LETTERS

and death indicated strong risks-similar in magnitude to 1 Peto R. Smoking and death: the past 40 years and the next 40. Smoking those found by later studies-whatever control BMJ 1994;309:937-9. (8 October.) 2 Proctor RN. . Medicne under the Nazis. Cam- Public health measures were taken more group was used (table); these risks were not bridge MA: Harvard University Press, 1988. than 40 years ago explicable in terms of changes in smoking 3 Reiter H. Ansprache bei der Er6ffnung des 1. Wissenschaft- behaviour induced by illness. Schoninger's study lichen Instituts zur Erforschung der Tabakgefahren an EDrroR,-Richard Peto discusses the failure of cost 173.30 of the 100000 Reichsmark given by der Friedrich-Schiller-Universitit Jena am 5 April 1941. public health with regard to smoking over the past from his personal funds to the Jena ManchenerMedizinische Wochenschnift 1941;88:697-9. 4 : stimulants endanger public health. JAMA 1939;12: 40 years,' but his account could have extended Institute. 2339-40. back further. He considers that the Medical 5 Lickint F. Tabak und Organismus. Stuttgart: Hippokrates- Research Council was, in 1957, "the first national Verlag, 1939. institution in the world to accept formally the 6 Davey Smith G, Strobele SA, Egger M. Smoking and health promotion in Nazi . I Epidemiol Community Health evidence that tobacco is a major cause of death." 1994;48:220-3. The formal recognition of this fact, however, was 7 Schairer E, Schoniger E. Lungenkrebs und Tabakverbrauch. explicitly made by many important national insti- ZeitschriftftilrKrebsforschung 1943;54:261-9. tutions in . The Public Health 8 Schoniger E. Lungenkrebs und Tabakverbrauch: Inaugural- Office and the German Medical Association, both Dissertation zur Erlangung des Dokworgrades einer Hohen Medizinischen Fakuklir der Friedrich-Schiler-Universitat zu under the leadership of Dr Gerhard Wagner, Jena. Jena: Friedrich-Schiller-Universitat, 1944. repeatedly issued precise pronouncements 9 Burleigh M, Wippermann W. The racial state: Germany 1933- regarding the dire health consequences of 1945. Cambridge: Cambridge University Press, 1991. smoking. By 1939 Wagner's successor, Dr 10 Bernhard P. Der Einfluj3 der Tabakgifte aufdie Gesundheit und die Fruchtbarkeit der Frau. Jena: Verlag von Gustav Fischer, Leonardo Conti, had established the Reich Bureau 1943. Against the Dangers ofAlcohol and Tobacco.2 The 11 Breyvogel W, ed. Piraten, Swings undjunge Garde: Judendwider- Reich Health Office also made numerous state- standitm Natialsozialismus. Bonn: Dietz, 1991. ments, which its president, Hans Reiter, reiterated 12 Peukert DJK. Inside Nazi Germany: conformity, opposition and at his inaugural address at the opening of the "first racism in everyday life. : BT Batsford, 1987. 13 Brenner H. A birth cohort analysis of the smoking epidemic in scientific institute for the struggle against the Health dangers of tobacco" at the University of Jena in West Germany.JYEpidemiol Community 1993;47:54-8. 1942.3 The , under the leadership of Robert Ley, ran many campaigns Edelweiss Pirates enjoy a cigarette: was smoking a way of showing disobedience in Nazi Germany? (Reproduced Author's reply highlighting the damaging effects of tobacco on from ref11 bypermission ofthepublisher) health. Like Adolf Hitler, Reiter, Ley, Wagner, EDITOR,-A qualitative difference exists between and Conti were, publicly at least, vehemently the strength of the scientific evidence about against smoking.245 The antismoking activities in Nazi Germany cigarette smoking and lung cancer that was avail- Recognition of the damaging effects of smoking were tied, institutionally and ideologically, to the able in the 1930s and that available in the 1950s. on health led to much antismoking legislation6; this racial hygiene movement,6 which was deeply Assertions by German institutions in the 1930s can included legislation banning smoking in public implicated in the campaigns against , hardly be described as "acceptance ofthe evidence" places by those under 18 and prohibiting both homosexuals, travellers, and those deemed to be as they predated even the early publications of the tobacco advertising and smoking in public mentally and physically defective.9 Tobacco was studies by Miiller and Schairer and Schoniger on buildings and on public transport. Pregnant considered to be a "genetic poison," decreasing lung cancer,12 and the evidence on other major women and those deemed to be sick because of fertility and increasing the incidence of chromo- diseases came still later. In retrospect, those and smoking had their tobacco rations withdrawn, and somal damage. other early publications, such as the analyses in there was serious discussion regarding whether The Jena Institute carried out both clinical and 1938 by Pearl of total mortality from smoking those sick with illness caused by smoking should animal research into these topics.'0 For the among holders of life insurance policies in the receive medical care equal to that given to patients anti-Nazi youth movements-the working class United States,3 should have been taken more whose illnesses were not considered to be self Edelweiss Pirates and the bourgeois Hamburg seriously by medical scientists, but even in the late inflicted.' Swing Youth alike-the constant cigarette seems 1940s they were not. Moreover, both of the to have been almost a badge of resistance (figure) German studies were small (with a total of 179 lung Odds ratios (95% confidence intervals) for lung cancer, and was referred to as a sure indicator of their cancers, as against 2150 in the publications in with different control groups, calculated from data of degeneracy in the surveillance reports produced by 1948-504). Thus what distinguished the official Schairer and Schi5niger'8 the ." 12 Indeed, one of the reasons for statements about smoking by the British and the relative failure of activities to prevent smoking other authorities in the late 1950s was that they Control group in Germany since the war"3 may be that the represented, for the first time, a consensus based Smoking General All other Cancers unrelated association of authoritarian antismoking efforts on substantial evidence. category population cancers to smoking* with the Nazi regime remained in popular memory Finally, there are no obvious ties between the for a long period.6 Nazi campaigns of racial murder and the current Very heavy 16-6 (4 to 91) 8-8 (2 to 49) 10O4 (3 to 58) antismoking movement-rather the reverse, in Heavy 5 8 (2 to 32) 5-6 (1 to 32) 6-6 (2 to 38) A travel grant from the Wellcome Trust allowed Medium 7-8 (2 to 42) 7 0 (2 to 38) 7-7 (23 to 42) inspection of relevant archives, including those of the fact, since Hitler before 1945 and tobacco since Moderate 1-6(0-4to9) 1-4(0-4to9) 1-4 (0 3 to 8) 1945 both caused about 50 million deaths.5 Non-smoker 10 10 10 Institute for the Struggle against the Dangers of Tobacco. RICHARD PETO Test for trend Professor ofmedical statistics and epidemiology (P value) <0 00001 <0 00001 <0 00001 GEORGE DAVEY SMITH Professor ofclinical epidemiology ICRF Cancer Studies Unit, University ofBristol, *As defined in study. Radcliffe Infirmary, Bristol BS8 2PR Oxford OX2 6HE SABINE STROBELE Research continued alongside these campaigns Teaching associate against smoking. Many would think that the case- Institute ofMedical Sociology, 1 Mailler FH. Tabakmissbrauch und Lungencarcinom. Zeitschrift control study of smoking and lung cancer con- University ofHamburg, fdrKrebsforschung 1939;49:57-85. Hamburg, 2 Schairer E, Schoniger E. Lungenkrebs und Tabakverbrauch. ducted by Sch6niger constituted clear evidence Germany ZeitschriftfarKrebsforchung 1943;54:261-9. that smoking was a major cause of death.'8 This MATTHIAS EGGER 3 Pearl R. Tobacco smoking and longevity. Science 1938;87:216-7. study avoided one of the methodological problems Senior research fellow 4 Doll R. Introduction and overview. In: Samet JM, ed. Epi- of later studies by including both general popula- Department ofSocial and demiology of lung cancer. New York: Marcel Dekker, 1994: tion controls controls Preventive Medicine, 1-14. and hospital and investigated University ofBeme, 5 Peto R, Lopez AD, Boreham J, Thun M, Heath C Jr. Mortality whether changes in smoking consequent on sick- CH-3012 Beme, from smoking in developed countnes 1950-2000. Oxford: Oxford ness could have biased the results. The findings University Press, 1994.

396 BMJ VOLUME 310 1 l FEBRUARY 1995