NIH Public Access
Author Manuscript
Addict Behav. Author manuscript; available in PMC 2014 September 01.
NIH-PA Author Manuscript
Published in final edited form as:
Addict Behav. 2013 September ; 38(9): 2415–2421. doi:10.1016/j.addbeh.2013.03.021.
Prospective Effects of Adolescent Indicators of Behavioral
Disinhibition on DSM-IV Alcohol, Tobacco, and Illicit Drug
Dependence in Young Adulthood
Rohan H. C. Palmera,b, Valerie S. Knopika,b, Soo Hyun Rheec, Christian J. Hopferd,c, Robin
C. Corleyc, Susan E. Youngd, Michael C. Stallingsc, and John K. Hewittc
aDivision of Behavioral Genetics, Rhode Island Hospital
bDepartment
cInstitute
of Psychiatry and Human Behavior at the Alpert Medical School of Brown University
for Behavioral Genetics, University of Colorado at Boulder, Boulder, CO
dDepartment
of Psychiatry, University of Colorado Denver School of Medicine, Denver, CO
NIH-PA Author Manuscript
Abstract
Objective—To identify robust predictors of drug dependence.
Methods—This longitudinal study included 2361 male and female twins from an ongoing
longitudinal study at the Center for Antisocial Drug Dependence (CADD) at the University of
Colorado Boulder and Denver campuses. Twins were recruited for the CADD project while they
were between the ages of 12 and 18. Participants in the current study were on average
approximately 15 years of age during the first wave of assessment and approximately 20 years of
age at the second wave of assessment. The average time between assessments was five years. A
structured interview was administered at each assessment to determine patterns of substance use
and Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; Fourth Edition) attention
deficit hyperactivity disorder (ADHD), conduct disorder (CD), and drug dependence symptoms.
Cloninger’s Tridimensional Personality Questionnaire was also used to assess novelty seeking
tendencies (NS). At the second wave of assessment, DSM-IV dependence symptoms were
reassessed using the same interview. Path analyses were used to examine direct and indirect
mechanisms linking psychopathology and drug outcomes.
NIH-PA Author Manuscript
Results—Adolescent substance use, CD, and NS predicted young adult substance dependence,
whereas the predictive effects of ADHD were few and inconsistent. Furthermore, CD and NS
effects were partially mediated by adolescent substance use.
© 2013 Elsevier Ltd. All rights reserved.
Correspondence: Rohan H. C. Palmer, Ph. D., Division of Behavioral Genetics at Rhode Island Hospital, Department of Psychiatry
and Human Behavior at Brown University, Bradley Hasbro Children’s Research Center, 1 Hoppin Street, Suite 204, Providence, RI
02903, Rohan_Palmer@Brown.edu, Tel: (401) 793-8395, Fax: (401) 793-8341.
Contributors
Authors John Hewitt, Susan Young, Michael Stalling, Soo Rhee, Robin Corley, and Christian Hopfer designed the CADD study and
wrote the protocol. Author Rohan palmer conducted literature searches and provided summaries of previous research studies, as well
as conducted the statistical analysis under the guidance of Dr. Valerie Knopik. Author Rohan Palmer wrote the first draft of the
manuscript and all authors contributed to and have approved the final manuscript.
Conflict of Interest
All of the listed authors declare that they have no conflicts of interests.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our
customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of
the resulting proof before it is published in its final citable form. Please note that during the production process errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Palmer et al.
Page 2
Conclusions—Adolescent conduct problems, novelty seeking, and drug use are important
indices of future drug problems. The strongest predictor was novelty seeking.
NIH-PA Author Manuscript
Keywords
Attention Deficit Hyperactivity Disorder; Conduct Disorder; Behavioral Disinhibition; Drug
Dependence; Alcohol Dependence; Tobacco Dependence
1. Introduction
Difficulty in inhibition of behavioral impulses results in an increased risk for the
development of substance use and substance use disorders (SUD; consisting of DSM-IV
abuse or dependence) (American Psychiatric Association, 2000). Individuals diagnosed with
conduct disorder (CD) or attention problems (i.e., deficits in attention or hyperactivityimpulsivity – as defined by DSM-IV ADHD) are more likely to use substances during
adolescence and develop a SUD during young adulthood (Charach, Yeung, Climans, &
Lillie, 2011). Likewise, high novelty seeking (NS; a tendency for high levels of exploratory
behavior, novel experiences, and immediate rewards), low harm avoidance, and low reward
dependence tendencies have been associated with the development of SUD (Wills, Vaccaro,
& McNamara, 1994). Further, early use of alcohol, tobacco, and illicit drugs increases the
likelihood of future drug problems (Grant & Dawson, 1997).
NIH-PA Author Manuscript
A growing number of studies (Iacono, Malone, & McGue, 2008) indicate that ADHD, CD,
and NS frequently co-occur and that the covariance among these traits can be best explained
by an underlying latent trait referred to as Behavioral Disinhibition (BD; i.e., “the inability
to inhibit behavior despite its social undesirability, and cascade of familial, educational,
psychological, and legal consequences”) (Young, Stallings, Corley, Krauter, & Hewitt,
2000). The high comorbidity among these disorders has prompted the need for more studies
that can clarify the relationship between early assessments of ADHD, CD, NS, and SU and
future drug problems. For instance, several reports have suggested that CD mediates the
association between the inattention and hyperactivity subscales of ADHD and young adult
drug problems (Brook, Duan, Zhang, Cohen, & Brook, 2008; Disney, Elkins, McGue, &
Iacono, 1999; Fergusson, Horwood, & Ridder, 2007). Of these studies, those that have
managed to adjust for the co-occurrence of CD and ADHD have concluded that ADHD
effects could be attributed to the fact that CD was not accounted for in the model. Still, it
remains unclear if any or all of these associations are due to their direct relationship with
substance dependence or with the early stages of substance initiation and use.
1.1 Associations between measures of BD and Substance Problems
NIH-PA Author Manuscript
Population-based studies suggest a more than chance association between adolescent
measures of BD and the development of future substance problems (Disney, et al., 1999;
Elkins, King, McGue, & Iacono, 2006; Elkins, McGue, & Iacono, 2007; Grekin, Sher, &
Wood, 2006; Jenkins, et al., 2011; Zucker, 2008); however, most studies are not prospective
in nature and fail to account for the shared liability among adolescent ADHD, CD, NS,
substance use, and substance problems (i.e., DSM-IV abuse or dependence) (Lee,
Humphreys, Flory, Liu, & Glass, 2011). At the time of this study, we reviewed the literature
and identified two longitudinal studies that examined the effect of both childhood/adolescent
ADHD and CD on young adult (i.e., 30 > age > 18) SUDs and a third study that also
included personality traits. In their study of 506 boys in the Pittsburgh Youth Study, Pardini
and colleagues (Pardini, White, & Stouthamer-Loeber, 2007) examined the effects of early
adolescent ADHD, CD, anxiety, and depression on young adult alcohol use disorders. The
authors concluded that while controlling for the comorbidity amongst all the adolescent
psychopathologies, CD was related to young adult alcohol use disorder symptoms but
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 3
NIH-PA Author Manuscript
ADHD was not. Furthermore, there was no interaction between ADHD and CD. Similar
findings were obtained by Fergusson and colleagues, who used a 25-year longitudinal study
of a New Zealand birth cohort to examine the link between CD and attention problems and
young adult substance use, abuse, and dependence (Fergusson, et al., 2007). In a separate
study, Tarter and colleagues used a sample of males assessed during adolescence and young
adulthood to identify pathways linking childhood hyperactivity to young adult substance use
disorder (Tarter, Kirisci, Feske, & Vanyukov, 2007). Tarter and colleagues discovered that
childhood hyperactivity is a “diathesis for externalizing disturbances” at young adulthood.
Furthermore, the link between childhood hyperactivity and young adult SUDs was mediated
by both neuroticism and conduct problems, thus suggesting that NS and CD carry a much
greater risk than ADHD.
1.2 Deriving robust estimates of effects in the context of adolescent drug use
NIH-PA Author Manuscript
NIH-PA Author Manuscript
Despite the growing number of studies linking childhood/adolescent psychopathology and
personality with adult drug outcomes, it is not known whether any of the patterns of
association described above are robust longitudinal effects, as most studies have failed to
account for the joint effects of adolescent substance use (a robust predictor of future drug
problems) (Grant & Dawson, 1997; Guy, Smith, & Bentler, 1994), which has been shown to
be elevated among teens with externalizing problems (August, et al., 2006). In order to
obtain robust estimates of effect between each of these early traits and young adult drug
outcome, it is necessary to analyze longitudinal studies that have assessed all of these traits
using a multivariate regression framework. A multivariate framework provides the
opportunity to explore two fundamental research questions linking early assessments of
ADHD, CD, and NS with future drug problems. Specifically, (1) Are adolescents with a
history of ADHD OR CD problems OR an exuberance for novelty at increased risk for
future drug problems, and (2) are those risk estimates, specifically related to the drug
problems themselves, OR are they driven by their association with other behaviors,
especially drug use, which on its own is capable of mimicking the characteristics of ADHD,
CD, and NS because it causes neurocognitive (e.g., decreased memory, attention and
speeded information processing, and executive functioning) and brain matter volume deficits
(e.g., hippocampal, prefrontal cortex, and white matter volume) in the developing brain
(Squeglia, Jacobus, & Tapert, 2009). For example, Squeglia and colleagues showed that
moderate to heavy alcohol use and high levels of hangover symptoms was associated with
reduced sustained attention in males and reduced visuospatial task performance (e.g.,
visuospatial memory) in females (Squeglia, Spadoni, Infante, Myers, & Tapert, 2009).
Overall, studies examining the effects of ADHD, CD, and NS on future drug problems need
to account for (1) shared variance between the traits, and (2) the effects of adolescent drug
use, as it can produce neural abnormalities that can perpetuate behavioral disadvantages that
increase the risk for drug use/problems.
1.3 Purpose of the current study
The purpose of this study was to address the ambiguity surrounding the predictive role of
BD indicators, especially since only a few prospective studies have considered their joint
effects and underlying comorbidity with early/adolescent substance use. We hypothesized
that early externalizing psychopathology (i.e., ADHD and CD symptomatology) and novelty
seeking tendencies are indicators of future drug dependence problems, over and above the
effects of early adolescent substance use. In addition to direct processes, we further
hypothesized that conduct and attention problems and novelty seeking influence the liability
to drug dependence by also influencing the level of drug use during adolescence (i.e.,
indirect mechanisms).
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 4
2. Materials and methods
2.1 Participants
NIH-PA Author Manuscript
NIH-PA Author Manuscript
The sample consisted of 2361 individual members of a twin pair (46% male) who were
drawn from the Center for Antisocial Drug Dependence (CADD) Study, an ongoing study at
the University of Colorado. The twins utilized in the CADD originate from two communitybased twin samples at the University of Colorado that are part of the much larger Colorado
Twin Registry based at the Institute for Behavioral Genetics at the University of Colorado at
Boulder. The two samples consist of the Longitudinal Twin Study (LTS) sample and the
Community Twin Sample (CTS). The CTS sample was open to all twins born in the state of
Colorado between 1979 and 1990, and additional in-migrating twins in the same age range
ascertained through Colorado school districts. The LTS included twins born in Colorado
between 1984 and 1990 who were initially tested prior to age 2 and who were followed
longitudinally (Rhea, Gross, Haberstick, & Corley, 2006, 2013); inclusion in this sample
depended on location and early twin and family characteristics. Twins from both samples
were recruited into the CADD while they were between the ages of 12 and 18 years of age.
Due to the longitudinal nature of the study, subjects from the LTS and CTS are currently
enrolled in five-year follow-ups of the original baseline assessment. Data for the current
study were drawn from the first and second waves of the CADD study. At the end of data
collection for Wave 2 during the year 2008, 100% of LTS participated at both waves and
93% of CTS twins participated at both waves.
Data for the current study are drawn from the first (Wave 1) and second (Wave 2) waves of
data collection. The sample is diverse and largely made up of Caucasians (87.12%) with
similar rates of males and females across different ages. At Wave 1, the average age of the
participants was 14.87 years (SD = 2.17). At Wave 2, the average age of assessment was
19.64 years (SD = 2.60). The interval between both waves of assessment was approximately
five years (mean = 5.22, SD = 1.06). Rates of substance use and abuse and dependence in
the CADD are similar to those typically observed in large population samples, such as the
Monitoring the Future Study and the National Survey on Drug Use and Health (R. H.
Palmer, et al., 2009).
2.2 Procedure
NIH-PA Author Manuscript
Data for the CADD project were obtained after identifying participants from the CTS and
LTS samples. Participants were assessed using diagnostic interviews as part of an entire day
of a battery of tests that included cognitive functioning, diagnostic interviews, and selfreports. The University of Colorado Boulder and Denver campuses Institutional Review
Boards approved all components of the CADD presented in this study. Additional details on
the recruitment and sample description of the twin samples in the CADD are available
elsewhere (Rhea, et al., 2006, 2013).
2.3 Psychiatric and Personality Assessments
DSM-IV symptoms of ADHD (i.e., nine symptoms of inattention and nine symptoms of
hyperactivity/impulsivity) and CD (15 symptoms) were measured during Wave 1 using the
Diagnostic Interview Schedule for Children Version IV (DISC-IV) (Shaffer, Fisher, Lucas,
Dulcan, & Schwab-Stone, 2000). The DISC-IV was used to assess DSM-IV symptoms and
diagnoses. Several ADHD that are part of the DSM-IV criteria were “Fidgets in seat”,
“Talks a lot”, “Easily Distracted”, and “Difficulties sustaining attention”. Several of the CD
criteria that were included were “Cruelty to animals”, “Bullying”, and “Breaking and
entering”. Computer algorithms were used to determine the lifetime symptom counts (i.e., a
sum of the criteria met by each respondent) of the hyperactivity-impulsivity (scores could
range from 0 to 9 symptoms) and inattention (participants could report no symptoms (i.e., 0)
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 5
or up to 9 symptoms) subscales of ADHD, and the total DSM-IV symptoms for CD
(participants scores could take on values from 0 to 15 symptoms).
NIH-PA Author Manuscript
Adolescent levels of NS were assessed using 18 items from Cloninger’s Tridimensional
Personality Questionnaire-Short Form (Heath, Cloninger, & Martin, 1994). Although the
TPQ assesses other measures of personality we decided to utilize only NS for this study
because of prior evidence. A sample of items asked by the questionnaire included, “I often
try new things just for fun or thrills, even if most people think it is a waste of time” and
(reversed) “I hate to make decisions based only on my first impressions”. The mean of the
NS items endorsed was used for the current analyses.
NIH-PA Author Manuscript
The Composite International Diagnostic Interview-Substance Abuse Module (CIDI-SAM)
was used to obtain information on substance experimentation, repeated use (defined as: use
of alcohol more than five times, smoked 20 or more cigarettes or used tobacco-based
products such as snuff, tobacco pipe, or cigars, and use of marijuana and the other illicit
drugs more than five times), frequency of use, and lifetime problems (i.e., DSM-IV
symptoms levels and diagnoses of drug abuse or dependence) during waves 1 and 2. For the
purposes of this study, adolescent substance use and problems were defined as the number
of substances repeatedly used, and the number of substances with problems, respectively.
These measures provided indication of the extent of substance use and problems during
adolescence.
Wave 2 (i.e., young adult) levels of lifetime DSM-IV alcohol, tobacco, and illicit (i.e.,
cannabis, sedative, hallucinogen, amphetamine, opioid, inhalant, phencyclidine, club drugs,
and cocaine) drug dependence symptoms endorsed were tallied to create a symptom count
variable (with scores ranging from zero to seven per substance). The measure Dependence
Vulnerability (DV; which is a polysubstance dependence vulnerability index (Button, et al.,
2006), was also constructed (DV = total number of DSM-IV dependence symptoms across
all 11 substances / total number of substances used).
Wave 1 (adolescent) predictors included: (1) the hyperactivity-impulsivity (HI) and (2)
Inattention (INATT) subscales of ADHD, (3) CD, (4) substance use (SU) and problems
(SP), and (5) NS. Wave 2 (young adult) dependent variables were: alcohol, tobacco, and
illicit drug dependence and dependence vulnerability (DV).
2.4 Analysis plan
NIH-PA Author Manuscript
Data management and descriptive analyses were conducted using SAS (Version 9.2)(SAS,
2002). To correct for departures from normality often seen in count phenotypes, the data
were rank normalized after adjusting for age effects within each gender (Young, et al.,
2000). Subsequent analyses were conducted in MPlus ® (Muthen & Muthen, 1998–2012)
that has the capabilities to handle multiple variables of varying distributions (ordinal,
continuous, count, etc.) and complex survey data (e.g., subpopulation analysis, clustering,
stratification, etc.).
To test the main research question (Figure 1), path analyses were conducted in MPlus using
the rank normalized [mean=0, standard deviation (STD) = 1] scores for the Wave 2 outcome
variables and the Wave 1 independent variables. Effect estimates for Figure 1 were derived
using path analysis to determine how this set of variables (i.e., ADHD hyperactivity/
impulsivity, ADHD attention problems, CD, NS, and SU) is related to future alcohol,
tobacco, and illicit drug dependence (i.e., the total effect). In addition, we tested the
hypothesis that a portion of the total effect of these variables (i.e., ADHD, CD, and NS)
results from an indirect effect through adolescent substance use. Further, because the
occurrence of substance use disorders (i.e., DSM-IV substance abuse or dependence) is
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 6
NIH-PA Author Manuscript
common during adolescence, the model also controlled for these effects to account for any
bias in our estimates that might result from early and persistent patterns of problematic
substance use. Finally the model also controlled for age at time of assessment during Wave
1, the gender [coded as male (1) vs. female (0)] of the participants, and the racial
identification of the respondent (coded as Caucasian vs. Other). Age and gender interaction
effects were also tested and are reported when detected. MPlus models were estimated while
accounting for the fact that participants within the sample are clustered within families (i.e.,
MPlus computed standard errors that accounted for the non-independent nature of the
observations).
3. Results
3.1 Mean-level Gender Differences
NIH-PA Author Manuscript
Table 1 presents the untransformed means and standard deviations for each trait. Robust
Poisson regression was used to determine gender differences in the mean at each
assessment. At Wave 1, males reported higher mean levels of ADHD, CD, NS, and alcohol
dependence symptoms. At Wave 2, females reported higher mean levels of tobacco
dependence symptoms, while males endorsed more alcohol dependence symptoms than
females. Males also scored higher on the DV measure. With the exception of NS, most of
the observed data were right skewed. Analyses presented in the Results section employed
normalized measures for each of these variables.
3.2 Prospective effects of BD indicators on young adult drug dependence
Table 2 shows the effects of ADHD, CD, SU, and NS on young adult alcohol, tobacco, and
illicit drug dependence and dependence vulnerability (DV). Details on other covariates in
the model are presented for completeness. The first half of Table 2 shows the association of
the Wave 1 predictors with adolescent substance use. High levels of ADHD attention
problems (ATT), novelty seeking (NS), and conduct disorder (CD) were associated with
greater levels of substance involvement at Wave 1. These significant associations suggested
that adolescent substance use may provide a mechanism for a part of or all of the association
of ATT, CD, and NS with drug dependence.
NIH-PA Author Manuscript
The pattern of results across the right portion of Table 2, outline the effect of each predictor
on drug dependence outcomes at Wave 2. CD, NS, and substance use at Wave 1 were the
most consistent predictors of substance problems at Wave 2. Taking into account the effects
of all other covariates in the model, increasing levels of ADHD attention problems was
associated with higher scores on the DV measure. The hyperactivity/impulsivity (HI)
subscale of ADHD was not predictive of alcohol, tobacco, or illicit drug dependence
symptoms at Wave 2. There was an association between symptom levels on the ADHD
attention problems subscale and Wave 2 illicit drug dependence and DV. Higher levels of
CD were associated with alcohol, and illicit drug dependence and DV. Further, the increase
in illicit drug dependence problems due to higher CD levels further increased with higher
levels of age. NS predicted increased symptom levels for all Wave 2 outcomes. Being male
or Caucasian did not have any effect on drug dependence at Wave 2.
Table 3 presents the partitioning of the total effect of each Wave 1 predictor (HI, ATT, CD,
and NS) into indirect effects via substance use (as indicated in Figure 1) and direct effects.
In regards to alcohol problems at Wave 2, the total effect for both subscales of ADHD
suggested no overall effect; however, there was an indirect effect of ATT on alcohol
dependence mediated by substance use. Notably, due to the lack of a total main effect for
ATT it is difficult to interpret its indirect relationship on alcohol dependence other than to
say that it may operate via adolescent substance use. Substance use also mediated a part of
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 7
CD’s and NS’s effect on alcohol problems, although there were also direct effects,
suggesting only partial mediation.
NIH-PA Author Manuscript
For Tobacco dependence, there was limited evidence to suggest that either ADHD subscale
influenced the level of tobacco problems. Similar to alcohol, the specific indirect effect for
ATT on tobacco via substance use was significant suggesting that such associations may be
operating via adolescent substance use. CD and NS had significant total effects on tobacco
dependence. Further, all of CD’s total effect was mediated by the level of substance use
during adolescence. The significant effect of NS on tobacco dependence was only partially
mediated by substance use, as there was evidence to suggest direct effects linking NS and
tobacco dependence.
In the model for Illicit drug dependence, ATT, CD, and NS were significant predictors.
However, unlike the results for alcohol and tobacco, the coefficients for specific indirect
effects suggested no mediation via substance use. Consequently, direct effects of each
predictor drove almost all of the total effects.
Lastly, ATT, CD, and NS significantly predicted DV, such that higher scores were
associated with higher DV scores. For this general liability index for drug dependence, ATT,
CD, and NS have specific indirect and direct effects on the liability for drug dependence,
suggesting partial mediation.
NIH-PA Author Manuscript
Overall, these results suggest that CD and NS are the most robust predictors of alcohol,
tobacco, and illicit DSM-IV dependence symptoms and DV. Further, their effect on
substance dependence (except illicit drug dependence) is partially mediated by adolescent
substance use.
4. Discussion
This study is a unique attempt to examine the prospective relationship between adolescent
psychopathology (i.e., ADHD and CD) and personality (i.e., NS) and young adult drug
dependence in the context of adolescent substance use. It also explores the mediating effects
of adolescent substance use, thus providing a useful mechanism for how early externalizing
psychopathology is related to later drug dependence. Regression analyses in our communitybased sample of twins showed that after accounting for the overlap among the adolescent
measures, the indicators of BD were differentially predictive of young adult substance
dependence. Moreover, in addition to effects mediated by substance use there are unique and
independent effects of CD and NS on drug dependence.
NIH-PA Author Manuscript
4.1 Consistency with previous research
Our results were similar to those of Pardini and colleagues’ study, which showed that
conduct disorder symptoms during late adolescence were predictive of alcohol dependence
(Pardini, et al., 2007). The results are also consistent with the aforementioned longitudinal
studies that also suggest that CD and NS mediate the associations between childhood/
adolescent measures of inattention, hyperactivity-impulsivity and ADHD and later substance
problems. This paper adds to the literature by demonstrating the effects of CD and NS on
DSM-IV drug dependence while also accounting for the mediating effects of early drug use.
Our results were similar to the results presented by Elkins and colleagues (Elkins, et al.,
2007), whose adjusted models showed no effect of either subscale of ADHD on alcohol
abuse/dependence, but a significant effect of CD. On the other hand, Elkins and colleagues
were able to detect strong positive associations between HI and tobacco dependence,
whereas our study did not. It is important to note, however, that Elkins and colleagues did
not account for the effects of early drug use and problems or their mediating effects.
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 8
4.2 Implications of findings for alcohol and other drug studies
NIH-PA Author Manuscript
NIH-PA Author Manuscript
To put the results of this study into perspective, behavioral problems in childhood (in
particular conduct disorder) increase the risk of developing problems with multiple
substances and future drug dependence. The results of this study suggest that treatment of
conduct problems during childhood or early adolescence or the management of novelty
seeking tendencies may help to reduce the risk of substance use during adolescence and
problematic substance use. The fact that these effects are only partially mediated by drug use
suggests that drug use is not the sole cause of later dependence symptoms. This supports the
findings of conditioning in the brain due to repeated substance use (Volkow, Wang, Fowler,
& Tomasi, 2012), as well as the hypothesis that individuals who present with early
externalizing psychopathology are more susceptible to the effects of substances of abuse.
Consequently, these findings have serious implications for the development of treatment and
prevention efforts, as well as future molecular genetics research. Behavioral or
pharmacological treatment of disruptive disorders in children and adolescents is likely to
have lasting effects across multiple disruptive psychopathologies due to the common thread
that underlies ADHD, CD, and NS - the inability to plan out actions, inhibit actions, and
consider the implications of actions (impulsivity) (Miller, Stephen, & Tudway, 2004). For
instance, preliminary findings from our lab recently determined that higher levels of CD and
ADHD symptoms are associated with higher levels of initial sensitivity (e.g., subjective and
autonomic experiences, such as reports of pleasure, liking the taste, nausea, heart rate, etc.)
to alcohol and tobacco during adolescence, which suggests that these individuals may be
primed to be more responsive to substances of abuse (Bidwell, et al., 2012; R.H. Palmer, et
al., 2012; Wills, et al., 1994).
NIH-PA Author Manuscript
Though clinically defined traits, our strongest predictors, CD and NS, represent an early
indication of neurobiological profiles that may be more susceptible to the effects of
substances of abuse (e.g., poor inhibitory control as evidenced by impulsive decision making
and poor executive control as evidenced by low response inhibition). This is suggested by
the fact that drug addicted individuals and individuals with a family history of drug
addiction present with similar cognitive and neuropsychological deficits (Iacono, et al.,
2008). Among the many neurological domains that influence drug addiction (Volkow, et al.,
2012), the prefrontal cortex (PFC) is a common structure that links ADHD, CD, NS, and
SUDs (Arnsten, 2009). The PFC is necessary for event planning, emotional regulation,
attention, and regulating behavior. Variation in PFC functioning, as manifested by these
childhood/adolescent disorders represents a singular component of a larger network of
factors that contribute to the liability to substance addiction. It is clear from the findings of
this study, that understanding and identifying these associations will help to build this
network and generate new hypotheses for addiction research. These findings also have
important implications for building genetic models of addiction liability. To date, a number
of twin studies have demonstrated an evidence of unique and shared genetic mechanisms
between externalizing behavior and substance problems (Button, et al., 2006; Edwards &
Kendler, 2012; Knopik, Heath, Bucholz, Madden, & Waldron, 2009). For instance, several
studies have suggested that both novelty seeking and drug seeking behaviors are mediated
by the mesolimbic dopamine system making it an ideal starting point for genetic studies. A
testament of the validity of this hypothesis has been evidence from pharmacological studies
which have shown that blockage of dopamine receptors using dopamine antagonists, such as
haloperidol, blocks/reduces both drug seeking and novelty seeking behaviors (Bardo,
Donohew, & Harrington, 1996). Recent studies also suggest that females with the CatecholO-methyltransferase (COMT; a protein responsible for the degradation of dopamine in the
synapse) Met/Met genotypes have a higher mean-level of NS symptoms compared to
females with the Val/Val or Val/Met genotypes (Golimbet, Alfimova, Gritsenko, & Ebstein,
2007), although these relationships may be sample dependent and should be replicated
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 9
NIH-PA Author Manuscript
across larger samples and similar measures. NS has also been shown to mediate the
association between DRD4 and drinking (Laucht, Becker, Blomeyer, & Schmidt, 2007) and
smoking behaviors (Laucht, Becker, El-Faddagh, Hohm, & Schmidt, 2005), as well as the
association between COMT and the age of onset of drug use (Li, et al., 2011). Further, the
Val 158 Met polymorphism in the COMT gene also influences the level of CD and ADHD
symptoms in males (DeYoung, et al., 2010). Overall, the evidence suggests that variation
within the mesolimbic dopamine system influences the liability to drug use, CD, ADHD,
and novelty seeking tendencies, however, further research with larger samples and using
multiple measures is necessary to obtain robust estimates of association as many studies
have been unable to conduct multivariate analyses of the type presented in this manuscript.
Such investigations are becoming increasingly possible with the developmental research
projects, such as the Genes Environment Development Initiative (Minnesota Center for
Twin and Family Research) and the Center for Antisocial Drug Dependence, which combine
whole genome genotyping with extensive behavioral assessments.
4.3 Limitations of research
NIH-PA Author Manuscript
A notable limitation, that is not unique to this type of study, is that not everyone in the study
has passed through the “age of risk” (i.e., ages 21–25) (Substance Abuse and Mental Health
Services Administration, 2011) for developing a substance use disorder; however, most of
our subjects are within the age range where the highest levels of substance use and disorders
have been observed. In the future, we plan to overcome this limitation by using data
collected from the third wave of assessment of respondents at the Center for Antisocial Drug
Dependence, and similar longitudinal studies. Despite this limitation, the current findings
support previous work highlighting NS tendencies as a powerful early indicator of future
SUDs. In addition, it suggests that understanding the mechanisms underlying ADHD and
CD may help to inform our understanding of the shared mechanisms underlying substances
of abuse, as well as unique mechanisms for some substances. Another limiting factor is the
limited inclusion of socio-demographic variables other than race, such as socioeconomic,
and peer and family substance use. Additional research in samples with detailed
environmental variables is still needed to better estimate these effects in such contexts.
5. Conclusions
NIH-PA Author Manuscript
In summary, this investigation indicates that in a model including CD, adolescent substance
use and problems, ADHD subscales, and NS, levels of adolescent conduct problems,
substance use, and an innate tendency to seek out novel stimuli (i.e., NS) are the most robust
indicators of young adult substance dependence. In conclusion, the characterization of
personality and psychopathology via prospective longitudinal studies is an integral
component in our understanding of the liabilities to alcohol and drug addiction.
Acknowledgments
Role of Funding Sources
Funding for this study was provided by AA021113, MH019927, MH063207, HD010333, DA011015, DA021913,
and DA023134. NIAAA, NIMH, NIDA, and NICHD had no role in the study design, collection, analysis or
interpretation of the data, writing the manuscript, or the decision to submit the paper for publication.
References
American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4.
Washington, DC: Author; 2000. text rev
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 10
NIH-PA Author Manuscript
NIH-PA Author Manuscript
NIH-PA Author Manuscript
Arnsten AF. Toward a new understanding of attention-deficit hyperactivity disorder pathophysiology:
an important role for prefrontal cortex dysfunction. CNS Drugs. 2009; 23(Suppl 1):33–41.
[PubMed: 19621976]
August GJ, Winters KC, Realmuto GM, Fahnhorst T, Botzet A, Lee S. Prospective study of adolescent
drug use among community samples of ADHD and non-ADHD participants. J Am Acad Child
Adolesc Psychiatry. 2006; 45:824–832. [PubMed: 16832319]
Bardo MT, Donohew RL, Harrington NG. Psychobiology of novelty seeking and drug seeking
behavior. Behav Brain Res. 1996; 77:23–43. [PubMed: 8762157]
Bidwell LC, Palmer RHC, Health AC, Madden PAF, Bucholz K, McGeary JE, Knopik VS. Common
and specific genetic effects on ADHD and Initial Sensitivity to Cigarettes in Female Adolescent
Twins. Behav Genet. 2012; 42:920.
Brook JS, Duan T, Zhang C, Cohen PR, Brook DW. The association between attention deficit
hyperactivity disorder in adolescence and smoking in adulthood. Am J Addict. 2008; 17:54–59.
[PubMed: 18214723]
Button TM, Hewitt JK, Rhee SH, Young SE, Corley RP, Stallings MC. Examination of the causes of
covariation between conduct disorder symptoms and vulnerability to drug dependence. Twin Res
Hum Genet. 2006; 9:38–45. [PubMed: 16611466]
Charach A, Yeung E, Climans T, Lillie E. Childhood attention-deficit/hyperactivity disorder and future
substance use disorders: comparative meta-analyses. J Am Acad Child Adolesc Psychiatry. 2011;
50:9–21. [PubMed: 21156266]
DeYoung CG, Getchell M, Koposov RA, Yrigollen CM, Haeffel GJ, af Klinteberg B, Oreland L,
Ruchkin VV, Pakstis AJ, Grigorenko EL. Variation in the catechol-O-methyltransferase Val 158
Met polymorphism associated with conduct disorder and ADHD symptoms, among adolescent male
delinquents. Psychiatr Genet. 2010; 20:20–24. [PubMed: 19997043]
Disney ER, Elkins IJ, McGue M, Iacono WG. Effects of ADHD, conduct disorder, and gender on
substance use and abuse in adolescence. Am J Psychiatry. 1999; 156:1515–1521. [PubMed:
10518160]
Edwards AC, Kendler KS. Twin study of the relationship between adolescent attention-deficit/
hyperactivity disorder and adult alcohol dependence. J Stud Alcohol Drugs. 2012; 73:185–194.
[PubMed: 22333326]
Elkins IJ, King SM, McGue M, Iacono WG. Personality traits and the development of nicotine,
alcohol, and illicit drug disorders: prospective links from adolescence to young adulthood. J
Abnorm Psychol. 2006; 115:26–39. [PubMed: 16492093]
Elkins IJ, McGue M, Iacono WG. Prospective effects of attention-deficit/hyperactivity disorder,
conduct disorder, and sex on adolescent substance use and abuse. Arch Gen Psychiatry. 2007;
64:1145–1152. [PubMed: 17909126]
Fergusson DM, Horwood LJ, Ridder EM. Conduct and attentional problems in childhood and
adolescence and later substance use, abuse and dependence: results of a 25-year longitudinal
study. Drug Alcohol Depend. 2007; 88(Suppl 1):S14–26. [PubMed: 17292565]
Golimbet VE, Alfimova MV, Gritsenko IK, Ebstein RP. Relationship between dopamine system genes
and extraversion and novelty seeking. Neurosci Behav Physiol. 2007; 37:601–606. [PubMed:
17657431]
Grant BF, Dawson DA. Age at onset of alcohol use and its association with DSM-IV alcohol abuse
and dependence: results from the National Longitudinal Alcohol Epidemiologic Survey. J Subst
Abuse. 1997; 9:103–110. [PubMed: 9494942]
Grekin ER, Sher KJ, Wood PK. Personality and substance dependence symptoms: modeling
substance-specific traits. Psychol Addict Behav. 2006; 20:415–424. [PubMed: 17176176]
Guy SM, Smith GM, Bentler PM. Consequences of adolescent drug use and personality factors on
adult drug use. J Drug Educ. 1994; 24:109–132. [PubMed: 7931922]
Heath AC, Cloninger CR, Martin NG. Testing a model for the genetic structure of personality: a
comparison of the personality systems of Cloninger and Eysenck. J Pers Soc Psychol. 1994;
66:762–775. [PubMed: 8189351]
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 11
NIH-PA Author Manuscript
NIH-PA Author Manuscript
NIH-PA Author Manuscript
Iacono WG, Malone SM, McGue M. Behavioral disinhibition and the development of early-onset
addiction: common and specific influences. Annu Rev Clin Psychol. 2008; 4:325–348. [PubMed:
18370620]
Jenkins MB, Agrawal A, Lynskey MT, Nelson EC, Madden PA, Bucholz KK, Heath AC. Correlates of
alcohol abuse/dependence in early-onset alcohol-using women. Am J Addict. 2011; 20:429–434.
[PubMed: 21838841]
Knopik VS, Heath AC, Bucholz KK, Madden PA, Waldron M. Genetic and environmental influences
on externalizing behavior and alcohol problems in adolescence: a female twin study. Pharmacol
Biochem Behav. 2009; 93:313–321. [PubMed: 19341765]
Laucht M, Becker K, Blomeyer D, Schmidt MH. Novelty seeking involved in mediating the
association between the dopamine D4 receptor gene exon III polymorphism and heavy drinking in
male adolescents: results from a high-risk community sample. Biol Psychiatry. 2007; 61:87–92.
[PubMed: 16945348]
Laucht M, Becker K, El-Faddagh M, Hohm E, Schmidt MH. Association of the DRD4 exon III
polymorphism with smoking in fifteen-year-olds: a mediating role for novelty seeking? J Am
Acad Child Adolesc Psychiatry. 2005; 44:477–484. [PubMed: 15843770]
Lee SS, Humphreys KL, Flory K, Liu R, Glass K. Prospective association of childhood attentiondeficit/hyperactivity disorder (ADHD) and substance use and abuse/dependence: a meta-analytic
review. Clin Psychol Rev. 2011; 31:328–341. [PubMed: 21382538]
Li T, Yu S, Du J, Chen H, Jiang H, Xu K, Fu Y, Wang D, Zhao M. Role of novelty seeking personality
traits as mediator of the association between COMT and onset age of drug use in Chinese heroin
dependent patients. PLoS One. 2011; 6:e22923. [PubMed: 21857968]
Miller E, Stephen J, Tudway J. Assessing the component structure of four self-report measures of
impulsivity. Personality & Individual Differences. 2004; 37:349–358.
Muthen, LK.; Muthen, BO. MPlus User’s Guide. 7. Los Angeles, CA: Muthen & Muthen; 1998–2012.
Palmer RH, Bidwell C, Bucholz K, Madden P, Heath AC, McGeary JE, Knopik VS. The Role of
Conduct Disorder in Explaining the Common Genetic Influences on Sensitivity to Alcohol and
Tobacco. Behav Genet. 2012; 42:958.
Palmer RH, Young SE, Hopfer CJ, Corley RP, Stallings MC, Crowley TJ, Hewitt JK. Developmental
epidemiology of drug use and abuse in adolescence and young adulthood: Evidence of generalized
risk. Drug Alcohol Depend. 2009; 102:78–87. [PubMed: 19250776]
Pardini D, White HR, Stouthamer-Loeber M. Early adolescent psychopathology as a predictor of
alcohol use disorders by young adulthood. Drug Alcohol Depend. 2007; 88(Suppl 1):S38–49.
[PubMed: 17257781]
Rhea SA, Gross AA, Haberstick BC, Corley RP. Colorado Twin Registry. Twin Res Hum Genet.
2006; 9:941–949. [PubMed: 17254434]
Rhea SA, Gross AA, Haberstick BC, Corley RP. Colorado twin registry: an update. Twin Res Hum
Genet. 2013; 16:351–357. [PubMed: 23092589]
SAS. System for Windows. Cary, NC, USA: SAS Institute Inc; 2002. Copyright © 2002 – 2003
Shaffer D, Fisher P, Lucas CP, Dulcan MK, Schwab-Stone ME. NIMH Diagnostic Interview Schedule
for Children Version IV (NIMH DISC-IV): description, differences from previous versions, and
reliability of some common diagnoses. J Am Acad Child Adolesc Psychiatry. 2000; 39:28–38.
[PubMed: 10638065]
Squeglia LM, Jacobus J, Tapert SF. The influence of substance use on adolescent brain development.
Clin EEG Neurosci. 2009; 40:31–38. [PubMed: 19278130]
Squeglia LM, Spadoni AD, Infante MA, Myers MG, Tapert SF. Initiating moderate to heavy alcohol
use predicts changes in neuropsychological functioning for adolescent girls and boys. Psychol
Addict Behav. 2009; 23:715–722. [PubMed: 20025379]
Substance Abuse and Mental Health Services Administration. N S H- Results from the 2010 National
Survey on Drug Use and Health: Summary of National Findings, HHS Publication No (SMA) 11–
4658. Rockville, MD: Substance Abuse and Mental Health Services Administration; 2011. Results
from the 2010 National Survey on Drug Use and Health: Summary of National Findings.
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 12
NIH-PA Author Manuscript
Tarter RE, Kirisci L, Feske U, Vanyukov M. Modeling the pathways linking childhood hyperactivity
and substance use disorder in young adulthood. Psychol Addict Behav. 2007; 21:266–271.
[PubMed: 17563150]
Volkow ND, Wang GJ, Fowler JS, Tomasi D. Addiction circuitry in the human brain. Annu Rev
Pharmacol Toxicol. 2012; 52:321–336. [PubMed: 21961707]
Wills TA, Vaccaro D, McNamara G. Novelty seeking, risk taking, and related constructs as predictors
of adolescent substance use: an application of Cloninger’s theory. J Subst Abuse. 1994; 6:1–20.
[PubMed: 8081104]
Young SE, Stallings MC, Corley RP, Krauter KS, Hewitt JK. Genetic and environmental influences on
behavioral disinhibition. Am J Med Genet. 2000; 96:684–695. [PubMed: 11054778]
Zucker RA. Anticipating problem alcohol use developmentally from childhood into middle adulthood:
what have we learned? Addiction. 2008; 103(Suppl 1):100–108. [PubMed: 18426543]
NIH-PA Author Manuscript
NIH-PA Author Manuscript
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 13
Highlights
NIH-PA Author Manuscript
•
The predictive effects of adolescent indicators of Behavioral Disinhibition
varied by substance.
•
ADHD subscales were positive associated with illicit drug dependence.
•
Conduct disorder and novelty seeking were predictive of alcohol, tobacco, and
cannabis dependence. These effects were partially mediated by adolescent
substance use.
•
The strongest predictor of future drug problems was novelty seeking.
NIH-PA Author Manuscript
NIH-PA Author Manuscript
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Palmer et al.
Page 14
NIH-PA Author Manuscript
NIH-PA Author Manuscript
Figure 1.
Path diagram depicting direct (solid lines) and indirect (dashed lines) relationships tested in
the current study. For simplicity, indirect pathways via substance use are only shown for
alcohol dependence symptoms at Wave 2. Other covariates included in the model were age
at Wave 1, gender, and race (none of which are shown for simplicity). Abbreviations: ATT,
ADHD attention problems, HI – ADHD hyperactivity/impulsivity, CD – DSM-IV conduct
disorder symptoms, DV – Dependence vulnerability, NS – novelty seeking.
NIH-PA Author Manuscript
Addict Behav. Author manuscript; available in PMC 2014 September 01.
NIH-PA Author Manuscript
NIH-PA Author Manuscript
NIH-PA Author Manuscript
Table 1
Females
Behavior
N
Males
Means Test
Mean (STD)
Skew
Kurtosis
N
Mean (STD)
Skew
Kurtosis
Z-value
Palmer et al.
Mean level of untransformed scores for each phenotype by gender
Wave 1 independent variables (adolescence)
Addict Behav. Author manuscript; available in PMC 2014 September 01.
ADHD - Hyperactivity/Impulsivity
1268
.18 (.91)
5.51
30.15
1093
.35 (1.25)
3.78
13.00
3.43 b
ADHD - Inattention
1268
.21 (.98)
5.04
24.87
1093
.43 (1.40)
3.27
9.32
4.16 b
Conduct Disorder
1268
.59 (.95)
2.42
8.50
1093
1.00 (1.37)
2.25
8.08
7.41 c
Novelty Seeking
1249
.43 (.19)
.32
−.28
1080
.47 (.19)
0.11
−.37
5.03 c
# of substances used
1268
.53 (1.11)
2.71
8.31
1093
.48 (.98)
2.99
12.13
−0.82
# of substances with problems
1268
.17 (.57)
4.21
20.24
1093
.16 (.60)
5.68
40.46
−0.39
Wave 2 (young adulthood)dependent variables
Alcohol Dependence Sx
984
1.01 (1.54)
1.91
3.55
901
1.34 (1.63)
1.30
1.11
3.92 c
Tobacco Dependence Sx
379
2.65 (1.94)
.18
−1.13
529
2.24 (2.03)
0.47
−1.03
−2.80 b
Illicit Drug Dependence Sx
379
1.95 (3.02)
2.32
6.57
452
1.97 (3.23)
3.59
22.28
.10
DV
1268
1.02 (1.23)
1.29
1.37
904
1.19 (1.21)
1.07
.95
2.78 b
Note: For gender difference tests using the transformed scores.
a
p<.05,
b
c
p<.01,
p<.001.
STD - standard deviation, Sx - Symptoms.
Page 15
NIH-PA Author Manuscript
NIH-PA Author Manuscript
NIH-PA Author Manuscript
Table 2
Addict Behav. Author manuscript; available in PMC 2014 September 01.
Wave 1 (Adolescence)
Wave 2 (Young Adult) Outcomes
Wave 1 (adolescent) variables
Substance Use
Alcohol
ADHD - Hyperactivity/Impulsivity
.000 (.024)
ADHD - Inattention
.118 (.044) b
Conduct disorder
.350 (.026) c
Novelty seeking
Tobacco
Illicit
DV
−.0228 (.024)
.002 (.037)
−.024 (.035)
−.009 (.024)
.023 (.039)
.049 (.041)
.123 (.048) b
.077 (.034) a
.096 (.027) c
.032 (.033)
.135 (.036) a
.093 (.026) c
.123 (.020) c
.141 (.022) c
.129 (.033) a
.110 (.033) b
.164 (.022) c
Substance use
-
.131 (.031) c
.128 (.038) b
.057 (.043)
.160 (.032) c
Substance problems
-
.078 (.037) a
.117 (.037 ) b
.125 (.044) b
.141 (.033) c
Age
−.074 (.045)
−.093 (.040) a
−.121 (.044 ) b
−.139 (.050) b
.056 (.034)
Gender
−.004 (.021)
−.014 (.022)
.029 (.036)
−.013 (.031)
−.007 (.022)
Race
.002 (.023)
.023 (.022)
.010 (.029)
.017 (.027)
−.013 (.020)
Palmer et al.
Standardized regression coefficients (standard error) from the models predicting young adult drug dependence
Note: ADHD: Attention Deficit Hyperactivity Disorder. Standardized coefficients are given with standard errors in parentheses. - indicates parameters that were not estimated in the model.
a
p<.05,
b
c
p<.01,
p<.001.
Page 16
NIH-PA Author Manuscript
NIH-PA Author Manuscript
NIH-PA Author Manuscript
Table 3
Palmer et al.
Standardized total, indirect, and direct effects on wave 2 outcomes
Standardized Effects on Wave 2 (Young Adulthood) Outcomes
Total Effect
Models
Specific Indirect Effect (via Use)
Direct Effect
Estimate
S.E.
P-Value
Estimate
S.E.
P-Value
Estimate
S.E.
P-Value
−.028
.025
.240
.000
.003
1.000
−.028
.024
.250
Alcohol
Addict Behav. Author manuscript; available in PMC 2014 September 01.
ADHD - Hyperactivity/Impulsivity
ADHD - Inattention
.038
.040
.339
.015
.007
.021
.023
.039
.558
Conduct Disorder
.142
.027
<.001
.046
.012
<.001
.096
.027
<.001
Novelty Seeking
.157
.022
<.001
.016
.005
.001
.141
.022
<.001
Tobacco
ADHD - Hyperactivity/Impulsivity
.002
.037
.983
.000
.003
1.000
.002
.037
.960
ADHD - Inattention
.064
.043
.134
.015
.007
.037
.049
.041
.233
Conduct Disorder
.077
.033
.019
.045
.014
.001
.032
.033
.329
Novelty Seeking
.145
.033
<.001
.016
.005
.003
.129
.033
<.001
−.024
.035
.491
.000
.001
1.000
−.024
.035
.489
Illicit
ADHD - Hyperactivity/Impulsivity
ADHD - Inattention
.130
.049
.008
.007
.006
.235
.123
.048
.011
Conduct Disorder
.155
.034
<.001
.020
.015
.191
.135
.036
<.001
Novelty Seeking
.117
.033
<.001
.007
.005
.197
.110
.033
.001
ADHD - Hyperactivity/Impulsivity
−.009
.025
.717
.000
.004
1.000
−.009
.024
.712
ADHD - Inattention
.096
.035
.006
.019
.008
.016
.077
.034
.021
Conduct Disorder
.149
.026
<.001
.056
.012
<.001
.093
.026
<.001
Novelty Seeking
.184
.022
<.001
.020
.005
<.001
.164
.022
<.001
DV
Standardized parameter estimates are shown along with their standard error and corresponding p-value for each model predicting the Wave 2 dependent variables.
Page 17