Electrocardiographic abnormalities in patients with cerebrovascular disease

THE JOURNAL OF

CLINICAL PSYCHIATRY

C E L E B R A T I N G O V E R 6 0 Y E A R S O F S E R V I C E T O P S Y C H I A T R I S T S

VOLUME 61 JUNE 2000 NUMBER 6

VO

L. 61, N

O. 6 JU

NE

2000T

HE

JOU

RN

AL

OF

CL

INIC

AL

PSY

CH

IAT

RY

ww

w.psychiatrist.com

PA

GE

S 399–461

© Copyright 2000 Physicians Postgraduate Press, Inc.

One personal copy may be printed

Partial Response, Nonresponse,and Relapse With Selective Serotonin

Reuptake Inhibitors in Major Depression:A Survey of Current “Next-Step” Practices

Steffany J. Fredman, Maurizio Fava, Allison S. Kienke,Candace N. White, Andrew A. Nierenberg,

and Jerrold F. Rosenbaum

CME ARTICLE

Memory Impairment inThose Who Attempted Suicideby Benzodiazepine Overdose

Bas Verwey, Paul Eling, Henk Wientjes,and Frans G. Zitman



THE JOURNAL OF

CLINICAL PSYCHIATRY

LETTERS TO THE EDITOR


The Journal of Clinical Psychiatry(ISSN 0160-6689) is publishedmonthly by Physicians PostgraduatePress, Inc. Address correspondenceto P.O. Box 752870, Memphis, TN38175-2870; phone inquiries to(901) 751-3800. Subscription rates:

domestic, $77.00; foreign, $132.00; student, $55.00.Single copies of the Journal or supplements are$10.00 (including postage and handling);prepayment is requested.

Periodical postage paid at Memphis, TN, andadditional mailing offices. POSTMASTER:Send address changes to Circulation Department,Physicians Postgraduate Press, Inc., P.O. Box752870, Memphis, TN 38175-2870.

The Journal of Clinical Psychiatry is indexed inIndex Medicus, Excerpta Medica, PsychologicalAbstracts, Current Contents, Science CitationIndex, Hospital Literature Index, BiologicalAbstracts, Cumulative Index of Allied Healthand Nursing, International Nursing, IndustrialMedicine, Chemical Abstracts, Adolescent MentalHealth Abstracts, and Alcohol and Alcohol ScienceProblems Database.

The Journal of Clinical Psychiatry is owned andpublished by Physicians Postgraduate Press, Inc.,6555 Quince Road, Suite 501, Memphis, TN 38119.Physicians Postgraduate Press, Inc., does not holditself responsible for statements made bycontributors in the Journal.

The Journal of Clinical Psychiatry requiresthe express transfer of copyright to PhysiciansPostgraduate Press, Inc., so that the author(s) andPhysicians Postgraduate Press, Inc., are protectedfrom misuse of copyrighted material. Allsubmissions must be accompanied by acopyright transfer signed by all authors.

A $3.00 fee per article for photocopying materialappearing in the Journal is payable to PhysiciansPostgraduate Press. This fee is waived in lieu of aspecific charge to be determined on a case-by-casebasis for classroom and library use; it is theresponsibility of the institution or individualto obtain permission for such use.

All advertising accepted conforms to high medicalstandards. Acceptance does not imply endorsementby Physicians Postgraduate Press, Inc.

Visit our Web site at www.psychiatrist.com.

© Copyright 2000 Physicians Postgraduate Press,Inc. All rights reserved.

BOARD OF EDITORS

Editor in Chief ............ Alan J. Gelenberg, M.D.Deputy Editor ............. Eric M. Reiman, M.D.

Deputy Editor ............. Michael H. Ebert, M.D.

PUBLICATION STAFFPublisher ............. John S. Shelton, Ph.D.

Publisher Emeritus .......... Irving SheltonAssistant Publisher ......... Judy BeachAdvertising Director ....... J. ShermanManaging Editor ............. Meg M. WatersSenior Editor ................... Kathryn HedgesAssistant Editor ............... David M. TwomblyAssistant Editor ............... Jennifer L. GentrySenior Proofreader .......... Kathy E. ShassereProofreader ...................... David SpicerProduction Assistant ....... Susan K. BeatonCirculation Manager ....... Sharon Landers

Publications Manager .............. C. J. FirestonePublications Coordinator ........ Anna KwapienGraphic Arts Manager ............. Faye M. GreenAsst. Graphic Arts Manager ... Erik AdamecGraphics Assistant ................... F. Grant Whittle, Ph.D.Graphics Assistant ................... Andrea E. BlevinsCME Director ........................... Jane B. EcksteinCME Coordinator .................... Diane BrunnerProject Manager ....................... Becky Brown DerychProject Associate ..................... Patricia P. Duckworth, M.D.Project Assistant ...................... John R. Reed

Geoffrey L. Ahern, M.D., Ph.D.Hagop S. Akiskal, M.D.Judith V. Becker, Ph.D.Joseph Biederman, M.D.Jonathan F. Borus, M.D.Rubin Bressler, M.D.Andrew W. Brotman, M.D.James L. Claghorn, M.D.Paula J. Clayton, M.D.Jonathan O. Cole, M.D.Pedro L. Delgado, M.D.William E. Falk, M.D.Ellen Frank, Ph.D.Jack M. Gorman, M.D.John H. Greist, M.D.George T. Grossberg, M.D.Bernard I. Grosser, M.D.John G. Gunderson, M.D.David B. Herzog, M.D.Robert M.A. Hirschfeld, M.D.James I. Hudson, M.D.James W. Jefferson, M.D.Dilip V. Jeste, M.D.John M. Kane, M.D.T. Byram Karasu, M.D.Paul E. Keck, Jr., M.D.Samuel J. Keith, M.DMartin B. Keller, M.D.Donald F. Klein, M.D.Philip W. Lavori, Ph.D.R. Bruce Lydiard, M.D., Ph.D.Stephen R. Marder, M.D.

EDITORIAL BOARD

Susan L. McElroy, M.D.Herbert Y. Meltzer, M.D.Roger E. Meyer, M.D.Mark J. Mills, M.D., J.D.Erwin B. Montgomery, Jr., M.D.George B. Murray, M.D.Henry A. Nasrallah, M.D.Mark H. Pollack, M.D.Harrison G. Pope, Jr., M.D.Michael K. Popkin, M.D.Robert M. Post, M.D.Robert F. Prien, Ph.D.Frederic M. Quitkin, M.D.John C. Racy, M.D.Judith L. Rapoport, M.D.Steven A. Rasmussen, M.D.Elliott Richelson, M.D.Steven P. Roose, M.D.Patricia I. Rosebush, M.D., F.R.C.P.C.Jerrold F. Rosenbaum, M.D.A. John Rush, Jr., M.D.José M. Santiago, M.D.Nina R. Schooler, Ph.D.Mogens Schou, M.D.George M. Simpson, M.D.Gary W. Small, M.D.Susan E. Swedo, M.D.John A. Talbott, M.D.Michael E. Thase, M.D.Richard D. Weiner, M.D., Ph.D.Roger D. Weiss, M.D.

CME ADVISORY BOARDMichael H. Ebert, M.D., ChairBenjamin L. Beatus, Jr., M.D.John M. Downs, M.D.Alan J. Gelenberg, M.D.Radwan F. Haykal, M.D.

Paul King, M.D.Joseph A. Lieberman, III, M.D., M.P.H.Frank W. Ling, M.D.J. Sloan Manning, M.D.Richard C. Shelton, M.D.




THE JOURNAL OF

CLINICAL PSYCHIATRY

LETTERS TO THE EDITORLETTERS TO THE EDITOR

BRAINSTORMS

401 The 7 Habits of Highly Effective Psychopharmacologists,Part 3: Sharpen the Saw With Selective Choices ofContinuing Medical Education Programs.

Stephen M. Stahl

ORIGINAL ARTICLES

403 Partial Response, Nonresponse, and Relapse WithSelective Serotonin Reuptake Inhibitors in MajorDepression: A Survey of Current “Next-Step” Practices.Steffany J. Fredman, Maurizio Fava, Allison S. Kienke,Candace N. White, Andrew A. Nierenberg, andJerrold F. Rosenbaum

409 Prior Benzodiazepine Exposure andBenzodiazepine Treatment Outcome.

Karl Rickels and Ellen W. Freeman

414 Low-Dose Amphetamine Salts and AdultAttention-Deficit/Hyperactivity Disorder.Joseph P. Horrigan and L. Jarrett Barnhill

418 The Brief Psychiatric Rating Scale as an AcuteInpatient Outcome Measurement Tool: A Pilot Study.

Roy V. Varner, Y. Richard Chen, Alan C. Swann, andFrederick G. Moeller

422 A Comparison of Persons WithEarly- Versus Late-Onset Panic Attacks.

David A. Katerndahl and Melissa Talamantes

428 Childhood Abuse and Platelet Tritiated-ParoxetineBinding in Bulimia Nervosa: Implications ofBorderline Personality Disorder.Howard Steiger, Stéphanie Léonard, N. M. K. Ng Ying Kin,Cécile Ladouceur, Dhunraj Ramdoyal, and Simon N. Young

436 Effects of Dextroamphetamine on Depressionand Fatigue in Men With HIV: A Double-Blind,Placebo-Controlled Trial.

Glenn J. Wagner and Richard Rabkin

441 Electrocardiographic Abnormalitiesin Patients Treated With Clozapine.

Ung Gu Kang, Jun Soo Kwon, Yong Min Ahn, Sun Ju Chung,Jee Hyun Ha, Young Jin Koo, and Yong Sik Kim

CME ARTICLE 455

456 Memory Impairment in ThoseWho Attempted Suicide byBenzodiazepine Overdose.

Bas Verwey, Paul Eling, Henk Wientjes,and Frans G. Zitman

LETTERS TO THE EDITOR

447 WATCHERS: RecognizingGeneralized Anxiety Disorder.Mark J. Berber

447 Gold Versus Silver: The Issue ofFunctional Versus SymptomaticRecovery in Depression.

Subas Chandra Pradhan

• Reply by Stephen M. Stahl

448 SIADH With Multiple Antidepressantsin a Geriatric Patient.David M. Blass and Vincent E. Pearson

449 Free Drug Fraction VersusFree Drug Concentration.

Ronald Pies

• Reply by C. Lindsay DeVane andBruce G. Pollock

449 Zolpidem-Related Delirium:A Case Report.

Oliver Freudenreich and Matthew Menza

450 Bupropion-Tranylcypromine Combinationfor Treatment-Refractory Depression.Joseph M. Pierre and Michael J. Gitlin

451 Inappropriate Masturbationand Schizophrenia.

John O. Brooks III and Manoj V. Waikar

BOOK REVIEWS 452

INFORMATION FOR AUTHORS

see January 2000, pages 58–60 or our Web siteat www.psychiatrist.com/author.htm

http://www.psychiatrist.com/author.htm



B R A I N S T O R M SClinical Neuroscience Update

402 J Clin Psychiatry 61:6, June 2000

BRAINSTORMS is a monthly section of The Journalof Clinical Psychiatry aimed at providing updates ofnovel concepts emerging from the neurosciences thathave relevance to the practicing psychiatrist.

From the Clinical Neuroscience ResearchCenter in San Diego and the Department ofPsychiatry at the University of California San Diego.

Reprint requests to: Stephen M. Stahl, M.D.,Ph.D., Editor, BRAINSTORMS, 8899 University CenterLane, Suite 130, San Diego, CA 92122.

The 7 Habits of Highly Effective Psychopharmacologists, Part 3

Sharpen the Saw With Selective Choicesof Continuing Medical Education Programs

Stephen M. Stahl, M.D., Ph.D.

Issue: Highly effective practitioners of psychopharmacology recognize that theymust continually “sharpen their saw” as they cut through the diagnosing andprescribing decisions of daily practice. Choosing the right continuing medicaleducation programs is critical to attaining balanced self-renewal in this era ofrapidly expanding knowledge about neuroscience and new therapeutic options.

his feature is the third in aseries of articles1,2 show-ing how Steven Covey’shighly acclaimed prin-

ciples3 can be applied to develophabits that make one a highly effec-tive psychopharmacologist. Psycho-pharmacology is a rapidly expandingfield, with an ever-increasing num-ber of therapeutic options availablefor clinical practice.4

Sawing Through BiasThe information explosion in the

neurosciences and psychiatry de-mands that modern practitionerscontinually cut through this flow ofnew information in order to updatetheir diagnosing and prescribingskills. To do this, they must periodi-

cally “sharpen their saw” by develop-ing a strategy for mastering the use ofnew drugs in a setting where many ofthe educational opportunities are ei-ther biased or inefficient.

Continuing medical education(CME), a requirement for licensingof practitioners, has grown into amultibillion dollar industry fundedlargely by pharmaceutical companiesand regulated by the U.S. Food andDrug Administration (FDA) and theAccreditation Council for Continu-ing Medical Education of the Ameri-can Medical Association (ACCME).5

A great deal of debate has sur-rounded the potential commercialbias of some CME programs becauseindustry-sponsored events, travel,samples, luncheons, and gifts do infact result in the addition of newdrugs to formularies and an increasein prescription rates of the sponsor’sdrug.5,6 However, we can limit thepotentially irrational prescribing bysharpening our minds throughdiscriminatory thinking. The highlyeffective psychopharmacologist ex-ploits the plethora of educational op-

portunities among sponsors by se-lecting unbiased programs or goingback and forth between commer-cially sponsored programs. Thesestrategies work to eliminate our con-fusion and help us cut through theprograms being offered, like a filesharpening a dull saw, so that biasedprograms eventually cancel them-selves out.

Sawing Through Educational DesignMuch less attention has been paid

to the relative ineffectiveness of theeducational design of widely usedCME delivery methods. This is par-ticularly distressing because the ex-tremely busy practitioners of todaycannot make use of the efficientlearning methods that exist in otherfields of adult education when CMEactivities often do not incorporatethem. For example, if the goal is tochange physician’s diagnosing andprescribing practices, the most com-mon CME delivery methods, such asconferences, will have little directimpact because physicians face acord of new data equipped with a

T

401



B R A I N S T O R M SClinical Neuroscience Update

J Clin Psychiatry 61:6, June 2000 403

dull saw.7 More effective meth-ods do exist but are not widelyused. These include systematicpractice-based interventions,outreach visits, and a change inconferences to incorporate mul-timedia and advanced principlesof adult education such as repeti-tion and interaction.7

We busy practitioners cannotspend all of our time sharpeningthe saw, for we need to spendmost of our time using the saw.Thus, highly effective psycho-pharmacologists select CME ac-tivities according to the teachingmethods being offered as well asthe information being presented.

Reading is of course neces-sary, but it is one of the least ef-ficient methods of learning (10%retention of new information) com-pared with other learning methods.8,9

Retention rates must reach at least70% to saw completely through newmaterial and master it. The usualstrategy for a practitioner to masternew information is by repeated expo-sure until 70% retention is reached—by reading and then rereading, byattending lecture after lecture andconference after conference. How-ever, by selecting educationally effi-cient as well as unbiased CMEactivities, one can sharpen the sawmuch more efficiently and quicklyreturn to the work of cutting throughclinical practice decisions.

Retention from lecture without au-diovisuals is only 5%, the lowest rate.If a lecture is a traditional 59-minute,59-slide “data dump” with good au-diovisual support, there will be 20%retention, especially if the speaker re-alizes that only 7% of the messageshould be in words.10 Excellentspeakers, in fact, recognize that 38%

of the message is in pace and inflec-tion of delivery and 55% of the mes-sage in their body language. Theyexploit this fact to get the best reten-tion rates, which are still very lowfor a traditional lecture format.Significantly more retention of newinformation occurs if it is deliveredby demonstrations or discussiongroups,8,9 but these are almost neveremployed at CME conferences.

The very highest retention ratesoccur with “practice by doing”(75%) and “immediate use of learn-ing” (90%) methods. These are beingincorporated into many adult educa-tion programs outside of medicine,especially with the use of multi-media technologies and interactiveaudience-response keypads. By pre-senting information to visual learnersthrough multimedia animations, evi-dence-based learning can be greatlyenhanced. Furthermore, we can prac-tice by doing and have immediateuse of learning if video vignettes are

used for case-based learning, en-abling us to first make diagnos-ing or prescribing decisionsduring the CME course. This canbe accomplished by executingprescribing decisions via keypadsduring the course and then apply-ing this new knowledge soon af-ter in our clinical practice.

SummaryThe highly effective psycho-

pharmacologist will develop thehabit of “sharpening the saw” byclever selection of unbiased andefficient CME programs that in-corporate the most thoughtful ap-plications of the principles ofadult education to enhance reten-tion rates after a single exposure.

REFERENCES

1. Stahl SM. The 7 habits of highly effective psy-chopharmacologists: overview [BRAINSTORMS].J Clin Psychiatry 2000;61:242–243

2. Stahl SM. The 7 habits of highly effective psy-chopharmacologists, part 2: begin with the endin mind [BRAINSTORMS]. J Clin Psychiatry 2000;61:327–328

3. Covey SR. The 7 Habits of Highly EffectivePeople: Powerful Lessons in Personal Change.New York, NY: Simon & Schuster; 1990

4. Stahl SM. Essential Psychopharmacology. 2nded. New York, NY: Cambridge UniversityPress; 2000

5. Wazana A. Physicians and the pharmaceuticalindustry: is a gift ever just a gift? JAMA 2000;283:373–380

6. LaRossa J. The bias among us. TEN: The Eco-nomics of Neuroscience 2000;2:17

7. Davis DA, Thomson MA, Oxman AD, et al.Changing physician performance: a systematicreview of the effect of continuing medical edu-cation strategies [see comments]. JAMA 1995;274:700–705

8. Brookfield SD. Understanding and FacilitatingAdult Learning. San Francisco, Calif: Jossey-Bass Press; 1986

9. Bligh DA. What’s the Use of Lectures? SanFrancisco, Calif: Jossey-Bass Press; 2000

10. Mehrabian A, Ferris SR. Inference of attitudesfrom nonverbal communication in two chan-nels. J Consult Psychol 1967;31:248–252

Take-Home Points♦ Principles of adult education can be

applied to continuing medical education(CME) to increase retention rates andfacilitate adaptation of new knowledgeinto clinical practice as updateddiagnosing and prescribing behavior.

♦ Reading and lectures are the leastefficient yet most widely used deliverymethods for continuing medicaleducation.

♦ The most efficient CME programsincorporate features of moderneducational design, emphasizingrepetition via animations, evidence-based learning, and case-based learningas well as interaction utilizingmultimedia and audience-responsekeypads.

402


Fredman et al.


One personal copy may be printedirst-line treatment of depression has shifted overthe past decade from the tricyclic antidepressants

Partial Response, Nonresponse, and RelapseWith Selective Serotonin Reuptake Inhibitors in

Major Depression: A Survey of Current “Next-Step” Practices

Steffany J. Fredman, B.A.; Maurizio Fava, M.D.;Allison S. Kienke, B.A.; Candace N. White, M.Ed.;

Andrew A. Nierenberg, M.D.; and Jerrold F. Rosenbaum, M.D.

Background: Many patients treated for majordepression require more than one antidepressanttrial to achieve or sustain response. However, theliterature provides few treatment algorithms oreffectiveness studies that empirically support“next-step” options available to clinicians.We conducted a survey of psychiatrists andother medical specialists who treat depressionto ascertain what clinicians actually do whenfaced with patients who suboptimally respondto an adequate course of selective serotoninreuptake inhibitor (SSRI) therapy.

Method: Attendees at a psychopharmacologycourse (N = 801) were queried about their topchoices for antidepressant-treatment nonre-sponders: a minimal responder after 4 weeksof adequate SSRI treatment, a partial responderafter 8 weeks of adequate SSRI therapy, anonresponder after 8 weeks of adequate SSRItherapy, and a relapser on long-term SSRI mainte-nance therapy. Choices included raising the dose,augmenting or combining with another agent,switching to a second SSRI, or switching toa non-SSRI agent.

Results: 432 (54%) of the surveys were re-turned. Raising the dose was the most frequentlyreported next-step strategy for a patient withminimal response after 4 weeks of adequateSSRI therapy, partial response after 8 weeks ofadequate SSRI therapy, and relapse on long-termSSRI therapy. Switching to a non-SSRI agent wasthe most frequently chosen option for non-responders to an adequate trial of SSRI therapy.

Conclusion: Our findings suggest that clini-cians select different next-step strategies whenpatients are nonresponders versus when patientsare partial responders or relapsers.

(J Clin Psychiatry 2000;61:403–408)

F(TCAs) to the selective serotonin reuptake inhibitors(SSRIs). Although antidepressant efficacy is comparableamong the different classes of drugs,1 the SSRIs appear tobe more effective than TCAs in clinical practice due totheir relatively greater safety and tolerability.2 Nonethe-less, 29% to 46% of depressed patients fail to respond fullyto antidepressant medication. Specifically, it has been sug-gested that 12% to 15% are partial responders and 19% to34% are nonresponders.3

The “next-step” strategies available to clinicians—rais-ing the dose, augmenting or combining with other agents,and switching classes of antidepressants or switching toanother antidepressant within the same class—may beemployed for partial responders, nonresponders, andrelapsers during SSRI treatment.4 Unfortunately, the lit-erature provides few “real world” effectiveness studiesof patients resistant to SSRIs.5 Although randomized clin-ical trials are important because they provide investigators

Received June 21, 1999; accepted Dec. 6, 1999. From the Mood andAnxiety Disorders Program (Mss. Fredman, Kienke, and White and Dr.Rosenbaum) and the Depression Clinical and Research Program (Drs.Fava and Nierenberg) at Massachusetts General Hospital, and HarvardMedical School (Drs. Fava, Nierenberg, and Rosenbaum), Boston, Mass.

Financial disclosure: Dr. Fava has received grant/research supportfrom Eli Lilly, SmithKline Beecham, Pfizer, Wyeth-Ayerst, Organon, Bristol-Myers Squibb, Pharmacia & Upjohn, Glaxo Wellcome, Knoll, Roche,Synthelabo, Lorex, Solvay, Sanofi, Lichtwer, Parke-Davis, Novartis,Pharmavite, Abbott, and Somerset; has received honoraria from Eli Lilly,SmithKline Beecham, Pfizer, Wyeth-Ayerst, Organon, Bristol-Myers Squibb,Pharmacia & Upjohn, Glaxo Wellcome, Solvay, Forest, Janssen, Lundbeck,Knoll, Parke-Davis, and Somerset. Dr. Nierenberg has received grant/research support from Bristol-Myers Squibb, Forest, Glaxo Wellcome, EliLilly, Lichtwer, Organon, Pfizer, Sanofi, and Wyeth-Ayerst; and has receivedhonoraria from Bristol-Myers Squibb, Forest, Glaxo Wellcome, Eli Lilly,Organon, Pfizer, and Wyeth-Ayerst. Dr. Rosenbaum has received grant/research support from Bristol-Myers Squibb, Eli Lilly, Pfizer, Sanofi, Forest,Organon, Lichtwer, Wyeth-Ayerst, Parke-Davis, Janssen, Pharmacia &Upjohn, Abbott, Merck, and Millenium Pharmaceuticals; is on the advisoryboard of Bristol-Myers Squibb, Eli Lilly, Forest, Organon, Wyeth-Ayerst,and Pharmacia & Upjohn; and has received speaker support from Bristol-Myers Squibb and Forest.

Reprint requests to: Jerrold F. Rosenbaum, M.D., Department ofPsychiatry, Massachusetts General Hospital, 15 Parkman St., WAC 812,Boston, MA 02114 (e-mail: [email protected]).

403

J Clin Psychiatry 61:6, June 2000

Next-Step Practices for Nonresponse in Depression

405



with methodological controls over potentially confound-ing factors and maximize internal validity, they typicallyrequire relatively rarefied research patient populations,such as those who are drug free and without significantpsychiatric comorbidity, and mainly have power onlyto examine one drug’s efficacy over placebo instead ofcomparing multiple alternate strategies simultaneously.

The design of research studies of clinical relevance topractitioners may be enhanced by knowing what strate-gies community-based clinicians actually employ in theirown practices. Surveying practitioners who treat depres-sion is one potentially informative means of gatheringthis information. Several surveys have been conductedpreviously. Nearly 10 years ago, Nierenberg6 queried 118Northeastern psychiatrists who were presented with acase vignette of a depressed patient who had failed to re-spond to 4 weeks of nortriptyline at 100 mg daily andwere asked what treatment they would use next. Lithiumaugmentation was the most popular strategy among psy-chiatrists in that study. One limitation of that survey isthat it was conducted prior to the widespread use ofSSRIs first-line for the treatment of depression; therefore,next-step strategies described in that study were mostlikely specific to TCA failure and may not reflect currentprescribing practices.

Shergill and Katona7 conducted a similar survey in theUnited Kingdom in which the clinical vignette was of apatient refractory to initial treatment with amitriptyline,150 mg daily. Unlike in the Nierenberg survey,6 the mostpopular next-step strategies were to raise the amitripty-line dose or switch to an SSRI. Mischoulon and col-leagues8 conducted a small study in which they asked 20expert psychopharmacologists in Boston, Mass., to ratethe perceived effectiveness of different augmentationstrategies when an SSRI failed to produce or sustain re-sponse among depressed patients. They found that bu-propion, methylphenidate, and dextroamphetamine wereperceived as the most effective agents to add to the treat-ment regimen. This sample, however, was limited to psy-chopharmacologists who practiced mainly in an aca-demic teaching hospital in the same city. Byrne andRothschild9 also surveyed psychiatrists in Massachusettsto ascertain their top choices in the treatment of break-through depressive symptoms among patients taking 20mg of fluoxetine, 100 mg of sertraline, 100 mg of nor-triptyline, or 40 mg of fluoxetine. The authors reportedthat for all drugs and doses, raising the dose was the mostcommonly chosen strategy among those surveyed. As inthe Nierenberg6 study, this study surveyed clinicians froma limited geographical region and posed only one clinicalvignette.

The goal of the present study was to extend the find-ings of earlier surveys by presenting several differentclinical vignettes of treatment-resistant depressed pa-tients to psychiatrists and other practitioners working in a

wide variety of practice settings (private practice, hospi-tal, health maintenance organization [HMO], academiccenters) across a broad geographical distribution in theUnited States.

METHOD

Attendees at an annual psychopharmacology coursesponsored by Harvard Medical School (N = 801) wereasked to fill out a questionnaire prior to the lectures ondepression and its treatment. Approximately 790 ques-tionnaires were distributed, and 432 (54% of 801 attend-ees) were returned. The questionnaire consisted of 4 clini-cal vignettes of depressed patients who did not achievefull response to treatment with an SSRI: a patient who hasminimal response after 4 weeks of adequate SSRI treat-ment, a patient who is partially responsive after 8 weeksof adequate SSRI treatment, a patient who is nonrespon-sive after 8 weeks of adequate SSRI treatment, and a pa-tient who relapses while taking long-term SSRI therapy(see Appendix 1 for list of questions). Adequate dose wasspecified as 20 mg/day of fluoxetine, 100 mg/day of ser-traline, or 20 mg/day of paroxetine.

Clinicians were asked to indicate and rank their top 3choices for next-step strategies in response to each of the4 vignettes. For the minimal responder vignette, thechoices were wait more time and observe the patient, raisethe dose, or add another agent. For the partial responder,nonresponder, and relapser vignettes, options includedraising the SSRI dose, augmenting, or switching agents.Clinicians were also able to fill in which agents theywould augment with or switch to, if that option wasamong their top choices.

The data were harvested and grouped in the followingmanner:

Medical specialty. Attendees identified themselves asworking in 1 of following medical specialties: psychiatry,internal medicine/primary care, other medical specialty,or nonmedical specialty (e.g., nurse, psychologist).

Practice setting. Practice setting was considered hos-pital based or non–hospital based. Hospital-based prac-tices referred to those in which clinicians worked inhospitals, academic teaching centers, community clinics,and mental health treatment centers. Non–hospital-basedpractices included those where clinicians worked exclu-sively in private practice (solo or group) or at an HMO.

Geographical region. The city and state where the cli-nicians practice were classified as belonging to 1 of 4 re-gions: New England, MidAtlantic/South/Southeast, West(Midwest, Northwest, Southwest, and West Coast), andOther (locations outside of the contiguous United States,i.e., Alaska, Hawaii, Canada, and Puerto Rico).

Next-step options. For the partial responder, nonre-sponder, and relapser vignettes, respondents’ first choicesfor next-step options were classified as 1 of the following:

404


Fredman et al.



raise the dose, augment/combine (add another agent thatmay or may not be an antidepressant), switch to anotherSSRI, or switch to a non-SSRI antidepressant. When theoption of choosing to switch to a non-SSRI agent was se-lected, the agents listed were classified as dual-actingagents (venlafaxine, mirtazapine, and clomipramine), bu-propion, nefazodone, a TCA, a monoamine oxidase in-hibitor (MAOI), or a non-SSRI agent (if the respondentwas nonspecific in regard to which non-SSRI agent tochoose). Some observations were dropped if the respon-dent did not discriminate among his or her first, second, orthird choices or did not complete a section of the survey.

Statistical comparisons across clinical vignettes. Inaddition to descriptive statistics, we conducted pairwisecomparisons using the McNemar chi-square analysis withBonferroni corrections for multiple comparisons. Statisti-cal significance was therefore set at p < .004.

RESULTS

Of the 801 individuals registered for the course, 681were physicians (630 psychiatrists, 33 internists/primarycare physicians/D.O.’s, and 18 other medical specialists).One hundred twenty of the attendees were not physicians,including 70 nurses, 20 clinical psychologists, and 30 othernonmedical specialists. Among the 432 attendees who com-pleted and returned the survey, 93% of those who indicatedtheir medical specialty in our sample identified themselvesas psychiatrists and in practice for a mean ± SD durationof 16.7 ± 10.6 years. Sixty-three percent of respondentsidentified themselves as men and 37% as women.

Table 1. First-Choice Next-Step Strategies After 8 Weeksa

PartialResponders Nonresponder Relapser(N = 412) (N = 392) (N = 384)

Treatment Option N % N % N %

Raise dose 338 82 104 27 306 80Augment/combine 56 14 49 12 35 9

2nd SSRI 0 0 0 0 1 0.3TCA 10 2 6 2 6 2Dual-acting agent 4 1 2 0.5 0 0Bupropion 15 4 13 3 12 3Non-SSRI/atypical

antidepressant/other antidepressant 1 0.2 3 0.8 1 0.3

Nonspecific agent 5 1 4 1 2 0.5Stimulant/

dopaminergic agent 0 0 1 0.3 0 0Lithium/

mood stabilizer/benzodiazepine 13 3 11 3 10 3

Thyroid 4 1 5 1 1 0.3Buspirone/pindolol 4 1 4 1 2 0.5

Switch 18 4 239 61 43 112nd SSRI 6 1 67 17 12 3Non-SSRI 12 3 172 44 31 8

TCA 4 1 20 5 7 2MAOI/RIMA 0 0 0 0 1 0.3Dual-acting agent 2 0.5 53 14 7 2Bupropion 0 0 46 12 4 1Non-antidepressant 0 0 2 0.5 1 0.3Nefazodone 0 0 6 2 0 0Unspecified

non-SSRI agent 6 1 45 11 11 3aAbbreviations: MAOI = monoamine oxidase inhibitor,RIMA = reversible inhibitor of monoamine oxidase, SSRI = selectiveserotonin reuptake inhibitor, TCA = tricyclic antidepressant. Total Nsare different and < 432 because some questionnaires were incompleteor uninterpretable with respect to the options chosen in response toeach vignette.

Minimal responder. The most commonly endorsedstrategy for a minimal responder after 4 weeks ofSSRI therapy was to raise the dose, which was se-lected by 80% of the respondents. The options ofwaiting more time and adding a second agent werechosen at rates of 16% and 4%, respectively.

Partial responder. The most commonly chosen first-choice strategy for a partial responder after 8weeks of adequate SSRI treatment was to raisethe SSRI dose (82%), followed (in order of de-creasing preference) by augmentation/combination(14%), switching to a non-SSRI agent (3%), andswitching to another SSRI (1%) (Figure 1). Themost popular augmentation/combination agentswere bupropion (4%), lithium/mood stabilizer/benzodiazepine (3%), and a TCA (2%) (Table 1).

Nonresponder. Switching to a non-SSRI agent was themost popular choice for the treatment of a non-responder after 8 weeks of adequate SSRI treatmentand was selected by 44%. Raising the dose was thenext most popular strategy (27%), followed byswitching to another SSRI (17%) and augmenting

100

90

80

70

60

50

40

30

20

10

0Raise Augment Switch Switch to

SSRIsDose Non-SSRI

Pre

fere

nce

(%)

**

† † ‡‡

Partial ResponderNonresponderRelapser

Figure 1. Next Step After 8 Weeks of Treatmenta

aAbbreviation: SSRI = selective serotonin reuptake inhibitor.*p < .004: Raise dose: partial vs. nonresponder; relapser vs.nonresponder.†p < .004: Switch to SSRI: partial vs. nonresponder; relapser vs.nonresponder.‡p < .004: Switch to non-SSRI: partial vs. nonresponder; relapser vs.nonresponder.

405



407



with another agent (12%) (see Figure 1). Amongthose switching to a non-SSRI agent, the medica-tions selected in order of decreasing preferencewere a dual-acting agent (14%), bupropion (12%),unspecified non-SSRI agent (11%), a TCA (5%),and nefazodone (2%). As in the partial respondervignette, the most frequently chosen augmentation/combination agents were bupropion (3%), lithium/mood stabilizer/benzodiazepine (3%), and a TCA(2%) (see Table 1).

Relapser. Raising the dose was the most commonlychosen first-choice strategy for the treatment ofpatients who relapse while on long-term SSRItreatment (80%). Augmentation was chosen by9%, switching to a non-SSRI agent by 8%, andswitching to another SSRI by 3% (see Figure 1).Similar to the partial responder and nonrespondervignettes, the favored augmentation/combinationagents for a relapser were bupropion (3%), lith-ium/mood stabilizer/benzodiazepine (3%), and aTCA (2%). The most popular agents to switch towere a second SSRI (3%), a non-SSRI agent (3%),and a dual-acting agent (2%) (see Table 1).

For each of the 4 clinical vignettes, rankings forfirst-choice next-step strategies remained the same aftercontrolling for gender, geographical region, and practicesetting (i.e., there were no significant differences in first-choice next-step treatments across gender, geographicalregions, and practice settings). Therefore, we are confi-dent that our results are not based on these biases amongrespondents in our sample.

Figure 1 shows our pairwise comparisons, which re-vealed statistically significant differences (p < .004, sig-nificant after Bonferroni correction) in the rates of choos-ing particular strategies depending on the patient vignettepresented. Differences were observed in the rates ofraising the dose, switching to an SSRI, and switching toa non-SSRI agent between the partial responder and non-responder vignettes as well as between the nonresponderand relapser vignettes. More specifically, raising the dosewas chosen more frequently in the case of a partial re-sponder than a nonresponder; similarly, raising the dosewas chosen more frequently in the case of a relapserthan a nonresponder. However, switching to a non-SSRIagent was chosen more often in the case of a nonre-sponder than either a partial responder or a relapser.A similar trend emerged with respect to switching toanother SSRI. No statistically significant differenceswere detected in the rates of augmentation/combinationstrategies in response to the 3 vignettes, nor were theresignificant differences between the partial responderand relapser vignettes with respect to the rates of choos-ing to raise the dose, switch to an SSRI, or switch to anon-SSRI agent.

DISCUSSION

On the basis of the findings of this survey, it appearsthat psychiatrists approach the treatment of depressed pa-tients differently, depending on the particular type of fail-ure to respond. When asked about the first-choice next-step options for a minimal responder after 4 weeks, apartial responder after 8 weeks, and a patient who relapseswhile on long-term medication treatment, 80% or more ofrespondents indicated that their first choice would be toraise the SSRI dose. This was in contrast to the vignette inwhich the patient is a nonresponder after 8 weeks of treat-ment. In response to this vignette, clinicians indicated thatthe most popular next-step would be to switch to a non-SSRI agent. Raising the SSRI dose for a nonresponderwas selected by only 27% of those surveyed, substantiallyless than the rates of choosing this option for the other 3vignettes.

One caveat in interpreting these results is that thesedata represent the perceptions clinicians have of their ownprescribing practices, and they are not necessarily a truereflection of their actual behavior. Because we relied onclinician self-report in response to hypothetical clinicalvignettes, we cannot absolutely infer that the cliniciansactually do what they say they would. Additionally, only54% of the course attendees responded to our survey;therefore, we are unable to say with certainty that the re-sponses we received generalize to all clinicians at thecourse. In the absence of sociodemographic data on thecourse participants who did not complete the survey, wecannot rule out the possibility that a significant selectionbias occurred. Nonetheless, the response rate of 54% iscomparable with or better than the rates for other largesurveys of clinicians’ approaches to treatment-resistantdepression, which reported rates of 63%,7 56%,9 and24.5%.10 Lastly, the course attendees, from whom oursample respondents were derived, were clinicians whowere particularly interested and motivated to learn aboutpsychopharmacology. Having devoted 3 days and a mini-mum (tuition only) of $650 to attend the course, these cli-nicians may not be representative of community practitio-ners in terms of their knowledge of psychopharmacologyor the extent to which they use pharmacologic agents inthe treatment of depression.

These limitations notwithstanding, there are severalpossible implications of these findings. First, this studyunderscores the importance for research reports on treat-ment resistance to clearly specify the type of suboptimalresponse on which results are based. Depending on thetype and level of response of a particular patient, clini-cians appear to choose different therapeutic strategies.Given that physicians may employ certain treatment op-tions based on findings from studies reported in the litera-ture and that research findings may be translated intoclinical practice, there is a clear need for operationalized

406


Fredman et al.



definitions of treatment resistance. In particular, it isimportant to distinguish between partial response andnonresponse.

Second, our findings suggest that antidepressant pre-scribing trends change over time, perhaps influenced byresults from clinical research and expert practitioner rec-ommendations. The results from our study are somewhatdifferent from those of Nierenberg6 with respect to treat-ing antidepressant nonresponders. For instance, the mostpopular strategy in the Nierenberg6 study was to add lith-ium to nortriptyline (34%), whereas in our study, augment-ing an SSRI with lithium was chosen by fewer than 3%.There also appears to be an increased trend in the use ofbupropion, as reflected by the fact that only 1 person inthe Nierenberg study6 (0.8%) chose to switch to bupropion,whereas this option was chosen by 12% of our sample, andaugmentation/combination with bupropion was chosen by3% of our sample. In addition, bupropion was consistentlythe first-choice augmentation/combination agent for par-tial responders, nonresponders, and relapsers.

Third, it appears that clinicians’ use of differentialnext-step strategies is actually empirically supported inthe literature. For minimal responders and partial re-sponders, the vast majority of respondents in our surveyindicated that they would raise the dose. In a controlledstudy, Fava and colleagues11 found that among partial re-sponders to fluoxetine, 20 mg/day, raising the dose to 40to 60 mg/day was significantly more effective than addingeither lithium or desipramine.

Consistent with psychiatrists’ responses reported byByrne and Rothschild9 in a survey of psychiatrists aboutthe treatment of breakthrough depressive symptoms dur-ing maintenance treatment, most clinicians in our surveyreported that they would raise the dose if the patient ex-perienced a relapse or recurrence while on longer-termantidepressant therapy. A trial by Fava and colleagues12

provides preliminary support for this approach as well.In an open study of 18 patients who relapsed with fluoxe-tine, 20 mg/day, the authors observed that 83% of patientsresponded when the fluoxetine dose was raised to 40mg/day.

Moreover, duration of treatment appears to be associ-ated with a clear contrast in the choice of next-step treat-ments for minimal and nonresponders. An overwhelmingmajority (80%) of respondents chose to raise the dose forminimal responders after 4 weeks of treatment, while only27% selected the same treatment option for nonre-sponders after 8 weeks of treatment. In addition, amongclinicians surveyed in our sample, the most popularchoice for nonresponders after 8 weeks of SSRI treatmentwas to switch to a non-SSRI agent (44%). This is consis-

tent with empirical findings that suggest the usefulness ofthis approach in patients resistant to SSRIs.13–15

The results of our survey may be of use in designingadditional controlled clinical research trials that empiri-cally compare next-step strategies commonly used in thepharmacologic treatment of depression. Our findings sug-gest that it may be useful to conduct head-to-head studiesthat directly compare raising the dose with augmentationand combination as well as with switching to a differentagent in all 3 subpopulations of depressed patients.

Drug names: amitriptyline (Elavil and others), bupropion (Wellbutrin),buspirone (BuSpar), clomipramine (Anafranil and others), desipramine(Norpramin and others), dextroamphetamine (Dexedrine and others),fluoxetine (Prozac), methylphenidate (Ritalin and others), mirtazapine(Remeron), nefazodone (Serzone), nortriptyline (Pamelor and others),paroxetine (Paxil), sertraline (Zoloft), venlafaxine (Effexor).

REFERENCES

1. Steffens DC, Krishnan KR, Helms MJ. Are SSRIs better than TCAs? com-parison of SSRIs and TCAs: a meta-analysis. Depress Anxiety 1997;6:10–18

2. Simon GE, VonKorff M, Heiligenstein JH, et al. Initial antidepressantchoice in primary care: effectiveness and cost of fluoxetine vs tricyclic an-tidepressants. JAMA 1996;276:1301–1302

3. Fava M, Davidson KG. Definition and epidemiology of treatment-resistantdepression. Pyschiatr Clin North Am 1996;19:179–200

4. Fava M, Kaji J, Davidson K. Pharmacologic strategies for treatment-resistant major depression. In: Pollack MH, Otto MW, Rosenbaum JF, eds.Challenges in Clinical Practice: Pharmacologic and Psychosocial Strate-gies. New York, NY: Guilford Press; 1996:3–30

5. Crismon ML, Trivedi M, Pigott TA, et al. The Texas Medication AlgorithmProject: report of the Texas Consensus Conference Panel on MedicationTreatment of Major Depressive Disorder. J Clin Psychiatry 1999;60:142–156

6. Nierenberg AA. Treatment choice after one antidepressant fails: a surveyof Northeastern psychiatrists. J Clin Psychiatry 1991;52:383–385

7. Shergill SS, Katona CLE. Pharmacological choices after one antidepres-sant fails: a survey of UK psychiatrists. J Affect Disord 1997;43:19–25

8. Mischoulon D, Fava M, Rosenbaum JF. Strategies for augmentation ofSSRI treatment: a survey of an academic psychopharmacology practice.Harvard Rev Psychiatry 1999;6:322–326

9. Byrne S, Rothschild AJ. Psychiatrists’ responses to failure of maintenancetherapy with antidepressants. Psychiatr Serv 1997;48:835–837

10. Chaimowitz GA, Links PS, Padgett RW, et al. Treatment-resistant depres-sion: a survey of practice habits of Canadian psychiatrists. Can J Psychi-atry 1991;36:353–356

11. Fava M, Rosenbaum JF, McGrath PJ, et al. Lithium and tricyclic augmen-tation of fluoxetine treatment for resistant major depression: a double-blind, controlled study. Am J Psychiatry 1994;151:1372–1374

12. Fava M, Rappe SM, Pava JA, et al. Relapse in patients on long-term fluox-etine treatment: response to increased fluoxetine dose. J Clin Psychiatry1995;56:52–55

13. Fava M, Dunner DL, Griest JH, et al. An open-label study with mirtazapinein depressed patients who are SSRI treatment failures. In: New ResearchProgram and Abstracts of the 152nd Annual Meeting of the American Psy-chiatric Association; May 19, 1999; Washington, DC. Abstract NR431:186

14. Nierenberg AA, Feighner JP, Rudolph R, et al. Venlafaxine for treatment-resistant unipolar depression. J Clin Psychopharmacol 1994;14:419–423

15. Thase ME, Zajecka J, Kornstein SG, et al. Nefazodone treatment of pa-tients with poor response to SSRIs. Presented at the 37th annual meeting ofthe American College of Neuropsychopharmacology (ACNP); Dec 14–18,1998; Los Croabas, Puerto Rico

407



409



Appendix 1: Enhancing Response to Treatment

1. You have a patient on an adequate dose of an SSRI (20 mgProzac, 20 mg Paxil, or 100 mg Zoloft) for 4 weeks, but thepatient has only shown minimal improvement. Please rank thefollowing strategies for usual next step treatments in order ofpreference (from 1 to 3).

_____ wait more time and observe the patient

_____ raise the dose

_____ add another agent

2. You have a patient on an SSRI for more than two months andconsider the patient to be a partial responder. Please rank thefollowing strategies for usual next step treatments in order ofpreference (from 1 to 3).

_____ raise the dose (please fill in below)

_____ augment or combine with _____________________



_____ switch to ______

3. You have a patient on an SSRI for more than two months andconsider the patient to be a non-responder. Please rank thefollowing strategies for usual next step treatments in orderof preference (from 1 to 3).

_____ raise the dose




_____ switch to ______

4. You have a patient on an SSRI for more than two months, andthe patient was a responder but then worsened clinicallywhile still on medication. Please rank the following strategiesfor usual next step treatments in order of preference(from 1 to 3).

_____ raise the dose (please fill in below)




_____ switch to ______

408


Prior Benzodiazepine Exposure and Treatment Outcome

409



e have previously proposed that at least 50% ofpatients with chronic generalized anxiety disor-

Prior Benzodiazepine Exposure andBenzodiazepine Treatment Outcome

Karl Rickels, M.D., and Ellen W. Freeman, Ph.D.

Background: We examined discontinuationsymptoms following brief benzodiazepinetherapy (8 weeks) and intermittent benzodiaze-pine therapy (2 weeks with at least 2 weekswithout drug) and associations with prior ben-zodiazepine use. The hypothesis was that priorbenzodiazepine use would predispose patientsto more severe discontinuation symptoms.

Method: Data were drawn from 3 double-blind, randomized, placebo-controlled, publishedtreatment trials: alprazolam for patients with pre-menstrual syndrome (PMS) and diazepam andlorazepam for patients with generalized anxietydisorder (GAD). The PMS group provided pro-spective daily symptom ratings, which allowedongoing investigation of effects of prior treat-ment. In the GAD groups, taper outcome wasexamined after 8 weeks of benzodiazepinetherapy as a function of prior benzodiazepineuse and as a function of time since last prior ben-zodiazepine use. Symptom scores were analyzedusing t statistics in the PMS group and analysisof covariance with 8-week scores as the covariatein the GAD groups.

Results: The PMS subjects reported no in-crease in symptom scores and no significantdifference from placebo-treated subjects duringtaper and discontinuation of alprazolam in thefollicular phase of each treatment cycle. In theGAD trials, the results of treatment discontinu-ation did not differ significantly as a function ofpresence or absence of prior benzodiazepine useor as a function of time since last benzodiazepineuse.

Conclusion: These preliminary data fail tosupport the hypothesis that prior benzodiazepineuse predisposes patients to more severe discon-tinuation symptoms when treatment is brief anddoses are low.


Wder (GAD) and many other patients who experience anxi-ety sufficient to impair functioning would benefit frombrief benzodiazepine treatment (up to 8 weeks) or inter-mittent therapy1 (up to 2 weeks with at least 2 weeks with-out the drug at repeated intervals). This contention is nowfurther supported by data from patients with premenstrualsyndrome (PMS), whose cyclic patterns of 1 to 2 weeks ofsymptoms followed by 2 to 3 weeks of symptom remis-sion in each menstrual cycle provide a model for intermit-tent therapies.

Of the drugs available for intermittent treatment ofanxiety disorders, the benzodiazepines hold the advantageof rapid onset, consistent efficacy, ease of use, and a widemargin of safety when prescribed for only a few weeks.2

However, these clear advantages are compromised by theproblem of rebound anxiety with discontinuation of thebenzodiazepine, which can occur following even 4 weeksof treatment.3,4 Rebound phenomena frequently cannot bedifferentiated by the patient or the physician from theoriginal anxiety and may often lead to unnecessary long-term drug therapy.

Alprazolam was an effective treatment for approxi-mately 40% of clearly diagnosed PMS patients,5 who tookthe medication repeatedly for 2-week intervals during thesymptomatic premenstrual phase of the menstrual cycleand stopped taking the medication during the 2-week fol-licular phase of the cycle. These results contrasted withthose of GAD patients, who were treated for 6 monthswith a benzodiazepine and experienced significantlymore discontinuation symptoms if they had prior benzo-diazepine use compared with patients with no prior ben-zodiazepine use.6 We have now followed these intriguingfindings by examining 3 data sets: (1) patients who metclearly defined criteria for PMS and were treated premen-strually only with alprazolam compared with oral micron-ized progesterone and placebo,5 (2) GAD patients whowere treated for 8 weeks with diazepam compared withgepirone and placebo,7 and (3) GAD patients who weretreated for 8 weeks with lorazepam compared withipsapirone and placebo.8,9 In the latter 2 studies, the diaze-pam results clearly showed that, after 8 weeks of treat-ment, abrupt discontinuation of diazepam resulted in sig-nificantly more discontinuation symptoms compared with

Received May 16, 1999; accepted Oct. 22, 1999. From theDepartments of Psychiatry and Obstetrics and Gynecology, University ofPennsylvania Medical Center, Philadelphia.

Supported in part by grants HD 18633 from the National Institutes ofHealth and MH 08957 from the National Institute of Mental Health.

Beatriz Garcia-Espana, M.A., and Felipe Garcia-Espana, Ph.D.,conducted the computer analyses.

Reprint requests to: Ellen W. Freeman, Ph.D., Department ofObstetrics and Gynecology, 2 Dulles Bldg., Mudd Suite, University ofPennsylvania Medical Center, 3400 Spruce St., Philadelphia, PA 19104.


Rickels and Freeman



discontinuation of either gepirone or placebo, but no in-crease above baseline symptoms was observed.7 Simi-larly, after 8 weeks of lorazepam treatment followed by a50% taper for 4 days and then complete discontinuation,lorazepam produced significantly more discontinuationsymptoms than placebo or ipsapirone.8,9

This report further examines discontinuation symptomsin relation to the intermittent dosing that was prescribedfor the PMS patients and in relation to prior benzodiaze-pine use in the GAD patients. In the PMS trial, the sub-jects experienced symptoms premenstrually only, andnone had been taking benzodiazepines prior to the trial.Therefore, the PMS group allowed prospective obser-vation of both the first ending of benzodiazepine use andrepeated dosing at approximately 2-week intervals notconfounded by a possible return of original anxiety symp-toms. In the 2 GAD trials, patients who had stopped ben-zodiazepine treatment more than 1 month before the trialwere compared with patients who stopped benzodiazepinetreatment at 1 month or less before the trial, and thesegroups were compared with benzodiazepine-naive patientsto determine the association of prior benzodiazepine usewith discontinuation symptoms. The data reported herewere not examined in the previous reports of these trials.

METHOD

The subject selection criteria and procedures are fullydescribed in previous reports. In the alprazolam group,5

subjects with PMS were randomly assigned to treatmentwith alprazolam (N = 55) or placebo (N = 54) for 3 monthsof double-blind treatment with 3 additional months of op-tional maintenance (double-blinded). The demographiccharacteristics did not significantly differ between thealprazolam- and placebo-treated groups. The mean ± SDage of the subjects was 34 ± 6 years; the mean duration ofPMS was 10 ± 8 years. Eighty-two percent had educationbeyond high school; 62% reported a family history of men-tal illness as recorded in the medical history.

Alprazolam was administered under double-blind con-ditions from day 18 of the menstrual cycle to the first dayof menses, with a taper on the first 2 menstrual days. Theinitial dose was 0.75 mg/day, taken in divided doses 3times daily. The dose could be increased to 1 mg/day inthe first cycle and further increased to a maximum of 3mg/day in the second and third cycles. The mean ± SD al-prazolam dose after 3 months of treatment was 1.5 ± 0.5mg/day. Subjects continued the same dose taken at the endof the 3 months of acute treatment for the 3 months ofmaintenance.

Symptom severity was assessed by subjects’ daily rat-ings of premenstrual symptoms, using the University ofPennsylvania Daily Symptom Rating form (DSR).10 Eachof 17 symptoms was rated daily on a 5-point scale rangingfrom 0 (none) to 4 (severe). The ratings for each day were

summed for a total daily score. A mean of the total dailyscore was obtained for the premenstrual and postmen-strual days indicated. The pretreatment mean of the DSRscores in the 3 screen cycles was 36 ± 34 postmenstrually(days 5–10) and 144 ± 53 premenstrually (days 23–28)(p < .001).

In the diazepam group,7 41 patients with GADwere randomly assigned to 8 weeks of double-blind treat-ment, 20 with prior benzodiazepine use and 21 with noprior benzodiazepine use (pretaper daily doses were18.0 ± 11.2 and 21.2 ± 11.7 mg, respectively). The demo-graphic characteristics did not differ between the prior useand no prior use groups. The mean age of the subjects was43 ± 15 years; 59% had an episode duration of ≥ 1 year;61% were female.

In the lorazepam group,8,9 55 patients with GAD wererandomly assigned to 8 weeks of double-blind treatment:13 with prior benzodiazepine use and 42 with no priorbenzodiazepine use (pretaper daily doses were 3.9 ± 1.2and 3.6 ± 1.2 mg, respectively). The demographic back-ground characteristics did not differ between the 2 groups.The mean age of the subjects was 40 ± 12 years; 24% hadan episode duration of ≥ 1 year; 38% were female. Out-come measures in both GAD trials were the HamiltonRating Scale for Anxiety (HAM-A)11 and the Physicians’Withdrawal Checklist (PWC).12

In the PMS study,5 the mean DSR scores were com-pared between the alprazolam and placebo groups usingStudent t tests. In the GAD studies,7–9 symptom scores ineach of the treatment groups were analyzed using analysisof covariance with the 8-week scores as the covariate.Background variable comparisons were tested with t sta-tistics for continuous variables and with the chi-squaretest for frequency distributions. In all analyses, resultswere considered significant at p < .05 with 2-tailed inter-pretation. The statistical software package was SAS.13

RESULTS

Intermittent Alprazolam Use in PMSFigure 1 shows the total DSR scores for the last pre-

menstrual days, the discontinuation period, and subse-quent postmenstrual days after 1 month and 5 months ofdouble-blind therapy. Comparing the response of alprazo-lam versus placebo, the data clearly indicate that patientsfor whom alprazolam was prescribed for 14 days andstopped with a 1- to 2-day taper developed no discontinu-ation symptoms.

Review of each of the 17 DSR symptoms showed thatonly the insomnia item evidenced an increase after taper-ing and stopping alprazolam on days 1 to 3 of the men-strual cycle following the first treatment period. The in-crease in insomnia was greater in the alprazolam groupcompared with the placebo-treated subjects, but did notreach statistical significance with the Bonferroni correc-



411


One personal copy may be printedtion, and subsequently decreased in each treatment periodand did not differ from placebo-treated subjects. Insomniais a possible withdrawal symptom, but this single findingamong 17 daily-rated symptoms, which diminished ratherthan increased over time, could also be due to chance.

To determine whether subjects dropped out because ofdiscontinuation symptoms, we reviewed the reasons fordropout and reported side effects in each treatment cycle.No subject was withdrawn from the study because of ad-verse events. No subject requested to continue taking al-prazolam beyond the 2-week intervals of each menstrualcycle as prescribed by the protocol. Only 2 subjects saidthat the reason for dropping out of the study was side ef-fects. The side effects reported by all dropouts (N = 20)included fatigue (N = 11), swelling (N = 5), breast tender-ness (N = 3), increased appetite (N = 2), and one each fordizziness, diarrhea, constipation, and increased cyclelength, but no symptoms viewed as discontinuation symp-toms. Forty percent (8/20) of the dropouts experienced noside effects.

Taper Outcome After 8 Weeks ofBenzodiazepine Treatment for GAD

In both the diazepam and the lorazepam trials, the re-sults of treatment discontinuation did not differ signifi-cantly as a function of presence or absence of prior benzo-diazepine use (Table 1). We then compared patients whoseprevious benzodiazepine use occurred more than 1 monthbefore study treatment with those who had used benzo-diazepines within 1 month (but not within 2 weeks) beforeentering the 8-week double-blind treatment phase of thestudy. The prediction was that patients who had stoppedtaking their benzodiazepine within the previous monthwould have more severe discontinuation symptoms thanpatients who had stopped benzodiazepine use more than1 month ago.14,15 This prediction was not confirmed.HAM-A scores at peak during taper showed a pattern op-posite to the prediction. In the lorazepam study, the pa-tients who stopped benzodiazepine use within a month ofthe study had less severe symptoms at peak during tapercompared with the subjects who discontinued benzodiaze-

Table 1. Severity of Discontinuation Symptoms After 8 Weeks of Benzodiazepine Therapy for GeneralizedAnxiety Disorder as a Function of Presence or Absence of Prior Benzodiazepine Usea

Prior Diazepam Useb Prior Lorazepam Usec

Yes No Fd p Yes No F pMeasure (N = 20) (N = 21) (df = 2,40) Value (N = 13) (N = 42) (df = 2,54) Value

HAM-A peak during taper 11.7 ± 8.5 17.4 ± 9.3 3.36 .08 16.2 ± 9.1 17.1 ± 8.8 0.00 .99PWC peak during taper 9.5 ± 9.1 14.1 ± 9.9 2.26 .14 18.9 ± 14.2 21.2 ± 17.0 0.18 .67aAbbreviations: HAM-A = Hamilton Rating Scale for Anxiety, PWC = Physicians’ Withdrawal Checklist.bData from Rickels et al.7cData from Cutler et al8 and Mandos et al.9dAnalysis of covariance with score at 8 weeks as the covariate; mean ± SD scores (unadjusted) are shown.

aData from Freeman et al.5 DSR scores are means of the sum of daily symptom ratings for the days indicated. Seventeen symptoms were rated0 (none) to 4 (severe); range, 0–68 per day.bTreatment cycle 1: alprazolam N = 55, placebo N = 54.cTreatment cycle 5: N = 35 in each group.

30

20

10

026–28 1–3 4–5 8–10

Treatment Cycle 1, db

DS

R T

otal

Sco

re30

20

10

0

Treatment Cycle 5, dc

DS

R T

otal

Sco

re

p < .07

26–28 1–3 4–5 8–10

AlprazolamPlacebo

AlprazolamPlacebo

Figure 1. Total Daily Symptom Rating (DSR) Scores for Alprazolam- and Placebo-Treated Groupsa



413



pine use more than 1 month before the study (F = 7.03,df = 2,12; p = .02). The corresponding comparison in thediazepam group showed a trend in the same direction,with recent discontinuers having the less severe symp-toms, but did not reach statistical significance (F = 2.43,df = 2,19; p = .14). These preliminary data suggest thatthe duration of prior benzodiazepine use for these condi-tions has no bearing on discontinuation symptoms.

DISCUSSION

These preliminary data refute the hypothesis that priorbenzodiazepine use, given either continuously or inter-mittently, predisposes patients to more severe discontinu-ation symptoms following brief benzodiazepine therapy.In fact, the data show that even for a short half-life benzo-diazepine such as alprazolam, up to 2 weeks of treatmentcan be offered over multiple time periods when the brieftreatment period is followed by at least 2 weeks with nobenzodiazepine medication.

When GAD patients were treated for at least 8 weekswith a benzodiazepine, discontinuation symptoms clearlyoccurred, although the symptoms were relatively mild. Inthe lorazepam group, in which the patients were seenweekly, the symptoms were clearly more marked 4 daysafter starting the taper than after 10 days, when they didnot differ from the symptoms reported in the placebogroup.9 Therefore, it appears that benzodiazepine treat-ment even up to 8 weeks does not increase the risk of se-vere discontinuation symptoms. Although discontinuationsymptoms did occur in this time frame, they were mild, ofbrief duration, and typically manageable.

In keeping with the known patterns of withdrawal forshorter versus longer half-life benzodiazepines,12 theshort half-life drugs were tapered at discontinuation inthese studies, whereas the longer half-life drug wasstopped abruptly. Lorazepam was tapered for 4 days athalf the dose while diazepam was discontinued with notaper. With the taper adjustment for the short half-life,withdrawal symptoms subsided at about the same time inboth GAD groups, regardless of the half-life of the medi-cation. In both treatment groups, peak withdrawal symp-toms were reported at the 1-week visit after ending treat-ment, the withdrawal symptoms were less severe than thebaseline symptom levels, and the withdrawal symptomssubsided by the 2-week visit.

We had previously found that discontinuation symp-toms were more marked in patients with prior benzo-diazepine use.6 Again, the previous study examinedlonger-term treatment (6 months), in contrast to thepresent studies, in which treatment was of brief duration(2–8 week treatment periods). On the basis of presentdata, we speculate that prior benzodiazepine use has lessbearing for short treatment periods than for longer dura-tions of use. Regardless of prior use, the present results do

not pertain to longer durations of benzodiazepine treat-ment, which have well-documented discontinuation prob-lems, e.g., 6 months of treatment with clorazepate6 or anybenzodiazepine use for 1 year or longer.12

The results with PMS patients, who had no history ofsedative, alcohol, or drug abuse and were symptomaticonly 1 to 2 weeks in each month, suggest that intermittentadministration of benzodiazepines can be a useful treat-ment for PMS and GAD. Although serotonergic antide-pressants are an effective treatment for a larger proportionof PMS patients,16,17 the response rate to selective seroto-nin reuptake inhibitors is approximately 60%, leaving astill-sizeable patient group who may benefit from othermedications such as benzodiazepines. Conversely, whenanxious patients must be treated for more than 8 weeksrather than for a brief time period, the physician shouldconsider anxiolytics other than benzodiazepines, such asbuspirone18 or antidepressants.19

It is important to emphasize that the findings pertain tothe diagnoses studied, i.e., generalized anxiety disorderand premenstrual syndrome. Benzodiazepines should notbe given to patients with current or past substance abuseor clear comorbid personality disorders. While the GADstudies in this report included patients 19 to 78 years ofage, the age data are insufficient to indicate prescriptionfor the elderly. Also, these results relate only to the dosesused, which were equivalent in the 3 studies (1 mg ofalprazolam = 2 mg of lorazepam or 10 mg of diazepam).Further study is needed to determine whether these find-ings pertain to higher doses.

In conclusion, these preliminary data suggest that foranxiety problems that can be appropriately treated in ashort time interval with low-to-moderate benzodiazepinedoses, benzodiazepine use for 2 weeks or less has fewwithdrawal symptoms, whereas 8 weeks of use has mildand manageable withdrawal symptoms, in contrast to thewell-documented withdrawal problems that occur withlonger periods of benzodiazepine use. Prior benzodiaze-pine use does not appear to affect withdrawal in short-term use, again in contrast to longer periods of use inwhich prior benzodiazepine use is associated with greaterwithdrawal problems. In all cases, when benzodiazepinesare prescribed, the physician should discuss discontinu-ation management with the patient at the outset. More-over, medication for anxiety should be prescribed not as apanacea for all the patients’ problems, which cannot besolved by drug treatments alone, but as an agent to reduceanxiety and enable more effective problem solving bypatients themselves. Even today, after more than 30 yearsof use, the benzodiazepines have not been replaced asthe treatment of choice for brief treatment of anxiouspatients.1

Drug names: alprazolam (Xanax and others), buspirone (BuSpar), clor-azepate (Tranxene), diazepam (Valium and others), lorazepam (Ativanand others).

412


Rickels and Freeman



REFERENCES

1. Rickels K, Schweizer E. The spectrum of generalized anxiety in clinicalpractice: the role of short-term, intermittent treatment. Br J Psychiatry1998;173(suppl 34):49–54

2. Hollister LE, Müller-Oerlinghausen B, Rickels K, et al. Clinical uses ofbenzodiazepines. J Clin Psychopharmacol 1993;13(6, suppl 1):1S–169S

3. Fontaine R, Chouinard G, Annable A. Rebound anxiety in anxious patientsafter abrupt withdrawal of benzodiazepine treatment. Am J Psychiatry1984;141:848–852

4. Rickels K, Fox IL, Greenblatt DJ, et al. Clorazepate and lorazepam:clinical improvement and rebound anxiety. Am J Psychiatry 1988;145:312–317

5. Freeman EW, Rickels K, Sondheimer SJ, et al. A double-blind trial of oralprogesterone, alprazolam and placebo in treatment of severe premenstrualsyndrome. JAMA 1995;274:51–57

6. Rickels K, Schweizer E, Csanalosi I, et al. Long-term treatment of anxietyand risk of withdrawal: prospective comparison of clorazepate and bu-spirone. Arch Gen Psychiatry 1988;45:444–450

7. Rickels K, Schweizer E, DeMartinis N, et al. Gepirone and diazepamin generalized anxiety disorder: a placebo-controlled trial. J Clin Psycho-pharmacol 1997;12:272–277

8. Cutler NR, Sramek JJ, Keppel Hesselink JM, et al. A double-blindplacebo-controlled study comparing the efficacy and safety of ipsapironeversus lorazepam in patients with generalized anxiety disorder: a prospec-tive multicenter trial. J Clin Psychopharmacol 1993;13:429–437

9. Mandos LA, Rickels K, Cutler N, et al. Placebo-controlled comparison ofthe clinical effects of rapid discontinuation of ipsapirone and lorazepamafter 8 weeks of treatment for generalized anxiety disorder. Int Clin Psy-

chopharmacol 1995;10:251–25610. Freeman EW, DeRubeis RJ, Rickels K. Reliability and validity of a daily

diary for premenstrual syndrome. Psychiatry Res 1996;65:97–10611. Hamilton M. The assessment of anxiety states by rating. Br J Med Psychol

1959;32:50–5512. Rickels K, Schweizer E, Case WG, et al. Long-term therapeutic use of

benzodiazepines, I: effects of abrupt discontinuation. Arch Gen Psychiatry1990;47:899–907

13. SAS/STAT for Personal Computers. Version 6.04. Cary, NC: SAS Insti-tute; 1990

14. Korkmaz S, Carlsson S, Wahlström G. Development of physical depen-dence to lorazepam in rats: the effects of repeated short treatments, doseand age. Eur Neuropsychopharmacol 1998;8:175–181

15. Lukas SE, Griffiths RR. Precipitated diazepam withdrawal in baboons: ef-fects of dose and duration of diazepam exposure. Eur J Pharmacol 1984;100:163–171

16. Steiner M, Judge R, Kumar R. Serotonin re-uptake inhibitors in the treat-ment of premenstrual dysphoria: current state of knowledge. Int JPsychiatr Clin Pract 1997;1:241–274

17. Yonkers KA, Halbreich U, Freeman E, et al. Symptomatic improvementof premenstrual dysphoric disorder with sertraline treatment. JAMA 1997;278:983–988

18. Fulton B, Brogden B. Buspirone: an updated review of its clinical pharma-cology, therapeutic pharmacology, and therapeutic applications. AdisDrug Evaluation. CNS Drugs 1997;1:68–88

19. Rickels K, Downing R, Schweizer E, et al. Antidepressants for the treat-ment of generalized anxiety disorder: a placebo-controlled comparison ofimipramine, trazodone and diazepam. Arch Gen Psychiatry 1993;50:884–895

413


Adderall for Adult ADHD

415



ttention-deficit/hyperactivity disorder (ADHD)can involve impairment across a variety of do-

Low-Dose Amphetamine Salts andAdult Attention-Deficit/Hyperactivity Disorder

Joseph P. Horrigan, M.D., and L. Jarrett Barnhill, M.D.

Background: Effective treatments forattention-deficit/hyperactivity disorder (ADHD)in adults are still being defined. Pediatric studieshave suggested that a mixed amphetamine saltproduct (Adderall) is safe and effective in thetreatment of childhood forms of ADHD. Pres-ently, there are no reports in the scientific litera-ture concerning the safety and efficacy ofAdderall in adults with ADHD, which is thefocus of this study.

Method: Twenty-four outpatients (meanage = 33.3 years) with DSM-IV ADHD wereadministered Adderall in an open-label fashion,starting at 5 mg p.o. b.i.d., with titration accord-ing to clinical response, across a 16-week period.Relatives or spouses of each patient completedserial checklists (including the Copeland Symp-tom Checklist and the Brown Attention-DeficitDisorder Scales). Prospectively collected datawere analyzed retrospectively.

Results: Thirteen patients (54%) responded ina positive manner to Adderall, based on ClinicalGlobal Impressions-Improvement scale scores.The mean end dose for responders was 10.77mg/day (0.14 mg/kg/day). An intent-to-treatanalysis revealed a decrease in the meanCopeland score from 99.05 to 63.3 (p < .001),while the mean Brown score dropped from76.75 to 50.85 (p < .0001). Nine patients (38%)were poor responders or nonresponders toAdderall. Acute anxiety symptoms occurred in4 of 7 patients with a comorbid anxiety diagnosis.

Conclusion: Adderall may be an effectiveagent for the treatment of adult forms of ADHD,with positive responses occurring at relativelylow doses, at least for some individuals. How-ever, Adderall may precipitate anxiety in vulner-able individuals. Further study is required.


Received Oct. 10, 1998; accepted Nov. 19, 1999. From the Departmentof Psychiatry, University of North Carolina at Chapel Hill, Chapel Hill.

Reprint requests to: Joseph P. Horrigan, M.D., Department ofPsychiatry, CB#7160, University of North Carolina at Chapel Hill, ChapelHill, NC 27599-7160 (e-mail: [email protected]).

Amains, with presenting signs and symptoms including hy-peractivity, impulsivity, inattention, distractibility, and lowfrustration tolerance. The prevalence of ADHD in child-hood may be over 10%, and symptoms persist into adult-hood (to a degree that causes continuing impairment) inas many as 65% of these individuals.1–8 While efforts havebeen made to develop sensitive and specific diagnostic in-ventories for ADHD in adulthood,9–15 the diagnostic pro-cess itself often presents the most challenging aspect of theclinical management of adults with this disorder.16,17 How-ever, substantial evidence suggests that adult formsof ADHD can be reliably diagnosed and effectivelytreated.18,19 Despite this progress, the range of effectivetreatments for adults with ADHD has yet to be defined,particularly with regard to possible pharmacotherapeuticapproaches.

Childhood treatment studies have been more extensive.For instance, psychostimulants represent the mainstay ofpharmacotherapeutic treatment for childhood forms ofADHD.20,21 The first controlled trial involving the use ofamphetamines in children with ADHD (or at least a phe-notypic equivalent of this disorder) was published in1937.22 In subsequent years, the findings of numerousclinical trials have been published involving psychostim-ulants in children and adolescents with ADHD, including22 studies involving amphetamines.23

In March 1996, a mixed amphetamine salt product(Adderall, a psychostimulant that contains d-amphetamineand l-amphetamine) received U.S. Food and Drug Admin-istration approval for unrestricted use in the treatment ofADHD.24 Clinical reports have emerged concerning the ef-ficacy of this mixed amphetamine salt product in child-hood forms of ADHD.25–27 In these childhood studies,Adderall appeared to be both effective and well tolerated.In addition, Adderall seemed to have a unique clinical pro-file, at least when compared with the standard preparationof methylphenidate. More specifically, these studies dem-onstrated both a unique time-course effect (the time ofpeak effect and the duration of action increased as the doseof Adderall was increased) as well as a marked preferencefor Adderall on the part of caretakers and teachers. In con-junction, no unusual or serious side effects were notedwith Adderall treatments.

414


Horrigan and Barnhill



Empirical studies have suggested that psychostimu-lants are effective in the treatment of adult ADHD.28–34

Given the encouraging findings concerning Adderall inchildhood ADHD, an exploration of the potential applica-tions of Adderall in the adult population appears to bewarranted.

METHOD

Twenty-four adult outpatients with ADHD (DSM-IV314.01, combined type) were diagnosed and treated withAdderall during a 12-month period at a university-basedneuropsychiatric clinic. Demographic, medical, psychiat-ric, and outcome data were collected on these patients ina prospective manner, as they are for all patients at theclinic, and the data were analyzed retrospectively via achart review.

The ADHD diagnosis was rendered following a struc-tured and semistructured interview of both the patient andat least one first-degree relative or the patient’s spouse (ifhe or she was married). Standardized checklists (includ-ing the Wender Utah Rating Scale,9,10 the Copeland Symp-tom Checklist for Adult Attention Deficit Disorders,12

and the Brown Attention-Deficit Disorder Scales15) werecompleted as part of the diagnostic process, and the com-monly accepted threshold scores described by the authorsof these checklists were applied prior to formal renderingof the ADHD diagnosis.

The diagnostic process included a complete review ofall medical and psychiatric records, a complete physicalexamination, and baseline laboratory work (including acomplete blood count with differential, serum electrolytelevels, liver function tests, thyrotropin levels, thyroidpanel, urinalysis, a urine toxicology screen, and in selectcases, an electrocardiogram). The clinical algorithmsfor the use of psychostimulant medications in this clinicconsidered hypertension (resting systolic blood pres-sure > 140 mm Hg, diastolic blood pressure > 90 mm Hg)and a myocardial infarction in the preceding 12 months tobe contraindications to psychostimulant therapy. Addi-tional exclusion criteria included any of the following:mental retardation, any active substance use disorder,schizophrenia or a psychotic disorder, bipolar disorder, acurrent episode of major depression, and a cluster B per-sonality disorder. The diagnosticians were the authors,who are dually board-certified in general as well as childand adolescent psychiatry.

Each patient was administered Adderall in an open-label fashion following written, informed consent. Theform of Adderall utilized was the scored 10-mg tablet,which contains 2.5 mg each of d-amphetamine sulfate;d,l-amphetamine sulfate; d,l-amphetamine aspartate; andd-amphetamine saccharate. At the time of initiation ofAdderall therapy, 4 of the patients were also taking a fixeddose of a selective serotonin reuptake inhibitor (sertraline

or venlafaxine), with no change in dose throughout theAdderall trial.

The Adderall was initiated at a dose of 5 mg p.o. b.i.d.(breakfast and lunch), and the dose was titrated accordingto clinical response over the next 16 weeks, employing atwice-daily dosing schedule throughout the study. Serialchecklists (including the Copeland Symptom Checklist12

and the Brown Attention-Deficit Disorder Scales15) werecompleted every 2 weeks by the patient along with his/herspouse or a first-degree relative. Structured side effectchecklists were completed on a weekly basis. Telephonecontacts occurred every 2 weeks, while clinic visits wereconducted on a monthly basis (at which time ClinicalGlobal Impressions scales35 were also completed by thephysician). The clinical endpoint of dosage adjustmentoccurred when intolerable side effects intervened or thepatient and his/her family member, in conjunction withthe prescribing physician, determined that maximumclinical benefits had occurred. It should be noted that nospecific target end dose was established a priori, anddosage adjustments were typically conducted every 2weeks as needed, in 5-mg increments. Treatment-relatedchanges were evaluated in a within-subjects manner,using paired t tests at the α = .01 level.

RESULTS

In this study, there were 12 men (mean ± SDage = 33.25 ± 9.73 years) and 12 women (mean ± SDage = 33.42 ± 11.89 years). The mean number of years ofpostsecondary schooling was 2.94 ± 1.84 years, corre-sponding to a mean Hollingshead-Redlich EducationCode36 of 2.61 ± 0.78. The mean Hollingshead-RedlichHighest Occupation Code36 was 3.46 ± 0.83. The comor-bid diagnoses at the time of initiation of Adderall treat-ment are detailed in Table 1.

Utilizing the criterion of a score of 1 (“very much im-proved”) or 2 (“much improved”) on the Clinical GlobalImpressions-Improvement scale (CGI-I) as a marker forpositive response, 13 (54%) of the patients were positiveresponders to Adderall after 16 weeks. The specific CGI-Iscores included 10 (42%) who were “very much im-proved” and 3 (12%) who were “much improved.” An

Table 1. Comorbid DiagnosesDiagnosis N %

Reading disorder 10 42Mathematics disorder 9 38Disorder of written expression 5 21Dysthymia 5 21Depressive disorder NOS 5 21Social phobia 3 12Anxiety disorder NOS 3 12Panic disorder 2 8Eating disorder NOS 1 4Developmental coordination disorder 1 4

415


Adderall for Adult ADHD

417



additional 2 (8%) were “minimally improved.” In thisgroup of 15 (positive or minimal responders), few side ef-fects were noted: 2 patients experienced a mild decrease inappetite, 1 experienced initial insomnia, and 1 experiencedmild sedation 4 to 5 hours after a given dose (Table 2 in-cludes all side effects observed). In addition, in this group,the mean ± SD effective dose was 10.33 ± 4.10 mg/day or0.14 ± 0.06 mg/kg/day, while the modal dose remained at5 mg p.o. b.i.d. For the 13 that obtained a CGI-I score of 1or 2, the mean effective dose was 0.14 ± 0.06 mg/kg/day,corresponding to 10.77 ± 4.3 mg/day.

In the above-described cohort of 15 individuals, theCopeland dimensions of inattention/distractibility, impul-sivity, noncompliance, and underactivity were most notice-ably affected by the Adderall treatment. In terms of meanpercent change for these various dimensions, inattention/distractibility decreased by 41%, impulsivity by 33%, non-compliance by 30%, underactivity by 27%, emotional dif-ficulties by 26%, underachievement/disorganization by24%, overactivity by 22%, “impaired family relationships”by 18%, and “poor peer relations” by 15%.

In the poor or nonresponder group of 9 patients (38%),all experienced side effects (including a more substantialdecrease in appetite, gastrointestinal upset, migraine head-ache, and heart palpitations; see Table 2), prompting eachone to discontinue the Adderall within the first 3 weeks ofthe trial. Of note, 4 of 7 patients in the study with a comor-bid anxiety diagnosis experienced near-immediate symp-toms of acute anxiety (including diaphoresis, tremor, short-ness of breath, and an impending sense of doom) at thestart of Adderall therapy. This adverse response led to dis-continuation of Adderall in each case within the first 24hours. Accordingly, these 4 patients did not provide check-list scores subsequent to the initiation of Adderall, al-though they did complete side effect checklists.

Intent-to-treat data analyses (utilizing a last-observation-carried-forward approach) were conducted on the outcomemeasures from the 20 patients that continued takingAdderall for more than 2 weeks. Utilizing this approach,Adderall continued to manifest a favorable response.In this cohort of 20, the mean ± SD Copeland scoredropped from 99.05 ± 27.02 to 63.3 ± 35.19 (t = 4.83,

2-tailed, p < .001), while the mean Brown score droppedfrom 76.75 ± 17.77 to 50.85 ± 27.66 (t = 5.20, 2-tailed,p < .0001). For this intent-to-treat analysis, the meanCGI-I score at the end of 16 weeks was 2.30 ± 1.66. Inaddition, the mean Adderall dose was 10.50 ± 4.34mg/day, corresponding to 0.13 ± 0.06 mg/kg/day.

DISCUSSION

This study has a variety of weaknesses. The design wasopen-label, which allowed for variable titration (dose ad-justment) of the Adderall. The sample size was relativelysmall, which diminishes the generalizability of the find-ings. The exclusion criteria may have diluted the general-izability even further, given that the majority of patientswith adult ADHD frequently have more complicated pat-terns of comorbidity.37 The outcome measures have notbeen well standardized, outside of the CGI scale. In thisstudy, no control group was included, and there were noblind raters. Finally, 3 of the positive responders were tak-ing fixed doses of either sertraline or venlafaxine duringthe course of this study. This may have been a confound-ing variable, since there is preliminary evidence that ven-lafaxine may be mildly helpful with some dimensions ofADHD.23

Despite these shortcomings, the results of this pilotstudy suggest that, in select cases, Adderall may be an ef-fective treatment for the adult form of ADHD. This appar-ent efficacy is in agreement with the preceding studies in-volving methylphenidate and adult ADHD.29,31–33 Oneinteresting finding in this study is that a small majority ofthe patients (15/24, 62%) achieved maximum clinical ben-efit on a relatively low dose of Adderall (0.14 mg/kg/day).This finding is in contrast to those of the more recent stud-ies involving methylphenidate, which have suggested thatmore substantial dosing (in the range of 1 mg/kg/day) isrequired before consistent positive responses are ob-served.31,32 One possible explanation may be that the doseof Adderall was not adequately advanced in the nonre-sponders in this study (although all of these individualsexperienced side effects at low doses and would not toler-ate further dose increases).

Further studies addressing the potential efficacy ofAdderall in adult ADHD would be worthwhile. One par-ticularly interesting avenue of exploration may involvethe specific time course of response to Adderall in thispopulation. In addition, it may ultimately be helpful toperform more systematic comparison studies between thevarious psychostimulants, not only to determine relativeefficacy, but also to discern if there are preferential pat-terns of response. If the latter proves true, this may ulti-mately assist with appropriate medication choice at thestart of drug therapy. Equally informative would be ananalysis of the impact of the various psychostimulants onthe different dimensions of adult ADHD.

Table 2. Side Effects (all patients)N %

Anxiety (generalized) 5 21Acute anxiety (panic) 4 17Decreased appetite 3 12Irritability 3 12Stomachache 2 8Dysphoria 2 8Insomnia 2 8Tremor 2 8Sedation 2 8Headache 1 4Motor tic 1 4Dizziness 1 4

416


Horrigan and Barnhill



An example of the idiosyncratic response to psycho-stimulant treatment from this study would be that 7 of 9nonresponders to Adderall were successfully crossed overto an alternative psychostimulant, with essentially no sideeffects noted. Four of these patients proved to be positiveresponders to dextroamphetamine sulfate, while the re-maining 3 responded positively to methylphenidate. Thekey difference between Adderall and dextroamphetaminesulfate is the l-amphetamine component of Adderall.Given the side effects observed in this study, it is possiblethat the levo (l-) isomer of amphetamine is relativelymore anxiogenic compared with the dextro (d-) isomer, atleast in individuals vulnerable to anxiety.38

None of the patients reported (or appeared to experi-ence) euphoria or tolerance. However, Adderall did notprove to be free of side effects, as discussed above. Whilethe findings of this pilot study are promising, further stud-ies will be required to determine the potential utility ofAdderall in the management of individuals with adultADHD.

Drug names: dextroamphetamine (Dexedrine and others), methyl-phenidate (Ritalin and others), sertraline (Zoloft), venlafaxine(Effexor).

REFERENCES

1. Goldman LS, Genel M, Bezman RJ, et al. Diagnosis and treatment ofattention-deficit/hyperactivity disorder in children and adolescents. JAMA1998;279:1100–1107

2. Wender PH, Reimherr FW, Wood DR. Attention deficit disorder (“mini-mal brain dysfunction”) in adults. Arch Gen Psychiatry 1981;38:449–456

3. Spencer T, Biederman J, Wilens T, et al. Is attention-deficit hyperactivitydisorder in adults a valid disorder? Harv Rev Psychiatry 1994;1:326–335

4. Faraone SV, Biederman J, Mick E, et al. Attention deficit hyperactivitydisorder in a multigenerational pedigree. Biol Psychiatry 1996;39:906–908

5. Biederman J, Faraone S, Milberger S, et al. Predictors of persistence andremission of ADHD into adolescence: results from a 4-year prospectivefollow-up study. J Am Acad Child Adolesc Psychiatry 1996;35:343–351

6. Roy-Byrne P, Scheele L, Brinkley J, et al. Adult attention-deficit hyperac-tivity disorder: assessment guidelines based on clinical presentation to aspecialty clinic. Compr Psychiatry 1997;38:133–140

7. Dulcan M. Practice parameters for the assessment and treatment of chil-dren, adolescents, and adults with attention-deficit/hyperactivity disorder.J Am Acad Child Adolesc Psychiatry 1997;36(10 suppl):85S–121S

8. Seidman LJ, Biederman J, Weber W, et al. Neuropsychological function inadults with attention-deficit hyperactivity disorder. Biol Psychiatry 1998;44:260–268

9. Ward MF, Wender PH, Reimherr FW. The Wender Utah Rating Scale: anaid in the retrospective diagnosis of childhood attention deficit hyperactiv-ity disorder. Am J Psychiatry 1993;150:885–890

10. Stein MA, Sandoval R, Szumowski E, et al. Psychometric characteristicsof the Wender Utah Rating Scale (WURS): reliability and factor structurefor men and women. Psychopharmacol Bull 1995;31:425–433

11. Burt KL, Parks-Charney R, Schwean VL. The AD/HD skills and strategiesprogram: a program for AD/HD adults in postsecondary education. Can JSch Psychol 1996;12:122–134

12. Copeland ED. Copeland Symptom Checklist for Adult Attention DeficitDisorders. Atlanta, Ga: Southeastern Psychological Institute; 1989

13. Triolo SJ, Murphy KR. Attention-Deficit Scales for Adults. New York,

NY: Brunner/Mazel; 199614. McCarney SB, Anderson PD. The Adult Attention Deficit Disorders

Evaluation Scale: Home Version Technical Manual. Columbia, Mo:Hawthorne Educational Services; 1996

15. Brown TE. Brown Attention-Deficit Disorder Scales. San Antonio, Tex:The Psychological Corporation; 1996

16. Zametkin AJ, Ernst M. Problems in the management of attention-deficit-hyperactivity disorder. N Engl J Med 1999;340:40–46

17. Spencer T, Biederman J, Wilens TE, et al. Adults with attention-deficit/hyperactivity disorder: a controversial diagnosis. J Clin Psychiatry 1998;59(suppl 7): 59–68

18. Barkley RA, Biederman J. Toward a broader definition of the age-of-onsetcriterion for attention-deficit hyperactivity disorder. J Am Acad ChildAdolesc Psychiatry 1997;36:1204–1210

19. Conners CK, Jett JL. Attention Deficit Hyperactivity Disorder (in Adultsand Children). Kansas City, Mo: Compact Clinicals; 1999

20. Barkley RA. A review of stimulant drug research with hyperactive chil-dren. J Child Psychol Psychiatry 1977;18:137–165

21. Dulcan MK. Using psychostimulants to treat behavioral disorders of chil-dren and adolescents. J Child Adolesc Psychopharmacol 1990;1:7–20

22. Bradley C. The behavior of children receiving benzedrine. Am J Psychi-atry 1937;94:577–585

23. Biederman J. Attention-deficit/hyperactivity disorder: a life-span perspec-tive. J Clin Psychiatry 1998;59(suppl 7):4–16

24. Popper CW. The story of four salts. J Child Adolesc Psychopharmacol1994;4:217–223

25. Swanson J, Wigal S, Greenhill L, et al. Objective and subjective measuresof the pharmacodynamic effects of Adderall in the treatment of childrenwith ADHD in a controlled laboratory classroom setting. Psycho-pharmacol Bull 1998;34:55–60

26. Swanson JM, Wigal S, Greenhill LL, et al. Analog classroom assessmentof Adderall in children with ADHD. J Am Acad Child Adolesc Psychiatry1998;37:519–526

27. Pelham WE, Aronoff HR, Midlam JK et al. A comparison of Ritalin andAdderall: efficacy and time-course in children with attention-deficit/hyperactivity disorder. Pediatr Online 1999;103(4):e43. Available at:http://www.pediatrics.org/cgi/content/abstract/103/e43. Accessed April10, 1999

28. Wender PH, Reimherr FW, Wood DR. Attention deficit disorder (‘minimalbrain dysfunction’) in adults: a replication study of diagnosis and drugtreatment. Arch Gen Psychiatry 1981;38:449–456

29. Wender PH, Reimherr FW, Wood D, et al. A controlled study of methyl-phenidate in the treatment of attention deficit disorder, residual type, inadults. Am J Psychiatry 1985;142:547–552

30. Matochik JA, Liebenauer LL, King AC, et al. Cerebral glucose metabo-lism in adults with attention deficit hyperactivity disorder after chronicstimulant treatment. Am J Psychiatry 1994;151:658–664

31. Spencer T, Wilens T, Biederman J, et al. A double-blind, crossover com-parison of methylphenidate and placebo in adults with childhood-onsetattention-deficit hyperactivity disorder. Arch Gen Psychiatry 1995;52:434–443

32. Wilens TE, Biederman J, Spencer TJ, et al. Pharmacotherapy of adult at-tention deficit/hyperactivity disorder: a review. J Clin Psychopharmacol1995;15:270–279

33. Spencer T, Biederman J, Wilens T, et al. Pharmacotherapy of attention-deficit hyperactivity disorder across the life cycle. J Am Acad ChildAdolesc Psychiatry 1996;35:409–432

34. Heiligenstein E, Anders J. Pemoline in adult attention deficit hyperactivitydisorder: predictors of nonresponse. J Am Coll Health 1997;45:225–229

35. Guy W. ECDEU Assessment Manual for Psychopharmacology. US DeptHealth, Education, and Welfare publication (ADM) 76-338. Rockville,Md: National Institute of Mental Health; 1976:218–222

36. Hollingshead AB, Redlich FC. Social Class and Mental Illness: A Com-munity Survey. New York, NY: John Wiley and Sons; 1958

37. Hornig M. Addressing comorbidity in adults with attention-deficit/hyperactivity disorder. J Clin Psychiatry 1998;59(suppl 7):69–75

38. Smith RC, Davis JM. Comparative effects of d-amphetamine,l-amphetamine, and methylphenidate on mood in man. PsychopharmacolBerl 1977;53:1–12

417

https://www.researchgate.net/publication/238337425_The_ADHD_Skills_and_Strategies_Program_A_Program_for_ADHD_Adults_in_Postsecondary_Education?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/247564933_The_Story_of_Four_Salts?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/247564933_The_Story_of_Four_Salts?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13605164_Addressing_comorbidity_in_adults_with_attention-deficithyperactivity_disorder?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13605164_Addressing_comorbidity_in_adults_with_attention-deficithyperactivity_disorder?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13928921_Toward_a_Broader_Definition_of_the_Age-of-Onset_Criterion_for_Attention-Deficit_Hyperactivity_Disorder?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/13605163_Adults_with_attention-deficithyperactivity_disorder_A_controversial_diagnosis?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/13409617_Problems_in_the_Management_of_Attention-Deficit-Hyperactivity_Disorder?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13409617_Problems_in_the_Management_of_Attention-Deficit-Hyperactivity_Disorder?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13605157_Attention-DeficitHyperactivity_Disorder_A_life-span_perspective?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13605157_Attention-DeficitHyperactivity_Disorder_A_life-span_perspective?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14142208_Pemoline_in_Adult_Attention_Deficit_Hyperactivity_Disorder_Predictors_of_Nonresponse?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14142208_Pemoline_in_Adult_Attention_Deficit_Hyperactivity_Disorder_Predictors_of_Nonresponse?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/16188290_Attention_Deficit_Disorder_Minimal_Brain_Dysfunction_in_Adults_A_Replication_Study_of_Diagnosis_and_Drug_Treatment?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/51336197_Diagnosis_and_treatment_of_attention-deficithyperactivity_disorder_in_children_and_adolescents_Council_on_Scientific_Affairs_American_Medical_Association?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/14705379_Ward_MF_Wender_PH_Reimherr_FW_The_Wender_Utah_Rating_Scale_an_aid_in_the_retrospective_diagnosis_of_childhood_attention_deficit_hyperactivity_disorder_Am_J_Psychiatry_150_885-890?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/15708502_Psychometric_characteristics_of_the_Wender_Utah_Rating_Scale_WURS_Reliability_and_factor_structure_for_men_and_women?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/284497517_Minimal_brain_dysfunction_attention_deficit_disorder_in_adults?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/284497517_Minimal_brain_dysfunction_attention_deficit_disorder_in_adults?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/285483617_Pharmacotherapy_of_attention_deficit_hyperactivity_disorder_accros_life_cycle?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/13892054_Practice_Parameters_for_the_Assessment_and_Treatment_of_Children_Adolescents_and_Adults_With_Attention-DeficitHyperactivity_Disorder?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/13570754_Neuropsychological_function_in_adults_with_attention-deficit_hyperactivity_disorder_Biol_Psychiatry_44_260-268?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/22810641_A_review_of_stimulant_drugs_Research_with_hyperactive_children?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/22810641_A_review_of_stimulant_drugs_Research_with_hyperactive_children?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1


The BPRS as an Outcome Measurement Tool

419



Background: Recent guidelines for length ofstay at psychiatric hospitals may have an unac-ceptable impact on patient outcome at discharge.A valid measurement tool is needed to evaluatesignificant patient change during brief hospital-ization, typically 7 days, and to provide early pre-diction of unfavorable short-term outcome. Thisstudy examines the utility of the Brief PsychiatricRating Scale (BPRS) as such a tool.

Method: During a 2-month testing period, theBPRS was administered to 87 successive adultsadmitted to an acute general psychiatric inpatientunit at admission, 2 days, 7 days, and weeklythereafter until discharge. Total BPRS scores and4 subscores were used in the data analysis, whichincluded paired t tests and correlation analyses.

Results: Mean BPRS total scores demonstratesignificant (p < .001) patient improvement at days2, 7, and 14 of the hospital stay. Changes insubscores and their relationship to eventual out-come vary across diagnostic groups.

Conclusion: The BPRS appears to be a usefulinpatient outcome measure since it is capable ofdemonstrating significant change during briefstays of 1 week or less. Subscale scores may pro-vide more specific prediction of change and mayhelp clarify outcome in individual patients whoshow insignificant change by total score.


Received June 25, 1999; accepted Nov. 8, 1999. From the Departmentof Psychiatry and Behavioral Sciences, the University of Texas, Houston,Medical School, Houston.

Reprint requests to: Roy V. Varner, M.D., 2800 S. MacGregor, Houston,TX 77021 (e-mail: [email protected]).

ressures to shorten the length of hospital stays havebeen joined by health care industry demands for

The Brief Psychiatric Rating Scale as an Acute InpatientOutcome Measurement Tool: A Pilot Study

Roy V. Varner, M.D.; Y. Richard Chen, Ph.D.;Alan C. Swann, M.D.; and Frederick G. Moeller, M.D.

Ppatient-outcome data.1 Unfortunately, shorter length ofstay, typically targeted at 7 days or less, may not allowtime for full response to treatment compared with thelonger stays of the past.2 Given the need for a short lengthof stay, it is important to distinguish those patients withearly favorable outcome from those requiring longer ormore vigorous treatment.

The Brief Psychiatric Rating Scale (BPRS) has been inuse since 1962, but was developed primarily for con-trolled psychoactive drug trials and similar research.3,4

Further refinement during its first 18 years of use, in-cluding questions of reliability and validity, has beensummarized.5 Recent literature has suggested broader,more routine clinical uses for the BPRS.6–15 Several ex-amples illustrate possible use in supporting treatment out-come accountability as an outcome measure in psychoso-cial rehabilitation programs,6 as a potential correlate ofsubjective quality-of-life appraisal in schizophrenic pa-tients,7 as a predictor of inpatient readmission,12 and as anaid to the determination of suicidality in schizophrenia,15

among others. We were unable to find a reference report-ing clinical use of the BPRS over the brief time spans thatare now the norm for acute inpatient episodes. We hy-pothesized that the BPRS could demonstrate significantchange in patient symptoms during a brief hospitalization,that early BPRS change could help predict the appropri-ateness of early discharge and/or degree of later improve-ment, and that subscale scores might be of more valuethan the total score for some individual patients, accord-ing to diagnosis, in demonstrating such early response.

METHOD

Data were collected for about 8 weeks during Marchand April 1996 from consecutive patients newly admittedto a single 23-bed general adult inpatient unit in the Uni-versity of Texas Harris County Psychiatric Center. Fourstaff psychiatric nurses were selected to administer theBPRS to the study patients as part of their routine nursingduties. All 4 had participated in extensive group trainingunder the direction of a staff psychiatrist with 20 years ofBPRS experience.16 Interrater reliability was establishedduring training by using videotaped clinical interviewsfollowed by a rating/discussion format using the 18-item“anchored” version of the BPRS.17 The anchored BPRS isreported to aid reliability by the addition of expandeddefinitions and instructions for each scale item and wasused throughout the study. In the present study, no formalmeasurement of reliability was made, and there was nosystematic blinding for previous ratings. However, earlierratings were not available at the time of subsequent rat-

418


Varner et al.



ings, and the nurse raters were blind to thegoal of the project and its hypotheses. Eachnurse rater was the same from admission toeach subsequent testing interval since theyserved as primary nurses assigned to spe-cific patients. All ratings were reviewed forapparent clinical validity by 1 of 2 attend-ing psychiatrists, both with BPRS experi-ence. It was anticipated that the size of thebaseline patient sample would graduallydecrease at subsequent rating points, re-flecting variable length of stay in this natu-ralistic study. Therefore, the cutoff intervalfor data analysis was ultimately set at 14days. The baseline sample (N = 87) con-sisted of 41 women and 46 men with anethnic distribution of white (N = 42), Afri-can American (N = 28), Hispanic (N = 13),and other (N = 4). The mean ± SD age was34.6 ± 9 years, mean number of total hos-pitalizations was 2.2 (range, 1–9), and themean study length of stay was 16.8 ± 8.9days (range, 1–42 days).

The baseline BPRS was administered toall 87 patients as near to admission as possible, on days 2and 7, and weekly thereafter until discharge. Patients re-ceived typical psychoactive drug treatment: a neurolepticfor schizophrenia; a mood stabilizer, often in combinationwith a neuroleptic, for mania; and an antidepressant for non-bipolar depression. Similarly appropriate medication wasused for a heterogeneous “other” patient category, mainlypatients with substance-induced psychotic or mood disor-ders and an assortment of not otherwise specified (NOS)conditions. The reported final DSM-IV discharge diagnoseswere a consensus of attending physician, nurse rater, socialworker, and psychiatric resident diagnoses. Discharge cri-teria were based solely on observed clinical improvementand attainment of individualized treatment plan goals; theBPRS scores were not considered. For data analysis, the 4general diagnostic categories derived from each patient’sdischarge DSM-IV primary diagnosis were schizophrenia(N = 13); mania, including the mixed subcategory (N = 30);depression, including major depression and depressive epi-sode of bipolar (N = 20); and other (N = 24). The “other”category contained diagnostic groups insufficient in size tobe included separately in the data analysis. However, theywere included in the “all-patients” category (i.e., the entirebaseline sample of 87 patients) to determine the importanceof factor scores or changes in scores regardless of diagno-sis. All data analyses were based on the original 0 to 6 scor-ing system for each of the 18 BPRS items where 0 meansnot present or not observed, and 6 means extremely severe.Additionally, we included 4 subscales that had been derivedfrom early BPRS development.4 Each subscale representsthe sum score of 3 items: (1) thinking disturbance: con-

ceptual disorganization, hallucinatory behavior, unusualthought content; (2) withdrawal-retardation: emotionalwithdrawal, motor retardation, blunted affect; (3) hostility-suspiciousness: hostility, suspiciousness, uncooperative-ness; and (4) anxiety-depression: anxiety, guilt feelings, de-pressive mood.

For the first step of data analysis, significance for thedifferences between baseline mean BPRS total scores andday-2, day-7, and day-14 mean BPRS total scores wastested for the all-patients category using the paired t test.Since patients with differing diagnoses are presumed tohave varied clinical presentations and rates of improve-ment, the next step was to repeat the same analysis foreach of 3 identified diagnostic categories. The differencesfor all 4 BPRS subscale scores between baseline and theother 3 timepoints were also studied with the same ana-lytic strategy. The statistical significance level was de-fined as p < .05. A correlational analysis was conducted todetermine if score differences between baseline and day 2could predict differences between day 2 and day 14.

RESULTS

Mean ± SD total scores and subscores at 4 rating pointsover 2 weeks for 4 diagnosis-based categories are pre-sented in Table 1. While obvious differences in magnitudeof statistical significance appear throughout this report,the relatively small sizes of the study samples precludeconclusive interpretation of these differences.

BPRS total scores demonstrate significant change frombaseline to each subsequent rating interval for the all-

Table 1. Brief Psychiatric Rating Scale Mean ± SD Total Scores and SubscaleScores by Patient CategoryPatient Category Baseline Day 2 Day 7 Day 14All patients N = 87 N = 82 N = 73 N = 58

Total score 18.20 ± 9.79 13.91 ± 9.78*** 9.54 ± 8.63*** 7.40 ± 7.62***Thinking disturbance 3.67 ± 3.63 2.39 ± 2.81*** 1.79 ± 2.73*** 1.19 ± 2.13***Withdrawal-retardation 2.93 ± 3.36 2.46 ± 2.99 1.86 ± 2.99 1.43 ± 2.51**Hostility-suspiciousness 3.22 ± 2.98 1.98 ± 2.81** 1.33 ± 2.82*** 1.19 ± 2.28***Anxiety-depression 5.57 ± 4.66 4.52 ± 4.44 2.40 ± 2.90*** 1.97 ± 2.74***

Schizophrenia N = 13 N = 13 N = 10 N = 7Total score 19.62 ± 8.76 14.15 ± 10.20 8.80 ± 7.53* 5.00 ± 5.66***Thinking disturbance 4.38 ± 2.88 3.31 ± 2.52 2.30 ± 2.50 0.57 ± 1.51*Withdrawal-retardation 3.08 ± 3.93 2.23 ± 2.13 2.30 ± 2.37 1.86 ± 2.40Hostility-suspiciousness 5.62 ± 3.61 2.62 ± 3.71 1.70 ± 3.76* 0.43 ± 1.11***Anxiety-depression 2.92 ± 3.39 2.92 ± 3.97 1.00 ± 2.21 1.57 ± 2.70

Mania N = 30 N = 27 N = 25 N = 22Total score 19.50 ± 12.60 15.63 ± 11.48** 11.64 ± 9.45** 6.91 ± 5.63***Thinking disturbance 4.47 ± 4.60 3.11 ± 3.43* 2.20 ± 2.65** 1.23 ± 2.11***Withdrawal-retardation 2.80 ± 3.56 3.26 ± 3.95 1.96 ± 3.70 1.00 ± 1.55*Hostility-suspiciousness 3.93 ± 3.18 2.44 ± 3.07* 1.68 ± 2.85** 1.64 ± 1.92*Anxiety-depression 4.77 ± 4.93 3.56 ± 4.31* 2.28 ± 2.10** 1.14 ± 1.31**

Depression N = 20 N = 20 N = 18 N = 14Total score 17.30 ± 8.77 14.35 ± 8.77 8.94 ± 9.38* 8.29 ± 10.44Thinking disturbance 2.55 ± 2.77 1.10 ± 1.83* 0.83 ± 1.82* 0.71 ± 1.50*Withdrawal-retardation 3.30 ± 3.53 2.15 ± 2.24 2.28 ± 3.18 2.00 ± 4.00Hostility-suspiciousness 2.20 ± 2.24 1.60 ± 2.77 1.28 ± 3.14 0.93 ± 3.48Anxiety-depression 7.90 ± 3.94 7.55 ± 4.74 3.50 ± 3.48*** 3.57 ± 4.08**

*p < .05. ** p < .01. *** p < .001.

419


The BPRS as an Outcome Measurement Tool

421



change at day 14 for the all-patients category inboth total and subscale scores. The reason for therelationship between the day 2 BPRS score andthe day 14 score is unclear. Two possible reasonsfor the association include cooperation with treat-ment and response to medications. It is possiblethat many patients responded positively to thestructure and supervision of the hospital and weremore likely to take medication. Another possibleexplanation involves response to medication.Some authors have hypothesized that there aresubpopulations of patients who show a rapid re-sponse to neuroleptics.18 Our results show thatearly change has the potential to predict eventualoutcome. Further studies should investigatewhether there is a threshold for lack of change atday 2 that could predict that a patient would beunlikely to improve during a brief hospitalization.

Since change in BPRS total and subscale scorefrom 7 to 14 days is relatively less than the changefrom baseline to day 7, a 7-day hospitalizationmay be sufficient before discharging many pa-tients, especially those demonstrating substantialimprovement at day 2. Some patients will requirea stay exceeding 7 days owing to clinical reasonsor psychosocial factors not related to rating scaledimensions. Extensive use of the BPRS by othersin controlled schizophrenia drug trials was usuallybased on a minimum of a 21-day stay to demon-strate adequate improvement.5

3. Subscale scores, rather than total scores, mighthelp predict which patient diagnostic categorieswould show the best early response during a briefhospitalization. Total BPRS score in schizophre-nia has been suspected of adding error variancesince it measures many items not typically associ-ated with schizophrenia,19 and we have speculatedthat this might apply to other diagnoses as well.Total BPRS score change might also be inconclu-sive for certain patients whose lower (i.e., lessimpaired) baseline total scores have reduced po-tential for dramatic improvement compared withhigher baseline scores. The differing pattern ofsubscore improvement according to diagnosis (seeTable 1) suggests potential advantage over totalscores for such patients. While subscores forthinking disturbance do not demonstrate a sig-nificant change until day 14 for schizophrenia, asignificant change does occur at day 2 in bothmania and depression, possibly correlating betterwith clinical improvement than would total scorein some patients with those latter diagnoses.Hostility-suspiciousness subscore demonstratesa significant change at day 7 for schizophrenia,perhaps a more useful measure for some patients

patients category. A similar finding is present for patientswith mania. Schizophrenic and depressed patients do notdemonstrate significant change until day 7. Schizophrenicpatients continue to demonstrate significant change at day14, while depressed patients demonstrate essentially nofurther change during the same interval.

Three of the subscores—thinking disturbance, hostility-suspiciousness, and anxiety-depression—demonstrate sig-nificant change at day 7 for the all-patients category,but differ according to diagnostic category: schizophrenicpatients demonstrate significant change only for thehostility-suspiciousness subscore; manic patients demon-strate significance for all 3 of these subscores, even at day2; depressed patients demonstrate significant change forthinking disturbance, none for hostility-suspiciousness,and an unequivocal change for anxiety-depression. Noneof the 4 patient categories demonstrate significant changeat day 7 for withdrawal-retardation.

A correlation analysis of BPRS total score differencebetween baseline and day 2 for the all-patients categorydemonstrated high correlation with the differencebetween baseline and day 14 (r = 0.53, p < .0001). A simi-lar high correlation was present for each of the 4subscores (thinking disturbance: r = 0.33, p = .0021;withdrawal-retardation: r = 0.46, p = .0001; hostility-suspiciousness: r = 0.46, p = .0001; anxiety-depression:r = 0.48, p = .0001).

DISCUSSION

Despite previous BPRS use, primarily in psychophar-macology research, we sought to investigate its applica-tion exclusively as a clinical inpatient tool. We hoped todemonstrate the following possibilities.

1. The BPRS could demonstrate significant changeduring brief, acute psychiatric lengths of inpatientstay, typically 7 days. Total BPRS scores indicatethat the BPRS can demonstrate significant changeby day 7 in all 4 patient categories and even asearly as day 2 for the all-patients and maniacategories. Significant change from baseline isalso seen at day 14 for all-patients, schizophrenia,and mania categories, but not for the depressioncategory, which remains unchanged, possibly ow-ing to the delayed full effect of antidepressantmedications.

2. Early change might predict both the appropriate-ness of early discharge and degree of later im-provement. The correlation between change at day2 and at day 14 was investigated to determinewhether the extent of initial change would poten-tially be able to predict the need for longer hospi-talization. Correlation analysis reveals that earlychange at day 2 correlates with further significant

420


Varner et al.



in this diagnostic group. Similarly, the anxiety-depression subscore is possibly a more useful out-come measure than total BPRS score for somedepressed patients at day 7. Since withdrawal-retardation shows the weakest treatment effect ofthe 4 BPRS subscales, it does not appear to be use-ful in demonstrating early response in any of thepatient categories, whether by itself or comparedwith total score.

The fact that there was no systematic blinding for pre-vious ratings was a potential source of bias since ratersmight tend to overstate improvement in patients on theirservice. Several factors, however, mitigated against suchbias: (1) as described in Method, earlier ratings were un-available to the raters at the time of subsequent ratings,so, in effect, raters were blinded; (2) raters were unawareof the goal or hypotheses of the study; and (3) amount andtime course of improvement varied by diagnostic groupand BPRS subscales. While the absence of a rigorous for-mal blind was not ideal, we believe it did not alter the ba-sic nature of our results. Furthermore, a formal blindingstrategy might have distorted the clinical structure of theunit and reduced the generalizability of the findings toroutine clinical settings.

The purpose of this study was to demonstrate the abil-ity of the BPRS to quantify numerically what may seemapparent intuitively. Clinical intuition that a patient hassomehow “improved” is no longer sufficient for maintain-ing funding sources needed for inpatient care. Our searchwas for a standard, relatively user-friendly instrumentsuch as the 18-item anchored BPRS that would providereasonable quantitative outcome support within the briefperiods of hospitalization now demanded. We feel thatthis study shows the BPRS to have great promise for ful-filling that need. Ultimate value of the subscales com-pared with total scores for some patients grouped by diag-nosis awaits correlation with various parameters in largerclinical samples, perhaps using other recent versions ofthis well-known instrument.

REFERENCES

1. Berman WH, Rosen CS, Hurt SW, et al. Toto, we’re not in Kansas any-more: measuring and using outcomes in behavioral health care. ClinPsychol Sci Pract 1998;5:115–133

2. Lieberman PB, Wiitala SA, Elliott B, et al. Decreasing length of stay: arethere effects on outcomes of psychiatric hospitalization? Am J Psychiatry1998;155:905–909

3. Overall JE, Gorham DR. The Brief Psychiatric Rating Scale. Psychol Rep1962;10:799–812

4. Overall JE. The Brief Psychiatric Rating Scale in psychopharmacology re-search. In: Pichot P, ed. Psychological Measurements in Psychopharma-cology. Modern Problems in Pharmacopsychiatry, vol 7. Basel, Switzer-land: Karger; 1974:67–78

5. Hedlund JL, Vieweg BW. The Brief Psychiatric Rating Scale (BPRS):a comprehensive review. J Oper Psychiatry 1980;11:48–65

6. Inch R, Crossley M, Keegan D, et al. Use of the Brief Psychiatric RatingScale to measure success in a psychosocial day program. Psychiatr Serv1997;48:1195–1197

7. Packer S, Husted J, Cohen S, et al. Psychopathology and quality of life inschizophrenia. J Psychiatry Neurosci 1997;22:231–234

8. Silverstein ML, Mavrolefteros G, Close D. BPRS syndrome scales duringthe course of an episode of psychiatric illness. J Clin Psychol 1997;53:455–458

9. Burger GK, Calsyn RJ, Morse GA, et al. Factor structure of the expandedBrief Psychiatric Rating Scale. J Clin Psychol 1997;53:451–454

10. Velligan DI, Mahurin RK, Diamond PL, et al. The functional significanceof symptomatology and cognitive function in schizophrenia. SchizophrRes 1997;25:21–31

11. Imbornoni S, Donenwirth K, Orr B, et al. Objective measurement of psy-chiatric symptoms: a quality improvement process for inpatient care.Jt Comm J Qual Improv 1997;23:183–195

12. Swett C. Symptom severity and number of previous psychiatric admis-sions as predictors of readmission. Psychiatr Serv 1995;46:482–485

13. Priebe S, Gruyters T. Patients’ and caregivers’ initial assessments of day-hospital treatment and course of symptoms. Comp Psychiatry 1994;35:234–238

14. Carmin CN, Ownby RL. The relationship between discharge readiness in-ventory scales and the Brief Psychiatric Rating Scale. Hosp CommunityPsychiatry 1995;45:248–252

15. Young AS, Nuechterlein KH, Mintz J, et al. Suicidal ideation and suicideattempts in recent-onset schizophrenia. Schizophr Bull 1998;24:629–634

16. Acorn S. Use of the Brief Psychiatric Rating Scale by nurses. J PsychosocNurs Ment Health Serv 1993;31:9–12

17. Woerner MG, Mannuzza S, Kane JM. Anchoring the BPRS: an aid toimproved reliability. Psychopharmacol Bull 1988;24:112–117

18. Garver DL, Steinberg JL, McDermott BE, et al. Etiologic heterogeneity ofthe psychoses: is there a dopamine psychosis? Neuropsychopharmacology1997;16:191–201

19. Harvey PD, Davidson M, White L, et al. Empirical evaluation of the facto-rial structure of clinical symptoms in schizophrenia: effects of typicalneuroleptics on the Brief Psychiatric Rating Scale. Biol Psychiatry 1996;40:755–760

421

https://www.researchgate.net/publication/15570821_Symptom_severity_and_number_of_previous_psychiatric_admissions_as_predictors_of_readmission?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15570821_Symptom_severity_and_number_of_previous_psychiatric_admissions_as_predictors_of_readmission?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15154362_Patients'_and_caregivers'_initial_assessments_of_day-hospital_treatment_and_course_of_symptoms?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/14844535_Use_of_the_brief_psychiatric_rating_scale_by_nurses?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14844535_Use_of_the_brief_psychiatric_rating_scale_by_nurses?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13433576_Suicidal_Ideation_and_Suicide_Attempts_in_Recent-Onset_Schizophrenia?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13433576_Suicidal_Ideation_and_Suicide_Attempts_in_Recent-Onset_Schizophrenia?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13963428_BPRS_Syndrome_scales_during_the_course_of_an_episode_of_psychiatric_illness?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/13958282_Psychopathology_and_Quality_of_Life_in_Schizophrenia?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13958282_Psychopathology_and_Quality_of_Life_in_Schizophrenia?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/19767848_Anchoring_the_BPRS_An_Aid_to_Improved_Reliability''?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/19767848_Anchoring_the_BPRS_An_Aid_to_Improved_Reliability''?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13934950_Use_of_the_Brief_Psychiatric_Rating_Scale_to_measure_success_in_a_psychosocial_day_program?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/270749496_The_Brief_Psychiatric_Rating-Scale?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/270749496_The_Brief_Psychiatric_Rating-Scale?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14075058_Objective_measurement_of_psychiatric_symptoms_a_quality_improvement_process_for_in-patient_care?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14075058_Objective_measurement_of_psychiatric_symptoms_a_quality_improvement_process_for_in-patient_care?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15011461_The_Relationship_Between_Discharge_Readiness_Inventory_Scales_and_the_Brief_Psychiatric_Rating_Scale?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/232477427_The_Brief_Psychiatric_Rating_Scale_in_psychopharmacology_research?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1





Early- vs. Late-Onset Panic Attacks

423



he literature on panic attacks in the elderly is lim-ited and focuses on late-onset panic attacks. In fact,

A Comparison of Persons WithEarly- Versus Late-Onset Panic Attacks

David A. Katerndahl, M.D., M.A., and Melissa Talamantes, M.S.

Background: Although some evidence sug-gests a bimodal age at onset of panic attacks, theliterature comparing subjects with early versus lateonset is limited. Previous work suggests that peo-ple with late-onset panic attacks may have fewerpanic symptoms and exhibit less avoidance. Thisstudy sought to compare late-onset panic attacksand early-onset panic attacks with regard to(1) comorbidity, (2) health care utilization,and (3) illness behaviors and coping.

Method: This community-based study involvedinterviewing randomly selected adults for the pres-ence of DSM-III-R panic attacks. If panic attackswere confirmed, subjects were asked questionsconcerning panic characteristics, psychiatric co-morbidity, symptom perceptions, illness attitudes,coping, and family characteristics. Subjects report-ing early-onset panic (panic onset < 50 years ofage) were compared with those reporting late onset(onset ≥ 50 years of age). Significant univariateanalyses were controlled for differences in age,panic duration, and socioeconomic status by usinganalysis of covariance and logistic regression.

Results: Subjects with late-onset panic attacks(N = 9) utilized the mental health sector less, butwere more likely to present to family physiciansfor their worst panic. Patients with late-onset panicfelt that choking and numbness more strongly dis-rupted function, but felt less strongly that eitherdepersonalization or sweating disrupted function.Subjects with late-onset had fewer comorbid con-ditions and lower Symptom Checklist-90 scores.Late-onset subjects also had less hypochondriasisand thanatophobia while coping less throughavoidance or wishful thinking.

Conclusion: Late-onset panic attacks areassociated with less mental health utilization,lower levels of comorbidity, less hypochondriasis,and a greater number of positive coping behaviors.


Received April 1, 1999; accepted Nov. 19, 1999. From the Departmentof Family Practice, University of Texas Health Science Center at SanAntonio.

Supported by a grant from the Upjohn Company, Kalamazoo, Mich.(Dr. Katerndahl).

Reprint requests to: David A. Katerndahl, M.D., M.A., Department ofFamily Practice, University of Texas Health Science Center at San Antonio,7703 Floyd Curl Dr., San Antonio, TX 78284-7795.

Tis the late-onset distinction a valid one? Although studiesof late- versus early-onset panic attacks have used differ-ent definitions of late onset,1–3 there is some suggestion ofa bimodal distribution in age at onset. In an unpublishedcommunity-based study, Marron et al.4 found the nadir inonset between the ages of 40 and 44 years after a peak inyoung adulthood and a second peak at age 55 to 59 years.Sheehan et al.5 found, in patients from a university psy-chiatry clinic, a discontinuity in rates of onset betweenages 37 and 44 years with 8% reporting onset after age 44years. Using hazard rates, Burke et al.6 found that onsetrates in the community leveled off at 45 to 49 years old.The observation in these studies of discontinuity in onsetsuggests that there may indeed be a discrete group of sub-jects with panic of late onset.

Except for case series, studies comparing late-onsetand early-onset panic are limited and focus on panic dis-order instead of panic attacks, and their criteria for late on-set differ. However, some early- versus late-onset differ-ences have been found. One pilot study that compared thephenomenology of early-onset with late-onset panic at-tacks found that older volunteers from the communitywho reported onset of panic attacks after age 54 years ex-hibited fewer panic symptoms. In addition, the older re-spondents exhibited less avoidance behavior.1 Differencesbetween late-onset and early-onset panic disorder re-vealed that the late-onset group of panic disorder patientswere more likely to be widowed.2 Raj et al.3 found that60% of elders with panic disorder in a geriatric psychiatryservice reported onset after age 60 years. No differenceswere noted in the prevalences of depression, agoraphobia,or illness phobia. Similarly, 13 elders with panic disorderbeginning after age 60 years seen in a department of psy-chiatry clinic had clinical and demographic profiles simi-lar to those with early onset.7 Thus, criteria for late-onsetpanic differ across studies, but early- versus late-onset dif-ferences have sometimes been found.

Although some differences have been reported, thecharacteristics compared to date are limited, are oftenconflicting, and use arbitrary criteria to define late-onsetpanic. Hassan and Pollard7 noted the need for research todetermine whether early and late onset in panic disordersubjects are clinically distinct entities, to explain why

422


Katerndahl and Talamantes



panic onset is delayed until older adulthood in late-onsetpanic, and to determine whether treatment approachesshould be different in late-onset versus early-onset panicdisorder. The purpose of the present community-basedstudy was to (1) compare the severity of and comorbidityin panic attacks between late- and early-onset respon-dents, (2) evaluate differences between the 2 groups withregard to their presentation and utilization of health ser-vices for panic attacks, and (3) compare the differencesin illness behavior and coping strategies between early-onset and late-onset panic patients.

METHOD

The Panic Attack Care-Seeking Threshold (PACT)study was conducted in San Antonio, Tex., utilizing meth-ods similar to those of the Epidemiologic Catchment Areastudy.8 Eighteen census tracts were used; the number ofsubjects interviewed in each tract differed so that thesample would be representative of the U.S. population inage, gender, and race. Because Hispanics form a majorityin the San Antonio area, ethnic representativeness was notattempted. After the identification of randomly selectedstarting points within each tract, clusters of 3 dwellingswere selected at intervals of 8. Individuals within thedwelling were randomly selected using the Kish method.Details of the methodology used have been presentedelsewhere.9

The initial screening consisted of a structured inter-view using the panic disorder portion of the StructuredClinical Interview of the DSM-III-R (SCID),10 as well asa demographic information sheet, administered by 1 of 2trained research assistants. The bilingual lay research as-sistants were trained in subject recruitment, instrumentadministration, and application of DSM-III-R criteria.Using volunteers, the research assistants administered theinstruments during the supervised training period and in-terrater reliability comparing the research assistants wasassessed using κ statistics and intraclass correlations.

Those subjects meeting DSM-III-R criteria for panic at-tacks were asked to participate in an extensive interviewregarding their access to health care, health care utiliza-tion, panic characteristics, coexisting psychiatric problemsusing the SCID10 and Symptom Checklist-90 (SCL-90),11

symptom perceptions using the Symptom PerceptionScales,12 illness attitudes and behaviors using the IllnessAttitude Scales13 and Illness Behavior Questionnaire,14

family variables using the Family Inventory of LifeEvents15 and the Duke Social Support and Stress Scale,16

and coping using the Ways of Coping Checklist.17 Subjectswere classified as having panic attacks if they hadever experienced unexpected and unexplained discreteperiods of intense fear or discomfort consisting of at least4 panic symptoms. Subjects with panic attacks thus in-cluded those with attacks frequent enough to meet

DSM-III-R criteria for panic disorder (i.e., 4 attacks in 4weeks or at least 1 attack with persistent fear), as well aspersons with less frequent attacks. Of the 1683 individu-als contacted, 1266 (75%) agreed to be screened. Refusalto be screened was not related to age, gender, socioeco-nomic status, income, education, or language, but washighest in non-Hispanic whites (26.3%) and lowest in His-panics (12.2%). Of the 1266 screened individuals, 119(9%) met criteria for panic attacks, and 97 agreed toparticipate in the long interview. Those agreeing to com-plete the long interview did not differ from those refusingin terms of language, ethnicity, gender, socioeconomicstatus, income, education, duration of panic, or phobicavoidance.

Using the change point test, we found a significant(D = 0.680, p ≤ .001) cut point in subjects reporting onsetof panic after the age of 46 years; no subjects reportedonset between ages 46 and 50. Subjects were consideredto have late-onset panic attacks if their first attack oc-curred at 50 years or older. For this analysis, 9 subjectswere identified as having late-onset panic attacks.

Univariate analyses were conducted using chi-squarewith Yates correction for discontinuity and t testing. Be-cause the 2 groups differed in age, socioeconomic status,and duration of panic, significant univariate analyseswere reanalyzed using logistic regression and analysis ofcovariance (ANCOVA), adjusting for age, socioeconomicstatus, and panic duration when these variables correlatedwith the dependent measure. A p value ≤ .05 was deemedsignificant. Owing to the large number of univariateanalyses, a Bonferroni correction was also calculated(p ≤ .00023) to adjust for multiple testing.

RESULTS

Demographic data of the subjects participating in thisstudy are presented in Table 1. The overall prevalence ofpanic attacks in persons 50 years and older was 5.8%, withequal numbers in the early-onset panic attack and late-onset panic attack groups. Two late-onset subjects met cri-teria for panic disorder (22%) as opposed to 45% (N = 40)in the early-onset group. Among all 97 patients with panicattacks, the most common chronic problems reported werehypertension (N = 10), arthritis (N = 9), diabetes (N = 6),heart disease (N = 6), and allergies (N = 5). No subjectsreported Parkinson’s disease or a history of cerebrovascu-lar accident, and only 3 reported chronic lung disease(asthma).

Table 2 provides the total health care utilization over a2-month period in general. The early-onset group pre-sented to more health care sites in general (t = 2.62,p = .012) and more mental health sites in particular(t = 4.74, p = .0001) than did the late-onset group. Over a2-month period, the early-onset group had higher utiliza-tion rates in the mental health sector (t = 3.24, p = .0002).

423



425



As opposed to the early-onset group, none of the late-onset group presented to a mental health clinic for panicattacks (Fisher exact, p = .042). None of the utilizationvariables in Table 2 correlated with duration, age, or so-cioeconomic status. At the time of their worst attacks,67% (N = 6) of the late-onset respondents considered us-ing a family physician as opposed to 31% (N = 27) of theearly-onset group (χ2 = 4.02, p = .045). More late-onsetrespondents actually used a family physician than did theearly-onset respondents: 33% (N = 3) versus 10% (N = 9;χ2 = 4.02, p = .045).

Table 3 reveals the comparisons of lifetime comor-bidity between the early-onset and late-onset groups. Sub-jects with “phobic avoidance” avoided situations becauseof their panic attacks and included those with agoraphobia

as well as less severe avoidance. Age, socioeconomicstatus, and duration of panic were not related to the pres-ence of specific psychiatric disorders. The total number ofcomorbid psychiatric conditions was greater in the early-onset patients (early-onset mean = 3.75 disorders vs. late-onset mean = 1.56 disorders; t = 4.23, p = .0001). Theonly SCL-90 scale that failed to reach statistical signifi-cance was the somatization scale. In each case, the early-onset group scored higher on each of the scales (Table 4).The late-onset respondents had a shorter duration of panicattacks (mean ± SD = 4.5 ± 5.89 years) versus the early-onset respondents (mean ± SD = 11.84 ± 12.02 years;t = 3.10, p = .007).

The late-onset group experienced less severepanic symptomatology in sweating (ANCOVA F = 5.45,p = .0058). The remainder of symptom severity analyseswere not significant. The symptom perceptions in Table 5describe the patients’ perception of the severity of symp-toms that required treatment or were disruptive to func-

Table 1. Demographic Characteristics of 97 SubjectsWith Either Early- or Late-Onset Panic Attacksa

EOP Group LOP GroupCharacteristic (N = 88) (N = 9)

Age, mean ± SD, y 36.9 ± 11.2b,c 68.6 ± 11.9Gender, female, N (%) 71 (80.7) 5 (55.6)Ethnicity/race, N (%)

Hispanic 49 (55.7) 5 (55.6)White 26 (29.6) 3 (33.3)Black 13 (14.8) 1 (11.1)

Marital status, N (%)Single 25 (28.4) 1 (11.1)Married 42 (47.7) 6 (66.7)Marital lossd 21 (23.9) 2 (22.2)

Hollingshead socioeconomicstatus, mean ± SD 60.9 ± 19.2e 78.1 ± 14.4

No. of chronic illnesses,mean ± SD 1.2 ± 1.3 2.0 ± 1.5

aAbbreviations: EOP = early-onset panic attacks, LOP = late-onsetpanic attacks.bt = –8.06, p < .0001.cSignificant by Bonferroni correction (p < .00023).dMarital loss = divorced, separated, or widowed.et = –2.62, p = .0103.

Table 2. Health Care Utilization During 2-Month Perioda

Total Utilizationb

EOP Group LOP Group(N = 88) (N = 9) Statistics

Site Mean SD Mean SD t p

Hospital emergencyroom 0.18 0.56 0.33 0.50 NS

Minor emergencyroom 0.06 0.35 0.11 0.33 NS

Clinic 0.16 0.58 0.00 2.55 .0125Physician’s office

Family physician 0.58 1.54 1.00 1.32 NSCardiologist 0.14 0.51 0.11 0.33 NSInternist 0.06 0.23 0.11 0.33 NSPsychiatrist 0.18 0.60 0.00 2.85 .0054

Psychologist 0.18 0.95 0.00 NSSocial worker 0.06 0.53 0.00 NSAmbulance 0.06 0.28 0.11 0.33 NSaAbbreviations: EOP = early-onset panic attacks, LOP = late-onsetpanic attacks.bUtilization calculated as number of visits per 2-month period.

Table 3. Comparison of Comorbidity in SubjectsWith Late- and Early-Onset Panic Attacksa

EOP Group LOP Group(N = 88) (N = 9) Statistics

Disorder N % N % χ2 p

Substance abuseAlcohol 30 34.1 1 11.1 NSDrugs 16 18.2 0 0 NS

Phobic avoidance 35 39.8 1 11.1 NSDepression 60 68.2 4 44.4 NSObsessive-compulsive

disorder 52 59.1 2 22.2 NSSocial phobia 40 45.5 2 22.2 NSSimple phobia 57 64.8 1 11.1 9.78 .002Generalized anxiety

disorder 69 78.4 4 44.4 4.83 .028aAbbreviations: EOP = early-onset panic attacks, LOP = late-onsetpanic attacks.

Table 4. Psychiatric Symptomatology in SubjectsWith Late- and Early-Onset Panic Attacksa

Mean Scores Statistics

EOP Group LOP GroupSCL-90 Scale (N = 88) (N = 9) t or (F) p

Somatization 1.99 1.68 NSObsessive-

compulsive 2.02 1.56 2.35 .030Interpersonal

sensitivity 1.83 0.62 6.92 .0001b

Depression 2.10 1.19 2.60 .011Anger 1.59 0.74 4.20 .0005Phobic

anxiety 1.69 0.60 (10.41) .0001c

Paranoia 1.77 0.79 6.03 .0001b

Psychosis 1.36 0.55 4.67 .0002b

Anxiety 2.13 1.12 (6.18) .003c

aAbbreviations: EOP = early-onset panic attacks, LOP = late-onsetpanic attacks, SCL-90 = Symptom Checklist-90.bSignificant by Bonferroni correction (p ≤ .00023).cAnalysis of covariance.

424




One personal copy may be printedtioning. In addition, the early-onset group was more likelyto believe that hot or cold flashes were severe symptoms(t = 2.18, p = .023), but indicated that chest pain was aless severe symptom (t = 2.60, p = .013). The illness be-haviors and coping mechanisms exhibited by both theearly-onset and late-onset groups can be seen in Table 5. Areview of the behaviors and attitudes shows that the late-onset respondents were more likely to report that thanato-phobia was not as serious a concern than the early-onsetrespondents. The late-onset respondents were also lesslikely to engage in negative coping mechanisms such aswishful thinking and avoidance.

DISCUSSION

This study found that whereas patients with early-onset panic attacks use psychiatrists more, late-onsetpatients use the general health sector. The high utilizationof the general health sector may reflect differences insymptom perceptions, medical morbidity, or likelihood ofhaving a personal physician. Because late-onset patients

perceived that somatic symptoms are more likely to bedisruptive to functioning, they may tend to seek help fromthe general health sector. Conversely, because the early-onset patients perceive that psychological symptoms(feelings of unreality) are more likely to need treatmentand are disruptive to functioning, they tend to use themental health sector. In addition, the late-onset group re-ported having more chronic illnesses than the early-onsetgroup, suggesting that they may tend to attribute their so-matic symptoms to one or more of their chronic illnesses.This finding agrees with previous studies suggesting apreference for the general health sector by the elderly.18

The severity of panic symptoms as measured by sever-ity of sweating during attacks and psychiatric comorbidityare greater in the early-onset panic attack group. Thesefindings generally agree with the previous studies on late-onset panic disorder.1,3 This may explain the higher men-tal health utilization in our early-onset group.

If there is a second peak in panic onset late in life, whymight this happen? Previous work suggests that the onsetof panic disorder in general is triggered by an object loss,19

stimulant use,20 or hormonal imbalance.19 Although thenumber of life events prior to panic onset was not in-creased, those events are more likely to be personal innature and viewed more negatively.21 The age at onset cor-related with the number of negative events and ratings ofpositive events.22 Specific events may include physical ill-ness,7,22,23 surgery,3 diagnosis of cancer,3 financial loss,3,7

or loss/illness of spouse.3,7 Late-onset (≥ 65 years old) ago-raphobia is more frequently precipitated by a threat to thepatient’s life than by bereavement or illness.24 Because thelate-onset respondents in this study had fewer life eventsat the time of their worst attack, were not widowed, andreported lower levels of family stress, an object loss (i.e.,death of a spouse, retirement, loss of friends) did not ex-plain the delayed onset of panic attacks in these patients.

Similarly, the lower prevalence of substance abuse inthe late-onset group suggests that stimulant use is notlikely to be the precipitant event. However, medicationscould trigger panic onset. The problem of polypharmacyin the elderly raises further possibilities for medication-related induction of panic attacks. Because neurolepticsare frequently used in the elderly and are known to exac-erbate panic symptoms, they may be a precipitating factorfor panic attacks. This study did not collect data concern-ing medication use.

Concerning hormonal causes, the female subjects inthe late-onset group could be expected to be at high riskfor hormonal imbalances such as postmenopausal symp-toms and osteoporosis. However, these disorders were notreported by women in this study.

Previous work suggests that most elders with anxietyhave psychological causes, but a variety of metabolic,neurologic, cardiac, and pharmacologic disorders can beresponsible.25 Because the locus ceruleus is an important

Table 5. Significant Differences in Symptom Perceptions,Illness Behaviors, and Coping Independent of Age,Socioeconomic Status, and Duration (N = 97)a

Mean Score

EOP LOPItems/ Group Group Statistics

Measure Scale (N = 88) (N = 9) t or (F) p

Symptom perceptions(SPS) (24 scales)b

Unreality needstreatment 1 3.58 5.67 2.34 .0216

Numbness disruptsfunctionality 1 3.25 1.56 2.11 .0372

Choking disruptsfunctionality 1 2.27 1.22 3.33 .0023

Unreality disruptsfunctionality 1 3.42 6.00 2.94 .0041

Sweating disruptsfunctionality 1 4.41 6.11 2.01 .0477

Illness behaviorsand attitudes(17 scales)General

hypochondriasis(IBQc) 9 4.33 1.44 (5.39) .0061d

Whiteleyhypochondriasis(IBQc) 14 7.51 4.78 2.15 .0341

Thanatophobia(IASc) 3 6.08 3.11 2.05 .0436

Coping style (WCCLc)(8 scales)Wishful thinking 8 2.35 1.75 2.08 .0430Avoidance 10 1.64 1.25 2.28 .0248

aAbbreviations: EOP = early-onset panic attacks, IAS = IllnessAttitude Scales, IBQ = Illness Behavior Questionnaire,LOP = late-onset panic attacks, SPS = Symptom Perception Scale,WCCL = Ways of Coping Checklist.bScore of 1 = most severe, 7 = least severe.cHigher score signifies greater severity.dAnalysis of covariance.

425



427



center in several neurotransmitter pathways and may beimportant in panic disorder, brain aging processes couldbe related to late-onset panic despite decreased activity inthe locus ceruleus. The onset of panic disorder and agora-phobia in elders has been linked to Parkinson’s disease,26

stroke recovery,27 and chronic obstructive pulmonary dis-ease,28 but none of these conditions were prevalent in thisstudy. Consequently, this study cannot explain the etiol-ogy of the late-onset attacks in these subjects.

This study revealed that, in general, patients with late-onset panic attacks exhibited more positive illness behav-iors, attitudes, and coping strategies than those with earlyonset. Although previous work failed to find a late- versusearly-onset difference in illness phobia,3 elder subjectswith phobias report more hypochondriasis and higher lev-els of worry, anxiety, and social unease than controls.23

The fact that symptoms are less severe and comorbidityless frequent in late-onset subjects may explain their posi-tive attitudes. However, conversely, the positive attitudesand coping in the late-onset subjects may explain thelower levels of symptom severity and psychiatric comor-bidity. In addition, these lower levels may in themselvessuggest why panic attacks did not occur earlier in thesepatients. Previous studies have noted an association be-tween positive coping strategies and lower levels of anxi-ety, depression, and phobias in panic disorder patients,whereas negative strategies are associated with higherlevels.29 Coping strategies may lessen or exacerbate theeffects of life events.28

In general, the results of this study agree with thosepredicted by the developmental diatheses-stress model.30

As biological vulnerability in terms of physical illnessand changes in neurotransmitter systems increases withage, psychological vulnerability decreases as copingskills and positive illness attitudes improve. However,stressful life events decrease in mid-life, but generallyrise later in life. Other than drop in socioeconomic status,increased life events were not observed in this study.Hence, early-onset subjects may develop panic in re-sponse to stressful life events during a time of psychologi-cal vulnerability. On the other hand, late-onset panic at-tacks may develop in response to increasing life eventsand biological vulnerability despite diminished psycho-logical vulnerability.

This study has several limitations. Because of the largeportion of Hispanics in our sample, the results may not begeneralizable to other populations. Comparison with pre-vious studies is problematic because the study popula-tions differ. This study focused on panic attacks and wasnot limited to those with panic disorder. However, most ofthe previous literature focuses on panic disorder. Thismay limit conclusions about comorbidity, hypochondria-sis, poor coping strategies, etc., seen in panic disorderpopulations. Questions concerning panic onset and healthcare utilization rely on patient recall, which may be bi-

ased. In addition, the number of univariate analyses con-ducted will increase the alpha level. Finally, the paucityof subjects with late-onset panic attacks lowers the statis-tical power of these analyses and may lead to a type II er-ror. Although we have attempted to determine whetherobserved differences were due to the differences in age,duration of panic, or socioeconomic status, early-late dif-ferences could still be due to these factors.

In conclusion, late-onset and early-onset subjects dif-fered in psychiatric comorbidity, illness behaviors andcoping, and health care utilization. This study should beinterpreted with caution, but suggests that late-onsetpanic attacks represent a less severe disorder and mayrepresent a distinct syndrome. There is clearly a need forfurther research on the etiology of panic attacks and panicdisorder of late onset to improve our understanding of thisphenomenon.

REFERENCES

1. Sheikh JI, King RJ, Taylor CB. Comparative phenomenology of early-onset versus late-onset panic attacks: a pilot study. Am J Psychiatry 1991;148:1231–1233

2. Beitman BD, Kushner MA. Late-onset panic disorder: evidence from astudy of patients with chest pain and normal cardiac functions. Int JPsychiatr Med 1991;21:29–35

3. Raj BA, Corvea MH, Dagon EM. Clinical characteristics of panic disorderin the elderly. J Clin Psychiatry 1993;54:150–155

4. Marron JT, Isklander TN, Kohen JC. Community based survey to deter-mine the prevalence of panic attacks. Presented at the North AmericanPrimary Care Research Group meeting; 1985; Seattle, Wash

5. Sheehan DV, Sheehan KE, Minichiello WE. Age of onset of phobic disor-ders. Compr Psychiatry 1981;22:544–553

6. Burke KC, Burke JD Jr, Regier DA, et al. Age of onset of selected mentaldisorders. Arch Gen Psychiatry 1990;47:511–528

7. Hassan R, Pollard CA. Late-life-onset panic disorder. J Geriatr PsychiatryNeurol 1994;7:86–90

8. Eaton WW, Holzer CE III, Von Korff M, et al. Design of the Epidemio-logic Catchment Area surveys: the control and measurement of error. ArchGen Psychiatry 1984;41:942–948

9. Realini J, Katerndahl D. Factors affecting threshold for seeking care. J AmBoard Fam Pract 1993;6:215–223

10. Spitzer RL, Williams JBW. Structured Clinical Interview for DSM-III-R,Upjohn Version, Revised. New York, NY: New York State PsychiatricInstitute; 1989

11. Derogatis LR, Lipman RS, Rickels K, et al. The Hopkins SymptomChecklist (HSCL): a self-report symptom inventory. Behav Sci 1974;19:1–15

12. Jones RA, Wiese HJ, Moore RW, et al. On the perceived meaning ofsymptoms. Med Care 1981;19:710–717

13. Kellner R. Illness Attitude Scales. Albuquerque, NM: University of NewMexico; 1983

14. Pilowski I, Spence ND. Manual for the Illness Behavior Questionnaire.Adelaide, South Australia: University of Adelaide; 1981

15. McCubbin HI, Patterson JM, Wilaon L. Family Inventory of Life Eventsand Changes. St. Paul, Minn: Department of Family Social Science, Uni-versity of Minnesota; 1990

16. Duke Social Support and Stress Scale. Durham, NC: Duke University;1986

17. Vitaliano PP. Manual for Appraisal Dimensions Scale and Revised Waysof Coping Checklist. Seattle, Wash: Department of Psychiatry, Universityof Washington; 1985

18. Buller R, Winter P, Amering M, et al. Center differences and cross-national invariance in help-seeking for panic disorder. Soc PsychiatryPsychiatr Epidemiol 1992;27:135–141

19. Klein DF. Anxiety reconceptualized. In: Klein DF, Rabking JG, eds.Anxiety. New York, NY: Raven Press; 1981

426

https://www.researchgate.net/publication/21090285_Late_Onset_Panic_Disorder_Evidence_from_a_Study_of_Patients_with_Chest_Pain_and_Normal_Cardiac_Evaluations?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/15933220_On_the_Perceived_Meaning_of_Symptoms?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15933220_On_the_Perceived_Meaning_of_Symptoms?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/283440357_Family_inventory_of_life_events_and_changes?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/14711909_The_clinical_characteristics_of_panic_disorder_in_the_elderly_A_retrospective_study?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14711909_The_clinical_characteristics_of_panic_disorder_in_the_elderly_A_retrospective_study?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14696610_Factors_affecting_the_threshold_for_seeking_care_The_Panic_Attack_Care-Seeking_Threshold_PACT_Study?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14696610_Factors_affecting_the_threshold_for_seeking_care_The_Panic_Attack_Care-Seeking_Threshold_PACT_Study?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15890570_Age_of_Onset_of_Phobic_Disorders_A_Reevaluation?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15890570_Age_of_Onset_of_Phobic_Disorders_A_Reevaluation?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1





20. Aronson TA, Craig TJ. Cocaine precipitation of panic disorder. Am JPsychiatry 1986;143:643–645

21. Roy-Byrne PP, Geraci M, Uhde TW. Life events and the onset of panicdisorder. Am J Psychiatry 1986;43:1424–1427

22. Lteif GN, Mavissakalian MR. Life events and panic disorder/agoraphobia.Compr Psychiatry 1995;36:118–122

23. Lindesay J. Phobic disorders in the elderly. Br J Psychiatry 1991;159:531–541

24. Livingston G, Warkin V, Milne B, et al. Natural history of depression andanxiety disorders in older people. J Affect Disord 1997;46:255–262

25. Schaffer CB, Donlon PT. Medical causes of psychiatric symptoms in theelderly. Clin Gerontologist 1983;1:3–18

26. Stein MB, Heuser IJ, Juncos JL, et al. Anxiety disorders in patients withParkinson’s disease. Am J Psychiatry 1990;147:217–220

27. Burvill PW, Johnson GA, Jamrozik KD, et al. Anxiety disorders afterstroke. Br J Psychiatry 1995;166:328–332

28. Karajgi B, Rifkin A, Doddi S, et al. The prevalence of anxiety disordersin patients with chronic obstructive pulmonary disease. Am J Psychiatry1990;147:200–201

29. Kendler KS, Keissler RC, Heath AC, et al. Coping. Physiol Med 1991;21:337–346

30. Gatz M, Kasl-Godley JE, Kavel MJ. Aging and mental disorders. In:Birren JE, Schaie KW, eds. Handbook of the Psychology of Aging. 4th ed.New York, NY: Academic Press; 1996:365–382

427


Abuse and Paroxetine Binding in Bulimia

429



ulimia nervosa is defined by dietary dyscontrol andbodily concerns, but is generally a polysympto-

Background: Co-occurrence of bulimia ner-vosa and borderline personality disorder has beenattributed to shared factors, including childhoodabuse and disturbances in central serotonin(5-hydroxytryptamine; 5-HT) mechanisms. Toexplore this notion, we conducted a controlledassessment of childhood abuse and 5-HT functionin bulimics with and without borderline personal-ity disorder.

Method: Forty patients with bulimia nervosa,confirmed with the Eating Disorders Examinationinterview (14 with borderline personality disorderand 26 without), and 25 normal-eater controlswere assessed for clinical symptoms (eating dis-turbances, mood lability, impulsivity, and disso-ciation) and childhood sexual and physical abuse.We also conducted tests of platelet tritiated-paroxetine binding in blood samples from 27 ofthe bulimics (11 with borderline personality dis-order and 16 without) and 16 of the controls.

Results: Relative to normal eaters, bulimicsshowed greater affective instability, overall im-pulsivity, and a history of physical abuse. How-ever, borderline bulimics alone showed elevatedmotor impulsivity, dissociation, and rates ofsexual abuse. Paroxetine-binding tests indicatedno differences attributable to comorbid borderlinepersonality disorder, instead linking bulimianervosa with or without borderline personalitydisorder to substantially reduced 5-HT transporterdensity.

Conclusion: Results suggest relatively au-tonomous pathologic entities: one, relevant tobulimia nervosa, being associated with abnormal5-HT transporter function and affective instabil-ity, but relatively independent of childhood sexualabuse; another, relevant to borderline personalitydisorder, onto which sexual abuse, dissociativesymptoms, and behavioral impulsivity converge.We propose that abnormal 5-HT function may,however, constitute one basis for the frequentco-occurrence of bulimic and borderlinedisturbances.


Childhood Abuse andPlatelet Tritiated-Paroxetine Binding in Bulimia Nervosa:

Implications of Borderline Personality Disorder

Howard Steiger, Ph.D.; Stéphanie Léonard, M.Sc.; N. M. K. Ng Ying Kin, Ph.D.;Cécile Ladouceur, B.A.; Dhunraj Ramdoyal, M.Sc.; and Simon N. Young, Ph.D.

Received Feb. 24, 1999; accepted Nov. 8, 1999. From the EatingDisorders Program, Douglas Hospital, Verdun (Dr. Steiger and Mss.Léonard and Ladouceur); the Psychiatry Department, McGill University,Montreal (Drs. Steiger, Ng, and Young and Ms. Léonard); and the ResearchCentre, Douglas Hospital, Verdun (Drs. Steiger and Ng and Mr.Ramdoyal), Quebec, Canada.

Some of the data reported here were collected as part of a Masters inPsychiatry thesis prepared by Stéphanie Léonard at McGill University,with additional work conducted on a grant from the Fonds de la Rechercheen Santé du Québec and the Conseil Québécois de la Recherche Sociale(no. RS3019).

Preliminary findings from this study were presented at meetings of theCanadian Council of Neuropsychopharmacology, Montreal, Quebec,Canada, on June 8, 1998, and of the Eating Disorders Research Society,Boston, Mass., on November 7, 1998.

Reprint requests to: Howard Steiger, Ph.D., Eating Disorders Unit,Douglas Hospital, 6875 LaSalle Blvd., Verdun (Quebec) Canada, H4H1R3 (e-mail: [email protected]).

Bmatic syndrome with a strongly characterological flavor.From 20% to 30% of persons with bulimia nervosa are,for example, reported to have borderline personality dis-order,1,2 for which dysregulation of affects, impulsivity,recurrent self-harm, and transient dissociative states arepathognomonic.3 Co-aggregation of bulimia nervosa and“borderline-spectrum” pathology has been attributed toshared factors—thought to explain concurrent dysregu-lation of impulse controls and mood and eating behav-iors1,4,5—and recent attention has focused on (1) child-hood sexual and physical abuse6 and (2) disturbances incentral serotonin (5-hydroxytryptamine; 5-HT) mecha-nisms.4,5 The present study examined the specificity of as-sociation, for bulimic and borderline syndromes, of child-hood abuse and 5-HT disturbances.

Developmental abuse and bulimia nervosa. Studiesindicate 30% to 45% of persons with bulimia nervosareport childhood sexual abuse, and more still, physicalabuse.6 Such associations need not, however, imply abulimia-specific link, given studies (1) reporting height-ened prevalences of childhood abuse in bulimic individu-als showing comorbid personality pathology, and espe-cially borderline personality disorder,7,8 and (2) showinghalf or more of patients with borderline personality disor-der to have a positive history of childhood sexual abuse.9

In light of such findings, the question arises: Is childhood

428


Steiger et al.



abuse associated with bulimia nervosa, or with borderlinepersonality disorder found in only some bulimic patients?The present study addressed this question.

Serotonin dysfunction and bulimia nervosa. Evi-dence shows 5-HT to moderate mood, impulsive behavior,and satiety,4,5 and this creates a rationale for the hypothesisthat central 5-HT mechanisms act in the predisposition to(or perpetuation of) bulimia nervosa. Empirical supportfor this notion has been impressive. Jimerson et al.10 foundhigh-frequency binge eaters (in a normal-weight bulimicsample) to have significantly lower levels of 5-HT me-tabolites in cerebrospinal fluid than did low-frequencybinge eaters or controls. Goldbloom et al.11 reported 22 ac-tive bulimics to have higher platelet 5-HT uptake ratesthan did 20 age-matched controls, and interpreted this toimply an adaptation to reduced 5-HT. Similarly, severalstudies in bulimia nervosa have documented blunted pro-lactin responses to 5-HT agonists or partial agonists4 (im-plying down-regulation at postsynaptic 5-HT sites). Fi-nally, the selective serotonin reuptake inhibitor (SSRI)fluoxetine is found to yield clinically significant reduc-tions in binge-eating episodes.12 While such findings indi-cate association between bulimia nervosa and 5-HTanomalies, they need not imply bulimia-specific effects.Compared with healthy controls, patients with borderlinepersonality disorder also show signs of decreased 5-HTtone, or anomalous hormonal responses to 5-HT agonists,9

and clinical trials show fluoxetine to be effective in treat-ment of dysphoria, impulsivity, and self-mutilation insome patients who have this disorder.13 The possibility ex-ists, therefore, that “borderline” phenomena may accountfor some aspects of the 5-HT anomalies observed in bu-limia nervosa. Our study also addressed this second issue.

Limited data are available that bear upon the implica-tions of borderline features for 5-HT function in bulimianervosa: Verkes and colleagues14 found bulimics with bor-derline personality disorder (N = 5) to show elevatedplatelet 5-HT content relative to bulimics without border-line personality disorder (N = 10) and argued that thismight reflect increased uptake associated with reducedcirculating 5-HT. Likewise, Waller and colleagues15 notedself-reportedly impulsive bulimics in a small (N = 6)sample to show greater blunting of prolactin responsesfollowing buspirone treatment (which they presumed tobe largely a 5-HT1A agonist).

The present study. A first goal in this study was to de-termine whether bulimics with and without borderline per-sonality disorder spanned a continuum of disturbances(with respect to psychiatric symptomatology, childhoodabuse, and 5-HT function) or showed distinct areas of dis-turbance (as might suggest distinct psychopathologicspectra). Another goal of this study was to allow an explo-ration into the association between abuse history and5-HT function. We assessed borderline personality disor-der, childhood abuse, and eating symptoms by structured

interview, and concurrent psychiatric symptoms (affectiveinstability, impulsivity, and dissociation) by question-naire. Serotonin function was assessed by measuring bind-ing, in blood platelets, of the selective 5-HT reuptake in-hibitor [3H]-paroxetine.

There are various reasons for the assumption that plate-let paroxetine binding models central 5-HT transporter(or reuptake) mechanisms16,17: (1) Platelets possess high af-finity-uptake sites for 5-HT, which seem morphologicallyand kinetically comparable with 5-HT reuptake sites inbrain.16,17 (2) Platelet binding is selectively associated withbinding in brain tissue.16 (3) Antidepressant response indepressed outpatients coincides with normalization of 5-HTreuptake inhibitor binding in the periphery.18 (4) Plateletparoxetine binding has been applied as a model of 5-HTfunction in various clinical syndromes.19,20 While ours is (toour knowledge) the first application of paroxetine bindingin bulimia nervosa, Marazziti and colleagues21 have usedplatelet imipramine binding as a model of 5-HT functionand found transporter density (Bmax), but not affinity (Kd),to be reduced in bulimic versus nonbulimic women.

METHOD

ParticipantsBulimic group. Forty women with bulimia nervosa

were recruited through a specialized outpatient service.Eating-disorder status was confirmed at the start of thestudy using the Eating Disorders Examination (EDE) in-terview.22 On the basis of the EDE, 33 (82.5%) womenmet Diagnostic and Statistical Manual for Mental Disor-ders, Fourth Edition (DSM-IV)3 criteria for bulimia ner-vosa, purging subtype; 1 (2.5%) for bulimia nervosa,nonpurging subtype; and 4 (10.0%) for a subclinicalbulimia nervosa purge type (bingeing once versus therequisite twice weekly). According to interviews, our bu-limic participants binged on a mean ± SD of 16.96 ± 7.03days monthly at a frequency of 24.55 ± 14.35 episodesmonthly. Those who vomited did so on a mean of16.74 ± 9.95 days monthly, at a mean frequency of47.17 ± 50.80 times monthly. Mean ± SD age and bodymass index (BMI) in this sample were 26.30 ± 6.19 yearsand 22.01 ± 3.48 kg/m2, respectively.

Normal-eater control. Members of the normal-eatercontrol group (N = 25) were recruited through advertise-ments or university classes and were admitted to the studyif they had no past or present eating disorder upon inter-view and no overt psychiatric history upon inquiry. All de-nied bingeing, purging, or use of psychoactive medica-tions. Mean ± SD age and BMI in this group were20.80 ± 3.69 years and 20.72 ± 1.85 kg/m2, respectively.

MeasuresRating scales. Well-known interviews and question-

naires were selected for demonstrated psychometric

429



431



strengths and relevance to constructs of interest. Weused the EDE22 interview and the Eating Attitudes Test(EAT-26)23 to tap clinical eating-disorder symptoms andBMI to reflect nutritional status. We also measured per-sonality disorders using the Structured Clinical Interviewfor DSM-IV Axis II (SCID-II),24 which we used to clas-sify all patients as either having or not having borderlinepersonality disorder. The borderline personality disordercriterion referring to overeating was excluded. Interraterreliability checks on a subsample of 17 interviews (se-lected pseudorandomly to represent adequate numbers ofprobable “borderline” and “nonborderline” diagnoses)yielded a kappa of 0.68 (representing 88.2% agreement)for a borderline/nonborderline distinction.

Additional psychopathologic characteristics were evalu-ated using the Dissociative Experiences Scale (DES)25; theBarrat Impulsivity Scale (BIS; version 10),26 producingscores measuring cognitive, motor, and nonplanning im-pulsivity; and the affective instability subscale from the Di-mensional Assessment for Personality Pathology-BasicQuestionnaire (DAPP-BQ).27 Finally, to assess childhoodabuse, we used the Childhood Trauma Interview (CTI).28

We used CTI severity and age indices to isolate experiencesinvolving frankly inappropriate sexual or physical contactsoccurring prior to age 13 years and then up to age 18 years.Given a bilingual population, we employed official, vali-dated French translations of the DES and EAT-26 and de-veloped French translations for other scales using carefulforward and back translation techniques. On global indi-ces, translations were psychometrically equivalent to cor-responding English questionnaires.

Paroxetine binding. Blood samples were always drawnbetween 8:30 and 9:00 a.m., after an overnight fast. Par-ticipants were asked to refrain from alcohol or nonpre-scription drug use for 48 hours prior to testing and frombinge eating for 24 hours prior to testing. Whole blood wascollected in Vacutainer tubes containing the anticoagulantEDTA and kept on ice (for no more than 30 minutes) untilplatelets were isolated by differential centrifugation. Plate-let rich plasma was first isolated at 280g for 15 min at 4°C.Platelets were then isolated from the platelet rich plasmaat 18,000g for 15 min. Next, the pellets were washed inbuffer containing EDTA/Tris/NaCl, pH 7.5, and homog-enized using a Polytron (Brinkman Instruments, Roxdale,Ontario, Canada). The lysed membranes were stored in asmall volume of buffer at –80°C until analyzed. Bloodwork was done under blind conditions. The binding ex-periment was performed as described by Langer et al.29

Lysed membranes (0.8 to 2.0 mg protein) were incubatedin a Tris/EDTA/NaCl/KCl buffer containing 0.05 to 10 nMof [3H]-paroxetine (26.5 Ci/mmol [980.5 GBq], NEN [LifeScience Products, Boston, Mass.]) for 90 min at 20°C. Thebound and free ligands were separated by filtration onGF/B Whatman filters, washed 3 times with buffer, andcounted. Specific binding, determined by incubating

[3H]-paroxetine in the presence and absence of an excessamount of citalopram (3 µM), was found to be between70% to 90% of total binding. The apparent Bmax and Kd

were obtained by Scatchard analysis of binding curves forthe different concentrations of [3H]-paroxetine.

ProcedureAll participants provided written informed consent

for research. Measures of psychopathology and child-hood abuse were obtained from all participants, andblood samples from a subset of 27 bulimics (11 with bor-derline personality disorder [BN/BPD] and 16 without[BN/nonBPD]) and 16 normal-eater control (NC) partici-pants. The 5-HT indices thus represented diagnostic clas-sifications well. Potential sources of extraneous variationon 5-HT measures necessitate controls or comment:(1) Contraceptive use: Given reports suggesting absenceof marked effects of oral contraceptives on blood 5-HTindices,30 we did not treat contraceptive use as an exclu-sion criterion. We did, however, test for differences (onparoxetine-binding indices) among individuals who wereor were not taking contraceptives and found no significanteffects. (2) Seasonal effects: Seasonal variations havebeen observed on various 5-HT indices, with studies inhealthy volunteers reported to yield reduced paroxetinebinding in summer/fall.31 Our recruitment of participantswas skewed over time in such a way that any bias due toseasonal variations should have run toward reduced bind-ing in normal controls versus bulimics. Nevertheless, weapplied statistical controls for possible confounds due toseasonal effects, using previously published values31 forseasonal variations in platelet paroxetine binding (see Re-sults). (3) Menses: To optimize sample size, we combinedone group of participants tested during follicular phaseonly with another in whom testing took place on non-menstrual days. We tested for (and ruled out) potentialconfounding effects of menstrual phase on paroxetine-binding findings. (4) Medication: Six cases providingblood samples (5 BN/BPD and 1 BN/nonBPD) hadstarted medication (always an SSRI) at the time of recruit-ment. To optimize sample sizes for data on childhoodabuse, we retained these participants and applied statisti-cal procedures (described below) to rule out confoundsattributable to medication effects. We note, also, that a re-cent report indicates absence of acute effects of variousantidepressants (including paroxetine) upon platelet par-oxetine binding in healthy volunteers.32

RESULTS

Descriptive DataAccording to SCID-II criteria, none of our NC partici-

pants had borderline personality disorder. A more sizablenumber of our bulimic participants met borderline person-ality disorder criteria (N = 14; 35.0%). When a borderline

430


Steiger et al.



personality disorder diagnosis was present,we assigned the participant to the BN/BPDgroup, and when not, to the BN/nonBPDgroup. Mean ± SD age (26.50 ± 6.25 and26.20 ± 6.28 years, respectively) did not dif-ferentiate BN/BPD from BN/nonBPDgroups. Bulimic participants were, however,slightly (and significantly) older than werecontrol participants (F = 7.96, df = 2,62;p < .01). Where the age variable was corre-lated with other indices (affective instability,BIS attention and nonplanning, and Bmax),findings were confirmed using analyses ofcovariance with age as a covariate. BMIyielded no group differences: mean ± SD val-ues across BN/BPD, BN/nonBPD, and NCgroups were 21.75 ± 3.16, 22.15 ± 3.70, and20.72 ± 1.85 kg/m2, respectively.

Eating SymptomsTable 1 shows mean ± SD scores for

BN/BPD, BN/nonBPD, and NC groups onEDE mean monthly binge and vomit indices(the latter values computed for cases whowere vomiters only) and the EAT-26. Resultsof t tests revealed significant borderline/nonborderline differences on mean monthlybinge episodes, borderline patients showingthe higher frequencies. No corresponding dif-ferences were obtained on measures of meandays of bingeing or vomiting per month, or of mean vom-iting episodes per month. On EAT-26 scores, analysis ofvariance (ANOVA) revealed a significant group effect (seeTable 1), Newman-Keuls tests indicating reliable bulimicversus nonbulimic differences but no borderline/non-borderline differences.

Psychiatric SymptomsTable 1 also provides mean ± SD results for BN/BPD,

BN/nonBPD, and NC groups on measures of dissociation(total score), impulsivity (motor, cognitive, nonplanning,and total scores), and affective instability (total score).One-way multivariate ANOVA on the total dissociation,impulsivity, and affective instability scores yielded anomnibus group effect (Wilks lambda = 11.07, df = 6,120;p < .001), and we therefore proceeded to univariateANOVAs. Reliable univariate group effects were obtainedon all but the nonplanning impulsivity variable (see Table1); those on affective instability (F = 20.80, df = 2,61;p < .001) and cognitive impulsivity (F = 24.18, df = 2,61;p < .001) remained after age effects were removedthrough analyses of covariance (ANCOVAs). Nonsignifi-cant results on nonplanning impulsivity were unchangedwhen age effects were removed through ANCOVAs.Group comparisons (Newman-Keuls) showed the follow-

ing: on dissociation and motor impulsivity (arguably themost pathognomonic features of borderline personalitydisorder measured), pathologic elevations occurred inBN/BPD cases, but not in BN/nonBPD cases.

Childhood AbuseTable 2 shows numbers (and proportions) of partici-

pants in each group who reported sexual abuse, physicalabuse, or any abuse (i.e., either form of abuse), both be-fore age 13 years and up to age 18 years. Group effects(or trends) were obtained for data reflecting sexual abuseprior to age 13 (χ2 = 5.22, df = 2, p < .08) and up to age18 (χ2 = 10.22, df = 2, p < .01). Pairwise group compari-sons for prevalences prior to age 13 were (given low fre-quencies in some cells) conducted using Fisher exacttests, and a significant difference was obtained betweenBN/BPD and NC groups (p < .03) alone. Hence, elevatedchildhood sexual abuse seemed to be characteristiclargely of BN/BPD cases and only nonsignificantly el-evated among BN/nonBPD bulimics.

To further explore an apparent association betweensexual abuse and borderline personality disorder, we con-ducted an analysis to reflect associations between eachdiagnostic classification and type of abuse, computingproportions of cases in each group who reported

Table 1. Mean ± SD for Borderline-Bulimic (BN/BPD),Nonborderline-Bulimic (BN/nonBPD), and Normal-Eater Control (NC)Groups on Indices of Eating and Psychopathologic Symptoms†

Group

BN/BPD BN/nonBPD NC(N = 14) (N = 26) (N = 25) Statistic

Variable Mean SD Mean SD Mean SD t (df = 38)

Binge days/mo 19.18 6.52 15.77 7.12 0.00 0.00 –1.49Binge episodes/mo 30.81a 14.45 21.18b 13.37 0.00 0.00 –2.11*

(N = 13) (N = 22) (N = 25) t (df = 33)

Vomiting days/mo 17.67 9.96 16.18 10.14 0.00 0.00 –0.42Vomiting

episodes/mo 63.58a 58.77 37.47b 44.04 0.00 0.00 –1.50

F(N = 14) (N = 26) (N = 25) (df = 2,62)

Eating AttitudesTest 39.21a 11.72 36.00a 15.03 4.19b 4.33 65.96**

Dissociation(DES) 22.73a 14.79 10.95b 11.41 8.47b 4.44 9.12**

Motor impulsivity(BIS) 27.08a 4.36 23.23b 3.58 20.76b 4.12 11.37**

Cognitiveimpulsivity (BIS) 21.68a 2.25 19.54b 2.23 16.28c 2.26 28.61**

Nonplanningimpulsivity (BIS) 27.79 3.74 28.32 4.69 25.26 5.14 2.80

Total impulsivity(BIS) 76.56a 7.48 71.09a 7.67 62.30b 9.14 14.90**

Affective instability(DAPP-BQ) 62.66a 9.36 58.58a 13.15 37.86b 12.13 26.54**

†Abbreviations: BIS = Barrat Impulsivity Scale, DAPP-BQ = DimensionalAssessment for Personality Pathology-Basic Questionnaire, DES = DissociativeExperiences Scale.a,b,cMeans with different letters in their superscripts differ at the .05 level or better.*p < .05.**p < .001.

431



433



intrafamilial abuse (involving a first-degree relative asperpetrator) or extrafamilial abuse (involving another per-petrator), and dividing each type into less-severe forms(involving nongenital contacts) and more-severe forms(involving genital contact). In this analysis, cases report-ing more than one class of abuse were counted more thanonce. Respective proportions of cases in BN/BPD,BN/nonBPD, and NC groups who reported each class ofabuse prior to age 13 were as follows: less-severeintrafamilial abuse: 35.7%, 15.4%, and 12.0%; more-severe intrafamilial abuse: 21.4%, 0.0%, and 0.0%; less-severe extrafamilial abuse: 14.3%, 7.7%, and 4.0%;more-severe extrafamilial abuse: 0.0%, 7.7%, and 0.0%.The pattern of results links intrafamilial abuse, especiallyin more-severe forms, with the BN/BPD classification(and not with BN/nonBPD).

Chi-square analysis also showed significant group ef-fects for physical abuse prior to age 13 (χ2 = 21.55,df = 2, p < .001) and up to age 18 (χ2 = 25.61, df = 2,p < .001). Here, exact tests (performed on values prior toage 13) differentiated BN/nonBPD cases and BN/BPDcases from NC cases (p < .01 and p < .001, respectively),and BN/nonBPD and BN/BPD groups from each other(p < .05). Results thus indicated elevated physical abuse

in both bulimic groups, although here, too, BN/BPD casesshowed extreme rates (see Table 2). Finally, we comparedthe groups for proportions of any abuse (combined sexualand physical abuse), and found significant effects forabuse prior to age 13 (χ2 = 16.28, df = 2, p < .001) andup to age 18 (χ2 = 18.16, df = 2, p < .001). Here, exacttests (for abuse prior to age 13) differentiated BN/BPDfrom control (p < .001) and BN/nonBPD from control(p < .02), and tended to differentiate BN/BPD fromBN/nonBPD groups (p = .10).

Paroxetine BindingResults reflecting receptor Bmax and Kd are shown for

the 3 groups (11 BN/BPD, 16 BN/nonBPD, and 16 NCcases) in Table 3. One-way ANOVAs revealed a signifi-cant group effect on Bmax, but not on Kd (see Table 3).Group contrasts indicated mean Bmax for both bulimicgroups to be significantly lower than that for the NCgroup. BN/BPD versus BN/nonBPD differences were not,however, obtained. To ensure that the group effect ob-tained on Bmax was not a function of age (which was corre-lated with Bmax), affective problems (known to be associ-ated with 5-HT function), or seasonal variations in plateletparoxetine binding,31 we repeated the analysis using ascovariates age, affective instability, and finally, seasonof testing. In line with reported findings,31 we coded sea-son as a dichotomous winter/spring (high-binding) versussummer/fall (low-binding) distinction. Although covar-iates never yielded significant effects, group effects ineach case remained significant: covarying age: F = 6.24,df = 2,39; p < .004; covarying affective instability:F = 5.58, df = 2,39; p < .01; and covarying season:F = 5.12, df = 2,39; p < .02.

Similarly, to verify the possible impact of medicationon Bmax values, we repeated the ANOVA on Bmax ondata from unmedicated subjects only (6 BN/BPD, 15BN/nonBPD, and 16 NC). The group effect remained reli-able (F = 6.55, df = 2,34; p < .005), with correspondingvalues for BN/BPD, BN/nonBPD, and NC groups being562.00 ± 146.39, 580.00 ± 365.97, and 1047.25 ± 467.95fmol/mg protein, respectively. Newman-Keuls compari-sons again indicated reliable differences between bulimicsand normal eaters, but no borderline/nonborderline differ-ences. Hence, results were quite comparable with thoseobtained in our full sample (see Table 3). As a finaltest, we computed mean Bmax scores for medicated (N = 6)and unmedicated (N = 21) participants who had bulimianervosa. Resulting values (467.67 ± 195.82 and574.86 ± 314.93 fmol/mg protein, respectively) did notdiffer (t = –0.79, df = 25, NS).

Association Between Transporter Densityand Other Indices

To explore possible links between altered transporterdensity and nutritional factors, we computed correlations

Table 2. Number and Percentage of Cases in BN/BPD,BN/nonBPD, and NC Groups Reporting Sexual Abuse,Physical Abuse, or Either Form of Abuse Prior toAge 13 Years and up to Age 18 Years†

Sexual Abuse Physical Abuse Any Abuse

Group N % N % N %

BN/BPD (N = 14)Prior to age 13 y 7 50.0 10 71.4 11 78.6Up to age 18 y 9 64.3 11 78.6 12 85.7

BN/nonBPD (N = 26)Prior to age 13 y 6 23.1 9 34.6 13 50.0Up to age 18 y 6 23.1 10 38.5 13 50.0

NC (N = 25)Prior to age 13 y 4 16.0 1 4.0 4 16.0Up to age 18 y 4 16.0 1 4.0 4 16.0

†Abbreviations: BN/BPD = borderline-bulimic,BN/nonBPD = nonborderline-bulimic, NC = normal-eater control.

Table 3. Mean ± SD for BN/BPD, BN/nonBPD, and NCGroups on Bmax and Kd Indices From Platelet[3H]-Paroxetine-Binding Tests†

BN/BPD BN/nonBPD NC(N = 11) (N = 16) (N = 16) F

Variable Mean SD Mean SD Mean SD (df = 2,40)

Mean Bmax(fmol/mgprotein) 529.09b 172.84 566.13b 357.89 1047.25a 467.95 9.02*

Mean Kd 0.29 0.25 0.32 0.30 0.28 0.25 .07†Abbreviations: Bmax = transporter density,BN/BPD = borderline-bulimic,BN/nonBPD = nonborderline-bulimic, Kd = binding affinity constant,NC = normal-eater control.a,bMeans with different letters in their superscripts differ at the .05level or better.*p < .001.

432


Steiger et al.



(in subjects with bulimia nervosa only) between Bmax val-ues and indices of nutritional status (BMI) and severity ofeating-disorder symptoms (EAT-26 and mean monthlybinge days, vomit days, binge episodes, and vomit epi-sodes). None of the resulting correlations (0.28, –0.02,–0.11, –0.11, –0.20, –0.24, respectively) were significant,implying absence of direct connection between eating be-haviors and reduced transporter density. We also exploredcorrelations (and partial correlations, after removing vari-ance due to bulimic versus nonbulimic status) betweenBmax values and presence of childhood abuse (coded di-chotomously as present or absent). Resulting correlationsfor variables reflecting presence or absence of sexualabuse prior to age 13 (r = –0.18; partial r = –0.17) or up toage 18 (r = –0.21; partial r = –0.14) and for physical abuseprior to age 13 (r = –0.11; partial r = 0.15) or up to age 18(r = –0.15; partial r = 0.12) were nonsignificant. Hence,presence of childhood abuse did not seem to be stronglypredictive of alterations in paroxetine binding.

DISCUSSION

Clinical SymptomsOn certain psychopathologic indices applied in this

study, we found rather clear evidence of a phenomenologi-cal discontinuity between bulimic patients with and with-out borderline personality disorder. Relative to normal-eater control participants, bulimics with the disorderdisplayed remarkable levels of motor impulsivity and dis-sociation; bulimics without it, on the other hand, showedno striking (or statistically significant) elevations on thesecharacteristics (see Table 1). Hence, the bulimic patientswith borderline personality disorder seemed to show a rela-tively unique propensity toward psychopathology of a be-haviorally impulsive or dissociative type. With respect toimpulsive/dissociative potentials, our findings thereforesuggest that bulimia nervosa and borderline personalitydisorder represent rather distinct psychopathologic spectra.

Corroborating the same theme, eating-symptom mea-sures provided evidence of a similar separation betweenbulimic and borderline components of disturbance. Al-though bulimics with borderline personality disordertended to binge more repeatedly when they did binge, allbulimics otherwise tended to display comparable propor-tions of days per month on which they binged and vomitedand similar levels of attitudinal distortion pertaining toeating (on the EAT-26). These trends again point to theconclusion that eating-specific and characterologicalcomponents of disturbance in bulimia nervosa are rela-tively independent and parallel several previous reportsthat have suggested absence of overall differences in bu-limic symptoms attributable to Axis II comorbidity.1 Nev-ertheless, to explain trends toward more dyscontrolledbingeing observed in borderline patients, we propose thatbulimic manifestations in patients with borderline person-

ality disorder, although existing independently of the bor-derline personality pathology per se, may be shaped or ex-aggerated by certain borderline characteristics (like im-pulsivity).

In contrast to the preceding, measures of cognitive im-pulsivity and affective instability differentiated bulimicgroups from normal controls, but yielded no marked bor-derline/nonborderline distinctions (see Table 1). One im-plication here, we assume, may be that there exists a pro-pensity toward labile moods and unreflectiveness in evennonborderline bulimics.

Childhood AbuseConsistent with the proposal (raised above) that bu-

limia nervosa and borderline personality disorder repre-sent independent psychopathologic structures, we foundbulimia nervosa in the presence of borderline personalitydisorder to coincide with substantially greater risk ofchildhood sexual abuse (especially in intrafamilial forms)than did bulimia nervosa without borderline personalitydisorder. Indeed, in the present findings, risk of childhoodsexual abuse was negligibly higher among bulimics whowere nonborderline (i.e., less characterologically dis-turbed) than it was among our normal-eater control par-ticipants. Such findings replicate previous results thathave shown childhood sexual abuse to be more typical ofpersons who have bulimia with comorbid personality pa-thology, and especially in those with borderline personal-ity disorder.10,11 We infer from these that childhood sexualabuse may have a more specific relevance to personalitypathology (in particular, borderline personality disorder)than to bulimia nervosa. Results on indices of physicalabuse differed somewhat in showing a progressive in-crease in prevalence of abuse across normal eaters, non-borderline bulimics, and bulimics with borderline symp-toms. While such findings highlight the pertinence ofabuse experiences for bulimic syndromes (and probablyfor many forms of maladjustment), even here, a particu-larly strong association was indicated between borderlinepersonality disorder and history of abuse. We add, as anote, that our data on childhood abuse do not support in-ferences about causality. In other words, it remains to as-certain whether findings imply causal effects of abuse forborderline personality disorder or isolate abuse as amarker of processes associated with vulnerability to bor-derline personality disorder.

Paroxetine BindingParoxetine-binding tests yielded a somewhat different

pattern of findings, showing Bmax (i.e., platelet 5-HT reup-take) to be significantly (and substantially) lower in bothof our bulimic groups than it was in normal-eater controls,without differing across borderline and nonborderline bu-limic subsamples (see Table 3). The pattern of group dif-ferences described seemed, furthermore, to exist indepen-

433



435



dently of seasonal variations, medication effects, associa-tions with childhood abuse, or indices of nutritional status(i.e., BMI, EAT-26 scores, binge/vomit frequencies). Thepreceding invites the conjecture that we may be observinga serotonergic anomaly that is implicated relatively ubiq-uitously in bulimia nervosa and upon which presence orabsence of borderline personality disorder has little im-pact. In interpreting these results, we remain aware thatplatelet measures need not reflect brain 5-HT functionsunder all circumstances. Nonetheless, if (for reasons re-viewed earlier16–21) we assume that platelet binding oftenprovides an approximation to central mechanisms, thenour findings might be taken to imply generally reduceddensity of the 5-HT transporter in bulimia nervosa—afinding that would be compatible with various other ob-servations indicating reduced 5-HT tone in bulimia ner-vosa.4,5 We can envisage several accounts for the appar-ently reduced reuptake of 5-HT observed. One couldargue that we are observing a compensatory reduction in5-HT reuptake, associated with dietary factors that pro-duce periodic excesses in circulating 5-HT (e.g., repeatedoverloading with 5-HT precursors during binge episodes).While such explanations may be viable, our findings indi-cate relative independence of 5-HT findings from nutri-tional indices and mitigate against any account couchedsolely in terms of dietary sequelae. Alternatively, wemight be observing an adaptive reduction in 5-HT reup-take, corresponding to a constitutional deficit in 5-HTavailability, or 5-HT reuptake levels that are themselves,for constitutional or other reasons, simply too low.Present findings provide no strong indications for prefer-ence among these alternative explanations. Regardless,our findings link bulimia nervosa quite strongly to a re-duction in platelet paroxetine binding. If findings withparoxetine binding correspond to underactivity at the cen-tral presynaptic 5-HT terminal, our results might justifyuse of SSRIs in bulimia nervosa treatment, as such treat-ment would presumably boost 5-HT activity at an appro-priate locus in the system.

In the absence of a comparison group composed of pa-tients with borderline personality disorder but without bu-limia nervosa, we cannot ascertain whether our results in-dicate a 5-HT anomaly that is associated specifically withbulimia nervosa (and hence found uniformly across ourborderline and nonborderline bulimic groups) or isequipresent in borderline personality disorder and bulimianervosa alike (i.e., present in both syndromes, without ad-ditive effects when the 2 are present concurrently). How-ever, previous evidence of reduced tritiated-paroxetinebinding in at least certain individuals with personalitydisorders33 encourages us to speculate that we may beobserving an anomaly that is common to both bulimianervosa and borderline personality disorder, and further-more, that this anomaly may account for frequent co-aggregation between borderline and bulimic syndromes.

Indeed, we speculate that the apparently reduced 5-HTreuptake observed here could reflect a common end stateassociated with bulimic eating but resulting from differentprocesses in different individuals. In some individuals,vulnerability to binge eating may arise from reduced5-HT activity (and corresponding, adaptive reduction indensity of 5-HT transporter sites) resulting, in part, fromsuch factors as prolonged or excessive dieting. Availableevidence indicates that dieting can alter 5-HT function inthe fashion described4,5 and might be responsible for5-HT–mediated effects conducive to dietary dyscontrol.Conversely, individuals with borderline personality disor-der might show a primary disturbance in 5-HT functionthat could underlie these patients’ unique proclivities to-ward trait impulsivity, mood dysregulation, and relatedsymptomatology. If the hypothetical disturbance includedalterations in 5-HT mechanisms regulating appetitive be-havior, it might account for concurrent susceptibility inpatients with borderline personality disorder to problemswith satiation (or binge eating). This proposal might ac-count for a relatively generalized involvement of 5-HTdisturbances in bulimia nervosa, regardless of comor-bidity, and a special affinity between bulimia nervosa anddisorders like borderline personality disorder that are pre-sumed to be, in part, 5-HT mediated.

We add a note concerning a limitation of the presentstudy. In interpreting findings showing borderline/nonborderline distinctions, it is necessary to consider thepossible impact of uncontrolled effects arising from co-morbid Axis I disorders (e.g., major depression or post-traumatic stress disorder) that have not been accountedfor here. While concern about such influences is indeedlegitimate, we believe that our findings, even if they re-flect such confounds, are likely still to be informativefrom a purely phenomenological standpoint about thatgroup of bulimic patients who meet formal borderlinepersonality disorder criteria, and about shared and uniquefactors (developmental and neurobiological) that may co-incide with phenomenologies of a bulimic and borderlinetype. Nonetheless, future work in bulimia nervosa needs,wherever possible, to include fuller controls for variousforms of psychiatric comorbidity.

Drug names: buspirone (BuSpar), fluoxetine (Prozac), paroxetine(Paxil).

REFERENCES

1. Steiger H, Séguin J. Eating disorders: anorexia nervosa and bulimianervosa. In: Millon T, Blaney PH, Davis RD, eds. Oxford Textbookof Psychopathology. New York, NY: Oxford University Press; 1999:365–389

2. Vitousek K, Manke F. Personality variables and disorders in anorexia ner-vosa and bulimia nervosa. J Abnorm Psychol 1994;103:137–147

3. American Psychiatric Association. Diagnostic and Statistical Manual ofMental Disorders, Fourth Edition. Washington, DC: American PsychiatricAssociation; 1994

4. Brewerton TD. Toward a unified theory of serotonin dysregulation in eat-ing and related disorders. Psychoneuroendocrinology 1995;20:561–590

434

https://www.researchgate.net/publication/14615148_Toward_a_unified_theory_of_serotonin_disturbances_in_eating_and_related_disorders?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14615148_Toward_a_unified_theory_of_serotonin_disturbances_in_eating_and_related_disorders?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/297518926_Eating_Disorders_Anorexia_nervosa_and_bulimia?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1





Steiger et al.



5. Wolfe BE, Metzger ED, Jimerson DC. Research update on serotonin func-tion in bulimia nervosa and anorexia nervosa. Psychopharmacol Bull1997;33:345–354

6. Wonderlich SA, Brewerton TD, Jocic Z, et al. Relationship of childhoodsexual abuse and eating disorders. J Am Acad Child Adolesc Psychiatry1997;36:1107–1115

7. Fullerton DT, Wonderlich SA, Gosnell BA. Clinical characteristics of eat-ing disorder patients who report sexual or physical abuse. Int J Eat Disord1995:17:243–249

8. Steiger H, Jabalpurwala S, Champagne J. Axis-II comorbidity anddevelopmental adversity in bulimia nervosa. J Nerv Ment Dis 1996;184:555–560

9. Paris J. Borderline Personality Disorder: A Multidimensional Approach.Washington, DC: American Psychiatric Press; 1994

10. Jimerson DC, Lesem MD, Kaye WH, et al. Low serotonin and dopaminemetabolite concentrations in cerebrospinal fluid from bulimic patientswith frequent binge episodes. Arch Gen Psychiatry 1992;49:132–138

11. Goldbloom DS, Hicks LK, Garfinkel PE. Platelet serotonin uptake in bu-limia nervosa. Biol Psychiatry 1990;28:644–647

12. Fluoxetine Bulimia Nervosa Collaborative Study Group. Fluoxetine in thetreatment of bulimia nervosa. Arch Gen Psychiatry 1992;49:139–147

13. Cornelius JR, Soloff PH, Perel JM, et al. Fluoxetine trial in borderline per-sonality disorder. Psychopharmacol Bull 1990;26:149–152

14. Verkes RJ, Pijl H, Meinders AE, et al. Borderline personality, impulsive-ness, and platelet monoamine oxidase measures in bulimia nervosa andrecurrent suicidal behavior. Biol Psychiatry 1996;40:173–180

15. Waller DA, Sheinberger AL, Gullion C, et al. Impulsivity and neuroendo-crine response to buspirone in bulimia nervosa. Biol Psychiatry 1996;39:371–374

16. Stahl SM. Platelets as pharmacological models for the receptors and bio-chemistry of monoaminergic neurons. In: Longnecker GS, ed. Platelets:Physiology and Pharmacology. Orlando, Fla: Academic Press; 1985:307–340

17. Lesch KP, Wolozin BL, Murphy DL, et al. Primary structure of the humanplatelet serotonin uptake site: identity with the brain serotonin transporter.J Neurochem 1993;60:2319–2322

18. Freeman AM, Stankovic MD, Bradley RJ, et al. Tritiated platelet imipra-mine binding and treatment response in depressed outpatients. Depression1993;1:20–23

19. Hrdina PD. Platelet serotonergic markers in psychiatric disorders: use,

abuse and limitations. J Psychiatr Neurosci 1994;19:87–9020. Meltzer HY, Arora RC. Platelet serotonin studies in affective disorder: evi-

dence for a serotonergic abnormality. In: Sandler M, Coppen A, Harnett S,eds. 5-Hydroxytryptamine in Psychiatry: A Spectrum of Ideas. New York,NY: Oxford University Press; 1991:50–89

21. Marazziti D, Macchi E, Rotondo A, et al. Involvement of the serotoninsystem in bulimia. Life Sci 1988;43:2123–2126

22. Fairburn C, Cooper P. The Eating Disorders Examination. 12th ed. In:Fairburn C, Wilson G, eds. Binge Eating: Nature, Assessment and Treat-ment. New York, NY: Guilford Press; 1993:317–360

23. Garner DM, Olmsted M, Bohr Y, et al. The Eating Attitudes Test: psycho-metric features and clinical correlates. Psychol Med 1982;12:871–878

24. First MB, Spitzer RL, Gibbon M, et al. Structured Clinical Interview forDSM-IV Axis-II Personality Disorders (SCID-II, Version 2.0). New York,NY: Biometric Research, New York State Psychiatric Institute; 1996

25. Bernstein EM, Putnam FW. Development, reliability, and validity of a dis-sociation scale. J Nerv Ment Dis 1986;174:727–735

26. Barrat E. Impulsive subtraits: arousal and information processing. In:Spence JT, Izard CE, eds. Motivation, Emotion and Personality. NewYork, NY: Elsevier; 1985:137–146

27. Livesley WJ, Jackson DN, Schroeder ML. Factorial structure of traits de-lineating personality disorders in clinical and general population samples.J Abnorm Psychol 1992;101:432–440

28. Bernstein DP, Fink L, Handelsman L, et al. Initial reliability and validity ofa new retrospective measure of child abuse and neglect. Am J Psychiatry1994;151:1132–1136

29. Langer SZ, Schoemaker H, Segonac A. (3H)-Paroxetine binding tohuman platelets: clues to selectivity paradox from temperature studies. BrJ Psychiatry 1985;96:303–309

30. Ortiz J, Artigas F, Gelpi E. Serotonergic status in human blood. Life Sci1988;43:983–990

31. D’Hondt P, Maes M, Leysen JE, et al. Seasonal variation in platelet[3H]paroxetine binding in healthy volunteers: relationship to climaticvariables. Neuropsychopharmacology 1996;15:187–198

32. Kasahara T, Ishigooka J, Nagata E, et al. Long-lasting inhibition of 5-HTuptake of platelets in subjects treated with duloxetine, a potential anti-depressant. Nihon Shinkei Seishin Yakurigaku Zasshi 1996;16:25–31

33. Coccaro EF, Kavoussi RJ, Sheline YI, et al. Impulsive aggression in per-sonality disorder correlates with tritiated paroxetine binding in the platelet.Arch Gen Psychiatry 1996;53:531–536

435


Dextroamphetamine in Men With HIV

437



tandard antidepressants have been shown to be effec-tive in treating depression in patients with the human

Effects of Dextroamphetamine onDepression and Fatigue in Men With HIV:A Double-Blind, Placebo-Controlled Trial

Glenn J. Wagner, Ph.D., and Richard Rabkin, M.D.

Background: This report documents findingsfrom a small placebo-controlled trial of dextro-amphetamine for depression and fatigue in menwith the human immunodeficiency virus (HIV).Dextroamphetamine offers the potential forrapid onset of effect and activation properties,both of which are important to persons withmedical illness and an uncertain, but limited,life expectancy.

Method: Primary inclusion criteria includedthe presence of a DSM-IV depressive disorder,debilitating fatigue, and no history of dependenceon stimulants. The study consisted of a 2-weekrandomized, placebo-controlled trial, with theblind maintained until week 8 for responders,followed by open treatment through the comple-tion of 6 months.

Results: Of 23 men who entered the study,22 completed the 2-week trial. Intent-to-treatanalysis indicated that 73% of patients (8/11)randomly assigned to dextroamphetaminereported significant improvement in mood andenergy, compared with 25% (3/12) among place-bo patients (Fisher exact test, p < .05). Bothclinician- and self-administered measures indi-cated significantly improved mood, energy, andquality of life among patients taking dextroam-phetamine. There was no evidence of the devel-opment of tolerance of, abuse of, or dependenceon the medication.

Conclusion: These results suggest that dextro-amphetamine is a potentially effective, fast-actingantidepressant treatment for HIV patients withdepression and debilitating fatigue.


Received May 10, 1999; accepted Nov. 8, 1999. From the Departmentof Psychiatry, New York State Psychiatric Institute, New York (bothauthors); and RAND, Santa Monica, Calif. (Dr. Wagner).

Funded by National Institute of Mental Health grant 5 RO3 MH54931-02 (Dr. Wagner).

Reprint requests to: Glenn J. Wagner, Ph.D., RAND, 1700 Main St.,Santa Monica, CA 90407 (e-mail: [email protected]).

Simmunodeficiency virus (HIV), including those with late-stage illness.1–4 However, some patients report improvedmood but continued low energy, which often accompaniesadvanced HIV illness.5 An alternative treatment for suchpatients is psychostimulant medication, with its potentialadvantages including rapid onset of effect (2–3 days), ac-tivation properties, and absence of anticholinergic side ef-fects. Case reports, chart reviews, and our own open-labelstudy of depressed HIV patients treated with psycho-stimulants suggest that psychostimulants are well toleratedin this population and that treatment improves mood, psy-chomotor activity, and cognitive functioning.6–9 However,we are not aware of any placebo-controlled trials ofpsychostimulants involving HIV patients.

Early placebo-controlled studies involving medicallyhealthy patients with primary depression failed to find adifference between psychostimulants and placebo, largelybecause of high response rates for both.10,11 With the ad-vent of tricyclic antidepressants (e.g., imipramine) and se-lective serotonin reuptake inhibitors (e.g., fluoxetine, ser-traline), psychostimulants are now primarily used as anaugmentation to standard antidepressants.10–12 However,psychostimulants have been found to be effective as theprimary antidepressant treatment in subpopulations in-cluding patients with depression and apathy, depressedpatients with organic impairment, and depressed medi-cally ill patients.11,13–15

The primary concerns about the use of psycho-stimulants such as dextroamphetamine are the risks ofabuse/dependence and development of tolerance. Wilburet al.16 reported habituation effects as efficacy declinedover time; however, many studies have reported no evi-dence of tolerance or habituation with duration of treat-ment ranging from months to years.10,17 There have beenno reports of abuse or dependence on psychostimulanttreatment in either HIV or non-HIV patients undermedical supervision. Other adverse reactions may includeincreased anxiety, insomnia, and overstimulation, allof which are usually transient or reversible by dosereduction. Psychostimulants have been associated withappetite/weight loss; however, in studies of HIV patients,appetite stimulation has also been reported.6,7,9

436


Wagner and Rabkin



We conducted a 2-week randomized, double-blind,placebo-controlled trial followed by up to 6 months ofopen-label treatment to assess the efficacy of dextro-amphetamine in treating depression and debilitating fa-tigue in people with HIV. A secondary aim of the studywas to study the effects of dextroamphetamine on cogni-tive function.

METHOD

RecruitmentPatients were recruited indirectly, having responded to

general notices about ongoing treatment studies for de-pression or fatigue that were placed in local HIV and ac-quired immunodeficiency syndrome (AIDS) newslettersor posted at local AIDS organizations. It was decided notto mention the specific nature of the treatment in postingsor notices to limit the risk of enrolling those who soughtthe treatment for recreational use. Instead, the psycholo-gists conducting the initial evaluations for other depres-sion treatment studies suggested this option when appro-priate. Enrollment was slower than expected (the samplewas recruited over a period of nearly 4 years) for otherreasons as well: patients with a history of stimulant (in-cluding cocaine) abuse/dependence were excluded, whicheliminated a significant proportion of the HIV communitythat we serve, and recent advancements in HIV treatments(i.e., protease inhibitors and combination antiretroviraltherapy) led to significant improvement in the health ofpeople with HIV, including those seeking treatmentthrough our studies, and consequently the complaint ofdebilitating fatigue was much less frequent.

Eligibility CriteriaInclusion criteria included age of 18 to 65 years, hav-

ing a DSM-IV depressive disorder diagnosis, and report-ing debilitating fatigue. Exclusion criteria included ahistory of abuse/dependence on stimulants (including co-caine) or recent (within past 6 months) abuse/dependenceon any substance, psychotic symptoms, serious suiciderisk, bipolar disorder, and current use of psychotropicmedications. Patients were required to provide written,informed consent after study procedures and possible sideeffects were explained, and their primary care physicianswere required to sign a statement that there were no medi-cal contraindications to study participation.

ProcedureThe study consisted of a 2-week randomized, double-

blind, placebo-controlled trial, followed by up to 24 weeksof open-label treatment. For responders at week 2, theblind was maintained until week 8 or relapse in terms ofmood or energy. For nonresponders at week 2, the codewas broken, and those taking placebo were started on dex-troamphetamine treatment, while those taking dextroam-

phetamine were offered standard antidepressants. Patientswere instructed to follow a titrated dose schedule: b.i.d. inthe morning and midafternoon with a daily dose starting at5 mg. In the absence of clinical improvement and limitingside effects, the dose was increased every 2 days in unitsof 2.5 mg/day with a maximum dosage of 40 mg/day. Pa-tients were seen weekly until response was achieved, pluspatients were contacted by telephone every other day tomonitor dosage and response until optimal dosage levelwas reached; thereafter, visits were every 2 weeks.

MeasuresDSM-IV depressive diagnoses were assessed with the

Structured Clinical Interview for DSM-IV (SCID).18 The21-item clinician-rated Hamilton Rating Scale for Depres-sion (HAM-D)19 was used to assess severity of depressivesymptoms. Additional measures of mood included the self-report Brief Symptom Inventory (BSI)20 and Beck Hope-lessness Scale (BHS),21 as well as a 10-point visual analogscale (VAS) from 1, “very depressed,” to 10, “very happy.”Fatigue was assessed with 7 items from the Chalder Fa-tigue Scale (CFS)22 and a 10-point VAS with the anchorsbeing 1, “very low energy,” and 10, “very high energy.”

The clinician-rated Clinical Global Impressions scale(CGI)23 provided a global rating of severity and degree ofimprovement for depression and energy level separately.For ratings of improvement, a score of 1, “very much im-proved,” or 2, “much improved,” was considered a re-sponse, whereas a score of 3, “minimally improved,” orgreater was nonresponse. Patients were classified as re-sponders if they responded to treatment in terms of bothmood and energy.

The neuropsychological tests used to assess cognitivefunction included the Trail-Making Test (A and B),24 a mea-sure of psychomotor speed, executive function, and cogni-tive flexibility; and the digit symbol subscale of theWechsler Adult Intelligence Scales-Revised (WAIS-R),25 ameasure of psychomotor speed and learning. Quality of lifewas assessed using the 16-item self-report Quality of LifeEnjoyment and Satisfaction Questionnaire (Q-L-S-Q)26 andthe 5-item clinician-rated Quality of Life Index (QLI) bySpitzer et al.27 Overall physical functioning was measuredusing the Karnofsky Performance Index.28

The SCID was administered at baseline; all other as-sessments were administered at baseline and weeks 2, 8,16, and 26, with the exception of the side effects form,which was completed at each visit.

Statistical AnalysisDescriptive statistics were used to describe the sample.

Independent t tests (2-tailed) and chi-square or Fisherexact tests were used to compare treatment groups oncontinuous and categorical variables, respectively. Pairedt tests (2-tailed) were used to assess change followingtreatment.

437



439



RESULTS

Sample CharacteristicsTwenty-three men entered the study between August

1995 and March 1999; women were eligible, but none en-rolled. Mean ± SD age was 41 ± 8 years; most were white(N = 16; 70%), unemployed (N = 20; 87%), and had atleast some college education (N = 17; 74%). As a whole,the sample had advanced HIV illness; 74% (N = 17) metcriteria for an AIDS diagnosis (according to 1993 criteriafrom the Centers for Disease Control), including 70%(N = 16) who had a history of a major opportunistic infec-tion. The mean CD4 cell count at baseline was 251 ± 210cells/mm3, with 52% (N = 12) having counts below 200.The mean number of medications used by the sample atbaseline was 7 ± 5 (range, 2–23). Fifteen men (65%) wereon combination antiretroviral therapy, including a pro-tease inhibitor. Most (N = 19; 83%) of the sample scored80 or above on the Karnofsky Performance Index, whichindicates a level of functioning sufficient to “carry outnormal activity with effort; some signs or symptoms ofillness are present.”

Baseline SymptomatologyTwelve men (52%) were diagnosed with a current ma-

jor depression, 3 (13%) were diagnosed with dysthymia,and 8 (35%) had either subthreshold major depression orminor depression. The mean ± SD HAM-D score at base-line was 14.9 ± 4.2 (range, 8–23). The mean score on theBSI depression subscale was 1.88 ± 0.77, and the totalscore was 1.13 ± 0.45; these scores are similar to the nor-mative means (1.80 for depression subscale, 1.32 for totalscore) for psychiatric outpatients. The mean score on theBHS was 9.8 ± 6.1, which is approximately equivalent tothe normative mean for depressed samples; 57% (N = 13)scored above 9, which is considered a cutoff for moderateto severe levels of hopelessness.

The mean ± SD score on the CFS was 28.2 ± 4.0; 70%(N = 16) had “clinical fatigue” (defined as a score of 28 orhigher; possible range of scores is 7–35). Nine men per-formed in the impaired range (defined as scoring morethan 2 standard deviations below the normative mean)on either the Trail-Making test A or test B (N = 9; 39%),and 2 (9%) did so on the digit symbol subscale of theWAIS-R; overall, 10 subjects (44%) scored in the im-paired range on at least 1 of these tests.

OutcomeAll but 1 of the 23 men completed the 2-week double-

blind trial. The only dropout discontinued owing to sideeffects (increased anxiety) from placebo. Of the 22 com-pleters, 11 were randomly assigned to dextroamphet-amine and 11 to placebo. Eight completers (73%) as-signed to dextroamphetamine responded to treatment inboth mood and energy, compared with 3 patients (27%)

on placebo (Fisher exact test, p < .10). In an intent-to-treatanalysis with the 1 placebo dropout included as a non-responder, the group difference becomes statistically sig-nificant (Fisher exact test, p < .05). Response to energyand mood were concordant for all but 1 patient who wastaking dextroamphetamine and responded in terms of en-ergy but not mood. The mean ± SD daily dose at week 2for those randomly assigned to dextroamphetamine was22 ± 9 mg, with the most common daily dose being 30mg. Once optimal dosage was obtained (usually within 2weeks of beginning dextroamphetamine), all but 2 pa-tients remained at that dose throughout the study. At week26, the dose ranged from 10 mg/day to 40 mg/day with amean daily dose of 26 ± 12 mg.

All 3 nonresponders to dextroamphetamine had majordepression, compared with only 2 of the 8 responders;however, response was not associated with HAM-D scoreat baseline. CD4 cell count at baseline was not associatedwith response among those randomly assigned to dextro-amphetamine, but placebo patients with higher CD4 cellcounts were more likely to be responders (mean ± SDCD4 = 546 ± 141 cells/mm3) than nonresponders(mean ± SD CD4 = 180 ± 133 cells/mm3; t = 4.0, p < .01).

Among completers on dextroamphetamine treatment,significant improvement in depressive symptoms as mea-sured by both the clinician-rated HAM-D and self-reportBSI depression subscale as well as reduced fatigue andimproved quality of life and physical functioning werefound at week 2 compared with baseline (Table 1). Therewere indicators of mild improvement (trend level of sig-nificance) in cognitive functioning. Among patients onplacebo, fewer parameters showed improvement; therewere indications of improved mood and fatigue, but notquality of life and physical functioning (Table 2).

Eleven patients were taking dextroamphetamine atweek 8, including 5 dextroamphetamine responders, 5placebo nonresponders, and 1 placebo responder wholater relapsed; an additional 4 patients (2 dextroamphet-amine responders and 2 placebo responders at week 2)discontinued treatment prior to week 8 because of side ef-fects. Among these 11, significant improvement in de-pressive symptoms, energy level, and quality of life werefound at week 8 compared with baseline (or week 2 forplacebo responders who started dextroamphetamine; datanot shown). Nine of these 11 men went on to complete the6-month trial; of the remaining 2, 1 died shortly afterweek 8 from leukemia and the other discontinued treat-ment at week 16 because of a relapse in depression.

Side effects. Although none of the patients randomlyassigned to dextroamphetamine discontinued treatmentprior to week 2, 4 patients (including 2 placebo non-responders who started dextroamphetamine) later dis-continued dextroamphetamine because of side effects(overstimulation, heart palpitations, sleep deprivation).The most common treatment-emergent side effects were

438


Wagner and Rabkin



overstimulation, insomnia, and loss of appetite and/orweight, with each reported by 5 patients (22%) at somepoint during the study. Other less common side effects in-cluded heart palpitations and headaches, with each re-ported by 3 patients. In general, side effects were tran-sient, reversible, and well managed with dose reduction;no serious medical side effects were reported.

DISCUSSION

Results suggest that dextroamphetamine is a poten-tially effective, fast-acting antidepressant treatment forHIV patients with depression and debilitating fatigue. Al-though the numbers are small, the data suggest that treat-ment may be less likely to be effective for patients withmore severe depression (major depression). Among thoserandomly assigned to dextroamphetamine, less than half(2/5) of the patients diagnosed with major depressionwere responders, compared with all 6 of the patients withmore mild depression (i.e., dysthymia, subthreshold de-pression, or minor depression). Since improvement inmood nearly always coincided with improved energy, wewere unable to assess whether dextroamphetamine has anantidepressant effect independent of its activation proper-ties. Measures of cognitive function suggested a mildtreatment effect; the trends we found may have been sta-tistically significant with a larger sample size.

Although psychostimulants are not a first-line standardantidepressant to be used with most patients, for de-pressed patients with fatigue and medical illness the po-tential for the treatment to enhance overall functioningand quality of life makes it a viable treatment option. Pa-tients described being able to resume activities of dailyliving including grocery shopping, cleaning the house,and visiting with friends, and at least 2 patients creditedthe treatment for their ability to return to work. Since re-turn to work is more widely discussed as HIV becomesmore of a chronic illness, treatments such as psycho-stimulants that can reduce the barriers of fatigue and de-pression become relevant to a broader range of people liv-ing with HIV.

Psychostimulant use is controversial and often notconsidered a treatment option for several reasons includ-ing the potential for abuse and dependence, developmentof tolerance and withdrawal reactions, and restrictiveregulations regarding its prescription since it is a con-trolled substance. It is our clinical impression that theseconcerns, along with the illicit “street drug” use ofamphetamines, have resulted in a stigma attached topsychostimulants that has both patients and cliniciansskeptical about their value and appropriateness; a numberof patients in our trial who benefited from treatmentfound it difficult to convince their doctors to continue thedextroamphetamine prescriptions following the comple-tion of their study participation. Results from this studydid not support these concerns; 10 of 15 responders todextroamphetamine (includes nonresponders to placebo

Table 1. Change in Symptom Domains Following 2 Weeks ofDextroamphetamine Treatment Among Completers (N = 11)a

Baseline Week 2

Symptom Domain Mean SD Mean SD t p Value

MoodHAM-D 15.1 3.6 7.5 3.9 5.3 .000BSI total score 1.18 0.49 0.73 0.60 2.0 .077BSI depression

subscale 2.00 0.85 0.98 1.04 3.1 .015BHS 8.4 5.9 8.2 6.9 0.2 .810VAS item 3.9 0.9 6.0 1.9 2.8 .018

EnergyCFS 26.8 4.9 18.3 3.9 5.5 .000VAS item 3.3 1.3 6.7 2.3 8.4 .000

Quality of lifeQ-LES-Q 39.6 6.2 52.2 12.1 4.2 .002QLI 5.3 1.3 7.0 2.4 2.3 .042Karnofsky Performance

Index 82.7 7.9 92.7 7.9 3.3 .008Cognitive functioning

Trail-Making Test A(percentile) 50.3 30.1 64.2 20.2 1.3 .216

Trail Making Test B(percentile) 45.6 20.7 65.3 21.1 2.2 .063

WAIS-R, digit symbolsubscale(age-adjusted mean) 10.2 1.9 11.5 1.8 1.9 .090

aAbbreviations: BHS = Beck Hopelessness Scale, BSI = BriefSymptom Inventory, CFS = Chalder Fatigue Scale,HAM-D = Hamilton Rating Scale for Depression, Q-LES-Q = Qualityof Life Enjoyment and Satisfaction Questionnaire, QLI = Quality ofLife Index, VAS = visual analog scale, WAIS-R = Wechsler AdultIntelligence Scales-Revised.

Table 2. Change in Symptom Domains Following 2 Weeks ofPlacebo Among Completers (N = 11)a

Baseline Week 2

Symptom Domain Mean SD Mean SD t p Value

MoodHAM-D 14.4 4.8 9.4 6.1 2.5 .033BSI total score 1.03 0.46 0.71 0.27 2.4 .036BSI depression

subscale 1.77 0.75 1.29 0.72 1.7 .120BHS 10.1 6.7 9.5 6.6 0.6 .572VAS item 4.1 1.9 4.9 2.3 1.6 .146

EnergyCFS 29.2 3.1 25.4 4.4 2.6 .026VAS item 3.2 1.2 4.3 2.1 1.7 .126

Quality of lifeQ-LES-Q 40.3 7.9 42.6 12.1 0.8 .424QLI 4.7 2.0 6.0 2.4 1.7 .116Karnofsky Performance

Index 80.9 11.4 81.8 15.4 0.6 .588Cognitive functioning

Trail-Making Test A(percentile) 44.6 31.6 61.2 32.5 2.3 .048

Trail-Making Test B(percentile) 55.9 31.4 59.2 34.7 0.3 .736

WAIS-R, digit symbolsubscale(age-adjusted mean) 10.5 2.4 10.9 1.8 0.6 .563

aAbbreviations are explained in the footnote to Table 1.

439



441



who switched to dextroamphetamine) maintained their re-sponse throughout their participation in the 6-month trial.Although an additional 4 patients discontinued treatmentdue to side effects, only 1 patient experienced a relapse indepression after having been a responder, and improve-ment in energy level was consistently maintained. At leastin the short term, there was little evidence of tolerance de-velopment in the study; although 2 patients required a40--g/day dose to obtain a therapeutic effect, patients didnot require ever increasing doses to maintain response.Some patients who had reached a dose of 30 or 40 mg/dayreduced the dose slightly over time without a loss of ef-fect. Alternatives to psychostimulants for use in treatingpatients with histories of drug addiction include hormonaltherapies such as testosterone29 for men and dehydro-epiandrosterone (DHEA)30 for men and women, both ofwhich we have found to have positive effects on moodand energy in open-label trials with HIV patients.

Overall, the treatment was tolerated well by thesample. The efficacy and well-tolerated nature of thetreatment are exemplified by the intention of all but one ofthe patients who completed the entire 26-week trial tocontinue the treatment with their primary care physician.The one patient who chose not to continue successfullyweaned himself off treatment without a significant loss ofeffect in terms of energy or mood. This patient’s experi-ence highlights the possibility that long-term treatmentmay not be necessary for some patients, particularly asHIV treatments continue to improve. Extended follow-upassessments are needed in future research to address is-sues related to safety, especially concerning tolerance andabuse/dependence.

In closing, a larger controlled trial that includeswomen is needed to assess treatment efficacy for severeas well as moderate and mild depression and whether dex-troamphetamine can improve mood independent of en-ergy level.

Drug names: dextroamphetamine (Dexedrine and others), fluoxetine(Prozac), sertraline (Zoloft).

REFERENCES

1. Rabkin JG, Wagner G, Rabkin R. Sertraline effects on mood and immunestatus in patients with major depression and HIV illness: an open trial.J Clin Psychiatry 1994;55:433–439

2. Rabkin JG, Wagner GJ, Rabkin R. Fluoxetine treatment for depression inpatients with HIV/AIDS: a randomized placebo-controlled trial. Am JPsychiatry 1999;156:101–107

3. Rabkin JG, Rabkin R, Harrison W, et al. Effects of imipramine on moodand immune status in depressed patients with HIV illness. Am J Psychiatry1994;151:516–523

4. Levine S, Anderson D, Bystritsky A, et al. A report of eight HIV-seroposi-tive patients with major depression responding to fluoxetine. J AIDS1990;3:1074–1077

5. Darko DF, McCutchan JA, Kripke DF, et al. Fatigue, sleep disturbance,disability and indices of progression of HIV infection. Am J Psychiatry1992;149:514–520

6. Holmes VF, Fernandez F, Levy JK. Psychostimulant response in AIDS-related complex patients. J Clin Psychiatry 1989;50:5–8

7. Fernandez F, Levy JL, Galizzi H. Response of HIV-related depression topsychostimulants: case reports. Hosp Community Psychiatry 1988;39:628–631

8. Angrist B, d’Hollosy M, Sanfilipo M. Central nervous system stimulantsas symptomatic treatments for AIDS-related neuropsychiatric impairment.J Clin Psychopharmacol 1992;2:268–272

9. Wagner G, Rabkin JG, Rabkin R. Dextroamphetamine as a treatment fordepression and low energy in AIDS patients: a pilot study. J PsychosomRes 1997;42:113–116

10. Satel SL, Nelson JC. Stimulants in the treatment of depression: a criticaloverview. J Clin Psychiatry 1989;50:241–249

11. Nelson JC. Psychostimulants in the treatment of depression. Prim Psychi-atry 1995;2:26–27

12. Stoll AL, Pillay SS, Diamond L, et al. Methylphenidate augmentation ofserotonin selective reuptake inhibitors: a case series. J Clin Psychiatry1996;57:72–76

13. Wallace AE, Kofoed LL, West AN. Double-blind, placebo-controlled trialof methylphenidate in older, depressed, medically ill patients. Am J Psy-chiatry 1995;152:929–931

14. Frierson RL, Wey JJ, Tabler JB. Psychostimulants for depression in medi-cally ill. Am Fam Physician 1991;43:163–170

15. Masand P, Pickett P, Murray GB. Psychostimulants for secondary depres-sion in medical illness. Psychosomatics 1991;32:203–208

16. Wilbur DL, MacLean AR, Allen EV. Clinical observations on the effect ofbenzedrine sulfate. JAMA 1937;109:549–554

17. Fernandez F, Adams F, Holmes VF. Methylphenidate for depressive disor-ders in cancer patients. Psychosomatics 1987;28:455–462

18. Spitzer RL, Williams JBW, Gibbon M, et al. Structured Clinical Interviewfor DSM-IV (SCID). New York, NY: Biometric Research, New York StatePsychiatric Institute; 1994

19. Hamilton M. Development of a rating scale for primary depressive illness.Br J Soc Clin Psychol 1967;6:278–296

20. Derogatis LR, Melisaratos M. The Brief Symptom Inventory: an introduc-tory report. Psychol Med 1983;13:595–605

21. Beck AT, Weissman A, Lester D, et al. The measurement of pessimism: thehopelessness scale. J Consult Clin Psychol 1974;42:861–865

22. Chalder T, Berelowitz G, Pawlikowska T, et al. Development of a fatiguescale. J Psychosom Res 1993;37:147–153

23. Guy W. ECDEU Assessment Manual for Psychopharmacology. US DeptHealth, Education, and Welfare publication (ADM) 76-338. Rockville,Md: National Institute of Mental Health; 1976:218–222

24. Army Individual Test Battery. Manual of Directions and Scoring. Wash-ington, DC: War Department, Adjutant General’s Office; 1944

25. Wechsler D. Wechsler Adult Intelligence Scales-Revised. New York, NY:Psychological Corp; 1981

26. Endicott J, Nee J, Harrison W, et al. Quality of Life Enjoyment and Satis-faction Questionnaire: a new measure. Psychopharmacol Bull 1993;29:321–326

27. Spitzer WO, Dobson AJ, Hall J, et al. Measuring the quality of life of can-cer patients: a concise QL-Index for use by physicians. J Chronic Dis1981;34:585–597

28. Schag CC, Heinrich RL, Ganz PA. Karnofsky Performance Status revis-ited: reliability, validity, and guidelines. J Clin Oncol 1984;2:187–193

29. Rabkin JG, Wagner G, Rabkin R. Testosterone therapy for HIV+ men withand without hypogonadism. J Clin Psychopharmacol 1999;19:19–27

30. Rabkin JG, Ferrando SJ, Wagner GJ, et al. DHEA treatment for HIV+ pa-tients: effects on mood, androgenic and anabolic parameters. Psycho-neuroendocrinology 2000;25:53–68

440

https://www.researchgate.net/publication/16496003_Karnofsky_Status_Revisited_Reliability_Validity_and_Guidelines?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/16496003_Karnofsky_Status_Revisited_Reliability_Validity_and_Guidelines?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14058892_Dextroamphetamine_as_a_treatment_for_depression_and_low_energy_in_AIDS_patients_A_pilot_study?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/21127955_Psychostimulants_for_Secondary_Depression_in_Medical_Illness?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/21127955_Psychostimulants_for_Secondary_Depression_in_Medical_Illness?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/21168493_Psychostimulants_for_depression_in_the_medically_ill?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/21168493_Psychostimulants_for_depression_in_the_medically_ill?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/21601765_Fatigue_sleep_disturbance_disability_and_indices_of_progression_of_HIV_infection?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/15236760_Effects_of_sertraline_on_mood_and_immune_status_in_patients_with_major_depression_and_HIV_illness_An_open_trial?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/15444319_Double-blind_placebo-controlled_trial_of_methylphenidate_in_older_depressed_medically_ill_patients?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/20587900_Stimulants_in_the_Treatment_of_Depression_A_Critical_Overview?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/20587900_Stimulants_in_the_Treatment_of_Depression_A_Critical_Overview?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/20513111_Psychostimulant_Response_in_AIDS-Related_Complex_Patients?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/20513111_Psychostimulant_Response_in_AIDS-Related_Complex_Patients?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/19753256_Response_of_HIV-Related_Depression_to_Psychostimulants_Case_Reports?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/243770205_Wechsler_Adult_Intelligence_Scale-Revised_UK?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/243770205_Wechsler_Adult_Intelligence_Scale-Revised_UK?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14608523_Methylphenidate_augmentation_of_serotonin_selective_reuptake_inhibitors_A_case_series?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/20941567_A_report_of_eight_HIV-seropositive_patients_with_major_depression_responding_to_fluoxetine?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/283922626_Development_of_a_fatigue_scale?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/283922626_Development_of_a_fatigue_scale?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/16572932_The_Brief_Symptom_Inventory_BSI_An_introductory_report?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/16572932_The_Brief_Symptom_Inventory_BSI_An_introductory_report?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/19722885_Methylphenidate_for_depressive_disorders_in_cancer_patients_An_alternative_to_standard_antidepressants?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/19722885_Methylphenidate_for_depressive_disorders_in_cancer_patients_An_alternative_to_standard_antidepressants?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/null?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/15284503_Effects_of_fluoxetine_on_mood_and_immune_status_in_depressed_patients_with_HIV_illness?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/17106538_Development_of_a_Rating_Scale_for_Primary_Depressive_Illness?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/17106538_Development_of_a_Rating_Scale_for_Primary_Depressive_Illness?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1


ECG Abnormalities and Clozapine

441



lthough clozapine is effective for the treatment ofschizophrenia and other psychiatric conditions,

Electrocardiographic Abnormalitiesin Patients Treated With Clozapine

Ung Gu Kang, M.D., Ph.D.; Jun Soo Kwon, M.D., Ph.D.;Yong Min Ahn, M.D., Ph.D.; Sun Ju Chung, M.D.; Jee Hyun Ha, M.D.;

Young Jin Koo, M.D.; and Yong Sik Kim, M.D., Ph.D.

Background: Cardiovascular side effects of cloza-pine are not uncommon, but few systematic studiesof these effects have been performed. In this study,we reviewed data on the electrocardiographic (ECG)abnormalities in patients treated with clozapine.

Method: Sixty-one patients treated with clozapinewere selected from the Seoul National University Hos-pital Treatment-Resistant Schizophrenia Clinic. A ret-rospective chart review was conducted to identifyECG abnormalities and cardiovascular side effects.

Results: The prevalence of ECG abnormalities inpatients who had been using antipsychotics other thanclozapine was 13.6% at baseline, which increased sig-nificantly to 31.1% after commencement of clozapinetreatment. Among the 53 patients without baselineECG abnormalities, 13 showed new-onset ECG abnor-malities after using clozapine. Normal ECG under pre-vious antipsychotic medication reduced the risk ofnew-onset ECG abnormalities, whereas increased agewas found to increase the risk. The occurrence of or-thostatic hypotension or tachycardia was not relatedto the development of ECG abnormalities. Most of thenewly developed abnormalities had little clinical sig-nificance, and they tended to occur during the initialphase of treatment. In 10 patients, ECGs normalizeddespite the continued use of clozapine. Clozapineincreased corrected QT interval (QTc) in a dose-dependent fashion; however, the clinical significanceof this observation is uncertain. Pathologic prolonga-tion of QTc was found to be rare.

Conclusion: Although a substantial portion of pa-tients treated with clozapine developed ECG abnor-malities, most of the abnormalities were benign anddid not hinder further treatment.


Received July 28, 1999; accepted Dec. 14, 1999. From the Departmentof Psychiatry, Seoul National University College of Medicine, Seoul, Korea(Drs. Kang, Kwon, Ahn, Ha, Koo, and Kim); and the Department ofNeuropsychiatry, Gil Medical Center, Gachon Medical College, Incheon,Korea (Dr. Chung).

Supported by Seoul National University Hospital Research Grant02-97-337.

The authors thank Dr. I. H. Chae, M.D., Ph.D., from the Department ofInternal Medicine, Seoul National University College of Medicine, forcritical review of this article.

Reprint requests to: Yong Sik Kim, M.D., Ph.D., Department ofPsychiatry, Seoul National University College of Medicine, 28 Yongon-Dong, Chongno-Gu, 110-744 Seoul, Korea (e-mail: [email protected]).

Avarious adverse effects limit its use as a first-line treat-ment. The most serious adverse effect of clozapine isagranulocytosis. However, the prevalence of this compli-cation is as low as 0.38% to 0.91%,1,2 and the standardizedmonitoring protocol for its early detection coupled withimproved management strategies involving, for example,the use of cytokines has decreased mortality.2 Cardiovas-cular side effects, including tachycardia and orthostatichypotension, are more frequent. These side effects arerelatively common and can adversely affect the patients’quality of life, but little attention is paid to them becausethey are believed to be benign. However, unexpected andsometimes fatal cardiovascular side effects such as hyper-tension,3–5 pericarditis or ncocarditis,6,7 cardiomyopathy,8

ST elevation mimicking ischemic heart disease,9 and ar-rhythmias10 have also been reported in patients using clo-zapine. These reports demonstrate the need for systemicstudy of the cardiac and electrocardiographic effectsof clozapine. In this study, we reviewed the data on theelectrocardiographic abnormalities detected in patientstreated with clozapine.

METHOD

Cases were selected from patients enrolled at theSeoul National University Hospital Treatment-ResistantSchizophrenia Clinic between January 1996 and June1998. A total of 121 cases were reviewed. Electrocardio-grams (ECGs) were taken before starting the clozapinetreatment and repeated during the titration and mainte-nance periods. Although the timing of the ECGs was leftto the clinicians, guidelines were given. Most of the pa-tients underwent ECG within a few days before the firstdose of clozapine, 1 week after starting clozapine, andagain when the dose was raised to approximately 400mg/day (and when clinically indicated). From this groupof patients, those without a medical history of previousheart disease, with at least 2 ECGs (including the onetaken during pretreatment), and with available vital signrecordings were selected, and their medical records werereviewed. A total of 61 patients (36 men and 25 women)



443



were finally selected. The age distribution was from 18 to54 years; mean age was 31.8 years for men and 31.3 yearsfor women. Diagnoses included schizophrenia, schizo-affective disorder, and bipolar disorder, according toDSM-IV criteria. All of the patients were Korean. Forty-four patients were receiving antipsychotic medicationwhen the baseline ECG was taken. Drugs used includedchlorpromazine, haloperidol, loxapine, pimozide, risperi-done, sulpiride, and thioridazine. Most of the patientsused high-potency drugs, and their mean chlorpromazine-equivalent dose was 451 mg/day.

ECG readings were performed primarily using theautomatic reading system and were inspected by experi-enced cardiologists in cases that suggested any abnormal-ity. Cardiovascular variables (heart rate and blood pres-sure) were transcribed from regular vital sign recordsperformed by nurses. Tachycardia was defined as a rest-ing heart rate ≥ 100 beats per minute at any time duringthe observation period after clozapine use. Orthostatic hy-potension was defined as ≥ 20 mm Hg change betweenthe supine and standing systolic blood pressure at anytime during the same period.

We first calculated the prevalence of ECG abnormali-ties under various conditions and then examined the ab-normal ECG findings. Sinus tachycardia and axis devia-tions were not considered to be abnormal findings sincethe former was very common and the latter had question-able clinical significance. We then analyzed the relation-ship between ECG abnormalities and demographic vari-ables and cardiovascular side effects. Next, we analyzedthe relationship between change in the corrected QT inter-val (QTc), the dosage of clozapine, and the duration ofclozapine use. Since QTc varies widely among individu-als, QTc changes from the baseline (before clozapine use)

were used as a measure. Because pimozide11,12 and thio-ridazine13–15 are well known for their ability to increaseQTc or induce torsade de pointes, patients who were tak-ing these drugs at the time of any ECG recording were ex-cluded from QTc analysis. Recordings from patients usingβ-blockers, antidepressants, or diuretics were also ex-cluded. None of the patients took sertindole, which is wellknown for its ability to increase the QT interval.16

RESULTS

Prevalence and Characteristics of ECG AbnormalitiesECG abnormalities present during the baseline re-

cording. Among the 61 patients, 8 showed baseline ECGabnormalities before the initiation of clozapine treatment.Among these, 6 were receiving antipsychotic medication.The prevalence of ECG abnormalities in the antipsychoticusers was 13.6% (6/44), while that in the antipsychotic-free patients was 11.8% (2/17), a difference that wasnot significant (p = .846, chi-square test; Table 1). Themean ± SD chlorpromazine-equivalent dose was not dif-ferent between normal (376 ± 428.5 mg/day) and abnor-mal (550 ± 597.5 mg/day) baseline ECG groups (p = .668,Student t test).

The abnormalities included nonspecific ST-T changes,conduction abnormalities, left atrial enlargement, leftventricular hypertrophy, and a pattern suggesting myocar-dial infarction. Follow-up ECGs, recorded after the startof clozapine use, were normal in 2 patients. The other pa-tients were found to have persistent abnormalities whenthe ECGs were rechecked after clozapine use for 30 daysto > 2 years. Two showed new-onset conduction abnor-malities during follow-up (Table 2).

Prevalence of ECG abnormalities in patients usingclozapine. The prevalence of ECG abnormalities in pa-tients using clozapine was 31.1% (19/61), including thosewith baseline ECG abnormalities but excluding those whoreverted to normal ECG after clozapine use. This resultwas significantly different from the baseline prevalenceof ECG abnormalities observed when the same groupof patients were treated with the other antipsychotics(p = .038, chi-square test; see Table 1). None of the pa-tients using clozapine were concomitantly using otherantipsychotics when ECGs were taken (Tables 2 and 3).

New-onset ECG abnormalities after clozapine use.Among the 53 patients with normal baseline ECG find-ings, new-onset ECG abnormalities after clozapine usewere detected in 13 (24.5%). In patients who had been us-ing antipsychotics during baseline recording, the preva-lence was 15.8% (6/38), whereas in drug-free patients, itwas 46.7% (7/15). This difference was statistically sig-nificant (p = .019, chi-square test; see Table 1). Of the 6patients who had been using antipsychotics and had ab-normal ECG at baseline, 4 had persistent ECG abnormali-ties after clozapine use.

Table 1. Prevalence of ECG Abnormalities*

Baseline ECG After Antipsychotics at Baseline ECG

Combination ECG Clozapine Yes No Total

1 Normal Normal 32 (a) 8 (b) 402 Normal Abnormal 6 (c) 7 (d) 133 Abnormal Normal 2 (e) 0 (f) 24 Abnormal Abnormal 4 (g) 2 (h) 6Total 44 (i) 17 (j) 61*Abbreviation: ECG = electrocardiogram.Prevalence of ECG abnormalities in users of antipsychotics other thanclozapine: k = (e + g)/i = 0.136.Prevalence of ECG abnormalities in antipsychotic-free patientsl = (f + h)/j = 0.118.Prevalence of ECG abnormalities in any clozapine user:m = (c + d + g + h)/(i + j) = 0.311.Prevalence of new-onset ECG abnormalities after clozapine use

In total patients (c + d)/(a + b + c + d) = 0.245.In patients who had been antipsychotic-free at baseline:n = d/(b + d) = 0.467.In patients who had been using other antipsychotics at baseline:o = c/(a + c) = 0.158.

No significant difference between k and l (p = .846, chi-square test).No significant difference between l and m (p = .111, chi-square test).Significant difference between k and m (p = .038, chi-square test).Significant difference between n and o (p = .019, chi-square test).

442


Kang et al.


One personal copy may be printedIn 6 cases, the abnormalities were of little clinical sig-nificance. Five cases showed nonspecific T wave abnor-mality, and 1 showed left atrial enlargement. However, theothers had potentially significant findings. Two casesshowed ventricular hypertrophy with ischemic patterns,3 showed findings suggesting inferior wall infarction,1 showed an acute pericarditis pattern, and 1 showed sig-nificant QTc prolongation (QTc = 533 ms). Nevertheless,none of these patients experienced clinically significantsubjective symptoms or objective signs of cardiovasculardisease requiring specific intervention. Apparently, theischemic, infarction, and pericarditis patterns were spuri-ous findings. In the patient with prolonged QTc, the QTcreverted to the normal range (430 ms) 1 month later, de-spite the patient’s having received a gradually increaseddosage of clozapine (see Table 3).

Timing and DosageMost of the new-onset abnormalities (10/13) were de-

tected during the upward titration period between 3 and 67days after clozapine treatment commencement, in the dos-

age range between 12.5 and 400 mg/day. In these cases,incremental titration of clozapine dosage was successfullyaccomplished without problems even after the onset ofECG abnormalities. The maintenance doses for those withand without ECG abnormalities were not different (318.1mg/day for normal vs. 298.1 mg/day for abnormal ECGgroups, p = .618, Student t test; Table 4). However, in 3patients, ECG abnormalities were first detected by inciden-tal ECG during the stable maintenance period. Their pre-vious recordings under the same or even higher doses werenormal. Follow-up data after the onset of ECG abnormali-ties were available for all patients, and in 10 of these 13patients, ECG findings reverted to normal despite the con-tinued use of clozapine. In those who developed ECG ab-normalities later during the maintenance period, only 1patient had normalized ECG on follow-up (see Table 3).

Demographic Factors Relatedto the New-Onset ECG Abnormalities

For male patients, the prevalence of new-onset ECGabnormalities after clozapine use was 20% (6/30) (see

Table 2. Patients With Baseline ECG Abnormalities Before Clozapine Treatmenta

Dose of Clozapine ConcomitantAge Drugs Used Follow-Up at Follow-Up Drugs at(y) Sex Abnormality at Baseline ECG Interval (mg/d) Follow-Up Follow-Up Result

27 M Ventricular preexcitation Pimozide 42 d 225 ... Normalized(WPW pattern type A)

27 F Left atrial enlargement Risperidone, clonazepam, 45 d 362.5 Lorazepam Normalizedlorazepam

43 M NST ... 25 d 300 ... NST, prolonged QTc41 d 450 ... NST, incomplete RBBB

26 M LVH Risperidone, diazepam > 1 y 350 Lorazepam Occasional VPC38 M Nonspecific conduction Sulpiride, pimozide, trihexy- 30 d 75 ... Persisted

abnormality phenidyl, clonazepam46 F NSST ... > 1 y 450 ... Persisted39 M R/O inferior wall infarction Haloperidol, benztropine > 10 mo 500 ... Persisted41 M LVH, first-degree AV block Haloperidol, carbamazepine > 2 y 550 Lorazepam PersistedaAbbreviations: ECG = electrocardiogram, LVH = left ventricular hypertrophy, NSST = nonspecific ST segment change, NST = nonspecific T wavechange, QTc = corrected QT interval, RBBB = right branch bundle block, R/O = rule out, VPC = ventricular premature contraction,WPW = Wolff-Parkinson-White.

Table 3. Patients With New-Onset ECG Abnormalities After Clozapine Treatmenta

Dose at ECG Dose atAge Days After Recording Concomitant Treatment Follow-Up Follow-Up(y) Sex Abnormality Clozapine Use (mg/d) Drugs Phase (mg/d) Result

54 F R/O inferior wall infarction 3 25 Clonazepam Titration 100 Normalized38 M Acute pericarditis 4 75 ... Titration 400 Normalized38 M R/O inferior wall infarction 20 125 ... Titration 325 Normalized47 M Left atrial enlargement 21 68.75 ... Titration 100 Normalized34 F RVH, ST-T abnormality 21 200 ... Titration 337.5 Normalized

(R/O inferior ischemia)18 F NST 30 175 Valproic acid Titration 250 Normalized28 F NST 31 225 Lorazepam Titration 400 Normalized30 F Prolonged QTc (533 ms) 39 137.5 ... Titration 200 Normalized31 M R/O inferior wall infarction 40 325 ... Titration 425 Normalized31 M NST 67 400 ... Titration 525 Persisted36 M LVH, ST elevation 141 550 ... Maintenance 262.5 Normalized49 F NST 219 112.5 ... Maintenance 112.5 Persisted36 F NST 408 250 ... Maintenance 250 PersistedaAbbreviations: RVH = right ventricular hypertrophy; for the other abbreviations, see Table 2.

443



445



Table 4). It was found to be 30.4% (7/23) in femalepatients. This difference between men and women wasnot statistically significant (p = .382, chi-square test). Themean age of those with new-onset ECG abnormalitiesafter clozapine use was 36.2 years (men, 36.8 years;women, 35.6 years), and the mean age of those withoutECG abnormalities was 29.3 years (men, 29.5 years;women, 28.8 years). Statistical analysis showed meaning-ful age differences for the total group (p = .028, Student ttest). When the subjects were separated into groups ofmen and women, this significance disappeared for women(p = .029 for men, p = .210 for women), probably owingto insufficient sample numbers.

Relationship to Cardiovascular Side EffectsOrthostatic hypotension and tachycardia were ana-

lyzed for a possible relationship to the occurrence of ECGabnormalities (Table 5). Neither of them was found to besignificantly related to new-onset ECG abnormalities af-ter clozapine use.

QTc Change With ClozapineWe analyzed the effect of clozapine on QTc using lin-

ear regression analysis of the QTc change from the base-line recording (∆QTc) and the dosage of clozapine andduration of clozapine use. We excluded data from patientswho had been taking β-blockers, antidepressants, diuret-ics, pimozide, or thioridazine at any recording. Excludingbaseline recordings (where ∆QTc, duration, and dose

were zero by definition), 165 recordings from 51 patientswere analyzed. We found a significant positive linearrelationship between the dosage of clozapine and ∆QTc(β = .029, t = 2.168, p = .032, Figure 1A). When the ap-parent “outlier” (see Figure 1A, asterisk) was excluded,the significance increased (β = .033, t = 2.558, p = .011).However, prolongation of QTc into the abnormal rangewas found in only 2 patients (patient 3 in Table 2 and pa-tient 8 in Table 3). The duration of clozapine use had nosignificant relationship to ∆QTc (p = .565, Figure 1B).When recordings from patients with normal baselineECGs were selected (total of 142 recordings from 45 pa-tients), the relationship between the dosage and ∆QTc be-came more significant (β = .247, t = 2.894, p = .004).

DISCUSSION

To date, no systematic studies on ECG changes inpatients treated with clozapine have been published. How-ever, many case reports suggest that clozapine causes vari-ous cardiac effects. Since our study was naturalistic andretrospective, we could not assign control groups, andregular data on plasma concentrations of clozapine wereunavailable. Although we had guidelines for ECG check-up during clozapine use, it was not a strict or mandatoryprotocol. These factors caused limitations in terms of theinterpretation of our data. However, the patients’ baselineECG data served as a control, and our study may beregarded as natural crossover in design. The researcherswere blind to ECG status during case selection, and no sys-tematic bias during the selection of the subjects is likely.Therefore, we believe that this study gives a reasonableestimate of the general characteristics of clozapine-induced ECG abnormalities.

According to this study, a substantial portion (31.1%)of patients using clozapine had ECG abnormalities otherthan sinus tachycardia or axis deviations, especially at theinitial stage of treatment. Other classical antipsychoticsalso cause a high incidence of ECG abnormalities. Chlor-promazine at as low a dosage as 125 mg/day may induceECG abnormalities in 34.1% of patients,17 a result whichis similar to ours. However, a direct comparison could notbe performed because different criteria were used to de-fine ECG abnormality.

Table 4. Demographic Profiles and Dosage of Clozapinea

Variable Women Men Total

Total N 25 36 61Age, mean, y 31.3 31.8 31.6b

Maintenance dose, mg/d 310.0 336.8 325.8c

Abnormal baseline ECG, N 2 6 8Age, mean, y 36.5 35.7 35.9d

Maintenance dose, mg/d 450.0 395.8 409.4e

Normal baseline ECG, N 23 30 53Age, mean, y 30.9 31.0 30.9d

Maintenance dose, mg/d 297.8 325.0 313.2e

Normal baseline andfollow-up ECG, N 16 24 40f

Age, mean, y 28.8 29.5 29.3g

Maintenance dose, mg/d 317.2 318.7 318.1h

New-onset ECG abnormalityafter clozapine, N 7 6 13f

Age, mean, y 35.6 36.8 36.2g

Maintenance dose, mg/d 253.6 350.0 298.1h

aAbbreviation: ECG = electrocardiogram.bNo significant difference between men and women (p = .838, Student t test).cNo significant difference between men and women (p = .373, Student t test).dNo significant difference between those with normal and abnormal baselineECG (p = .189, Student t test).eNo significant difference between those with normal and abnormal baselineECG (p = .089, Student t test).fNo significant difference in the prevalence of new-onset ECG abnormalitiesbetween male and female (p = .382, chi-square test).gSignificant difference between those with and without new-onset ECGabnormalities after clozapine use (p = .028, Student t test).hNo significant difference between those with and without new-onset ECGabnormalities after clozapine use (p = .688, Student t test).

Table 5. Relationship of New-Onset ECG Abnormalities toCardiovascular Side Effectsa

OrthostaticHypotension* Tachycardia†

Condition Absent Present Absent Present

No ECG abnormalities(N = 40) 32 8 18 22

ECG abnormalities (N = 13) 8 5 6 7aAbbreviation: ECG = electrocardiogram.*p = .178, chi-square test.†p = .942, chi-square test.

444


Kang et al.



Our baseline ECG data showed the prevalence of ECGabnormalities in the other antipsychotic users at a level of13.6%, which was found to increase significantly afterclozapine use. This suggested that clozapine could induceECG abnormalities more frequently than other antipsy-chotics. However, since the prevalence of ECG abnor-malities in the clozapine users decreased after prolongeduse and we could not control the duration of antipsychoticuse in the baseline recording, we cannot definitely statewhether clozapine usage causes a higher rate of long-lasting ECG abnormalities.

Among those patients with normal baseline ECG find-ings, 24.5% of patients showed new-onset ECG abnor-malities after clozapine use. Those without baseline anti-psychotic medication had a particularly higher incidenceof new-onset ECG abnormalities after clozapine use,which suggested that those showing no ECG abnormali-ties with previous antipsychotics may also have a lowerprobability of developing abnormal ECGs after clozapinetreatment.

In our sample, nonspecific abnormality in the T wavewas the most common finding. This suggests that cloza-pine affects cardiac repolarization. Abnormal patternssuggesting ischemic heart disease were also found, butthe clinical status of the patients suggested that these find-ings were spurious. A similar case has been reported byKetch et al.9 These false-positive ischemic patterns oc-curred relatively frequently in our cohort. However, thisfinding may not be unique to clozapine, since similarfindings have been described in patients with heterocyclicantidepressant overdoses.18 Acute pericarditis after cloza-pine treatment has also been reported,6 and 1 of our pa-tients did produce ECG results suggestive of acute peri-carditis. However, this too proved to be a false-positive

finding, since the patient had no signs of cardiac prob-lems. In fact, most of the new-onset ECG abnormalitieswere benign.

Most ECG abnormalities were detected during thetitration phase when the dosage was being increased. Ab-normalities were found as early as after a few days ofcommencement and at doses as low as 25 mg/day. How-ever, there seemed to be an adaptation to the electro-cardiographic effects of clozapine, and by cautiously in-creasing the dose, most patients could tolerate muchhigher maintenance dosages without showing persistentECG abnormalities. In many cases, ischemia-like patternsreverted to normal during follow-up. A case with in-creased QTc showed no clinically significant arrhythmia,and QTc also reverted to the normal range during follow-up. Other changes such as nonspecific T wave changesand ventricular hypertrophy also tended to revert to nor-mal on follow-up. However, some patients developednew-onset ECG abnormalities long after a stable mainte-nance dosage was reached.

Although women carry an increased risk of drug-induced cardiac arrhythmia,19 this gender effect was notobserved for clozapine-induced ECG changes. It is wellknown that the geriatric population is at an increased riskof medication-induced cardiac problems. Although noneof our patients were in the geriatric age group, increasedage may be a risk factor for the development of ECG ab-normalities after clozapine use. Clinically detectable car-diovascular side effects such as orthostatic hypotension orsinus tachycardia were not related to ECG abnormalities.

Clozapine increased QTc in a dose-dependent manner.Increased QTc in patients treated with other antipsy-chotics has also been reported,16,20–22 which suggests thatantipsychotics may induce fatal arrhythmias, such as

aMultiple ECG recordings from 51 patients were pooled. Linear regression analysis showed a significant positive linear relationship between thedosage of clozapine and ∆QTc (β = .029, t = 2.168, p = .032).bLinear regression analysis failed to show a linear relationship between the duration of clozapine use and ∆QTc (p = .565).* = outlier.

Figure 1. Relationship Between the Change in Corrected QT Interval (∆QTc) and (A) the Daily Dosage of Clozapine and(B) the Duration of Clozapine Usea,b

150

100

50

0

–50

150

100

50

0

–500 100 200 300 400 500 600 700 0 30 60 90 120 150 180

*

Dose of Clozapine (mg/d) Duration of Clozapine Use (d)

∆QT

c (m

s)

∆QT

c (m

s)

A B

445



447



torsade de pointes.13–15 Although no reports exist of fatalarrhythmia in patients using clozapine, our results showedthat use of clozapine might be a risk factor for the devel-opment of potentially fatal arrhythmia. However, in oursample of 61 patients, abnormally prolonged QTc wasrare and was found only in 2 patients, one of whom hadbaseline ECG abnormality whereas the other showed pro-longed QTc that normalized despite the continued use ofclozapine. Although the duration of clozapine use was ap-parently related to the increase in QTc, this was not statis-tically significant and seemed to be related to the fact thatthe dosage of clozapine was increased over time.

We have recently observed a nonfatal ventriculartachycardia in a patient using clozapine, but in this patientQTc was within the normal range just a few hours beforethe onset of ventricular tachycardia (this patient had beenreferred to our clinic after the beginning of the data analy-sis and was thus excluded from the analysis). A case offatal arrhythmia not preceded by increased QTc has alsobeen reported in a haloperidol user.23 Because the QTcconstantly fluctuates, a single normal ECG may not ruleout the possible prolongation.22 Moreover, recent findingscast doubt on the usefulness of QTc measurement as apredictor of fatal arrhythmia.21,22 Various methods ofmodifying the correction of QT24,25 or QT dispersion26

have been believed to be more valuable for the predictionof arrhythmia. These variables have not been analyzed inthis article. Thus, more studies may be needed to elucidatethe potential arrhythmogenecity of clozapine.

In conclusion, a substantial number of patients treatedwith clozapine developed ECG abnormalities. Most ofthese abnormalities occurred during the initial stage oftreatment. Increased age was the main risk factor fornew-onset ECG abnormalities after clozapine treatment,whereas normal ECG with previous antipsychotics maylower the risk. Most of the abnormalities were essentiallybenign and subsided despite the continued use or even in-creased dosages of clozapine. Therefore, we believe thatECG abnormalities generally do not hinder clozapinetreatment if treatment is accompanied by judicious mon-itoring. Although clozapine increased QTc in a dose-dependent fashion, most of the QTc measurements werestill within the normal range. Thus, the significance ofQTc as a predictor of the future occurrence of ventriculararrhythmia in patients taking clozapine is uncertain.

Drug names: benztropine (Cogentin and others), carbamazepine (Teg-retol and others), chlorpromazine (Thorazine and others), clonazepam(Klonopin and others), clozapine (Clozaril and others), diazepam (Val-ium and others), haloperidol (Haldol and others), lorazepam (Ativanand others), loxapine (Loxitane and others), pimozide (Orap), risperi-done (Risperdal), thioridazine (Mellaril and others), trihexyphenidyl(Artane and others), valproic acid (Depakene).

REFERENCES

1. Liberman JA, Alvir JMJ. A report of clozapine induced agranulocytosisin the United States: incidence and risk factors. Drug Saf 1992;7(suppl 1):1–2

2. Honigfeld G. Effects of the clozapine national registry system on inci-dence of deaths related to agranulocytosis. Psychiatr Serv 1996;47:52–56

3. Ennis LM, Parker RM. Paradoxical hypertension associated with cloza-pine [letter]. Med J Aust 1997;166:278

4. George TP, Winther LC. Hypertension after initiation of clozapine. Am JPsychiatry 1996;153:1368–1369

5. Shiwach RS. Treatment of clozapine induced hypertension and possiblemechanisms. Clin Neuropharmacol 1998;21:139–140

6. Khalil N. Peri-/Myokarditis unter Clozapin (Leponex). Psychiatr Prax1997;24:46

7. Bandelow B, Degner D, Kreusch U, et al. Myocarditis under therapy withclozapine. Schizophr Res 1995;17:293–294

8. Leo RJ, Kreeger JL, Kim KY. Cardiomyopathy associated with clozapine.Ann Pharmacother 1996;30:603–605

9. Ketch J, Herd A, Ludwig L. ST segment elevations without myocardialinfarction in a patient on clozapine [letter]. Am J Emerg Med 1996;14:111–112

10. Beale MD, Pritchett JT, Kellner CH. Supraventricular tachycardia in apatient receiving ECT, clozapine, and caffeine. Convuls Ther 1994;10:228–231

11. Desta Z, Kerbusch T, Flockhart DA. Effect of clarithromycin on the phar-macokinetics and pharmacodynamics of pimozide in healthy poorand extensive metabolizers of cytochrome P450 2D6 (CYP2D6). ClinPharmacol Ther 1999;65:10–20

12. Tueth MJ, Cheong JA. Clinical uses of pimozide. South Med J 1993;86:344–349

13. Schmidt W, Lang K. Life-threatening dysrhythmias in severe thioridazinepoisoning treated with physostigmine and transient atrial pacing. Crit CareMed 1997;25:1925–1930

14. Roden DM. Torsade de pointes. Clin Cardiol 1993;16:683–68615. Liberatore MA, Robinson DS. Torsade de pointes: a mechanism for sud-

den death associated with neuroleptic drug therapy? J Clin Psychophar-macol 1984;4:143–146

16. Brown GR, Radford JM. Sertindole hydrochloride: a novel antipsychoticmedication with a favorable side effect profile. South Med J 1997;90:691–693

17. Chouinard G, Annable L, Melancon, et al. Chlorpromazine-induced elec-trocardiogram abnormalities. Int J Clin Pharmacol Biopharm 1975;11:327–331

18. Bolognesi R, Tsialtas D, Vasini P, et al. Abnormal ventricular repolariza-tion mimicking myocardial infarction after heterocyclic antidepressantoverdose. Am J Cardiol 1997;79:242–245

19. Ebert SN, Liu XK, Woosley RL. Female gender as a risk factor for drug-induced cardiac arrhythmias: evaluation of clinical and experimental evi-dence. J Womens Health 1998;7:547–557

20. Warner JP, Barnes TR, Henry JA. Electrocardiographic changes in patientsreceiving neuroleptic medication. Acta Psychiatr Scand 1996;93:311–313

21. Sramek JJ, Cutler NR, Kurtz NM, et al. Optimizing the Development ofAntipsychotic Drugs. Chichester, England: John Wiley & Sons; 1997:53–99

22. Stahl SM. Psychopharmacology of Antipsychotics. London, England:Martin Dunitz Ltd; 1999

23. Jackson T, Ditmanson L, Phibbs B. Torsade de pointes and low-dose oralhaloperidol. Arch Intern Med 1997;157:2013–2015

24. Fei L, Statters DJ, Anderson MH, et al. Is there an abnormal QT interval insudden cardiac death survivors with a “normal” QTc? Am Heart J 1994;128:73–76

25. Funck-Brentano C, Jaillon P. Rate-corrected QT interval: techniques andlimitations. Am J Cardiol 1993;72:17B–22B

26. Stoletniy LN, Pai SM, Platt ML, et al. QT dispersion as a noninvasivepredictor of inducible ventricular tachycardia. J Electrocardiol 1999;32:173–177

446

https://www.researchgate.net/publication/14374833_Hypertension_after_initiation_of_clozapine?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14374833_Hypertension_after_initiation_of_clozapine?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14746988_Clinical_Uses_of_Pimozide?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14746988_Clinical_Uses_of_Pimozide?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14570700_ST_segment_elevations_without_myocardial_infarction_in_a_patient_on_clozapine_Letter?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/14537145_Myocarditis_under_therapy_with_clozapine_Letter?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14537145_Myocarditis_under_therapy_with_clozapine_Letter?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14137518_Paradoxical_hypertension_associated_with_clozapine?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14137518_Paradoxical_hypertension_associated_with_clozapine?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14942043_Rate-corrected_QT_interval_techniques_AND_limitations?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14942043_Rate-corrected_QT_interval_techniques_AND_limitations?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13993503_Sertindole_Hydrochloride_A_Novel_Antipsychotic_Medication_With_a_Favorable_Side_Effect_Profile?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/14490385_Electrocardiographic_changes_in_patients_receiving_neuroleptic_medication?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14490385_Electrocardiographic_changes_in_patients_receiving_neuroleptic_medication?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13634933_Female_Gender_as_a_Risk_Factor_for_Drug-Induced_Cardiac_Arrhythmias_Evaluation_of_Clinical_and_Experimental_Evidence?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/13705306_Treatment_of_clozapine_induced_hypertension_and_possible_mechanisms?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13705306_Treatment_of_clozapine_induced_hypertension_and_possible_mechanisms?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13862657_Life-threatening_dysrhythmias_in_severe_thioridazine_poisoning_treated_with_physostigmine_and_transient_atrial_pacing?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/15365483_Supraventricular_tachycardia_in_a_patient_receiving_ECT_clozapine_and_caffeine?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/14282507_Effects_of_the_clozapine_national_registry_system_on_incidence_of_deaths_related_to_agranulocytosis?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14282507_Effects_of_the_clozapine_national_registry_system_on_incidence_of_deaths_related_to_agranulocytosis?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13912596_Torsades_de_pointes_and_low_dose_oral_Haloperidol?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13912596_Torsades_de_pointes_and_low_dose_oral_Haloperidol?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14410110_Cardiomyopathy_associated_with_clozapine?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14410110_Cardiomyopathy_associated_with_clozapine?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14026289_Abnormal_Ventricular_Repolarization_Mimicking_Myocardial_Infarction_After_Heterocyclic_Antidepressant_Overdose?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/17040046_Torsade_de_pointes_A_mechanism_for_sudden_death_associated_with_neuroleptic_drug_therapy?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/22056140_Chlorpromazine_induced_electrocardiogram_abnormalities?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1




Letters to the Editor




WATCHERS:Recognizing Generalized Anxiety Disorder

Sir: The recent article by Davidson et al.1 raises awarenessof generalized anxiety disorder (GAD), one of the most com-monly diagnosed psychiatric illnesses in the United States, withan estimated life prevalence of 5.1%. To improve recognition ofGAD, I have introduced the mnemonic WATCHERS, whichdescribes the core symptoms of this condition as given in theDiagnostic and Statistical Manual of Mental Disorders, FourthEdition.2

The first 2 letters signify the Worry and Anxiety about eventsor activities that are difficult to control and that impair function-ing for at least 6 months. The remaining letters reflect the 6symptoms of which at least 3 must be present to diagnose thiscommon condition: Tension in muscles, Concentration diffi-culty, Hyperarousal (or irritability), Energy loss, Restlessness,and Sleep disturbance. It is hoped that increased recognition ofGAD will lead to better management of these patients with im-provement in well-being and quality of life.

REFERENCES

1. Davidson JRT, DuPont RL, Hedges D, et al. Efficacy, safety, andtolerability of venlafaxine extended release and buspirone in outpa-tients with generalized anxiety disorder. J Clin Psychiatry 1999;60:528–535

2. American Psychiatric Association. Diagnostic and Statistical Manualof Mental Disorders, Fourth Edition. Washington, DC: AmericanPsychiatric Association; 1994:432–436

Mark J. Berber, M.B., M.R.C.Psych., F.R.C.P.C.Markham Stouffville Hospital

Markham, Ontario, Canada

Gold Versus Silver: The Issue of FunctionalVersus Symptomatic Recovery in Depression

Sir: The BRAINSTORMS article in the April 1999 issue of theJournal by Dr. Stahl1 emphasizes an important yet frequentlyneglected aspect of clinical psychiatry, i.e., undertreatment ofmood disorders. Like most of his articles, this is a no-nonsensepiece highlighting the psychopharmacologic angle, especiallyfrom a clinician’s viewpoint. However, the following points areto be considered while appreciating his arguments:

First, the term recovery may be divided into at least 2 cat-egories, i.e., symptomatic and functional. Symptomatic recov-ery, sometimes equated with response to treatment, consists ofimprovement in psychopathology as assessed through com-monly used rating scales such as the Hamilton Rating Scale forDepression. Functional recovery, on the other hand, representstrue recovery and well-being of the patient, which are not prop-erly assessed through conventional rating scales.

Next, the assumption that psychopharmacologic interventionscan bring sustained resolution of symptoms in depressive disor-ders has not been fully substantiated. On the other hand, there isan increasing body of research to support the view that depres-sion might be responsible for permanent changes in personality,2,3

residual features, psychosocial dysfunction, and disability thatmay persist long after successful completion of pharmaco-therapy.4,5 Combined pharmacotherapy and psychosocial inter-vention strategies may be more effective than either modalityalone, especially if administered at a proper treatment stage.

Finally, in the current context, Rush6 made a similar state-ment in one of his editorials concerning the need for differ-entiating partial recovery from complete remission in mooddisorders. He has compared partial recovery in depression tolowered blood pressure (as opposed to normotension) in hyper-tension or partial control of blood sugar (as opposed to full con-trol) in diabetes mellitus.

REFERENCES

1. Stahl SM. Why settle for silver, when you can go for gold? responsevs recovery as goal of antidepressant therapy [BRAINSTORMS]. J ClinPsychiatry 1999;60:213–214

2. Angst J, Kupfer DJ, Rosenbaum JF. Recovery from depression: riskor reality? Acta Psychiatr Scand 1996;93:413–419

3. Shea T, Leon A, Muller T, et al. Does major depression result in last-ing personality change? Am J Psychiatry 1996;153:1404–1410

4. Fava GA, Grandi S, Zielenzny M, et al. Cognitive-behavioral treat-ment of residual symptoms in primary major depressive disorders.Am J Psychiatry 1994;151:1295–1299

5. Pradhan SC, Sinha VK, Singh TB. Psycho-social dysfunction inpatients after recovering from mania and depression. Int J RehabRes 1999;22:303–309

6. Rush AJ. Assessing outcome in practice: a paradigm shift? Curr OpinPsychiatry 1996;9:1–2

Subas Chandra Pradhan, M.D.Institute of Human Behavior & Allied Sciences

Delhi, India

Dr. Stahl Replies

Sir: Dr. Pradhan provides thoughtful input regarding the ap-propriate threshold of treatment response to target for patientswith affective and anxiety disorders. By reminding us that scor-ing low on a rating scale is not the same thing as being “well,”his comments underscore the need for clinicians to have differ-ent outcome targets (i.e., true recovery and well-being) thanclinical investigators, whose studies target 50% or greater re-duction of symptoms.1 After all, who would accept a 50% reduc-tion of infectious organisms for an antibiotic or a 50% reductionof tumor cells in malignancies as appropriate outcome targets inthese areas of medicine?

Dr. Pradhan also adds the important perspective to my ar-ticle1 that psychotherapy combined with pharmacotherapy may

447

https://www.researchgate.net/publication/232113390_Assessing_outcome_in_practice_A_paradigm_shift?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/232113390_Assessing_outcome_in_practice_A_paradigm_shift?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14373236_Recovery_from_depression_Risk_or_reality?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14373236_Recovery_from_depression_Risk_or_reality?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15130731_Cognitive_behavioural_treatment_of_residual_symptoms_in_primary_depressive_disorder?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15130731_Cognitive_behavioural_treatment_of_residual_symptoms_in_primary_depressive_disorder?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1




https://www.researchgate.net/publication/297904455_Efficacy_safety_and_tolerability_of_venlafaxine_extended_release_and_buspirone_in_outpatients_with_generalized_anxiety_disorder?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1




https://www.researchgate.net/publication/44413696_Diagnostic_and_Statistical_Manual_of_Mental_Disorders_Fourth_Edition_DSM-IV?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1







be the ultimate route to a complete recovery from depression,given the psychosocial “damage” that this illness can cause interms of personality, interpersonal dysfunction, vocational dis-ability, and even increased liability to relapse and nonresponseto future antidepressant treatment. This idea is especiallytimely, since interesting new studies2,3 have now appeared sub-sequent to the publication of my article that show for the firsttime that psychotherapy can indeed be synergistic with antide-pressant treatments, a point well known to practicing clinicians,but only recently being proved in controlled trials.

The point of all this discussion is that clinicians need to“raise the bar” for our expectations of treatment outcomes andnot “settle for silver when we can go for gold” for our patientswith affective and anxiety disorders. We are currently in an ex-citing era when the use of antidepressants, psychotherapy, com-binations of antidepressants, or combinations of antidepressantsand psychotherapy in a sequential pattern are capable of extin-guishing symptoms of depression and thus delivering not onlysymptomatic relief, but also functional recovery in a large pro-portion of depressed patients.

REFERENCES

1. Stahl SM. Why settle for silver, when you can go for gold? responsevs. recovery as the goal of antidepressant therapy [BRAINSTORMS].J Clin Psychiatry 1999;60:213–214

2. Reynolds CF, Frank E, Perel JM, et al. Nortriptyline and interpersonalpsychotherapy as maintenance therapies for recurrent majordepression: a randomized controlled trial in patients older than59 years. JAMA 1999;281:39–45

3. Keller MB, McCullough JP, Rush AJ, et al. Nefazodone HCl, cog-nitive behavioral analysis system of psychotherapy and combinationtherapy for the acute treatment of chronic depression. In: NewResearch Program and Abstracts of the 152nd Annual Meetingof the American Psychiatric Association; May 19, 1999;Washington, DC. Abstract NR405:178

Stephen M. Stahl, M.D., Ph.D.Clinical Neuroscience Research Center and

University of California, San DiegoSan Diego, California

SIADH With Multiple Antidepressantsin a Geriatric Patient

Sir: The syndrome of inappropriate secretion of antidiuretichormone (SIADH) is a known side effect of psychopharmaco-therapy, having been reported with nearly all psychotropicdrugs.1 The mechanism of drug-induced SIADH is thought to bemediated serotonergically, but the mechanism is not known forcertain at this time.2 It is often difficult to predict which patientswill develop this complication, although certain risk factors areknown, such as age ≥ 65 years, concomitant diuretic use, andsmoking.3,4 The following case illustrates a propensity towardSIADH in a single patient.

Case report. Ms. A, a 70-year-old white woman, was admit-ted to our inpatient service for electroconvulsive therapy (ECT)for recurrent severe major depressive disorder with psychoticfeatures (DSM-IV criteria). At the time of admission, she wasbeing treated with mirtazapine, 15 mg daily; risperidone, 3 mgdaily; and trazodone, 100 mg at bedtime. Serum sodium con-centration at the time of admission was 126 mmol/L (allsamples drawn at 6:00 a.m.). Ms. A’s urine and serum osmolar-ity, as well as urine sodium concentration, was consistent with

SIADH. Her psychotropic medications were rapidly tapered inpreparation for the ECT. Within 4 or 5 days, her sodium concen-tration returned to the 132- to 134-mmol/L range, where it re-mained throughout the duration of ECT.

Ms. A responded well to ECT, and after her ninth treatment,she was started on venlafaxine (immediate-release form), 37.5mg b.i.d., which was increased to 75 mg b.i.d. after 2 days. Twodays later, her serum sodium concentration dropped to 131mmol/L and further decreased to 125 mmol/L after 2 more days.Venlafaxine was immediately discontinued, fluid restrictionwas instituted, and her sodium concentration promptly returnedto 135 mmol/L. Of note, at no time during the hospitalizationwas there evidence of increased fluid intake by Ms. A.

A trial of bupropion (immediate-release form), 75 mg b.i.d.,was then initiated in the hope of avoiding the serotonin system.After 4 days of therapy, Ms. A’s sodium concentration remainedwithin normal limits, and she was discharged to our GeriatricPsychiatry Day Hospital. Two weeks after discharge, her serumsodium concentration was 137 mmol/L and she was doing well.

A review of Ms. A’s chart from her previous admission 6months prior to this one revealed that her admission sodiumconcentration, while she was being treated with desipramine,was 126 mmol/L, which rapidly corrected upon cessation ofdesipramine treatment. No other laboratory data are availablefor Ms. A.

There are a number of interesting points to this case. Firstly,it is one of the few reported cases of venlafaxine-inducedSIADH.5–8 Secondly, it highlights the fact that not only do idio-syncratic interactions exist between a particular agent and a par-ticular patient leading to SIADH, but certain individuals may beparticularly predisposed to the development of SIADH while ontreatment with a variety of agents. This patient had documentedSIADH on treatment with mirtazapine and risperidone, devel-oped rapid onset of hyponatremia on venlafaxine treatment, andhad documented hyponatremia on desipramine treatment thatreversed rapidly upon cessation of desipramine. Further studyof patients predisposed to the development of SIADH may helpus understand the mechanism of medication-induced SIADH.This case also serves to remind clinicians that in assessing amental status change after the initiation of a new psychotropictreatment, SIADH must be on the differential diagnosis, par-ticularly in patients with known risk factors for the developmentof SIADH.

The final intriguing aspect of this case is the possibility thatbupropion may be an agent that has less of a propensity to in-duce SIADH than other antidepressants. As mentioned above,some evidence suggests that antidepressant-induced SIADH ismediated through the serotonergic system. Bupropion is thoughtto have exclusive activity in the dopaminergic system and tocompletely avoid serotonin.9 Further research is necessary toinvestigate whether or not this is a unique property of bupropionthat can be used to clinical advantage.

REFERENCES

1. Spigset O, Hedenmalm K. Hyponatremia and the syndrome ofinappropriate antidiuretic hormone secretion (SIADH) inducedby psychotropic drugs. Drug Saf 1995;12:209–225

2. Spigset O, Mjorndal T. The effect of fluvoxamine on serum prolactinand serum sodium concentrations: relation to platelet 5-HT2A receptorstatus. J Clin Psychopharmacol 1997;17:292–297

3. Liu BA, Mittman N, Knowles SR, et al. Hyponatremia and thesyndrome of inappropriate secretion of antidiuretic hormone associ-ated with the use of selective serotonin reuptake inhibitors: a reviewof spontaneous reports. Can Med Assoc J 1996;155:519–527

4. Settle EC Jr. Antidepressant drugs: disturbing and potentially

448

https://www.researchgate.net/publication/14400981_Hyponatremia_and_the_syndrome_of_inappropriate_secretion_of_antidiuretic_hormone_associated_with_the_use_of_Selective_Serotonin_Reuptake_Inhibitors_a_review_of_spontaneous_reports?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1




https://www.researchgate.net/publication/13978896_The_effect_of_fluvoxamine_on_serum_prolactin_and_serum_sodium_concentrations_relation_to_platelet_5-HT2A_receptor_status?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1







dangerous adverse effects. J Clin Psychiatry 1998;59(suppl 16):25–30 5. Gupta AK, Saravay SM. Venlafaxine-induced hyponatremia.

J Clin Psychopharmacol 1997;17:223–235 6. Meynaar IA, Peeters AJ, Mulder AH, et al. Syndrome of inappropri-

ate ADH secretion attributed to the serotonin re-uptake inhibitors,venlafaxine and paroxetine. Neth J Med 1997;50:243–245

7. Ranieri P, Franzoni S, Rozzini R, et al. Venlafaxine-induced resetosmostat syndrome: case of a 79-year-old depressed woman.J Geriatr Psychiatry Neurol 1997;10:75–78

8. Masood GR, Karki SD, Patterson WR, et al. Hyponatremia withvenlafaxine. Ann Pharmacother 1998;32:49–51

9. Davidson JRT, Connor KM. Bupropion sustained release: atherapeutic overview. J Clin Psychiatry 1998;59(suppl 4):25–31

David M. Blass, M.D.Vincent E. Pearson, Pharm.D.

Johns Hopkins HospitalBaltimore, Maryland

Free Drug Fraction VersusFree Drug Concentration

Sir: In their article on pharmacokinetic considerations in theelderly, DeVane and Pollock1 present a wealth of valuable infor-mation for the clinician. However, a small point needs clarifica-tion. The authors state—as do many textbooks and articles onthis subject—that an increase in plasma binding proteins in theelderly “can lead to a diminished free concentration of basicdrugs” (italics mine). This notion is also repeated in Table 1 intheir article. In fact, however, changes in plasma binding pro-teins (such as α-acid glycoprotein) have no lasting effect on thefree concentration of a drug, which is the absolute amount ofthe drug available to affect target sites and cause clinical effects.Rather, increases or decreases in binding proteins alter the freefraction of drug, i.e., the proportion that is not bound to protein.

Alterations in free fraction change how we interpret thetherapeutic and toxic ranges for the agent in question, but do notmaterially affect the actual amount of the drug available to acton target organs.2,3 In practice, this means that an elderly patientwith elevated α-acid glycoprotein levels would show increasedtotal (free plus protein-bound) drug concentration, decreasedfree fraction of the drug, and an unchanged free drug concentra-tion. It would be a mistake, in such a case, to reduce the dose ofthe agent simply on the basis of an elevated total drug con-centration; rather, the toxic range for the drug would need to beadjusted upward.2,4 All this stems from the fact that the free con-centration of a drug is dependent solely on the dosing rate andthe intrinsic clearance of the agent. For confused clinicians, thearticle by Greenblatt et al.2 provides some helpful diagrams onthis point.

REFERENCES

1. DeVane CL, Pollock BG. Pharmacokinetic considerations of antide-pressant use in the elderly. J Clin Psychiatry 1999;60(suppl 20):38–44

2. Greenblatt DJ, Sellers EM, Koch-Weser J. Importance of proteinbinding for the interpretation of serum or plasma drug concentrations.J Clin Pharmacol 1982;22:259–263

3. Hansten PD, Horn JR. Mechanisms of drug interactions: proteinbinding displacement. Drug Interact Newsletter 1989;9:449–456

4. Pies RW. Handbook of Essential Psychopharmacology. Washington,DC: American Psychiatric Press; 1999:95–96

Ronald Pies, M.D.Tufts University School of Medicine

Lexington, Massachusetts

Drs. DeVane and Pollock Reply

Sir: We thank Dr. Pies for his thoughtful comments andagree that the free, non–protein-bound concentration of drugs inplasma is more closely correlated with drug effect than totalconcentration. In addition, for many drugs and situations, in-cluding drug-drug interactions that result in displacement ofdrugs from plasma protein binding sites, the increase in freedrug concentration is transient. An exception to this generaliza-tion would be a drug with a high hepatic intrinsic clearancewhose plasma protein binding is nonrestrictive for eliminationand that is administered intravenously.1 Few drugs are charac-terized in this manner. For many drugs therapeutically useful inpsychiatry, a change in drug binding to plasma proteins resultsin an altered free fraction that is offset by a change in total drugclearance by the liver, resulting ultimately in an altered totaldrug concentration in plasma but a free drug concentration simi-lar to that which existed before a change in binding. We in-tended to simplify this often confusing point in our summary ofthe effects of aging on the disposition of antidepressants,2 andwe regret any unintentional implication that aging results in asustained change in free drug concentration. We have previ-ously emphasized the importance of free drug concentrationmonitoring3,4; unfortunately, the techniques to do this are notwidely available.

REFERENCES

1. Wilkinson GR, Shand DG. A physiological approach to hepatic drugclearance [commentary]. Clin Pharmacol Ther 1975;18:377–390

2. DeVane CL, Pollock BG. Pharmacokinetic considerations of antide-pressant use in the elderly. J Clin Psychiatry 1999;60(suppl 20):38–44

3. Pollock BG. Geriatric psychiatry: psychopharmacology: generalprinciples. In: Sadock BJ, Sadock VA, eds. Kaplan & Sadock’sComprehensive Textbook of Psychiatry, VII. Baltimore, Md:Williams & Wilkins; 2000:3086–3090

4. DeVane CL. Principles of pharmacokinetics and pharmacodynamics.In: Schatzberg AF, Nemeroff CB, eds. American Psychiatric PressTextbook of Psychopharmacology. 2nd ed. Washington, DC:American Psychiatric Press; 1998:155–169

C. Lindsay DeVane, Pharm.D.Medical University of South Carolina

Charleston, South CarolinaBruce G. Pollock, M.D., Ph.D.

Western Psychiatric Institute and ClinicPittsburgh, Pennsylvania

Zolpidem-Related Delirium:A Case Report

Sir: Zolpidem is a nonbenzodiazepine hypnotic that is usedin the treatment of short-term insomnia. Although zolpidem isgenerally safe and well tolerated, a number of cases of percep-tual disturbances and psychosis have been described.1,2 At least2 cases of probable zolpidem-related delirium have been re-ported.3,4 We report a case of delirium that developed in a pa-tient with low-grade fever after zolpidem was added to herpsychotropic medications.

Case report. Ms. A, a 26-year-old woman, was treated at apsychiatric inpatient unit for psychotic depression. She had noformal thought disorder, showed no perceptual disturbances,and was cognitively intact. Ten days into the hospitalization,while stabilized on fluoxetine, 20 mg p.o. q.a.m.; risperidone, 3

449



https://www.researchgate.net/publication/21990472_A_physiological_approach_to_hepatic_drug_clearance_Clin_Pharm_Ther?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/21990472_A_physiological_approach_to_hepatic_drug_clearance_Clin_Pharm_Ther?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/16091028_Importance_of_Protein_Binding_for_the_Interpretation_of_Serum_or_Plasma_Drug_Concentrations?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1







mg p.o. b.i.d.; and benztropine, 1 mg p.o. b.i.d., Ms. A devel-oped flu-like symptoms with a sore throat, watery eyes, mal-aise, and a temperature of 99.2°F (37.3°C). A mild increase inmuscle tone had been detected after the initiation of risperidonethat did not resolve after benztropine was added. Three days af-ter the onset of the viral syndrome, she requested zolpidem forsleep and received a 10-mg dose. Thirty minutes after ingestion,Ms. A was found agitated and confused, rambling about want-ing to go to the beach. Her speech was disorganized, and shehad visual hallucinations. Her gait was described as “ataxic,”and no signs of meningeal irritation were noted. Her tempera-ture was 99.2°F (37.3°C), her pulse 114 b.p.m., and her bloodpressure 116/78 mm Hg. When she was evaluated the nextmorning, her delirium had cleared. Only some lethargy re-mained for most of the day. Her white blood cell count was 6.2,her creatinine kinase level was 54 U/L (range, 41–117 U/L), andher mild muscular rigidity remained unchanged.

This case serves as a reminder that delirium is often due tomultiple etiologies. Although the potential for benzodiazepinehypnotics such as triazolam to cause delirium is well recog-nized, even fairly benign agents such as zolpidem can contrib-ute to the development of delirium in susceptible individuals.Fever alone can cause delirium, and a case of delirium resultingfrom risperidone use was described in an elderly woman.5 Ourpatient, however, had had a low-grade fever for several days,and she had been treated with risperidone for 2 weeks. More-over, the patient became delirious 30 minutes after zolpidemwas given, consistent with its pharmacokinetic properties. Thistime course speaks against an acute drug-drug interaction in-volving the cytochrome P450 system. However, the use offluoxetine could, through some pharmacodynamic interaction,have led to a predisposing brain state that created “fertile soil”for a later reaction to zolpidem. Cases with prolonged zolpidem-associated hallucinations in patients taking selective serotoninreuptake inhibitors have been described,6 even though no sig-nificant pharmacodynamic or kinetic interaction was found in aformal study.7 We considered neuroleptic malignant syndrome,since this patient could have been at higher risk for a neurolep-tic malignant syndrome–like or serotonin syndrome–like reac-tion prior to receiving zolpidem.

It is notable that this case involved a woman, as have themajority of reported cases of zolpidem-related mental statuschanges. As Markowitz and Brewerton2 point out in their report,young women achieve an almost 50% higher plasma zolpidemlevel compared with men with a given dose.

In these days of short inpatient stays and aggressive poly-pharmacy, many patients are treated with more than one agent.Medications that are safe in healthy individuals and routine set-tings should be used cautiously in ill patients receiving multiplemedications. The nonbenzodiazepine hypnotic zolpidem is noexception to this rule, and the lowest effective dose should beused. Special note should be taken with women, who might re-quire a lower dose of zolpidem than men.

REFERENCES

1. Pies RW. Dose-related sensory distortions with zolpidem [letter].J Clin Psychiatry 1995;56:35–36

2. Markowitz JS, Brewerton TD. Zolpidem-induced psychosis.Ann Clin Psychiatry 1996;8:89–91

3. Katz SE. Possible paroxetine-zolpidem interaction [letter].Am J Psychiatry 1995;152:1689

4. Hoyler CL, Tekell JL, Silva JA. Zolpidem-induced agitation anddisorganization. Gen Hosp Psychiatry 1996;18:452–453

5. Tavcar R, Dernovsek MZ. Risperidone-induced delirium [letter].Can J Psychiatry 1998;43:194

6. Elko CJ, Burgess JL, Robertson WO. Zolpidem-associated hallucina-tions and serotonin reuptake inhibition: a possible interaction.J Toxicol Clin Toxicol 1998;36:195–203

7. Piergies AA, Sweet J, Johnson M, et al. The effect of co-administration of zolpidem with fluoxetine: pharmocokinetics andpharmacodynamics. Int J Clin Pharmacol Ther 1996;34:178–183

Oliver Freudenreich, M.D.Massachusetts General Hospital

Boston, MassachusettsMatthew Menza, M.D.

Robert Wood Johnson Medical SchoolPiscataway, New Jersey

Bupropion-Tranylcypromine Combination forTreatment-Refractory Depression

Sir: We report the safe and successful use of combined bu-propion and tranylcypromine in a case of treatment-refractorydepression.

Case report. Ms. A, a 27-year-old woman with chronic ma-jor depression (DSM-III-R criteria), was treated in 1990 at age17 with imipramine up to 150 mg/day. In March 1994, after asuicide attempt by imipramine overdose, paroxetine, 20 mg/day,replaced imipramine. On paroxetine, 30 mg/day, and amitripty-line, 50 mg q.h.s., she experienced only mild improvement at6 months. By 1995, Ms. A was taking paroxetine, 30 mg/day;nortriptyline, 35 mg/day; and temazepam, 15 mg q.h.s. (with aplasma nortriptyline level of 71 ng/mL). Despite the addition ofbrief trials of methylphenidate, 5 mg b.i.d., her mood remaineddepressed.

In August 1995, bupropion was initiated with the gradualwithdrawal of nortriptyline and paroxetine. Trazodone, 25 mgq.h.s., replaced temazepam. By October 1995, Ms. A was takingbupropion, 150 mg b.i.d.; trazodone, 100 mg q.h.s.; and loraze-pam, 0.5 mg t.i.d., with only a partial response. A trial of adjunc-tive liothyronine (T3), 25–75 µg/day, failed to yield sustainedimprovement, and her dysphoria, fatigue, and insomnia insteadworsened. Tranylcypromine was added and upon titration to 50mg/day, Ms. A reported a gradual return to euthymia with reso-lution of her long-standing depressive symptoms.

Bupropion was tapered and discontinued in March 1996, butwithin 2 weeks, Ms. A noted a return of depressive symptoms.Bupropion was therefore restarted with good results. On oneoccasion, Ms. A developed symptomatic hypertension after eat-ing cheese, but it was managed with nifedipine at home. Otherthan this episode, she was normotensive at all checkups. By thesummer of 1996, T3, trazodone, and lorazepam were withdrawn.She remained on a regimen of tranylcypromine, 60 mg/day (40mg in the morning and 20 mg at noon), and bupropion sustainedrelease (SR), 150 mg b.i.d., with sustained euthymia. Transientand mild periods of stress-related dysphoria or insomnia weremanageable with low-dose lorazepam. In November 1997, tran-ylcypromine was withdrawn in order for Ms. A to undergo asurgical procedure. During the 2 weeks off tranylcyprominetreatment, she experienced an acute worsening of mood symp-toms that quickly resolved with its reintroduction. Two yearslater, she remains on tranylcypromine, 60 mg/day, and bupro-pion SR, 150 mg b.i.d., without relapse of depression. She con-tinues to keep nifedipine in the event of a hypertensive crisis,but aside from her 1 episode associated with ingestion of cheese3 years ago, she has had no further blood pressure elevationsover the course of her treatment.

450

https://www.researchgate.net/publication/14397922_Zolpidem-Induced_Psychosis?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15363564_Dose-related_sensory_distortions_with_zolpidem_2?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15363564_Dose-related_sensory_distortions_with_zolpidem_2?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/51326354_Risperidone_induced_delirium?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/51326354_Risperidone_induced_delirium?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15712740_Possible_paroxetine-zolpidem_interaction?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1





Ms. A’s case is remarkable for chronic depression unrespon-sive to treatment with a selective serotonin reuptake inhibitorand tricyclic antidepressants, as well as adjunctive methyl-phenidate and T3. The addition of tranylcypromine to bupropionfinally resulted in a sustained remission of her depression. De-pressive symptoms returned whenever either antidepressantwas withdrawn over the course of treatment, thereby emphasiz-ing the apparent necessity of both antidepressants for treatmentresponse. Ms. A’s poor response to multiple antidepressant tri-als, including both a tricyclic and tranylcypromine alone, classi-fies her depression at Stage IV resistance as defined by Thaseand Rush.1 The treatment of choice for Stage IV resistant de-pression is electroconvulsive therapy (ECT). 1 ECT was notadministered to Ms. A, owing to her desire to continue employ-ment and outpatient management of her depression.

While no controlled double-blind studies support the use ofcombination antidepressant therapy in treatment-resistant de-pression,2 this practice is supported by anecdotal evidence andgeneral clinical opinion. Ms. A’s case adds to this body of evi-dence and suggests that combination antidepressant therapy de-serves greater study in refractory depression. Although thecombination of bupropion and a monoamine oxidase inhibitor isnot usually recommended due to the risk of hypertensive crisis,3

the cautious administration of tranylcypromine and bupropiontogether may be a safe and effective strategy in some cases oftreatment-resistant depression.

REFERENCES

1. Thase ME, Rush AJ. When at first you don’t succeed: sequentialstrategies for antidepressant nonresponders. J Clin Psychiatry1997;58(suppl 13):23–29

2. Nelson JC. Overcoming treatment resistance in depression. J ClinPsychiatry 1998;59(suppl 16):13–19

3. Kaplan HI, Sadock BJ, eds. Comprehensive Textbook of Psychiatry.6th ed. Baltimore, Md: Lippincott, Williams & Wilkins; 1995

Joseph M. Pierre, M.D.Michael J. Gitlin, M.D.

UCLA Neuropsychiatric InstituteLos Angeles, California

Inappropriate Masturbationand Schizophrenia

Sir: Inappropriate masturbation and sexual preoccupation inschizophrenia can have disastrous consequences for a patient.We report a case of treatment-resistant schizophrenia in whichinappropriate masturbation presented legal and housing prob-lems for the patient and did not resolve with treatment of hispsychosis. Behavioral redirection was of no avail, so we resortedto oral medroxyprogesterone acetate (MPA) in an attempt to de-crease the incidence of masturbation.

Case report. Mr. A is a 42-year-old white man diagnosedwith paranoid schizophrenia and alcohol/cocaine abuse (in re-mission). He was transferred to an inpatient ward from jail,where he was sent after a neighbor complained about his urinat-ing off the porch of his board and care home. The police discov-ered that Mr. A was a sex offender and had not registered hiscurrent address. He had a history of inappropriate masturbationand sexual advances toward females, which led to his convic-tion after he exposed his genitalia to a minor.

During his recent jail stay, clozapine, which he had been tak-ing for 1 year, was discontinued and he became grossly psy-chotic to the point of near-mutism and lying in his own feces.On admission to the hospital, his clozapine was resumed, and,even though his psychotic symptoms lessened, he continuedto masturbate openly in the ward several times per week. Al-though he was receiving clozapine, 750 mg/day; divalproexsodium, 2500 mg/day (blood level = 108 µg/mL); and trazo-done, 100 mg/day, his behavior did not change, and he often re-ported “seeing” nude women dance around his bed in themorning, which he acknowledged were not “real.”

It was at this time that we considered beginning MPA. Pre-liminary laboratory values revealed that Mr. A’s plasma testos-terone (418 ng/dL), luteinizing hormone (6.4 mIU/mL), andfollicle-stimulating hormone (4.3 mIU/mL) levels were allwithin normal limits. After giving consent, Mr. A began oralMPA, 30 mg/day. Over the course of the following week, his“visions” of frolicking nude women dissipated and his mastur-bation decreased to the point of nonexistence. Additionally, heno longer wore his masturbation “attire,” which comprisedsweat pants, a hat, and dark sunglasses. His mood remained eu-thymic throughout this period, and his testosterone level after 1week of MPA treatment was 285 ng/dL. He was accepted into acommunity housing project.

Antiandrogenic agents such as MPA are often used toreduce sexual drive and improve sexual conduct in individualssuch as Mr. A. MPA inhibits luteinizing hormone and follicle-stimulating hormone, preventing the release of testosteronefrom the testes and thus decreasing libido and sexual arousal.1

Yet MPA, unlike some antiandrogenics, has not been reported tocause feminization.2 Additionally, MPA administration can de-crease the frequency of erotic fantasies without altering plasmatestosterone, luteinizing hormone, or follicle-stimulating hor-mone levels, suggesting the possibility of a direct central ner-vous system effect.1 This potential dissociation between MPA’sefficacy and testosterone is particularly important in a psychiat-ric setting because low testosterone levels can lead to depres-sion.3

After starting MPA, Mr. A reported a decline in the frequencyof his psychotic sexual fantasies, and although his testosteronelevel decreased, it remained within the normal range. Patientssuccessfully treated with MPA for troubling fantasies have beenshown to experience no adverse effect on penile response. 1 Thepreservation of sexual function allows for continued appropriatesexual activity while decreasing problematic sexual behavior—a combination that makes the use of MPA in such a setting moreconsistent with a medical treatment than a punitive measure.

REFERENCES

1. Richer M, Crismon ML. Pharmacotherapy of sexual offenders. AnnPharmacother 1993;27:316–319

2. Berlin FS, Meinecke CF. Treatment of sex offenders withantiandrogenic medication: conceptualization, review of treatmentmodalities, and preliminary findings. Am J Psychiatry 1981;138:601–607

3. Seidman SN, Walsh BT. Testosterone and depression in aging men.Am J Geriatr Psychiatry 1999;7:18–33

John O. Brooks III, Ph.D., M.D.Dept. of Veterans Affairs Health Care System-Palo Alto

Palo Alto, CaliforniaManoj V. Waikar, M.D.

Stanford University School of MedicineStanford, California

451



https://www.researchgate.net/publication/14746482_Pharmacotherapy_of_sexual_offenders?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/14746482_Pharmacotherapy_of_sexual_offenders?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/16741841_Treatment_of_sex_offenders_with_antiandrogenic_medication_Conceptualization_review_of_treatment_modalities_and_preliminary_findingsAmerican_Journal_of_Psychiatry_138_601-607?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1




https://www.researchgate.net/publication/13489332_Overcoming_treatment_resistance_in_depression?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13489332_Overcoming_treatment_resistance_in_depression?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/13827112_Thase_ME_Rush_AJ_When_at_first_you_don't_succeed_sequential_strategies_for_antidepressant_nonresponders_J_Clin_Psychiat_58Suppl_13_23-29?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1




Book Reviews



Book ReviewsMichael H. Ebert, M.D., Editor

Countertransference Issues in Psychiatric Treatment(Review of Psychiatry Series, Vol. 18)edited by Glen O. Gabbard, M.D. Washington, D.C.,American Psychiatric Press, 1999, 127 pages,$26.50 (paperback).

In an era during which medication management of patientsbased on a descriptive diagnosis derived from DSM-IV has insome circles become the clinical foundation of psychiatry,Gabbard’s brief book on issues in countertransference is awelcome and essential addition to contemporary clinical psy-chiatry.

Freud first introduced the concept of countertransferencein 1910, noting, “We have become aware of the ‘counter-transference,’ which arises in him [the physician] as a result ofthe patient’s influence on his unconscious feelings. . . . ”1 Issuesof the psychiatrist’s countertranferences to patients can and dooccur whether the treatment focuses on medication management,psychotherapy, or a combination of both treatment modalities.We today share Freud’s view “ . . . that he [the therapist] shall rec-ognize this countertransference in himself and overcome it.”

Gabbard and his contributors first offer an expanded theo-retical basis for understanding countertransference. After ac-complishing this task, they develop models as to how practi-tioners can observe their countertransference reactions and thenuse their self-knowledge to minimize such countertransferencesfrom interfering with treating their patients.

It should be noted that in their contributions, Glen Gabbard,John Maltsberger, Francis Varghese, and Brian Kelly all base asubstantial element of their theoretical understanding of thecomplex unconscious elements that may create countertrans-ferences on the original contributions of Melanie Klein. Not allcontributors to this difficult subject matter would explain coun-tertransference on a Kleinian model.

Independent of the theoretical views presented, the authorspresent a clearly written view of the varied clinical problemspsychiatrists face in varied settings that can create counter-transferences. Marcia Goin, in her chapter “Countertransfer-ence Is General Psychiatry,” gives the reader an informative yeteasy-to-read guide to using awareness of one’s countertrans-ferences to assist treatment. All too frequently, psychiatrists be-come defensive when discussing countertransferences, forget-ting its inevitability in our work. Understanding our counter-transferences informs us about ourselves and our patients.

The final chapter of this volume focuses not on the generalissues we face in psychiatric practice but on a major socialquestion—physician-assisted suicide. The authors argue thatany physicians assisting in the suicides of their patients arelikely to be dealing with countertransference issues or en-actment with their patients and not functioning in the bestinterest of their patients. Although the authors provide a litera-ture review and theoretical support for their contentions that

physician-assisted suicide is a grossly inappropriate act for aphysician, an issue of this magnitude needs a much broader re-view. The authors are, however, correct in articulating the viewthat one cannot discuss physician-assisted suicide without com-prehending the role of countertransference in the physician’sbehavior.

In sum, this volume is useful in keeping us focused on thedoctor-patient relationship wherever the site of treatment orwhatever the modalities utilized.

REFERENCE

1. Freud S. The future prospects of psycho-analytic therapy (1910).In: Strachey J, ed. The Standard Edition of the Complete Works ofSigmund Freud, vol. 11. London, England: Hogarth Press; 1957:144–145

Sidney H. Weissman, M.D.Loyola University

Chicago, Illinois

The Chemical Dependence TreatmentDocumentation Sourcebookby James R. Finley, M.A., and Brenda S. Lenz, M.S. New York,N.Y., John Wiley & Sons, Inc., 1999, 278 pages, $49.95.

This book is a “how-to” manual for developing and structur-ing a substance abuse treatment program (SATP). It is writtenby 2 master’s degree–level addiction therapists (who also doeducation and SATP program management) and includes a com-prehensive collection of clinical, educational, and program-management material.

The text is divided into 3 basic sections. The first section(chapters 1 and 2) addresses the basic issues of evaluating theneed for a SATP, as well as local resources, goals, costs, and re-ferral sources. It also explains the initial steps in beginning aSATP. The second section (chapters 3–7) provides numerous ad-ministrative aids for establishing a SATP such as referenceforms, assessment procedures, consent and release forms, andvarious other administrative forms (attendance sheets, behav-ioral contracts, and so on). It is particularly helpful that the formshave also been provided on a computer disk (which comes withthe book) so that readers can create their own copies. The thirdsection (chapters 8–11) is primarily concerned with enhancingSATP programs. Chapter 8 includes a few ideas for outcomemeasurements, but this potentially very important chapter is un-fortunately quite limited in content. Other chapters in this sec-tion contain information for conducting groups and teaching onissues such as substances of abuse, stress, anger, codependency,comorbidity, loss, domestic violence, and family depression.

https://www.researchgate.net/publication/290796543_The_future_prospects_of_psychoanalytic_therapy_1910?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1


Book Reviews



Overall, this book will be extremely useful to addictiontherapists and hospital administrators who are trying to createand/or enhance a SATP. However, some information that mighthave been useful to addiction psychiatrists, such as detoxifica-tion protocols, pharmacologic treatments for relapse preven-tion, clinical outcome measurements, and comorbidity issues, isweak or not included.

John R. Hubbard, M.D., Ph.D.Vanderbilt University School of Medicine

Nashville, Tennessee

The Recognition and Management of Early Psychosis:A Preventive Approachedited by Patrick D. McGorry and Henry J. Jackson.New York, N.Y., Cambridge University Press, 1999,495 pages, $90.00.

Growing evidence suggests that early recognition and treat-ment of the schizophrenic psychoses may be of critical impor-tance in the reduction of subsequent morbidity. The present vol-ume, edited by McGorry and Jackson, presents a comprehen-sive review of the psychosocial literature relevant to detectionand treatment of the schizophrenias, together with a descriptionof a rehabilitation program designed to reduce subsequentschizophrenic morbidity.

The centerpiece of the argument is that the schizophrenia pro-cess is most active, and schizophrenia is most amenable to suc-cessful intervention, within a relatively short period that extendsfrom the beginning of the prodromal period to 5 years after theemergence of frank psychotic symptoms. The contributing au-thors bewail the common failure of early diagnosis, with psy-chotic symptoms often present years before the individual comesto the attention of mental health professionals. They also notethe often extended prodromal period, with evidence of deterio-rating psychosocial interests and skills. These are the optimumtimes for intervention to arrest or partially arrest the process thatsubsequently tends to become relatively fixed, with permanentdamage to both the biological and psychosocial substrates.

The Melbourne program (Early Psychosis Prevention andIntervention Center [EPPIC] and its Early Psychosis Assess-ment and Community Treatment Team [EPACT]) attempts tooptimize initial management of the patient’s first psychotic epi-sode, both maximizing chances for full recovery and minimiz-ing the potential for future relapses and morbidity. The protocolfirst attempts to provide the patient with immediate access tomental health evaluation (rather than weeks or months of shuf-fling from one professional to another), prompt use of low-doseneuroleptics (2–4 mg/day equivalents of haloperidol) preferablyin a community, home-based setting, together with cognitiveand family interventions aimed at rapid and sustained recoveryfrom the episode. Recovery is further enhanced by the use of theEPPIC day programs, with social, recreational, vocational, cre-ative expression, health promotion, and personal skills develop-ment streams.

As readily acknowledged by the authors, a paucity of con-trolled studies documents the efficacy of such interventions.The extensive literature reviews and the EPPIC program itselfare largely descriptive. Intuitively, this is an empathic programthat should provide almost every available chance for recoveryfrom first-episode psychosis. Yet the skeptics, including stateand local legislatures, might not be convinced in the absence ofhard, controlled efficacy data.

It is striking that the concept of identifying those at risk andintervening prior to the emergence of frank psychosis is paidlimited attention here. That much of the biological and psychoso-cial damage in schizophrenia is turned on and results in perma-nent destructive effects even during the prodromal period is in-creasingly accepted. The advent of new atypical antipsychoticswith low drug-induced morbidity now permits experimentalpharmacologic treatment interventions in high-risk populationsduring early-to-mid prodromal periods, although the results ofsuch pharmacologic trials may not be known for some time.Similarly, serious strategies need to be attempted in the sphere ofpsychosocial intervention early in the prodromal periods withat-risk populations. It is likely that, within the next few years,genomic scans will aid the identification of such at-risk individu-als, leading to an opportunity for truly preventive rather than re-storative treatments.

David L. Garver, M.D.University of Louisville School of Medicine

Louisville, Kentucky

Personality-Guided Therapyby Theodore Millon. New York, N.Y., John Wiley & Sons, 1999,776 pages, $65.00.

Theodore Millon’s Personality-Guided Therapy will prob-ably be viewed as somewhat of an opus of this singular and pro-digious contributor to our understanding of personality and sys-tematic attempts to influence human behavior. With 713 pages oftext, expanding to 776 pages including references and index, anda $65.00 quoted cost, the book is hefty in size but moderatelypriced by contemporary standards.

The book begins with a section on “Foundations” that gives abrief overview of treatment modalities, including contemporarytrends and 2 chapters focusing on Millon’s personal contribu-tions to integrating psychotherapeutic strategies in a systematicmanner to plan treatment for humans as opposed to diseasestates. The review is strong in its treatment of individual thera-pies but expectably weaker in handling of group psychotherapy.(Specifically, there was no mention of David Spiegel’s work withexpressive supportive group psychotherapy or the group analyticcontributions of Bion.) I also wish that he would have spent moretime on the admittedly less developed but extraordinarily prom-ising contributions of Kandel and Cloninger. The “Foundations”section does best what the author can do best. The chapters onpersonality-guided synergism and planning personality-guidedtreatment provide a succinct and highly useful summary of theauthor’s singular contributions to organized efforts to help healpeculiarly human suffering.

For the reader already immersed in Millon’s work, thisbook will largely be a helpful review. For the reader who is anestablished clinician but unfamiliar with psychological aspects ofpsychopathology, the book provides a systematic, thorough ori-entation to move beyond conceptualizing psychopathology as onlya neurobiological event. It will help such readers to gain a frame-work upon which to organize comprehensive treatment. Millon’swork is systematic enough that even the most committed biologi-cal reductionist will have to (grudgingly perhaps) acknowledgethat there is a systematic approach toward understanding and re-sponding to the psychological aspects of psychopathology.

The reader who is in the best position to benefit from this bookis an experienced psychotherapist familiar with 2 or more differ-ent conceptual frameworks for structuring psychotherapeutic re-


Book Reviews



lationships. This reader will easily grasp the ways in whichMillon’s efforts to integrate conceptual models and connect themto a developing treatment relationship can enhance therapeuticresults. A reader with less sophistication will need a supervisoryrelationship to achieve maximum benefit from this book. For ex-ample, in psychiatric and psychology training situations, a semi-nar leader with previous experience utilizing Millon’s concep-tual framework will be helpful to produce a learning environ-ment in which students, residents, or fellows can benefit from aclear but complicated scheme. As an alternative, a learninggroup with 2 or more respectful and competent faculty with dif-ferent backgrounds would need to collaborate with each otherand the learners to make the best use of this book. One approachwould be for several sessions of a case conference in a mooddisorder clinic to review the “Foundations” section and thenpresent clinical material from the clinic while reading the chap-ter on mood-related syndromes. After a session or 2 of that fo-cus, the faculty might identify where pharmacologic treatmentis providing only partial response because of the confounding ef-fects of a personality disorder. The group could then use a caseconference to discuss a clinical presentation and the relevantchapter: for example, a schizoid patient whose mood improve-ment has not resulted in constructive social contacts.

In summary, this is a sophisticated and comprehensive bookthat urges us to recognize the limits of descriptive diagnosis-based treatment while not discarding the advances made by theDSM approach. Millon challenges us to provide treatments that“derive logically from the theory of that particular person.” Thisbook is a helpful tool in that worthy effort.

James W. Lomax, II, M.D.Baylor College of Medicine

Houston, Texas

The Selfish Brain: Learning From Addictionby Robert L. DuPont, M.D. Washington, D.C.,American Psychiatric Press, 1997, 553 pages, $26.95.

This is a rather unusual book for the reader accustomed tomost scientific books. It begins with 5 pages of endorsementsand a foreword by Betty Ford. Then, after a brief, well-writtenpreface that outlines the structure of the book, the author spendsanother 7 pages on his biography. Dr. DuPont states in the pref-ace that this book is for therapists, counselors, and educators,which might explain the lack of references to support its manyassertions, except for footnotes dealing with some of the statis-tical data.

The book comprises 3 sections. In the first section, “Think-ing About Addiction,” the author briefly touches on the geneticsand character traits that predispose individuals to addiction. Hestresses the disease concept of addiction, although later heseems to back off the traditional disease concept by extendingthe “disease” to family and even society. The chapters on his-tory of drug abuse and the current “scene” of drug use are wellwritten and provide a wealth of information for anyone con-cerned with addiction. However, the exact sources of some ofthe statistical data are difficult to ascertain.

In the section “The Brain and Addiction,” the discussion ofthe neuropharmacology and neurophysiology of the brain is

rather superficial, but probably sufficient for the target audi-ence. Chapters 5 and 6 do a reasonably good job in describingspecific drugs, from nicotine and alcohol to the hallucinogens,and appropriately stress the highly addictive nature of cocaineand its central nervous system and peripheral deleterious ef-fects. The section on marijuana lacks data to support many ofthe assertions. It reminds one of the “fried egg” commercials afew years ago designed to turn young people against drugs,which were totally rejected and ridiculed by the target audience.There are also some inaccurate statements in the section. Forexample, state-dependent learning is defined as “brain learningto function” with presence of alcohol, rather than the specificprocess in which an animal under the influence of a drug canlearn but not recall the material when sober.

The last section, “Overcoming Addiction,” consists of 7chapters. The chapter on “The Addict’s Career” clearly outlinesthe stages of addiction and has a good discussion of type I andtype II addicts as well as an excellent discussion on “hitting bot-tom.” However, the author’s use of personality and personalitydisorders is problematic. Temperament traits like high noveltyseeking and low harm avoidance are labeled character disor-ders, and antisocial personality, which has very clear criteria inthe DSM-IV, is rather loosely used in the discussion.

The chapter on “Prevention of Addiction” has a useful guideto “drug-proofing your children” and for identifying high-riskyouth. However, the terms temperament and character disorderare used interchangeably. The beneficial, well-documented ef-fects of modest (up to 2 drinks per day) drinking on reducingcoronary disease and stroke are rather quickly dismissed in fa-vor of a focus on excessive drinking.

The chapters on intervention and on 12-step programs arewell written and should provide useful data to therapists andcounselors. Particularly significant is the author’s stress of theimportance of after care in any treatment process.

In the chapter on drug policy, the author admits up front thatthe ideas expressed are controversial. He extensively discussesthe advantages and disadvantages of drug legalization, with mostemphasis on the disadvantages. Harm reductions, includingclean-needle programs, are dismissed by the author. Dr. DuPont’sview on drug policy is probably best expressed in his concept ofthe “drug abuse chain,” where drug production/cultivation andthe drug user, i.e., the addict, are the main targets of sanctions.He recommends strict sanctions for users.

In the chapter on future of addiction, he admits that “cropsubstitution” does not work and that only eradication and de-struction will work. He proposes that crop substitution on thesupply side of the chain is analogous to education and treatmenton the demand side and that without sanctions, treatment andeducation are ineffective.

This book may be useful to the stated target audience oftherapists, counselors, and educators, but I do not find much torecommend it to psychiatrists or neuroscientists interested inaddiction. It is unfortunate that Dr. DuPont sees no other op-tions except for the U.S. government’s quasi-military approachto the problems of drug abuse and addiction. If addiction is trulya disease, then treatment, not punishment, should be the ulti-mate goal to which most of our resources are dedicated.

John J. Straumanis, M.D.Tulane University School of Medicine

New Orleans, Louisiana


Sponsored by Physicians Postgraduate Press

CME: CATEGORY 1

CME ACTIVITY© Copyright 2000 Physicians Postgraduate Press, Inc.


This activity has been planned and implemented in accordance with the Essential Areas and Policiesof the Accreditation Council for Continuing Medical Education (ACCME). To obtain credit, pleaseread the following article and complete the posttest as instructed on page 460.

CME ObjectivesAfter completing this CME activity, physicians practicing clinical psychiatry should be able to:• Describe the potential cognitive effects of benzodiazepines in patients admitted to the hospital

after a suicide attempt with benzodiazepines.• Develop a strategy for an initial psychiatric interview in a patient who has attempted suicide

with benzodiazepines.

Statement of Need and PurposeBenzodiazepines, which are still used extensively, are now recognized as major drugs of abuse andaddiction. Physicians responding to questionnaires in the Journal and its related CME activities haverequested updated information on the risks and benefits of benzodiazepines. This CME activity wascreated to address that need. There are no prerequisites for participation in this CME activity.

Accreditation StatementPhysicians Postgraduate Press is accredited by the ACCME to sponsor continuing medical educationfor physicians and takes responsibility for the content, quality, and scientific integrity of this activity.

Credit DesignationPhysicians Postgraduate Press designates this educational activity for a maximum of 1 hour ofCategory 1 credit toward the American Medical Association Physician’s Recognition Award. Eachphysician should claim only those hours of credit that he/she actually spent in the educational activity.

Date of Original Release/ReviewThis article was published in June 2000 and is eligible for CME credit through November 30, 2000.The latest review of this material was May 2000.

Faculty DisclosureIn the spirit of full disclosure and in compliance with all ACCME Essential Areas and Policies, allfaculty for this CME activity were asked to complete a full disclosure statement. The informationreceived is as follows:

None of the authors of this article has significant commercial relationships to disclose relative tothe presentation.

Discussion of Investigational InformationFaculty may be presenting investigational information about pharmaceutical agents that is outsideU.S. Food and Drug Administration–approved labeling. This information is intended solelyas continuing medical education and is not intended to promote off-label use of any of thesemedications. Please consult the paragraph titles “Disclosure of Off-Label Usage” at the endof the article for a complete listing of off-label or investigational uses.

455




CME: ARTICLE

t has been demonstrated in many studies that benzo-diazepines can induce memory impairment.1,2 A com-

mon finding is anterograde amnesia, the phenomenonwhereby information presented after benzodiazepineshave been taken is poorly remembered.3,4 Severity and du-ration of the amnesia are dependent on the particular ben-zodiazepine used, the dosage, the route of administration,the time (postdrug) at which information is presented andretrieval is required, and characteristics of the subjectpopulation tested.5–7 The memory task used is also of im-portance. For instance, recall tasks are more sensitive thanrecognition tasks, and verbal tasks are more sensitive thanvisual tasks.2 In longitudinal studies, it was found thatonly partial tolerance to the amnesic effects develops; inverbal recall tests, tolerance to the effects of benzodiaze-pines was found on immediate but not delayed (15–20minutes) recall in normal volunteers8 as well as in patientswith generalized anxiety disorder.9 Impaired delayed re-call was also found in chronic benzodiazepine users afteracute administration of the medication.10

Anesthetists welcome the amnesic effects since theycause patients to forget unpleasant operative procedures.However, in other situations, these effects are a reason forconcern. In the large group of people who are taking ben-zodiazepines on a daily basis, anterograde amnesia mayinterfere with the ability to function optimally, and in pa-tients taking benzodiazepines who are also being treatedwith psychotherapy, anterograde amnesia may impair re-membrance of what happened during a therapy session.11

In a case report,12 it has been suggested that anterogradeamnesia also interferes with the efficacy of psychiatricconsultation in patients admitted to the hospital after a sui-cide attempt with benzodiazepines. Since benzodiazepinesare involved in almost half of the suicide attempts,13 wedecided to further explore possible memory impairmentfollowing a benzodiazepine overdose. We examined sys-tematically whether anterograde amnesia occurs in a groupof patients consecutively admitted to the hospital becauseof a suicide attempt involving benzodiazepines. We inves-

Memory Impairment in Those WhoAttempted Suicide by Benzodiazepine Overdose

Bas Verwey, M.D.; Paul Eling, Ph.D.;Henk Wientjes, M.Sc.; and Frans G. Zitman, M.D., Ph.D.

IBackground: A prospective study was con-

ducted to investigate the presence of antero-grade amnesia in those who attempted suicideby benzodiazepine overdose and to study thecorrelation with sedation.

Method: In 43 patients who attempted sui-cide by taking benzodiazepines, memory wastested with a 15-word memory recall task. Theimmediate and delayed recall on the first dayafter admission (day 1) and 24 hours later (day2) were rated. Each patient and the interviewerscored the patient’s degree of sedation on avisual analogue scale. Patients also had totry to recognize, from photographs, the psy-chiatrist with whom they had spoken theday before.

Results: The ratings of immediate and de-layed recall were significantly lower on day 1than on day 2. Subjective ratings of sedationof the patients were not significantly higherthan the ratings of the observer. Less than halfof the patients recognized the psychiatrists andknew that they were the ones they had spokento the day before.

Conclusion: Anterograde amnesia ispresent in suicide attempters who take over-doses of benzodiazepines. The implicationsof this finding for the assessment of suicideattempters during admission are discussed.


Received March 1, 1999; accepted Jan 16, 2000. From ZiekenhuisRijnstate, Arnhem (Dr. Verwey); and the Departments of Psychology (Dr.Eling) and Psychiatry (Dr. Wientjes and Professor Zitman), University ofNijmegen, Nijmegen, the Netherlands.

Reprint requests to: Bas Verwey, M.D., Ziekenhuis Rijnstate, Postbus9555, 6800 TA Arnhem, the Netherlands.

456


CME: ARTICLE



tigated verbal recall with the 15-word test, because sucha test has been applied frequently in the evaluation ofbenzodiazepine-induced anterograde amnesia. This test isthought to be highly sensitive to the presence of antero-grade amnesia induced by benzodiazepines.14 Scores on ahighly standardized verbal recall test do not necessarilyreflect clinically significant memory impairment. There-fore, we added a visual recognition task with a more im-mediately obvious relevance to clinical practice: a photorecognition test to investigate whether patients were ableto recognize the face of the psychiatric resident who ex-amined them. In clinical practice, it is often assumed thatthe degree of anterograde amnesia can be predicted fromthe degree of sedation as estimated by the psychiatrist.However, the results of laboratory studies are equivocal.5

Therefore, in this study we also assessed the degree of se-dation as estimated by the patient and the psychiatric resi-dent who examined the patient and related these estima-tions to the degree of memory impairment.

METHOD

Material and ProcedureFrom September 1, 1994, until November 25, 1996,

all patients admitted to the Rijnstate general hospital(Arnhem, the Netherlands) because of attempted suicideby taking an overdose of benzodiazepines were prospec-tively studied. The medical ethics committee of the hospi-tal approved the study.

SubjectsAll patients with benzodiazepines in their blood were

included. Additional use of other drugs and nondrug meth-ods in the suicide attempt were not exclusion criteria, butall patients that had alcohol in their blood were excluded.

Also, patients meeting the following criteria were ex-cluded: age younger than 18 years (because the tests usedwere not developed for younger people), a diagnosis ofdementia or amnestic disorder before admission (cogni-tive dysfunction may interfere with the drug effects), al-cohol dependence or abuse (cognitive dysfunction andimpaired liver function may interfere with the drug ef-fects), delirium according to DSM-IV criteria, and/or in-ability to read or understand test instructions and items onthe verbal recall test.

Assessments1. The “15-woorden” test (15-word test) is the Dutch

equivalent of the Rey Auditory Verbal Learningtest, a verbal recall test often used in research on

anterograde amnesia caused by benzodiazepines.15

We adapted the procedure slightly by administer-ing the list of words only once on day 1 and onceon day 2. On both days, the research assistant in-structed the patient in the use of a computerizedversion of the test. The patient was asked to re-member words that were presented successivelyon the computer screen for a period of 2 secondseach. Immediately after the presentation of 15 un-related words, the patient was asked to recall asmany words as possible (immediate recall). After15 minutes the patient was asked again to recallas many words as possible (delayed recall). Thetest was administered on days 1 and 2. To preventlearning, patients were presented different butequivalent lists on these days.

2. A photograph recognition test was used as a visualrecognition test. The patient was presented with aseries of 6 photographs of faces with a variety inthe features such as hair, spectacles, and so on, 1of which showed the resident in psychiatry whoexamined the patient. This procedure is used informal police investigations in the Netherlands.The patient had to answer 3 questions: (1) Do yourecognize anyone? (2) Who do you recognize?(3) How do you know this person? This test wasadministered on day 2 only. Patients who wereknown to the resident, possibly because of earliersuicide attempts, were excluded from this test.

3. Degree of sedation was rated on a 10-cm visualanalogue scale (VAS). On day 1, the patient as wellas the resident in psychiatry filled in this scale. Alow score indicates a high level of sedation.

4. Blood alcohol concentration was measured withgas liquid chromatography.

5. Presence of benzodiazepines in blood was ascer-tained by the immunochemical method TDxFLx(Abbott Laboratories, Abbott Park, Ill.)

ProcedureAt Rijnstate hospital, all patients admitted because of a

suicide attempt are seen by a psychiatric resident for aroutine clinical interview at least 12 hours after admit-tance, provided the patient is sufficiently alert for psychi-atric consultation. For this study, at the end of the inter-view, the psychiatric resident informed patients who hadused benzodiazepines in the suicide attempt about thestudy and asked them to participate. When informed con-sent was obtained, a research assistant (an experiencedconsultation-liaison nurse) immediately started assess-

457


CME: ARTICLE



ments (day 1). Twenty-four hours later, the research assis-tant again performed assessments (as day 2).

Statistical AnalysisData were analyzed using the paired t test; the level of

significance was p ≤ .05. Relationships between memoryvariables and subjective sedation ratings were assessedusing Pearson correlation coefficients.

RESULTS

Forty-seven patients were eligible for the study. Threerefused to cooperate, and 1 had to be excluded becauseof insufficient cooperation. In this article, data on the re-maining 43 patients (9 men and 34 women; mean ± SDage = 39.1 ± 12.4 years) are presented. The following ben-zodiazepines had been used: oxazepam (N = 16), temaze-pam (N = 11), diazepam (N = 10), clorazepate (N = 4),alprazolam (N = 6), flurazepam (N = 4), lormetazepam(N = 3), lorazepam (N = 1), clobazam (N = 1), flunitraze-pam (N = 2), and zopiclone (N = 1). Eleven patients hadused more than one benzodiazepine. Six patients had alsoused an antidepressant agent. Three others had taken aneuroleptic; 1 of them had used 3 neuroleptics.

Verbal Recall With the 15-Word TestImmediate recall on day 1 was significantly lower than

on day 2. Delayed recall was also significantly lower onday 1 than on day 2. The results are presented in Table 1.

Visual Recall With the Photograph Recognition TestLess than half of the patients recognized the psychiat-

ric resident from the photograph and knew that he was theone they spoke with the day before (the recognizers)(Table 2). Scores of the recognizers on the verbal recalltest were not significantly different from the scores of thenonrecognizers. Likewise, the subjective ratings of seda-tion of the recognizers were not significantly differentfrom the ratings of the nonrecognizers.

SedationOn day 1, patients rated themselves as more sedated

than the residents rated them, but the difference was not

significant (mean ± SD VAS score = 5.09 ± 2.22 and5.80 ± 2.15, respectively; p = .142). A low correlationwas found between subjective ratings of the patients andimmediate or delayed recall on day 1 (Pearson r = 0.12,p = .463 and Pearson r = 0.08, p = .576, respectively). Ahigher correlation was found between ratings of the resi-dent and immediate and delayed recall on day 1 (Pearsonr = 0.5, p = .00 and Pearson r = 0.47, p = .00, respec-tively). When the 9 patients who had also used antide-pressants or neuroleptics were excluded, the results didnot differ from those of the whole group.

DISCUSSION

This study is the first to demonstrate memory impair-ment in patients who made a suicide attempt by takingbenzodiazepines. First of all, in a verbal recall test, pa-tients performed more poorly on the first day of admit-tance to the hospital than on the second day in immediateand delayed recall. These results are in agreement withthe existing wide body of literature in which verbal recalltest results demonstrate memory impairment after the in-take of benzodiazepines in a laboratory setting.5 Sec-ondly, impairment was found in a photo recognition task.This result is surprising because a number of factorsmake memory loss in this task less probable. Patientswere in contact much longer with the psychiatric residentthan with the words of the verbal recall test, given thatpsychiatric examination of these patients takes 45 min-utes on the average.16 Moreover, a psychiatric interviewimmediately after a suicide attempt is often a very emo-tional experience for the patient, implying that the inter-view and the interviewer are more meaningful and thusmore likely to be remembered than the words of the ver-bal recall test. In addition, in laboratory studies visualtasks are much less sensitive than verbal tasks, andrecognition tasks are less sensitive than recall tasks.3

The results of our study also show that anterograde am-nesia is not necessarily accompanied by a lowering ofconsciousness.

Table 2. Results of Interviewer Recognition FromPhotograph (N = 35)a

Question Yes/Correct No/Incorrect

Do you recognize anybody? 22 13Who do you recognize? 20 2How do you know this person? 12 8aSome patients had previously received treatment from the residentpsychiatrist and therefore did not participate in the photographrecognition test.

Table 1. Recall Ratings on the 15-Word Test (N = 43)Day 1 Day 2 p Value

Recall Mean SD Mean SD (t test)

Immediate 4.91 1.76 6.00 1.93 .002Delayed 3.02 1.82 4.26 1.77 .000

458


CME: ARTICLE



Because 9 patients had used psychopharmacologicagents in addition to benzodiazepines, we also studiedthose who had only taken benzodiazepines. The resultsdescribed above were also found in this subgroup of pa-tients who made a suicide attempt with benzodiazepinesonly, and it makes it more probable that benzodiazepinesare the main factor in the memory disturbances. The factthat the type of memory loss found here is typical for ben-zodiazepines also implies that the role of benzodiazepinesis an important factor. However, these arguments do notexclude the possibility that the impairment of memory re-sults from the turmoil caused by acute admittance to thehospital and the diagnostic and treatment procedures inthe acute treatment ward. On the other hand, stress canalso induce the amelioration of memory. Only a compari-son between patients having made a suicide attempt withbenzodiazepines and patients having made an attemptwithout benzodiazepines can provide a definite answer.At least in our hospital, such a control group is difficult toobtain because most suicide attempters use benzodiaze-pines, and patients who do not use them almost alwaysleave the hospital within a few hours after arrival withouthaving been seen by a psychiatrist.

Notwithstanding these shortcomings, the results sup-port the hypothesis that the efficacy of psychiatric consul-tation in patients who made a suicide attempt with benzo-diazepines can be compromised by memory impairment,even in patients who do not seem to be sedated. Withmany patients who have made a suicide attempt, arrange-ments for follow-up care and rules of life have to be nego-tiated. It is, of course, very important that the patients re-member these arrangements later on. In accordance withcase reports, our study lends support to the notion that pa-tients who made a suicide attempt with benzodiazepinesare likely to forget such arrangements. This is even true ifthe patient does not look sedated during the interview andseems to be cooperative with and responsive to the psy-chiatrist. Therefore, it is preferable to make arrangementslater, although this is not always possible in clinical prac-tice. We recommend providing essential information inwriting as well and, whenever possible, drawing signifi-

cant others into the discussions about the arrangementsthat have to be made.

Drug names: alprazolam (Xanax and others), clorazepate (Tranxene),diazepam (Valium and others), lorazepam (Ativan and others), oxaze-pam (Serax and others), temazepam (Restoril and others).

Disclosure of off-label usage: The authors have determined that, to thebest of their knowledge, no investigational information about pharma-ceutical agents has been presented in this article that is outside U.S.Food and Drug Administration–approved labeling.

REFERENCES

1. Barbee JG. Memory, benzodiazepines, and anxiety: integration of theoreti-cal and clinical perspectives. J Clin Psychiatry 1993;54(10, suppl):86–97

2. Curran HV. Tranquillising memories: a review of the effects of benzodiaze-pines on human memory. Biol Psychol 1986;23:179–213

3. Patat A, Klein MJ, Hucher M. Effects of single oral doses of clobazam,diazepam and lorazepam on performance tasks and memory. J ClinPharmacol 1987;32:461–466

4. Soldatos CR, Kales A, Bixler EO. Behavioral side effects of benzodiaze-pine hypnotics. Clin Neuropharmacol 1985;8:S112–S117

5. Curran HV. Benzodiazepines, memory and mood: a review. Psychophar-macology (Berl) 1991;105:1–8

6. Ghoneim MM, Hinrichs JV, Mewaldt SP. Dose-response analysis of the be-havioral effects of diazepam, I: learning and memory. Psychopharmacol-ogy (Berl) 1984;82:291–295

7. Preston GC, Broks P, Traub M, et al. Effects of lorazepam on memory, at-tention and sedation in man. Psychopharmacology (Berl) 1988;95:208–215

8. Ghoneim MM, Mewaldt SP, Berie JL, et al. Memory and performance ef-fects of single and 3-week administration of diazepam. Psychopharmacol-ogy (Berl) 1981;73:147–151

9. Morton S, Lader M. Studies with alpidem in normal volunteers and anxiouspatients. Pharmacopsychiatry 1990;23(suppl 3):120–123

10. Lucki I, Rickels K. The behavioral effects of benzodiazepines followinglong-term use. Psychopharmacol Bull 1986;22:424–433

11. Curran HV, Bond A, O’Sullivan G, et al. Memory functions, alprazolamand exposure therapy: a controlled longitudinal study of agoraphobia withpanic disorder. Psychol Med 1994;24:969–976

12. Laan WJ, Verwey B. Automutilatie: een zinvol gesprek op de juiste tijd.[Self mutilation: a meaningful discussion at the right time]. Ned TijdschrGeneeskd 1992;136:353–355

13. Michel K, Waeber V, Valach L, et al. A comparison of the drugs taken infatal and nonfatal self-poisoning. Acta Psychiatr Scand 1994;90:184–189

14. Ghoneim MM, Mewaldt SP. Benzodiazepines and human memory: a re-view. Anesthesiology 1990;72:926–938

15. Lezak MD. Neuropsychological Assessment. 2nd ed. New York, NY: Ox-ford University Press; 1983

16. Verwey B, Koopmans GT, Opmeer BC, et al. Richtlijnen en de nalevingdaarvan bij de psychiatrische opvang van suicidepogers in algemeneziekenhuizen. [Guidelines and their observance in the psychiatric care offailed suicides in general hospitals.] Ned Tijdschr Geneeskd 1997;141:1338–1342

459

https://www.researchgate.net/publication/15389548_A_comparison_of_the_drugs_taken_in_fatal_and_nonfatal_sel-poisoning?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/15389548_A_comparison_of_the_drugs_taken_in_fatal_and_nonfatal_sel-poisoning?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/16768121_Dose-response_analysis_of_the_behavioral_effects_of_diazepam_I_Learning_and_Memory?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/20277732_The_behavioral_effects_of_benzodiazepines_following_long-term_use?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/20277732_The_behavioral_effects_of_benzodiazepines_following_long-term_use?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/20276949_Tranquilising_memories_a_review_of_the_effects_of_benzodiazepines_on_human_memory_Biol_Psychol?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/20276949_Tranquilising_memories_a_review_of_the_effects_of_benzodiazepines_on_human_memory_Biol_Psychol?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/20017889_Effect_of_lorazepam_on_memory_attention_and_sedation_in_men?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/20017889_Effect_of_lorazepam_on_memory_attention_and_sedation_in_men?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/20269689_Effects_of_single_dose_of_clobazam_diazepam_lorazepam_on_performance_tasks_and_memory?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1



https://www.researchgate.net/publication/21182078_Studies_with_Alpidem_in_Normal_Volunteers_and_Anxious_Patients?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/21182078_Studies_with_Alpidem_in_Normal_Volunteers_and_Anxious_Patients?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/21471819_Curran_H_Benzodiazepines_memory_and_mood_a_review?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1

https://www.researchgate.net/publication/21471819_Curran_H_Benzodiazepines_memory_and_mood_a_review?el=1_x_8&enrichId=rgreq-da26f289-0bc0-496b-a389-faf7440c0491&enrichSource=Y292ZXJQYWdlOzEyNDE4NjY2O0FTOjE5NTg5NDA0MjQ2ODM1MkAxNDIzNzE2MTg3MzI1




CME: POSTTESTMemory Impairment in Those Who

Attempted Suicide by Benzodiazepine Overdose

1. Anterograde amnesia means:a. Diminished capacity to learn new informationb. Diminished capacity to retrieve old informationc. Diminished capacity to rehearse new informationd. Diminished capacity to consolidate old information

2. Anterograde amnesia is not a common finding involunteers after a single dose of the following drug:a. Diazepamb. Lorazepamc. Paracetamold. Midazolam

3. What is the prevalence of benzodiazepine use amongpeople hospitalized because of a suicide attempt?a. 10%b. 20%c. 35%d. 50%

4. Choose the right statement about tests used todemonstrate anterograde amnesia induced bybenzodiazepines:a. Visual tests are more sensitive than verbal tests.b. Recall tests are more sensitive than recognition tests.c. Recall tests have shown tolerance to the amnestic

effects.d. Test have shown that sedation is the main factor of the

amnestic effects.

5. Which test is most commonly used to demonstrateanterograde amnesia in users of benzodiazepines?a. 15-word testb. Minnesota Multiphasic Personality Inventory (MMPI)c. Symptom Checklist-90 (SCL-90)d. Wechsler Adult Intelligence Scale (WAIS)

6. A consulting psychiatrist who visited a patient within24 hours after admittance to the ward because of asuicide attempt with benzodiazepines should takeinto account that the next day:a. The patient may not recognize the consulting

psychiatrist.b. The patient may not remember what he had discussed

with the psychiatrist.c. The patient may neither recognize the psychiatrist nor

remember what was discussed.d. The patient will recognize the psychiatrist and

remember what was discussed.

7. When making arrangements for follow-up care forpatients who attempt suicide by benzodiazepineoverdose, the psychiatrist:a. Should ask the patient to repeat the arrangements

immediately after they were explained to her or himb. Should provide the information in writingc. Should inform the patient about the consequences of

memory impairmentd. Should only discuss the arrangements with the family of

the patient

Physicians may receive up to 1 hour of Category 1 credittoward the American Medical Association Physician’sRecognition Award by reading the article starting on page456 and correctly answering at least 70% of the questions inthe posttest that follows.

1. Read each question carefully and circle the correctcorresponding answer on the Registration form.

2. Type or print your full name and address and SocialSecurity, phone, and fax numbers in the spaces provided.

3. Send the Registration form along with a check, moneyorder, or credit card payment in the amount of $10 to:Physicians Postgraduate Press, Office of CME, P.O. Box752870, Memphis, TN 38175-2870.

4. For a credit certificate to be issued, answers must bepostmarked by the deadline shown on the CMERegistration form. After that date, correct answers to theposttest will be printed in the next issue of the Journal.All replies and results are confidential. Answer sheets,

once graded, will not be returned. Unanswered questions willbe considered incorrect and so scored. Your exact score canbe ascertained by comparing your answers with the correctanswers to the posttest, which will be printed in the Journalissue after the submission deadline. The PhysiciansPostgraduate Press Office of CME will keep only a record ofparticipation, which indicates the completion of the activityand the designated number of Category 1 credit hours thathave been awarded.

Instructions

Answers to the December 1999 CME posttest1. b 2. a 3. d 4. b 5. b 6. d 7. b 8. d

460


TEAR OUT AND MAIL THIS PAGE, ALONG WITH YOUR PAYMENT, TO:PHYSICIANS POSTGRADUATE PRESS • OFFICE OF CME • P.O. BOX 752870 • MEMPHIS, TN 38175-2870IF YOU ARE PAYING BY CREDIT CARD, YOU MAY FAX THIS PAGE TO: OFFICE OF CME AT 901-751-3444

QUESTIONS? CALL 1-800-489-1001 X123

CME: REGISTRATION/EVALUATIONMemory Impairment in Those Who

Attempted Suicide by Benzodiazepine Overdose



Circle the one correct answer for each question.1. a b c d2. a b c d3. a b c d4. a b c d5. a b c d6. a b c d

7. a b c d

Print or typeName ________________________________________

Social Security number _______ – _____ – _________(for CME credit recording purposes)

Degree __________Specialty _____________________

Affiliation _____________________________________

Address _______________________________________

City, State, Zip _________________________________

Phone ( ) _________________________________

Fax ( ) ___________________________________

E-mail ________________________________________

Hospital: ❏ Private Practice: ❏ Resident: ❏ Intern: ❏

Deadline for mailingFor credit to be received, the envelope must be postmarked

no later than November 30, 2000.

Keeping a copy for your filesRetain a copy of your answers and compare them with the

correct answers, which will be published after the submissiondeadline.

PaymentA $10 payment must accompany this form. You may pay by

check, money order, or credit card (Visa or MasterCard). Makecheck or money order payable to Physicians PostgraduatePress. If paying by credit card, please provide the informationbelow.

Check one: ❏ Visa ❏ MasterCard

Card number ___________________________________

Expiration date _________________________________

Your signature _________________________________

Please evaluate the effectiveness of this CME activity byanswering the following questions.

1. Was the educational content relevant to the statededucational objectives? ❏ Yes ❏ No

2. Did this activity provide information that is useful inyour clinical practice? ❏ Yes ❏ No

3. Was the format of this activity appropriate for the contentbeing presented? ❏ Yes ❏ No

4. Did the method of presentation hold your interest andmake the material easy to understand? ❏ Yes ❏ No

5. Achievement of educational objectives:

A. Enabled me to describe the potential cognitive effectsof benzodiazepines in patients admitted to thehospital after a suicide attempt with benzodiazepines.❏ Yes ❏ No

B. Enabled me to develop a strategy for an initialpsychiatric interview in a patient who has attemptedsuicide with benzodiazepines. ❏ Yes ❏ No

6. Did this CME activity provide a balanced, scientificallyrigorous presentation of therapeutic options related to thetopic, without commercial bias? ❏ Yes ❏ No

7. Does the information you received from this CMEactivity confirm the way you presently manage yourpatients? ❏ Yes ❏ No

8. Does the information you received from this CMEactivity change the way you will manage your patients inthe future? ❏ Yes ❏ No

9. Please offer comments and/or suggested topics for futureCME activities.

___________________________________________

___________________________________________

___________________________________________

___________________________________________

10. How much time did you spend completing this CMEactivity?

___________________________________________

461

Electrocardiographic abnormalities in patients with cerebrovascular disease

Documents

Transcript of Electrocardiographic abnormalities in patients with cerebrovascular disease