Tobacco Epidemiology and the Challenge of Multiple EtiologiesTobacco Epidemiology and the Challenge of Multiple Etiologies

In the late 19th century, as tobacco consumption began a dramatic increase, medical debate about its health consequences intensified. Since the publication of King James' Counterblaste to Tobacco in 1604, physicians and other observers had debated whether tobacco smoking helped or harmed a person's health.¹ This debate took on a different form in the late 19th century. Following the development of bacteriology, physicians concentrated on identifying the specific etiologies of pathological processes. Armed with Koch's postulates and sophisticated laboratory techniques, they deciphered the causes of many infectious diseases. But when physicians applied these new methods to the old problem of tobacco use, they immediately ran into trouble.

The debates about tobacco amblyopia demonstrate what happened. Although first described in the early 1800s and defined in 1830 by William Mackenzie, MD,² tobacco amblyopia remained a contested diagnostic category in the 1890s. Most physicians agreed on the basic symptoms: diminishing color perception and visual acuity in a central scotoma, possibly progressing to blindness if left untreated.³

However, they disagreed about the etiologic role of tobacco. Amblyopia, like neurasthenia and chlorosis, was an amorphous diagnostic category, defined by symptoms, not cause. This ambiguity of causality created insoluble controversy. While tobacco use was implicated in most cases, physicians remained unsure whether it was a sufficient or even a necessary condition.

The author of the accompanying 1898 JAMA article, William B. Meany, MD, of Louisville, Kentucky, displayed protests that were typical in his day. Meany argued that physicians' "rudimentary" insight did not allow them to trace the disease to tobacco: "Is it not a fact that in the majority of instances we are unable with assurance to assign a cause?" He listed many other factors that could cause amblyopia, including tuberculosis, alcohol, embolism, epilepsy, starvation, dysmenorrhea, and railroad accidents. Other physicians cited similar problems. After all, many tobacco smokers remained healthy, many nonsmokers suffered these symptoms, and some patients' symptoms disappeared even as they continued to smoke.⁴ This etiologic ambiguity fueled Meany's claim that tobacco intake had been arbitrarily singled out from many other equally responsible causes. He believed that this focus on tobacco, the so-called filthy habit, was the sort of moral sentiment that physicians should purge from medical science.

Meany's arguments capture our attention for many reasons. First, the diagnosis of tobacco amblyopia, "an entity out of the class of the ordinary clinical syndrome," has continued to puzzle physicians.² According to an 1897 report, tobacco amblyopia was diagnosed in 1% of hospital patients during its heyday.³ Now it is diagnosed rarely. Many theories of causation have been explored, including the theory of chronic cyanide intoxication in the presence of vitamin B₁₂ deficiency.² However, physicians still do not fully understand the nature of the declining incidence, nor can they fully explain the pathophysiological role of tobacco.⁵ Despite this uncertainty, physicians did implicate tobacco amblyopia as a contributing cause of the 1991 epidemic of optic neuropathy in Cuba.⁶

Second, Meany's arguments demonstrate the challenges faced by physicians who have sought to expose the health consequences of tobacco. Meany argued that amblyopia, a multifactorial disease, should not be blamed on tobacco, a single cause. Although this specific debate faded from prominence as medical and public attention increasingly focused on the carcinogenic properties of tobacco smoke, the underlying reasons for debate continued. Tobacco advocates simply repeated Meany's critique of cause: lung cancer, like other tobacco-related diseases, is multifactorial and should not be blamed on tobacco. As with tobacco amblyopia, laboratory science could not unravel the complex multiple etiologies of chronic diseases. Although independent voices of concern existed within the medical profession,⁷ they lacked authoritative evidence, and the marketing programs of the tobacco industry ran unopposed.

Resolution of the uncertainty about the risks of tobacco came from two intertwined developments. First, as medical attention shifted from tobacco amblyopia and smoker's cough to lung cancer and coronary artery disease, researchers became better able to test their hypotheses. Lung cancer was a discrete diagnosis, with increasingly standardized and reproducible diagnostic criteria. This diagnostic specificity enabled more accurate measurement, which formed the basis of strong causal hypotheses that could be tested through epidemiologic research.

Second, epidemiologists developed new study designs. Instead of using laboratory methods to identify single causes of infectious diseases, researchers used probabilistic methods to identify individual contributions to the multiple causes of chronic diseases. The development of large, prospective cohort studies in the 1950s gave researchers the data they needed to calculate the relative risks of specific behaviors or exposures. Starting with the British Medical Research Council's 1951 study of lung cancer in physicians, epidemiologists began to produce unequivocal evidence of the harm of tobacco.^{8 - 9}

Although diagnostic categories and causal assumptions remain critical conundrums for medical research, the debates over tobacco have now been categorically resolved. Even tobacco company executives have been forced to accept that smoking is harmful. Smoking is now recognized as the leading cause of preventable morbidity and mortality in the United States. This progress in epidemiologic investigation has helped turn the medical profession and the public powerfully against smoking. Both groups accept the harms of secondhand smoke, the addictive nature of nicotine, the vulnerability and exploitation of children, and the knowledge and intent of the tobacco industry.¹⁰ Now, rather than debating whether tobacco causes harm, we debate how best to address the impressive harms we have come to understand so thoroughly.