Potential for Genetics to Promote Public Health: Title and subTitle BreakGenetics Research on Smoking Suggests Caution About Expectations

The number of DNA-based tests available for use in clinical care has rapidly increased over the past decade, and this trend has implications for public health and preventive medicine. In the 1990s, Coughlin¹ described an “emerging paradigm of disease prevention—the identification and modification of environmental risk factors among persons susceptible to disease due to genotype.” This optimistic vision of genetically based prevention was applied to lung cancer when Collins introduced “John,” who joins a support group of persons at genetically high risk of complications of smoking and thereby kicks the habit.² Yet a closer look raises questions about public health and research priorities.³

Clearly, smoking-induced lung cancer, first described by Rottman⁴ in 1898, continues to be a major public health concern. Lung cancer is the leading cause of cancer-related death in the United States.⁵ Smoking cessation reduces the chance of a subsequent lung cancer diagnosis and reduces risk of chronic obstructive lung disease, coronary artery disease, asthma, and other smoking-related conditions.⁶ A decades-long effort shows limited but tangible success in cessation interventions targeting individual smokers.⁷ Even brief physician counseling results in increased rates of smoking cessation, and there is a dose-response relationship between counseling intensity and effectiveness. Pharmacologic approaches to cessation may double the rate of cessation.⁸

However, from a public health perspective, societal-level approaches, such as bans, taxes, and government-sponsored antismoking campaigns, provide more substantial utility. These approaches offer the potential for a dual benefit: decreasing smoking prevalence and, in the case of smoking bans, protecting vulnerable populations from secondary exposure to smoke. In spite of early resistance, bans have been feasible⁹ and beneficial to employee¹⁰ and community¹¹ health. Higher cigarette prices decrease youth smoking,¹² and the combined effect of taxes, a ban, and an antismoking campaign in New York City contributed to the recent 11% decline in smoking prevalence there.¹³

How could genetics augment these public health efforts? Researchers have investigated the genetics of both smoking behavior and lung cancer. Genotypes that modulate smoking status, initiation, cessation, quantity, and treatment response have been identified (Table).^{14 - 18} For example, impaired nicotine metabolism is associated with less tobacco initiation and improved cessation rates, presumably because relatively sustained levels of nicotine decrease the craving for cigarettes due to falling nicotine levels.

In addition, variants in a number of genes, including several of the glutathione S-transferase group, are associated with altered risk of lung cancer. The variant most consistently associated with increased risk is the null allele of glutathione S-transferase M1 (GSTM1). While individual studies vary significantly, a meta-analysis incorporating data from 43 relevant studies found a modest but significant increase in risk (odds ratio, 1.17; 95% confidence interval, 1.07-1.27) for those with the null polymorphism.¹⁹

These data suggest 2 possible mechanisms by which genetic testing might improve rates of smoking cessation. Genetic testing could be used to inform individuals of higher risk of lung cancer and thereby increase motivation to quit, as proposed in Collins' example.² Or genetic testing could conceivably identify candidates for more intensive cessation programs, either on the basis of increased cancer risk or because their genotype helped identify a more effective cessation technique, such as a specific drug therapy.

Determining whether these approaches provide a public health benefit requires attention to 2 questions. First, how does genetic knowledge improve existing individual or societal interventions? Studies have evaluated whether genetic testing enhances cessation rates among susceptible smokers.^{20 - 21} Results indicate that knowledge of a small personal increase in risk is insufficient to facilitate smokers' quitting, consistent with evolving evidence that genetic risk information may be ineffectual in motivating behavior change²² or potentially may even be harmful by inducing fatalism, feelings of impotency, or loss of willpower.²³ In addition, current evidence that nicotine patch therapy might be effectively tailored by knowledge of genotype is inconclusive.¹⁸ So far, regarding smoking behavior and lung cancer, the postulated benefit of genetic testing has not materialized.

Second, presuming a future demonstration of incremental increases in cessation rates due to genetic testing, will these increases justify the risks? The potential for harm from this testing paradigm must be scrutinized from a variety of perspectives. The risks to smokers seem considerable. Knowledge of an adverse polymorphism could cause substantial anxiety, with long-standing repercussions. In addition, lack of cessation success might be blamed on genetic factors, decreasing residual motivation to quit. Alternatively, if a genetic test failed to identify an increased risk of smoking-related disease, the result might be considered a justification to start smoking or serve as a tobacco industry defense against liability. One can also envision misleading marketing of genetic tests to smokers (“Is cigarette smoking safe for you? This test can help you decide.”) or of unproven therapies toward those with adverse polymorphism (“This pill may be particularly effective for those with gene variant X.”). Would cigarettes ultimately be marketed to those with reduced genetic risk?

If genetic testing were used clinically to predict likelihood of quitting or risk of smoking-related disease, other harms might also occur. Health care systems might focus efforts on those at highest risk, neglecting others. Physicians might be less motivated to encourage cessation in some smokers if a genetic profile predicted poor response to multiple treatment modes. In the absence of prohibiting legislation, it is even possible that health insurers would charge higher rates for those with a predisposition to lung cancer, irrespective of smoking status. In the case of smoking, the appropriate public health stance seems clear: smoking cessation should be encouraged uniformly. From this perspective, the potential harms of genetic testing could be significant.

Are there other rationales for the study of gene variants associated with smoking or smoking-related diseases? One rationale is that such research will lead to a better understanding of substance abuse and potentially lead to innovative therapeutics with applicability beyond smoking. Another might be to promote discovery of novel biological pathways contributing to lung cancer or other smoking-associated illness. Broadly, study of the genetics of smoking might lead to fundamental insights into gene-gene and gene-environment interactions. Any of these applications of research into the genetics of lung cancer and associated behaviors are defensible. However, the ultimate potential of each approach as a research investment requires rigorous assessment. In essence, does the smoking and/or lung cancer genetic model provide a unique or particularly productive opportunity for understanding the problem? If so, how will the research be generalizable beyond the public health problem of smoking?

Genetics studies should be acknowledged as currently unlikely to lead to improved lung cancer prevention compared with the proven, non–genetic-based strategies for smoking cessation outlined above. Even as an interim measure, genetic testing should be secondary in funding and emphasis to universal measures to decrease smoking. And given the obvious dangers of tobacco and the associated imperative to eliminate it, research undertaken purely to unravel mechanisms of tobacco-related cancer is difficult to justify, unless the research can be shown to provide insight into fundamental cancer pathophysiology. The burden of proof lies with researchers, who should be willing to justify why research on the genetics of smoking, with its inherent limitations, is in the public interest.

Genetics research with a rationally applied focus should be supported.²⁴ Therefore, it is reasonable to ask that genetics research goals related to smoking be stated explicitly. The goals should either add convincing practical value to existing antitobacco efforts or demonstrate that the smoking or lung cancer model affords a unique venue for advancing knowledge. The goals also should reflect a rigorous consideration of the potential for adverse downstream effects. For example, increasingly popular genome-wide association studies have thus far demonstrated only modest success (ie, logarithmic odds [LOD] score of 2.7 for smoking dependence²⁵ ) in identifying influential candidate genes. What is the likelihood that further studies will ultimately lead to a gene variant (or gene variants) with both overwhelming influence and ability to affect significant changes in behavior commensurate with community-based interventions? This is an inefficient strategy at best, with low likelihood of yield. Furthermore, it distracts from efforts to reduce smoking in the population at large.

Given the unavoidable need to prioritize research funding, research on the genetics of smoking and related disease requires careful scrutiny. We join others in cautioning against “genocentric views,” which may lead to inappropriate expectations rather than substantive progress toward improving health outcomes.²⁶ The example of genetics research related to smoking is a telling case in point.