160 Tobacco Dependence
(FTQ; Fagerström, 1978) and its more recent variant,
the Fagerström Test for Nicotine Dependence (FTND;
Heatherton, Kozlowski, Frecker, & Fagerström, 1991), or the
DSMcriteria, an approach deriving from the need to include
tobacco dependence in psychiatric nomenclature and classi“-
cation and which attempts to adhere to classic de“nitions of
drug dependence (American Psychiatric Association, 1994;
Robins, Helzer, Cottler, & Goldring, 1988). Although both
paper-and-pencil and psychiatric diagnostic approaches have
provided reliable de“nitions for use in many dif ferent types
of studies, neither of the current existing approaches relied on
test development approaches well grounded in psychometric
theory. Moreover, even though dependence may consist of
several components (Lombardo et al., 1988), factor analyses
of the FTQ consistently result in one large factor marked pri-
marily by the number of cigarettes smoked per day (Kendler
et al., 1999; Lichtenstein & Mermelstein, 1986). More re-
cently, we (Hudmon, Marks, Pomerleau, Brigham, & Swan
2000; Shiffman, Hickcoz, Gnys, Paty, & Kassel, 1995) pre-
sented multidimensional scales for assessing tobacco depen-
dence; however, these measures have not yet been published.
While several twin studies claim to have identi“ed a heritable
component to dependence, they have relied on indirect mea-
sures such as continued smoking after becoming a regular
smoker (Heath & Martin, 1993; Madden et al., 1999; True
et al., 1997) or amount smoked (Koopmans et al., 1997;
Swan, Carmelli, & Cardon, 1996, 1997). The few genetic
analyzes involving the FTQ (Kendler et al., 1999; Straub
et al., 1999) of which we are aware have identi“ed a heritable
component for •dependenceŽ (Kendler et al., 1999) as well as
tentative linkage with regions on six different chromosomes
(Straub et al., 1999), while leaving the question of which
components of tobacco dependence have the most or least
genetic in”uence unresolved.
The Potential Importance of
Gene-Behavior-Environment
Interactions in Tobacco Dependence
A second feature of Figure 7.1 presented at the beginning of
this chapter that needs further investigation concerns the
plethora of environmental conditions that are recognized to
play a role in the acquisition, maintenance, cessation, and
relapse of smoking (USDHHS, 1994). Simply put, previous
genetic investigations of smoking have not adequately incor-
porated environmental measures into their study designs.
Often overlooked is the fact that behavior genetic studies,
while consistently “nding that about 60% of variation in
smoking is attributable to genetic sources, have also found
that the remaining 40% is due to environmental influences
(Sullivan & Kendler, 1999). Despite the importance of envi-
ronmental in”uences in explaining variation in smoking, in-
variably, genetic investigations can only speculate as to the
speci“c environmental sources of variation because they
were not assessed. Evidence has been obtained for a role for
gene-environment interactions in alcoholism and conduct
disorder, supporting the notion that the coupling of a genetic
susceptibility with a high-risk environment increases risk for
pathology to a greater extent than would be expected in the
presence of either factor alone or in additive combination
(Legrand, McGue, & Iacono, 1999). Religious upbringing,
for example, was recently identi“ed as a signi“cant mediat-
ing factor reducing the risk for initiation of alcohol use
(Koopmans et al., 1997).
A “ nal feature of Figure 7.1 concerns preexisting condi-
tions or comorbidities. As described earlier, it is known that
certain individual characteristics increase the likelihood of
becoming, remaining, or becoming again a smoker. Factors
such as depression, attention de“cit disorder, and schizo-
phrenia may well map into and contribute to a number of life-
time tobacco use milestones, trajectories, and/or dimensions
of tobacco dependence (Hughes, 1999). While each of these
comorbidities may have heritable components (Heath &
Madden, 1995), we are only at the beginning of our under-
standing of the extent to which they share a genetic correla-
tion with tobacco dependence phenotypes. Although Kendler
and colleagues (1993) identi“ed a common genetic substrate
to depression and smoking in female twins, we need to know
more about the speci“cs of this relationship before this in-
formation can inform prevention and treatment efforts. For
example, of the possible tobacco dependence dimensions,
which ones are most strongly associated with depression?
What are the genetic correlations between depression and
speci“c dimensions of dependence?PUBLIC HEALTH IMPERATIVE FOR
COLLABORATIVE, TRANSDISCIPLINARY
RESEARCHMore integration of the psychosocial and biogenetic lines of
research is needed because tobacco dependence is a complex
genetic trait. In this context, we use the term complexityas
de“ned by the “eld of genetic epidemiology (Lander &
Schork, 1994; Ottman, 1990, 1995, 1996). A complex trait
has the following features: (a) it has reduced penetrance (i.e.,
not everyone with a susceptibility gene(s) will become
dependent on tobacco); (b) genetic heterogeneity is involved
(i.e., a different set of susceptibility genes may contribute
to the likelihood of becoming a smoker in different people);