Anxiety in Health Behaviors and Physical Illness

SERIES IN ANXIETY AND RELATED DISORDERSSeries Editor: Martin M. Antony, Professor, Department of Psychology,

Ryerson University, Toronto, Ontario, Canada

ACCEPTANCE AND MINDFULNESS-BASED APPROACHES TO ANXIETYConceptualization and TreatmentEdited by Susan M. Orsillo and Lizabeth Roemer

CONCEPTS AND CONTROVERSIES IN OBSESSIVE-COMPULSIVE DISORDEREdited by Jonathan S. Abramowitz and Arthur C. Houts

SOCIAL ANXIETY AND SOCIAL PHOBIA IN YOUTHCharacteristics, Assessment, and Psychological TreatmentChristopher A. Kearney

TREATING HEALTH ANXIETY AND FEAR OF DEATHA Practitioner’s GuidePatricia Furer, John R. Walker, and Murray B. Stein

TREATING TRICHOTILLOMANIACognitive-Behavioral Therapy for Hairpulling and Related ProblemsMartin E. Franklin and David F. Tolin

ANXIETY IN HEALTH BEHAVIORS AND PHYSICAL ILLNESS

A Continuation Order Plan is available for this series. A continuation order will bring delivery of each new volume immediately upon publication. Volumes are billed only upon actual shipment. For further information please contact the publisher.

Edited by Michael J. Zvolensky and Jasper A. J. Smits

Anxiety in Health Behaviors and Physical

Illness

Michael J. ZvolenskyUniversity of Vermont

Burlington, Vermont, USA

Jasper A. J. SmitsSouthern Methodist University

Dallas, Texas, USA

Michael J. Zvolensky, Ph.D.Department of PsychologyUniversity of VermontJohn Dewey Hall2 Colchester AvenueBurlington, VT [email protected]

Jasper A. J. Smits, Ph.D.Department of PsychologySouthern Methodist University6424 Hilltop LaneDallas, TX 75205USA

ISBN 978-0-387-74752-1 e-ISBN 978-0-387-74753-8

Library of Congress Control Number: 2007935078

© 2008 Springer Science+Business Media, LLCAll rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden.The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights.

Printed on acid-free paper.

9 8 7 6 5 4 3 2 1

springer.com

[email protected]

Preface

Research has been accumulating on the prevalence and nature of the co-occurrence between various forms of anxiety disorders and problematic healthbehaviors as well as physical illness. This research has significant implicationsfor both those interested and affected by anxiety as well as physical healthfactors. Yet, it is striking that there has been little systematic integration of thishealth-oriented research in contemporary science and practice on anxiety andits disorders. This relative neglect is unfortunate given that the co-occurrence ofanxiety and health problems is a major public health priority when measuredboth in human and financial terms.

The overarching aim of this book is to provide a single resource that offerscurrent theoretical perspectives and cutting eZdge reviews of scientific researchon health behaviors and physical illness in relation to anxiety and its disorders.A critical analysis of this emerging literature is needed to help move this fieldforward, making this proposed volume timely. The specific objectives of thisedited book are to (1) provide a review of the literature on the link betweenanxiety and certain health behaviors and processes as well as physical illness;(2) present contemporary theories of their co-occurrence and interplay (e.g.,onset, maintenance, and relapse); and (3) provide an analysis of recent researchin regard to therapeutic models for targeting these problems.

The book is organized into two general sections. In the first part of the book,prototypical health behaviors – smoking, alcohol, illicit substance use, exercise,and sleep – are discussed in relation to anxiety and its disorders. In the secondpart of the book, the association between anxiety psychopathology and physi-cal health conditions – chronic pain, cardiovascular disease, asthma, HIV/AIDS – and their treatment are covered. In this same section, the potentialrole of puberty and the menstrual cycle in the onset and maintenance of anxietypsychopathology are discussed.

Inspection of the excellent and comprehensive works has yielded a number ofbroad-based conclusions relevant to informing research and practice for anxi-ety disorders. First, there is consistent empirical evidence that medical problemsand poor health behaviors are overrepresented among persons with anxietydisorders, and vice versa. Thus, there is a pressing need to marshal informationon anxiety-health processes to better serve this population. Second, as each

v

contribution makes clear, there is uniform evidence that both health behaviorsand physical illness can, and do, affect the nature of anxiety psychopathology.Yet, the exact nature of these associations depends on the specific disorder andhealth factor in question. And finally, a variety of the chapters make clear thatpersons suffering from anxiety psychopathology and poor health behaviors ormedical illnesses may need specialized interventions to prompt clinical change.That is, traditional interventions may not be ideally suited or maximize clinicalbenefit for this population.

For us, the present book offers the opportunity to appreciate the importanceand complexity involved with the study of anxiety disorders. For many years,health behaviors and medical illnesses have been a neglected facet of anxietydisorder research and practice. The contributions in this book help drive homethe message that such neglect is unwarranted, and that by working to betterunderstand the enigmas between health status and functioning and anxietypsychopathology, significant clinically-relevant strides can likely be achieved.We hope the present book helps move such work forward and bring a betterquality of life and reduced morbidity to persons with anxiety disorders in thefuture.

We owe gratitude to many people who have helped us complete this project.First among these are the experts who authored the chapters. We would like tothank them for their hard work and dedication. We also appreciate the com-ments and suggestions of Dr. Martin Antony, the editor of the Series in Anxietyand Related Disorders, and the assistance of Anna Tobias of Spinger with thepublishing of this book. Lastly, we continue to be appreciative of our respectivefamily members, Heidi and Jack Zvolensky and Jill and Stella Smits, for theirlove and support.

April 2007 Michael J. Zvolensky, Ph.D.University of Vermont

Jasper A. J. Smits, Ph.D.Southern Methodist University

vi Preface

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

Part I: Health Behaviors and Anxiety Disorders

Tobacco Use and Panic Psychopathology: Current Status and Future

Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3Michael J. Zvolensky, Theresa Leyro, Amit Bernstein,Matthew T. Feldner, Andrew R. Yartz, Kimberly Babson, andMarcel O. Bonn-Miller

Alcohol Use and Anxiety Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29Brigitte C. Sabourin and Sherry H. Stewart

Illicit Drug Use Across the Anxiety Disorders: Prevalence, Underlying

Mechanisms, and Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Matthew T. Tull, David E. Baruch, Michelle S. Duplinsky,and C. W. Lejuez

The Promise of Exercise Interventions for the Anxiety Disorders . . . . . . . . 81Jasper A. J. Smits, Angela C. Berry, Mark B. Powers, Tracy L. Greer,and Michael W. Otto

Anxiety and Insomnia: Theoretical Relationship and Future

Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105Thomas W. Uhde and Bernadette M. Cortese

Part II: Physical Conditions and Anxiety Disorders

Anxiety Disorders and Physical Illness Comorbidity: An Overview . . . . . . 131Tanya Sala, Brian J. Cox, and Jitender Sareen

vii

The Relation Between Puberty and Adolescent Anxiety: Theory

and Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155Ellen W. Leen-Feldner, Laura E. Reardon, Chris Hayward, andRose C. Smith

Anxiety, Anxiety Disorders, and the Menstrual Cycle . . . . . . . . . . . . . . . . 181Sandra T. Sigmon and Janell G. Schartel

Pain and Anxiety Disorders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207Gordon J.G. Asmundson,Murray P. Abrams, andKelsey C. Collimore

Asthma in Anxiety and Its Disorders: Overview and Synthesis . . . . . . . . . . 237Lisa S. Elwood and Bunmi O. Olatunji

Cardiovascular Disease and Anxiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279Kamila S. White

HIV and Anxiety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317Conall O’Cleirigh, Trevor A. Hart, and Carolyn A. James

Physical Illness and Treatment of Anxiety Disorders: A Review . . . . . . . . . 341Norman B. Schmidt, Meghan E. Keough, Lora Rose Hunter, and AnnP. Funk

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367

viii Contents

Contributors

Murray P. Abrams, B.A., Anxietyand Illness Behaviours Laboratory,University of Regina

Gordon J.G. Asmundson, Ph.D.Anxiety and Illness BehavioursLaboratory, University of Regina

Kimberly Babson, B.A., Departmentof Psychology, University ofArkansas

David E. Baruch, Center forAddictions, Personality, and EmotionResearch and the University ofMaryland

Amit Bernstein, Ph.D., VeteransAffairs Palo Alto Health Care Systemand University of Vermont

Angela C. Berry, M.A., Departmentof Psychology, Southern MethodistUniversity

Marcel Bonn-Miller, B.A.,Department of Psychology,University of Vermont

Kelsey C. Collimore, B.S., Anxietyand Illness Behaviours Laboratory,University of Regina

Bernadette M. Cortese, Ph.D.,Department of Psychiatry, Penn StateCollege of Medicine and HersheyMedical Center

Brian J. Cox Ph.D., Department ofCommunity Health Sciences andDepartment of Psychiatry,University of Manitoba

Michelle S. Duplinsky, Center forAddictions, Personality, and EmotionResearch and the University ofMaryland

Lisa S. Elwood, M.A., Department ofPsychology, University of Arkansas

Matthew T. Feldner, Ph.D.,Department of Psychology,University of Arkansas

Ann P. Funk, M.A., Department ofPsychology, FloridaState University

Tracy L. Greer, Ph.D., Department ofPsychiatry, University of TexasSouthwestern Medical Center atDallas

ix

Trevor A. Hart, Ph.D.,Department of Psychology,York University

Chris Hayward, M.D., M.P.H.,Department of Psychiatry andBehavioral Sciences, StanfordUniversity

Lora Rose Hunter, B.A., Departmentof Psychology, Florida StateUniversity

Carolyn A. James, M.A., Departmentof Psychology, York University,Toronto

Meghan E. Keough, M.S.,Department of Psychology, FloridaState University

Ellen W. Leen-Feldner, Ph.D.,Department of Psychology,University of Arkansas

Carl W. Lejuez, Ph.D., Center forAddictions, Personality, and EmotionResearch and the University ofMaryland

Teresa Leyro, B.A., Department ofPsychology, University of Vermont

Conall O’Cleirigh, Ph.D.,Massachusetts General Hospital/Harvard Medical School and FenwayCommunity Health

Bunmi O. Olatunji, Ph.D.,Department of Psychology,Vanderbilt University

Michael W. Otto, Ph.D., Center forAnxiety and Related Disorders,Boston University

Mark B. Powers, Ph.D., Departmentof Psychology, University ofAmsterdam

Laura E. Reardon, M.A.,Department of Psychology,University of Arkansas

Brigitte C. Sabourin, B.A.,Department of Psychology,Dalhousie University

Tanya Sala, M.D., FRCPC,Department of Psychiatry, Universityof Manitoba

Jitender Sareen B.Sc., M.D., FRCPC,Department of Community HealthSciences and Department ofPsychiatry, University of Manitoba

Janell G. Schartel, M.A., Departmentof Psychology, University of Maine

Norman B. Schmidt, Ph.D.,Department of Psychology, FloridaState University

Sandra T. Sigmon, Ph.D.Department of Psychology,University of Maine

Rose C. Smith, B.A.,Department of Psychology,University of Arkansas

Jasper A. J. Smits, Ph.D.,Department of Psychology, SouthernMethodist University

Sherry H. Stewart, Ph.D.,Departments of Psychiatry andPsychology, Dalhousie University

x Contributors

Matthew T. Tull, Ph.D., Centerfor Addictions, Personality,and Emotion Research and theUniversity of Maryland

Thomas W. Uhde, M.D.Department of Psychiatry,Penn State College of Medicineand Hershey Medical Center

Kamila S. White, Ph.D., Universityof Missouri-Saint Louis

Andrew R. Yartz, Ph.D., Departmentof Psychology, University of Vermont

Michael J. Zvolensky, Ph.D.,Department of Psychology,University of Vermont

Contributors xi

Part I

Health Behaviors and Anxiety Disorders

Tobacco Use and Panic Psychopathology: Current

Status and Future Directions

Michael J. Zvolensky, Teresa Leyro, Amit Bernstein, Matthew T. Feldner,

Andrew R. Yartz, Kimberly Babson, and Marcel O. Bonn-Miller

Recently, there has been increased effort to better understand linkages between

tobacco use and the anxiety disorders (Feldner, Babson, & Zvolensky, 2007;

Morissette, Tull, Gulliver, Kamholz, & Zimering, 2007; Morrell & Cohen,

2006; Zvolensky, Bernstein, Marshall, & Feldner, 2006; Zvolensky, Feldner,

Leen-Feldner, & McLeish, 2005; Zvolensky, Schmidt, & Stewart, 2003). These

efforts are theoretically and clinically important because substance use pro-

blems frequently co-occur with anxiety psychopathology, and anxiety-related

factors often play a role in tobacco use and dependence (Morissette et al.,

2007; Morrell & Cohen, 2006; Zvolensky, Bernstein et al., 2006). However, our

understanding of the explanatory nature of these comorbid relations is only

beginning to emerge. The purpose of the present chapter is to provide a current

review of extant empirical work pertaining to the inter-relations between

tobacco use and panic psychopathology. We expressly and specifically focus

on panic psychopathology, rather than anxiety disorders more broadly, as

most of the work on tobacco and anxiety relations to date has focused on

panic.This chapter is organized into four sections. First, we briefly describe panic

psychopathology. Second, we review risk factor terminology developed by

Kraemer, Kazdin, and Offord (1997), in order to establish a nomenclature for

conceptualizing interrelations between tobacco and panic psychopathology.

Third, we describe tobacco use and common patterns of use. Fourth, we discuss

the nature of comorbidity between tobacco use and panic psychopathology and

review and evaluate the related empirical evidence. We focus both on the role of

tobacco use in the onset and maintenance of panic psychopathology, and the

role of panic factors and processes in the onset and maintenance of smoking.

Within each of these sections, we identify gaps in the existing literature and

highlight formative questions for future research.

Michael J. ZvolenskyUniversity of Vermont, John Dewey Hall, Burlington, VT 05405-0134, Tel: 802-656-8994,Fax: [email protected]

M. J. Zvolensky, J. A. Smits (eds.),Anxiety in Health Behaviors and Physical Illness.� Springer 2008

3

Panic Psychopathology

The term ‘‘panic psychopathology’’ is used in this chapter to denote panicattacks, panic disorder, and agoraphobia (with or without panic disorder).Panic attacks are a subjective sense of extreme fear or impending doom accom-

panied by an autonomic nervous system surge and a strong flight-or-fightaction tendency (Barlow, Brown, & Craske, 1994). Recent estimates of unex-pected (‘‘out of the blue’’) panic attacks in representative samples suggest thatapproximately 20% of individuals experience such attacks at one point in theirlives and 11.2% in the past 12-months (Kessler, Chiu, Jin, Ruscio, Shear, &Walters, 2006). Thus, panic attacks are a relatively common psychologicalexperience and many people experience panic attacks without necessarily devel-oping panic disorder (i.e., nonclinical panic attacks; Norton, Cox, & Malan,1992). Typically, individuals who experience nonclinical panic attacks do notexperience these attacks as ‘‘spontaneous’’ or ‘‘uncued’’ (as is generally the casein panic disorder), but rather in certain contexts such as stressful or threateningsocial situations (Norton, 1989). Panic attacks also occur among thosewith other types of psychopathology (i.e., beyond panic disorder; Bryant &Panasetis, 2005). Panic attack onset can occur across the lifespan, but earlyonset tends to first occur between the ages of 12–13 years (Hayward et al., 1992;Warren & Zgourides, 1988; please see ‘‘Developmental Course’’ section belowfor further details).

Panic disorder involves recurrent unexpected panic attacks and anxiousapprehension about the possibility of experiencing future panic episodes(American Psychiatric Association, 2000). Lifetime estimates of panic disorderwithout agoraphobia are 3.7% and 1.1% for panic disorder with agoraphobia(Kessler et al., 2006). Twelve-month estimates for panic disorder (with orwithout agoraphobia) are approximately 2.8% (Kessler et al., 2006). Thus,panic disorder is a relatively common psychiatric disorder both in terms oflifetime and 12-month prevalence rates. This clinical condition is generallyregarded as a disorder of adulthood with a median age of onset of 24 (Burke,Burke, Regier, & Rae, 1990), although emerging research has noted thatanother possible ‘‘peak onset period’’ may be between ages 45–54 years(Burke et al., 1990). Panic disorder with and without agoraphobia is associatedwith a chronic, fluctuating course and high rates of both psychiatric comorbid-ity and substance use disorders (Zvolensky, Bernstein et al., 2006).

Although not all persons with panic disorder will meet diagnostic criteria foragoraphobia, individuals with panic disorder often show signs of avoidingpotentially threatening situations in which a panic attack might occur (Feldner,Zvolensky, & Leen-Feldner, 2004). Agoraphobia reflects a pattern of behaviorcharacterized by consistent avoidance of threatening situations where a panicattack or high anxiety is perceived to be likely (e.g., limited options to escape),or experiencing marked emotional distress when in such situations. Avoidancebehavior can be multifaceted, with a wide range of stimuli that are perceived as

4 M. J. Zvolensky et al.

threat-relevant (e.g., anything from being in crowds to certain substances like

caffeine; Feldner et al., 2004). Although agoraphobia does not necessarily

require the presence of panic attacks or panic disorder (Fava, Grandi, &

Canestrari, 1988), researchers frequently conceptualize agoraphobia as a com-

plication of (severe) panic disorder (Barlow, 2002). However, it is noteworthy

that this approach has been increasingly called into question in recent years, as

research indicates that there may be differing forms of agoraphobic avoidance

(Hayward & Wilson, in press). Regardless, agoraphobia with or without panic

disorder often is related to higher rates of clinically significant life impairment

and severity of illness (Kessler et al., 2006). The onset of agoraphobia with or

without panic disorder is not as firmly established as that of panic attacks and

panic disorder; though research suggests it likely occurs later in life than the

onset of panic attacks and panic disorder (Lindesay, 1991).

Vulnerability Nomenclature

Explicit delineation of terminology facilitates efforts to understand the nature

of associations between tobacco use and panic psychopathology. Led by the

work of Kraemer and colleagues, groundbreaking conceptual strides have

created a clearer understanding of various risk processes (Kazdin, Kraemer,

Kessler, Kupfer, & Offord, 1997; Kraemer et al., 1997; Kraemer, Lowe, and

Kupfer, 2005; Kraemer, Stice, Kazdin, Offord, &Kupfer, 2001). Specifically, as

is reviewed in this section, Kraemer and colleagues have worked to standardize

operational definitions for risk processes to increase the clarity and consistency

with which these factors are communicated across studies.Risk factors. A risk factor is a variable that is related to, and temporally

precedes, an unwanted outcome (Kraemer et al., 1997). Causal risk factors

reflect variables that, when modified in some way (e.g., through an interven-

tion), produce change (increase or decrease) in the dependent variable of

interest (Kraemer et al., 1997). Controlled research designs are necessary to

document causal effects because they can rule out competing alternative

explanations (e.g., ‘‘confounding variables’’). Proxy risk factors are variables

that are related to an outcome of interest, but this association is due to the proxy

risk factor’s relation with another causal risk factor (Kraemer et al., 2001).

Thus, change in a proxy risk factor would not yield corresponding systematic

change in an outcome variable; accordingly, a proxy risk factor may ‘‘mark’’

risk, but not explain or account for such risk.Due to the importance of the ability to change a risk factor, both risk and

proxy risk factors often are further categorized on the basis of whether or not

they aremalleable (i.e., can be changed or altered). When a risk factor cannot be

changed, it can be classified as a fixed marker, whereas when it can be changed,

it can be classified as a variable risk factor (Kraemer et al., 2005).

Panic and Tobacco 5

A risk factor also can be contrasted with a maintenance factor. A mainte-nance factor is a variable that predicts the persistence of an existing conditionover time among individuals already demonstrating the outcome (Stice, 2002).In theory, the same categorization scheme could be applied to maintenancefactors in terms of whether or not they are causal or proxy maintenance factors(Kazdin et al., 1997). Moreover, a risk factor also may subsequently function asa maintenance factor.

Qualifying conditions for risk factor effects. Clarifying relations betweenvulnerability processes and outcomes represents only the ‘‘first step’’ in a largerprocess of risk factor research. That is, this step represents a focus on ‘‘maineffects,’’ but does not explicate how, when, or among whom a specified riskprocess unfolds (see Zvolensky, Schmidt, Bernstein, & Keough, 2006). We donot delve fully into these explanatory processes within the present chapter, asextant work has largely focused on questions of main effects at this earlydevelopmental stage of this area of study.

Tobacco Use: Definition, Nature, and Prevalence

Cigarette smoking. Cigarette smoking is widely recognized as the most popularform of tobacco use and as a major public health problem (Windle & Windle,1999). Indeed, cigarette smoking remains a leading preventable cause of deathand disability in the United States (Centers for Disease Control and Prevention[CDC], 1994). Smoking is considered a key factor in various types of medicalillness, including heart disease, a variety of pulmonary diseases (e.g., chronicobstructive pulmonary disease), and many types of cancer (CDC, 1994, 2002).For instance, smoking is responsible for almost 31%of all cancer-related deaths(American Cancer Society [ACS], 2006a). Although smoking increases one’srisk for developing many of these lethal medical diseases, quitting smokingdecreases the risk of developing such problems and may increase the survivaltime among persons who have already developed such medical problems(Samet, 1992).

Despite a reduction in smoking prevalence over the past 25 years, approxi-mately 45–48 million (approximately 22% to 25%) adults in the U.S. currentlysmoke (CDC, 1996). Though nearly 70%of these smokers aremotivated to quit(CDC, 2002), approximately 90–95%of smokers who do try to quit smoking ontheir own (Cohen et al., 1989), and 60–80% who attend treatment programs,relapse to smoking (CDC, 2002). Additional work suggests that 64% of youth(adolescents) report having tried cigarettes and 14% have smoked frequently inthe past month (i.e., 20 out of the past 30 days; CDC, 2002). Thus, there isevidence that smoking is not only a major source of death and disability, butthat once started it is often difficult to stop.

Smokeless tobacco use. Although cigarettes are the most common type oftobacco use, there also is a significant population of smokeless tobacco


users in the U.S. and other regions of the world (e.g., India; ACS, 1999).Epidemiological data in the U.S., for example, suggest that approximately3% of individuals have used some form of smokeless tobacco – either snuff(finely ground, shredded tobacco) or chewing tobacco – in the past month(ACS, 1999). These rates of use are elevated among young Caucasian malescompared to females, and among those in southeastern and north central statescompared to other regions, as well as rural compared to urban settings(Hatsukami & Severson, 1999). For example, past work suggests that thehighest rates of current use (16.6%) are among Caucasian males 18–25 yearsof age (CDC, 1994).

In general, rates of smokeless tobacco use are noteworthy because smokelesstobacco contains carcinogens (e.g., tobacco-specific nitrosamines [TSNAs];National Cancer Institute and National Institute of Health (NCI/NIH), 2006)known to be causal agents in lung, oral, esophageal, liver, and pancreaticcancer (About: Smoking Cessation, 2006b). In addition, smokeless tobaccouse is associated with greater nicotine absorption and for longer periods oftime (e.g., stays in the bloodstream for greater durations of time) than cigarettesmoking (ACS, 2006; NCI/NIH 2006b). Similar to cigarettes, smokelesstobacco use can lead to increased rates of physical disease (e.g., oral cancer,gum disease) and nicotine addiction (ACS, 1999). Whereas the risks associatedwith cigarette use have become well-publicized in the U.S., public awarenessabout the dangers of smokeless tobacco remains limited. Indeed, many perceivesmokeless tobacco as a ‘‘risk-free alternative’’ to cigarette smoking (CDC,1994). Although as many as 50% of smokeless tobacco users report wantingto quit (ACS, 2006a), as with cigarettes, rates of relapse remain high (Hatsu-kami, Jensen, Allen, Grillo, & Bliss, 1996). In a recent review of smokelesstobacco treatment programs, Hatsukami and Severson (1999) estimated a 3–6month abstinent rate of 12%–30% with intensive behavioral treatments fairingbetter than other intervention options. Thus, similar to cigarette use, smokelesstobacco use is an addictive behavior that is difficult to quit and more intensivecare appears to yield relatively better outcomes.

Prevalence of Comorbid Tobacco Use and Panic Psychopathology

There have been both representative surveys and community-based studiesfocused on addressing the extent of the co-occurrence between tobacco useand panic psychopathology. To date, this work has largely centered on cigarettesmoking rather than smokeless tobacco. Additionally, from a historical per-spective, the vast majority of early work in this domain did not focus on panicor other specific anxiety conditions, but rather, addressed anxiety disorders as asingle ‘‘class of problems’’ (e.g., Breslau, 1995; Brown, Lewinsohn, Seeley, &Wagner, 1996; Costello, Erkanli, Federman, & Angold, 1999; Degenhardt,Hall, & Lynskey, 2001; Hughes, Hatsukami, Mitchell, & Dalgren, 1986;

Panic and Tobacco 7

Kandel, Huang, & Davies, 2001; Kandel et al., 1997; Merikangas et al., 1998;Tilley, 1987). Although such work has importantly directed scientific attentionto tobacco-anxiety relations at the broadest level, it is limited in demarcatingspecific rates of co-occurrence for (particular) disorders of interest, such aspanic. Thus, for the sake of explanatory specificity, we focus our summary oninvestigations that expressly distinguished panic psychopathology from otheranxiety disorders.

Comorbidity prevalence. The majority of studies have focused on document-ing rates of smoking among persons with panic psychopathology. The criteriafor smoking behavior has varied across investigations. Moreover, treatment-based recruitment strategies have been most commonly employed, perhapsmaking these data somewhat less generalizable to the overall smoking popula-tion. Nonetheless, among treatment-seeking adults, several studies havereported that current daily smoking among patients with panic psychopathology(either panic disorder or agoraphobia or both) ranged from 19% Panic(Baker-Morissette, Gulliver, Wiegel, & Barlow, 2004) to 57% (Himle, Thyer, &Fischer, 1988), with the vast majority of investigations falling between 30% to50% (Amering et al., 1999; Lopes et al., 2002; McCabe et al., 2004; Pohl,Yeragani, Balon, Lycaki, & McBride, 1992). These rates of daily smoking aretypically higher than comparison groups involving persons without psychiatricproblems and typically higher, or as high as, rates among persons with otheranxiety or mood disorders (McCabe et al., 2004). Thus, these data collectivelysuggest that smoking is a relatively common unhealthy behavior among treat-ment-seeking individuals with panic psychopathology.

Studies focused on non-treatment seekers are currently limited. Of theavailable work, one study focused on youth (Hayward, Killen, & Taylor,1989; 95 9th graders in public schools) and the other on college students(Valentiner, Mounts, & Deacon, 2004, n = 337). In both investigations, indi-viduals with panic attacks, but not necessarily panic disorder or agoraphobia,had higher rates of cigarette use on a ‘‘regular basis’’ (Hayward et al., 1989;Valentiner et al., 2004). For example, Hayward et al. (1989) reported that ofthose with a lifetime history of panic attacks, 77% had engaged in ‘‘experi-mental’’ or ‘‘regular’’ cigarette use compared with 48% of adolescents without alifetime history of panic attacks. These results, albeit highly limited in overallscope, generally parallel those of the treatment-oriented investigations notedearlier in terms of documenting elevated use prevalence among individuals withpanic problems.

Another set of investigations has utilized representative sampling methods toexplore the nature of tobacco use among those with panic psychopathology(Covey, Hughes, Glassman, Blazer, & George, 1994; Farrell et al., 2001; Lasseret al., 2000). In perhaps the most comprehensive and well-known of theseinvestigations, Lasser et al. (2000) examined smoking status according topsychiatric diagnoses using data from the National Comorbidity Survey (NCS),a nationally representative study that used structured clinical interviews to docu-ment mental illness (Kessler et al., 1994). Participants were 4,411 individuals aged


15 to 54 years. Among individuals diagnosed with panic attacks, panic disorder,

and agoraphobia in their lifetime, 38%, 35%, and 38% were current smokers,

respectively. These rates were significantly greater than rates of current smoking

among individuals without mental illness. By comparison, 36% of individuals

with a lifetime history of major depression and 49% of individuals with a lifetime

diagnosis of drug abuse or dependence were current smokers. Rates of lifetime

smoking among persons with a lifetime history of panic psychopathology

(i.e., panic attacks, panic disorder, or agoraphobia) ranged from 58% to 61%.

When diagnostic status in the past month was used as the grouping variable,

current rates of smoking were 46% among persons with panic attacks, 42%

among persons with panic disorder, and 48.1% among persons with agorapho-

bia. It is noteworthy that as number of mental diagnoses increased (ranging from

0 to 4 or more), the percentage of heavy (i.e., peak consumption exceeding

24 cigarettes a day) compared to relatively lighter (i.e., peak consumption less

than 24 cigarettes per day) smokers increased. Overall, these data, coupled with

the treatment-seeking data noted earlier, suggest that smoking occurs at relatively

higher rates among those with panic psychopathology compared to those with no

mental illness.Whereas the studies just reviewed focused on smoking among those with

panic psychopathology, other investigations have sought to evaluate rates of

panic psychopathology among smokers (Black, Zimmerman, & Coryell, 1999;

Breslau, Kilbey, & Andreski, 1991; Goodwin, Zvolensky, & Keyes, 2007;

Nelson & Wittchen, 1998). All of these investigations except that by Black

and colleagues (1999), which involved community-based recruitment, involved

some sort of representative sampling strategy. In contrast to the studies

reviewed earlier, these investigations attempt to understand panic within the

context of tobacco dependence and severity. Here, across studies, results

indicate that among those persons meeting criteria for more addictive use of

cigarettes (e.g., nicotine dependence), there is a greater prevalence of panic

psychopathology (Breslau et al., 1991). For example, Breslau and colleagues

(1991) found that 6.6% of persons meeting criteria for moderate dependence,

4.8% of those with mild dependence, and 2.4% of those with no dependence

had a lifetime diagnosis of panic disorder. Nelson andWittchen (1998) similarly

found that among participants endorsing a lifetime history of smoking (yes/no),

7.6% met lifetime diagnostic criteria for panic attacks, 2% for panic disorder,

and 4.4% for agoraphobia. These rates of panic psychopathology were

significantly greater than those reported among nonsmokers; 2.4% had a

panic attack history, 0.7% had panic disorder, and 1.6% had agoraphobia.

Smokers with a lifetime nicotine dependence diagnosis compared to smokers

without such a diagnosis evidenced greater rates of panic attacks (11.3% versus

4.0%), panic disorder (2.2% versus 1.8%), and agoraphobia (6.4% versus

2.5%). It should be noted that a similar, albeit not uniform, pattern of findings

was apparent for individuals with other psychiatric disorders (e.g., alcohol

dependence, drug dependence).

Panic and Tobacco 9

In the only study to focus on cigarette and smokeless tobacco use, Goodwin

and colleagues (2007) found that among a representative sample from the U.S.,

rates of past-year panic attacks (with or without agoraphobia) were greatest

among smokers with nicotine dependence (6.7%), followed by cigarette use

with no dependence (2.2%), both of which were greater than among those with

no past year cigarette or smokeless tobacco use (1.5%). Being dependent on

smokeless tobacco (1.9%) was largely comparable to no past year tobacco use,

both of which were greater than smokeless tobacco use without nicotine

dependence (0.6%).Overall, the extant literature suggests that heavier rates of cigarette use

(greater degrees of dependence) are associated with a greater rate of comorbid-

ity with panic psychopathology. Although limited, this pattern of findings does

not yet seem to be apparent for smokeless tobacco use, suggesting that factors

related to the ‘‘mode of administration’’ may be an important domain to further

understand in tobacco-panic linkages.Future directions. Though there are many avenues for future inquiry into the

nature of tobacco-panic comorbidity, here we highlight a few domains of

primary importance based upon the gaps in the existing literature. Before

specific recommendations are made, it is striking to point out that, to the best

of our knowledge, none of the past work focused on comorbidity issues has

been a priori oriented on tobacco-panic relations. Thus, it is, perhaps, not

surprising that some of the assessment approaches used in past work may not

be fully comprehensive or geared towards maximizing information about the

nature of the co-occurrence of these specific behavioral problems. As such, a

first-step in improving research in this domain would be to design evaluations

specifically focused on better understanding tobacco-panic comorbidity.

Beyond this general issue, there are at least three specific points within this

domain that would be particularly useful to address.First, only one study has provided data on smokeless tobacco and panic

comorbidity. Thus, to foster further empirical knowledge in this domain, it is

necessary to complete investigations wherein multiple forms of tobacco use are

assessed to provide information on both cigarette use and smokeless tobacco.

Aside from providing much needed data on smokeless tobacco and

psychopathology, this type of work would help to define the parameters of

tobacco-panic relations more generally. In this same context, it would be

advisable to clarify the extent to which the observed co-occurrence rates

between tobacco use and panic psychopathology are similar to, or different

from, other health behaviors (e.g., alcohol use, physical exercise). In general,

research suggests smoking is strongly positively related to alcohol and other

substance use and negatively related to exercise (Zvolensky & Bernstein, 2005).

This work is necessary because it would further explicate the degree to which

tobacco use is or is not unique to the co-occurrence of panic psychopathology.

As the present book illustrates, research in the health-anxiety linkage is only

now emerging.


Second, there are a number of issues central to the generalizability of the

reviewed investigations. The large majority of these studies have nottilized

representative samples and therefore selection biases may be operative. It also

is noteworthy that only one study (Hayward et al., 1989) focused exclusively on

youth. Therefore, it is not possible to generalize the present tobacco-panic

relations to other segments of the lifespan (e.g., adolescents) or the various

stages of tobacco use (e.g., initiation, maintenance) that would presumably be

apparent across different age ranges. Additionally, there are very limited data

on tobacco-panic psychopathology linkages from a cross-national perspective.

As factors that govern tobacco use may vary across communities and cultures,

and in conjunction with the world-wide public health impact of both smoking

and panic problems, it is important to extend work in this area to more diverse

global populations.Finally, existing work has largely utilized limited assessments of smoking

and panic psychopathology. Due to the focus on the ‘‘presence or absence of

daily smoking,’’ little is known about the nature or topography of smoking

behavior in terms of its association with panic psychopathology (e.g., age

of onset, age of daily use, amount used when smoking the heaviest). This

work would be improved by broadening smoking assessments to include a

more detailed account of smoking history. It also may be productive to

incorporate a multidimensional approach that takes into consideration

theoretically-relevant motivational processes underlying cigarette use

(Piper et al., 2004). This type of approach could be particularly valuable when

examining linkages between smoking and panic psychopathology, whereby

motivation to smoke to avoid negative affect might be a formative psychologi-

cal process (Zvolensky, Schmidt, Antony et al., 2005).

Nature of the Associations Between Tobacco Use

and Panic Psychopathology

Developmental course. As a basis for understanding the nature of the relations

between tobacco use and panic psychopathology, it is important to first clarify

their developmental course. Representative data on the age of onset of panic

attacks and smoking provide a means to evaluate temporal sequence.

Research suggests that the onset of daily smoking typically occurs between

the ages of 15 and 20 and rarely after age 25 (Breslau, Johnson, & Hiripi, 2001).

For example, the CDC reports that in the United States, approximately 3,900

adolescents between the ages of 12 and 17 years initiate cigarette smoking each

day (CDC, 2004), and an additional 1,500 become daily cigarette smokers each

day (Substance Abuse and Mental Health Services Administration, 2005).

Early studies of smokeless tobacco use indicate a mean age of onset between

10 and 12 years (i.e., Gottlieb, Pope, Rickert, & Hardin, 1993).

Panic and Tobacco 11

Studies examining the typical age of onset for panic attacks also suggest that

such problems often first occur in adolescence. For example, Goodwin and

Gotlib (2004) reported that the mean age of panic attack onset was 13.4 years

(n=1,285; age range 9–17). Other studies based on community or school

samples have found similar results, with the modal age of onset of panic attacks

being 12 years old (Hayward et al., 1992; Warren & Zgourides, 1988), and

clinical samples report a slightly younger age of panic attack onset (Alessi &

Magen, 1988; Black & Robbins, 1990). These data suggest that, across studies,

panic attack onset tends to first occur between the ages of 12–13 years. One

important interpretative caveat to these investigations is that they focus

exclusively on youth and expressly do not sample from a larger age range.

Thus, it is possible that the ‘‘average’’ age of onset of panic attacks may be

different if the sampling strategy incorporated adults.Based on available indirect data from smoking and panic attack age of onset

studies, it appears that in many instances the typical age of onset of panic

attacks precedes the typical age of onset of smoking. However, retrospective

reports of smokers with ‘‘active panic problems’’ are not entirely consistent with

this perspective. For example, Amering and colleagues (1999) examined 102

consecutive panic disorder patients with or without agoraphobia attending an

academic treatment clinic in Austria. Participants were diagnosed using the

SCID-III-R (First, Spitzer, Gibbon, & Williams, 1995) and interviewed about

their smoking status. Individuals presenting with ‘‘severe somatic illness’’

and comorbid depression and other psychiatric illnesses were excluded

from the study. Amering and colleagues (1999) reported that the onset of

smoking preceded the onset of panic disorder (cf. panic attacks) by 12.3 years

(SD=9.4) in a community sample of individuals with the condition (n=102).

Bernstein, Zvolensky, Schmidt, and Sachs-Ericsson (2007) directly evaluated

onset patterns among 4,409 adults (Mage = 33.1, SD= 10.7, females = 2,221)

from the NCS (Kessler et al., 1997). Results indicated that among cases with a

lifetime history of comorbid daily smoking and panic attacks (n = 167), the

onset of daily smoking (M=16.0 years, SD=3.0) preceded the onset of panic

attacks (M = 27.8 years, SD = 7.6) in the majority, but not among all, of the

comorbid cases (63.7%, n=106). A relatively large minority of comorbid cases

(33%; n=55) reported that panic attacks (M=11.4 years, SD=5.2) preceded

the onset of daily smoking (M= 18.2 years, SD= 4.7). The concurrent (same

year) onset of these two problems appeared rarely (3.3%, n = 6). Also, as the

pattern of ages of onset above illustrate, daily smoking demonstrated a rela-

tively consistent mean age of onset (mid to late adolescence) across comorbid

sub-samples and the uni-morbid sub-sample of smokers (age 18.5 years). In

contrast, the mean ages of onset of panic attacks differed markedly between the

comorbid sub-samples and the uni-morbid sub-sample of nonsmokers with

panic attacks (age 20.3 years). Overall, while data focused expressly on devel-

opmental course and smoking-panic psychopathology is limited, extant studies

suggest that the majority of cases may involve smoking preceding panic attacks.


Future directions. Again, it is important to highlight that work in thisdomain, although important for illuminating basic facets of tobacco-panicrelations, is currently limited. First, to the best of our knowledge, no studieshave examined the developmental course for smokeless tobacco. Future pro-spective work is therefore needed to expand this area of knowledge. Second,future study is needed to more carefully examine the putative smoking-to-panicand panic-to-smoking developmental courses. Third, smoking often occurs in acontext of other substance use patterns. Thus, it may be useful to understandthe developmental relations of tobacco and panic within the larger develop-mental context of panic-substance use comorbidity. In this sense, understand-ing the relative degree of ‘‘specificity’’ of initial findings vis a vis other substanceuse patterns and problems as well as health behaviors (e.g., physical exercise)may be a fruitful next research step.

Current knowledge regarding tobacco use and panic psychopathology.Cross-sectional studies that have utilized interview and self-report methodshave uniformly indicated that smoking, compared to non-smoking, is asso-ciated with more panic-relevant symptoms and impairment among nonclinical(Zvolensky, Forsyth, Fuse, Feldner, & Leen-Feldner, 2002) and clinical(McCabe et al., 2004; Zvolensky, Eifert, Feldner, & Leen-Feldner, 2003)samples. For example, Zvolensky, Schmidt and McCreary (2003) found thattreatment-seeking smokers with panic disorder compared to nonsmokers withpanic disorder reported more severe and intense anxiety symptoms, greaterinterview-based overall severity ratings of panic symptoms, and more socialimpairment. In these investigations, effects did not vary by gender, age, or otherforms of substance use.Moreover, there is emerging evidence that these types ofeffects are relatively specific to panic disorder and psychopathology thatfrequently co-occurs with panic (e.g., posttraumatic stress disorder; Feldneret al., 2007). For example, Morissette and colleagues (2006) found that smokerswith anxiety disorders, as compared to their non-smoking counterparts,reported higher levels of anxiety sensitivity (i.e., fear of anxiety andbodily-related sensations;McNally, 2002), anxiety symptoms, and agoraphobicavoidance. However, this association was specific to panic disorder and notevident for any of the other studied anxiety disorders, which did not includeposttraumatic stress disorder (Morissette et al., 2006).

Laboratory studies, although less common, have yielded similar findings.Zvolensky and colleagues (2004), for example, employed a voluntary hyperven-tilation paradigm to examine associations between smoking and panic-relevantfearful responding to bodily sensations. Results indicated smokers with panicdisorder reported greater levels of anxiety than smokers without panic disorderat baseline, and also showed greater increases in anxiety during thepost-challenge assessment and recovery periods. Although smokers with, versuswithout, panic disorder did not differ on baseline or post-challenge anxiety,smokers with, compared to without, panic disorder, demonstrated sloweraffective recovery from the challenge. These results indicate that smokingcompared to not smoking is related to greater affective distress in response to


panic-relevant cues even among those with panic disorder. Less attention hasbeen focused on determining the relation between smoking rate and level ofaffective distress or impairment. Yet, a number of studies, some involvingprospective measurement (discussed in greater detail below), have found thatsmoking rate is related to greater degrees of panic-specific emotional symptoms(e.g., panic-relevant avoidance; Breslau&Klein, 1999; Goodwin, Lewinsohn, &Seeley, 2005; Johnson et al., 2000; McLeish, Zvolensky, & Bucossi, 2007;Zvolensky, Kotov, Antipova, & Schmidt, 2003). Thus, there is empiricalevidence that both smoking status and rate are related to increased risk forpanic-relevant emotional vulnerability.

Cross-sectional tests also have helped to clarify factors that may affect thesmoking-panic relation. In one study of epidemiologically-defined (i.e., repre-sentative) adult residents of Moscow (n = 95 daily smokers from a largersample of about 400 persons; Zvolensky, Kotov et al., 2003), anxiety sensitivitymoderated the effects of cigarettes smoked per day (m = 15) on level ofagoraphobic avoidance. This significant interaction accounted for approxi-mately 10%of unique variance after controlling for their respective main effectsand the theoretically-relevant factors of problematic alcohol use and negativeaffectivity. No interaction, however, was found for panic attacks, potentiallydue to the fact that assessment of this factor was restricted to the past (mostrecent) week to enhance the validity of panic reports (but probably truncatingvariability). Similar moderating effects have been evident for perceived healthamong young adult daily smokers (McLeish, Zvolensky, Bonn-Miller, &Bernstein, 2006), and for neuroticism among a representative sample of adultsmokers (Zvolensky, Sachs-Ericsson, Feldner, Schmidt, & Bowman, 2006).Overall, these findings suggest smokers are not a homogeneous group in regardto their risk for panic problems and that individual differences in anxietysensitivity (or other cognitive-affective factors like perceived health or neuroti-cism) may be key factors in accounting for such differences.

Moderating effects for anxiety sensitivity also have been evident inbetween-group tests involving smokers and nonsmokers. For example, thecombination of high levels of anxiety sensitivity and a positive current smokingstatus predicted panic symptoms and somatic complaints, but not depressivesymptoms in a biological challenge test (Leen-Feldner et al., 2007). Again, suchfindings suggest that anxiety sensitivity (and possibly other factors) may mod-erate the relation between smoking and prototypical panic psychopathologyvariables (panic attacks and somatic complaints) even after controlling forgender and negative affectivity. Moreover, these associations are specific topanic-relevant processes. In a re-analysis of the Russian epidemiological studyreported earlier, Zvolensky and colleagues extended this smoking and anxietysensitivity effect (Zvolensky, Kotov, Bonn-Miller, Schmidt, & Antipova, inpress). Here, anxiety sensitivity, again, moderated the association of smokingstatus with indices of anxiety symptoms; effects were evident after controllingfor the variance accounted for by alcohol use problems, environmental stress(past month), and gender.


Although cross-sectional data are informative in the study of tobacco-panicpsychopathology relations, their utility also is limited. Prospective studies offerunique insight into the nature of the observed relations over time, and byextension, the order or temporal sequence of the associations. Researchershave evaluated the association between smoking and risk of panicpsychopathology in a number of studies. Breslau and Klein (1999) tested theassociation between daily smoking and risk for panic attacks and panicdisorder. Participants were drawn from two separate epidemiologically-defineddata sets. Across both data sets, results indicated that there was a significantlifetime and prospective association between daily smoking and onset of panicattacks and panic disorder; daily smokers were almost 4 times more likely toexperience panic attacks and 13 times more likely to develop panic disorderafter controlling for major depression and gender. Additionally, amongindividuals who continued to smoke, compared to those who had quit, therewas a significantly increased risk for experiencing a panic attack and panicdisorder. Johnson and colleagues (2000) also found that anxiety disordersduring adolescence were not significantly related to smoking in youngadulthood. However, smoking in adolescence increased the risk for developingagoraphobia and panic disorder during early adulthood. These effects wereobserved above and beyond the variance accounted for by temperament, familyhistory of psychopathology, drug/alcohol use and other theoretically-relevantfactors. Specifically, adolescents who were heavy smokers were 15.6 timesmore likely to develop panic disorder in early adulthood than non-smokers.Interestingly, adolescents who smoked fewer than 20 cigarettes per day were notat elevated risk for the development of (later) anxiety disorders,potentially suggesting, once again, that heavier smoking levels impart greaterpanic-related risk.

In another prospective study recently completed in Germany, 2,500 partici-pants (ages 14–24 years at baseline) were evaluated over 4 years (Isensee,Wittchen, Stein, Hofler, & Lieb, 2003). Compared with all other levels ofsmoking, dependent regular smokers at baseline were significantly more likelyto develop panic attacks and panic disorder, and a similar pattern was observedfor agoraphobia. Similarly, Breslau, Novak, and Kessler (2004) evaluated dailysmoking and subsequent onset of psychiatric disorders. Results indicatedthat the onset of panic disorder (odds ratio = 2.6) and agoraphobia(odds ratio = 4.4) were associated with pre-existing daily smoking aftercontrolling for age, gender, ethnicity, and educational level. Additionally,after controlling for pre-existing psychiatric disorders and sociodemographiccharacteristics, current nicotine dependent smokers were significantly morelikely to have panic disorder compared to current non-dependent smokersand former smokers. Importantly, the likelihood of panic disorder andagoraphobia was significantly reduced as time since quitting increased; theseeffects were specific to these conditions and not other psychiatric disorders(e.g., major depressive disorder), suggesting quitting smoking likely decreasesthe risk of developing panic problems, an issue that is discussed in greater detail


later in the chapter. More recently, Goodwin and colleagues (2005) replicatedthe results of Breslau and Klein (1999), Johnson et al. (2000), and Isensee et al.(2003) by finding that daily smoking during adolescence was associated with anincreased risk for panic attacks and panic disorder in young adulthood.Moreover, the observed effects were no longer evident after controlling forparental smoking and anxiety disorder status, suggesting that these familyhistory characteristics may be formative in the linkages between smoking andpanic psychopathology.

Prospective tests examining moderating factors in the tobacco use-panicrelation are very limited. In the only study to date on this topic, McLeishand colleagues (2007) evaluated the moderating role of anxiety sensitivity inthe relation between smoking rate and panic vulnerability variables among acommunity-based sample of 125 daily smokers (60 females; Mage =26.02 years). Findings indicate that the interaction between anxiety sensitiv-ity and smoking rate significantly predicted concurrent agoraphobic avoid-ance (3.2% of unique variance) and change in levels of anticipatory anxietyabout bodily sensations during the 3-month follow-up period (4.7% uniquevariance). Smokers high in anxiety sensitivity who also smoked at greaterrates reported the highest levels of avoidance and greatest increase in antici-patory anxiety. These data, in accord with cross-sectional findings (Leen-Feldner et al., 2007; Zvolensky, Kotov et al., 2003), once again suggest thatanxiety sensitivity is an important individual difference factor that, whencoupled with higher rates of smoking, is associated with greater levels ofavoidance and anticipatory anxiety among daily smokers, both of whichcontribute to the development of panic psychopathology.

Overall, research co-addressing smoking and panic psychopathologysuggests that smoking can be considered a variable risk factor for panicproblems. Indeed, existing work provides evidence regarding relations withpanic problems based on cross-sectional and prospective studies, but it isnoteworthy that this work is rarely multi-method in its approach. To havemore confidence in smoking-panic psychopathology relations, the incorpora-tion of multi-method assessment protocols would be an important nextresearch step. Additionally, evidence from cross-sectional, and to a lesserextent, prospective studies indicates that fears of internal sensations (anxietysensitivity) and perhaps other ‘‘affect-amplifiers’’ (e.g., perceived health,neuroticism) may moderate smoking-panic processes.

Future directions. There is a rapidly developing empirical literature ontobacco-use and panic psychopathology relations. Such scientific interest inthis work underscores its public health relevance and potential clinical implica-tions (see Zvolensky, Bernstein, Yartz, McLeish, & Feldner, in press, for anexpanded discussion of treatment implications of tobacco-panic relations). Atthe same time, this literature remains relatively under-developed and there are anumber of key areas in need of future study.

First, as in the area of comorbidity prevalence studies reviewed earlier, thereis a dearth of data on smokeless tobacco-panic relations. Virtually no scientific


data exists on this important topic, making it a fertile area for future explora-tion. Second, available data suggest daily smoking tends to precede the onset ofpanic attacks in the majority of cases, although direct evaluations with panicdisorder and agoraphobia have not been completed. Given that smoking can bechanged via intervention (Abrams et al., 2003), there is evidence of its potentialmalleability, and hence, possible application to prevention programs for panicpsychopathology. Overall, then, evidence that changing cigarette smoking rateor smoking cessation will alter the future risk of panic psychopathology from apreventative standpoint is lacking. Thus, it is currently not clear if smokingrepresents a variable marker or a variable causal risk factor for panic psycho-pathology. To clarify this issue, it is important for future research to examinechanges in smoking prospectively following experimental manipulation(e.g., smoking cessation intervention; Zvolensky, Schmidt, Bernstein, &Keough, 2006). Third, research has yet to examine the possibility that sharedor common risk factors may further explain the development of comorbidtobacco use and panic. It is theoretically possible that certain biological,psychological, and social factors may partially underlie the etiology andmaintenance of these behavior problems. And finally, while there is a growingliterature on moderating factors, there has been little scientific attention tomediators of smoking-panic associations and therefore almost no empiricalknowledge exists pertaining to the putative causal mechanisms of interest.Intensifying the focus on mediators of smoking-panic relations is a clinically-relevant and timely task. Specifically, clarification of key mechanisms throughwhich smoking achieves its panicogenic effects will stimulate the developmentof targeted interventions focused on therapeutic processes, and help to establishsuch processes (e.g., emotional reactivity) as important in the etiology and/ormaintenance of panic-related problems. Only Breslau and Klein (1999)conducted exploratory analyses of possible mediators by evaluating therole of lung disease. Although medical illness is one useful process to betterunderstand, other factors such as perceived health, affect tolerance, varioustrajectories of emotional distress (e.g., delayed recovery), withdrawalsymptoms, and avoidance-oriented smoking patterns are all examples oftheoretically-relevant factors deserving of future study (see Zvolensky &Bernstein, 2005, for an expanded discussion).

Current knowledge regarding the relation between panic psychopathology.pre-morbid panic risk variables, and tobacco use. Although much of the mosthighly publicized work on smoking and panic psychopathology pertains to thepotential role of smoking in the onset or maintenance of panic problems, panicvulnerability characteristics, broadly encompassing both pre-morbid variablesand full-blown panic problems, may conversely impact smoking behavior(Zvolensky & Bernstein, 2005). Work in this domain has focused on empiricalevidence related to smoking cessation outcome, expression of withdrawalsymptoms, and motivational and cognitive processes related to smokingbehavior (e.g., outcome expectancies). A major strength of work in this domainis that study of smoking has involved measurement of various facets of


smoking-related problems and processes (i.e., motivational processes) asopposed to focusing more narrowly on rates of tobacco use. This multidimen-sional conceptualization and measurement of smoking behavior is central totheoretical and clinical advances relevant to tobacco-panic relations.Moreover,it provides Moreover, it provides a means by which to further examine the roleof panic-specific, and other anxiety factors (e.g., traumatic event exposure,negative affectivity), in smoking initiation (Bernstein, Zvolensky et al., 2007;Feldner et al., 2007; Isensee et al., 2003).

Research has shown that affective vulnerability factors like panic psycho-pathology may be related to problems in quitting smoking. Lasser et al. (2000),for example, reported quit rates (i.e., proportion of lifetime smokers who werenot current smokers) in relation to psychiatric diagnosis among a representativesample from the U.S. Using 1-month diagnostic status as a criterion point, thequit rate was 29% for persons with panic attacks, 32% for those with panicdisorder, and 23% for persons with agoraphobia. Individuals with both panicattacks and agoraphobia in the past month were significantly less successful inquitting smoking compared to individuals with no mental illness (42%). Coveyand colleagues (1994) reported conceptually similar findings that panic psycho-pathology may be related to poorer success in quitting, although it is not clearwhether this effect is more robust than the types of association(s) betweenpsychiatric disorders more generally and quit success. Other cross-sectionalfield and laboratory work using a community sample has found that dailysmokers with a history of panic attacks, but no axis I histories, reportedsignificantly shorter average quit attempt histories, measured in days, com-pared to smokers without panic (Zvolensky, Lejuez, Kahler, & Brown, 2004).Similar results have been observed in laboratory investigations (Zvolensky,Feldner, Eifert, & Brown, 2001). Although limited by cross-sectional design,and by extension, possible reporting errors (e.g., recall biases), these dataprovide evidence of a relation between panic psychopathology variables andproblems in quitting.

A related line of work has focused on anxiety sensitivity and success inquitting smoking. Anxiety sensitivity tends to be elevated among individualswho fear anxiety and arousal-related sensations such as panic disorder(Bernstein & Zvolensky, 2007). In the earliest study in this domain, Brown,Kahler, Zvolensky, Lejuez, and Ramsey (2001) examined a subset of data froma randomized controlled clinical trial comparing standard smoking cessationtreatment versus standard smoking cessation plus cognitive-behavioral treat-ment for depression in smokers with past major depressive disorder. In thisinvestigation, the association between anxiety sensitivity and relapse during theearly stages of a quit attempt (e.g., first week), when individuals are most apt toexperience symptoms of anxiety (Hughes, Higgins, & Hatsukami, 1990), wasexamined. Anxiety sensitivity was significantly associated with increased oddsof lapsing during the first week after quit day (odds ratio = 2.0). Subsequentwork has conceptually replicated and extended the results of Brown andcolleagues (2001). For example, Zvolensky, Bonn-Miller, Bernstein, and


Marshall (2006) found anxiety sensitivity was significantly associated with

increased risk of early smoking relapse among a community sample of daily

smokers; these effects were evident above and beyond smoking rate and nega-

tive affectivity. Suchwork has recently been extended to low-level smokers from

Mexico, adding cross-national empirical support (Zvolensky, Bernstein, et al.,

in press). Collectively, there is a growing amount of empirical evidence suggest-

ing that panic psychopathology or pre-morbid panic-relevant variables such as

elevated anxiety sensitivity is related to early relapse problems, and possibly,

lower rates of overall success in quitting. Here again, controlled, prospective

studies are an important next research step, as they would remove concerns that

observed effects to date are attributable to reporting biases.A closely related line of inquiry has suggested that anxiety sensitivity is

related to motivation to quit, barriers to quitting, and reasons for quitting.

For example, Zvolensky, Baker and colleagues (2004) found anxiety sensitivity

was related to higher levels of current motivation to quit smoking among adult

daily smokers (Mage = 20.4], Mcigarettes per day = 10.2); effects were not attri-

butable to other theoretically-relevant factors (e.g., gender, smoking rate;

Zvolensky, Baker et al., 2004). These findings may at first seem counterintuitive

in that it seems logical that individuals with high levels of anxiety sensitivity

would be less likely to express interest or motivation in quitting due to the

feared negative consequences related to quitting (e.g., withdrawal symptoms,

emotional dyscontrol). Yet, related work suggests that smokers who worry

about the negative health-related effects of smoking may engage in more quit-

ting behavior (Dijkstra & Brosschot, 2003). From this perspective, high anxiety

sensitivity smokers may be more apt to perceive a personal vulnerability to the

negative effects of smoking (e.g., health risks), and as such, express greater

motivation to quit (Zvolensky & Bernstein, 2005) despite their greater difficulty

in successfully doing so (Brown et al., 2001). In line with this reasoning,

Zvolensky, Vujanovic and colleagues (2007) more recently examined the rela-

tions between anxiety sensitivity and (1)motivation to quit smoking, (2) barriers

to smoking cessation, and (3) reasons for quitting smoking among 329

(160 females; Mage = 26.08 years, SD = 10.92) adult daily smokers. After

covarying for theoretically-relevant variables (negative affectivity, gender, axis

I psychopathology, non-clinical panic attack history, number of cigarettes

smoked per day, and current levels of alcohol consumption), anxiety sensitivity

was significantly incrementally related to level of motivation to quit smoking, as

well as perceived barriers to quitting smoking. Additionally, after accounting

for the variance explained by other theoretically relevant variables, anxiety

sensitivity was significantly associated with self control reasons for quitting

smoking (intrinsic factors) as well as immediate reinforcement and social influ-

ence reasons for quitting (extrinsic factors). These results provide empirical

evidence that anxiety sensitivity is uniquely related to level of motivation to quit

smoking, perceived barriers to quitting, and certain intrinsic and extrinsic

reasons for quitting.


Panic psychopathology or pre-morbid risk factors also appear to be relatedto severity of acute nicotine withdrawal. In an early study in this domain,Breslau, Kilbey, and Andreski (1992) found tobacco withdrawal symptoms ina sample of young adults were significantly elevated among smokers with ‘‘anyanxiety disorder’’ compared to individuals without a history of such disorders;however, specific anxiety diagnoses were not provided, rendering unclear thespecificity of such results to panic psychopathology per se. Zvolensky andcolleagues (2004) found that daily smokers with a history of panic attacksreported significantly more intense anxiety-related withdrawal symptoms(anxiety, restlessness, difficulty concentrating, and irritability) compared tosmokers without such a history; no differences were evident for the othertobacco withdrawal symptoms (e.g., increased appetite). In another study,Zvolensky, Baker et al. (2004) tested whether anxiety sensitivity predicted theintensity of withdrawal symptoms during the first week of daily smokers’ mostrecent quit attempt. Results indicated that anxiety sensitivity predicted theintensity of nicotine withdrawal symptoms during the first week of smokers’most recent quit attempt, and this effect was above and beyond varianceaccounted for by negative affectivity, panic attack history, gender, cigarettesper day, and age of smoking onset, accounting for 16% of unique variance inwithdrawal symptoms. This work is promising in suggesting panic-specificfactors are related to an enhanced reactivity to nicotine withdrawal symptoms.Experimental work, which is now underway in our laboratory, is necessary toprovide an additional degree of confidence in such conclusions.

Another facet of evidence in support of a panic-tobacco relation is apparentfrom motivational and outcome expectancy research. In regard to smoking-related motivational processes, there is a large empirical literature documentingthat smokers often attribute their smoking, at least in part, to its mood-regulatingfunctions and believe that smoking will reduce negative affect states (Parrott,1999). Due to their affective vulnerability, smokers with panic-relevant vulner-abilities (i.e., high anxiety sensitivity) may be particularly motivated to smoke toescape from emotional distress elicited by acute nicotine withdrawal ornon-withdrawal states (e.g., anticipatory anxiety; Zvolensky & Bernstein,2005). A number of cross-sectional studies support this theory. Specifically,studies have indicated that anxiety sensitivity is associated with coping-orientedsmoking motives among young adults with no history of psychopathology(Novak, Burgess, Clark, Zvolensky, & Brown, 2003; Stewart, Karp, Pihl, &Peterson, 1997; Zvolensky, Bonn-Miller et al., 2006), adolescents (Comeau,Stewart, & Loba, 2001), and individuals with a past history of major depression(Brown, Kahler et al., 2001). Zvolensky, Feldner, Leen-Feldner et al. (2004)report conceptually similar findings for relations between anxiety sensitivityand negative-reinforcement outcome expectancies for smoking. The Comeauet al. (2001) investigation, in particular, is noteworthy in that anxiety sensitivitymoderated the relation between trait anxiety (frequency of anxiety symptoms)and use of cigarettes to cope with affective distress, reporting a stronger relation-ship between anxiety and use of cigarettes to cope with negative emotions


Anxiety in Health Behaviors and Physical Illness · 2013-07-23 · A Practitioner’s Guide...

Documents

Transcript of Anxiety in Health Behaviors and Physical Illness · 2013-07-23 · A Practitioner’s Guide...