Correspondence

Genetic Influence on Smoking

N Engl J Med 1993; 328:353-354February 4, 1993

Article

To the Editor:

The paper by Carmelli et al. (Sept. 17 issue)1 describing genetic influences on smoking among male subjects in the National Academy of Sciences-National Research Council Twin Registry shows the value of studies of twins in elucidating genetic and environmental influences on important health-related traits and behavior. The large sample provided an opportunity to detect an apparently small but statistically significant excess concordance in monozygotic twins as compared with dizygotic twins, a finding consistent with the existence of genetic influences on this important risk factor. The results of this study must, however, be interpreted with caution. Monozygotic twins are potentially more similar than dizygotic twins with respect to environmental as well as genetic factors. We documented this in a study of female twins and risk factors for coronary heart disease in which monozygotic twins were more alike than dizygotic twins with respect to education, exercise, alcohol and coffee consumption, and use of specific medications2. Fabsitz et al.3 also observed greater similarity among monozygotic twins with respect to various dietary measures in a subgroup of the same male twins studied by Carmelli et al.1. Although these factors may have some small genetic component, they are largely environmental in nature. As a result of differential environmental covariance, the apparent contribution of genetic influences to a trait can be spuriously overestimated in comparisons of monozygotic and dizygotic twins. It is certainly possible that such a bias was present in the report on smoking in male twins.

Second, any analysis of heritability depends on both the specific population examined and prevailing environmental conditions4. The sample of male twins used in this analysis is distinctive in a number of ways. Obviously, they were all men, and the results may or may not be generalizable to women. The subjects were also all veterans of the U.S. armed services, an environment in which smoking was extremely common. It is not clear that the same results would be found today in people in whom exposure to tobacco is much less common.

Thus, although the results reported by Carmelli and colleagues1 are consistent with a genetic influence on smoking behavior, this influence may be overestimated. The paramount importance of environmental exposure should not be dismissed.

Melissa A. Austin, Ph.D.
University of Washington, Seattle, WA 98195

Beth Newman, Ph.D.
University of North Carolina, Chapel Hill, NC 27599

4 References

1. 1

   Carmelli D, Swan GE, Robinette D, Fabsitz R. Genetic influence on smoking -- a study of male twins. N Engl J Med 1992;327:829-833
   Full Text | Web of Science | Medline

2. 2

   Austin MA, King M-C, Bawol RD, Hulley SB, Friedman GD. Risk factors for coronary heart disease in adult female twins: genetic heritability and shared environmental influences. Am J Epidemiol 1987;125:308-318
   Web of Science | Medline

3. 3

   Fabsitz RR, Garrison RJ, Feinleib M, Hjortland M. A twin analysis of dietary intake: evidence for a need to control for possible environmental differences in MZ and DZ twins. Behav Genet 1987;8:15-25
   CrossRef | Web of Science

4. 4

   Quantitative characters, polygenic inheritance, and environmental interactions. In: Cavalli-Sforza LL, Bodmer WF. The genetics of human populations. San Francisco: W.H. Freeman, 1971:597-602.
   

To the Editor:

Carmelli et al. make a fundamental error common to many studies of human behavior based on analyses of concordance in twins. This is the confusion of familial with genetic traits. Studies of twins may demonstrate a familial trait, but they do not necessarily indicate a genetically transmissible phenotype and are particularly unlikely to do so in the case of multifactorial traits, such as the tendency toward cigarette smoking or toward drinking.

Although we may recognize a subgroup of genetically defined traits that are included within the set of familial traits, there is also included a much larger subgroup of traits that result from entirely coincidental juxtapositions of genes. These traits may be held in common by blood relatives because they share these genes, but the traits themselves are not genetically determined in any meaningful sense. Consider a hypothetical study of concordance for happiness in twins. One might reasonably predict, on the basis of certain truly genetic characteristics known to have bearing on happiness, such as physical appearance and disease susceptibility, that concordance of indexes would be higher in the monozygotic cohort. This does not mean that happiness is genetically determined or that there is a “happiness gene”; rather, it identifies a vast cohort of genes that, by coincidence rather than linkage, tend to produce a complex phenotype in a specific environment.

It seems reasonable that smoking behavior, like happiness, might have a familial predisposition: things such as bronchial asthma, anxiety, and athletic ability have an impact on smoking, and these are characteristics with truly genetic components. Identical twins share these traits and thus have a higher concordance for smoking behavior and other covarying phenotypes. But this does not mean that smoking behavior itself is genetically transmissible, nor does it mean that a single cluster of genes is particularly important in determining whether or how people smoke. It is possible, of course, that certain genes may be directly relevant to smoking behavior. But such a question would be far better addressed by examining behavioral correlates with likely biochemical indexes, such as nicotine metabolism or nicotinic-receptor density, than by conducting studies of twins in which the truly genetic signal, if there is one at all, is buried beneath a mountain of coincidental “familial noise.”

Daniel Levy, M.D., Ph.D.
Public Health Service Hospital, Tuba City, AZ 86045

Author/Editor Response

The authors reply:

To the Editor: For those who consider genetic factors to be important primarily in the context of the inheritance of disease, the notion that there may also be a substantial genetic influence on normal variation in human behavior (e.g., smoking practices) may seem counterintuitive. A number of studies of twins,1 adoptees and their biologic and nonbiologic relatives,2 and twins reared apart3 have, however, demonstrated a substantial genetic contribution to individual differences in normal patterns of behavior. Our study provided evidence of the contribution of genetic influences to different aspects of smoking behavior in one of the largest cohorts of male twins studied to date. We agree with Dr. Levy that “biochemical indexes, such as nicotine metabolism or nicotinic-receptor density” would be traits closer to the gene products responsible for the addiction to nicotine. A prerequisite, however, for the development of nicotine addiction is the initial exposure to tobacco and its continued use. People at high genetic risk for addiction to nicotine will not become addicted if they remain lifetime abstainers. Thus, it is conceivable that genetically determined differences in lifetime smoking practices, such as the initiation and maintenance of the smoking habit over extended periods, constitute important routes by which genetic susceptibility to nicotine addiction is expressed and can be studied. Smoking is an important epidemiologic risk factor for a wide range of adverse medical conditions. Understanding the inheritance of this risk factor may thus be important for efforts at prevention and early intervention.

Drs. Austin and Newman are correct in pointing out that heritability estimates derived from studies of twins may be spuriously inflated as a result of environmental influences shared by twin brothers. Because of this possibility, we did not estimate heritability directly in our study. Instead, we limited our analyses to testing for genetic variance in different stages of the smoking process.

Dorit Carmelli, Ph.D.
Gary E. Swan, Ph.D.
Health Sciences Program, SRI International, Menlo Park, CA 94025

3 References

1. 1

   Eaves LJ, Eysenck HJ, Martin NG. Genes, culture, and personality: an empirical approach. London: Academic Press, 1989.
   

2. 2

   Loehlin JC, Horn JM, Willerman L. Heredity, environment, and personality change: evidence from the Texas Adoption Project. J Pers 1990;58:221-243
   CrossRef | Web of Science | Medline

3. 3

   Bouchard TJ Jr, Lykken DT, McGue M, Segal NL, Tellegen A. Sources of human psychological differences: the Minnesota Study of Twins Reared Apart. Science 1990;250:223-228
   CrossRef | Web of Science | Medline

Citing Articles (2)

Citing Articles

1. 1

   Yechiel Friedlander, Melissa A. Austin, Beth Newman, Karen Edwards, Elizabeth J. Mayer-Davis, Mary-Claire King. (1997) Heritability of Longitudinal Changes in Coronary-Heart-Disease Risk Factors in Women Twins. The American Journal of Human Genetics 60:6, 1502-1512
   CrossRef

2. 2

   Karen L. Edwards, Melissa A. Austin, Gail Pairitz Jarvik. (1995) Evidence for genetic influences on smoking in adult women twins. Clinical Genetics 47:5, 236-244
   CrossRef