Tobacco Documents Online

! ~ Decemf"20, 1982 MEMORANDUM TO: W. T. Hoyt ~ FROM: Hilda Zahn SUBJECT: American Public Health Association CC: WDH SCS RFG RCH Annual Meeting, Montreal, Nov. 14-18', 1982 The annual meeting of the APHA covered all aspects of its theme, "Aging and Public Health," a very topical subject. Smoking issues are always in evidence at APHA's annual gatherings and this year's session was not different; as in re- cent years, little that was new was on the program. An~exception was a report"-by REIMERT T. RAVENHOLT,-long- time antismoker, on the "bodywid~e spectrum of diseases caused~by the 1918 swine influenza: pandemic. The virus, he postulated, compromised~many organs in the body, notably those of the res- piratory and central nervous systems and the heart. Though~ smoking is the greatest danger to the heart, he said, altered susceptibility to many of the diseases associated with these organs may have begun with the 1918 pandemic. (See No. 1 below) The lengthy 1981 study by RICHAR'.D DOLL and RICHARD PETO,was politely shredded by PE'TER'F. INFANTE of the Occupational Safety and Health Administration (OSHA). His remarks were personal, not OSHA's, he said~. (Perhaps his reticence was due to recen~t occa- sions of reprimands or worse for government scientists making, comments inimical to commercial interests.), He accused Doll and Peto of ignoring evidence and using,inconsistent criteria ini their work. Nor was he pleased1with ERNST WYN!DER, GIO GORI, JOHN! HIGGINSON1, and CALUM MUIR (the last-named, it will be re- membered, an epidemiologist at the International Agenicy for Re- searchion Canicer in Lyon, France). (See No. 2 below.) SAXONIGRAHAM of Roswell Park Memorial Institute in Buffalo, N.Y., gave.his auldience a little jolt when he said'that reti- noids (vitaminiA analogs) anid vitamin C increa,sed the risk of developing prostate cancer. This conclusion, he.noted, was at odds with TAKESHI HIRAYAMA.'s reported findings. (See No,. 3'be- low.) One recenit study of interest, from San Antonio, found that employment per se for women had a very positive effect on their HDL cholesterol an~d lowered triglycerides. These favorable dif- ferences may be more significant for corona~ry well-being than eonard _" ]_ __ PUBLIC REIATlON$.COUNSEL~ /WaIn~16dI?.'AIV ssoc.iates,~lnG .. (P. 0. BOX 223) 13 LINCOLN ROAD • GREAT NECK, N.Y. 1102L •(212) 895-7445

I . 2. C the "unfavorable differences in cigarette smoking," according,to~ the author. (See Nlo. 4 below.) Press co-verage was not extensive, consisting mostly of re- porters from local newspapers and from a variety of medical and' paramedical publications. A reporter from the Washingtonibureau of the "New York Daily News" showed up, but she had a freelance assignment from,"Medical World News." (An article oniReimert appeared in the magazine's Dec. 20 issue.) A freelanice writer came from Philadelphia to cover for the news sectioniof the "Journal of the American Medical Association." The highlights: 1. "Late effects of 1918 influenza" -- The pathiologic pic- ture of the 1918 influenza victim shows a tremiendous impact on the respiratory, cardiovascular and central nervous systems, said Reimert, who is director, World Health Suurveys, Centers for Disease-Conitrol, Rockville, Md. The flu probably has con- tributed to incr.eases in a "bodywide spectrum of diseases."' He presented a stuldy of death records in Seattle-King County, Washington, of mortality trends for many diseases inithe U.S. since 19001and of the literature. (He also referred, to a recent article by him and WILLIAM H. FOEGE in the Oct. 16, 1980, "Lancet" titled "191!8 influenza, encepohalitis lethargica, parkinsonism.") The probable cause of the flu was the swine virus, which propagates primarily in the respiratory epithelium and then is disseminated tolother tissues. But, he said, the first site of onslaught is the respiratory system,. In 1920, it was found that epithelial proliferation was solextensive, it couldn't always be differentiated' histologically from invasive cancer. In 1925, it was suggested that the flu might have contributed'to the then, incipient rise in lung cancer. About 20 years.ago, Ravenholt and Foege, after reviewing lung cancer death records covering the period 1881-1960, noticed the evolving,age-splecific trends of lung cancer and suggested that the 191!8 flu may have been a contributory factor. It was found that total mortality fromilung,cancer had risenirapidly in older males since 1930. It increased steadily by about 50% every 5 years from 1935-55 and by 23% from 1955-60. T'his sLackening rate of increase, the result of the leveling off of mortality rates among younger males, suggested at the time that the lung cancer epidemic was nearing its crest, or least a plateau. "Analysis of mortality from cancer of the trachea, bronchus and lung by birth cohort and age at death for white males in the U.S. during 1947-77 shows flattened death rate curves for birth cohorta born after 1922'. Hence, although smoking has certainly been the dominant cause of lung cancer (the same principles apply to cancers throughout the respiratory tract, and to emphysema) during this century, suspicion lingers that the 1918'

s, r C 3. influenza (and' other influenzas) may have laid a ground-work of alteredisusceptibility for the more direct role of tobacco in the production of the pandemic of lung cancer..." R'avenholt said. ~ "Overall, heart disease trends are compatible with a pos- sible late effect of 1918 flu,-- with the falling age-adjusted rates during recent years possible due to the die-off of an in- fluenza-damaged cohort. But there is ample evidence that the smoking of cigarettes is a foremost direct cause of coronary heart disease (CHD); and much of the recent fall in CHD mortal- ity is no doubt due to improved control of smoking and~hyper-. ension," he continued. The epidemliologic evidence he presented is "suggestive,"i he acknowledged, and needs intensive work, but he feels the 1918 flu can serve as a kind of "'R'osetta Stonie" for diseases that_ have been pandlemic since 1918. 2. "The contribution of occupation to environmental cancer" -- INFANTE, Washington, D.C. Infante prefaced his talk with the explanation that his remarks were personal opinionis, not OSHA"s, the agency for which he works. He proceeded to examine and criticize the value of some recent studies of occupation as it relates to cancer risk. Estimates for determining the contribution of life style, occu- pational exposure and other factors have been limited by "use of no criteria or inconsistent criteria," he said. He cited a 1977 report by WYNDER'and GORI which stated that from 1-10% of all cancers were caused by occupatiorral exposure. No methodology was given, he said'. Nor was there documentation or methodology in other research papers, including one by MUIR and HIGGINSON. A 1978'government report attributed 20-30% of cancer to present and~ future occupational exposure. The report dealt with only six occupational carcinogens: asbestos, arsenic, benzene, chromium, niickel anid petroleum fractions. Even so, it was crit- icized by the American Industrial Health Council and Doll and P'eto for overestimating thie current occupational exposure of these substances because cancer today is caused by past expo- sures that were higher than current ones. That was correct, Infante said, but 39 additional substances inithe workplace withi either a confirmed or suggested causal relationship to cancer weren't included inithe estimate. Animal studies weren't even considered. "These factors lead to a bias for an understatement of the occupational contribution to the cancer burden,` he said. O G% Doll and Peto's large and'comprehensive report had excel,- _j lent documentatiion on cancer causation, he said, but there were ~W some gaps for occupational factors. They used inconsistent cri- co teria on how much cancer is caused by smoking, infection, etc. Q) Even they agreed that their estimates were shaky and of uncer- CO

C ~ 4. tain reliability. Doll and Peto estimated 30% of all cancers were due to :'smoking. Ther-e are probliems with their methodology in this re- gard:, Inifante continued',. People who smoke do other things -- and, we don't know about smoking, and synergism and! smoking. "For ex- ample, if they took nonsmoking rates and apply it to smoking rates for asbestos workers, they wouldlconclud'e a 50-fold risk in lung cancer, and'this all due to cigarette smoke (sic). From data'we now know, of that 5:0-fold risk, 80% of it could be elim- inated ilf asbestos exposure was eliminatedi. This shows the haz- ardis of.the approach Doll and Peto used~." Th~e Briton~s supported their smoking,etiology claimiby cit- ing animal studies, experimental bioassays andishort-term muta- genicity tests. Yet, they complained, since experimental studies on alcohol and cancer were negative, this was an example of how unreliable experimental studies are. But their figures on diet were supported by experimental results, the same kind for which their alcohol findings were denigrated!. Inifante continiued to criticize the Doll-Peto report: it contended that 10% of cancers are caused by infection, but this is not based oniany concrete evidence; its occupational cancer findings were based~on epid'emiologic results from 18 substances in the workplace, with confirmed evidence in humans, but.it ig- nored 27 additional substances where evidence either has been confirmed or suggested!; it ignored experimental evidence en- tirely for an additional 85 substances. A different study worked on by DolJl and Peto concluded that 10%0, plus-minus 5%, of cancer deaths resulted from occupation,. On the basis of the quality of data available, this wasn't an unreasonable estimate. This mid-range estimate results in 20-60,&00 cancer deaths yearly that are related to the occupa- tional setting. "Why are we tolerating this?" Infante asked. 3. "Diet in the epidem~iology of cancer of the prostate" - GRAHA,M, Buffalol, N.Y. After reviewing questionnaires on. diet given to bothiin- and outpatients inithe years 1957-65 at Roswell Park, Graham selected for study 311 white males with histologically proven prostate cancer and 294 control's. Like some earlier studies, he found no differences in cases and controls with regard to socioeconomic status, weight, hor- O monal status, or ingestion or cruciferous vegetables or grains. ~ What he did find was a higher risk for prostate cancer with higher intake of vitamins A and C. The results were statistical- w .C+ ~ ly significant. Th~is apparent enhancement of risk associated . with retinloids was a"surprising departure," he said. Past stud- ..':O

r r C 5. ies have showed a reduction~in risk with the use of retinoid~s for cancer of the mouth,, larynx, lung, bladder, breast, and cervix. All except blad'der and! breast are of squamous cell der- ivation. "Itk;s possible that risk reduction associated with retinoids is specific for tumors of squamous or similar types (of cancer)~," he said. Prostate tumiors are glandular in origin. "Our findings are at odds withitholse of Hirayama who ob- served a lower incidence among men who frequently at-e green and yellow vegetables," Graham said!. "'Our differences may be attrib- utable to the methodologies of the two studies." His group found~greater.use of fatty foods among prostate cancer patients, as have other investigators; Hirayama's large cohort study didn't find this. It's important to note that risk associated with diet has almost always been small and that there have been far too few studies to warrant firm conclusions, he said. But it's reason- able to suggest that specific nutrients may reduce the risk of certain kinds of cancer and increase the risk of, or not affect, other kinds. Graham agreed with and quoted ROSWELL BOUTWELL's statement: "We conclude that the prevention of cancer by retinoids is not universal." He feels that the field of cancer and nutrition is much more complex than suspected. The effects of nutrients may vary from disease to disease. 4. "Work status and women's protection against coronary heart disease (CHD)'" -- HE'LEN P. HAZUDA, San Antonio. . This study was prompted by the limited evidence available concerning the relationship between employment and coronary health in women,. Data were collected between October 1979 and August 1981 ass part of the San Antonio Heart Study. A community-based sample of Miexican-American and Anglo-American women was used with socio- economic status as a control. Cardiovascular risk factors ex- amined were obesity, lipids and lipoproteins, blood pressure, and smoking. Levels of HDL chol!esterol, reported to protect against CHD, were significantly higher among currently employed womenithan housewives in both ethinic groups, while levels of triglycerides, a positive factor for CHD, were significantly lower. A significant difference between working and'nonworking women was found for cigarettes smoked per day. Contrary tolfind- ings for lipids and~lipoproteinis, this difference favoredihouse- wives, especially Anglo-Americans. Since smoking appears to have a greater negative impact on CHD inimen than women, the favor-

C 6. able difference in lipids and lipoproteins experienced by em- ployed women may be more significant for coronary health than ,the unfavorable difference in cigarette smoking. The levels of total serum cholesterol, systolic and dia- stolic blood~pressure, and obesity were quite similar in cur- rently employed women and housewives. 5. "Smoking and lung,disease in textile workers" -- MARY KORNIGUTH, Nashville, Tenn. When a textile workers tries to bring a compensation case for byssinosis or brown lung disease to a hearing, the case is usually lost if the person has a history of smoking or is cur- rentl'y smoking, Korniguth saiid . It's a foregone conclusion withisome persons that one can't have irreversible effects of brown lung disease unless one smokes. It's assumed~ that cottonidust reacts similarly to asbes- tos -- effects are multiplied with smoking. Korniguthicited somee studies that refute this assumption. A survey was done of clothing,workers by the Tennessee BrowniLung,Associiation. Byssinosis has been associated with workers handling raw cotton. However, these were garment workers and yet 30% of those who came for screening had some pulmonary dysfunction. "Though smoking appears to reduce FEV1i (forced expiratory volume in one second!) in a large number of the workers, those who niever smoked'i had! r•ediucedi flow rates too liarge to igniore," Kornguth said. Forced vital capacity (FVC, another pulmonary function measurement) was reduced 5% by controlling for smoking. Smokers had an additional 1.5% reduction in FVC. This meant that the effect of the dust on FVC was four times higher than the effect of smoking, she said (sic). The opposite effect was experienced for FEV. FEV was reduced 0.5% (per year) from the processed dust and 1.4% from smoking. On average, the workers had been employed for 22 years in low dust clothing manufacturing. The number of people in the study was small, but Korniguthithought the results showed~ the confounding effect of smoking on pulmonary function. Beyond smoking, there was a significant relationship be- tween reduced'pulmonary function and dyes in the materials and chemicals for permanently pres'sed'fabrics. This relationship was stronger than that seen for either smoking or cotton dust.. She said' the garment workers she studied hadinot been con- sidered to be susceptible to occupational respiratory disease.

r i 6. "Oral contraceptive use and circulatory systemidisease mortality in T'aiwan"'-- L.P. CH'0W, Baltimore. He was absent and the paper was given by EZATOLLAH KEYVAN, a: colleague. ~ This case-control study of oral contraceptive (0C)' use in Taiwan wa;s undertaken because of the adverse finding;s in western countries. Since OCs are an effective means of populatiion con- trol inideveloping countries, continued~ use has raiised~ concern among health workers. Th~e study, using death certificates and family interviews, covered 306 women who had died of circulatory systemidisease (stroke,, ischemic, etc.) 305 deceased controls who died of other causes and 611 healthy controls. Overall, preliminary results did not show a strong associ- ation between OC use and circulatory diseases. Probable explana- tions could~includ~e genetic makeup a:nid the low rates of ciga- .. rette smoking among Taiwan~ese women (~it's only about 5%). In.ad- dition, the prevalence of OC use was found to be low (15%)' and the duration short -- between 14-2'1 months. Larger samples of greater duration are needed to show more significant results. 7. "Cancer incidence and mortality among U.S. Chinese" - FRANCES B. LOCKE, Bethesda, Md". This poster presentation outlined findings from data ob- tained in th~e years 1960-77. Age-adjusted incidence rates for cancer of the liver, thyroid and nasopharynx were higher for Chinese of'both sexes than for whites of both sexes. Chinese maLes h~ad lower rates for prostate, bladder, lung, melanoma, nervous system, lymphoma, and leukemia than white males. Chinese women h~ad lower rates for breast, uterus, ovary, bladder, melanoma, lymphoma, andileulkemia than white women. Chinese males in Hawaii had higher rates of prostate and thyroid cancer than those in San Francisco; those in San Fran- cisco had higher rates of lung,and liver cancer. Rates areg rising for Chinese males for lung, prostate and stomach cancer; for females it's lung, breast and uterus. (Locke an~d her associate, HAITUNG KING, told me that Chinese smoke less than whites.) 8. "Risk of myocardial infarction in relation to person- ality type in men and women under 50 yea:rs of age" -- SUSAN HELMRICHI, Brookline, Mass. Is a particular behavior pattern, Type A, related to risk of myocardial infarction (1MI)?'Daita collected'by the Drug Epi- demiology Unit in Cambridge, Mass., from two hiospital-based, case-control studies of first, nonfatal MIs were analyzed to

test, or least further confirm, thiis theory. From180 hospitals in the northeast U.S., 1,357 male Mil pa- tients under age 55 were comipared with 2,502 controls. From 155 h~ospitals in the same area, 250 female MI patients under 50 years were compared with 675 controls. Both groups were con- trolled for possible confounding factors including age, family history, diabete&, and cigarette smoking. Type A behavior was foundito be an independent risk factor for MI in both sexes. Type A's effect seemed significantly stronger for meniin all groups. For women, the effect was just barely significant,though it seemed stronger among hypertensives, working womeniand particularly those with more than1five child~ren. Helmrich cautioned that Type A was not the same as stress, but was an habitual response to circumstances such as excesses of drive, hostility, impatience, and'competitiveness. 9. ",Drinking, smoking an~d women's health: a Canadian per- spective" -- ROBERTA G. FERRENCE, Toronto. - Compared to their U.S. counterparts, Canadian women,, es- pecially in Quebec, are more likely to drink, but not heavily. They're also more like to smoke, but the total' consumption of cigarettes isn't higher than for U.S. women. Ferrence's study found that the mortality rate for lung cancer in Canadian women was substantially lower than in U.S. women. In 1976-77, the rate for U.S. women1was 17.2 per 100,000, fifth highest of 46 countries. Canada: was 11th with 12.5 per 100,000. Cirrhosis rates for Canadians were only slightly lower thanifor U.S. women. There's "much evidence that levels of use of both alcohol and tobacco are connected to availability of disposable income," F'errence said. Yet, a Canadian Public Health Association survey found that smoking among doctors varied from 44% in Newfoundland to 8% iniBritish Columbia. Smoking programs should be aimed at women in blue and white collar work because they'll be the last to give up smoking, she said. However, she added that these women were most likely to be exposed to toxic agents that act synergistically with tobacco. 10. "Cigarette smoking and~blood lead levels in occupationi- ally exposediworkers" -- C. PERRY BROWN, Los Angeles.. Among,workers in two car battery manufacturing plants, smokers were found to have significantly higher.blood lead lev- els than nonsmokers. Brown theorized that, in smokers, decreased cilia activity and mucus production, coupled with in~creased per- meability of the alveolar walls, allowed~more lead to reach the

blood. Smelters.in a lead recycling plant also were studied. Even~ with the use of respirators, these workers had blood lead levels thalt were much higher than thiose in battery plant workers be- cause of the huge amount of lead in the air of the plant. There was no difference between smokers and:1 nonismokers among the smelters. Brown attributed' this to the overwhielming lead' expo- sure that reduced the smoking factor to insignificance. He concl'udled that anyone who works with a respirable par- ticle such as lead, cadmium or asbestos should not smoke. In addition, production is affected because workers with excessive blood levels of particulate matter must stopiwork until the lev- els are lowered, according to federal rules. This study, completed in 19,81, coveredi489 men in three southern California plants. 11. "Predictors of sudden coroniary death inithe Framingham Study"' -- ARTHUR SCHATZKIN, Boston. Sudden unexpected death (SUD) accounits for nearly half all annual coronary deaths in males and for a third in females, Schatzkin said'. Significant predictors for SUDiin males are systolic blood pressure, ventricular rate, age, cholesterol (,for younger men), cigarette smoking, and left venitricular hypiertrophy (LVH). Risk profiles for women are very different. Blood pressure, weight and smoking are not significant risks for women. What is significant are hematocrit (~a very strong predictor)i, vital cap!- a:city and cholesterol (only marginally). Going from low risk'to high risk profiles, risks for men increase tremendously. The rise for women isn't nearly as dira- matic. LVH, which is highly significant in mien, may be picking up some unimeasured impact of high blood pressure, making,it a bet- ter ind'icator of hypertension thanicasual blood pressure. The presence of LVH may also indicate a damaged myocardium. Schatzkin said'he belileved SUiDs are probably preventable. cC' 12. "Bladd'~ eradd'er cancer mortality in New Jersey counties and %J its correlation with chemical industries" -- G. REZA NIAJEM, :~ Parsippany, N. J. C1'i Preliminary findings on recent trends in bladder cancer mortallity in: New Jersey counties were presented~. The years 1968'- 77 were compared to the 1950-69 period in the state, to the Surveillance, Epidemiology and End Results (SEER')! survey, and to

10. U.S. cancer mortality from 1973-76. Age-adjusted bladder cancer death rates for 20'of the 21 counties in New Jersey were greater than both U.S. and SEER rates for metropolitan areas and' states. In recent years ('1970-77) the gap between the rates for U.S. and New Jersey whites has narrowed. This isn't true for nonwhites. E'stimiates vary, but up to 30%p of the bladder cancers can be ascribed to occupational and' chemical exposure. In one Newark hospital, 60% of the bladder cancer patients lived in the northeast corner of the state which has dense concentrations of chemical, dye and leather manufacturing facilities. Salem County, where the highest percentage of the popula- tion works inithe chemical industry, rarnkedifirst in bladder cancer mortality among the 21 counties. It had double the rate of the rest of the stat e. Female bladder cancer was not associated with employment in the chemical industry, Najem said. This is consistent with the hypothesis that smoking is the major risk factor for women,. The role of smoking in men is complicated by the additional ma jor risk of individual occupational exposure. 13. "Respiratory morbidity in bridge and tunnel toll col- lectors" -- ROBERT G. EVANS, St. Louis. Ventilatory function tests were given during a seven-year period to the New York Triborough Bridge and Tunnel Authority workers at the request of their union. N~o controls were used. It was found that pulmonary function declined with~the number of years worked. All age-corrected pulmonary values were significantly higher in those wh~o never smoked compared~to smokers. Toll collectors who smoked and worked for more than 20' years had the lowest measures of lung function and the.highiest prevalence of respiratory symptoms. Carboxyhemoglobin levels were significantly higher for tun- niel workers tha:n bridge workers, but analysis of pulmonary func- tion data didn't indlicaite any difference between the two sets of workers. 14. JACK FRANKEL,, Tampa, Fla., reported on his CTR'-spon- soired~"Feasibility study of a diagnostic test to identify indi- viduals at risk for cancer." He gave details of his work onion- cogenic virus transformation,of cultured skin fibroplasts from lung cancer patients and f'rom 'relatives in families with neuro- fibromatosis ("elephant man's disease"). The technique could provide "a genetic basis for advising individuals at risk for lung and other cancers in terms of avoidance of chemicals and physical agents in the environment epidemiologically linked'to carcinogenesis, and recommendations for intensive monitoring to