University of Nebraska - Lincoln University of Nebraska - Lincoln
DigitalCommons@University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln
Faculty Publications, Department of Psychology Psychology, Department of
August 2004
Impulsivity and smoking relapse Impulsivity and smoking relapse
Neal Doran
University of Illinois at Chicago
Bonnie Spring
University of Illinois at Chicago
Dennis E. McChargue
University of Nebraska-Lincoln
, dmcchargue2@unl.edu
Michele Pergadia
Washington University, St. Louis, MO
Malia Richmond
University of Illinois at Chicago
Follow this and additional works at: https://digitalcommons.unl.edu/psychfacpub
Part of the Psychiatry and Psychology Commons
Doran, Neal; Spring, Bonnie; McChargue, Dennis E.; Pergadia, Michele; and Richmond, Malia, "Impulsivity
and smoking relapse" (2004).
Faculty Publications, Department of Psychology
. 288.
https://digitalcommons.unl.edu/psychfacpub/288
This Article is brought to you for free and open access by the Psychology, Department of at
DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Faculty Publications,
Department of Psychology by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln.
Impulsivity and smoking relapse
Neal Doran, Bonnie Spring, Dennis McChargue, Michele Pergadia,
Malia Richmond
[Received 12 November 2003; accepted 5 March 2004]
Previous research has shown that elevated trait-impulsivity heightens the risk for initiating tobacco use and indicates
that nicotine may be disproportionately rewarding for more impulsive persons. However, the influence of impulsivity
on the ability to maintain nicotine abstinence has not been studied. The present study tested the hypothesis that a
higher level of trait-impulsivity would predict a more rapid relapse to smoking following 48 hr of nicotine abstinence.
Participants were euthymic, regular smokers (N~45), with a history of at least one major depressive episode, who
participated in a paid smoking cessation study with biological challenge (tryptophan depletion). Treatment involved a
1-day skills training workshop followed by 48 hr of bioverified abstinence and weekly follow-up for 1 month.
Regression analyses indicated that elevated impulsivity predicted shorter time to relapse following the workshop after
controlling for treatment condition, baseline nicotine dependence, and age (b~2.39, R
2
change~.147, p~.011).
Greater impulsivity predicted more rapid relapse to smoking, which mediational analyses indicated could not be
explained by positive affect, negative affect, or craving. Findings suggest a need to identify alternative mechanisms to
explain impulsive smokers’ increased difficulty in maintaining abstinence and to develop targeted treatments that
address the special needs of smokers high in impulsivity.
Introduction
Approximately 23% of U.S. adults are regular
cigarette smokers (Center for Disease Control and
Prevention [CDC], 2002). Although 70% of regular
smokers report the desire to quit, few unassisted quit
attempts are successful (CDC, 2002). Individual
differences in some personality traits appear to
increase susceptibility to relapse (e.g., Gilbert,
Crauthers, Mooney, McClernon, & Jensen, 1999;
Ward, Klesges, & Halpern, 1997). Traits that predis-
pose toward negative affect have been the best
explored, guided by the hypothesis that some smokers
use nicotine to alleviate negative moods (e.g., Gilbert
et al., 1998). Traits that predispose toward appetitive
or rewarding behaviors have not been as well studied
for their contribution to tobacco dependence. The aim
of the present study was to examine the impact on
abstinence maintenance of one such trait, impulsivity,
defined as a chronic difficulty delaying gratification or
inhibiting response to rewarding stimuli (Monterosso
& Ainslie, 1999).
Trait-impulsivity refers to a long-standing pattern
of behaviors that reflect preference for readily
available rewards (Bickel, Odum, & Madden, 1999;
Mitchell, 1999; Monterosso & Ainslie, 1999) coupled
with marked difficulty delaying or resisting response
to such rewards (Arnett & Newman, 2000; Corr,
Pickering, & Gray, 1995; Monterosso & Ainslie, 1999;
Patterson & Newman, 1993). Some research suggests
that highly trait-impulsive individuals are more likely
to smoke cigarettes than are their less impulsive peers
(Bickel et al., 1999; Kassel, Shiffman, Gnys, Paty, &
Zettler-Segal, 1994; Mitchell, 1999). More impulsive
people may experience the rewarding aspects of drug
use more intensely than others do, as evidenced by the
finding of a positive association between impulsivity
and euphoric mood following cocaine administration
(Cascella et al., 1994).
Related traits associated with a propensity toward
approach behavior have been shown to be predictors
of smoking initiation and escalation, but linkages to
relapse have not been examined. For example,
sensation seeking has been shown to predict smoking
ISSN 1462-2203 print/ISSN 1469-994X online # 2004 Society for Research on Nicotine and Tobacco
DOI: 10.1080/14622200410001727939
Neal Doran, M.A. and Malia Richmond, M.S., University of
Illinois at Chicago; Bonnie Spring, Ph.D. and Dennis McChargue,
Ph.D., University of Illinois at Chicago and Hines VA Hospital;
Michele Pergadia, Ph.D., Washington University, St. Louis, MO.
Correspondence: Neal Doran, M.A., Department of Psychology (M/
C 285), University of Illinois at Chicago, 1007 W. Harrison Street,
Chicago, IL 60607, USA. Tel.: z1 (312)-355 0651; Fax: z1 (312)-355
2155; E-mail: ndoran1@uic.edu
Nicotine & Tobacco Research Volume 6, Number 4 (August 2004) 641–647
initiation and to be inversely related to the probability
of spontaneous cessation (Lipkus, Barefoot, Williams,
& Siegler, 1994). Sensation seeking also may moderate
the association between social variables (e.g., peer
pressure) and substance use (Slater, 2003). Smoking
also has been associated with higher levels of novelty
seeking (Heath, Madden, Slutske, & Martin, 1995;
Tercyak & Audrain-McGovern, 2003).
Thus, the few extant studies are consistent with the
premise that highly trait-impulsive people are initially
drawn to use cigarettes for their rewarding properties.
Little is known, however, about how impulsivity
influences smoking cessation. Heightened relapse rates
among impulsive smokers can be hypothesized on the
grounds that impulsive smokers will have difficulty
finding rewards that are as portable and as convenient
to administer as cigarettes. The acute decrement in the
level of reward access that immediately follows
cessation should, therefore, place them at heightened
risk of relapse. Accordingly, the primary aim of the
present study was to test the hypothesis that smokers
with higher trait-impulsivity would relapse more
quickly than their less impulsive peers following a
quit attempt.
We also tested hypotheses about three psychological
mechanisms that might mediate disproportionately
rapid relapse for impulsive smokers. Candidate
contributors to relapse were decreased positive
affect, increased craving, and increased negative
affect. Because decreased reward consumption may
lead to a drop in pleasant mood states (e.g., positive
affect), trait-impulsive smokers may relapse dispro-
portionately quickly at least partially as a result of
decreased postquit positive affect.
Decreased reward consumption resulting from
nicotine deprivation also may result in increased
cigarette craving. Kruedelbach, McCormick, Schulz,
and Grueneich (1993) found that among inpatient
substance abusers, greater impulsivity was associated
with higher drug cravings and with increased risk of
relapse. Pharmacological challenges that interfere with
the rewarding aspects of smoking have been found to
elicit higher levels of craving among more impulsive
smokers relative to other smokers (Reuter & Netter,
2001; Reuter et al., 2002). Consequently, we also
posited that more impulsive smokers’ quicker relapse
would be at least partially mediated by disproportio-
nately high levels of postquit craving relative to their
less impulsive peers.
Nicotine abstinence is believed to be a stressor that
activates depressive vulnerability (e.g., depression
proneness), exacerbating withdrawal-related negative
affect during the first 48 hr of abstinence and
promoting smoking relapse (e.g., Lerman et al.,
2002; Niaura, Shadel, Britt, & Abrams, 2002). It is
unclear whether trait-impulsive smokers also are prone
to experience heightened negative affect when stressed.
Because trait-impulsivity has been traditionally viewed
as a predisposition toward appetitive behaviors
(Monterosso & Ainslie, 1999), one might assume
that trait-impulsive smokers would be motivated
primarily to pursue reward and positive affect rather
than to dispel withdrawal-related negative affect.
However, evidence suggests that negative affect may
influence the relationship between trait-impulsivity
and smoking relapse. For example, among inpatients
in a substance abuse clinic, negative affect was
associated with especially strong drug craving and
relapse for those high in impulsivity (Kruedelbach
et al., 1993). Similarly, Hussong and Chassin (1994)
found that more impulsive adolescents were especially
likely to use drugs when they were experiencing
elevated negative affect.
Relatively little research has explored the possibility
that positive moods may be associated with smoking,
perhaps because positive and negative affect have been
conceptualized as opposite poles of a single con-
tinuum. Positive affect is conceptualized as a pleasant,
energized mood state reflecting feelings of elation,
excitement, enthusiasm, and peppiness (Watson &
Tellegen, 1985). Although positive and negative affect
are correlated, they appear to be distinct (Watson,
Clark, & Tellegen, 1988) and to have different neural
bases (Davidson, 1992) and psychological correlates
(Watson et al., 1988). Consequently, secondary aims
of the present study were to examine whether more
rapid relapse among more impulsive smokers would
be mediated partially by heightened postquit negative
affect as well as by increased craving and decreased
positive affect.
Method
Participants
The present research was part of a larger study
designed to test the efficacy of a 1-day workshop
intervention to promote tobacco abstinence. The
intervention taught smoking cessation and mood
management skills that were practiced during psycho-
logical and biological induction of transient negative
mood via guided imagery and tryptophan depletion.
Participants (N~45) were recruited from the commu-
nity through radio advertisements and through flyers
posted in the community and were paid to remain
abstinent for 48 hr following the workshop. Absti-
nence was verified by expired carbon monoxide and
salivary cotinine. The average participant was aged
41.02 years (SD~12.4, range~18–63); smoked
approximately one pack per day (M~19.57,
SD~7.7), had smoked for 23.16 years (SD~12.3);
and reported moderate nicotine dependence, as
measured by the Fagerstro
¨
m Test for Nicotine
Dependence (M~4.95, SD~2.3). Owing to the
constraints of the parent study, all participants had
a history of at least one episode of major depression.
642 IMPULSIVITY AND SMOKING RELAPSE
To be eligible, participants were required to be at least
18 years of age and to have smoked 10 or more
cigarettes per day for the past year. Individuals
receiving other smoking cessation treatment (e.g.,
other formalized treatment programs, nicotine repla-
cement therapy, bupropion) were excluded, as were
individuals currently experiencing an episode of major
depression (as diagnosed by Structured Clinical
Interview for DSM-III-R interview; Spitzer, Williams,
Gibbon, & First, 1992). Candidates being treated with
psychotropic medication other than antidepressants
also were excluded from the study.
Procedure
The parent research, of which the present study was a
component, was a treatment analogue study designed
to explore whether the addition of mood management
skills training and rehearsal to smoking cessation skills
training would enhance abstinence. Participants were
randomized to one of three treatments (tryptophan
depletion plus mood management and cessation skills
training, placebo depletion with mood management
and cessation skills training, and cessation skills
training alone), each of which involved participation
in a 1-day group workshop facilitated by trained
therapists (Spring, Pergadia, Richmond, McChargue
& Doran, 2002). Workshops focused on training
participants in the use of smoking cessation and mood
management skills and included an experiential
component in which skills were practiced.
Following the workshop, participants were paid to
quit smoking for 48 hr. Participants visited the
laboratory twice during the 48-hr period, to complete
questionnaires and for verification of self-reported
smoking status. Participants then returned to the
laboratory weekly for 4 weeks to complete question-
naires and for assessment of smoking status.
Measures
Relapse to smoking.
Relapse status was assessed at
24 and 48 hr postquit and at four weekly follow-up
sessions. At each visit, participants reported the
number of cigarettes smoked since the previous visit.
Participants also provided a breath sample that was
analyzed immediately for exhalation carbon mono-
xide with an Ecolyzer (Model EC-50, Vitalograph
Corporation, Lenexa, Kansas). Finally, participants
who reported abstinence at the second and fourth
weekly follow-up visits provided saliva samples that
were analyzed for cotinine, the major metabolite of
nicotine. Saliva samples were analyzed via radio-
immunoassay. Participants were judged to have
relapsed if they reported any amount of smoking on
7 consecutive days or any smoking in each of 2 suc-
cessive weeks. Participants also were considered to
have relapsed if they reported abstinence but had
Ecolyzer values greater than 10 parts per million
carbon monoxide or cotinine values greater than
20 ng/ml. Time to relapse was operationalized as the
number of consecutive days between the day the
participant attended the workshop and the day on
which the criteria for relapse were met.
Nicotine dependence. The Fagerstro
¨
m Test for Nico-
tine Dependence (FTND; Heatherton, Kozlowski,
Frecker, & Fagerstro
¨
m, 1991) is a self-report scale
measuring level of nicotine dependence. Cumulative
FTND scores range from 0 to 10, with values of 7
or higher suggestive of clinically significant nicotine
dependence (Gilbert et al., 1998; Killen, Fortmann,
Newman, & Varady, 1991). Previous research has
shown an association between FTND scores and
other measures of nicotine dependence (e.g., mea-
sures of nicotine and cotinine), providing support
for the construct validity of the FTND (Heatherton
et al., 1991).
Trait-impulsivity. The Barratt Impulsiveness Scale,
version 11 (BIS-11; Patton, Stanford, & Barratt,
1995) provides a measure of trait-impulsivity. The
BIS-11 is a self-report questionnaire that asks parti-
cipants to rate how often a series of statements
applies to them. Cumulative scores range from 30
(low in trait-impulsivity) to 120 (high in trait-
impulsivity). The BIS-11 has been shown to be reli-
able in both clinical and community samples, with
Cronbach’s alpha coefficients ranging from .79 to
.83 (Patton et al., 1995). The BIS-11 is structured to
assess long-term patterns of behavior and has been
used to assess trait levels of impulsivity across a
variety of populations, including substance-dependent
individuals (e.g., Mitchell, 1999; Moeller et al., 2002;
Stanford, Greve, Boudreaux, Mathias, & Brumbelow,
1996). Previous research suggests that the BIS-11
may be a better predictor of the ability to delay
smoking than other measures of impulsivity and
related constructs (e.g., the impulsivity scale of the
Eysenck Personality Inventory, the experience-seeking
and thrill- and adventure-seeking subscales of the
Sensation-Seeking Scale, or the novelty-seeking scale
of the Tridimensional Personality Questionnaire;
Mitchell, 1999). Research has shown that individuals
who have heightened difficulty delaying gratification
on laboratory measures of impulsivity tend to have
elevated BIS-11 scores (Allen, Moeller, Rhoades, &
Cherek, 1998; Cherek, Moeller, Dougherty, &
Rhoades, 1997).
Cigarette craving. Cigarette cravings were assessed
via the Questionnaire of Smoking Urges (QSU; Tiffany
& Drobes, 1991). The QSU is a 32-item instrument
NICOTINE & TOBACCO RESEARCH 643
containing two subscales, one representing anticipa-
tion of deriving pleasure from smoking and the
other anticipation of relief from negative affect and
other symptoms of nicotine withdrawal (Davies,
Willner, & Morgan, 2000; Tiffany & Drobes, 1991).
The QSU has been shown to be sensitive to both
abstinence and exposure to smoking cues (Morgan,
Davies, & Willner, 1999), suggesting relatively high
construct validity. The QSU was administered prior
to the workshop to establish a baseline craving
score and was readministered once on each of the 2
days immediately following the workshop.
Positive and negative affect. Positive and negative
affect were assessed via the Profile of Mood States
(POMS; McNair, Lorr, & Droppleman, 1971), a
self-report questionnaire consisting of 65 mood-
related adjectives. The ‘‘right now’’ version was used
to determine participant moods at the time the test
was administered. The POMS contains six subscales:
Tension/anxiety, depression/dejection, anger/hostility,
vigor/activity, fatigue/inertia, and confusion/
bewilderment. All subscales have been shown to be
internally consistent (Cronbach’s alpha coefficients
approaching .90) and have evidence of construct
and predictive validity (McNair et al., 1971). The
POMS was administered three times (at breakfast,
lunch, and dinner) over the course of 1 day prior to
the workshop to obtain a baseline mood score and
was administered an additional six times (at break-
fast, lunch, and dinner) during the first 48 hr postquit.
The vigor/activity subscale was used as the primary
index of positive affect. Postcessation scores (three
scores daily for 2 days) were averaged to increase
reliability and create a composite postquit positive
affect score. The tension/anxiety, depression/dejec-
tion, and anger/hostility subscales were combined to
form an aggregate measure of negative affect. Post-
cessation scores were again averaged to increase
reliability and to create a composite postquit nega-
tive affect score.
Data analyses
Residual change scores were used to assess changes in
positive and negative affect and cigarette craving,
because they provide a measure of the magnitude of
change while controlling for baseline values. Pre-
liminary analyses examined potential covariates,
including nicotine dependence, treatment condition,
age, ethnicity, gender, number of past depressive
episodes, symptoms of depression at baseline, and
POMS vigor and dysphoria at baseline. No evidence
of multicollinearity was found. Variables associated
with time to relapse or trait-impulsivity were entered
as covariates in the analysis described below. All
hypotheses were tested via hierarchical regression.
Mediating hypotheses were tested using the proce-
dures recommended by Baron and Kenny (1986).
Results
The primary analysis tested the hypothesis that more
impulsive participants would relapse more quickly
than other participants. Time to relapse was entered
as the criterion variable. Nicotine dependence, treat-
ment condition, and age were entered on the first step,
as covariates, and trait-impulsivity was entered on the
second step. As expected, trait-impulsivity predicted
time to relapse: More impulsive participants relapsed
more quickly than other participants. Trait-impulsiv-
ity accounted for approximately 14.7% of the variance
in time to relapse after controlling for age, treatment
condition, and nicotine dependence (b~2.39, R
2
change~.147, p~.011; Table 1).
Contrary to expectations, the relationship between
trait-impulsivity and time to relapse was not mediated
by postquit change in craving or by postquit changes
in either positive affect or negative affect. Postquit
changes in positive and negative affect were not
associated with either trait-impulsivity or time to
relapse. Similarly, postquit changes in cigarette
craving were independent of both trait-impulsivity
and time to relapse. Each mediational analysis was
run again after substituting difference scores (i.e.,
baseline–postquit) for residual change scores; these
analyses also indicated that the relationship between
trait-impulsivity and time to relapse was not mediated
by postquit changes in craving, positive affect, or
negative affect. Further analyses indicated that the
association between trait-impulsivity and time to
relapse was not mediated by residual change scores
or difference scores in the individual negative affect
subscales of the POMS (i.e., depression/dejection,
tension/anxiety, and anger/hostility).
Finally, to enhance the clinical utility of the
findings, we examined the sensitivity and specificity
of a variety of BIS-11 cutoff scores in predicting
participants’ relapse status 30 days after the workshop
(Table 2). In the context of the present study,
Table 1. Hierarchical regression analysis examining the
influence of impulsivity on time to relapse (criterion variable:
days to relapse following the workshop).
Step Predictor variable Beta
R
2
change
F
change
p
value
1 Covariates .113 2.35 .110
Treatment condition .06 .711
Age .12 .466
FTND score 2.34 .039
2 Trait-impulsivity 2.39 .147 7.14 .011
FTND, Fagerstro
¨
m Test for Nicotine Dependence.
644 IMPULSIVITY AND SMOKING RELAPSE
