Document EEykyXMq6g8DpE7jGmDpDbBx

cumulative and Reversible Effects of Lifetime Smoking on' Simple Tests of Lung Function in DOUGLAS W. DOCKERY, FRANK E. SPEIZER, BENJAMIN G. FERRIS, JR., JAMES H. WARE, THOMAS A. LOUIS, and AVRON SPlRO Ill Introduction T h e 1984 Surgeon General's Report on the Health Consequences of Smoking (1) once again documented the association between smoking and chronic obstructive lung disease. The accumulated evidence from many studies leads to the following conclusion with iespect to chronic obstructivelung disease morbidity (I, Pg 9). Within a few p a r s after beginning to smoke, smokers experiencea higher prevalence of abnormal function in the small airways than nonsmokers. The prevalence of abnormal small airways function increases with age and duration of the smokinghabit, and is greater in heavy smokers than in light smokers. These abnormalities in function reflect inflammatory changes in the small airways and often reverse with the cessation of smoking. It is clear from a number of investigations of young smokers that even short periods of smoking can lead to small but significant reductions in forced expiratory volumes (2, 3) and peak flows (4). In adults, there appears to be a direct relationship between number of cigarettes smoked and level of lung function (5,6)or rate of decline of lung function (7). This report providesan estimate of the degree of lung function impairment associated with smoking, developed through an analysis of adults participating in a longitudinal study of the health effects of air pollution. The analysis builds on a previous report that described the decline of lung function with age among the never-smoking,asymptomatic adults participating in the study (8). Methods Sampling The sampling plan for this study, often called the Six Cities Study, has been described previously (8,9). Briefly, historical air pollution data were used irl 1974 to select 6 communities representing a range of air pollution levels. The 6 communities are Watertown, MA; Kingston-Harriman, TN; Steubenville, OH; a geographically defined portion of St. Louis, MO; Topeka, KS;and Portage, WI. Two cit- 286 m,, -,< b SUMMARY Data from a random aampls of 8,lW men and women trOm 6 US.ClUsr .(.a b cu,fH a modeldercrlbingthe effecls of cumulativeand cumntcigarolW d n g On pulmony tlon. The data show that smoksn 8uft.r an Irfeverdbls IOUo,f FVC a d whlch h by a linear functionof their cumuiattw clgarottesmokingaa meaaumdin p.clysimmFor a- male173cmlall, the estimatedlossof FEV, is7.4 mlfor eachpack-year smoked. Foratypical 161 cm MI, the estimatedeffect Is 4.4 mi per packwar. Cumnt cigarotte MIolclng addr .n deficit over and abova the cumulatlw offed of Ilfetime unokhrg. For any IIMlm pldcmn, s m o k n hme higher levels of FEV,, 123 ml for a typical man, 107 mf for a tvpkrl woman, olr,do current smokers of a pack per day (p < 0.001).A man who 8tarta smoking one p.dc of m,per day at 25 yr of age would at age 60,after 35 p k - y e a n of expowm, haw an expoclod equal to that of a man 69.4 yr of age who had m r smoked. llhe stopped Mloklng alw r( age, his expected level would incroase lo that of a 66.5-yroid never-smoker.Thk model astimalea how much lung function Is irmvenlbly lost by smoking, estlmtsrhow much ccn# mgainsd with cessalion of smoking, and predict8 the futuro IOU of lUng fUmtiOn in both AM REV RESPIR D S 1- 1m-a ies were studied each year starting in September 1974. Adults 25 to 74 yr of age were selected at random in each community and invited to come to a central office to complete a standardized respiratorydiseasequestionnaireand to perform pulmonary function tests on a water-filled 8-L recording spirometer (9). Those not attending the central office were seen in their homes when possible. In all cases, spirometry was performed in the sitting position without a noseclip. Height was measured in stocking feet against a wall with a right angle FEV,and FVC were measured in a stan- dard fashion (10, 11) and corrected to BTPS. 1The questionnaire was administered y a trained interviewer. A detailed smoking is- tory was recorded for each subject. Separate histories were collected for cigarettes, pipes, and cigars. (Data on hand-rolled tobacco also was collected but has been combined in this analysis with the cigarette data, assuming 1 g of tobacco per cigarette.) For each individual. total years of smoking was calculated separately for cigarettes, pipes, and cigars, using the age of smoking cessation (orcurrent age for current smokers) minus the reported age of first smoking and the number of years of reported abstention from smoking. Participants estimated weekly smoking amount for each year they had reported being a smoker. These values were then accumulated to producea total consumption of cigarettes, pipes, and cigars. Total cigarette consumption is represented here as the sum of the packs of cigarettes per day over -years of cigarette smoking, Le., pack-- Pipe tobacco consumption is rep- analogously as ounces of tobaccoper dayyears smoked, and cigar consumptionas ber per day over years smoked. C i smokers were defined as those subjectsm. ing more than 25 lifetimepacks, with curran cigarette smokers defined as those (Received in originalform November I8,1966nd in revised jorm August 10. 1987) From the Department of EnviromentalSdaa and Physiology and the Department of B d tics, Harvard School of Public Health, and (kr, ning Laboratory, Departmentof Medicine, and Women'sHospital, Harvard Medical scbod Boston, Massachusetts. Supportedinpartby GrantsES-OIlOSaad~ oooO2from the National Institute of Envirorrm, tal Health Sciences,Contract No. RP-1001fnnUb, -Electric Power Research Institute, and Coopa, tive Agreement No. CR 811650 from the mental Protection Agency. This report has not been subjected to tbe vironmentalProtectionAgency's requiredp e a d policy review and therefore does not necess&* flect the views of the agency, and no o f f i d5 dorsement should be inferred. `Presented in part at theAnnua1Meetingof* American Thoracic Society, Anaheim, C& Ma 12-15, 1985. ' Requests for reprints should be addrrswd Dr. Douglas W. Dockery, Assistant Profacoc Department of EnvironmentalScienceand ology, Harvard Schod of Public Health. tington Ave., Boston, MA 02115. 7 Or: LIFETIME SMOKING ON LUNG FUNCTION IN ADULTS I. .\ . , tham one cigaretteper day for the last ! month. Pipe smokers were similarly defined athosesubjectsreportingmore than 25 life- I time ounces of tobacco, and cigar smokers more than 25 lifetime cigars. smokerswere asked "Do/did you inhalethe I smoke?" The response was graded in 4 categories: Not at all, Slightly, Moderd y , and Deeply. A similar question was TABLE 1 PREVALENCE OF ClGARETfE SMOKING AMONG WHITE ADULTS IN THE SAMPLE Men Women Never (n) (96) Former (n) (%) Current Never (n) (96) (n) (%I Total 1.001 27.0 1.277 34.4 1.436 38.7 2,302 51.4 Age, Yr asked for pipes and cigars. 25-34 257 34.5 146 19.6 343 46.0 405 47.2 112 13.0 342 I Smokers were also asked "Doyou smoke f*-tip cigarettescurrently?"The response msgraded in 4 categories: Never, Less than I balfthe time, More than half the time, and Ahnys Brandsof tobaccoproductswere not 35-44 45-54 55-64 65-74 187 29.3 216 22.6 194 23.6 147 26.5 169 26.4 341 35.7 355 43.2 266 48.0 282 44.2 398 41.7 272 33.1 141 25.5 329 44.4 106 14.3 575 48.3 193 16.2 523 51.7 178 17.6 470 69.6 91 13.5 306 422 311 114 1 raorded. 287 39.8 41.3 35.5 30.7 16.9 I Statistical Methods Reviously reported analyses of pulmonary function among the 2,454 healthy, white, i ncm-smokers participating in this study (8) showd that when FEV, and FVC measure- I ~ t aresstandardized for height (HT) by dividing by HT'. the declineof lung function ritb age among never-smokers is described I brsimple linearfunctionof AGE and AGE'. i b h p o r t a n t l y , these analyses showed that *method of standardizationand only this I#bodeliminatedthe relationship between 1 WIor FVC and height. This report uses *trmemethod of standardizationfor height =ses its validity. 1 %make the results presented in this report Wrrinterpretabe,FEVJHT' and FVC/HT' *are all adjusted to the height of a typi- 1 mbiect. FEVI/HT2and FVC/HT2were rrdbipliedby a sex-specificstandardsquared --(I73 cm)'or 3.0m' for men and (161 a)'or 2.6 mz for women. Thus, a man of tall with an FEV, of 3.00 L would b*abeight-adjusted FEV, Of 3.00 L x (173 I -4178 cm)' = 2.83 L. In the remainder ol*npOrt, all referencesto FEV, and FVC I -btbese height-adjusted values.Age is cen- pulation-based sample con2 adults between 25 and 74 1 (96.3%) of whom had measurements that met acceptable pulmonary vers (11). The proportion of white subjects and the frequency of 18.3%, respectively, among men and 41.7, acceptable pulmonary function exami- 19.3, and 18.7% among women. nations was evenlydistributed across age For each person in the sample, the decades and cities (8). predicted height-adjusted lung function The age-specificprevalencesof smok- values, FEV, and FVC, were calculated ing of cigarettes,cigars, or pipes were very from a model fitted to the pulmonary close to estimated prevalences for the function values of asymptomatic never- 1974 U.S.population (12). Among the smokers (8) and subtracted from the ob- men, 38.7% were currently smoking cig- served height-adjusted valuesto produce arettes; of the women, 33.4% (table 1). pulmonary function residuals. Mean More men had stopped smoking ciga- values of these residuals were signifi- rettes (34.4%) than women (15.2%). The cantly less than zero, indicating lung frequency of current cigarette smoking function lower than predicted, as would was highest for men between 25 and 34 be expected given the effect of smokers yr of age, and for women between 35 and and symptomatic subjects in the sample 44 yr of age, and lowest in the oldest age (table 3). Subjects with a history of ciga- group, 65 to 74 yr. The frequency of rette smoking had lower mean residuals former smokers increased dramatically than did those who had never smoked; with age in men, from 19.6% in the youn- current smokers had lower mean residuals gest men to 48.0% in the oldest. Among than did ex-smokers(table 3). Mean lung women, there was no apparent trend in function residual was lower among men frequency of former smokers with age. than among women, but this should be Among the men, 7.2% also reported that expected since 73% of the men were cur- they currently smoked cigars, and 7.4% rent or former cigarette smokers versus that they formerly smoked cigars. Cur- 49% of the women. rent pipe smoking was reported by 7.8070, Plots of mean lung function residuals and former pipe smokingby 6.5%. Thus, against lifetime cumulative cigarette for combined smokingof cigarettes, cigars, pack-years show a linear relationship be- and pipes, 21.1% of the men were never smokers, 30.9% were former smokers, and 48.0% were currentsmokers. Among the women, only 3 subjects reported ever smoking cigars or pipes. The mean age was higher for the TABLE 2 MEAN (STANDARD DEVIATION) OF C l G A R m E SMOKING CHARACTERISTICS AMONG CURRENT AND FORMER former cigarette smokers than for the CIGARETTE SMOKERS current cigarettesmokers (table 2). Years of cigarettesmokingand total pack-years were higher among the current cigarette smokers. For those currently smoking cigarettes, the mean daily consumption was 1.3 pack for men and 1.1 pack for women; 54% of the male smokers and Aae, Yr Men Women Years smoked Men Women Current 46.5 (12.9) 46.3 (12.4) 28.4 (13.1) 24.7 (11.8) Fom 52.8 (12.7) 49.9 (12.7) 21.7 (13.3) 16.2 (12.3) 36% of the female smokers reported that Pack-years they currently smoked 1 pack per day. Men 31.8 (21.8) 25.3 (23.7) As expected, respiratory symptoms were Women 20.8 (15.3) 12.8 (16.7) reported more frequently among the cur- Current pacWday rent cigarette smokers than among the Men former or never-smokers: 51.7,27.3, and Women 1.3 (0.7) 1.1 (0.6) 0 0 - TABLE 3 MEAN RESIDUAL HEIGHT-ADJUSTED LUNG FUNCTON IN MILLILITERS AND (STANDARD ERRORS) COMPARED WITH EXPECTED VALUES FOR HEALTHY, NEVER-SMOKERS Men Women Totar C i m e Smoking Never Ex Current Respiratory Symptoms No YeS Number 3,714 1.001 1,277 1.436 2,440 1,274 Residual FN, -293 (10) - 34 (16) -257 (18) -508 (16) -165(11) -538 (19) Residual FVC - 231 (11) -27 (19) -219 (19) -384 (17) - 128 (12) -428 (20) Number 4,477 2.302 680 1,495 3,291 1.186 Residuat FEV, - 24 (8) - 54 (16) -234 (11) - 51 (6) -230 (13) Residual WC - 19 (6) -23 (17) -146(12) --12654 (7) (14) 4P*Yr 25-34 35-44 45-54 55-64 65-74 -746 128 (20) -68 (22) 859 -38 (13) -638 -222 (22) 174 (25) 741 -94 (15) -955 -304 (19) -251 (20) 1.190 122 (12) -- -821 416 (24) 35a (24) 1,012 120 (13) - -554 397 (29) -298 (29) 675 107 (16) - 13 (15) - 47(17) - 70 (13) -86 (14) - 89 (17) % * ,' tween mean residual and pack-years of smoking (figure I), suggesting a cumulative and irreversible effect of cigarette smoking that was directly related to the number of cigarettes smoked. Among never-smokers, the distribution of pulmonary function residuals (figure2) was very closeto Gaussian (that is, the normal distribution). As the cu- mulative pack-years of cigarette smoking increased, the mean of the distribution of residuals became more negative. In addition, the variabilityof pulmonary function measurements, measured by the interquartilerange, increased with cumulativepack-years of smoking. The distributions of pulmonary function residuals, however, remained closetoGaussian for all cumulative exposure groups, as has been shown previously by Burrows and coworkers (5). The subjects in the sample who reported having ever smoked cigarettes (n = 4,888) were divided into current cig- arette smokers (n = 2,931) and former cigarettesmokers (n = 1,957). Comparison of sex-specificresidual FEV, for the current cigarette smokers versus exsmokers (figure 3) showed a linear relationshipbetween pulmonary functionre- sidual and pack-years. For each cumulative smoking group, however, the mean residual was lower for current cigarette smokers than for ex-smokers. This implies an additional deficit in lung func- tion associated with active smoking which apparently is recovered over some time period after the subject stopped smoking. These data were fit by parallel straight lines (figure 3), with no significant difference between the regression coefficientson cumulativepack-years for current and ex-smokers (p > 0.10). The irreversible effect of cumulative packyears was estimated to be a loss of 8.9 ml/pack-years for men and 6.3 ml/pack- L I F E T I M E PACK-YEARS 4 L I F E T I M E PACK-YEARS 0. 0. -a2 0' - 0. v) -Wa -0. > WIL +E - 0 . v3-I) 8 -0. c 0I z -1. I J- 1. 4a2 - 0 . I v) Wa -0. u>. 0 - 0. va) 7 =n - 0 . 7y -I. - 1. FQ.1. Meand m (residual)of height-adjustedRICand F W from erpeded values forhealthynever smokers yB(8w litetime pack-yeam years for women. The additional, n y ~ . ible effect of active smoking wgp w. mated to be 200and 128 ml for mep apd fuwomen of typical size, respectively. A multiple regression riiodel of& dataset of height-adjusted FEV, FVC measurements (n = 8 , 1 9 1 ) ~ ~ to develop a unified model for n m ainggi.nTghaenmd othdeeleifnfcelcutdseodfAcGigEaraenttdem.m&., the previously identified agingmodd fa healthy never-smokers, plus 2 rn- of smoking exposure, cumulative podt. t years and current cigarette s m o w I packdday as regression variables (tab& i: 4). Because of apparent differenceink I! effect of cigarette smoking between tbc t sexes, separate models were fitted for am 3 and women. \ The estimatedeffect of each pack-- I of cigarette smoking on height-adjusted I FEV,for men was 7.4 ml/pack-years (u- i ble 4). Current smoking of 1 pack/& was estimated to contributean additionrl t deficit of 123 ml. For women, the eai. i mated effect of cumulative smokiag 00 height-adjusted FEV, was 4.4 ml/pect- years. Current smoking of 1 pack/dq V would produce an acute loss of 107 ad. c Thus, the estimated effect of curnulatin I. smoking in women was about 60% of E that for men, and the estimated effcn e of current smoking was slightly 1- l women. For both sexes, the effectsOfCU. i mulative smoking (pack-years) and cllr. 5 rent smoking @acks/day) were s d on FVC than on FEV, (table 4). Among the subjects in this sampk 533 f it reported smoking more than 25 0- of pipe tobacco during their life Tbd these were women. The mean CUI* tive burden wasP.7 ounce-years. Ofthat 1' d c I. eOF LIFETIME SMOKING ON LUNG FUNCTION IN ADULTS c s 33O0 20 --1 10 - 0 PK-YRS Or -3.0 s 3300 20 10 -1- 0-t -3.0 -30 - 20 -*\. 10 -2.0 -2.0 -1.0 -1.0 0.0 0.0 1.0 2.0 3 . 0 1-20 PK- YRS - 1.0 2.0 3.0 21-40 PK- YRS N.1539 IQR=.638 Fig. 2. Distribution of height-adjusted 0 ~ ~ u ~ norms almfodelf~or m - 3 . 0 - 2 . 0 -1.0 0.0 1.0 2.0 3.0 bdh sexes combined as a function of pK.IRdciyseianrtse.rqPuiaarntgilelerainngdeic.ates mean; 3o s 20 - 10 - 41-60 PK- YRS N.575 IQR=.699 07 -3.0 -2.0 -1.0 0.0 , a- 1.0 -2.0 3.0 -3O 1 20 s 10. 01 -3.0 30 - s 20 . - - 2.0 - 1.0 0.0 3 61-8@ PK-YRE N= I89 ]OR=. 881 1.0 2.0 3.0 81* P K - Y R S N.108 I O R = . 966 -- . -3.0 -2.0 V 1.0 I 0.0 l:o 2:o 3.0 HEIGHT ADJUSTED FEV, RESlDUkL (LITERS1 289 ence, -63 f 9 ml). Almost all of the smokers reported that they smoked filtered cigarettes:.91% of the women and 84% of the men. No association was found between lung function residualand use of filtered cigarettes. Discussion Many epidemiologic studies have found a progressive relative loss of pulmonary function with cumulative smoking. For example, Anderson and Ferris (6) found a progressive decreaseof FEV,/FVC with increasing pack-years of smoking in a random sample of 1,167 adults in Berlin, NH. In a study of a random sample of 2,369 whites 14 yr of age and older from "bcson, AZ, Burrows and coworkers ( 5 ) reported that the most important predictor of percent predicted FEV, was packyears of smoking. Beck and colleagues (13) studied pulmonary function in a random sample of 4,609 whites 7 yr of age and older from 3 communitiesin the eastern United States. The most important predictors of loss of FEV, compared with healthy never-smokers were duration of smoking and pack-years. Changes in pulmonary function have been observed even among smokerswith very short durations of smoking. Seeley with a history of pipe smoking, 290 (%.4%) reported pipe smoking at the of their examination. Their mean w,Wnsumption was (3.5 ounces/day. Simi561 subjectsreported smokingmore than25 cigars during their life. One was a Wman. The mean cumulative burden Wgs 40.7 cigar-years. Of these, 275 (49.1Q'o) reported cigw smoking at the of their examination. Their mean masumption was 2.7 cigarslday. Regression analysis of the men's &ht-adjusted FEV, and FVC showed tht after adjusting for age, agef, pack- R&%of cigarette smoking, and daily cig- Uratesmoking, there wasno association rriflr cumulative burden of either pipes pd&m(p > 0.20) 'W Of either pipes noorrcwigiathrsd(apil>y s0m.2o0k).- b i v a l e n t results were found when the of cumulative and current pipe *Cigar smokingwere estimatedseparately for never, former, and current cigarette Ofthe male smokers, 14%. and of the b a l e smokers, 30%, reported that they ed slightly or not at all. These sub- lower mean residual FEV, than e who inhaled moderately or ference, -39 f 18 ml). &iwere found for FVC (differ- aIJn =l 0 O -250 c- W> LL -500 0WI- =In -750 7 0a c -1ooc 0 y e:. ':'haEX-SMOKERS .' >' CURRENT SMOKERS i v 300 ;zoo E EXSMOKERS N.1277 0 vl 20 a El EIn -250 - > W LL -500 0Wc I$n -750 0a 3c -1000 I WX PACKYEARS 0 IO 20 30 YO 50 SO 65. WOMEN 300 n l h=I- 200 a 0 u LOO 0 EXSMOKERS N.680 --- 8 100 CURRENT SMOKERS N.1436 300 ::g a 220000 . 1000 , CURRENT SMOKERS N.1495 0 PRCKYERRS 0 *S IO 20 30 YO 50 60 6s. PRCKTERRS FQ.9 &x-specific mean height-adjustedFEV, residualswrsus pack-yeamfor Current and ex-smdcen.Distributions of number of subjects by packyears are also preeented. I I 1 I I i 290 TABLE 4 SIMPLIFIED MULTIPLE REGRESSION ESTIMATES (STANDARD ERRORS) FOR THE ENTIRE SAMPLE (3,714 MEN AND 4.477 WOMEN). AGE IS CENTERED AT 50 YR. Height-Adjusted FEV, Height-Adjusted FVC Men Women Men Women Constant 1AGE-W (YO [AGE-50)' (yP) Lifetime pack-years Current packdday Rat RMSE, mi* 3.477.6 (16.9) -34.7 (0.8) - 0.204 (0.053) - 7.4 (0.5) - 123.3 (14.2) 0.576 568 2.506.2 (9.4) -28.3 (0.5) -0.121 (0.033) - 4.4 (0.6) - 107.1 (13.1) 0.548 381 4,393.2 (18.7) - 34.6 (0.9) - 0.234 (0.W - 5.6 (0.6) -96.7 (15.8) 0.548 629 3.123.3 (10.7) -28.7 (0.5) - 0.270 (0.037) -2.9 (0.6) - 81.3 (14.9) 0.584 437 * HeQMadjusted FEV, Iml) esuatS observed FEV, divided by heightsquIved t i i (1.73m)) tor men and (1.61 m)) tor women. Height-adjusledFVC is defined equivsbnt)y. t R2 la squ8md combtion bst*resnprsdic(ed andobsewadEN, or FVC. *WS broot man a q g w r s d ~ e n m r , b e l w w und~obmrved FEV, or FVC. value for a never-smoke A). By 60yr of age, he 2.728 L, 382 ml (13% pected for a never-smo 1 pack per day until age FEV, would be 1.989 L and associates (2) examined teen-agers in high school, and reported that flow at 50% of vital capacity was lower in smokersthan in nonsmokers. Peters and Fenis(3) found reduced flow rates among collegeseniorswho smoked. Part of this reduction in function may be a reversible effect of active smoking. Beck and colleagues (13) found that current smoked-ad lower pulmonary functions than did t h b e who had stopped, after adjusting for cumulative smoking, whereas Burrows and coworkers (5) did not. -.. Severalstudies have shown that smoking a singlecigarettecausesan immediate increasein airwayrasistanceand adin expiratory flow. Sob01and associates (14) found significant acute changes in FEV, with smoking. Tockrnan and coworkers (15) suggested that small airways dysfunction occurs relatively soon after the onset of smoking and is not affected by duration of smoking. Buist and colleagues (16) and Manfreda and coworkers (17) suggestedthat smoking itself and not the quantity of cigarettes smoked is the crucial factor in the development of early lung function impairment. Pathologic studies have found smoking to produce both structural changes in the airways that indicate irreversible effectsand inflammatory responses that are likely to be reversible. In a classic report on the pathologic abnormalitiesof chronic bronchitis, Reid (18) reported that inflammatory responses were found in the large central airways in mild cases of chronic bronchitis, and were found in small airways in advanced disease. She suggestedthat excessivemucus in the airways may cause impairment of breathing, especially if peripheral bronchioles are involved. Berend and coworkers(19) did pulmonary function tests before lung resection and correlated the function tests with morphologic abnormalities. They found that structural changes in the peripheral airways were associated with abnormalitiesof both specialtests ofsmall airways function and spirometry, including FEV,. Inflammation was the most im- portant cause of obstruction to flow in the small airways. A reanalysis and expansion of this study (20) reported that the degree of inflammation correlated well with level of FEV,. Severalstudieshave reported that performanceon tests of small airways func- tion returned toward normal in cigarette smokerswithout significant chronic airflow obstruction who quit smoking (16, 21-25). Thus, the observation that there are both cumulative irreversible, and acute, reversible effects of smoking on FEV, and FVC is consistent with epidemiologic, physiologic, and pathologic studiesof small airwaysdysfunction. Toillustratethe magnitude of the com- bined acute and cumulativeeffectsof cig- arette smoking, consider a white man, 173 cm tall, with FEV,equal to the ex- pected value for his height at age 25 (4.218 L). If he begins smoking 1 pack of cigarettes per day at age 25 (figure 4, curve B), the model predicts that his FEV,will drop by 123 ml. With each year, his cumulative smoking burden will increase by 1 pack-year. This leads to an increas- studies of small airways dicatea time scaleof less smokers (approximately 1 velop serious respiratory 5.01Y.5 RGE IYERRSI Fig. 4. Expected for a man 173 cm tall (A)* is at 50th percentile at 25 yr of age and who does nd smoke, (e)who smokes 1 pack per day slatting a 25 yr of age,(c)who stops smoking at 60 yr d aga @I who smokes 2 packs per day starting at 25 yrdand (E) who is at the 5th percentibat 25 y r 0 t - d smokes 1 pack per day. m C l 8 OF UEETIME SMOKING ON LUNG FUNCTION IN ADULtS 291 Susceptible to the effects of cigarette smoking. Buist and Ducic (27) have described 3 possible ways that persons can come to have low values. First, their lung for their sex, height, age, and smoking habits. That is a hypothesis that can only be addressed with longitudinal, not with cross-sectional, observations. Burrows and coworkers ( 5 ) have ously smoked. The recovery of the acute loss is equal to 3.2 yr of aging for our male example. It would take this patient approximately 11yr to lose an additional 500 ml of FEV, if he continued to smoke Wwth and development may have reported that subjects with chronic 1 pack/day. If he stopped at age 60, he placed them at the lower end of the age- productive cough have steeper rates of would take approximately 17 years to specific distribution curve for the maxi- decline of FEV,, and therefore may rep- reach that same level. mum obtained lung function. Second, resent a susceptible subgroup. In this One class of patients at higher risk of they may have started out with normal analysis, a history of chronic productive disability from respiratory impairment 1 function, but had a faster than normal cough (chronic cough with chronic &line in function with age. Finally, they phlegm) was associated with lower levels are those on the low end of the distribution of pulmonary function for their age. 1 may have had a normal rate of decline of pulmonary function after adjusting While it is obvious that those with low with age plus a discrete loss of function for age and smoking, for both men (-283 lung functions compared with predicted caused by some insult, such as necrotiz- 30 ml for FEV, and -214 & 33 ml are at risk, the model presented allows ing pneumonia. for FVC) and women (-124 & 26 ml for the physician to plot the predicted track The model presented here assumesthat FEV, and - 98 f 30 ml for FVC), but of lung function decline for his or her all smokers lose pulmonary function at there was no significant difference in the patient given an initial observed value a rate determined by their sex, age, and loss of pulmonary function between and assuming continued cigarette smokinghabits. Differencesbetween sub- those subjects reporting and those not smoking. jects are assumed to be due to measure- reporting chronic productive cough. For the epidemiologist, this analysis ment error and random individual vari- Smoking is associated with both chronic provides guidance in efficiently control- Cltion. The model therefore implies that productive cough and steeper decline of ling for the effects of cigarette smoking those subjects who entered adulthood pulmonary function. Chronicproductive when evaluating the effect of environ- with lung function on the low end of the cough, or any respiratory symptom as- mental agents or other risk factors on dktribution of pulmonary function are sociated with smoking, should not be lung function. Most such studies do not atthehighest risk of disabilityassociated considered as a risk factor, but rather as have the luxury of many thousands of with respiratory impairment from ciga- an intermediate variable. The risk factor subjects randomly chosen from the com- rettesmoking. We showed earlier that a producing both symptoms and more munity. While it may not be awropriate 2pyr6d male, 173antall,who has never rapid loss of pulmonary function is to use the prediction equations presented rtj"e;dted and has an FEV, equal to the smoking. here for comparison to other samples, apedted mean value for a never-smoker, Clinically, this description of the ef- it is appropriate to use the functional nouM,if he smoked 1 pack of cigarettes fect of smoking impliesthat, on average, form presented to remove the effects of until 75 yr of age, have an ex- there is a modest additional loss of FEV, sex, age, and smoking from the analysis. of 1.989 L. Even smoking or FVC associated with cigarette smok- This analysisdescribesthe mean cross- h, his expected FEV, at age ing. The expected annual decline of sectional effect of smoking, but the be- would not be low enough height-adjusted FEV, caused by aging havior of an individual subject may not red clinicallyimpaired (28). alone for a 50-yr-old man is 34.7 ml/yr follow the cross-sectional model. These a similar person at 25 yr of and for a 50-yr-old woman is 28.3 ml/yr results do suggest that there should be the low end of the normal (table 4). The expected annual effect of some initial improvement of lung func- rexample, the 5th percentile for smoking 1 pack of cigarettes per day is tion among subjects who stop smoking, omatic never-smokers, based on an additional annual loss of 7.4 ml/yr and that their subsequent annual loss of sly published distributions for men and 4.4 ml/yr for women. Thus, lung function should be equal to that of omatic never-smokers (8) smoking a pack of cigarettes per day in- never smokers. The public health impli- (FEV, = 3.411 L), then after smoking creasesthe annual decline of FEV, by ap- cations of these findings are obvious. mv,1Packldayuntil 75 yr of age, his expected proximately one-fifth that of the effect Stopping smoking at any age reduces the would drop to 1.182 L.At this level of normal aging for a 50-yr-old man, and progression of lung function loss, and of FEV,, he would probably be com- one-sixth for a 50-yr-old woman. How- also provides an immediate improvement Ph&g of shortnessof breath and would ever, because this additional loss is cu- in lung function that may amount to a akClassified as severely impaired by the mulative, and irreversible, the net loss of significant fraction of the overall loss schema(28). Becausethe additional lung function becomes substantial after caused by smoking. b s of FEV, associated with smoking is 10or 20 yr of smoking. If a patient stops small compared with that in normal ag- smokingcigarettes,his annual rate of pul- k,the risk of developing respiratory im- monary function decline will return to Acknowledgment pairment based on this model is deter- that of a never-smoker. For the example The writers aregrateful to all those who have mined primarily by level of FEV, at maturity. Thus, this simple model is sufficient to show that only a small fraction of Smokerswill actuallydeveloppulmonary we have been considering of a man who smokes 1 pack per day who stops smoking at 60yr of age, the annual FEV, loss would be reduced by 17%, from 44.1 to 36.7 ml/yr. There should also be some assisted in the examination of the adults, especially those who have directed the data collection in the 6 communities: George Allen, Eric Bjsmstad, Steve Hancock, Charles Humble, Marcia Lyndon, Lynn McClelland. and Vicki Senape.Many others have worked function low enough to be impaired. It improvement in level of FEV, over a to prepare these data for analysis, including ' may be true that some proportion of period of approximately 1 yr. The size Hermia Chen, Renee Croxen, Martha Fay, Smokers have faster rates of pulmonary of this improvement is small in men, only StefanaFerris,David Glicksberg,Dean Harris, function loss than the population mean about 123 ml for each packlday previ- Ellie Root (deceased), and Janice Weener. References 1. US. Department of Health and Human Services. Thehealth consequencesof smoking: chronic obstructivelung disease Washington, Dc:USGPO, 1984. DHHS (PHS) 84-50205. 2. Seeley JE, Zuskin E, Bouhuys A. Cigarette smoking: objective evidence for lung damage in teen-agers. Science 1971; 17241-3. 3. Peters JM, Ferris BG Jr. Smoking, pulmonary function, and respiratory symptoms in a collegeage group. Am Rev Respir Dis 1967; 95:774-82. 4. Barkhouse CI. Peak expiratory flow in youths with varying cigarette smoking habits. Br J Med 1975; 1:360-2. 5. Burrows B, Knudson RJ. Cline MG, Lebowitz MD.Quantitative relationships betwem cigarette smokingand ventilatory function. Am Rev h p i r Dis 19777; II5:195-uH. 6. Anderson DO, Ferris BG Jr. Role of tobacco s m o k w in the causation of chronic respiratory disease N E n d J Mad 1%2; 262787-94. 7. Fletcher C. Pet0 R. Tinker C. SpeirerFE. The natural history of chronic bronchitis and emphysema. London: Oxford University Pres, 1976. 8. Dockery DW, Ware JH, Ferris BG Jr, et uL Distribution of FEV, and FVC in healthy. white adult never-smokersin six US.cities. Am Rev Respir Dis 198% 131:511-20. ology standardization project. Am Rev Respir Dis 1978; I18(Part 2):55-88. 11. Ferris BG Jr, Speizer FE, Bishop Y,Prang G, Weener J. Spirometry for an epidemiologicstudy: deriving optimum summarystatistics for each subject. Bull Eur PhysiopatholRespir 1978; 14145-66. 12. HarrisJE. Cigarette smoking amongsuccessivebirth cohortsof men and women in the United States during 1900-80. J Natl Cancer lnst 1983; 71~473-9. 13. Beck GJ, Doyle CA, Schachter EN. Smoking and lung function. Am Rev Respir Dis 1981; 123:149-55. 14. Sobol BJ, Van Voorhies L, Emirgil C. Detection of acute effects of cigarette smoking on airway dynamics: a critical and comparative study of pulmonary function tests. Thorax 19777;32312-6. IS. Tockman M, Menkes H. Cohen B, et UL A cornperisonof pulmonaryfunctionin malesmokers and nonsmokers. Am Rev Respir Dis 1976; 114 711-20. 16. Buist AS, Ghezu, H, Anthonisen NR, et 01. Relationshipbetween the single-breath N,test and age, sa,and smoking habit in three North American cities. Am Rev Respir Dis 1979; 12&305-18. 17. Manfreda J, Nelson N, Cherniack RM. PIWalence of respiratory abnormalities in a rural and an urban community. Am Rev Respir Dis 1978; 117:215-26. 18. Reid LM. Pathology of chronic bronchitis. with function. Lung, 1982; 160:115-36. -21. lngramRH,(YCainCEFnqueacyof compliance in apparently healthy m o b w non-smokers. Bull Eur Physiopathol Rcq& 7995-210. 1975; 5943-52.