PM

las

i

Dw

REVIEW

Citalopram, QTc Interval Prolongation, and Torsade dePointes. How Should We Apply the Recent FDA Ruling?W. Victor R. Vieweg, MD,a,b Mehrul Hasnain, MD,c Robert H. Howland, MD,d John M. Hettema, MD, PhD,a,e

Christopher Kogut, MD,a Mark A. Wood, MD,b Ananda K. Pandurangi, MDa

aDepartment of Psychiatry and bDepartment of Internal Medicine, Virginia Commonwealth University, Richmond; cDepartment ofsychiatry, Memorial University, St John’s, NL, Canada; dWestern Psychiatric Institute and Clinic, University of Pittsburgh School of

edicine, Pittsburgh, Pa; eVirginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Richmond.

E-mail address

0002-9343/$ -see fdoi:10.1016/j.amjm

ABSTRACT

Recently, both the manufacturer of citalopram and the US Food and Drug Administration have warnedhealth care providers and patients about new information implicating drug-induced QTc interval prolon-gation and torsade de pointes when using citalopram in doses �40 mg/day. This warning is not placed inthe context of either benefits or risks in real-world clinical practice, leaving clinicians with an untenablechoice between depriving patients of high-dose citalopram or malpractice litigation. We reviewed theliterature and found no cases of citalopram-induced sudden cardiac death among patients taking up to 60mg/day of citalopram and free of risk factors for QTc interval prolongation and torsade de pointes. Becausepsychotropic drug-induced sudden cardiac death is an outlier in the absence of identified risk factors forQTc interval prolongation and torsade de pointes, we do not believe current Phase 3 and Phase 4 studiesprovide sufficient information to limit current prescribing practices for citalopram (20 mg to 60 mg/day).We urge drug manufacturers and regulatory agencies to periodically publish full case reports of psycho-tropic drug-induced QTc interval prolongation, torsade de pointes, and sudden cardiac death so thatclinicians and investigators may better understand the clinical implications of prescribing such drugs ascitalopram.© 2012 Elsevier Inc. All rights reserved. • The American Journal of Medicine (2012) 125, 859-868

KEYWORDS: Citalopram; Drug-induced QTc interval prolongation; Risk factors; Torsade de pointes

On August 12, 2011, the manufacturer (Forest LaboratoriesInc., New York, NY) of citalopram (marketed as Celexa)published a Dear Healthcare Professional letter.1 In thatetter, under the heading WARNINGS QT Prolongationnd Torsade de Pointes, several key points were made, ashown in Table 1 (Part A). Twelve days later, a US Food

and Drug Administration (FDA) Safety Announcement in-formed health care professionals and patients that citalo-pram (Celexa) should no longer be used at doses above 40mg/day because in higher doses the drug can unfavorablyalter the electrical activity of the heart, and higher doses

Funding: None.Conflicts of Interest: None.Authorship: All authors had access to the data and a substantive role

n writing the manuscript.Requests for reprints should be addressed to Victor Vieweg, MD,

epartment of Psychiatry, Virginia Commonwealth University, 17 Runs-ick Drive, Richmond, VA 23238-5414.

: vvieweg@verizon.net

ront matter © 2012 Elsevier Inc. All rights reserved.ed.2011.12.002

provide no additional benefit in the treatment of depres-sion.2 Key additional information for health care profession-als appears in Table 1 (Part B), including the admonitionthat citalopram should not be prescribed at a dose �20mg/day for patients who are more than 60 years old.

A data summary in the FDA Safety Announcement2

reported that the FDA evaluated a “. . . thorough QT study[unpublished] assessing the effects of 20-mg and 60-mgdoses of citalopram in adults . . .” in a randomized, multi-center, double-blind, placebo-controlled, crossover studyinvolving 119 subjects taking citalopram 20 mg/day on day9 and citalopram 60 mg/day on day 22. This summary foundthat citalopram 20-mg dose increased the QT interval 8.5msec, and the 60-mg dose increased the QT interval 18.5msec. From these data, the FDA analysis estimated thatcitalopram 40 mg would increase the QT interval 12.6 msec.For the reader interested in what constitutes a “thoroughQT/QTc study,” see the article by Darpo.3 Briefly, a QTc

interval �500 msec (increased risk of developing torsade de

pvaLeLapVrJ

diyCpr

Fup

biany

r4

pjpo

860 The American Journal of Medicine, Vol 125, No 9, September 2012

pointes) or drug-induced change from baseline �60 msecconstitute clinically significant thresholds.

Citalopram, a selective serotonin reuptake inhibitor, wasapproved by the FDA to treat depression, marketed as Cel-exa, and became available in the US in July 1998. By May2002, its share of new prescrip-tions for selective serotonin re-uptake inhibitors was 17.5%.4 Itsatent expired in 2003. It was de-eloped in the 1980s by Lundbecknd marketed in the US by Forestaboratories. Its S-stereoisomer,scitalopram, marketed by Forestaboratories as Lexapro, becamevailable in the US in 2002. Itsatent will expire in March 2012.iibryd (vilazodone hydrochlo-

ide) was approved by the FDA inanuary 20115 to treat major de-

pressive disorder and is now mar-keted by Forest Laboratories inthe US. Both citalopram and esci-talopram constitute potential com-petitors for vilazodone.

Animal toxicology studies inthe 1980s demonstrated that cita-lopram, when used in very high (4times higher than prescribed inhumans) doses long term, com-monly caused QT interval prolongation and fatalities inbeagle dogs.6 It is metabolized in the liver to desmethylci-talopram via CYP 2C19 and CYP 3A4, and then to dides-methylcitalopram via CYP 2D6.7 High concentrations ofidesmethylcitalopram were singularly associated with QTnterval prolongation in beagle dogs. Although didesmeth-lcitalopram is ordinarily a minor metabolite in humans,PY 2D6 ultra-rapid metabolizers (about 2% of the USopulation) could have high concentrations with increasedisk of QT interval prolongation and torsade de pointes.

In 1997 (before FDA approval of citalopram in 1998),orest Laboratories studied mean QTc intervals in 12 vol-nteers taking citalopram 60 mg and 11 volunteers takinglacebo (healthy men ages 21-41 years).6 They found no

significant difference (P � .46) between these 2 groups ataseline or steady state. There was no treatment-by-timenteraction (P � .2) or correlation between the QTc intervalnd didesmethylcitalopram (highest level 39 nM). Theseegative findings were to be expected with low didesmeth-lcitalopram concentrations.

Åström-Lilja et al8 searched a national Swedish databasecovering January 1, 1991 to February 1, 2006 (about 15years and 61,788 adverse drug reactions) for drug-inducedtorsade de pointes. They found 101 suspected cases ofdrug-induced torsade de pointes, of which 5 died (2 wereconsidered to have died from torsade de pointes). Of 88eligible cases involving 116 drugs, citalopram was impli-

CLINICAL SIGNIF

● There was no cocitalopram aboveas prescribed wcreased risk of QTor torsade de po

● Because theseevents are sucprinciples are nothe cardiac risks

● Drug manufactagencies shouldcase reports aboinduced cardiotinterval prolongpointes.

cated in 9 cases and occurred as a concomitant drug in an

additional 5 reports. They concluded that citalopram neededfurther investigation to better define any association withtorsade de pointes.

Kannankeril et al9 and the accompanying editorial com-mentary10 recently noted that the QT interval-prolonging

properties of a drug better predictthe risk of torsade de pointes thanthe baseline QTc interval, all otherfactors being equal. The “thor-ough study” evaluated by theFDA2 reported that clinicianscould expect citalopram adminis-tration at doses of 20 mg, 40 mg,and 60 mg to lengthen the QTcinterval by 8.5 msec, 12.6 msec,and 18.5 msec, respectively.Based on these data, theoretically,treatment with 60 mg of citalo-pram is more likely to provokeQTc interval prolongation or tor-sade de pointes than is a 20-mg or40-mg dose. However, neither themanufacturer nor the FDA has of-fered data to support such aspeculation.

RATIONALE FOR THISSTUDY

Camm et al,11 in a review of psychotropic drug-inducedQTc interval prolongation and torsade de pointes, pointedout that in the absence of overdose or accompanying riskfactors, citalopram-induced QTc interval prolongation and,particularly, torsade de pointes in doses up to 60 mg/day,are exceedingly rare events. Even large, controlled studiesare insufficient to determine the risk of drug-induced tor-sade de pointes.12 Thus, standard statistical principals maynot be clinically useful in quantitating or even semi-quan-titating these drug-induced cardiac risks.

We are particularly concerned that clinicians might acton the FDA warning and potentially destabilize patientswho might be doing well on citalopram 60 mg/day or higheror depriving some patients of the benefits of high-dosecitalopram. We will discuss the STAR-D trial13 later in thiseview, noting that the average daily dose of citalopram was1 mg.

High doses of citalopram are commonly used to treatatients with anxiety disorders. In a successful trial of ad-unctive citalopram combined with clomipramine to manageatients with treatment-resistant obsessive-compulsive dis-rder, citalopram 60 mg/day was employed.14

Recently, we reviewed the proarrhythmic effects of psy-chotropic drugs using a case-report format but did not dis-cuss citalopram specifically.15,16 In an effort to better de-fend our view that citalopram is no more cardiotoxic thanother selective serotonin reuptake inhibitors, we culled from

CE

ing evidence thatg/day when used

sociated with in-rval prolongation

.

ctrocardiographictliers, statisticalful in determiningtalopram.

and regulatoryiodically publishsychotropic drug-y, including QTcand torsade de

ICAN

nvinc40 m

as asc inteintes

eleh out useof ci

urersper

ut poxicitation

the literature, using a case-report format, what we know

ldotr

Cceoaf

78

9

861Vieweg et al Citalopram-induced QTc Interval Prolongation

about the capacity of citalopram to prolong the QTc inter-val, leaving the patient vulnerable to polymorphic ventric-ular tachycardia and its subtype torsade de pointes. Thateffort appears below.

METHODSOn October 13, 2011, we entered “citalopram and QTc” (24articles) and then “citalopram and torsade” (11 articles) intoPubMed. We selected case reports with abstracts for moredetailed review and subsequent commentary. Three addi-tional case reports were found in the process.

RESULTSWe identified 9 case reports, as shown in Table 2. Sevensubjects were women (ages 21, 36, 46, 48, 54, 76, and 83years) and 2 subjects were men (ages 31 and 81 years). Thecase report abstracts appear followed by our comments inTable 2. Where necessary, our comments provided addi-tional clinical material not appearing in the abstracts. Nocase reports demonstrated citalopram-associated QTc inter-val prolongation without either citalopram overdose or as-sociated well-known risk factors for QTc interval prolon-

Table 1 Key Points Recently Made by Forest Laboratories andde Pointes, and Drug Dosing

Part A“WARNINGS QT Prolongation and Torsade de Pointes” from Forest1) Celexa (Forest Laboratories Inc., New York, NY) causes dose-d

Torsade de pointes has been reported postmarketing.2) Celexa should not be used in patients with congenital long Q3) Hypokalemia and hypomagnesemia should be corrected befor4) Electrocardiographic monitoring is recommended in patients

concomitant medications that prolong the QT interval.5) Dose escalations over 20 mg in CYP2C19 poor metabolizers o

inhibitor are not recommended.Part BThe US Food and Drug Administration (FDA) Safety Announcemencitalopram for health care professionals2

1) Citalopram causes dose-dependent QT interval prolongation.2) Citalopram should not be used in patients with congenital lo3) Patients with congestive heart failure, bradyarrhythmias, or p

concomitant illness or drugs, are at higher risk of developing4) Hypokalemia and hypomagnesemia should be corrected befor

clinically indicated.5) Consider more frequent electrocardiogram (ECG) monitoring i

on concomitant medications that prolong the QT interval.6) 20 mg per day is the maximum recommended dose for patien

2C19 poor metabolizers, or who are taking concomitant cimelead to increased blood levels of citalopram, increasing the r

) No dose adjustment is necessary for patients with mild or mo) Advise patients to contact a health care professional immedi

or rhythm while taking citalopram.) Report adverse events involving citalopram to the FDA MedW

bottom of the page.

gation and torsade de pointes.15,16

DISCUSSION

Specific Case Reports (Table 2)Case Report #1.17 The main risk factor was 400 mg cita-opram overdose. Although the QTc interval lengtheneduring citalopram overdose, it did not exceed the upper limitf normal for women in this age group (470 msec).15 Nei-her torsade de pointes nor any other cardiac arrhythmia waseported.

ase Report #2.18 The main risk factor was 400 mgitalopram overdose. The most common serious adverseffect of citalopram overdose is a generalized seizure, asccurred in this man. The authors did not attempt tossess ventricular repolarization, tell whether the Bazettormula19 was used to calculate the QTc interval, or

speculate about the cause of the initial QTc intervalprolongation. The Bazett formula overestimates the QTcinterval at rapid heart rates.20 The major citalopram car-diac effect was drug-induced left bundle branch block(LBBB). In LBBB, the QRS duration must meet or ex-ceed 120 msec (40 msec longer than the common QRSduration of 80 msec). This case may not be an example

out Citalopram-induced QTc Interval Prolongation and Torsade

atories Dear Healthcare Professional letter of August 12, 20111

ent QT prolongation and should not be dosed above 40 mg/day.

rome.tion of treatment and periodically monitored.

ongestive heart failure, bradyarrhythmias, or patients on

tients taking concomitant cimetidine or another CPY2C19

ugust 24, 2011 providing key additional information about

ram should no longer be prescribed at doses �40 mg per day.syndrome.osition to hypokalemia or hypomagnesemia because ofe de pointes.nistering citalopram. Electrolytes should be monitored as

nts with congestive heart failure, bradyarrhythmias, or patients

hepatic impairment, who are �60 years of age, who are CYP(Tagamet; GlaxoSmithKline, Clifton, NJ), because these factorsQT interval prolongation and torsade de pointes.renal impairment.

f they experience signs and symptoms of an abnormal heart rate

rogram, using the information in the “Contact Us” box at the

FDA ab

Laborepend

T synde initiawith c

r in pa

t on A

Citalopng QTredisptorsad

e admi

n patie

ts withtidineisk ofderateately i

atch p

of citalopram-induced QTc interval prolongation. See

Table 2 Case Reports of Citalopram Associated with QTc Interval Prolongation

Year & Case Report Abstract Comment

Case #1:2001 Catalano et al17

QTc interval prolongationassociated with citalopramoverdose: a case reportand literature review

Citalopram is a member of the selective serotonin reuptake inhibitor classof antidepressants. In 1998, citalopram was approved by the US Foodand Drug Administration for the treatment of major depression. Likethe other selective serotonin reuptake inhibitors, citalopram enjoys arelatively benign side-effect profile compared with the tricyclicantidepressants and the monoamine oxidase inhibitors. However,citalopram has been associated with electrocardiographic changesand seizures at doses �600 mg per day. Fatalities have occurred withcitalopram-only overdoses. We report the case of a healthy 21-year-old woman who developed QTc interval prolongation after ingestionof approximately 400 mg citalopram. We discuss the cardiac effectsof citalopram, review previous cases of citalopram overdose, anddiscuss treatment recommendations.

She was brought to Emergency Department 1 hour after ingesting 20 20-mgcitalopram (400 mg) tablets after drinking “a couple of beers” (bloodalcohol 121 mg/dL) and taking a single alprazolam tablet. She had beenprescribed 30 mg daily (11⁄2 20 mg). She was irritable and cooperatedpoorly while crying and yelling occasionally. Vital signs and physicalexamination were unremarkable. She underwent treatment with activatedcharcoal and gastric lavage. The first electrocardiogram showed QTcinterval of 380 msec and 2 h later, it was 438 msec. Her course in theICU was unremarkable. 7 h after overdose, the QTc interval was 450msec, and 6 h later it was 457 msec. The QTc intervals 20 and 21 h afteroverdose were 393 msec and 353 msec, respectively. She remained stableand was transferred to the inpatient Psychiatry Service 30 h afteroverdose.

Case #22003 Engelbretsen et al18

Cardiotoxicity and lateonset seizures withcitalopram overdose

A 31-year-old man ingested 400 mg of citalopram (Celexa) after anargument with his parents and girlfriend 13 h before presentation.Paramedics witnessed the patient having a generalized clonic seizure.The electrocardiogram (ECG) revealed a wide QRS complex,prolongation of the QTc interval, and left bundle branch block(LBBB) pattern. He was treated with sodium bicarbonate withresolution of these changes. The patient was continued on a sodiumbicarbonate infusion and demonstrated no further ECG abnormalities.Sodium bicarbonate should be considered as a treatment modality inpatients with ECG abnormalities of prolongation of QRS or QTcinterval after citalopram overdose.

The initial toxic effect of citalopram was a generalized seizure. He waspostictal when he arrived in the Emergency Department and could notprovide a reliable history. His first electrocardiogram showed a ventricularrate of 130 beats/min, QRS duration of 124 msec in a LBBB pattern, andQT/QTc intervals of 344/506 msec. His rapid heart rate likely made theBazett estimation of QTc interval prolongation less accurate.20 Hereceived sodium bicarbonate intravenously and 25 min later QRS durationwas 95 msec, LBBB resolved, and QT/QTc intervals were 386/495 msec.Serum electrolytes were unremarkable. Serum citalopram was 1940 ng/mL(therapeutic rate up to 120 ng/mL). His clinical course was subsequentlyunremarkable and he was transferred to the psychiatry service on thesecond hospital day.

Case #32008 Kanjanauthai et al21

Citalopram inducedtorsade de pointes, a rarelife-threatening side effect

Acquired long QT syndrome (LQTS) is a disorder caused by medications,electrolyte imbalances, and drug interactions. This syndrome isassociated with an increased risk of a characteristic life-threateningcardiac arrhythmia, known as torsade de pointes (TdP). In thesetting of LQTS, selective serotonin reuptake inhibitors (SSRIs) canprecipitate TdP. We report the first case of LQTS and TdP induced bycitalopram in the US. After discontinuation of citalopram, the QT/QTcinterval normalized after 3 days and resolved further episodes of TdP.Patients on citalopram should be monitored closely for QT/QTcinterval to prevent torsade de pointes.

An 81-year-old man with hypertension, diabetes mellitus, end-stage kidneydisease, and depression presented to the Emergency Department becauseof dizziness after renal dialysis. Laboratory studies including electrolytesand serum magnesium were normal. An episode of torsade de pointes waswitnessed in the Emergency Department (QT/QTc intervals 572/695msec). He was admitted to the intensive care unit for telemetry andexperienced a second episode of torsade de pointes. He had startedcitalopram 4 months earlier and the drug was stopped with subsequentnormalization of the QTc interval and no further episodes of torsade depointes.

862The

American

Journalof

Medicine,

Vol125,

No9,

September

2012

Table 2 Continued

Year & Case Report Abstract Comment

Case #42008 Tarabar et al22

Citalopram overdose: latepresentation of torsade depointes (TdP) with cardiacarrest

IntroductionCitalopram overdose may produce bradycardia, QT prolongation, and

torsade de pointes (TdP). A cardiotoxic metabolite may be responsiblefor the delayed onset of cardiotoxicity. Although some authoritiesrecommend a minimum of 24 h of observation following citalopramoverdose, a recent analysis suggested that dysrhythmias rarely occurbeyond 13 h postingestion. We present a case of citalopram overdosewith a substantially delayed onset of cardiac toxicity.

Case ReportA 36-year-old woman complained of shakiness, numbness in the arms, and

palpitations that began approximately 32 h after ingesting 50 (20-mg)tablets of citalopram. Her initial vital signs were: blood pressure 84/44mm Hg; pulse 102-150 beats/min; respirations 17 breaths/min;temperature 37.3° C (99.3°F). Her initial ECG showed sinus rhythm witha prolonged corrected QT interval (572 msec) with paroxysmal, self-limited runs of wide-complex tachycardia that appeared multifocal innature. Approximately 20 min after presentation, she experienced self-terminating TdP, with transient hypotension and loss of consciousness.Her serum citalopram concentration (33 h postingestion) was 477 ng/mL(therapeutic: 40-110 ng/mL); desmethylcitalopram concentration was123.2 ng/mL (therapeutic: 14-40 ng/mL). She was treated withmagnesium and lidocaine, and her corrected QT interval remainedabnormal for 24 h after presentation.

DiscussionCitalopram overdose can produce life-threatening cardiac toxicity with

a clinical onset that may be delayed beyond a routine observationperiod of 6 h. Once the QT interval is prolonged, it seems prudent toprolong the observation period.

1000 mg of citalopram coupled with hypokalemia (serum potassium 3.1mmol/l—probably secondary to anorexia nervosa and drugs taken to loseweight) led to QTc interval prolongation and torsade de pointes about32 h later. Her rapid heart rate likely made the Bazett estimation of QTcinterval prolongation less accurate.20 Both citalopram anddesmethylcitalopram were 3-4 times the therapeutic range. Although thetitle includes citalopram-induced cardiac arrest, such an event and itsmanagement were not documented in the article. Treatment includedintravenous magnesium sulfate (normal serum calcium and magnesiumlevels) with production of bigeminy, lidocaine (no effect), potassiumadministration, intravenous isoproterenol, and temporary transvenouspacemaker. 1 day after admission, the QTc interval was 529 msec, and24 h later it was 442 msec. She was discharged after full recovery.

Case #52009 Bruggisser et al23

Medication-associatedlong QT syndrome

We report on a 76-year-old woman who fainted on her way to therestroom during the night. At the Emergency Department, aprolonged QT-interval was noticed in addition to sinus bradycardiaand marginal hypokalemia. The QT-interval normalized promptly aftercitalopram was discontinued. Taking into account the clinical pictureand the ascertained orthostatic dysregulation, a diagnosis oforthostatic syncope was made. Because of the lengthened QT-interval, drug-induced torsade de pointes ventricular arrhythmia wasconsidered as a differential diagnosis. We describe that citalopramwas most probably the cause for the prolonged QT-interval. In ourarticle we discuss the pathophysiology of drug-induced long QTsyndrome (LQTS), the most important drugs involved, and finally theprophylaxis and treatment of a TdP ventricular arrhythmia.

In the Emergency Department, the electrocardiogram showed QTc intervalof 526 msec (Bazett19) along with sinus bradycardia of 55 beats/min.Serum potassium was 3.4 mmol/L (normal 3.5-5.0 mmol/L). She wastreated for depression and schizophrenia with medications includingcitalopram 20 mg daily and olanzapine 1.25 mg daily. She also wastaking donepezil 10 mg and Gingko daily for cognitive impairment. Forrheumatoid arthritis, she took prednisone and methotrexate. Fordyslipidemia, she took atorvastatin 1 mg daily. Once citalopram wasstopped, QT/QTc intervals normalized at 410/428 msec. Although notdocumented, torsade de pointes may best explain her dizziness.

863View

eget

alCitalopram

-inducedQTc

IntervalProlongation

Table 2 Continued

Year & Case Report Abstract Comment

Case #62011 Liotier & Coudoré25

Drug monitoring of a caseof citalopram overdosage

Selective serotonin reuptake inhibitors are widely prescribed drugswithout recognized cardiovascular risk. We report the case of a 54-year-old patient who developed QTc-interval prolongation, followedby ventricular fibrillation episodes, 10 h after admission to the ICU,in the setting of a citalopram overdose. Citalopram plasma valuesdecreased from 5.88 to 0.34 mg/L at 9 days postadmission. Thepatient was treated by oral activated charcoal, and final outcome wasfavorable.

A middle-aged woman overdosed with citalopram (20 mg/day) and zopiclone(7.5 mg/day). In the Emergency Department, she was hypotensive witharterial desaturation (Glasgow Coma Scale 8/15). Blood alcohol level was 200mg/L and she was acidotic (pH 7.15) and hypokalemic (2.9 mmol/L). Theinitial serum citalopram level was 5.88 mg/L (therapeutic range 0.01-0.2mg/L). The initial QT/QTc intervals were 440/620 msec in the setting ofLBBB. Management included activated charcoal and countershock for 2episodes of ventricular fibrillation. She recovered completely and wasdischarged on hospital day 16. The authors believe the meticulous trackingof serum citalopram recorded in this case offers insight into citalopramkinetics in the setting of drug intoxication.

Case #72011 Fayssoil et al26

Torsade de pointesinduced by citalopram andamiodarone

Long QT syndrome (LQTS) is a disorder of myocardial repolarizationcharacterized by a prolonged QT interval on the electrocardiogram(ECG). This disorder is associated with an increased risk of torsade depointes (TdP). We report a case of TdP induced by citalopram, aselective serotonin reuptake inhibitor (SSRI), taken in conjunctionwith amiodarone.

An 83-year-old woman with hypertension, atrial fibrillation, and depressiondeveloped palpitations and dizziness and was evaluated in the EmergencyDepartment. Daily medications included amiodarone 200 mg, vitamin Kantagonist, losartan 50 mg, a dihydropyridine calcium antagonist, andcitalopram 20 mg (taken for 3 weeks). Serum potassium was 3 mmol/L.The first ECG showed prominent T-wave inversion in the precordial leadsand a QTc interval of 526 msec. An echocardiogram showed leftventricular hypertrophy consistent with age and hypertension. Cardiacmonitoring displayed self-limited torsade de pointes followed by frequentpremature ventricular contractions. Management included magnesiumsulfate and potassium administration and discontinuation of citalopramand amiodarone. No further torsade de pointes was noted and the QTcinterval had normalized by hospital day 3. She was discharged to home.

Case #82011 de Gregorio et al27

Citalopram-induced longQT syndrome and torsadede pointes: role forconcomitant therapy anddisease

Although citalopram is considered to be a safe antidepressant drug, somecases of arrhythmic disorders, syncope, or intoxication leading tolethal outcomes have been described over the last few years.

We report the case of citalopram-induced long QT syndrome complicatedby torsade de pointes in a young woman, in whom the role forantihypertensive treatment with diuretics and intercurrent flu-likesyndrome is discussed.

A 48-year-old woman taking 40 mg citalopram daily for 5 years for depressionand furosemide 25 mg daily for hypertension presented to the EmergencyDepartment with recent syncope and acute gastroenteritis with 5-day historyof diarrhea. The initial electrocardiogram showed nonspecific ST-T changesand prolonged QT/QTc intervals (620/670 msec). Serum potassium was 2.1mmol/L (normal range 3.5-5.0 mmol/L) and serum magnesium was 1.6 mg/dL (normal range 1.5-3.8 mg/dL). 3 episodes of torsade de pointes occurredin the hospital during the first 24 h, with 2 of them requiringelectrocardioversion. Citalopram was stopped and potassium and magnesiumadministered intravenously during the first 3 hospital days. By the fifthhospital day, the electrocardiogram and electrolytes had normalized. Follow-up was unremarkable.

864The

American

Journalof

Medicine,

Vol125,

No9,

September

2012

diaa

astcfiomca

Cdp(cw

Cble

2Co

ntin

ued

r&

Case

Repo

rtAb

stra

ctCo

mm

ent

e#9

11U

nter

ecke

ret

al28

rela

tion

ofQT

cin

terv

allo

ngat

ion

and

seru

mel

ofci

talo

pram

afte

rox

icat

ion

-a

case

ort

Cita

lopr

am(C

IT)

isa

wid

ely

used

anti

depr

essa

ntth

atac

tsby

sele

ctiv

ese

roto

nin

reup

take

inhi

biti

on.

Itis

cons

ider

edto

besa

fer

than

tric

yclic

anti

depr

essa

nts

atth

erap

euti

cle

vels

,bu

tal

sow

ith

resp

ect

toin

toxi

cati

ons.

We

repo

rtth

eca

seof

a46

-yea

r-ol

dw

oman

who

inge

sted

,in

suic

idal

inte

ntio

n,14

00m

gCI

T.Du

ring

the

follo

win

gin

pati

ent

trea

tmen

t,re

peat

edEC

Gsan

dde

term

inat

ions

ofth

ese

rum

leve

lof

CIT

wer

epe

rfor

med

.In

itia

llyth

epa

tien

t’sse

rum

leve

lof

CIT

was

1231

ng/m

Lan

dQT

cin

terv

alw

as54

1.60

mse

c.It

took

12da

ysun

tilth

ese

rum

leve

lof

CIT

fell

belo

wth

eup

per

thre

shol

dof

the

reco

mm

ende

dth

erap

euti

cra

nge

(130

ng/m

L).

The

QTc

inte

rval

onth

esi

xth

day

afte

rth

ein

toxi

cati

onfo

rth

efir

stti

me

was

belo

w50

0m

sec.

The

QTc

inte

rval

corr

elat

edsi

gnifi

cant

lyw

ith

the

seru

mle

vel

ofCI

Taf

ter

into

xica

tion

(r�

0.94

3;P

�.0

05).

Alth

ough

CIT

ises

tim

ated

asa

safe

anti

depr

essa

ntre

gard

ing

seri

ous

adve

rse

effe

cts,

toxi

cdo

ses

can

lead

topo

tent

ially

haza

rdou

sEC

Gch

ange

sw

hich

,ac

cord

ing

toou

rfin

ding

s,co

rrel

ate

stro

ngly

wit

hth

ese

rum

leve

lof

the

drug

.

A46

-yea

r-ol

dw

oman

inte

ntio

nally

over

dose

dw

ith

70ta

blet

sof

cita

lopr

am20

mg

(140

0m

g)an

d10

tabl

ets

ofop

ipra

mol

50m

g(5

00m

g)(a

ntid

epre

ssan

tan

dan

xiol

ytic

used

inGe

rman

yan

dot

her

Euro

pean

coun

trie

san

dis

atr

icyc

lican

tide

pres

sant

but

does

not

act

asa

reup

take

inhi

bito

r—ra

ther

,it

acts

asa

sigm

are

cept

orag

onis

t).

The

usua

ldo

seof

opip

ram

olis

inth

e20

0m

g/da

yra

nge.

Opip

ram

olis

not

asso

ciat

edw

ith

QTc

inte

rval

prol

onga

tion

orto

rsad

ede

poin

tes.

43

Init

ial

seru

mm

agne

sium

was

slig

htly

decr

ease

d.Se

rial

elec

troc

ardi

ogra

ms

and

seru

mci

talo

pram

leve

lm

easu

rem

ents

took

plac

eov

er23

days

,sh

owin

ga

good

corr

elat

ion

betw

een

QTc

inte

rval

and

seru

mci

talo

pram

leve

ls.

865Vieweg et al Citalopram-induced QTc Interval Prolongation

our comments later about LBBB and ventricularrepolarization.

Case Report #3.21 The main risk factors contributing totorsade de pointes were advanced age, severe medical ill-nesses, and citalopram administration. When citalopramwas stopped, QTc interval prolongation and torsade depointes resolved. Chronic renal failure may have contrib-uted to elevated serum citalopram levels.

Case Report #4.22 The main risk factors were citalopramoverdose and hypokalemia. Even though the serum potas-sium level was low (3.1 mmol/L), the cardiac muscle con-centration of potassium is the real determinant of risk forQTc interval prolongation and torsade de pointes. Bothstopping citalopram and administrating potassium were as-sociated with normalization of the QTc interval and nofurther episodes of torsade de pointes.

Case Report #5.23 The main risk factors were citalopram,onepezil, olanzapine, sinus bradycardia—possibly donepezil-nduced—slight hypokalemia, and advanced age. Donepezillone may induce bradycardia, QTc interval prolongation,nd torsade de pointes.24

Case Report #6.25 The main risk factors were citalopramoverdose with documented elevated serum levels, hypoka-lemia (2.9 mmol/L), and LBBB. The QTc interval of 620msec provides compelling evidence of prolonged ventricu-lar repolarization in the setting of LBBB. However, onemay not be able to extrapolate this finding to subjects takingcitalopram without drug-induced LBBB. See our commentslater about LBBB and ventricular repolarization. Also, aGlasgow Coma Scale score of 8/15 may be sufficientlyunrepresentative of patients taking citalopram therapeuti-cally to exclude this case.

Case Report #7.26 The main risk factors were citalopram,miodarone, hypokalemia (3.0 mmol/L), advanced age, andevere medical illness. The most important risk factor fororsade de pointes in this case was hypokalemia. Citalopramomes in a distant second and may better follow the riskactors of being an elderly woman with severe medicalllness. The authors seem to implicate citalopram and ami-darone equally. QTc interval correction after only 3 daysay reduce the implication of amiodarone as an important

ause of torsade de pointes because this antiarrhythmicgent has a long half-life.

ase Report #8.27 Risk factors were citalopram, possibleehydration increasing citalopram serum concentration, hy-okalemia (2.1 mmol/L), and borderline serum magnesium1.6 mg/dL). Significant hypokalemia was the most likelyause of torsade de pointes given the 5-year history ofell-tolerated 40 mg citalopram daily.

ase Report #9.28 The main risk factor was citalopram

overdose, with documented very elevated serum levels al-Ta Ye

a

Cas

20 Cor

pro

lev

int

rep

tot

pcvirfeo

pdofdmptw

LadL

vv

rTceasi9

cp

866 The American Journal of Medicine, Vol 125, No 9, September 2012

most certainly explaining the QTc interval prolongation.Fortunately, torsade de pointes did not develop.

Case Report CommentaryNone of the case reports documented QTc interval prolon-gation or polymorphic ventricular tachycardia of the torsadede pointes type due to citalopram administration in the“therapeutic” dose range 20-60 mg, as documented in ear-lier package inserts without concomitant well-known riskfactors for QTc interval prolongation and torsade depointes.15 Statistical analysis alone of citalopram’s capacityo induce QTc interval prolongation compared with baseliner controls appears to be an inadequate database upon whicho make regulatory decisions.

LBBB and Ventricular RepolarizationCases #2 and #6 developed LBBB following citalopramoverdose. Ventricular depolarization (QRS duration—com-monly approximately 80 msec) occurs when specializedconduction pathways activate the left and right ventricles.Ventricular repolarization (JT interval—commonly ap-roximately 320 msec) derives largely from myocyte intra-ellular potassium shifts and takes about 4 times as long asentricular depolarization.15 Survival is in greater jeopardyf ventricular depolarization is prolonged than if ventricularepolarization is prolonged, even though there is more timeor risk factors to act during the latter event than the formervent because the pumping action of the heart and cardiacutput derive critically from ventricular depolarization.

In LBBB, ventricular depolarization follows an abnormalattern that moves more slowly than through normal con-uction tissue. Single or multiple lesions in the left bundler its branches may produce LBBB. Because depolarizationollows an abnormal pathway in LBBB, repolarization alsoeviates from the normal process. However, QT interval iseasured in units of time rather than as an organizational

attern. Therefore, there may or may not be a difference inhe time required for repolarization (JT interval) with andithout LBBB.Assessing QTc interval prolongation in the setting of

BBB has challenged cardiologists for many years.29 Ands we showed, LBBB may be intermittent or rate-depen-ent, with clear-cut normalization of the QTc interval whenBBB is not present.30

Talbot,31 almost 40 years ago, asserted that correction ofthe “actual” QTc interval in LBBB can be made by sub-tracting QRS prolongation (for example, 120 msec due toLBBB minus 80 msec � 40 msec) from the Bazett-formula-derived QTc interval. Das32 studied QT interval and repo-larization time in subjects with intraventricular conductiondelay and reported that LBBB did not prolong ventricularrepolarization. Calo et al33 looked at the effects of LBBB onentricular repolarization dispersion and found no JT inter-al lengthening in this setting.

More recently, Rautaharju et al34 studied 11,739 men and

women with normal ventricular conduction and 1251 sub-

jects with ventricular conduction defects. They pointed outthat the JT interval adjustment derived by subtracting QRSduration from the QTc interval was ventricular rate depen-dent and not recommended in the setting of LBBB. Theseauthors asserted that determining ventricular repolarizationin the setting of LBBB derives from a statistical modeloffered by them. Therefore, speculating about the magni-tude of citalopram-induced QTc interval prolongation in thesetting of LBBB is highly problematic at best.

Pfizer 054 StudyRelevant to our thesis is the Pfizer 054 study.35 In this study,ziprasidone administration (n � 31), at a steady-state doseof 160 mg/day without coadministration of a metabolicinhibitor, was associated with a mean 20.3-msec increase inthe QTc interval above baseline. Quetiapine administration(n � 27), at a steady-state dose of 750 mg/day with the CYP3A4 metabolic inhibitor ketoconazole co-administered, wasassociated with a mean 19.7-msec increase in the QTcinterval above baseline. We have been unable to find eitherisolated (no other risk factors) ziprasidone-induced or que-tiapine-induced torsade de pointes at therapeutic antipsy-chotic drug doses in recent literature reviews on thistopic.15,16,36

Other Important QTc Interval IssuesZhang et al37 assessed QT interval duration and mortalityate. The authors studied 7828 men and women from thehird National Health and Nutrition Examination Surveyonducted from 1988 through 1994 and assessed mortalitynd points through December 2006 (2291 deaths), yieldingn average follow-up of 13.7 years. They reported a U-haped curve to describe the association between the QTnterval and mortality end points. Subjects at or above the5th percentile of age-, sex-, race-, and R-R interval-cor-

rected QT interval (�439 msec) had a hazard ratio of 2.03(95% confidence interval, 1.46-2.81) for total mortalitycompared with subjects in the middle quintile of 401 to�410 msec. Subjects in the fifth percentile (�377 msec),compared with the middle quintile, had a hazard ratio of1.39 (95% confidence interval, 1.02-1.88) for total mortal-ity. That is, one can demonstrate increased mortality ateither end of the distribution of QT interval measurementsconsidered in the “normal” range. (Such findings may ex-plain the observation that chronically psychotic patientstaking a variety of antipsychotic drugs are twice as likely asthe general population to experience sudden cardiacdeath38,39). And, this observation further complicates thelinical implications of citalopram-induced QTc intervalrolongation.

Star*DCitalopram (20-60 mg) was selected as the initial treatmentin the STAR*D Study ($35-million National Institute ofMental Health-funded Sequenced Treatment Alternatives to

Relieve Depression [STAR*D] Study) conducted to deter-

irippd“e

2

3

867Vieweg et al Citalopram-induced QTc Interval Prolongation

mine the effectiveness of different treatments for patientswith major depression who had not responded to initialantidepressant treatment.13 STAR*D constituted the largestand longest study ever conducted to evaluate treatment ofdepression. Warden et al40 have provided a comprehensivereview of STAR*D findings. This “real-world” study hasbeen used by many clinicians in developing strategies totreat depression in a variety of settings. Compromising thisstudy by disregarding findings related to prescribing citalo-pram at a dose of 60 mg/day may represent a major loss inmanaging depression.

Howland41,42 recently reviewed citalopram cardiotoxic-ty in STAR*D, noting that more than half of patientseceived more than 40 mg/day of citalopram. STAR*D wasntended to be a “real-world” clinical trial in which 40% ofatients were recruited from primary care sites. As such,atients had comorbid medical conditions including cardiaciseases such as coronary heart disease. There was no clearcardiac” signal from the STAR*D adverse event profile,ven for doses of citalopram above 40 mg/day.

Study LimitationsOur article is based on fewer than 10 case reports. We havepublished extensively on the subject of psychotropic drug-induced QTc interval prolongation and torsade de pointes,but those publications have largely depended on cases ofour own and case reports in the literature. We have not beenable to apply statistical principles in our studies.

CONCLUSIONS

1) Contrary to the assertions of the manufacturer of citalo-pram1 and the FDA2 published in August 2011, wefound no convincing evidence that citalopram, whenused as prescribed in doses above 40 mg/day, was asso-ciated with an increased risk of QTc interval prolonga-tion and torsade de pointes, as long as clinicians attendedto well-known risk factors. And this drug at a 60 mg-dose is no more likely to induce QTc interval prolonga-tion than such drugs as ziprasidone and quetiapine. Alarge body of experience and data attest to the clinicaleffectiveness and cardiac safety of citalopram. Truncat-ing its use at this time seems premature at best.

) Our current approach to assessing drug-induced QTcinterval-related problems has serious limitations. Be-cause psychotropic drug-induced QTc interval prolonga-tion and torsade de pointes are such outliers, we do notbelieve statistical principles should guide patient man-agement for the foreseeable future.

) For such rare side effects, case reports would providemost useful practical information. We believe drug man-ufacturers and regulatory agencies should periodicallypublish case reports about psychotropic drug-inducedcardiotoxicity, including QTc interval prolongation and

torsade de pointes.

References1. Forest Laboratories. CELEXA (citalopram hydrobromide) causes dose

dependent QT prolongation (2011). Available at: http://www.frx.com/pdf/DearDrCelexa%20HCP%20Letter%2008_12_2011.pdf. AccessedFebruary 12, 2012.

2. US Food and Drug Administration. FDA Drug Safety Communication:Abnormal heart rhythms associated with high doses of Celexa (cita-lopram hydrobromide) (August 24, 2011). Available at: http://www.fda.gov/Drugs/DrugSafety/ucm269086.htm. Accessed February12, 2012.

3. Darpo B. The thorough QT/QTc study 4 years after the implementa-tion of the ICH E14 guidance. Br J Pharmacol. 2010;159:49-57.

4. Berfield S. A CEO and his son. The remarkable saga of HowardSolomon, his son Andrew, and his company, Forest Labs . . . a journeyof depression and hope. BusinessWeek. May 27, 2002.

5. Food and Drug Administration. FDA approves Viibryd to treat majordepressive disorder (2011). Available at: www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm240642.htm. Accessed February12, 2012.

6. Rasmussen SL, Overo KF, Tanghoj P. Cardiac safety of citalopram:prospective trials and retrospective analyses. J Clin Psychopharmacol.1999;19:407-415.

7. Jimmink A, Caminada K, Hunfeld NG, Touw DJ. Clinical toxicologyof citalopram after acute intoxication with the sole drug or in combi-nation with other drugs: overview of 26 cases. Ther Drug Monit.2008;30:365-371.

8. Åström-Lilja C, Odeberg JM, Ekman E, Hägg S. Drug-induced tor-sades de pointes: a review of the Swedish pharmacovigilance database.Pharmacoepidemiol Drug Saf. 2008;17:587-592.

9. Kannankeril PJ, Norris KJ, Carter S, Roden DM. Factors affecting thedegree of QT prolongation with drug challenge in a large cohort ofnormal volunteers. Heart Rhythm. 2011;8:1530-1534.

10. Ding C. Predicting the degree of drug-induced QT prolongation andthe risk for torsades de pointes. Heart Rhythm. 2011;8:1535-1536.

11. Camm AJ, Malik M, Yap YG. Acquired Long QT Syndrome. London:Blackwell Futura; 2004.

12. Wenzel-Seifert K, Wittmann M, Haen E. Torsade de pointes episodesunder treatment with selective serotonin reuptake inhibitors. Pharma-copsychiatry. 2010;43:279-281.

13. National Institute of Mental Health. Sequenced Treatment Alternativesto Relieve Depression (STAR*D) Study (2011). Available at: http://mentalhealth.gov/trials/practical/stard/index.shtml. Accessed February12, 2012.

14. Marazziti D, Golia F, Consoli G, et al. Effectiveness of long-termaugmentation with citalopram to clomipramine in treatment-resistantOCD patients. CNS Spectr. 2008;13:971-976.

15. Vieweg WVR, Wood MA, Fernandez A, Beatty-Brooks M, HasnainM, Pandurangi AK. Proarrhythmic risk with antipsychotic and antide-pressant drugs: implications in the elderly. Drugs Aging. 2009;26:997-1012.

16. Vieweg WVR, Hasnain M, Wood MA, Fernandez A, Lesnefsky EJ,Pandurangi AK. Cardiometabolic risks of antidepressant and antipsy-chotic drugs, Part 1. Proarrhythmic risks of antidepressant and anti-psychotic drugs. Psychiatr Times. 2011;28:50-55.

17. Catalano G, Catalano MC, Epstein MA, Tsambiras PE. QTc intervalprolongation associated with citalopram overdose: a case report andliterature review. Clin Neuropharmacol. 2001;24:158-162.

18. Engebretsen KM, Harris CR, Wood JE. Cardiotoxicity and late onsetseizures with citalopram overdose. J Emerg Med. 2003;25:163-166.

19. Bazett HC. An analysis of the time-relations of electrocardiograms.Heart. 1920;7:353-370.

20. Indik JH, Pearson EC, Fried K, Woosley RL. Bazett and Fridericia QTcorrection formulas interfere with measurement of drug-inducedchanges in QT interval. Heart Rhythm. 2006;3:1003-1007.

21. Kanjanauthai S, Kanluen T, Chareonthaitawee P. Citalopram inducedtorsade de pointes, a rare life threatening side effect. Int J Cardiol.

2008;131:e33-e34.

2

2

2

3

3

3

3

3

3

3

3

4

4

4

868 The American Journal of Medicine, Vol 125, No 9, September 2012

22. Tarabar A, Hoffman R, Nelson L. Citalopram overdose: late presen-tation of torsades de pointes (TdP) with cardiac arrest. J Med Toxicol.2008;4:101-105.

23. Bruggisser M, Bravo A, Bodmer M. Medication associated long QTsyndrome. Praxis (Bern 1994). 2009;98:1409-1415.

24. Tanaka A, Koga S, Hiramatsu Y. Donepezil-induced adverse sideeffects of cardiac rhythm: 2 cases report of atrioventricular block andTorsade de Pointes. Intern Med. 2009;48:1219-1223.

25. Liotier J, Coudoré F. Drug monitoring of a case of citalopram over-dosage. Drug Chem Toxicol. 2011;34:420-423.

6. Fayssoil A, Issi J, Guerbaa M, Raynaud JC, Heroguelle V. Torsade depointes induced by citalopram and amiodarone. Ann Cardiol Angéiol.2011;60:165-168.

27. de Gregorio C, Morabito G, Cerrito M, Dattilo G, Oreto G. Citalo-pram-induced long QT syndrome and torsade de pointes: role forconcomitant therapy and disease. Int J Cardiol. 2011;148:226-228.

8. Unterecker S, Warrings B, Deckert J, Pfuhlmann B. Correlation ofQTc interval prolongation and serum level of citalopram after intoxi-cation – a case report. Pharmacopsychiatry. 2012;45(1):30-34.

9. Lanjewar P, Pathak V, Lokhandwala Y. Issues in QT interval mea-surement. Indian Pacing Electrophysiol J. 2004;4:156-161.

0. Vieweg WVR, Stanton KC, Alpert JS, Hagan AD. Rate-dependent leftbundle branch block with angina pectoris and normal coronary arte-riograms. Chest. 1976;69:123-124.

1. Talbot S. QT interval in right and left bundle-branch block. Br Heart J.1973;35:288-291.

2. Das G. QT interval and repolarization time in patients with intraven-tricular conduction delay. J Electrocardiol. 1990;23:49-52.

3. Calo L, Levato M, Sciarra L, et al. The effects of a left bundle-branchblock on ventricular repolarization dispersion. Cardiologia. 1996;41:

1073-1077.

34. Rautaharju PM, Zhang ZM, Prineas R, Heiss G. Assessment of pro-longed QT and JT intervals in ventricular conduction defects. Am JCardiol. 2004;93:1017-1021.

35. Psychopharmacological Drugs Advisory Committee July 19. BriefingDocument for ZELDOX CAPSULES (Ziprasidone HCl) – Pfizer. Avail-able at: www.fda.gov/ohrms/dockets/ac/00/backgrd/3619b1a.pdf. Ac-cessed February 12, 2012.

6. Vieweg WVR, Schneider RK, Wood MA. Torsade de pointes in apatient with complex medical and psychiatric conditions receivinglow-dose quetiapine. Acta Psychiatr Scand. 2005;112:319-323.

7. Zhang Y, Post WS, Dalal D, Blasco-Colmenares E, Tomaselli GF,Guallar E. QT-interval duration and mortality rate: results from theThird National Health and Nutrition Examination Survey. Arch InternMed. 2011;171:1727-1733.

8. Ray WA, Meredith S, Thapa PB, Meador KG, Hall K, Murray KT.Antipsychotics and the risk of sudden cardiac death. Arch Gen Psy-chiatry. 2001;58:1161-1167.

9. Ray WA, Chung CP, Murray KT, Hall K, Stein CM. Atypical anti-psychotic drugs and the risk of sudden cardiac death. N Engl J Med.2009;360:225-235.

0. Warden D, Rush A, Trivedi M, Fava M, Wisniewski S. The STAR*Dproject results: a comprehensive review of findings. Curr PsychiatryRep. 2007;9:449-459.

1. Howland RH. A critical evaluation of the cardiac toxicity of citalo-pram: part 1. J Psychosoc Nurs Ment Health Serv. 2011;49(11):13-16.

2. Howland RH. A critical evaluation of the cardiac toxicity of citalo-pram: part 2. J Psychosoc Nurs Ment Health Serv. 2011;49(12)13-16.

43. Wenzel-Seifert K, Wittmann M, Haen E. QTc prolongation by psy-chotropic drugs and the risk of torsade de pointes. Dtsch Arztebl Int.

2011;108:687-693.