`I'ABLE l&--Continued Author, year of study, Number and type location, reference of p0pu1at10n Fmdings Khmla. 1964. Port Talbot, United Kingdom (1971) Schlesinger et al., 1966. 7,701 males employees Adjusted mean FEV, level (liters) in the ateel industry Never amokem /- Current smokers i 15 cigarettealday lb24 cigarett&day 25-34 cigar&tea/day 2 35 cigarettes/day 4.331 male civil servants, Mean value of the FEV,/FVC ratio Israel (1972) aged 45 or older Nonsmokera Jhmokem Current smokers 1-19 cigarettes/day - 3.70 3.57 3.48 3.41 3.37 - 76.0 74.3 73.9 2 20 cigarettes/day 72.7 Keateloot et al., 1968-1969, Belgium (2976) O'Donnell and de Hamel. 1969- 1970, New Zealand (1976) Linn et al., 1973, San Fran- cisco and Los Angeles, U.S. (1976) Rndelman et al.. year not stated. Baltimore. U.S. (1966) 4,961 males in the Belgian military, aged 15 to 59 1.079 male public sewante, up to age 65 644 male and female office workem aged 17 to 60 410 male volunteers, aged2otQ103 By multiple regression, FEW, reduced by 0.14 liters in smokers of 1-19 cigarettes daily and by 0.23 liters in smokere of 20 or more daily Beduoed mean FEV, in smokers of 10 or more cigarettes daily; increaeed prevalent of FEX, below 80 percent of predicted in smokers of more than two pa& daily By analysis of covariance, significant reduction of FJXV, in smokers compared with nonsmokers By partial regression analysis, significant reduction of FEW, in current and former cigarette smokem c 0 e TABLE 10.~-Continued Author. year of study, Number and type location, reference of population FlllilllgS Woolf and Suero, year not stated. Toronto (1971) Krumholz and Hedrick. year not steti, Dayton. U.S. (1973l 298 female volunteers employed at commercial film, aged 25-54 227 male executives. aged 3.544, selected to include nonsmokers (n = 136) and long- term smokers (n=91) Adjwted mean levels Nonsmokers Exsmokers Current smokers 70 cigarettes/week 71-140 cigarettes/week 2 140 cigarettes/week Mean values Nonsmokers Smokers EV, FEV,/FVC ratio 2.65 a.7 2.64 85.0 2.63 86.2 2.50 85.1 2.45 84.1 ___- EV, FEV,IFvC 3.80 71.3 3.42 73.6 Grimes and Hanes. year not stated, Los Angeles, U.S. (1973) Lefcoe and Wonnacott. year not stated. western Ontario, Canada ( 1974 1,059 male and female insurance company employees 1,072 males in four occupational groups By multiple regression, significant reduction of FEV, level in male smokers but not in female smokers By multiple regression. significant reduction of FEY, in current cigarette smoken Higgenbottam et al.. year not stated, London. England (1980) 18.403 male civil sewants. aged 4ota64 Reduced FEV, io cigarette smokers compared with nonsmokers, increased effect with increasing daily amount in current smokers (Table lo), even though people with symptomatic airflow obstruction may be likely to retire from their jobs. Recently, predictors of the incidence of airflow obstruction have been examined with multivariate techniques in data from popula- tion samples in Tecumseh, Michigan (Higgins et al. 19821, and in Tucson, Arizona (Lebowitz et al. 1984). In Tecumseh, the strongest predictors of airflow obstruction (defined as an FEVl less than 65 percent of predicted) were age, the number of cigarettes smoked daily, changing smoking habits, and the initial FEVl level (Higgins et al. 1982). The addition of other variables to the predictive model did not greatly improve its validity. In Tucson, these same variables, along with certain symptoms and illnesses, and skin test reactivity were significant predictors (Lebowitz et al. 1984). During the 10 years of followup of a population sample in Finland, incidence cases of chronic airflow obstruction (defined as FEV1/FVC ratio less than 60 percent) were observed only in those who continued to smoke (Huhti and Ikkala 1980). These studies of incidence highlight the importance of cigarette smoking in t.he etiology of airflow obstruc- tion; new cases are rare among nonsmokers. Dose-Response Relationships Dose-response relationships between FEVl level and the amount of cigarette smoking have been described with simple descriptive statistics and further characterized by multiple regression analysis. In cross-sectional data, the FEVl level varies inversely with the amount smoked. Although the variation in mean FEVl levels among strata of smoking appears clinically unimportant, the distributions of values in smokers and nonsmokers are quite different (Figure 4). Cigarette smokers more often have abnormal lung function, regard- less of the criteria applied to the population (Mueller et al. 1971; Knudson et al. 1976; Burrows et al. 1977a; Detels et al. 1979; Rokaw et al. 1980; Beck et al. 1981). This increased prevalence of abnormal function is a result of the skewed distribution of function in smokers, with a subgroup of the smokers showing a large decline rather than the entire group shifting by a small amount (Figure 4). As noted in this reference, however, there are decreasing numbers of smokers with FEVI above the mean for nonsmokers as pack-years increase, suggesting that all smokers are probably somewhat affected, even though only a minority eventually develop clinically significant airflow limitation. In several populations, the relationship between cigarette smoking and FEVl level has been examined in greater detail. Burrows et al. (1977a) used linear multiple regression analysis to examine the relationship between cigarette smoking and ventilatory function in a population sample in Tucson, Arizona. Pack-years, a cumulative- dose measure, was the strongest predictor of FEVi level among the 103 smoking variables considered. In currently smoking men and women, the FEVi declined by approximately 0.25 percent of the predict.ed value for each pack-year of cigarette smoking; the effect was of a similar magnitude in ex-smokers. Using data from three separate U.S. communities, Beck and colleagues (1981) assessed the importance of six separate smoking variables: amount smoked daily, use of filters, inhalation, age started, age stopped for ex-smokers, and cumulative pack-years. For the FEVi, the strongest predictors in male current smokers were the duration of smoking and the amount smoked; in female current smokers, only pack-year was statistically significant. The number of years of cessation was associated with FEVl in male but not in female ex-smokers. However, in both the multiple regression analysis reported by Beck et al. (1981) and that reported by Burrows et al. (1977a), the measured cigarette smoke variables accounted for only about 15 percent of the variation of age- and height-adjusted FEVi levels. Unmeasured aspects of cigarette smoking, other environmental exposures, and the characteristics of the smokers must contribute to the unexplained variation. A role for the type of cigarette smoked has not yet been established (USDHHS 19811, and the impact of differences in depth or pattern of inhalation and other aspects of the pattern of smoking remains to be investigated; they are discussed in more detail in the chapter on low tar and low nicotine cigarettes. Further studies of these aspects of cigarette smoking are needed to monitor the consequences of changing cigarettes. Factors Other Than Cigarette Smoking A number of risk factors other than cigarette smoking have been postulated as contributing to the development of airflow obstruction (Table 7). Of these, a definite role for a,-antitrypsin deficiency has been established, but only the small number of persons with homozygous deficiency incur markedly increased risk (Morse 1978). The current hypotheses on susceptibility to cigarette smoke postu- late roles for childhood respiratory illnesses (USDHEW 1979; Burrows and Taussig 1980; Samet et al. 19831, for endogenously determined hypersensitivity of the lung, and for other genetic and familial factors (Speizer and Tager 1979; USDHHS 1980aJ At present, these hypotheses remain largely untested. The data are similarly incomplete at present for the other factors listed as putative risk factors in Table 7. The status of each is briefly reviewed below. ABH Secretor Status Secretion of ABH antigens is a genetically determined trait that follows an autosomal dominant inheritance pattern; approximately 104 70 to 80 percent of the population excrete antigen into the body fluids (Cohen et al. 1980a). In a genetic-epidemiology study in Baltimore, Maryland (Cohen et al. 1980a), ABH nonsecretors had lower levels of FEVJFVC ratio and a higher proportion with FEVJFVC ratio below 69 percent. Studies in France (Kauffmann et al. 1982a, 1983) and in England (Haines et al. 1982) have confirmed reduced expiratory flow rates in ABH nonsecretors. In contrast, ABH secretor status did not predict the development of obstructive airways disease in the Tecumseh, Michigan, population (Higgins et al. 1982). Air Pollution Although exposure to air pollution at high levels may exacerbate the clinical condition of persons with chronic lung disease, a causal role for air pollution in the development of airflow obstruction has not been established (Tager and Speizer 1979; USDHHS 1980b). However, smoking is the major predictor for chronic airflow obstruction in areas of high as well as low atmospheric air pollution. Airways Hyperreactivity Orie and colleagues in the Netherlands (Orie et al. 1960) speculat- ed that bronchial hyperreactivity and allergy may predispose to asthma and chronic bronchitis. Findings from two small longitudinal studies have suggested that airways reactivity may influence indi- vidual susceptibility to cigarette smoke. Barter and colleagues followed 56 patients with mild chronic bronchitis during a 5-year period (Barter et al. 1974; Barter and Campbell 1976). The rate of decline of FEVl increased with the degree of airways reactivity, as measured by reversibility with isoproterenol or responsiveness to methacholine. Britt et al. (1980) measured change of FEVl in 20 young adult male relatives of patients with chronic obstructive pulmonary disease. The decline of FEVl was approximately five times larger in the nine subjects with a positive methacholine challenge test. In patients with clinically diagnosed airflow obstruc- tion, airways reactivity is also associated with more rapid decline of lung function (Kanner et al. 1979). Because airway reactivity would affect the FEV, directly as well as possibly influence the susceptibili- ty to smoke, it is difficult to ascertain from these data whether the relationship between airway reactivity and COLD is direct or spurious. Alcohol Consumption The epidemiological data on alcohol consumption are conflicting. A study of former alcoholics demonstrated an excess prevalence of lung function abnormalities, including airflow obstruction (Emergil 105 480-144 0 - 85 - 5 and Sobol 1977). In the Tucson population, alcohol consumption was a significant predictor of ventilatory function after the effect of smoking was controlled (Lebowitz 1981). The findings of an investiga- tion in Yugoslavia were similar (Saric et al. 1977). However, two large U.S. investigations did not demonstrate adverse effects of alcohol intake (Cohen et al. 1980b; Sparrow et al. 1983a). Cross-sectional data from the Tucson population suggest increased susceptibility to cigarette smoke in atopic people (Burrows et al. 1976). In subjects aged 15 to 54, the prevalence of an FEVAWC ratio below 90 percent of predicted value increased with skin test reactivity among both smokers and nonsmokers. Subsequent reports from this same study have not confirmed an overall relationship between FEVl level and atopy, but indicate that atopy may predis- pose to airfIow obstruction in a subset of the population (Burrows et al. 1977a, 1983). Burrows and coworkers (1981) also reported an increased level of IgE in smokers independent of their allergy skin test reactions, and the interrelationship of these factors is currently being examined. Childhood Respiratory Illness In a longitudinal investigation of 792 English working men, Fletcher and coworkers (Fletcher et al. 1976) found a cross-sectional association between childhood illness history and FEVl level. The decline of FEVl level during the study's longitudinal phase was not correlated with childhood illness variables. In contrast, a.nalyses of cross-sectional data from a population sample in Tucson suggested that childhood respiratory illnesses may increase susceptibility to cigarette smoke (Burrows et al. 1977b). In this population, people with a history of respiratory trouble before age 16 demonstrated excessive decline of ventilatory function with increasing age and with increasing cigarette consumption. Familial Factors Familial aggregation of lung function level, adjusted for age, height, and sex, has been demonstrated in populations in the United States and elsewhere (Higgins and Keller 1975; Tager et al. 1976; Schilling et al. 1977; Mueller et al. 1980). However, a recent report suggests that the familial aggregation of lung function may be a reflection of the familial aggregation of body habitus (Lebowitz et al. 1984). Relatively modest correlations of FEVI level have been demonstrated between siblings and between parent-child pairs. The role of familial factors is further supported by investigations demonstrating increased prevalence of airflow obstruction in rela- 106 tives of diseased subjects (Kueppers et al. 1977; Tager et al. 1978; Cohen 1980). This familial factor cannot be explained by familial resemblance of a,-antitrypsin phenotype or of ABH secretor status (Kueppers et al. 1977; Cohen 1980). In the Tecumseh population, however, family history of airflow obstruction did not predict the incidence of this disease. The results of twin studies are also consistent with genetic influences on FEVl level and suggest that genetic factors may influence susceptibility to cigarette smoke (Webster et al. 1979; Hankins et al. 1982; Hubert et al. 1982). Occupation Several population-based investigations suggest that occupational exposures other than those recognized as causing lung injury may have some effect on lung function level. In Tecumseh, mean age and height-adjusted FEVl scores in men were highest in farmers and lowest in laborers; the differences were not explained by smoking and were present in nonsmokers (Higgins et al. 1977). Similarly, in Tucson, men reporting employment in certain high risk industries or exposure to specific harmful agents had a higher prevalence of abnormal lung function (Lebowitz 1977a). In a Norwegian case- control study, men employed in workplaces characterized as polluted were at increased risk for clinically diagnosed emphysema (Kjuus et al. 1981). Longitudinal studies of industrial populations also show that occupational exposures may increase the rate of decline of FEVl (Jedrychowski 1979; Kauffmann et al. 1982b; Diem et al. 1982). For example, Kauffmann et al. (1982b) found that FEVl change during a 12-year period varied with job exposures in an employed industrial population. Effects of dust, gas, and heat were present, as was evidence for a dose-response relationship between increasing exposure and a greater rate of decline. In these studies, however, smoking effects were generally much greater than the occupational effects. Passive Exposure to Tobacco Smoke Passive exposure is discussed in detail elsewhere in this Report. Respiratory Illnesses In an 8-year followup study of London men, chest infections were not associated with a rate of FEVl decline (Fletcher et al. 1976). The findings of several smaller longitudinal studies were similarly negative with regard to respiratory infection (Howard 1970; John- ston et al. 1976). It is now apparent that mucus hypersecretion and airflow obstruction are separate pathophysiological entities that have a common cause-cigarette smoking (Fletcher et al. 1976; Peto et al. 1983). 107 Socioeconomic Status Weak effects of socioeconomic status on lung function level have been demonstrated in community samples in Tecumseh (Higgins et al. 1977) and in Tucson (Lebowitz 197713). In both populations, lung function appeared to be influenced independently by socioeconomic status indicators, even after controlling for cigarette smoking. In the Tecumseh study, FEVI increased slightly with increasing income and education level (Higgins et al. 1977); in the Tucson study, the proportion of people with an abnormal FEVl varied in a similar pattern with these indices (Lebowitz 1977al. Effects of socioeconomic status were present in nonsmokers in both investigations. Stebbings (19711, in a sample of nonsmokers in Hager&own, Maryland, also demonstrated an association between lung function level and socioeconomic status. In summary, there is evidence that a number of factors other than cigarette smoke may influence lung function, but the influence of these factors is small relative to the effect of smoking, and the major question is whether they can influence susceptibility to cigarette- induced lung injury rather than whether they, of themselves, result in lung disease in nonsmokers. Development of Airflow Obstruction At this time, the natural history of airflow obstruction has been only partially described; a population has not yet been followed from childhood to the development of airflow obstruction during adult- hood. However, the available data from separate investigations cover the entire course of the disease and support the conceptual model proposed in Figure 15. With aging, measures of function begin to deteriorate after age 25 to 30. In nonsmokers without respiratory disease, cross-sectional data generally show that the FEVl declines by 20 to 30 ml per year (Dickman et al. 1969; Morris et al. 1971; Cotes 1979; Crapo et al. 1981). Longitudinal data have been confirmatory (Tables 11 and 12). For example, Tockman (19791 measured the FEVi loss during an 8 year period in 399 male nonsmokers. In most, the FEVl declined at 25 ml annually; a few, with an initial FEVl lower than 2.5 1, lost 34 ml annually. Sufficient excessive loss leads to the development of airflow obstruction. However, many questions remain unanswered concern- ing this process of functional deterioriation. It is unclear whether the loss always occurs uniformly or if it develops in stages with intermittent and relatively steep declines (Bates 1979; Burrows 1981). The concept that the decline is nearly always gradual receives strong support from the findings of the &year longitudinal study conducted by Fletcher and coworkers (1976). In this investigation of 108 TABLE IL-Association between cigarette smoking and longitudinal change in lung function in selected population samples Author, years of study, Number and type location, reference of population Findings Higgins and Oldham, 1954-1959 Rhondda Fach. Wales (196.9~ 253 male miners, ex- miners, and nonmining controls Annual decline of indirect maximal breathing capacity (liters/min) Miners, ex-miners controls without pneumoconiosis Nonsmokers 1.6 0.8 Es-smokers 0.7 1.8 Current smokers 1-14 g/day 1.3 1.7 115g/day 1.6 2.2 Ashley et al.. 195S1968, Framingham. U.S. 11975) 399 men and 636 women. aged 37 to 69 in 1958 lo-year change in FEV,IFVC ratio (agestandardized to overall distribution for each sex) Men Nonsmokers 0.21 Gmtmued smokers -1.3 stopped. 1953-1963 0.51 Women -3.6 -4.1 -4.6 Higgins et al.. 1957-1966. 594 men, aged 25-34 Annual decline of FEVw, (ml/year) bv age and smoking in 1957 &ely. -- England (1968b) or 55-64 in 1957 2534 ;rs 5Gl yrs Nonsmokers 21 32 Exsmokem 29 44 Current smokers 1-14glday 37 54 2 15gida.v 38 37 E TABLE 1 l.-Continued Author, years of study. Number and type location, reference of population Huhtl and Ikkala. 196-1971. Harjavalta. Finland (1980) Wilhelmsen et al.. 1963-1967. Goteborg. Sweden (1969) 492 men and 671 women, aged 40 to 64 in 1961 313 men, aged 50 in 1963 Annual decline of FEV, (ml/year) Nonsmokers Ex-smokers Continued smokers Stopped. 196-1971 Annual decline of FEV, (ml/year) Nonsmokers Ex-smokers Current smokers I-llg/day 2 15glday StQDDd. 1963-1967 Men 33 45 44 51 43 33 70 70 40 Women 27 27 39 35 _ -- _ -. .,"II***.u~u Author. years of study. Number and type locatlon. reference of population Findings Oxho] et al.. 1963-1973, 269 men. aged 25 ctgslday 54 Annual decline of FEV, iml/year). adjusted for initial level Nonsmokers 40 Ex-smokers 44 Current smokers c 15 g/day 46 ? 15 g/day 51 5-year decline of FEV, as percent of mean, by 1973 smoking Nonsmokers 3 Ex-smokers 5 Current smokers 7 Annual decline of FEV, (ml/year) I.9821 - Nonsmokers Ex-smokers Continued smokers stopped, 1967-1977 37 39 49 48 E TABLE 12.-Continued Author. years of study, Number and type locatton. reference of population Findings Woolf and Zamel. years not gtven. Toronto. Canada r IY8O)xoI BOW et al, lYfxLlY68 to 1YtwlY74. Boston. II s ( I.WII 302 female volunteers, aged 25 to 54 at entry H50 male volunteers 5.year change m FEV, as percent of initial value Nonsmokers 15 Rx-smokers 0 8 Smokers <. 70 clgslweek 0.4 71-140 ctywveek `): 3.t,, 140 cigs! week 48 Annual declme of FEV, imllyearr. adjusted for age and mtttal level Nonsmokers 0.053 Ex-smokers 0.057 Current smokers 0085 Love and Mtllcr. 1957 to 1973 1.677 male coalmmers II-year declme tn FEV, Ilttersl Iaverage followup. 1 I years). UnIted Ktngdom ~l.Y&?b Nonsmokers 0.41 b-smokers 0.48 IntermIttent smokers 0.52 Current smokers 0.53 792 employed men, the individual patterns of temporal change of the FEVl were strongly variable, but the loss generally occurred gradually. Fletcher et al. further demonstrated that FEVl levei correlated with FEVl slope, a finding that they termed the "horse- racing effect." Correlation between slope and level would be antici- pated, if functional loss occurs gradually. This correlation has important implications for intervention; those losing FEVl more rapidly should become identifiable early as they develop a reduced FEVl level. Other studies, however, do not agree with either the pattern of FEVl decline or the "horse-racing" effect. Rapid declines to levels compatible with clinical disease or followed by a prolonged plateau have been described (Howard and Astin 1969; Howard 1970; Johnston et al. 1976). In a followup study of Canadian men with chronic bronchitis, steep declines of FEVl without subsequent improvement were frequently observed (Bates 1973). Additionally, correlation of FEVl level and slope has been. found in most other longitudinal investigations (Howard 1970; Petty et al. 1976; Huhti and Ikkala 1980; Bosse et al. 1981; Clement and van de Woestijne 1982; Kauffmann et al. 1982b), but not in all (Barter et al. 1974; Krzyzanowski 1980). Another unanswered question concerning functional deterioration is whether gradual decline occurs in a linear or a nonlinear fashion (Fletcher et al. 1976). Sufficient numbers of people have not yet been followed to distinguish alternative patterns, although the available data indicate acceleration of the decline with aging (Emergil et al. 1971; Fletcher et al. 1976). In spite of these uncertainties concerning the development of airflow obstruction, the available data indict cigarette smoking as the primary risk factor for excessive loss of FEVl (Tables 11 and 12). The findings in both general population samples (Table 11) and occupational and volunteer cohorts (Table 12) have been similar. Recent reports from Belgium (Bande et al. 1980; Clement and van de Woestijne 1982) and from Connecticut (Beck et al. 19821, not readily summarized in tabular form, also described a strong effect of smoking on FEVl decline. A few studies have not shown increased loss in cigarette smokers (Howard 1970; De Meyere and Vuylsteek 1971). Even in people with clinically diagnosed airflow obstruction, continued smoking maintains the excess decline of FEVl (Hughes et al. 19821, although not all findings are consistent (Ogilvie et al. 1973; Johnston et al. 1976). Dose-response relationships have been found in many investiga- tions between the amount smoked during followup and the FEVl decline (Tables 11 and 12). The reported increases from the lowest to the highest smoking categories range up to 10 to 15 ml annually. Although this additional loss in heavier smokers appears small, if sustained for long periods of time it would shorten the time interval 115 Never smoked or not suscepbbk? Its effects - Stopped at 45 Death 25 50 75 Age (years) FIGURE X-Risks for men with varying susceptibility to cigarette smoke and consequences of smoking cessation NOTE: + = death. SOURCE: Fletcher and Pet0 (1977). to the development of functional impairment. So far, favorable effects of filter tip smoking and declining tar content on the rate of decline have not been shown (Fletcher et al. 1976; Sparrow et al. 1983b). Generally, sustained smokers experience a greater loss than those who stop during followup. In the study by Fletcher et al. (1976) of London men, subjects who stopped smoking at the beginning of the followup period lost FEVI at the same rate as never smokers. The results of two U.S. studies of ex-smokers are similar (Bosse et al. 1981; Beck et al. 1982). This reduced loss in ex-smokers emphasizes the importance of active smoking and the immediate benefits of smoking cessation (Figure 24). Smokers with reduced FEVI may be protected from developing clinically significant loss by timely smoking cessation (Fletcher and Peto 1977). The distribution of FEVI decline has been characterized and described for some populations, including patient groups (Burrows and Earle 1969; Howard 1974; Barter et al. 19741, population samples (Milne 19781, and occupational cohorts (Howard 1970; Fletcher et al. 1976). Similar data are also available for the mid-maximum expira- tory flow, another measure of ventilatory function (Bates 1973; Woolf and Zamel 1980). In each of these investigations, the distribu- tion of FEVI decline is unimodal (Figure 25); that is, a distinct population with more rapid decline is not sharply separated from those with lesser rates. The modes and medians of the distributions 116 60 70 Mean 60 Nonsmokers and Q a-smokers ii 50 e 6 i 40 30 1 20 10 0 -160 -140 -120 -100 -60 -60 40 -20 0 +20 FIGURE 25.-Distribution of &year FEVl slope in 792 London men SOURCE: Fletcher et al. (1976). are generally negative, but some subjects have had positive slopes during the relatively brief followup period of investigations conduct- ed up to this time. The distributions tend to be skewed by subjects losing FEVl more rapidly. The proportion of cigarette smokers is increased among those in the tail of excess loss (Figure 25). For example, Clement and van de Woestijne (1982) examined subjects with excess FEVl decline in a prospective study of 2,406 members of the Belgian Air Force. Losses beyond those expected from nonsmokers affected 6 percent of nonsmokers, 7.5 percent of light smokers (< 20 cigarettes/day), and 12 percent of heavy smokers ( > 20 cigarettes/day). The shape of the distribution of FEVl decline has important implications for the development of airflow obstruction. Smokers are not sharply separated from nonsmokers (Figure 251, but more often lose FEVl at a rapid rate. Because of this spectrum of severity, not all smokers develop significant airflow obstruction. Although the fac- tors that lead to excessive loss in individual smokers remain uncertain, they may include differences in the pattern of smoking. It is apparent, however, that this susceptible minority can be protected by smoking cessation. 117 summary During the 20 years that have elapsed since the 1964 Surgeon General's Report, the relationship between cigarette smoking and airflow obstruction has been intensively investigated. Surveys of community samples and other groups have established that airflow obstruction is a common condition in the United States and elsewhere. In some populations, as high as 10 percent of adults are affected. Determinants of lung function level and of the prevalence of airflow obstruction have now been examined in many populations throughout the world. Cigarette smoking is the strongest predictor of abnormal measures of ventilatory function. A causal relationship between cigarette smoking and airflow obstruction is supported by the consistency of the many published reports, the strength of the association, and the evidence for dose-response. Many risk factors for airflow obstruction other than cigarette smoking have been postulated, including other harmful environmen- tal exposures and the inherent susceptibility of the smoker. Homozy- gous a,-antitrypsin deficiency can explain only a minute proportion of the disease burden. The development of airflo-w obstruction by only a minority of smokers indicates that the interaction of smoking with other factors may influence the risk for specific smokers. Current research emphasizes the potential roles of childhood respira- tory illness and airways hyperresponsiveness. Longitudinal studies have now partially described the prolonged natural history of airflow obstruction. Excessive loss of ventilatory function, beyond that expected from aging alone, results in the development of disease in cigarette smokers. Only a susceptible minority of cigarette smokers lose function at a rate that will eventually cause clinically significant impairment. For this group, timely smoking cessation can prevent the development of disease. 118 EMPHYSEMA JA-oduction Pulmonary emphysema is frequently present in the lungs of individuals with chronic obstructive lung disease. This section has three purposes: (1) to review the definition, types, and quantification of emphysema; (2) to summarize the physiological and radiographic feature of emphysema; and (3) to discuss critically the relationship of smoking to emphysema, based upon observations in people and in experimental animals. Current concepts of the pathogenesis of emphysema are reviewed elsewhere. Definition of Emphysema The generally accepted definition of emphysema is an anatomic condition of the lung characterized by abnormal dilation of air spaces distal to the terminal bronchioles accompanied by destruction of air space walls (American Thoracic Society 1962; Heard et al. 1979). Difficulties with this definition have been discussed by Thurlbeck (1983). Normal air space dimensions have not been determined, and criteria of destruction have not been defined. These limitations hamper attempts to investigate the earliest lesions of emphysema and the subtle effects of environmental agents on lung structure. Types of Emphysema British pathologists pointed out in the forties and fifties that emphysematous lesions in certain people involved the respiratory bronchioles, which appeared as grossly enlarged airspaces in the center of the primary lung lobules surrounded by normal lung. In other individuals, the alveolar ducts were involved early, and even mild involvement appeared grossly as a coarsening of the architec- ture of the entire lobule. They designated the two polar patterns of emphysema as centrilobular emphysema (CLE) and panlobular emphysema (PLE) (Heppleston and Leopold 1961). Many lungs either show both types of emphysema or are unclassifiable. Of 122 lungs with emphysema examined by one pulmonary pathologist, 73 were -considered mixed or unclassifiable and 49 were clearly CLE or PLE (Mitchell et al. 1970). When the agreement of three pathologists was required, only 27 of the original 122 lungs remained classifiable and 95 were mixed or could not be classified. There were no statistically significant differences between the groups classified as PLE or CLE in any clinical variables. The only nonsmokers in either group had CLE, and the proportion of light smokers (less than 25 pack-years) was very similar between groups. In this study and others (Anderson and Foraker 19731, CLE was most severe in the upper lobes and PLE was uniformly distributed. According to Thurlbeck (19761, a common 119 combination is CLE in the upper lobes and PLE in the lower lobes; where lobectomies are used for correlation, typing of emphysema is therefore a particularly empty exercise. When emphysema is far advanced, it is often impossible to recognize the site of the initial involvement. Thus, it is not clear whether the differences in prevalence of CLE and PLE are real or represent differences in interpretation by different observers. Several localized types of emphysema occur in areas around scar tissue (paracicatricial), along interlobar and interlobular septa (paraseptal), and as bullous lesions (which represent the most advanced and extreme distortion of normal lung structure). Bullous deformities occur with any type of emphysema, including CLE and PLE. Occasionally, bullous lesions occupy huge intrapulmonary volumes. Detection of Emphysema The detection of emphysema requires suitably prepared lung specimens. At a minimum, this means the lung must be fued in inflation (Thurlbeck 1964). Fume fixation or fixation by instillation of liquid fixative through the airways is satisfactory, but for optimal evaluation of the latter group, barium impregnation or paper- mounted whole-lung sections should be used. Because lungs with emphysema frequently also have some degree of intrinsic airways disease, the severity of emphysema and the clinical state of the patient may not correlate directly. Pathologists can easily recognize mild degrees of emphysema that are rarely associated with clinical disability. Quantification of Emphysema There are a number of techniques for quantifying the volume of lung involved with "obvious" emphysema that are adequately reproducible and correlate well with one another (Thurlbeck 1976; Bignon 1976). Semiquantitative or subjective scoring methods as well as point counting have been used. These approaches all require lungs inflated to a relevant volume, usually one approximating total lung capacity during life. This can be achieved by a distending pressure of 25 cm H& (Thurlbeck 1979; Berend et al. 1980). In the scoring method, the lung is divided into a number of units and the severity of emphysema in each unit is scored (mild, moderate, or severe receive 1,2, or 3 points, respectively). The scores for each unit are summed to give a total score for the lung (Ryder et al. 1969). Alternatively, lung slices may be matched by visual comparison to a set of graded standards to achieve an emphysema score (Thurlbeck et al. 1970). These methods include both severity and extent of emphysema, and although they involve subjective judgments, they have proved to be remarkably reproducible. 120 In the point counting approach, regularly spaced points are superimposed on a lung slice. Each point is recorded as falling on normal parenchyma, emphysematous parenchyma, or nonparenchy- ma (conducting airways or vessels). The volume proportion of emphysematous lung is recorded. This method can be objective (e.g., if an emphysematous space is taken to be one greater than 1 mm in diameter), but it includes only extent and not severity of emphyse- ma. Morphometric methods carried out on histologic sections, exempli- fied by the mean linear intercept (Lm) (Thurlbeck 1967a, b), are strictly objective, but they require careful attention to problems of sampling and are time consuming and insensitive to focal disease. For measurements of the Lm, histologic sections are made of blocks selected by stratified random sampling. The average distance between alveolar walls is determined from the number of intersec- tions of alveolar walls with a line of known length. The internal surface area of the lung can be calculated when the volume of the lung is known (Hasleton 1972). Pulmonary Function in Emphysema Because unequivocal proof of the presence of emphysema requires direct examination of lung tissue, the strategies used to characterize the pulmonary function abnormalities associated with emphysema have either involved comparison of functional data collected during life with autopsy or surgical material or have used measurements made exclusively on post-mortem specimens. Two important conclu- sions from these studies should be noted at the outset. First, impaired air flow during maximal expiratory maneuvers, as reflect- ed in reduced values for the FEV1, FEVIS, and FEF~~s, is neither sensitive nor specific for emphysema. It is possible to have severe emphysema without clinical obstructive lung disease (Thurlbeck 1977). It is also possible to have severe chronic obstructive lung disease without having emphysema, even though most patients with advanced chronic obstructive lung disease have some degree of emphysema (Mitchell et al. 1976). Second, none of the tests used to identify early obstructive lung disease, such as closing volume, the single breath NZ curve, or frequency dependence of compliance, distinguish diminished elastic recoil that may be related to emphyse- ma (see below) from increased resistance in small airways (Buist and Ducic 1979). Even the determination of density dependence of maximum expiratory airflow, once felt to be specific for detecting abnormalities in the caliber of small airways, is not immune to the effects of lung elastic recoil. A decreased effect on maximal expiratory air flow of using low density gas can be caused by decreased elastic recoil (Gelb and Zamel 1981). 121 Pulmonary function testing of individuals with proven emphyse- ma often shows increases of residual volume, functional residual capacity, and total lung capacity and decreases of maximal expirato- ry air flow (Boushy et al. 1971; Park et al. 1970; reviewed in Kidokoro et al. 1977). However, because individuals with emphysema commonly also have intrinsic airway disease (Casio et al. 1978) affecting the results of these pulmonary function tests in the same direction as emphysema, it is clear that these tests are not specific for emphysema. Accordingly, there has been interest in other, more distinctive tests. Among readily applicable tests, the diffusing capacity has proved to be directly related to the extent of emphyse- ma (Park et al. 1970; Boushy et al. 1971; Berend et al. 19791, presumably reflecting a diminution of internal surface area avail- able for gas exchange. The usefulness of the diffusing capacity to identify and estimate emphysema is limited, however, because the measurement is not sensitive to low grades of emphysema (Symonds et al. 1974) or specific for emphysema. Moreover, the results must be interpreted carefully in smokers because the values for diffusing capacity are lower than in nonsmokers, and the difference extends even to young smokers who are not likely to have emphysema (Enjeti et al. 1978; Miller et al. 1983). Mechanical Properties of the Lungs in Emphysema Measurements of the pressure-volume characteristics of the lung have generally been regarded as a reliable means of physiologically detecting and quantifying emphysema because (al patients with emphysema often have increased lung distensibility and correspond- ingly low transpulmonary pressures (loss of elastic recoil) and (b) the severity of emphysema has seemed to correlate with the change in elastic recoil. It has also been assumed that the regions of lung with emphysema are the cause of the decreased lung elastic recoil, an assumption that appears reasonable because elastic recoil results in part from surface forces at the air-liquid interface and there is less surface area in emphysema. Recent observations challenge these concepts. Berend and Thurl- beck (19821, using lungs obtained post mortem, could not demon- strate a relationship between indices of lung elasticity and the grade of emphysema in 48 lungs ranging in grade from 2 to 80 (on a scale of 100), and observed (Berend et al. 1981) in emphysematous lungs that the relative increase in compliance of the lower lobes was greater than the upper lobes, even though the emphysema was worse in the upper lobes. Others have also reported poor correlations between emphysema and elastic recoil. Silvers et al. (1980) found decreased elastic recoil and increased total lung capacity in excised human lungs with minimal emphysema, and Schuyler et al. (19781 noted in hamsters given small doses of elastase intravenously that there was 122 decreased lung elastic recoil at low lung volumes, although the lungs did not show morphometric changes. Guenter et al. (1981) noted that mild emphysema produced by pepsin caused greater changes in lung elasticity than similar degrees of lung destruction produced by endotoxin-induced repetitive leukocyte sequestration. They suggest- ed that these differences may be due to differences in the location of the connective tissue injury within the lung. Even among those who have reported an association between emphysema and elastic recoil, the correlations have been best when the emphysema was severe (Greaves and Colebatch 1980). Pare et al. (1982) found a correlation between emphysema grade and elastic properties of the lungs in 55 persons; however, in 5 whose surgically removed lung tissue received emphysema scores between 20 and 70 (out of a maximum of lOO), the elastic properties of the lungs tested preoperatively were indistinguishable from normal. While such discrepancies probably reflect the limitations of relating the overall elastic properties of both lungs to the morphology of a single lobe, it must also be recognized that the sensitivity of the pressure-volume diagram is limited, since a narrow range of pressure (to 20 cm HzO) depicts the average retractive force from millions of air spaces and the connective tissue network of the lung. From these recent findings it must be concluded that the relationship between elastic recoil and morphologic measures of emphysema is not highly predictable, and that the decrease of elastic recoil and increase of total lung capacity commonly seen in emphysematous lungs may not result entirely from abnormal mechanical properties in the areas showing emphysema. The mechanical abnormalities may also derive from areas that appear normal, although the possible reasons for this are obscure (reviewed by Thurlbeck 1983). An alternate explanation for this discordance between elastic recoil and morphologic emphysema may be the problems of sampling and grading intrinsic to these morphologic measures. The work of Michaels et al. (1979) introduces a further complexity to the use of pressur*volume curves as an indicator of emphysema. They found that inhalation of a bronchodilator shifted the curve of smokers in the direction of increased compliance, but had no effect in nonsmokers (Figure 26). Cessation of smoking had the same effect as a bronchodilator. These results were interpreted as indicating that smoking causes some peripheral airway units to constrict and become effectively closed. Thus, pressure-volume studies to detect early changes compatible with emphysema in smokers may give false negative results unless accompanied by studies with bronchodi- lators. 123 90 60 70 60 50 0 / I - .smoken p 6.0. [3--o Ncrwnckers M Nonsmokers p B.D I I I 1 4 6 12 16 P(stat)cmH& FIGURE 26.-The effect of nebulized bronchodilator on the pressure-volume characteristics of the lungs in 19 smokers (6 men and 13 women) and 16 nonsmokers (9 men and 7 women) NOTE: The mean age was approximately 40 years (range. 19 to 66) and smokers wed approximately 30 cigarettes per day. Male amoken, showed borderline significant differences in indicea of expiretory airflow and single breath Nz test data as mmpsred with the male nonsmokers. but there was no diRerace in these testa between female smokers and nonsmokers. As shown. smokers had significantly laes elastic recoil than nonsmokers. After the bronchcdiletor, the difference between smokers and nonsmokers increased further. particularly at high lung volume. B.D. = broncodilator; % pred. TLC = percent predicted total lung capacity: Rstat) = transpulmonary preeeure. *p