Tobacco Documents Online

ty rates between the two rg oups. In other words, the death rates were almost the same for both groups. Yet there was a difference 17 percentage points between the groups in the rates of smoking reduction.7 of As a British cardiologist noted in the British Medical Journal, this trial "ended with inconclusive results," leaving many unanswered questions.8 It should be clear, therefore, that the results of this study do not demonstrate that reductions in "risk factors," including smoking, lower the risk of death from heart disease. It may in fact be time for the medical community to reexamine the significance of the "risk factor" concept. Oslo The results of another intervention study, aimed at examining the effect of reduction of serum cholesterol and smoking on the in- cidence* of heart disease in Norway were reported in 1981 and 1982. The Oslo study group consisted of healthy middle-age men pre- sumed to be at "high risk" for heart disease. Men in the interven- tion group were counseled on diet and smoking to see if this could lower the incidence of first heart attacks, including those that are fatal. The authors found heart disease incidence was signifi- *Incidence is the number of new cases of disease that develop in a selected period of time. It is used to illustrate the rate at which diseases develop. 3 I

cantly reduced in the intervention group. The authors attributed this lowered risk of disease largely to reduced cholesterol levels. They said their calculations showed no statistically significant decline in risk as a result of reduced cigarette consumption.9 When these data were reanalyzed in an attempt to separate the possible effects of smoking reduction or cessation from those of improved diet on the incidence of new disease, the results were even more striking. Neither smoking status when the study began nor changes in amount smoked were found to be signficantly associa- ted with heart attacks. In fact, changes in amount smoked were "unassociated." In this update, even though the investigator speculated that smoking may play some role, smoking was found to be a "non-signifi- cant risk factor" for new cases of heart disease.l0 There is only one study in the literature of the results of intervention against smoking alone. In preliminary findings pub- lished in 1979, Geoffrey Rose and P. J. S. Hamilton reported no evidence of any reduction in overall death rates in a group whose smoking was reduced. The authors described this finding as disap- pointing.11 4

Four years later, they reported their 10-year results. Now the death rate for heart disease in the intervention group was not significantly lower than the rate for the normal care group. The heart disease mortality rate for the entire study population was higher than the national rate.12 It is difficult to understand', therefore, how the results of this study could be cited as support for the hypothesis that smoking causes heart disease. It should be clear that the findings from these "risk factor" intervention studies are inconsistent with the claim that smoking is causally related to heart disease. The results from Rose and Hamilton and from baseline studies like Friedman's, discussed be- low, suggest that the "risk factor" concept may have been misinter- preted. It is simple to do so, but it is highly suspect from a scientific point of view. There are those who, like the Surgeon General, label smoking as a cause of heart disease. They claim support in reports that former smokers experience significantly lower risks for heart dis- ease than continuing smokers. They conclude that smoking cessation is the reason for the lowered risk of disease, because they assume that former smokers are comparable to continuing smokers except for the change in their smoking habits. This is only an assumption. Such studies generally have N failed to compare the smokers and the former smokers at the same o ~_, 0 5 CO

baseline -- when all were still smokers. Comparing a variety of health, social and personal factors among continuing smokers, never smokers and former smokers before they stopped, Gary Friedman and co-authors found that the smokers who would stop later were different from continuing smokers in many characteristics believed to be related to the development of heart disease. In other words, the smokers who later stopped already may have been at a lower "risk" for heart disease than those smokers who would continue to smoke. Smokers who later quit showed statistically significant dif- ferences from continuing smokers in a number of physical symptoms, personal characteristics and miscellaneous traits. For example, when compared with that of persistent smokers, their prevalence of reported heart disease symptoms such as chest pain on exercise, shortness of breath and leg pain was more like that of the never smokers. In fact, those who would stop in the future were, overall, more like the never smokers than they were the continuing smokers. The authors concluded that smokers who later quit were not a typ- ical group of smokers.l3 But the study became controversial. In a 1980 American Heart N) cn a ~ ~ Journal editorial, Carl Seltzer, a co-author of the Friedman ~ w 0 m r_j 6

report, cited the Friedman study and the 1978 report by Rose and Hamilton as evidence that "the reversibility of the claimed risk of cigarettes to the smoker's life has not been demonstrated. "Accordingly," the editorial said, "it is reasonable to be- lieve there is no proof that stopping reduces the risk of heart disease."14 Friedman took another look at his study population and in a subsequent article claimed that further study showed that baseline differences between the groups could not explain the differences in heart disease mortality between persistent smokers and quitters.15 Seltzer and Philip Burch criticized the Friedman "update" for the methodologies employed and the conclusions reached. Seltzer noted Friedman's second analysis ignored the possible effects of other factors associated with risk of heart disease that are impor- tant for reaching conclusions that are accurate.l6 Burch concluded that there was no valid scientific evidence to prove that stopping smoking reduces the risk of death from heart disease. He wrote: "The conclusion that quitting smoking results in a substantial reduction in mortality is not established by the study of Friedman et al."17 IU t3} It should be clear that the results from the intervention and 0 ~ ~ ~ w 0 7 - w

the Friedman studies do not scientifically demonstrate that reducing smoking reduces the risk of death from heart disease. The MRFIT authors, in trying to explain the seven-year results of their study, wrote that "multifactor intervention received a less than optimal test" because of unexpected declines in "risk factor" levels and lower than expected mortality in the usual care group. They suggested that the recent reduction in heart disease mor- tality in the t1.S., "the reasons for which are still not totally understood," might explain the unexpected mortality decline in the usual care group.7 Time Trends In the U.S., heart disease death rates have been reported to be declining since 196R. This, however, has not been demonstrated to be attributable to changes in smoking habits. Attributing the decline to improved medical care or "risk fac- tor" change is open to criticism. "The truth of the matter is that we are uncertain as to the precise reason or reasons for the de- cline," a prominent government scientist conceded in 1982. "It has not been possible to obtain national data that would support an ex- planation," he admitted.1R - 8 -

Figure 2: "What underlies the recent decrease? While this question has been addressed by a number of authorities, no consensus has emerged. The question would be easier to answer if answers were available to a number of other questions. "Has the incidence, as well as the mortality rate, from CHD [coronary heart disease] been decreasing? "Have the presenting symptoms become, on the average, less severe? "Is there a lower case fatality rate following an initial myocardial infarction? Follow- ing subsequent infarctions? "Has the reinfarction rate decreased? "Has there been adecrease in the incidence of myocardial infarction following angina pectoris? "Is the ratio of sudden to nonsudden CHD death changing? "Data on these subjects are sparse and equivocal, but unless we know, or presume we know, answers to some of these questions, it becomes difficult to decide what to explain, let alone to apply tests of relevancy to proffered explanations." These are statements from a letter to the editor, American Heart Jour- nal, January 1982, by Tavia Gordon, consulting biometrician and former longtime supervisory statistician, Biometrics Research Branch, National Heart, Lung and Blood Institute, National Institutes of Health. Heart disease mortality did drop markedly between 1968 and 1978 in the TJ.S. Participants in a Conference on the Decline in Coronary Heart Disease Mortality, convened by the National Heart, Lung and Rlood Institute (NHLRI) in October 1978, agreed on one thing--that the decrease was real, not a result of artifacts or changes in death certificate coding.lA Their report said the reasons for the decline were not clear. And they remain speculative today. Last September, the Framingham study's Kannel noted that "the case for either improved treatment or risk-factor modification is difficult to sustain" in providing explanation for the continuing mortality decline.210 Two papers from the 1978 conference analyzed epidemiologic - 9 - N tJ1 0 ~ ~ .Q w 0 m cn

data to determine whether a decline in the percentage of smokers was related to the decrease in heart disease mortality rates. Roger Williams and co-authors suggested that recent trends in stopping smoking might be partially responsible for changes in U.S. males' mortality rates, but that change in smoking habits could not explain the substantial decline in the largely nonsmoking population of Utah. Nor could it explain why younger U.S. women showed the largest percentage decline in heart disease at a time when their smoking rates reportedly increased.21 These views were reinforced in Joel Kleinman's conference pa- per. He looked at the possible impact of smoking trends on heart disease mortality and reported that changes in smoking habits among women were not consistent with the decline.22 A year later, Klein- man and his co-authors wrote that changes other than smoking must account for the bulk of the decline in heart disease mortality.23 A principal problem in understanding the decline is the rela- tive lack of current U.S. data on heart disease incidence. The 1978 conference report called for additional research on heart disease rates to determine whether there had been changes in the frequency of nonfatal heart attacks. As Kannel commented, it is unfortunate that incidence data are "sparse" and "inconsistent."20 ~ .a 0 CO - 10 - oN

Figure 3: Smoking Prevalence and Rates Among Middle-Aged Females: United States, 1965 and 1976 Regular Cigarette Smokers Smokers of 15 or More Cigarettes/Day White Females 35-44 45-54 55-64 Black Females 35-44 45-54 55-64 Age 1965 LI 1976 ® Source: Reproduced from Kieinman, J.C., et ai., "Trends in Smoking and ischemic Heart Disease Mortali- ty," in Proceedings of the Conference on the Decline In Coronary Heart Disease Mortality, Havlik, R.J. and M. Feinleib (eds.), U.S. Pubiic Health Service, Department of Health, Education and Welfare, DHEW Publication No. (NIH) 79-1610, May 1979

If "risk factor" changes had played an important part in the decline in heart disease mortality, then they should also have had a role in reducing the incidence of the disease. By 1982, there was little evidence that the incidence had declined at the same time as the mortality rate. It has been suggested by some researchers that if there had been no actual decline in the incidence of heart disease, the re- ported decreases in the levels of "risk factors" might be far less important in reducing mortality than improved medical care. A 1982 editorial in the British Medical Journal added a signi- ficant note -- that the decline in TJ.S. mortality rates was appar- ent before there were substantial alterations in many factors, in- cluding smoking.24 This observation appeared in a discussion of findings from one of the very few incidence studies to appear in the scientific lit- erature, an examination by a group at Mayo Clinic of heart disease incidence in Rochester, Minnesota, over a 26-year period. The Mayo authors noted that the decline in the mortality rate there was similar to that observed nationally but that the decline in incidence preceded the decline in mortality by approximately 10 years. N Ln 0 ~ ~ ~ w 0 m w - 12 -