SMOKING AND INFECTIOUS DISEASES The category, influenza and pneumonia (ISC 480-493)) contributed to the excess mortality of smokers observed in six of seven prospective studies ( Chapter 8, Tables 19 and 26). Details sufficient to warrant conclusions about the nature of this association are not presented in these studies, nor has the apparent association been evaluated further by careful epidemiolopi- cal research. Studies adequate for examination of this association are available for only two categories of infectious diseases, upper respiratory viral illness ,and tuberculosis I 30 I . Experiments on transmission of common colds failed to demonstrate increased susceptibility in volunteers with a history of ciga- rette smoking (50 J . Jloreover, common colds were detected among 5,500 employees over a P-year period with approximately the same frequency in smokers and non-smokers (110). In a study of illness in a group of families under close observation for several years, the frequency and severity of common respiratory diseases, such as the common cold, rhinitis, laryngitis, acute bronchitis, and nonbacterial pharyngitis, were the same in cigarette smokers and non-smokers (21). Similar results were obtained by ques- tionnaires in an analysis of the frequency of common colds in a group (Jf college graduates followed over a 20-year period (85). 276 A number of studies have suggested a substantial relationship between smoking and pulmonary tuberculosis (55, 125, 133, 175). The possibility that the relationship is not a direct one needs further careful examination. Certain social factors, important to epidemiological assessment in tubercu- losis, have not been considered in detail in these studies. Of particular interest in this regard is a study (29) in which both cigarette and alcohol consumption were found to be in excess in tuberculosis patients as compared to the matched controls. The number of cigarettes consumed in the two groups was the same, however, at each level of alcohol intake. Matching by cigarette consumption failed to weaken the association between alcohol con- sumption and tuberculosis (29). Thus, the relationship between tubercu- losis and smoking in this study was only an indirect one: the association was found to occur between smoking and alcohol consumption and between alcohol consumption and tuberculosis, rather than between smoking and tuberculosis. Thus the association between smoking and the infectious diseases is con- fined at present to a single cause-of-death category : Influenza and pneumonia contribute to the excess deaths in cigarette smokers, but the data are insuffi- cient to evaluate this observation. In the limited number of studies avail- able, cigarette smoking has not been shown to contribute to the incidence or severity of either naturally acquired or experimentally induced upper respir- atory viral infections. CHRONIC BRONCHOPULMONARY DISEASES Mortality for certain respiratory diseases (bronchitis, bronchiectasis: chronic pulmonary fibrosis, chronic interstitial pneumonia, and emphysema) increased in the decade 1949-1959 (48) and continues to show an upward trend (132, 141) . In 1955, cancer of the lung was certified as the under- lying cause of death in 27,133 persons and chronic bronchopulmonary dis- eases in 11,480 persons. A tabulation of all diagnoses, both contributing as well as underlying causes of death, however, showed that cancer of the lung was entered upon a total of 28,123 death certificates, whereas the chronic bronchopulmonary diseases were certified as contributing to 32,051 deaths (47). The possibility that mortality data, as presently recorded, may under- estimate the role of chronic bronchopulmonary diseases through incorrect listing by the physician as contributory rather than the principal cause has also been suggested (115). Social security records in 1960 show that chronic bronchopulmonary dis- eases, particularly emphysema, ranked high among the conditions for which disability benefits were allowed to male workers 50 years of age or older in the United States (186). Chronic bronchitis and emphysema are the chronic bronchopulmonary diseases of greatest public health importance in the United States. They contribute to the excess mortality of cigarette smokers. but there is little information about the effects of smoking on the other chronic broncho- Pulmonary diseases. The scope of the subsequent remarks is limited there- fore to the possible relationship of smoking to chronic bronchitis and 27; emphysema. Since dexriptions of both were published long before ciga. rette smoking became commonplace (13, 14, II4), it seems reasonable to suggest at the outset that cigarette smoking alone 1s not the only cause oI chronic bronchitis and emphysema. Chronic Bronchitis and Emphysema DEFINITIONS Many definitions of chronic bronchitis and emphysema have been sue. gested. For the purposes of this report the definitions proposed by the American Thoracic Society (4) will be used: "Chronic bronchitis is a clinical disorder characterized by excessive mucous secretion in the bronchial tree. It is manifested by chronic or recurrent productive cough. Arbitrarily, these manifestations should be present on most days for a minimum of three months in the year and for not less than two successive years. Many diseases of the lung, e.g., tuberculosis, abscess, and of the bronchial tree, e.g., tumors, bronchiec. tasis, as well as certain cardiac diseases, may cause identical symptoms: furthermore, patients with chronic bronchitis may have other pulmonary or cardiac diseases as well. Th us, the diagnosis of chronic bronchitis can be made only by excluding these other bronchopulmonary or cardiac disorders as the sole cause for the symptoms." Tb is . . definltlon and classification of chronic bronchitis later considers complications. listing three: infection, airway obstruction, and pulmonary emphysema : "Emphysema is an anatomic alteration of the lung characterized by an abnormal enlargement of the air space distal to the terminal, non- respiratory bronchiole, accompanied by destructive changes of the alveolar walls." DIAGNOSIS The diagnosis of chronic bronchitis is based essentially on descriptions of clinical manifestations and is achieved by exclusion. Recollection and interpretation on the part of the subject are necessary. There is, no simple sensitive pulmonarv function test that will indicate which person has chronic bronchitis. A clinical diagnosis of emphysema, based on the clinical syndrome and certain changes in pulmonary function, is even less exact. The clinical features usually encountered in emphysema tend to be very similar to those found in chronic bronchitis. Most of the symptoms and signs and many of the physiological changes usually thought to indicate the presence of emphysema may result from airway obstruction due to bronchitis (66, 180). There is no completely satisfactory method of detecting emphysema by pulmonaq function testin, - and no pulmonary function test is specific for the detection of pathologic lesions of emphysema (521. The clinical detec. tion of emphysema is therefore not a simple matter, especially in the presence of chronic bronchitis. 27% The following, adapted from the American Thoracic Society-`s statement (4), epitomizes the situation for emphysema: Clinicopathologic correlations have demonstrated that certain per- sons who have this morphologic alteration at autopsy have symptoms of pulmonary insufficiency during life and die of this disease. Others show- ing qualitatively similar pathologic findings had no respiratory symp- toms during life and died of unrelated causes. In some persons, em- physema may be strongly suggested by the patient's symptoms and its existence predicted on clinical grounds with considerable accuracy. On the other hand, clinical manifestations identical with those of patients with emphysema may occur in persons who are not found to have this disease at autopsy but who have some other lung disease. Emphysema may exist without any clinical manifestations, and its clinical and func- tional alterations are not unique but occur in other pathologic conditions. RELATIONSHIP BETWEEN CHRONIC BRONCHITIS AND EMPHYSEMA Chronic bronchitis and emphysema frequently coexist, although one can be present without the other. A 1' ' c nucal continuum appears to extend from bronchitis at one end, through a mixture of the two conditions in the major- ity of cases, to emphysema at the other end (123). An alternative method of assessing the relationship is by study of patho- logical change. A 1 c ose relationship is found between chronic bronchitis and emphysema on purely morphologic grounds. Although emphysema occurred more frequently in patients with chronic bronchitis than could be accounted for by chance, the two conditions also occurred independently of one another (183). Three of the possible reasons why chronic bronchitis and emphysema are found in association more often than would be expected by chance are the presence of a common cause and causation each by the other. The protective mechanisms for the upper respiratory tract are cilia and a mucous sheath, and the lower respiratory tract mechanisms involve macrophages, the lymphatic system, and possibly the fluid lining of the alveoli. Although not yet proved, failure of the protective mechanisms of the upper respiratory tract might be expected to lead to chronic bronchitis and failure of the pro- tective mechanisms for the lower respiratory tract to emphysema, On this hypothetical basis, a common cause would not seem unlikely; noxious en- vironmental agents in gaseous or aerosol form would be likely to affect upper and lower respiratorv tracts simultaneously, perhaps with potentiation of the injury in the lower tract by particles. Several ways in which chronic bronchitis might cause or aggravate emphysema have been suggested: such as through trauma resulting from pressure changes induced in the thorax by cough (138) and by airway obstruction (114). Clinical evidence of bron- chitis preceded clinical evidence of emphysema in over 50 percent of cases in one continuing study (137). of chronic bron,chitis 153 I. Others suggest that emphysema may be a cause It seems likely that a common cause. causation of emphysema by chronic bronchitis, and causation of chronic bronchitis hy emphysema are all operatin g mechanisms, with varying importance in different populations and different individuals (123). 270 Evidence Relating Smoking to Chronic Bronchitis and Emphysema Experimental and pathological evidence bearing on the possible rela. tionship of smokin g to chronic bronchitis and emphysema has been pre- sented in an earlier section of this chapter. Epidemiological and clinical evidence relating smoking to these diseases will be considered here. EPIDEMIOLOGICAL EVIDENCE Chronic bronchitis and emphysema probably represent disorders of multi- ple causality. Such problems are particularly suited for analysis by the epidemiological method, especially with regard to the identification of causes and the disentanglement of their relations (140). Two types of studies, prevalence studies and prospective studies, will be considered. PREVALENCE STUDIES.-The most important epidemiological evidence available relating smoking to non-neoplastic respiratory diseases is found in the prevalence studies which concern the number of cases in a population at one point in time. The definitions and criteria for diagnosis of chronic bron- chitis and emphysema are not ideal for the purposes of these epidemiological surveys. The absence of standardized diagnostic methods in chronic bron- chitis and the non-specificity of clinical diagnostic criteria for emphysema have resulted in the use of prevalence of symptoms and signs of the respira- tory diseases under study as a basis for the surveys. Studies of the prevalence of chronic bronchitis and emphysema in the United Kingdom and in the United States over the last decade have developed highly reliable epidemiological methods. Because of the nature of the diseases in question, these surveys present results by the prevalence of specific symp- toms and signs, or combinations, rather than diagnostic labels of disease en- tities. Various levels or grades of severity of the symptoms or signs are defined and the data are obtained and handled in a standardized manner, permitting comparisons between different populations and communities; thus it becomes feasible to evaluate whether smoking is associated with cer- tain signs or symptoms to a greater extent than with other findings. ( 1. I Smoking and Respiratory Symptoms-( a.) Chronic Cough-The common phrase "smoker's cough" suggests that this symptom is popularly be- lieved to be associated with smoking. Several workers have investigated the relationship between smoking and cough; Table 1 lists surveys that tabulate the frequency of cough in smokers as compared with non-smokers. Several different types of populations have been surveyed; the purpose of presenting the findings together is to demonstrate the variation found among the differ- ent populations. The 1,456 mill workers studied by Balchum et al. ( 16) constituted the ran- dom sample of those who volunteered for chest X-rays and pulmonary func- tion tests. Of 1.198 smokers, 23.3 percent reported cough; of the 253 non- smokers, 10.2 percent reported cough. When the percentage of smokers re- porting cough is considered in each of several categories described by pack- years of smoking experience, a gradient was found for those reporting cough, ranging from 11 percent of those who smoked less than one pack-year of cigarettes up to 50 percent of the subjects with 60 or more pack-years of smoking experience. 280 TABLE I.-Summary of reports cm the predence of cough in relation to smoking Author I Year Numhrr of subjects Hrfrr- __- ence Smokers Non- ; smokers 23. R In. 2 31. 5 13. n n.6 4. 1 21. 2 78 20.6 1% 7 z 54. R qu 12 1 0 IR. 9 R. 3 64 IA 60 I 0 Boucot and others (251 considered the relationship in older men of smok- ing and chronic cough in a self-selected population 45 vears of age and older. Chronic cough was defined as cough existing for months or rears. Again. a considerably higher percentage of the smokers reported cough. and a clear- cut gradient was established according to amount of smoking. Bower (26) studied 172 men and women employed in a bank. This study is one of the few which included men and women working under similar con- ditions. Eighteen percent of 95 men and 17 percent of 77 women admitted to cough "more or less every day." Of the smokers, 27.6 percent admitted to daily cough (12 of 42 men, 9 of 34 women), whereas 4.1 percent of non- smokers admitted to this symptom (0 of 13 men, 2 of 36 women). Densen and others (44) presented findings in transit and postal employees. Persistent cough was reported by 21.2 percent of 2:530 smokers and 7.8 per- cent of 514 non-smokers. Fletcher and Tinker (67) studied male workers aged 30 to 59 in the British General Post Office and in the London Transport Executive. In the G.P.O., 18.7 percent of 166 smokers reported cough during the whole of the day in the winter, compared with none of 10 non-smokers. Among smokers of the L.T.E., 20.6 percent of 272 admitted to a comparable cough pattern whereas none of 30 non-smokers described such a cough pattern. Flick and Paton (68) in a study of patients excluding those with cardiac and respiratory disorders, found 55 percent of 157 smokers admitted to habitual cough compared with 10 percent of 51 non-smokers. After the first hundred patients, the admission to the study was weighted in the older age groups. The questioning was not as standardized as in some of the more recent surveys. Olsen and Gilson (148) ~ in their study comparing findings in population samples in Britain with those in Denmark, found cough in 32.1 percent of 162 British smokers and in 18.9 percent of 132 Danish smokers: the cor- responding figures for non-smokers was 0 percent of 11 and 8 percent of 24. Schoettlin (173) studied a group of veterans in a domiciliary and medi- cal-care center, mostly in the age group 45 to 74. The results for cough !"constantly present for two years or more") are presented in terms of 281 years of smoking. although the original figures were not published and are not included in Table 1. By recalculation, it appears that of those who smoked more than 10 years, 43.9 percent of 2,153 subjects had cough whereas 18.0 percent of 718 u-ho had smoked less than 10 years had cough. In the population samples quoted thus far. the percentage of smokers admitting to cough ranged from 17.3 percent to 55 percent, whereas the range for non-smokers was 0 percent to 13.0 percent. Two other studies show a considerably lower prevalence of cough both among smokers and non-smokers in two unusual types of population. Short and others I 176) reported the frequency with which unselected policyholders admitted to cough on periodic health examination, a time when they would be expected to minimize their symptoms. Of 1,292 smokers, 6.4 percent admitted to cough whereas 1.6 percent of non-smokers admitted to cough. In a study of a parachute brigade, Liebeschuetz (120) found 6.0 percent of 83 smokers and none of 52 non-smokers admitted to cough. The study of members of this unit with particularly high fitness standards was con- ducted at the time of discharge. Hammond ( 82) has presented the frequency of cough in smokers and has compared this with the frequency of cough among non-smokers. The subjects were asked to state whether they had a cough at the time of the questionnaire. They were also asked the question: "Have you had a cough over a period of many years?" They also were asked to estimate its severity as slight. moderate, or severe. The analysis of complaints has been reported so far for 43,068 questionnaires, 18,697 for men and 24,371 for women. For each age group and for both sexes, cough was significantly more common among those who smoked cigarettes. The percentage with cough (and the percentage with more than a slight cough ) increased rapidly with the num- her of cigarettes per day in both sexes and in all four age groups. Except for ex-smokers. the relationship between "chronic cough" and smoking habit was very much the same as the relationship between "present cough" and smoking habits. The proportion of male smokers with the complaint of cough \vas almost three times as ,areat as might have been expected on the basis of cough prevalence among non-smokers. For women: the ratio of observed-to-exl:ected smokers w-ith the complaint of cough was 2.5 to 1. The ratio of ohserved-to-expected numbers complaining of cough "more severe than slight" \\as 4.09 for males and 2.74 for females. The difference in frequent!- of the comf)laint of cough or of cough "more severe than slight" bet\\een smokers and non-smokers is statistically significant at the 0.001 level. The study sample was not a random sample of the population, but it pro\-idrs information about the relationship between smoking and various complaints for larger numbers of subjects than does any other study. The results again make it clear that a larger proportion of cigarette smokers are aware of couph than are non-smokers. In earh of the surve\s. smoking \cas found to be associated with the e\mptom of rough defined in a variety of ways. The studied populations varied considerably-from hospital patients. workers in dusty trades and clean offices. urban and rural population samples to members of a parachute brigade. Despite the diversity of these groups. it is surprising to note the consistency of the difference between smokers and non-smokers in regard 282 to cough. In each of the surveys, a larger proportion of the subjects ad- mitting to cough were smokers and about twice the proportion of smokers admitted to cough as non-smokers. (b.) Sputum.-Table 2 lists surveys in which the frequency of sputum pro- duction has been tabulated separately for smokers and non-smokers in preva- lence surveys. Most of the studies were considered in the section on cough and in Table 1. It is interesting that in most of these studies non-smokers report sputum production more frequently than cough. TABLE 2.--Summary of reports on the prevalence of sputum in relation to smoking London Transport .._._.........._.._. ~... ._ 1961 Post Offic% . . . . . .._......_.... -.~ _... ~.._ 1961 Flick..... .__..._._. ..__...._........_...... _ 1959 Olsen: ( Numhw of suhjrctn United Kingdom . . . . . . . . . . .._. . . .._..... ~- Denmark.-.........--.....~.-..-...-.--...- 1 Percentages standardized for age. T'wcfnt with sputum 30.4 34. 2 21.9 11.1 204 13 R 1 4". 3 ' lY.8 1 13. R 19.4 16. 9 7 0 1% i 10 II 64. 7 24. 5 27. 7 11.4 - 0 R. :3 Ferris and Anderson (61) studied a sample of the population of a town; their results are presented as percentages: standardized for age. The sample sizes were 542 males and 695 females. Among males 40.3 percent of smokers and 13.8 percent of non-smokers admitted to sputum production with the corresponding figures for females being 19.8 percent for smokers and 9.4 percent for non-smokers. Thus, sputum production in each of the diverse populations was found associated with smoking and a consistent difference between smokers and non-smokers was present in regard to sputum production. (c.1 Cough and Sputum.-The closely associated symptoms of cough and sputum have been combined in the results of a number of epidemiologic sur- veys. Table 3 shows the prevalence of cough and sputum in smokers and in non-smokers among samples studied. Of particular interest is the series of comparisons made by Higgins and his colleagues (88, 90, 92. 93, 95)) on samples drawn from contrasting pop- ulations, selected for their different backgrounds. Lapse rates were low, and a high degree of uniformity was achieved in the collection of informa- tion. In the disparate groups studied-including male and female subjects, older and younger, and varyin g in degree of dust exposure and exposure to rural or urban environment-the consistent direction and extent of the dif- ference between prevalence rates in smokers and non-smokers demonstrates a strong relationship hetween smokin g and productive cough in a variety of different situations. and the predominance of smoking as a determinant of these symptoms. 281 TABLE 3.-Summary of reports on the prevalence of cough and sputum i,, rehion to smoking Author i year (153) 1.400 (l.3) (156) .iz WC 91 ;:3 43 I 83 --I---- Non- 1 Smokers / x,,,, smokws 1 i SInok;.,, -__- __ --. 1: flf 33 ifi 364 1,468 451 46 81 52 - 23. 9 17.2 : I 4 !, 24.0 30.0 (, :i , 29.8 Ii 1 29.1 L. 44.7 ',4 :i , 11.0 6. 0 1 .* I (8 51.0 2.11 23.1 4. i 1% 6 4 !4 7. 2 u The percentages of symptoms noted by Oswald and Medvei (1501 are unusually high because occasional cough or sputum is included, in addi- tion to more frequent or persistent symptoms. The results are not shown in Table 3, which considers only smoking and cough with sputum; among males, 63.7 percent of 2,617 smokers and 47.7 percent of 985 non-smokers in Oswald and Medvei's study had cough or sputum. Among females, 63.2 percent of 970 smokers and 47.7 percent of 1.272 non-smokers admitted to either or both of these symptoms. Payne and Kjelsberg (153) presented data on respiratory symptoms, lung function, and smoking habits in the adult population of Tecumseh, Michigan, where a comprehensive epidemiological study is being made of the entire community. Cough and sputum were graded in severity as Grade I or Grade II, the latter being defined 3s both cough and phlegm, of which at least one was present throughout the day for three months in the year or longer. The prevalence of Grade II sy-mptoms is noted in Table 3. Dur- ing an interview period continued for 18 months, authors were able to show that the prevalence of symptoms did not vary significantly with the season of the year. Cough and sputum at the Grade II level were admitted to by 11 percent of 1.400 cigarette-smokin, m males, and 2 percent of 364 non- smoking males. The corresponding figures for females were 6 percent of 888 smokers and 2 percent of 1,X% non-smokers. These Grade II symptoms increased in prevalence with advancing age in men, and in women up to 49 j ears. It is interesting to note that lesser degrees of cough and sputum, classed 3s Grade I symptoms. showed little change in frequency after 19 years of age in either sex. In both sexes, Grade I symptoms of cough and sputum uere considerably more prevalent among smokers than among non- smokers-45 percent of 1,400 smokers and 19 percent of 364 non-smokers among the males, and 29 perucnt of 888 smokers and 17 percent of 1,468 non- smokers among the females. Phillips and his associates (156) studied two groups: one of male em- ployees in a steel-making plant, examined as part of an industrial hygiene program, and containing sub-groups with different types of industrial ex- posure, and a second group consisting of 300 patients in a Veterans Ad- ministration Hospital who were chosen at random, except for exclusion of cases of specific pulmonary diseases such as tuberculosis or tumor and cases of congestive heart failure. Chronic cough was defined as daily cough with sputum for a period of one year or more. Various possible environ- mental factors-geographic area, air pollution, specific work environment. and smoking-were considered. Fifty-one percent of 823 cigarette smokers were recorded as having cough, and 2 percent of 451 non-smokers. In a tabulation of chronic cough by age in decades, for cigarette smokers and non-smokers, it was shown that the increasing prevalence of chronic cough with age was much greater in the cigarette-smoking group. Read and Selby (159) in a mixed group of 302 subjects, some of them clinic patients, some patients' friends, and some hospital staff, found that male smokers admitted to cough or sputum ten times as often as did male non-smokers, and to cough and sputum five times as often. In their female subjects the ratios for these categories were eight to one and four to one. Liebeschuetz (120) in his study of parachute brigade members found. as might be expected, a much lower proportion of subjects with cough and sputum; these do not include subjects previously noted in Table 1 as having cough alone. Considering these surveys as a group, it appears that the presence of cough, sputum, or the two symptoms combined, is consistently more frequent among smokers than non-smokers, in a variety of samples drawn from populations differing so widely in other respects that this association ma) be taken to be a general one. TABLE 4.-Summary of reports on the prevalence of breathlessness in relation to smoking Refer- Number of subjects Smokers 1.19R 2. 530 272 166 222 93 is 20 315 z 1.400 Rx4 1. 292 Non- jmokers 14. 5 25.3 R. 5 9. 0 19. R 9. 7 9% 42 i 285 Some of these surveys are limited in one respect. and some in another. The degree to which bias has been avoided varies; several of the survevs quoted are open to criticism in this regard. but in others considerable pains have been taken to avoid any possihility of suggesting a relationship which mav not trulv exist. It would be Mrong to extrapolate from. say. a hospital population to the general public. but the groups surveyed vary enough that the evidence demonstrates clearly that cigarette smokers more often report symptoms of cough. sputum. or both. than do non-smokers. t d. I Breathlessness.-Table 4 summarizes the prevalence of breathlessn?ss as reported in surveys of various populations. Balchum and others (16) in their survey of mill workers. reported a greater prevalence of breathlessness among the smokers in their sample, Tabulation of the frequency of this complaint by pack-years of smoking experience showed a less smooth gradient than for prevalence of cough and sputum. Densen and others 144), who studied respiratory symptoms in transit workers and postmen in New York City. found that 25.3 percent of 2,530 smokers and 16.9 percent of 514 non-smokers admitted to breathlessness of Grade II or worse iindicated by positive answers to specific questions on the questionnaire). Fletcher and Tinker (67)) in a study of Transport Executive employees and Post Office employees, had only one non-smoker out of 40 complain of breathlessness. and 38 smokers out of 438. These figures are for workers complaining of dyspnea (a positive answer to the question, "DO you have to walk slower than most people on the level?" or "Do you have to stop after a mile or so on the level at your own pace?"). In the four studies by Higgins listed in the table. the difference in prevalence of breathlessness between smokers and non-smokers is more variable. In his study 188) in the agricultural district of the Vale of Glamorgan, the author bresents prevalence figures for the various symptoms among females in two age groups. those under age 45, and those over age 45. His reason for doing so is the considerable difference in frequency of the smoking habit between women in these two age-groups. In both the age groups of females, the prevalence of breathlessness is greater among the non-smokers. but the difference is not statistically significant. Female smokers in the over 45 age groups have rather more cough and sputum and wheeze than the non-smokers. but apparently have less breathlessness. In his study in Annandale (93 1 the prevalence of breathlessness among all men and all women studied MBS greater in the non-smokers than in the smokers. although the numbers of non-smoking men and of smoking women were small. When males aged 55 to 64 are considered. from the three surveys 1901, breathlessness is more prevalent among the smokers. and the same thing applies to the t\\ o different age groups of males studied in Staveley (92 I. Payne and Kjelsberg i 153 I : in their survey of a total community, ha\-e stated that among the men, cigarette smokers were affected more often with breathlessness at all ages. Among the women, cigarette-smokers had a higher prevalence of breathlessness than non-smokers below the age of 40, and above this age the non-smokers had a higher prevalence. Considering all ages together. twice the proportion of male smokers admitted shortness of breath 286 compared to non-smoking males; the prevalence of shortness of breath among females was the same for smokers and non-smokers. Short et al. (176)) in a study of answers to a questionnaire on routine medi- cal examination for insurance purposes, obtained a larger percentage of com- plaints of breathlessness amon g smokers than among non-smokers. Hammond 182) also presents figures for the frequency with which breath- lessness was noted in answer to a questionnaire by 18.697 men and 24.371 women. The relationship between breathlessness and smoking is less clear than the relationship between cough and smoking. A significantly greater proportion of complaints of breathlessness was encountered among male and female cigarette smokers, both for total complaint of breathlessness and complaint of breathlessness "more severe than slight." The ratio of ob- served-to-expected complaints of breathlessness among male smokers was 1.97 for the total number with this complaint, and 2.62 for those complain- ing of breathlessness more severe than slight. The ratios for females were 1.36 and 1.49. A consideration of the frequency of complaints of shortness of breath in smokers and in non-smokers, by age group and by sex, shows that the excess of breathlessness among cigarette smokers is greater and more consistent for men than for women. The older age groups of women show only a slight excess. Thus, the relationship between smoking and the symptom of breathless- rrcss is less general than the relationship between smoking and cough or sputum, which is found in all age-sex groups in a variety of different pop- ulations. For males the association is clear: male cigarette smokers com- plain of breathlessness more often than do non-smokers, particularly in the older age groups. Females present a less uniform pattern. In several sur- veys, females show a higher prevalence of breathlessness in non-smokers than in smokers, particularly in the older age-groups. The reasons for this sex difference have not been explained. (e.) Smoking and Chest Illness.-The percentage of smokers and non- smokers who reported chest illness in the three vears prior to the interview TABLE 5.-Summary of reports on history of chest illness in the past 3 years in relation to smoking I .-__ _-- .~~ 114-422 o-64-20 date is presented in Table 5. For men, the prevalence was consistently higher among smokers, and in one study (93), the association of smoking and chest illness was apparent for the younger (25-34) as well as the older males (5% 6.11. For female smokers and non-smokers, the prevalence of chest illness was about the same. cf.) Combinations of Symptoms.-A number of prevalence studies (7,54, 61, 62. 77, 150 ) hav-e reported results, either totally or in part, under diag nostic headings which cannot be translated into single symptoms. The symptom combinations and the names applied to them varied; some of the studies gave the percentages of smokers and non-smokers with "any" signs or symptoms rather than specified combinations. The results are presented in Table 6. TABLE 6.-Su.mmary of reports on the prevalence of combinations of certain symptoms in relation to smoking Smokers .___~ (7) 3.214 (54) 779 I::{ 340 209 NOD Smokers NOW smokers smokws -- 677 524 21.7 10.3 29.4 19. 5 ' 24.9 ' 7.3 ' 17.5 ' 9.4 42. 6 15. 0 20.0 10.0 43.0 31. 4 Per@ant with symptoms 16. 1 9. 7 15.4 9.1 Ashford and his colleagues I 71 found twice the proportion of "respira- tory symptoms" among Scottish coal mine workers who smoked than among those who did not smoke. "Respiratory symptoms" were regarded as pres- ent in those who have caugh or sputum all day for more than three months per year and walk slower than others on the level, or wheeze, or if the weather affects their chest. or if they have had a chest illness in the last three years. Those who had wheeze and who claimed the weather affected their chest were also classed under "respiratory symptoms." Edwards and others (541 I-resented the percentage of smokers and non- smokers with bronchitis. according to clinical assessment by one of 11 general practitioners coooerating in the survey. No attempt to standardize the diagnosis was reported. Of ii!, smokers. 29. I percent had "bronchitis" compared with 19.5 percent of 524 non-smokers. Ferris and Anderson (61: presented the prevalence of "irreversible ob- structive lung disease." which was defined as the report that wheezing or whistling in the chest occurred most days and nights, that the subject had to stop for breath when walking at his own pace on the level, or had a forced expiratory volume in the first second of expiration ( F.E.V. 1.0) of less than 00 percent of the total forced expiratory volume. According to this defi- nition. male smokers showed a 24.9 percent prevalence of irreversible 288 obstructive lung disease, compared with 7.3 percent of male non-smokers. The corresponding percentages for females were 17.5 percent and 9.4 per- cent. These percentages were age-standardized. ln a study conducted in a flax mill, Ferris, et al (621 presented the prev- alence of "chronic respiratory disease," defined as productive couEh on four days of the week, for three months of the year: for three successive years; or wheezing in the chest most days and nights; or breathlessness. of Grade III or more, in the winter; or asthma diagnosed by the physician at the time of the survey; or F.E.V. 1.0 less than 60 percent of forced vital capacity. rnder this definition, 42.6 percent of 54 male smokers and 15.0 percent of 20 male non-smokers had "chronic respiratorv disease." For females. the figures were 10.0 percent of 10 smokers and lo.0 percent of 60 non-smokers. Goldsmith and others c 77'). in their study of longshoremen. classified the subject as having a "respiratory condition" if he had ever had asthma or bronchitis, or currently was "troubled by constant coughing." With this definition, 43.0 percent of 1:238 moderate or heavy smokers had a respira- tory condition, compared with 31.4 percent of 744 non-smokers. Oswald and Medvei (1501, defining "bronchitis" as disability from acute exacerbations of chest symptoms, or breathlessness, or both, found a prel-- alence of 16.1 percent among 2,617 male smokers. and of 9.7 percent among 985 non-smokers. In their female subjects, 15.4 percent of 970 smokers compared with 9.1 percent of 1,272 non-smokers had "bronchitis." Although these various combinations of symptoms are not comparable. the consistency and extent of the differences between prevalence of symp- tom combinations in smokers and non-smokers are striking. (g.j Relationship between Symptoms or Signs and Amount Smoked.-In several surveys, smoking categories were based on the daily consumption or total lifetime consumption (16, 61, 67, 82. 90, 153, _ In the majority. the Prevalence of cough and sputum increased with amount smoked. A recent study (82) showed that those who smoked cigarettes of low nicotine content tended to cough less than those who smoked cigarettes of high nicotine con- tent. Other symptoms and measurements of pulmonary function show a less clear relationship between prevalence and amount smoked. fh.) Rebztionship bettceen Symptoms and Signs and Method of Smokinp.- The numbers of pipe and cigar smokers in many prevalen,ce studies are so small that conclusions about the effects of these methods of smoking are not reliable, but they all tend to show that pipe and cigar smokers are likely to be intermediate between non-smokers and cigarette smokers in prevalence of Symptoms and signs. (i.) Vent&tory Function-Pulmonary tests and the method of presenting results, though varyin g widely, are important features of the prevalence surveys. ln the study by Ashford and others 171 of 4,014 coal miners, the forced exPiratory volume in the first second of expiration (F.E.V. 1.0'1 of non- smokers was slightly higher than that of the smokers. and a small but sta- tistically siunilicant difference was found even after correction for differ- ences attributable to physique. No consistent relationship was reported between the amount smoked and the average F.E.V. 1.0. 289 Balchum and others (16) reported that 19.3 percent of 1,194 srnoken and 7.8 percent of 243 non-smokers had an "abnormal" test, an F.E.Vr. l.O of less than 70 percent. When the "abnormal" test was compared \,ith the number of pack-years of cigarettes smoked, a steady increase in tl,,. proportion of men with decreased F.E.V. 1.0 was found with inereasirlz pack-years. Ferris and Anderson (61) showed a progressive decrease in the ,,,,,a,1 F.E.V. 1.0 in successive age groups for male smokers, male non-smoker, and female non-smokers. In males. there was also a regular decrease i,, F.E.V. 1.0 within each age group with increase in the number of cigarrtt,., currently smoked. In females, there was little difference in the F.E.V. l.(, between smokers and non-smokers except in one age group. The peak expiratory flow rate showed a decrease with age and a decrease within th, age groups w.ith cigarette smoking. Chivers (36) showed that smoking. age. and height were correlated "ia. nificantly with the expiratory flow rate. The older and shorter men ha,1 greater impairment associated with smoking. Flick and Paton (68) demonstrated a distinct decline, beginning at about 40 years of age. in expiratory flow rate among smokers, but no appare,,t change among non-smokers until 70 years of age. Fletcher and Tinker (67), measuring expiratory flow rates by the Peak Flow Meter, found one group of smokers, but not another, had lower value. than the non-smokers. In a later paper (58). Fairbairn, Fletcher and Tinker reported that the Peak Flow Meter appeared to be a less satisfactory screen. ing test than the forced expiratory volume. Franklin and Lowell (73), in a study of 1,000 apparently healthy factor, workers, found the mean expiratory flow rate during the third quarter 01 maximal forced expiration to be approximately 20 percent less in "heal\ smokers" than in "light smokers." "Heavy smokers" were defined as thos, who had smoked 30 pack-years or more, and "light smokers" less than 111 pack-years. Higgins (88) showed a decrease in F.E.V. 0.75 among smokers of 15 grams or more of tobacco per day. compared with non-smokers and with those who smoked less than 15 grams a day. For this test, there was ncl significant difference between non-smokers and the lighter smoking group. Peak how measurements indicated a difference between heavy and light smokers. and also between nonsmokers and light smokers. In each lo-year age group over 45, the peak flow was lower in smokers than in non-smokers. but the numbers were small. Th ese differences are not explained by differ. ences in age, social class, or occupation. The difference between smoker5 and non-smokers in peak flow measurement was not seen in tests of women. Higgins ( 90) summarized the difference in F.E.V. 0.75 in a variety of different samples of the population. Tabulations for 16 different groups included miners and ex-miners in varying pneumoconiosis categories and non-miners in the same district, and agricultural workers in two different areas in Britain. In the 13 group.. s in which comparisons were feasible. non-smokers recorded a higher F.E.V. 0.75 than the smokers. The small over-all difference in means was recorded (as indirect Maximum Breathing Capacity) as 50 liters per minute, which was significant at the one percent level. By pooling subjects with different occupations in the older age groups, differences between light and heavy smokers were apparent, though not statistically significant. Higgins commented on a strong trend in the prevalence of persistent cough and sputum, with amount of tobacco smoked. without a significant trend in ventilatory capacity. His possible explanation of the difference is that smokers are more likely to give up smoking or re- duce the amount smoked, once their lung efficiency becomes impaired. than they are when their only symptoms are cough and sputum. In their study of miners and foundry workers in Staveley I 92 j. Higgins and his colleagues showed a decrease in the F.E.V. 0.75 in smokers. Non- smokers, light smokers. and heavy smokers ( 15 grams per da!- and over I ranked in that order for decreasing F.E.V. 0.75, both in men aged 25 to 3-l and in those aged 55 to 61. The d ff i erence between the non-smokers and the light smokers was smaller than the difference between the light and the heavy smokers in the younger age group; in the older age group the dif- ference was larger between non-smokers and light smokers. Olsen and Gilson ( 148) measured the F.E.V. 0.75 in a sample of a pop- ulation in Denmark for comparison with British population samples. Cig- arette smokers had a lower mean F.E.V. 0.75 than cigar smokers or pipe smokers who in turn had a higher mean than non-smokers. but these differ- ences were not statistically significant. If non-smokers. cigar smokers, and pipe smokers are grouped-together, non-cigarette smokers had a significantly higher mean F.E.V. 0.75 than the cigarette smokers. Payne and Kjelsberg (153)) who presented mean values of F.E.V. 1.0 for men and women by age group and by smoking category, found a lower mean value for cigarette smokers than for non-smokers in each age group of men Over 19. In the 16-to-19 ape group. cigarette smokers had a slightly higher mean value than non-smokers. A comparison of the mean values by age group for non-smokers and for cigarette smokers shows a decline with advancing Years in both, but more rapid in the cigarette smokers. Women also show a decline of F.E.V. 1.0 with advancing vears. but this is no more marked and no more rapid in the cigarette smokers than in the non-smokers. The reduction in F.E.V. 1.0 in cigarette smokers amounted to 7 percent and :< percent of the mean values in non-smoking men and women respectively when values adjusted to the over-all mean age of 40 years were compared. Read and Selby (159) measured peak flow rates in smokers with cough. and in smokers with cough and sputum. To a statistically significant extent. male smokers without cough or sputum showed a more rapid fall in peak ilow rate with age than expected. Male smokers with cough showed a still more rapid fall with age. and those with cough and sputum. the most rapid fall. Amount smoked had no obvious effect. Results were similar for women. Revotskie and his colleagues i 165), who grouped smokers in Framingham as never smoked, light smoker, medium smoker, and heavy smoker. found that the F.E.V. 1.0 measurements show a gradient from never smoked to heavy smoker in the "normal" subjects, both for males and females: in the other groups this gradient is not clear. The "Puffmeter" ratios tended in the same direction. but in less clear-cut fashion than the F.E.V. 1.0 measurements. 20 I Goldsmith and others (77) showed that smokers, regardless of amount smoked. have a slight diminution in the pulmonary function test results, Eden in the absence of respiratory symptoms. The total vital capacity was murh less sensitive in this regard than the F.E.V. 1.0 or the "Puffmeter" Longshoremen with "respiratory conditions," and particularly th reading, 0%~ with shortness of breath, had a more marked decrease in pulmonary function Cough was associated with the greatest diminution of pulmonary function measurement. The relationship between cigarette smoking and abnormal results of pul. monary function tests is more difficult to evaluate from the published SUf,.p,., than is the relationship between symptoms and cigarette smoking. pL,I. monary function test results are influenced by several factors, among which are age, physique: and perhaps occupation. When allowance is made for these factors. there appears to be a clear difference in the ventilatory funr. tion between smokers and non-smokers. In the majority of prevalence surveys, the subjects were not forbidden to smoke pcor to pulmonary function testing. Since acute alterations due to smoking might be mix. interpreted as due to a permanent abnormality, it is important to examine the magnitude and significance of the acute effects of smoking on pulmonary function. Rickerman and Raracb (2Oj found no consistent alterations in vital capacity or in maximum breathing capacity before and after their patients and normal subjects smoked three cigarettes. Simonsson (177) found a small decrease in the F.E.V. 1.0 in 13 ,,f 16 young subjects after smoking, and the difference for the group was statistically sic. nificant. No significant change was found in the total capacity. Several authors have studied more sensitive tests of airway resistance and lung co,,,. pliance. Eich. Gilbert and .4uchincloss (56) made compliance and airway resistaac? measurements, using an esophageal balloon technique, on a group of nine healthy ad& five of whom had respiratory symptoms. No difference was detected after one cigarette. In a group of emphysematous patients, a statistically significant increase in airflow rp. sistance was found, but without significant change in compliance. Attinger and others (8) reported no statistically significant difference in expiratop airflow resistance or compliance, but in a later study of subjects with pulmonary diseasr. significant physiological changes-increased mechanical resistance and increased work of breathing--were noted after smoking one or two cigarettes. Motley and Kuzman (142) studied the lung volumes, spirometry, blood gas exchaner. and pulmonary compliance in 141 subjects, before and after smoking two cigarettes. Not all of these measurements were made on all subjects. There was no significant change in the mean values of vital capacity performed after smoking, some subjects showing a d errease, and others an increase. Six of the normal subjects showed a decreased com- pliance after smoking. In 33 subjects with cardiac or respiratory disease, 17 had a sip. nificant decrease in romplianrr after smoking. The authors felt that a decrease in pul. monary romplianre was the only notable abnormality which followed smoking acut+ Forced rxpiratory volume and airflow rcsiktance studies were not included. Miller i ]34a), who constructed pressure-volume work loops, demonstrated increased airflow rP of the same age, and the ventilatory function of non-smokers in all agr groups was significantly higher than that of the smokers. Among smokcn of 50 years of age and above, the prevalence of pneumoconiosis tended to bP lowest among the men who smoked the most and highest among men whr, smoked the least. However, the prevalence of pneumoconiosis was higher in ex-smokers than among smokers and non-smokers, except in the oldest age group, suggesting that men with pneumoconiosis tend to reduce their tobacco consumption. The possibility that factors of selection eliminatp some persons with symptomatic pneumoconiosis from study groups should also be considered in the evaluation of these studies. In a sample of 11317 men aged 40 to 65 who worked in a variety of non. dusty and dusty environments, a greater prevalence of bronchitis (dail! cough for at least the preceding six months, productive of one teaspoon of sputum per day) was found in moderate and heavy smokers 127). Betweerl the non-smokers and the heavy smokers, a significant difference was found at all age levels, and also between non-smokers and moderate smokers except in the oldest age group. Although effects from dust exposures could be noted. it appeared that cigarette smokin, 01 was the dominant etiologic factor in "chronic bronchitis" in this selected group. Among alkaline dust workers it was found that the dusts in the working environment did cause some increase in respiratory illness but the sig. nificance of the dusts in the production of respiratory disability, either functional or pathological, was not as important as the number of cigarettes smoked daily (36). 298 In a study of 1.274 steel workers. non-smokers had a comparatively low incidence of chronic cough, regardless of their job classification or condi- tions of work or residence. Th ere was a direct relationship between chronic cough and the number of cigarettes smoked daily in each occupational category (156) . Cigarette smoking was of greater importance in deter- mining the prevalence of chronic cough than was the occupational exposure. In a study of New England flax mill workers, 161 subjects were subjected to a questionnaire and measurements of pulmonary function to determine the presence of "chronic non-specific respiratory disease." The prevalence of such a syndrome. based on a certain combination of symptoms or signs. was related to age. sex. smoking habits. years of exposure to dust. and estimated inhaled quantity of dust. The effect of smoking "far out-shadows any effect due to age or occupational exposure to dust" (62). The studies by Higgins and his colleagues (87. 88. 89. 91. 92 I show that smoking and occupational exposure are both related to the prevalence of chronic respiratory disease but do not allow quantitative assessment of their relative importance in the populations defined. As this series of studies was undertaken to demonstrate any effect from industrial exposure. and the popu- lations surveyed were such that exposure to occupational dusts was more varied than in the general population. the importance of the effect of smoking in this group of studies on the production of respiratorv symptoms is rather convincing (1231. The authors comment in one of the papers in this series: "So important is the influence of tobacco smoking that it is essential to allow for differences in smoking in comparable groups before drawing `conclusions about the importance of other factors." In a recent study of bituminous coal miners (103,) ex-smokers had pul- monary function results and prevalence of respiratory symptoms comparable to those of non-smokers; no impairment was attributed to pure pipe or cigar smoking. Cigarette smokers had the most symptoms of respiratory disease and. except for vital capacity, they had the lowest pulmonary function. The authors comment: ". . . although smoking definitely impairs pulmonary function, the impairment of pulmonary function by years worked under- ground is clear and separate from the effect of smoking." In a study of 7,404 metal mine workers, aged 35 years and older. a com- parison was made of the effects of 20 years' aging and smoking on pulmonary- ventilation, as measured by the F.E.V. 1.0 in individuals without X-ray evi- dence of silicosis. A decrease of 23 percent occurred with the process of aging 20 years. For heavy smokers (those who smoked for 25 years or more and now smoke more than 20 cigarettes a day 1, there was an additional de- cline of 10 percent over that of aging alone. "The decline in pulmonary function associated with heavy smoking was equivalent to the decline that comes about by the process of aging 10 years. For the entire group of metal mine workers, the reduction in pulmonary function assmiated with smoking was equivalent to half the effect of heavy smoking, or about five years of aging" (128). The population at risk from occupational exposure is relatively small com- pared to the population of cigarette smokers. Among occupational groups, cigarette smoking is an important variable that must be considered in all 299 studies of rhronic hronchopulmonary disease. the relative importance of cigarette smokin, In most studies, but not all, u is greater than occupational ex. posures in the production of symptoms and signs of chronic bronchitis ur emphysema. SUMMARY Tobacco smoke is a heterogenous mixture of a vast number of compounds. several of which have the ability to produce damage to the tracheobronchial tissues and lung parenchyma. Retention of inhaled cigarette smoke particles in the respiratory system of man is about 80-90 percent complete with breatl, holding of two-to-five seconds. Particles penetrate deeply into the respira. tory tract and are deposited on the surface of the terminal bronchi&. respiratory bronchioles, and pulmonary parenchyma. Little information j, available concerning the specific toxic properties of the particulate phaW components. Gas phase components probably have a diffuse though not uniform pattern of distribution. It seems likely on the basis of the physical characteristics of gas absorption and distribution, that a substantial portioa is retained along the upper bronchial tract. Certain of the gases known to be present in cigarette smoke are capable of producing pulmonary damaPe in experimental animals and man. Cigarette smoke produces significant functional alterations in the upper airways. Like several other agents, cigarette smoke can reduce or abolish ciliaty motility in experimental animals. Post-mortem examination of bronchi from smokers s.hows a decrease in the number of ciliated cells, shortening of the remaining cilia, and changes in goblet cells and mucouY glands. The implication of these morphological observations is that func. tional impairment would result. Cigarette smoke is also capable of interference with functions in the lower airways. In animal experiments, cigarette smoke appears to affect the phy. sical characteristics of the lung lining layer and to impair alveolar stability. Alveolar phagocytes ingest tobacco smoke components and assist in their re- moval from the lung. This phagocytic clearance mechanism decompensates under the stress of protracted high-level exposure to cigarette smoke and tn- bacco smoke components accumulate in the pulmonary parenchyma of experimental animals. The acute effects of cigarette smoking result in an increase in airway rr. sistance but clinical expression of this change in pulmonary function is ncft common. The chronic effects of cigarette smoking upon pulmonary func- tion are manifested mainly by a reduction in ventilatory function as measurer! by the forced expiratory volume. Histopathological alterations occur as a result of tobacco smoke exposure in the tracheohronchial tree and in the lung parenchyma of man. Changes regularly found in chronic bronchitis-increase in the number of goblet cells. and hypertrophy and hyperplasia of bronchial mucous glands-are more often present in the bronchi of smokers than non-smokers. In experimental animals, cigarette smoke consistently produces significant functional altera. 300