Tobacco Documents Online

The New Engla~nd .Journal of Medicine eCopyright, 1979. by the Massachusetts Medical Society t'olume 300 FEBRUARY 22, 1979 Number 8 CHANGES IN ' BRONCHIAL EPITHELIUM IN RELATION TO CIGARETTE SMOKING, 1'955-1960~ VS. 1!9'70-1'97'7 OscAR AvERBacx,, M.D., E. CuYLER hIAMhtoND, Sc.D., A.vD I.AwREticE GARfltixEL„ Ivf.A. I ( Abstract To test the hypothesis that the reduction in tar and nicotine content of cigarette smoke that began in the 1950's should be reflected by the histologic: changes in the bronchial epithelium of cigarettie srnokers; 20,424 sections taken at autopsy from! the bronehial"tubes of 445 men (non-lung-cancer deaths)i were examined microscopically ini random order. There were 211 men who diedd in 1955-19'80, of whom .154 smoked regularly,, and 234' men who died in 1I970- 1977; of whom 1i8't were regular smokers. Changes studied included basal-cell hyperplasia, loss of' cilia and' occurrence of cells with atypical nuclei. In both IN ai set of studies published some years ago,''' we ~`- found that several types of histologic changes inn bronchial epithelium occurred far more frequently ira. cigarette smokers than in nonsmokers, increased with amount of cigarette smoking and, among cigarette smokers, increased with~ advancing age. The same changes were found tess ft•equentl~ in former tlian in continuing cigarette smokers. The changes studied in- cluded basal-cell hyperplasia, loss of cilia in some areas and occurrence of cells with atypicad nuclei. These histologic findings paralleled' epidemioiogic findings nhat rates of death from lung cancer are many times~ higher among cigarette smokers than among nonsmokers, increase with amount of cigarettesrnok- ing and„ among persons with the same smoking histories, increase with advancing age.'-' Among former cigarette smokers, the rates~ decline with the length of time since giving, up the habit,`-' In a later study, we had beagle d'ogs inhale the smoke from filter-tipcigarettes and from cigarettes of the same brand but without' the filter.' The filter removed about half'of the total tar and sornewhat less than half of the nicotine contained in the smoke. Fewer pulmonary neoplasms were found in dogs thar smoked fiDter-tip cigarettes than in those tharsmoked non-filter-tip~ cigarettes. Invasive pulmonary tumors were found only in non-filter cigarette smokers. Two /~ From the Veterans Administration Medical Center„Eat Orangc. NJ:,the. ~ =ollege of Medicine and Dantistry of New Jcrsay, New Jersey Medical \~Sehoal, Newark„NJ, and the Department of Epidemiology and,Staustica, American Cancer Society. Inc.. New Y;ork.,NY (address reprint requests to Dr. Awerbach at 15 11% Veterans Adminisuation Mediuli Center, East Orange. NJ 07019). periods studied these~histblogia changes occurred far less frequently in nonsmokers than in cigarette smok- ers~and increased in frequency with amount of smok- ing, adjusted for age. Sections withd advanced histo- logic changes . in those dying in 1955*1980 oc- curred in 0'per cent of'nonsmokers, in 25per cent of, those smoking one to 19'cigarettes a day, in 112 per cent of those smoking 20 to 39 and in 22.5 per cent of those smoking 40+ cigarettes a day. In those who: died in 1970-1977 the percentages were 0, 0.1, 0:8,, and 22, respectively. (iN Engl J Med 300:381-386,, 1979) . of these were early squamous-edl~ bronchial car- cinomas, the.others being invasive bronchioloalveolar tumors. In a retrospectiivetpidezniolbgic.study„itwas found that lung cancer occurred less frequently in men who smoked filter-tip cigarettes: than in those whosmoked non-filter cigarettes,' In a prospective epidemiologic study, three groups of cigarette smokers were matched oni the basis of age, current number of' cigarettes smoked per day, age at which, they began smoking, race,, education,, residence (urban, vs. rural), oc- cupational exposure or non-exposure to dust, fumes and chemicals, and, past h~istioryoflheart di'sease! and lung cancer.30 The three groups were as follows: (A), men who smok.:d high tar/nicotine cigarettes; (B), men who smoked medium tar/nicotine cigarettes; and (C), men who: smoked relatively low tar/nicotine cigarettes. In subsequent years, death rates from lung cancer were highest in Group A and lbwesi in INA Group C. 0 Since 1954, when effective filters were first i!n- C trodut:ed„ there has been a large and continuing drop O in the tar and nicotine content of' the! mainstream stnoke of cigarettes consumed in the United' States.t't ~ .Even non-filter cigarettes deliver considerably less tar ~ and nicotine than those sold' in previous years - this cio ' reduction being achieved by selection of tobacco, the Q~ use of homogenized tobaccoiand various other means. jV. Indeed, the highest tar/nicotine brarlds on the markt:t today deliver less tar than the lowest tar brand of American cigarettes on the market before 1954. Thus, everyone who has been a habitual cigarette smoker for 25 years or longer must be smoking cigarettes with

. 382 THE NEW ENGLAND JOURNAL OF'MEDICI:NE' less tar and nicotine than, those formerly'smoked, and a large proportion of smokers have deliberately selected brands with reduced tar and nicotine.. This study was undertaken to determine whether the drop in tar andi nicotine has been reflected in less extensive histologic changes in the bronchial epi- thelium of cigarette smokers. Men who died in 1955'- 1960 were compared with men who died in 197CD•1977. M'A1IERIALS' AA D N1 E7H Ot)S During the period 1955-1960, we obtained a portion of the traehea and the lungs of e.ery person who died and came so autopsy at the 4 eterans' Mcdical Center of East Orange, New Jeney, and, during'a part'of this period, we also obtained''lungs from persons .who died!in several hospitals in upstate New York. We obtained in- fortnationon the,smoking histories and'occupations of'these sub- jects by interviewing members of their famtlicss lhis step provided material for our earlier studies of bronchial epithelium.''s'Collection of lungs from the same hospitals was resumed in 1970 and con• tinued through 1977'to obtain additional material for this study. Exactly the same technical methodi wex used in the two periods. They have been described in detail." Briefly, they, are as follows: As soon after death as possible, the specimen was removed ar autopsy and fixed by instillation of formalin solution down' the trachea. Later, the entire tracheobronchialtree was dissected'out and divided into 208 portions;,which were then embedded in paraf- fin."A thin histologie section was eue from each portion, mounted on a slide and stained with hematoxylin and eosin. In this study, as in some of our other studies, we examined sections from just 55 of' the 203 portions, which t.ere selected from the lower trachea or the main lobar or'segmentallbronchi according,to a set plan' (Fig. 1). S'ELEt:rIOK AND MA?CHT~C OE SUBJECTS We decided to confine this investigation to male subjects of just two sorts: men who had never smoked regularly and men who had smoked cigarettes regularly up to the time of their terminal illness. (Some of the latter had also smoked pipes or cigars.), Those who died of lung cancer were excludedl Alsoo excluded were men for whom we could' not obtain reliable information on smoking, and men whose oc- cupations were such that they may have been Figure 1. Schematic Diagram oEthe Tracheo'bronchial Tree. The circles indicate the locations from which 55' sections were selected for study. Feb. 22, 1979 Table 1. Distribution of Matched Pairs of Subjects According to Age at Death and Smoking Habits. Aoc 6awn(Y.) No. or' WATeneo IPAIRs TOTAL Nivtl S,WKID. f?eOK[D. S910[lDt?AOKED <1 FACI- 1-31ACx- 1F,P4OL-RIOtiURLT' AG[/DAY AGlS(DAT, . A6ltIflAlr «o S' 2' 2 1 0 40:-td 61 0 2 3' 1 45-49 18' 1 5 9 3 1 50+44 ' 11 2' 2 4 3 55-59 28 5~ 7 13' 3 60=64' 35 10' 4 10 11' 65-69 31 6 1II, 12' 2 7044' IS 61 41 S 3 75-79 16 6 3' 5 2 80i84' 12 7 31 2 0 85+ 2, I 0 I 0 Total pairs 182 46 43' 65 28 Total subjects 364 92 86I 130 56 exposed to dust or fumes that are said to have an ef- fect on pulbnonary tissues.. Presumably because of technical difficulties (but possibly because of pathologic alterations in some cases) some of the sections from the 1955-1960 cases were completely or almost completely denuded of epitheliutn; such sections were useless for a study of this sort. Therefore, all the sections were screened microscopically for denudation. Forty-fiive to 55 satisfactory sections were available for a majority of the subjects, and at least 35 were available for most of the others. We excluded cases with less than, 20 satisfactory sections. The age at, death, cause of death and, smoking habits of each of the potential subjects were put on - computer tape together with, identification of' the groupito which they belonged: "A" for those who die& 1'955-1'960 and "B" for those who died 197011977. For each, Group A subject, we'sea'rched for'a GrouplB sub- ject who had smoked the same number of cigarettes'a day and' was within the same five-year age group at time of death. In, most of the final matched pairs the two men were within one year of the same age at the time of death. Where two or more B subjects were found to match, an, A subject, the one closest in cause of death was s+dected'as the match. By this procedure we found 182 matched pairs of subjects, each pair consisting, of one A (1955-1960) subject an& one B (1970-1977) subject. Table t shows the distribution of these subjects according to age and smoking habits. It will be noted that some of the spaces in, Table 1' contain no subjects or only a few subjects. This setup is because the age-by-smoking-habits distribution of subjects who died' im 1970-1i977 was considerably dif- ferent from the distribution for those who died'l in, 1955-1'960. To build up numbers and achieve a better distribution according to age and smoking habits, wee supplemented the matched-pair cases by including, an additional 81 subjects; 29 from Group A and 52' from Group B, Thus a tota'1' of 445 subjects (364 ito matched pairs plhs 81 others) were included in the 1005+05296%11'

Vol. 300 Xo. a' SA'tII AND BRONCHIAL CH.1NCES - rAVERBACH ET AL 0 procedures described beiow, Their distribution ac- `%t:ording to age at death and by smoking habits is shown, in' Table' 2. In all, there were 20,424 sections that contained enough epithelium for evaluation, an average of 45.8 per case from the 445 subjects. A computer record containing the identification number was made for each of these sections, the 20,424'records were put in random order by random numbers, and while they were in this order, a serial number was addea to each record. A label containing, the serial number of each record was placed on the corresponding glass slide, after which the 20,424' slides were put in order by ' seriall nunnber.T,hus; the slides: were, put in ran'dom' order. While still in this order they were micro- C scoplcally examined by one of us ('l7. A.) who was. given no clue to whether a'slide was from the Group A or Group B and no clue to the smoking habits and age of the subject. Normal bronchial epithelium, as' seen inl a cross- section, consists of one row of basalU cells lying on the basement membrane and a single roM1V consisting of ciliated' columnar cells and goblet cells; no cells with atypical nuclea are present. Sections in which the en- tire epithelium fits this description were classified' as "no changes found." Many sections contained one or more areas in which was a change of some sort (i.e., hyperpiasia, loss of ciliai cells with atypical nuclei or a, eombination of these alterations). We called such, an ' " ' area a Two'or more lesions of different types lesion." were found in some sections. Each lesion, was described according to presence or absence of cilia, number of cell rows and proportion of cells with <40 3' 2 40~ c 6' 0 45-49 I8I 1, 50-54 11 2 s3-59 28' S' 60-64 49 13' 65-69 46 14 70-7a 18 6' 76-19 16 6' 80-84 12 7' Totals t 1 Si7 All 1970-1977 subjects (Group B): <x0 , 6 2 4o-s,t 7 0 43-.t9 20 3 50-54 37 3 3*r54 36 6 60-64 3s lo 65-69 34 6 ` 70l 21 6 75-79 21' a 80-8t 12 7 8s+ 5 2 Totals '154- 33 Table 2. Distribution of MumberofSubjects Aucording,to Age •.; at Death and Smoking, Habits. - TorAL M[vlR Srocap RIGULAt6Y All 1955-1960 subjects (Group A,k S}1oKdD <1 Pncc- .oeyDAY Stactn. 1-2 Pucc. ,ou/D.,r SroK[D 2+ PAc.• AaEsyDwY 2 1 0 2 3' I 3 9 3 2 4 3 7 13 3 12 13 111 111'. 12' 9 4 5 3 3 5 2 3 2 0 0 I 0 SII 68 J5' 2 1 1 2' 3 2' 3 9 3' 13 9 12' 7 13 10 4 no u 14 12 2', 7 5 3 3 8 2 3 2 01 1 1, I 1 3 i7 383 atypical nuclei. Additional information was also recorded - for example, the f nding of mitotic figures or unusual cells of any sort. For purposes of this report, we use two terms. The furst is "ba'sal-cell hyperplasia" - a lesion with cilia present but with three or more rows of basal cells. All, the cells may have normal nuclei or'sorne may hawe a''typicall nuclei. The second' is "lesion with eilia absent"t such lesions lack cilia. They m'ayy be norma'!' in other respects, or they may have a few or a great - many cells with atypical' nuclei. They may or may not fit, the description' ofst~ratification: or metaplasia. Photomicrographs of'such lesions have appeared inn published reports:'" As shown in Tables 3 and 4, basal-cell hyperplasia was subdivided according to degree of' hyperplasia (i1e., a total with three or more rows of'basad cells andd those'with six or more rows of'basal cells) and also ac- cording to percentage of cells with atypical' nuclei. Sections containing a lesion with cilia absent were further subdivided according, to percentage of cells with artypicalI nuclei. All the above information was recorded and trans- ferred'to computer tape together with information on the identity of the subject (GrouplA or B);, number of' cigarettes smoked per day and' age at death. FINDINGS IN AI:L SuBJzQ?s'. Table 3 shows the findings for all of the 445'subjects (211 in Group A and 234 in Group B)n including un- matched cases as well as matched cases. The results are presented in terms of percentage of sections con- taining lesions of each of'various types: Since the age distribution vaciedl in different amount-of-smoking categories, the percentages were'standard'ized, for age on the age distribution of'all the subjects in the study. 8aeal1CeII Hyperplasia As shown on the first line of Table 3„ in both Croup A (1955-1960) and Group B(r270,1977) the percentage of sections withibasal-cell hyperplasia was far higher in cigarette smokers than, in nonsmokers and increased with amount qf' cigarette smoking, Among nonsmokers,, Group A had somewhat fewer sections.with basal-cell'hyperplasia than Group B (3.8 per cent vs. 5.8 per cent). Among cigarette smokers the percentage of'sections with basal-cell hyperplasia was higher fbr Group A than for Group B. Relatively few sections had a, basal-cell hyperplasia to the degree of six or more rows of' basal cells. The percentage of sections with this finding,increased with •amount of cigarette smoking, and among, cigarette smokers, the percentages were far higher in Group A than in Group B. The next fbur lines in Table 3 show the percentage of sections with lesions defined as basal-cell hyper- plasial further classified according tio the proportion of cells with atypical nuclei. Basal-cell hyperplasias with as many as 110 per cent of the cells having atypicall nuclei were rarely found inisections from nonsmokers.

e ~; but were found in a large proportion~of'sections from cigarette smokers. The' percentage of sections with such' lesions increased with amount of cigarette smok- ,ing, and among cigarette' smokers was greater in, Group A than in Group B. Basal-cell hyperplasias with 70 per cent or more of the cells having a2ypicall nuclei were found in none of: thesections from nonsrnok'ers and in only a very small proportion of the sections from Group B cigarette. 'smokers. They were found in 0.1 per cent of sections from, Group' A subjects who smoked one to 19 ciga- rettes a day, in 12.2 per cent of sections from Group A subjects who smoked 20'to 3'9 cigarettes a day andlin 66:6' per cent of Group A subj ects who smoked 40 or more cigarettes a day. . . Lesions with Cilia Absent Sections having,lesions with cilia absent were found more frequently in, cigarette smokers than in non- smokers and increased in frequency with amount of' ; cigarette smoking. They were fbund''' more frequently in' sections from Group' A cigarette smokers than in sections from Group B cigarette smokers. Lesions lacking cilia and wit'h, 10 per cent, or more of the cells having, atypicall nuclei were found in 0 per cent of the sections from Group A nonsmok'ers and in <0.1 per cent of the sections from Group B non- smokers. smok'ers. They'were found in an appreciable percent.- age of' sections from cigarette smokers, and the percentage increased with armount, of' cigarette stnok- 109:.+ aalls'with atypinnnuclci -. •Tab1e 3. Percentage of Sections with Each of Several!Categories of Histologic Change, Classified Aecording, to Smoking Basal-csll hyperplisir. ~ : Total _ i 6+ rows 30%+ cells with atypical inuelei SO'8r:+ cclls with atypictl nucld 70%+ cella with atypical nuclei THE NEW ENIGLrluliD JOIJRD1n1L OF NtED'tt:IIYE Feb. 22, 1979 '~'r Y ing. Within each smoking class, more such lesions were found in Group A than in Group B subjects. Lesions lacking cilia and with 1100''per cent ofthe" s cells having atypical nuclei were foundlin none of'sec- tions from nonsmokers. Among, Group A cigarette smok'ers; the proportion of sections with such' lestonr rose with amount of smoking from 2.6 per cent (one to 19 cigarett'es ~ a day)~,. to 13~.2~ per~ cent (20 to, 39 ciga- rettes a day) to 22.5 per cent (40 or more cigarettes a day). Such lesions were found far less frequently in sections from Group B cigarette smokers but rose from 0.1 per cent (one to 19' cigarettes a day) to 0.8 ''.' g per cent (20 to 39 cigarettes a, day) to 2 2 per cent (40 . or more cigarettes a day) (Fig. 2). Four, of the lesions with cilia absent andi with 1100' per cent, of the' cells having atypical nueld' gave evidence of invasion (i.e., very early bronchial carci- -',l noma). Three of'these sections were in the Group B> subjects and one was in a Group, A subject. FINDINGS tN i1r1XTCHE'D PAIRS In this study, 364'of the 445 subjects were matched in pairs according, to age at death, but it is possible that the inclusion of 81 unmatched cases could have introduced some bias in comparison of findings in Group' A vs: Group, B. To avoid'! such bias, we under- took a matched4pair analysis confined to the matched'' subjects: Table 4 gives a summary of findings ineach of twoo sets of mat'chedi pairs: the 46 matched pairs of non- Habit:•' A'D/tnTto ~~ . ADNSTeD \tvan~sMoKm. S.MOCaD 1-191 RtastJ.at:r GwaEmTS/ DAr 7 kn.- . ADiusrto S ADrtnnn. % . SKO:ec -,0-39 S4ohw 40+ GwRrrrea/ GwriTTU/' OAT DAY A 8 A 8'A f. A 8. 3.81 5.9 ' 87.8' 63.11 93;2 76:2. 0 0.1 2.1 0.4 5.7 0.5 0.1 03' 87,6' 62.4 93.2 75:0' 01 0:A 77.2' 53.9 92.6 7II:5 0 0.1 56.7' 9.6' 8,t.1 26.3 0 0 0.1 0 12Z 0:1 ~ 0 Ls:ion with cilia abscnt 98:8' 86.3 13.0 0:8 98.8 86J 98.8' 85:1 98.6 56.1' 66:6' <0.11 Totall 53 4.2 118 8:8 22.5 10.i3 ~ 30.3 11.7 ' I0R,+ cells with 0' <0.1 13.8 8!S 22.3 9:8 30.3 1'1.7 atypical nuclei 30'7e+ ceits with 0 <0.1 12.9 7.6 22J 9.3 30.3 111.7 atypical nuclei r~ 10Si+ ceils with 01 0 10.0 2'? 21.9 6.0 -' 30.3 9.3 ' atypical nudei' 70qi+ celli with 01 0 2.6 0.11 14.6 0.8 28.6 2.2' atypical nuctei' 100% colls with 0 0 2.6 0.7 131' 0.8 22.3 2-1 ' atypialinuclci' Mo: of seetions No. of subjacts 2,:80 2.623 2~~,208 3,026 2:$S I I - 3:47I 57 53 S'II 61 68 73 1,413 35 2;217 47 •'Peteenu8eaadjustedforaptotliedianbuuono(a{sudc.thotalllsubjsasinilhestudy.Adsnotusubjaets.hodiedi nI9S5-19Mt;t8 subjaxs wheidied ie 197tk1977,

Vn1: 300 \o. 8 S\lOK211G 'A.`DI BRONCHIAL CHANGES -AL'iER'BiACH ET'AL 385 !,~Table 4. Comparison of M'att:hed Pairs of Subjects for Each of Several Categories of Findings: 46 Pairs Who Never Smoked ~{ , 'Reguiariy and 136 Pairs of Cigarette' Smokers. _ ~>. Fmm.e Tfnrta S.eoKla Rket twu•r Cie.am Smocas A- 8'-100R, A• 81• 0!i A-8' A>8 A<8 A'- 8-d00R A- 8'-0!f A-8 A>8 A<8. BaswGuell Ih'yperptasia: Total 0 9 0 6+ rows' 0 4i 0 Ir.+ eells with 0 35 0 atypicai nuclei 3046+cellsw'ith 0 36 0 atypical nuelei SO^e'+ ceils with' 0 43 0 atypical nuclei " 70'b+cellswith 0 46 0 uypioal nuclei Lesion with cilia absent Total 0 14 10°c'+ cells with, - 0 4~t 0 atypical nuclei 30'd+ tells with i 0 44 0 atypieal nuclei 5r.'+eelltwith' 0 416' 0 atqpical nuclei 70%+ cells with 1 0 46' 0 atypical nuciai I001ioccllrwith 0 46' 0 atypicai nuclei 100=collswith 0' 46 0 aty'picrl nuclei & invasion, •A daaoce+ wbjaus who dicd in 19b3+1960, A' B subjects who died in 19741977. l _. k ` ' smo ers and the 136 rs of cigarette matched pai smokers. Each subject was classified according to the percentage of' his 20 to 55 sections that shonved! a par- ticular change. This figure could vary from 0 to 1000 per cent. If all the sections oE both subjFcts showed that particular change, this pair was entered under 'A=B=100 per cent. If none of the sections in' either ". subject of a pair showed' a particular type of change,, r' this'pair was entered under the heading,A=B=0 per cent. Otherwise, if.the samepercentage'was found in. Group, A and Group B, this observation was entered under A=B'. If the A subj ect, had more changes than the B subject, this fact was entered under A>$. If B had more changes than A, this increase was entered under A<B. Nonsmokers For basal-cell hyperplasias A>B' in 118 ' pairs and A<B in, 1'9 pairs, and' for lesions with cilia absent, A> B in 15' pairs and A<B' in 1'7 pairs: Very few of the sections fromi the nonsmokers had' lesions with as many as 10 per cent or more of the cells having atypical nuclei. However, atypia was'somewhat more common in Croup B' than in Group A. For example, in the category, "basalLeell hyperplasia with 30 per cent or'rnore of the cells having atypical nuclei,'" A> B' in oniy one pair, and A<B in nine pairs. This dif- %rence between the two groups was not sufficient to (J.iraav a firm conclusion that the frequency of atypia in the' bronchial epithelium of nonsmokers increased with time from, 195'5-196& to 1970-1977. 16 19 6 0 0 112 18 0 2 0 52 0 72 12 2 9 6' 0 0 112 Iit 1' 9 6 1! 1 108 20 0 3 1 7 0 122 6 0 0 0 ' 6E' 0 47' 1' Cigarette Smokers For basal-cell hyperplasia, A>B' in 112 pairs and A<B in only 18 pairs, and for lesions with cilia absent A>B in 92 pairs and A<B' in 42 'pairs. These dif- ferences were statistically signi'ficant6 A large proportion of the sections Erom the cigarette smoker had lesions with 10 p'er' cent or more of the cells h'aving, atypicall nuclki„ and' atypia was more common in Group A than in Group B. For exampde, in the category "basal-cell hyperplasia with 10 per cent or more of cells having atypical nuclei," A>B in' 112 pairs and A<B in, 1 8,pairs; in the category"le- ~' xs~ 20' ~1, 15 0 0 0 2`~o_i A 81 A B gNEV~EgRp SM'OqK~EOx REGULArZLY tA PtMY 0.8 " B A SMOKED hI P~1_cyX3 Figure 2. Percentage of Sections withi Advancedl Lesions (Cilia Absent andiAll Nuclei Atypical) According to Smoking Habit imGroup A(1955-1960 Deaths) and Group B(1!970-1977 Deaths). The 95 per cent confidence limits for these percentages are as follows:' <1 pack, A 1.9-3.3. B 0.0-0.2: 1-2 packs, A 11.9- 14.5, B'0.5-1.1; andi2+ packs„A 20.3-2+4.7, B 1.6-2.8.

THE NEW ENGt11ND JOURNAL OF \'tEDICINE' sions with cilia' absent and 10 per cent or more of the cells having atypical nuclki," a'>B' in 93 pairs and' A<'B in 40! pairs. These differences were statistically significant. More' advanced lesions were found' far more frequently in Group A than in Group B. For ex- ample,, im the category "lesions with cilia absenr and 100 per cent of'thecells h'aving'aty.pical''nuclei;"A>'B in 84 pairs and A<B in only fiire pairs. In examining the thousands of sertions from cigarette smokers,,.+•efound just two eputhelial'lesionss of microscopic siiewith evidence of invasion of the un. derlying' tissue (i.e., very early carcinomas). One of these was in a Group' A subject and the other was in a Group B subject. Because of their very small size, they had' not been found at routine autopsy. DgscusszoN As described abose,, histologic changes of various sorts were found far more frequently in the bronchial epithelium of cigarette srnokers'w,ho died in the period 1955-1t?60 than, in those who died in, the period' 1970- 1977: The changes ranged from slight basal-cell' hy- perplasia withifew if any cells having atypical nuclki too lesions with cilia absent and a'lll the cel'ls having, atypical nuclei. Basal-cell hyperplasia, a reversiblk change, is pre- sutnably a reaction to some de!t:terious' factor; it is~ probably protective rather than harmful. Loss of cilia, a reversib'lechange, presumably results from the pres- ence of some deleterious factor; it is harmful' in that it destroys one of the mechanisms by' which foreign : material is ordinarily removedl from, the lungs. In our opinion, the occurrence of atypical nuclki is the first definite step along a road that' may eventually lead to carcinoma in, situ and from there to invasive carcinoma. This'statement is not to imply that all cells Ty with nuclei havec pttial th is , atypical neoplastioen;ere evadence" that lesions containing a great many ~ atypical cells are still generally reversible. Epitheliali lesions consisting, entirely of cells with atypicali nuclei (and lacking cilia) are nt-oplasrns by definition. Perhaps all such neoplasms, or~ perhaps~ only some of'them, have the potential of invading un- derlying tissues at some later date. Thosewirch suchah potential may properly be callkd carcinoma in situ. Presumably, the neoplasms in which atypia appears to, be most extreme are those with the greatest poten+ tsal for invasion. However, t~hisopinion is~ based upon _ subjective impressions rather than upon conclusive evidence. We are strongly of the opinion that the patho- genesis of bronchogenic carcinoma is a continuum starting' with the occurrence of a few cells with atypical nuclei, the number of'such cells gradually in- Feb. 22, 1979 creasing until all those in a small mass have atypical nuclei„ and from there some' eventually invading un- derlying tissue. At every step' along th'is' road, the process is reversib'le.' There is even evidence thatt some invasive' carcinomas may' disappear spon- taneously, though this is probably an extremely rare event.'= Alllthe evidence suggests that the'progression~from. a' few cellls' with atypical nuciei to lesions consisting entirely of cells with atypical nuclei' (neoplasms) oc- curs over a period of' many' years; it seem likely that carcinoma in situ usually'remains in situ for a, long time before invasion oceun: If so, the decline over time in the proportion of cigarette smokers with large numbers of atypical nuclei in their bronchial epi- thelium should presage a decline in the deathirates of' cigarette smokers from lung cancer at some' future date. The evidence from this study is consistent with evidence from epidemiologic studies'•1D' indicating, that: death rates from lung cancer are lower among men who smoke low talr/nicotine cigarettes than among men who smoke the same number of'hi'gh tar/' nicotine cigarettes per day, and with the fact that dur- ing the last 25' years, there has been algreat drop in thee tar and nicotine content of the smoke' from cigarettes consumed in the United States. R>:remNcFa 1. Auerbach O: Stout AP. Hammond EC, et ab Changes in bronchial epithdium in relation to ci;arette smoking, and in relatioa to lung cancer. N Engl J I Med 263:253-267; 1961 Z Auerbach o; Stout AP, Hammond' EC, u a17 Changes in bronchial epithelium in,relation to sea, age, residence,,smoking and pneumoniy ' N EngIIJ'Med:67:1'111-11d, 1962 3:, ATierbach O. 5tout AP, Hammond EC, et al: Bronchial epithelIum in former smokers. Nl'Engl J Mod 267:11!9-12Si 1962 4. Hammond EC„Horn D: Smoking and death rates - reporTon fony- four months of follow~up of1'37,783 mea.lrAMA 166:1159-1'172,•,I'29X- 1308. 1958 S5. Report of the Advisory Committee to the Surgeon General of the Public Health Sarrica Smoking and health (DHEW Publication No. 1103). Washington. DC. Government Printing Office. 1964 6 Kah'n,HA: The Dortt study of smoking,and mortality, among U.S. veterans: report on eight' and one-half'years of' observation: epi- demiulogial study of cancer and other chronic discases: ,llatl!Cancer Inst Monogr I9sl-1'2b4 1966 7. DoII R, Hill .AB: Morulity in,relation to smoking:, ten _vearsl'observa- tions of British doctors. Br hlbd J I:1'399-1410. 1964 8. Hammond EC. Auerbach 0. Garfinkel, L, et al: Effects of cigarette 9. smoking,on dogs. Arch Environ Health1 ?1:741D.768. 1970 Wyndtr EL, Mabuchi K. Beattie El 1rt The epidemiology of lung, ancerr recent trends: JAMA 213:2221-2228;,1970 10. Hammond EC. Garfinkel L Sridman iH. et ali "Tad' and nicotine con- tent of'eigareue smoke in relation to death,ntes. En.iron,Res 1'2:263- 274„ 1976 11. Wynder EL, Mushinski M;,Stellman S: The epidemiology of'the less' harmful eigarette, Smoking and Health. Part I(D;H!EW Publication No. (NIIHI 76-1I221); Washington. DC. Government Printing Office. 1976 12 Everson TC. Col! WH: Spontuteous Rezr_.,sion of'Canoer: Phila- delp,hia, WiB'Saunders, 1966 i' 10105052"96`~