Tobacco Documents Online

:'_aims made agairst cigarette _n_or.siste:t, selective, =*-e stater.err.s o-" the Surgeon mi-c :a.a ieart Stud_v. 683 STATEMENT OF SHELDON C. SOMMERS, M.D. I am Sheldon C. Sommers, M.D., a physician specializing in pathology, currently Clinical Professor of Pathology at Columbia University College of Physicians & Surgeons, New York, N.Y., and University of Southern California School of Medicine, Los Angeles, California. Also I am consultant in pathology, Lenox Hill Hospital, New York; Chairman, New York State Mental Hygiene Medical Review Board; and President-Elect, Arthur Purdy Stout Society of Surgical Pathologists. I am past president of the New England Pathological Society and New York Pathological Society. Since 1936, except for World War II years, I have been engaged in medical research with particular reference to cancer, endocrine and gastrointestinal diseases, with over 300 publications about 10 percent dealing with lung and lung cancer, and also some on pancreatic cancer. I am coedilor of Pathology Annual and Diagnostic Gynecology 6 Obstetrics and serve on the editorial boards of five other medical journals. My curriculum vitae and publication list are attached. For the past six months, I have served as Scientific Director, Council for Tobacco Research USA, Inc. This is a funding agency for biomedical research in the area of smoking and health, funded by tobacco manufacturers. The budget for research grants in 1982 is 7 million dollars. Applications are acted upon by a Scientific Advisory Board, and those approved with a favorable rating

»+.r..~.~.,y.. 684 -2- - are funded for up to three years with an opportunity for continuation grants thereafter.. The Council for Tobacco Research exerts no in- fluence upon the grantees, who may freely publish what they find as they choose. About eighty active grants now exist in .~' the U.S. and abroad. My appearance at this hearing is voluntary, and the opinions expressed are personal, not representing those of any organization. They are the result of over 45 years of study, investigation and practice in the field of pathology and clinical- pathologic correlations of diseases, some of which have been at- tributed to cigarette smoking. In the field of science, knowledge is gained through experiment and interpretation, the scientific method. A theory is proposed. Thereafter, experiments confirm or refute it. If the latter, a new theory is developed. It is a continuous evolutionary process, and needs a critical and open mind. One must be constantly alert for surprises, as Lewis Thomas wrote. Lung cancer is high in the list of statements in Bill HR 4957 attributing diseases to cigarette smoking. There are two general methods of investigating the cause of cancers in humans or animals. one is the biomedical method. Cause might be defined biologically as something both necessary and sufficient to cause a condition. Cigarette smoking is not a necessary factor in human lung cancer, which existed for centuries in radium miners before cigarettes were invented. Lung cancer accompanies scarring processes in the lung due to TB, connective tissue diseases and various other > abnormalities in nonsmokers Also, cigarette smc criterion of being sufficier of smokers, more than'908 of lung cancer. Hence cigarett sufficient in the developmerr biological definition is not Like many other dis to be multifactorial, which many things, in addition to searchers now agree on this ~~... (1) Heredity. Fam Other families have decrease (2) Sex and Race. -.:ng cancer than women.v Bla incidence. (3) Urban. Certai rates which cannot be accoun- c::aracterized by severe wint, high air pollution. "' (4) Occupation. Ac involve increased risk. Some (5) Immune competer re~duced immunity have develoF some family cases, and the ir

. . ..., rs ~ 3rtunity for continuation Research exerts no in- blish what they find . now exist in , 5 years of study, athology and clinical- which have been at-, is+gained through .c method. A theory is _ .i _ . )r refute'it. If the continuous evolutionary . . One must be constantly d JEC~c: °t.7 T~~ statements in Bill HR 4957 There are`two general rs in humans or animals. Se defined 'biologically to cause a condition. in human lung cancer, s before cigarettes rring processes in the and various other 685 -3- abnormalities in nonsmokers. Also, cigarette smoking fails to meet the causal criterion of being sufficient. The fact is that the vast majority of smokers, more than 908 of even heavy smokers, do not develop lung cancer. Hence cigarette smoking is neither necessary nor sufficient in the development of human lung cancer, and by the biological definition is not the cause. Like many other diseases of older age, lung cancer appears to be multifactorial, which means the disease is associated with many things, in addition to smoking. Practically all active re- searchers searchers now agree on this point. These include: (1) Heredity. Families with lung cancer are known. Other families have decreased lung function. (2) Sex and Race. Men have three to six times more lung cancer than women. Blacks and orientals differ from whites in. incidence. (3) Urban. Certain urban areas have high lung cancer rates which cannot be accounted for by smoking. These areas are characterized by severe winter weather inversion patterns with high air pollution. (4) Occupation. As already noted, some twenty occupations involve increased risk. Some workers smoke while others do not. (5) Immune competence. Individuals with demonstrably reduced immunity have developed lung cancer. This may explain some family cases, and the increase of cancer with age.

686 -4- (6) Hormones. Adrenal and sex hormones accelerate scme metabolic processes leading to human lung cancer. (7) Aging. The mean age of lung cancer diagnosis has been reported as about 67 years, and said to be rising to older ages in some populations. Currently, researchers do not know which, if any, of these or other factors play a role in the causation of lung cancer. The other approach to cancer causation investigation is epidemiologic and statistical. Edpidemiologic statistics involves a.z experimental group and a control group. For a valid comparison, the groups must be alike as nearly as possible in all respects except for the item being investigated. In studies of cigarette smoking, matching smokers and nonsmokers by sex and age was achieved, and it has been assumed that in all other respects the two groups were comparable. This is not true, since in body build, extroversion-introversion, marital history, alcohol use, use of nonprescription medications, police records, military records and other aspects, cigarette smokers are demonstrably different from nonsmokers. The fallacy of a one-to-one comparison of smokers and nonsmokers with respect to mortality was pointed out in a monograph by Rose and Bell in 1971. They studied predictors of longevity in war veterans from Boston, re-examined at intervals. One-on-one comparisons placed cigarette smok'_.Zg ri as a predictor of early death, like many other studies. Kulti_actorial statistical analysis .-spped smoking to somewh -~,sfaction with job" be __=ively undeveloped st< _ _~«eeping conclusions. where do the datz __zions come from? The -__=icates. Most are nc -_:i certificate is an au ssion of the coroner a scientific document. =c:apared to autopsy di. _: certificates have be~ studies in the U.S. anc _: show this large erro: The basic data foz --:ty and uncertain verif the ACS dropped their -___s to have 20% minimum -cline by half or twe-tn _ so few that new diseas, °~ectively investigated. blamed on cigarettes amc -,.:nity hospital routine c =:e responsible organism. Beside the inadequa

I sex hormones accelerate some i lung cancer. )f lung cancer diagnosis has '. said to be rising to older ._r,,: . , _ ..-...., _. .. nvolves an experimental group nparisoci, the groups must be aspects except for the item -q. . . .. .. .. igarette smoking, matching 3 was achieved, and it has been ie two groups were comparable. i, extroversion-introversion, -. _ . . __ . - . .. .- ac - . - _ . nonprescription medications, ._~. .. _3•._ ... .:cI. other aspects, cigarette -om nonsmokers. comparison of smokers and was pointed out in a monograph ad predictors of longevity in at intervals. One-on-one #1 as a predictor of early ifactorial statistical analysis -5- dropped smoking to somewhere below ;30 as a predictor, and "dis- satisfaction with job" became f1. The lesson is that in the present relatively undeveloped state of epidemiology to beware of facile and sweeping conclusions. Where do the data on deaths from lung cancer and other conditions come from? The diagnoses are largely from death certificates. Most are not supported by autopsy examinations. A death certificate is an authorization for burial not requiring permission of the coroner or medical examiner. It is a legal but not a scientific document: When death certificate diagnoses - are compared to autopsy diagnoses of lung cancer,.errors in the death certificates have been found to range from 30 to 60 percent. New studies in the U.S. and other countries in the past two years again show this large error. The basic data for lung cancer incidence thus are of poor quality and uncertain verification. The problem is getting worse since the ACS dropped their requirement for accreditation of hos- pitals to have 20% minimum autopsies. All pathologists have noted a decline by half or two-thirds of U.S. autopsies. Soon we may have so few that new diseases like legionnaires' disease cannot be effectively investigated. Recall that legionnaires' disease was blamed on cigarettes among other things, and that it was community hospital routine autopsies which permitted identification of the responsible organism. Beside the inadequate epidemiologic matching and the ~ C..+ ~ 0 T N O CD

688 -6- defective lung cancer data from death certificates, one other event has occurred which bears upon the annual quantification of lung cancer. The rules have been changed. In the revision of the International Causes of Death, Sth edition, called ICD-S to ICD-9, for reporting mortality the two rubrics: -Primary Lung Cancer and Lung Cancer Not Otherwise Specified (NOS) were for the first time combined. Lung cancer NOS could begin in lung or have spread to lung from many different body sites. Lung metastases are among the three most common lo- cations for all major internal canc=rs. What in essence ICD-9 did was to guarantee a continued increase in the reported lung cancer mortality, conveniently disregarding that half or more of these cancers began elsewhere in the body and spread to the lungs. Two last points about statistical epidemiology. Every textbook states them. Every active scientist knows them. Epi- demiologic studies by the nature of the mathematics so far developed deal mainly with random popoulations. But smokers are self selected, as are nonsmokers. Comparisons of selected populations using mathematics valid only for random populations cannot be expected to provide valid answers. Second, epidemiology cannot prove cause and effect. All it can demonstrate is a relationship. The nature of the re- lationship, causal or otherwise, has to be worked out by other metr.ods, usually experimental. -he CTR in 1970 _-.:,estigate whether cir Almost 14 millic To take account o_ ,sted for and vaccir these animals cc -.ccer, pure chemica_ '2s. Biochemically, __ze these carcinogc s developed "humar =o these pure chem= :'^ereafter, thoi -ole cigarette smoi :-,cotine and high =:eir whole lives, ccere were 11,000 = 0:K_ng and control : <Dosure, practical= of squamous cell : smoking. There a d penetrated into ----.-vely. _.^. ather experi: -_r=iaogenic chemica. whole lives. =-.ience found after ?'= =^e smoking wa:

'8 6- _h certificates, one other the annual quantification of hanged. ernational Causes of Death, for reporting mortality the :3 Lung Cancer Not Otherwise time combined. Lung cancer NOS to lung from many different ;ng the three most common 10- ars. What in essence ICD-9 =rease in the reported lung :garding that half or more of a body and spread to the istical epidemiology. Every scientist knows them. Epi- the mathematics so far developed 3. But smokers are self selected, selected populations using opulations cannot be expected : prove cause and effect. :ship. The nature of the re- ; to be worked out by other 689 -7- The CTR in 1970 undertook a large scale research program to investigate whether cigarette smoking causes lung cancer in animals. Almost 14 million dollars went into the project in twelve :.ears. To take account of heredity, inbred mice were used, and they ,..ere tested for and vaccinated against respiratory viruses. To show that these animals could develop the major types of human Lung cancer, pure chemical carcinogens were introduced down their tracheas. Biochemically, these particular mice were known to -:etabolize these carcinogens into the biologically active forms. Abou*_ 20% developed "human type" lung cancers as a result of ex- ?osure to these pure chemical carcinogens. Thereafter, thousands of mice were exposed daily to fresh whole cigarette smoke of either low nicotine and high tar .r high nicotine and high tar content up to maximum tolerance during their whole lives, up to 40 months in some cases. At one zoint, there were 11,000 animal manipulations per day, including sham smoking and control mice. After all these years of cigarette smoke exposure, practically zero lung cancers developed, and not one case of squamous cell carcinoma, the human cancer most often olamed on smoking. There was no question that the cigarette smoke had penetrated into their lungs, since this was worked out quantitatively. In other experiments, mice were primed with intratracheal pure carcinogenic chemicals and then exposed to cigarette smoke during their whole lives. No increase in lung cancer occurred over the incidence found after using pure chemicals alone, and in one experiment the smoking was associated with a reduced lung cancer

690 -8- rate. The numbers of animals and the comouterized information are so abundant that statistical analysis by modern experimental methods is still continuing. Some 40 publications have in part already appeared or are in prospect. In summary, a massive experiment to demonstrate that cigarette smoking can cause lung cancer in animals has proved negative. One knows how important an experimental model is in cancer research from the excitement that attended the claims years ago that cancers of lung or larynx had been produced by cigarette smoke in animals. However, no model so far develoned withstands an objective analysis of the pathology of the cancers. c.-ea, ...-. - L_ 19'_5, India, +?ol?s, rr-ec: So•.e-bc: 9. 1943, Edith . cun laude , 'arvard `:edCcal Sciool, 19G: ;,..'_versi-- Clinicr and 3es~dent ir. --.-..; .'.es_ce-.t in ?a:^o.egy, F . .~n= _- __ r. _.. . r._ _. :.. _ cenry. Ford ..__.. --_:.. Cert_•_oate in Path: -- - -c_-_ic9's• `:e_ Englanc Earvard canc - ..-" :Lassar..csetts Xe-oriai .. ._cc„ -occe•;- of .a_..clo;;ists .. -cr.a_ }:Os.,: :aSoratories, .. ___- ---= -f =a'=oracories, Lenox Hi'_ _- _.. _a_.ctcoy, Harvard J:ec alleged :._ :.c:essc: o:~Pa,^clogy, . -_ :roce<sc: o: Pat:olos?, Cn- -'.-: cancers, C. .. .:_[csso: cf . , E_...._~ Star, Croi:: de . remains a major biological mystery. Epidemiologic studies report a statistical association between cigarette smoking and lung cancer. However, the biomedical experimentation does not supnort the smoking causation hypothesis. ?achclc -itrastr::. 1991- -_.- ., _cal E-iev Boarc. --"-• Societq of Surgic In summary, lung cancer, like many other human

:terized information :y modern exnerimental _cations have in part to demonstrate that animals has proved :imental model is in _tended the claims ad been produced by Ddel so far developed :hology of the alleged :v other human cancers, ,miologic studies report = smoking and lung tation does not support 691 CC9_".ICL•L'.'H 4IT.`.E _:-.z..:..c ... So.-.ers, ...... ..L 7, 1916, Indianapol'-s, Indiana .._rr_ed: :lever.Scr 9, Edith Briggs, no children :.acvard Co aeoe, 1937,~cua laude ...- , -..cvard Medical School, 1941, cum laude ..•_ -;o L'~iversi- Clinics, L°=1-42 3es~d~_nt and Aes`cer.t ic Pathology, New England Deaconess Hospital, Boston, 1946-48 .s__5ta._ F,esi.cen[ i.. _ t:noicgY, Free Hcspital for Nor..en, Brcokline,."1ass., 1945 .-._____cr.t 3esi?e-: i.. ?at^.clcep, Boston L•ring-In Hospital, Bos^,or, 1948-49 ~es=ce-t i. ?az..o'.c;Y, Henry Ford Hespital, Detroit, 1949-50 -_-.er'_c=- ~o acd Certif.cate in Pathology (Clinical Pathology - Patholegic Anatocry), 1950 .:-soci=te =c_colc_z.s;, `!ev England Deaconess Hospital, Boston, 1950-53 Harvard Canczr Coraissi^ ,$osto.^., 1950-53 .a-sacr.•:setts `:er..orial Hospitals, Boston, 1953-51 _-- -- _-r.c Societ, of PatSe'_c;;ists; President. 1959-60 c___t, Scri(er.or al Hospital, La Joila, Cal-:fot-iia, 1961-63 __=-.,c=r.-_ -:rectcr of i.aberacor=es, Francis Delafield Hospital, New York, 1963-67; _-_'-58 --rectcr c: -aSorarozies, Leno>: Hill rospital, New York, 1968-d= =~-....-..c .a . Harvazd >tedica. School, 1948-49: instructor in Pathology, 1950-52; _=t_ _c .~t`:c_oa7, 1.1_-- Lecturer in Patholo5y, 1954-61 r-cfessc_ of P,:thciog;,J3oston Cniversity School of :4edicine, Boston, 1953-61 :=c:_sscr of ?at:olo5}•, tiniversity of Southern California School of Medicine, -:-=-=-_ =..,.ecscr of Colu-:h:a Cnzvcrsi;;r, College of Physicians & Surgeons, \.Y., of ?at:olo_y, 1965-5?; Clinical Professor of PatholugY, 1955 - ..___.a_ Uc:i,=_d States .. -y, 1943-46 ...o', Croi:c de Cuerreand Presidential Lnit Citation SecrctarY 1959-7'_; Presidcnt 1977-79 . "56 -; ^at!.oloZy Dec~nniala, 1975; Co-Lditor, -- 1979 - -~-:^oi c` Surr,ical Pathologp, Hu^nn Patholocy, Crologic nad'.olo,^,Y -']c;astr.~ctural Pathnlo,v :or TnSacco Research, 1°E7-; P.esearch Director, 19FP-72 Di= _ctor, 1981- cal Review Board, New York State, Chain~^.an, 197F- Soc-ietp of SurRical°atholegists, President-Elect, 1981 _ ?othalc- __ ___------ ___ ___ u:pcri-en:a1 :ieloh/ anL ."fedicina :Iecical Societ^, ::ew York - --a__'.nn of Path.olodisa ' _ticcl ?at;clocists _ _ccict•i -

692 PUBLISHED ARTICLES Stel3cn C. Sommers,' M.D. 1. Jacobs, J.L, and Sommers, S.C.: The specificity of formolized proteins. J. Icmiunol., 36: 531-54% 1937. 2. Menkin, V., Kadish, M.A. and Sommers, S.C.: Leukocytosis-promoting factor in inflammatory exudates of inen. Arch. Pzth., 33: 188-192, 1942. 3. LeCompte, P.M., Sommers, S.C. and Lathrop, F.D.: Tumor of carotid body type arising in the middle ear. Arch. Pathl., 44:78-81, 1947. 4. Warren, S. and Sommers, S.C.: Cicatrizing enteritis (regional ileitis) as a pathologic entity. Am, J, Path., 24:475-501, 1948. 5. Warren, S. and Sommers, S.C.: Giant-cell inclusions in cicatrizing enteritis. Proc. Soc. E.per. Biol. & Med „ 8: 461-463, 1948. 6. Warren, S. and Sommers, S.C.: Pathogenesis of ulcerative colitis. Am. J, Path., 25: 657-659, 1949. 7. Hertig, A,T, and Sommers, S.C.: Genesis of endometrial carcinoma. I, Study of prior biopsies. Cancer, 2: 946-956, 1949. 8. Soarsers, S.C., Hcrtig, A,T, and Bengloff, H.: Genesis of endometrial \ carcinoma. II. Cases 19 to 35 years old. Cencer, 2: 957-963, 1949. 1 9. Hertig, A.T., Soc¢ncrs, S,C, and Bengloff, H,: Genesis of endometrial carcinoma. III. Carcinoma in situ. Cancer, 2: 964-971, 1949. 10. Sommers, S,C „ Lwley, T,B. and Hertig, A.T.: A study of the placenta in pregnancy treated by stilbestrol. Am. J, Obst. 6: Gynec., 58: 1010, 1949. 11. Warren, S, and Sorarzers, S.C.: Proteolysis in intestinal disease. Gastroenterology, 14: 522-526, 1950. 12. Wyatt, J.P, and S.mners, S.C.: Chronic marrow failure, myelosc'-erasis and extramedullary henatopoiesis. Blood, 5: 329-347, 1950. 13. Mcissner, W,A, and Sorsvers, S.C.: Postpartum andometrial hyperplasia in diabetics treated with stilbestrol and progesterone. J, Clin. Endocrinol„ 10: 603-609, 1950. 14. Sommers, S,C, and Johnson, J,H.: Congenital tricuspid atresia. Am. Heart J., 41: 130-143, 1951. 15. Sommers, S.C., Wilson, J,C.. an poliomyelitis, J. E.:per. Mod„ 16. Werren, S., Holt, M,W, and Soc of ionizing radiation. Proc. S 17. Holt, M.W_ Sommera,.S,C, and for quantitative histochemical :8. Sommers, S,C,. and Tcloh. H_: 0 Arch. Path., 53: 160-166, 1952 Warren, S., Holt, M,W, and Sor., studies of radiation reaction. 30. McManus, R,G, and S=mers, S.C cortical stromal hyperplasia. 21.. Sz;mners, S,C, and Young, T.L.: 28: 673-689, 1952. Edwards, J.L. and Sommers, S.C J, Lab, and Clin. Med „ 40: 34 Chute, R,N, and Sommers, S.C.: parabiosis intoxication, Blood Hollander, A, and Sommers, S.C. 247: 634, 1952. Christensen, W,R „ Sommers, S,~ roentgen rays on the rabbit sk ccaers, S.C. Chute, R.N, and human cancer. I. Irradiated ra Chutc, R.N., Sommers, S.C. and hunan cancer. II. Hamster chee: S,amers, S.C., Sullivan, B.A. : hunan cancer. III. Chorioallan: Rcs., 12: 915-917, 1952.