The Cigarette Century

by Allan Brandt

  Rather than seeing the multiple ways of acquiring new knowledge as being at odds, Hill viewed them as complementary. Each approach would have particular advantages depending on the particular hypothesis and the possibilities for its evaluation.

  There are of course no grounds for antagonism between experiment and observation. The former, indeed, depends on observation but of a type that has the good fortune to be controlled at the experimenter’s will. In the world of public health and preventive medicine each will—or should—constantly react beneficially upon the other. Observation in the field suggests experiment; the experiment leads back to more, and better defined, observation. . . . However tangled the skein of causation one must, at least at first, try to unravel it in vivo.74

  This is what Hill’s mentor, Major Greenwood, called “the permeation of statistical research with experimental spirit.”75

  Critics of the epidemiological method could not see this clearly. But there had never been, as some would later claim, a single gold standard of disease causality. That the biomedical paradigm of single cause and single disease was a chimera was well understood by even its most vigorous advocates. And medical knowledge was always provisional and contingent. Just as drugs deemed “effective” do not work in every case, so too a cause of disease does not always result in disease. As Richard Doll would later explain, the epidemiologists had identified a cause of lung cancer (and other diseases), not the cause.76

  Some historians have suggested that modern epidemiological techniques were radically innovative and untried at the time.77 But this underestimates their deeper historical traditions, as well as the experimental basis of much of modern epidemiological technique. Certainly, the epidemiologists understood the necessity of controlling variables and limiting opportunities for bias—but it had also been shown that there were ample opportunities for these problems to intrude in laboratory investigations.78

  A great deal was at stake in the debates about tobacco and health in the 1950s. Epidemiologists and statisticians often pointed to the limitations of laboratory investigation for sorting out complex multiple causes of disease, even as they imported experimental techniques into their work. Physicians working in clinical settings noted that laboratory and epidemiological findings alike often failed to conform to their observations of patients and their symptoms, even as those pursuing laboratory and epidemiological investigations disparaged clinical observation as “anecdotal.”

  The question that troubled the entire medical community in the 1950s was what constituted adequate knowledge to act in the various realms of medicine, public health, business, and politics. No one doubted that it would be valuable to understand the most basic mechanisms of carcinogenesis. But neither would anyone in public health or medicine presume that all knowledge short of defining those mechanisms was inadequate or suspect. Medical and public health interventions had often been pursued with great benefit before causal mechanisms were known. John Snow’s heroic studies of cholera never identified the underlying organism, yet brought fundamental changes in urban infrastructure. The history of medicine is filled with examples of partial knowledge being used to reduce disease. Unlike some other domains of knowledge, the pragmatic demands of human health place a clock on investigation. The cost of acting must be balanced against the cost of not acting.

  The development of systematic knowledge about the harms of cigarette use illuminates the complex character of medical science in the mid-twentieth century. Typically, the debate about cigarettes and lung cancer is portrayed as a battle between laboratory and statistical science. But this is largely a particular historical construction offered by some of the protagonists in the debate, most notably the tobacco industry. It is an example of how powerful economic and industrial interests would deploy their resources to influence, delay, and disrupt normative scientific processes.

  By the mid-1950s, clinicians and researchers were largely convinced of the connection between cigarettes and cancer. In early 1955, the chair of the pathology department at the University of Michigan, Carl Weller, offered a comprehensive assessment of the evidence. As someone whose work depended on the visualization of disease-related changes in cell and tissue, Weller, like many pathologists, had previously been highly skeptical that smoking was a cause of lung cancer. “I early subscribed to the then prevailing opinion that protoplasm was relatively stable and the chromosomes particularly so,” he explained in his book, Causal Factors in Cancers of the Lung. Most pathologists had tended to accept the notion that rates of cancer, including lung cancer, varied little over time. “The rate of the incidence of cancer in any organ was considered to be a fairly constant function,” he wrote, “not readily influenced by environment, although numerous occupational carcinomas of the skin had been demonstrated.” Weller therefore concluded that the reports of rising rates of lung cancer were probably spurious. “I joined others in attempting to explain it by the aging of the population, by the advent of radiography, by clinical awareness, and by better diagnostic methods in general.” Ultimately, however, “these explanations proved inadequate and it was necessary to admit that some recently acquired feature of our way of life was very rapidly changing the incidence of pulmonary cancer.”79 The increase of the disease was real.

  Weller undertook a systematic review of all the studies to date as well as the criticisms. In Causal Factors in Cancers of the Lung he described his shift of perspective. “I have searched the literature for other reasonable explanations or for recognizable fallacies. I have found none of importance.” He concluded: “As of today, I must agree with many of the specialists in statistical analysis and in the endemiology of cancer, that this association has been established.”80 Weller understood the significance of this conclusion: it was now incumbent upon the medical profession to address the problem of smoking. “What is the next step?” he asked. “When will it be not only proper but requisite that the medical profession take cognizance?” Citing the long delay between English surgeon Percivall Pott’s recognition of the high rates of scrotal cancers among chimney sweeps and the eventual identification of carcinogens in soot, Weller urged immediate action. “May we show the same practical sense as our forefathers,” he wrote, “and not look for direct proofs which are out of reach before we transmit experience into practical measures.”81

  As physicians and scientists critically assessed what by now amounted to dozens of reports, they typically arrived at similar positions. The medical and scientific director of the American Cancer Society, Charles Cameron, followed the same path as Weller. The ACS presented itself as a voluntary health agency dedicated to funding research and public education regarding cancer prevention and treatment, and since the 1930s it had emphasized education to encourage early diagnosis and treatment. This agenda, based on highly touted medical technology, such as X-rays and surgery, had received strong support from both the medical profession and powerful donors. For ACS executives like Cameron, the findings about tobacco came as a mixed blessing. On one hand, it was valuable to know that smoking might lead to cancer. On the other hand, the news demanded a significant reorientation of the agency’s strategy and put it in the center of a potentially vicious contest between business and public health. Most importantly, the ACS did not want to take any position that might be deemed as usurping physicians’ professional prerogatives.

  In early 1952, for example, Cameron invited Evarts Graham to write an article on lung cancer for a book to be published by the ACS for the lay public. Graham accepted and, not surprisingly, emphasized the causal relationship between smoking and disease. Cameron wrote to request revisions in the manuscript:Could I ask you to redo this piece with emphasis on the need for frequent X-ray examinations of the chest. . . . I really think this would be of greater value than the emphasis on smoking which is the theme of the present article. I have no objection, of course, to your mentioning it but feel that the space given it should be decreased in favor of references to early diagnosis.82

  Yet a
few years later, Cameron had become convinced that smoking constituted a major cause of cancer, and he brought the ACS along with him. In 1956, in an article for the Atlantic Monthly entitled “Lung Cancer and Smoking: What We Really Know,” he wrote:Although the complicity of the cigarette in the present prevalence of cancer of the lung has not been proved to the satisfaction of everyone, yet the weight of the evidence against it is so serious as to demand of stewards of the public welfare that they make the evidence known to all. . . . There is in some quarters an unbecoming skepticism of statistics in general and of these remarkably consistent results in particular. By some—a diminishing band, as I see it—the findings are rejected because there is not “laboratory proof.”83

  But this standard, Cameron argued, was both unrealistic and unprecedented. “What is the nature of the proof which is demanded to establish the cancer-causing effect of cigarette smoking? If it is that smoke or another tobacco product must be shown to cause cancer of the lung under conditions of experimental control using living human subjects, then I hope the experiment will never be undertaken. No standards of proof in the entire world of research demand as much as that.”

  While still not prepared to “hold that smoking causes cancer of the lung,” Cameron nonetheless joined the growing medical consensus, concluding: “If the degree of association which has been established between cancer of the lung and smoking were shown to exist between cancer of the lung and say, eating spinach, no one would raise a hand against the proscription of spinach in the national diet.”84 The time had now arrived, he wrote, to act on this knowledge.

  By the mid-1950s, other astute observers of clinical medicine had come to agree. Assessing the evidence in September 1953, Joseph Garland, editor of the New England Journal of Medicine, noted that the most recent Doll and Hill publication (their prospective study) “yielded evidence of an association between cigarette smoking and lung cancer so strong as to be considered proof within the everyday meaning of the word.” Garland continued, “If similar data had incriminated a food contaminant that was not habit forming and was not supported by the advertising of a financial empire, there is little doubt that effective counter measures would have followed quickly.” He concluded, “The situation affords unusual opportunities for the vast tobacco interests to support impartial researches into the effects that their products may have on human health.” Leading figures in medical science now argued that the evidence was clear, convincing, and scientifically persuasive, and that physicians and public health officials had a responsibility to warn their patients and the public. They reasoned that medical knowledge incorporates social responsibility and that the findings about lung cancer and smoking had reached a level of significance and certainty that triggered these professional commitments.

  Many physicians, as they came to know and accept these findings, began to quit smoking. According to a study done in Massachusetts, nearly 52 percent of physicians reported being regular smokers in 1954, with over 30 percent smoking at least a pack a day. Just five years later, only 39 percent were regular smokers, and only 18 percent went through a pack or more per day.85 Evarts Graham attributed much of the remaining skepticism to the fact that many in the medical profession were smokers themselves. “Unfortunately,” he wrote in 1954,

  it has not been universally accepted and there are still many cigarette addicts among the medical profession who demand absolute proof. . . . The obstinacy of many of them in refusing to accept the existing evidence compels me to conclude that it is their own addiction to this drug habit which blinds them. They have eyes to see but they see not because of their unwillingness or inability to give up smoking. . . . I have never encountered any non-smoker who makes light of the evidence or is skeptical of the association between excessive smoking and lung cancer.86

  It is important to recognize just how popular smoking was at mid-century. The findings implicating smoking as a cause of disease and death were an indictment of an enormously popular behavior, difficult to moderate. This, Graham argued, constituted an important and powerful bias in the evaluation of the data.

  Survey research conducted by the ACS confirmed Graham’s perspective. Physicians who were heavy smokers were among the most skeptical of the research findings linking tobacco use to lung cancer. In 1955, Cameron, Horn, and David Kipnis surveyed members of the American Board of Thoracic Surgery, the American Board of Pathology, and the American Association for Cancer Research. Among those polled, 55 percent agreed with the statement that heavy smoking may lead to lung cancer; 32 percent expressed uncertainty; while only 5 percent disagreed. But among those surveyed who smoked a pack or more each day, only 31 percent agreed that “Heavy smoking may lead to lung cancer.” Among nonsmokers, the figure was more than 65 percent.87

  Faced with his own research findings, Graham had quit smoking, so he well understood the difficulty of withdrawing from nicotine. But his five decades of exposure to tobacco smoke would now confirm in the most personal and intimate way what his and Wynder’s research had so clearly demonstrated. In 1957, he wrote to his friend and colleague Alton Ochsner, “Perhaps you have heard that I have recently been a patient in the Barnes Hospital because of a bilateral bronchiogenic carcinoma which sneaked up on me like a thief in the night.”88 Ochsner, deeply shaken by the news, wrote back, “Thank you for your letter . . . which simply crushed me. It is a perfectly horrible thing to think that you have bronchiogenic carcinoma, a condition for which you have done so much.”89 Two weeks later, Graham died, a victim of the very disease that had been the center of his professional life. In the end, he became yet one more data point in the lethal history of smoking.

  As a result of several statistical surveys, the idea has

  arisen that there is a causal relationship between

  ZEPHYR and tobacco smoking, particularly cigarette

  smoking. Various hypotheses have been propounded, from

  time to time, as explanations of this conception. The two

  which seem most important at the present time are:

  (i) Tobacco smoke contains a substance or substances

  which may cause ZEPHYR

  (ii) Substances which can cause ZEPHYR are inhaled

  from the atmosphere, e.g. in the form of soot.1

  BRITISH AMERICAN TOBACCO, 1957

  I just don’t believe it. People are hearing the same old story, and the record is getting scratched.2

  BOWMAN GRAY, JR., 1960

  CEO, R.J. REYNOLDS

  Members of the Research Department have studied in detail cigarette smoke composition. Some of these findings have been published. However, much data remains unpublished because they are concerned with carcinogens and carcinogenic compounds. This raises an interesting question about the former compounds. If a tobacco company pled “Not guilty” or “Not proven” to the charge that cigarette smoke (or one of its constituents) is an etiological factor in the causation of lung cancer or some other disease, can the company justifiably assume the position that publication of data . . . should be withheld because such data might affect adversely the company’s economic status when the company has already implied in its plea that no such etiologic effect exists?3

  ALAN RODGMAN, 1962

  RESEARCH SCIENTIST, R. J. REYNOLDS

  CHAPTER 6

  Constructing Controversy

  THE IDENTIFICATION OF cigarette smoking as a cause of serious disease shook the tobacco industry to its core. For decades, tobacco companies had developed strategies for dealing with concerns about the health impact of smoking. From ads promising mildness to claims like “More Doctors Smoke Camels,” the companies had repeatedly sought to calm smokers’ medical anxieties. Such competitive claims were yet another vehicle to promote individual brands.

  By the early 1950s, however, it was abundantly clear that the evidence implicating cigarette smoking as a risk to health was now of a different order. First, the link between smoking and disease was categorical, outside the realm of individual clinical ju
dgment. Although physicians might advise individual smokers to “cut down,” no one could offer assurance that any level of smoking was safe. Second, the cigarette was tied to the most feared disease of mid-century: cancer.4 Earlier concerns about cough or scratchy throat gave way to the ominous medical data indicating that the “habit” could kill. No major industry had ever faced such a fundamental threat to its future.

  In this unprecedented crisis, the company executives came to recognize that traditional approaches to promotion and marketing had to change radically. The new scientific evidence would require a collective response if the industry was to survive. Unsubstantiated health claims proffered for individual brands would merely call attention to the problems with the cigarette, and they were sure to draw intense medical and scientific scrutiny if not regulatory intervention. In the early 1950s, despite decades of concerns about the health impacts of smoking, the industry possessed almost no internal capacity to assess the new scientific evidence. The research departments in each company were focused on product design and modification—small changes to enhance “mildness” or vary taste. Having brilliantly mastered the meaning and character of their product for more than half a century, the tobacco companies found that they had begun to lose control of the very cultural processes that they had so effectively utilized in creating the modern cigarette. Try though it might—often with some considerable success—the tobacco industry would never again unequivocally control the meaning of the cigarette. The scientific findings of the 1950s constituted a sea change in the history of smoking. Industry executives found themselves in uncharted waters, and the boat was leaking.