Restless LegsSyndrome Causedby Quetiapine

Successfully TreatedWith Ropinirole in2 Patients WithBipolar Disorder

To the Editors:

Quetiapine is an atypical antipsy-chotic medication indicated for thetreatment of bipolar disorder and schizo-phrenia.1 It has a higher affinity forserotonin (5-HT2A) receptors relative todopamine (D2) receptors, specifically inthe mesolimbic and mesocortical dopa-mine systems.2 The moderate affinityfor D2 receptors distinguishes quetiapinefrom the other atypical antipsychoticsthat as a class exhibit high D2 receptoraffinity. The incidence of extrapyrami-dal symptoms, a side effect of antipsy-chotic medication, is similar betweenquetiapine and placebo.3 Patients withexisting extrapyramidal symptomsbecause of other antipsychotic usageimproved when switched to quetiapine.4

Another movement/sleep disorder,restless legs syndrome (RLS), is a dis-tressing condition that can affect aperson’s quality of life. Restless legs syn-drome is a clinical diagnosis that mustmeet the 4 criteria set by the InternationalRestless Legs Syndrome Study Group:(1) the urge to move usually associatedwith uncomfortable leg sensations (par-esthesias), (2) the sensation is worse in theevening, (3) the sensation is worse withinactivity, and (4) movement results insymptom relief.5–7 It can be caused bypsychotropic medications8–11 as well asrelieved by psychotropic medications.4,12

Quetiapine has been reported to improveRLS4 as well as cause it.10

Ropinirole, a dopamine agonist withhigh relative in vitro specificity and fullintrinsic activity at the D2 and D3 dopa-mine receptor subtypes (with higher affi-

nity for the D3 receptor), has been shownto be effective in both the acute and main-tenance phases of treatment of RLS.13

This article presents 2 cases of RLSapparently caused by quetiapine effec-tively treated with ropinirole in patientswith bipolar disorder. The initial studieson the efficacy of ropinirole for RLSwere completed using relatively healthyparticipants. In clinical practice, RLSpatients may have complex medicalpresentations including taking multipleother medications, some of which maycause RLS symptoms. An importantquestion is whether ropinirole is alsoeffective in cases of apparent drug-induced RLS. These 2 cases suggestthat it is efficacious for treating RLSwithout negatively impacting patientsdiagnosed with bipolar disorder. Noadditional patients with RLS due toquetiapine unsuccessfully treated withropinirole have been noted in thispsychiatrist’s (M.U.) practice.

CASE 1Ms A. is a 53-year-old white woman

diagnosed with bipolar II disorder who hasreceived multiple medications over the last6 years to maintain her mood. She has chronicpain from a crushed lumbar vertebrae andherniated disk treated with chlorzoxazone. Shealso had a hysterectomy and takes estrogenreplacement. Her iron and ferritin levels werewithin normal limits. She uses no alcohol orrecreational drugs. She smokes 1 pack of cig-arettes per day and drinks 1 cup of coffee aday. She has had trials of olanzapine, ven-lafaxine, trazodone, gabapentin, risperidone,sodium valproate, fluoxetine, and lamotrigine.She had a brief trial of quetiapine in January2003 for 30 days but discontinued the medi-cation because it disrupted her sleep. At thattime, shewas not specifically questioned aboutsymptoms of RLS. She was readministeredquetiapine in November 2006 with increasingdoses to a stable dosage of 600 mg every nightat bedtime (qhs) while also taking clonazepam,lorazepam, bupropion, and oxcarbamazepine.She began to complain of poor sleep whentaking 200 mg of quetiapine. At the same time,she was taking narcotics for back pain. Whenher narcotic pain medicine was discontinuedafter spinal fusion surgery in July 2007, shedescribed classic symptoms of RLS as thecause of her disrupted sleep. She was admin-

istered ropinirole 0.25 mg qhs with some im-provement. When she reached a dose of 1 mgof ropinirole in November 2007, her RLSsymptoms abated within 2 days. She hastolerated the ropinirole with no side effectswhile continuing quetiapine and has devel-oped no tolerance to ropinirole. The patientdescribes a return of RLS symptoms if shemisses her dose for more than 3 days.

CASE 2Ms B. is a 52-year-old white woman

with a diagnosis of bipolar disorder with psy-chotic features. Ms B. has had 3 inpatientadmissions, one in 1998 for a major depressiveepisode, one in 1994 for a recurrent depres-sive episode, and one in 2003 for manic andpsychotic symptoms. She has been tried onmultiple psychiatric medications includingthioridazine, fluoxetine, lorazepam, trazodone,sodium valproate, risperidone, clonazepam,gabapentin, carbamazepine, venlafaxine, lith-ium carbonate, duloxetine, and escitalopram.She underwent a gastroplasty for weight reduc-tion in 1981, which was revised to a gastricbypass in 1997. She takes atenolol for hyper-tension. She has had 3 ferritin levels checkedin the past 2 years, all of which were normal.She did have a low iron level in August 2006,which was corrected by iron replacement. Sheuses no alcohol, nicotine, or other recreationaldrugs. She drinks 2 to 3 cups of coffee in themorning. She has been taking quetiapine sinceNovember 2005 in low doses. When quetia-pine was increased to 600 mg in April 2006,poor sleep began. She then described uncom-fortable leg sensations. In August 2006, shewas diagnosed as having RLS with docu-mented periodic leg movements during a sleepstudy in August 2007 performed while she wastaking quetiapine. Iron replacement did notrelieve the RLS symptoms. Because the patientresponded well to quetiapine, ropinirole wasadded. Her RLS symptoms resolved withropinirole 0.25 mg qhs in 2 to 3 days, with noworsening of her psychotic symptoms. Her cur-rent medications include quetiapine 300 mg qhs,oxcarbamazepine (since November 2005), lor-azepam on an as-needed basis, and ropinirole,which remains effective in treating her RLS.If the patient’s dosage of quetiapine increases,she requires a higher dose of ropinirole to re-lieve her RLS symptoms.

Both patients retrospectively completedthe RLS rating scale.7 The score of patient 1went from severe (29), 9 months before therating scale administration, to mild (2), andthe score of patient 2 went from severe (27),

Letters to the Editors

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008704

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

8 months before the rating scale administra-tion, to mild (1).

DISCUSSIONQuetiapine is a generally well-

tolerated atypical antipsychotic medica-tion with few reports of extrapyramidalsymptoms thought to be due to its rela-tively lowD2 receptor affinity.

2 There hasbeen 1 report of a sleep disorder, sonam-bulism, induced by quetiapine.14 Restlesslegs syndrome must be considered in thedifferential diagnosis of movement disor-der side effects in patients taking quetia-pine. Restless legs syndrome can bedifferentiated from akathisia as the latterdoes not have a circadian component. Inthese 2 patients, RLS was recognized asthe disorder disrupting sleep only after amore thorough sleep history was ob-tained from the patients. As a result, ro-pinirole was prescribed with significantsymptomatic relief of the patients’ RLSsymptoms. Ropinirole, at the low dosesprescribed in these 2 patients, did notcreate any untoward side effects itselfsuch as the potential for increased seda-tion or psychosis. The removal of theRLS symptoms by ropinirole allowedthese 2 complex patients with a historyof multiple medication failures to con-tinue on quetiapine, the medication thatwas most effective in treating their psy-chiatric symptoms.

The clinical significance of the casesreported here is to alert psychiatrists tothe potential development of RLS as aside effect of prescribing quetiapine andthus to take a complete sleep history.Many psychiatric patients such as theones presented here are complex and re-quire multiple medication trials before aneffective treatment regimen can be found.Ropinirole can be effective in treating atroublesome side effect of quetiapine thatallows patients to continue to benefitfrom quetiapine’s therapeutic effects.

ACKNOWLEDGMENTThe author thanks Robert D. Vor-

ona, MD, for his careful proofreading.

AUTHOR DISCLOSUREINFORMATION

There were no sources of support.

Maria R. Urbano, MD*J. Catesby Ware, PhD*y

*Department of Psychiatryand Behavioral Sciences

and yDepartment of Internal MedicineEastern Virginia Medical School

Norfolk, VAurbanomr@evms.edu

REFERENCES

1. Physicians Desk Reference. 2008. Montvale,NJ: Thomas Healthcare Inc; 2007:670.

2. Goldstein JM. Quetiapine fumarate (Sero-quel): a new atypical antipsychotic. DrugsToday. 1999;35:193–210.

3. Timdahl K, Carlsson A, Stening G. Ananalysis of safety and tolerability data fromcontrolled, comparative studies of quetiapinein patients with schizophrenia, focusing onextrapyramidal symptoms. Hum Psychophar-macol. 2007;22:315–325.

4. Cortese L, Caligiuri MP, Williams R, et al.Reduction in neuroleptic-induced movementdisorders after a switch to quetiapine inpatients with schizophrenia. J Clin Psycho-pharmacol. 2008;28:69–73.

5. Ferini-Strambi L. RLS-like symptoms: differ-ential diagnosis by history and clinical assess-ment. Sleep Med. 2007;8:S3–S6.

6. Hening WA, Allen RP. Restless legs syndrome(RLS): the continuing development of diag-nostic standards and severity measures. SleepMed. 2003;4:95–97.

7. Gamaldo CE, Earley CJ. Restless legs syndrome:a clinical update. Chest. 2006;130:1596–1604.

8. Kraus T, Schuld A, Pollmacher T. Periodic legmovements in sleep and restless legs syn-drome probably caused by olanzapine. J ClinPsychopharmacol. 1999;19;478–479.

9. Page RL, Ruscin JM, Bainbridge JL, et al.Restless legs syndrome induced by escitalo-pram: case report and review of the literature.Pharmacotherapy. 2008;28:271–280.

10. Pinninti NR, Rajnish M, Townsend J, et al.Periodic restless legs syndrome associatedwith quetiapine use: a case report. J ClinPsychopharmacol. 2005;25:617–618.

11. Wetter TC, Brunner J, Bronisch T. Restlesslegs syndrome probably induced by risperi-done treatment. Pharmacopsychiatry. 2002;35:109–111.

12. Kikukawa S. Effectiveness of aripiprazole intreatment of adults with attention deficitdisorder and restless legs syndrome. Int JNeuropsychopharmacol. 2008;11:439–440.

13. Bogan RK. Ropinirole treatment for restlesslegs syndrome. Expert Opin Pharmacother.2008;9:611–23.

14. Hafeez ZH, Kalinowski CM. Sonambulisminduced by quetiapine: two case reports and areview of the literature. CNS Spectr. 2007;12:910–912.

Oropharyngeal andFacial Dyskinesia AfterDiscontinuation of

Quetiapine

To the Editors:

Atypical antipsychotics, in particu-lar, clozapine and quetiapine, cause fewerextrapyramidal symptoms (EPSs) includ-ing tardive dyskinesia than conventionalantipsychotics.1 However, the effects ofthe sudden discontinuation of antipsy-chotics on extrapyramidal function areless known.2–4 To our knowledge, therehave been no reports, to date, of quetia-pine causing dyskinesia in response tosudden discontinuation. Here, we presenta case of oropharyngeal and facial dyski-nesia (OFD) after cessation of quetiapine,which was administered off-label as aug-mentation in a female patient with recur-rent depressive episodes.

CASE REPORTA 64-year-old woman with recurrent

depressive episodes was referred to theinpatient ward after attempted suicide byintoxication with 300 mg zolpidem andapproximately 80 g alcohol (7 dL red wineand 1 flask of medicinal spirit, blood alcoholconcentration 1.84 parts per thousand) andinitial treatment in an intensive care unit.Apart from depression, the patient sufferedfrom alcohol dependency with periodicdrinking and showed signs of an accentuatedpersonality with emotional instability anddependent traits. Her psychopharmacol-ogical medication on admission included40 mg citalopram and 20 mg zolpidem. Dueto chronic painful degenerative spondylosisand osteoporosis, she further received a cal-cium/vitamin D3 preparation on a daily basisand 10 mg alendronate sodium once a week.Furthermore, hermedication included 1000mgparacetamol and 50 mg tramadol on demand,as well as 20 mg esomeprazole daily as pro-phylaxis against gastroduodenal ulcers.

The psychiatric treatment consisted ofa combination of pharmacotherapy and cog-nitive behavioral therapy. Significant with-drawal symptoms due to the precedingepisodic alcohol consumption were notobserved. The antidepressant medicationwith citalopram was continued. Due topersistent affective instability with feelingsof anxiety, lack of prospects, ambivalence,and suicidal thoughts, an additive treat-ment with 25 mg quetiapine was initiated1 month after admission and increased step-wise to a daily dose of 175 mg throughouta period of 14 days. This regimen led toan overall improvement in affect and moodand was continued for 6 months. Thenonpsychopharmacological medication re-mained unchanged.

During a weekend holiday, the patientceased taking her psychopharmacological and

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 705

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

other medication and instead ingested anunknown amount of alcohol. She was laterfound by a family member and immediatelyreferred back to the psychiatric hospitalbecause of peculiar behavior. The breathalcohol concentration was 0.34 parts perthousand on readmission. The patientdescribed the sensation as a narrowing ofthe throat with difficulties in swallowing. Weobserved pronounced dyskinesia of themouth, tongue, and facial muscles as wellas dysarthria. A tentative diagnosis of oro-pharyngeal and facial withdrawal dyskinesiadue to the discontinuation of antipsychoticmedication was made, which led us to im-mediately administer 200 mg of quetiapine.In the following 8 hours, there was con-tinuous improvement of the OFD, whichhad completely disappeared after 24 hours.No other medication was given until 1 dayafter readmission to verify whether the dis-continuation of quetiapine was causative ofthe OFD.

DISCUSSIONThis is the first report on OFD as

a probable discontinuation effect afterquetiapine treatment. It is unlikely thatthe discontinuation of citalopram, eso-meprazole, or the pharmacological treat-ment for osteoporosis and degenerativespondylosis caused the extrapyramidalsyndrome, because all signs and symp-toms disappeared after readministrationof quetiapine only. Moreover, the afore-mentioned drugs are not known to affectthe extrapyramidal motor system, whichhas a crucial role in generating motordisturbances. Although quetiapine is be-lieved to have low D2-receptor–bindingaffinity, possible mechanisms leading toEPSs as a discontinuation effect mayinclude hypersensitivity of D2 receptorsin the nigrostriatal system after long-term postsynaptic blockade.5 Moreover,there is recent evidence that vulnerabil-ity to antipsychotic-induced movementdisturbances in patients with affectivedisorders is higher than in those withschizophrenia.6

In conclusion, the current case un-derlines the importance of being awareof the possibility of EPSs, even withquetiapine, after abrupt cessation of an-tipsychotic intake. Thus, besides the ef-forts to achieve good compliance, slowreduction of antipsychotics is recom-mended together with regular monitor-ing of EPSs, particularly of the tongue,as an early indicator.3,7

AUTHOR DISCLOSUREINFORMATION

The authors declare no conflict ofinterest or sources of support.

Tillmann H.C. Kruger, MD*yAndreas Horvath, MDyMirja Schnieber, MDyErich Seifritz, MDyz*Clinic for Psychiatry

Social Psychiatry and PsychotherapyMedical School of Hannover

Hannover, GermanyySanatorium Kilchberg

Private Psychiatric ClinicKilchberg

and zUniversity Hospital and Policlinicof Psychiatry

University of BernBern, Switzerland

krueger.tillmann@mh-hannover.de

REFERENCES1. Timdahl K, Carlsson A, Stening G. An anal-

ysis of safety and tolerability data from con-trolled, comparative studies of quetiapine inpatients with schizophrenia, focusing on extra-pyramidal symptoms. Hum Psychopharmacol.2007;22:315–325.

2. Ehrt U, Fritze F, Aarsland D. Respiratorydyskinesia as discontinuation effect of risper-idone. J Clin Psychopharmacol. 2005;25:609.

3. Skidmore F, Reich SG. Tardive dystonia. CurrTreat Options Neurol. 2005;7:231–236.

4. Tranter R, Healy D. Neuroleptic discontinua-tion syndromes. J Psychopharmacol. 1998;12:401–406.

5. Margolese HC, Chouinard G, Kolivakis TT, et al.Tardive dyskinesia in the era of typical andatypical antipsychotics. Part 1: pathophysiologyand mechanisms of induction. Can J Psychiatry.2005;50:541–547.

6. Gao K, Kemp DE, Ganocy SJ, et al.Antipsychotic-induced extrapyramidal sideeffects in bipolar disorder and schizophrenia:a systematic review. J Clin Psychopharmacol.2008;28:203–209.

7. Hennings JM, Krause E, Botzel K, et al. Suc-cessful treatment of tardive lingual dystonia withbotulinum toxin: case report and review of theliterature. Prog Neuropsychopharmacol BiolPsychiatry. 2007;32:1167–1171.

Leukopenia andNeutropenia AfterIntoxication WithDiphenhydramine(Nytol) During

Clozapine Treatment

To the Editors:Clozapine is an atypical antipsy-

chotic used for the treatment of schizo-phrenia and is known to be significantlysuperior to other antipsychotics in re-fractory schizophrenia.1 However, 0.8%to 2.9% of the patients taking clozapinehave leukopenia and neutropenia, andthis side effect may prevent continuingtreatment with clozapine.2 Some authorshave reported that polypharmacy mayincrease the risk of neutropenia andleukopenia. We wish to report a case inwhich leukopenia and neutropeniaoccurred in a patient treated with cloza-pine a few days after intoxication withdiphenhydramine.

CASEMr P.T., a 45-year-old white man, was

diagnosed with paranoid schizophrenia sincethe age of 26 years. He presented a multipledrug addiction between the age of 13 and30 years. His medical history includes anappendectomy and amygdalectomy. Havingbeen refractory to multiple antipsychotics,the patient was offered clozapine to controlhis psychotic symptoms. Thereafter, heshowed a marked improvement in his behav-ior, an increased concentration capability,and noticeable reduction of psychotic symp-toms. His hematologic analysis showednormal white blood cell (WBC) and neutro-phil counts.

Eighteen months after clozapine treat-ment was started, the patient presented asudden leukopenia (3.0 � 109/L) and neu-tropenia (0.96 � 109/L). Eventually, thepatient admitted taking, 3 days earlier, 16tablets of Nytol (diphenhydramine) 25 mgfor insomnia, an antihistamine availablewithout a prescription. Other blood cellsand platelets counts were normal. The patientdid not present infectious symptom this daynor in the previous days. Two weeks earlier,the WBC and neutrophil counts were 6.90 �109/L and 5.39 � 109/L, respectively. At thetime of the leukopenia, Mr P.T. was takingclozapine 400 mg/d, risperidone 4 mg/d,gabapentin 1200 mg/d, and clonazepam4 mg/d, and this medication had been admin-istered at the same dosage for the previous14 months.

Clozapine was discontinued the sameday the Bred alert[ occurred. The hemato-logic report of the following day confirmedthe results of the previous day, showing aleukopenia of 3.30 � 109/L and a neutrope-nia of 1.19 � 109/L. Gradually, his WBCcount increased, reaching normal values 4days later. In the days after clozapinediscontinuation, Mr P.T. was more agitated,

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins706

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

disorganized, psychotic, and anxious. Twoweeks later, clozapine, in combination withrisperidone, clonazepam, and divalproex, wassuccessfully reintroduced, and the patienthas not presented neutropenia nor leukopeniafor the past 13 months.

DISCUSSIONThis report illustrates the case of

a patient who had taken clozapine for18 months without a drop in his WBCcount but presented a leukopenia andneutropenia after diphenhydramine(Nytol) intoxication. Most but not allcases of clozapine-induced neutropeniaand leukopenia occur in the first 18weeks of treatment, emphasizing theunexpected appearance of low WBCcount after 18 months of treatment withclozapine.3 A few cases of neutropeniahave been reported a few years afterinitiating clozapine treatment.4,5 How-ever, in our patient, the short periodbetween the ingestion of diphenhydra-mine and WBC count drop suggests thatthis intoxication along with clozapineand other psychotropic drugs has pro-voked leukopenia and neutropenia. Fur-thermore, WBC reached normal values7 days after the intoxication, and cloza-pine was successfully reintroduced with-out causing any modification in theWBC count for 13 months. In mostcases of clozapine-induced neutropenia,a second trial of clozapine may causeneutropenia with a shorter latencyperiod than the first episode.6 Thesuccessful rechallenge demonstrated thatdiphenhydramine contributed to thehematologic disorder. This being said,we cannot completely refute the possi-bility that the drop in WBC was causedby clozapine alone and coincided withthe intoxication with diphenhydramine.

No WBC disorder has yet beendocumented with the use of diphenhy-dramine,7 although our case reportsuggests that this can happen whendiphenhydramine is added to clozapine.Similarly, leukopenia or neutropeniahave been observed after the recentaddition of risperidone,2,8 sodium val-proate,9 paroxetine,10 and erythromycin11

to clozapine.Therefore, in the case of patients

with refractory schizophrenia presentinga neutropenia or a leukopenia whileunder treatment with clozapine, clini-cians should consider stopping a medi-

cation recently administered that mayhave contributed or precipitated neutro-penia instead of interrupting clozapinetreatment. Also, our case report indi-cates that the patient should be asked ifhe took a nonprescription medication. Itis important for physicians to considerthe effect of polypharmacy and thecontext in which neutropenia and leu-kopenia occurred to evaluate if there is apossibility of clozapine rechallenge. Asuccessful rechallenge would never-theless necessitate careful hematologicmonitoring.

Karine Charbonneau, BSc*Pierre Landry, MD, PhD, FRCPCy

*Sciences BiomedicalesFaculte de Medicine

Universite de Montrealand yHopital Louis-H. LafontaineModule de Psychopharmacologie

Montreal, Quebec, Canadaplandry@ssss.gouv.qc.ca

REFERENCES1. Kane J, Honigfeld G, Singer J, et al. Clozapine

for the treatment-resistant schizophrenic: adouble-blind comparison with chlorproma-zine. Arch Gen Psychiatry. 1988;45:789–796.

2. Senechal A, Landry P, Deschamps R, et al.Neutropenia in a patient treated with clozapinein combination with other psychotropic drugs.Encephale. 2002;28:567–569.

3. Munro J, O’Sullivan D, Andrews C, et al. Ac-tively monitoring of 12760 clozapine recipientsin the UK and Ireland: beyond pharmacovi-gilance. Br J Psychiatry. 1999;175:576–580.

4. Bhanji NH, Margolese HC, Chouinard G,et al. Late-onset agranulocytosis in a patientwith schizophrenia after 17 months of clo-zapine treatment. J Clin Psychopharmacol.2003;23:522–523.

5. Small JG, Weber MC, Klapper MH, et al. Re-challenge of late-onset neutropenia with cloza-pine. J Clin Psychopharmacol. 2005;25:185–186.

6. Dunk LR, Annan LJ, Andrews CD. Rechal-lenge with clozapine following leucopenia orneutropenia during previous therapy. Br JPsychiatry. 2006;188:255–263.

7. Andersohn F, Konzen C, Garbe E. Systematicreview: agranulocytosis induced by nonche-motherapy drugs. Ann Intern Med. 2007;146:657–665.

8. Godleski LS, Sernyak MJ. Agranulocytosisafter addition of risperidone to clozapine treat-ment. Am J Psychiatry. 1996;153:735–736.

9. Pantelis C, Adesanya A. Increased risk ofneutropenia and agranulocytosis with sodiumvalproate used adjunctively with clozapine.Aust NZ Psychiatry. 2001;35:544–545.

10. George TP, Innamorato L, Sernak M, et al.Leukopenia associated with addition of parox-etine to clozapine. J Clin Psychiatry. 1998;59:31.

11. Usiskin SI, Nicolson R, Lenane M, et al. Re-treatment with clozapine after erythromycin-

induced neutropenia. Am J Psychiatry. 2000;157:1021.

Atypical Antipsychoticsand MetabolicOutcomes in

Chinese Patients

A Comparison ofOlanzapine and

Risperidone

To the Editors:There is increasing evidence to

suggest that atypical antipsychotics areassociated with different metabolic com-plications such as weight gain, hyper-lipidemia, and diabetes mellitus.1,2

Studies in western samples suggesteda more deteriorated metabolic risk fac-tor profile in patients treated with olan-zapine than risperidone.3,4 Atypicalantipsychotics associated with meta-bolic side effects in the Chinese havenot been under thorough study. Thiscross-sectional comparison study wasdesigned to determine the metabolicoutcomes in Chinese schizophreniapatients who are treated with olanzapineor risperidone.

Patients, aged 18 to 65 years, re-ceiving olanzapine or risperidone inthe outpatient department of a universityteaching hospital in Hong Kong wereinvited to participate in the study fromApril to August 2005. Patients with sig-nificant physical diseases or receivinganticonvulsants, antidepressants, lithium,or antipsychotics other than olanzapineor risperidone were excluded. A trainedresearch assistant measured the bodyweight, body height, blood pressure, andheart rate when they attended follow-upduring the study period. Subjects werefasted after midnight, and their plasmalipid and glucose levels were checkedby standard laboratory methods on thesame day. Information on smoking, drink-ing habits, and family history of diabe-tes mellitus was collected. Pretreatmentweight, clinical data, and demographicdata were retrieved from the case records.Subjects were categorized into groupsaccording to the number of the Natio-nal Cholesterol Education Program

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 707

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Adult Treatment Panel III definition formetabolic syndrome.5 Ordinal regres-sion analysis was performed for clinicaland demographic variables in predictingmetabolic syndrome using the StatisticalPackage for the Social Sciences, version14.0 (SPSS Inc, Chicago, Ill).

Seventy-five subjects (45 olanzapine-treated and 30 risperidone-treated) withDiagnostic and Statistical Manual ofMental Disorders diagnosis of schizo-phrenia consented and completed thestudy. Their demographic and clinicalcharacteristics are listed in Table 1. Ac-cording to National Cholesterol Educa-tion Program Adult Treatment Panel IIIdefinition of metabolic syndrome, 11subjects (14.7%) were found to havemetabolic syndrome, and 47 subjects(62.7%) fulfilled 1 or more criteria. Therewere 12 (50%) and 3 (12.5%) subjectstreated with olanzapine and risperidone,respectively, found to have dyslipidemia.One olanzapine-treated subject (4.2%)and 2 risperidone-treated subjects (8.3%)were found to have impaired fastingglucose. As listed in Table 2, female sex,mean dosage of antipsychotics, meanblood pressure, total cholesterol level,high-density lipoprotein (HDL) choles-terol level, triglycerides level, and post-treatment body weight were statisticallysignificant factors in predicting meta-

bolic syndrome. Other variables were sta-tistically not significant (P > 0.05), withage and smoking habit having P = 0.051and 0.056, respectively.

DISCUSSIONMetabolic syndrome is an increas-

ing and important problem in Asia,but local findings among psychiatricpatients are limited.6 This study dem-onstrated the commonness of metabolicsyndrome and dyslipidemia among Chi-nese patients with schizophrenia. Thehigh proportion of the metabolic prob-lem untreated in Chinese patients withschizophrenia raises the need for in-creased attention to monitoring of phys-ical condition.7 We also identified sexand mean dosage of antipsychoticsas risk factors, in addition to the corefeatures, of metabolic syndrome. This isconsistent with prevalence study in Asiashowing differential effect of sex inmetabolic syndrome. The positive asso-ciation of mean dosage of antipsycho-tics with metabolic problem may beexplained by the interaction between en-vironmental and genetic factors.6

In contrast to the literature,8 thepresent study showed no statisticallysignificant difference between olanza-pine and risperidone in associationwith metabolic syndrome. This may beexplained by the heterogeneity of sub-jects between the 2 groups and small

TABLE 1. Demographic and Clinical Characteristics

All Olanzapine Risperidone

Total no. subjects 75 45 30

No. female 40 (53.3%) 19 (42.2%) 21 (70.0%)

Age, yrs 34.8 T 11.1 34.1 T 10.8 35.9 T 11.7

Smoking habit, no (%) 18 (24.0%) 13 (29%) 5 (16.7%)

Drinking habit, no (%) 8 (10.7%) 7 (15.6%) 1 (3.3%)

Family history of diabetes, no (%) 24 (32.0%) 14 (31.1%) 10 (33.3%)

Mean dosage, mg/d 12.2 T 7.6 4.9 T 3.2

Duration of treatment, mo 37.0 T 21.5 39.9 T 20.7 32.5 T 22.3

Pretreatment body weight, kg 59.2 T 9.5 57.6 T 9.0 61.4 T 10.4

Posttreatment body weight, kg 64.6 T 10.3 64.6 T 10.5 64.6 T 10.1

Weight change, kg 5.4 T 6.5 7.0 T 6.3 3.1 T 6.2

Total cholesterol, mmol/L 5.2 T 1.2 5.5 T 1.3 4.7 T 0.9

HDL cholesterol, mmol/L 1.4 T 0.4 1.4 T 0.4 1.4 T 0.4

Triglycerides, mmol/L 1.55 T 0.96 1.76 T 0.68 1.22 T 0.95

Glucose, mmol/L 5.2 T 0.6 5.2 T 0.6 5.2 T 0.7

Mean blood pressure, mm Hg 87.8 T 12.0 90.8 T 11.0 83.2 T 12.2

Body height, m 1.61 T 0.08 1.62 T 0.09 1.60 T 0.07

Waist circumference, cm 87.5 T 10.2 87.8 T 10.2 87.0 T 10.4

Hip circumference, cm 99.2 T 6.8 98.0 T 7.1 101.1 T 5.9

TABLE 2. Ordinal Regression Analysis of Clinical and Demographic Variables inPredicting Metabolic Syndrome

Variables Estimate

95% Confidence Interval

PLower Bound Upper Bound

Female sex* 2.914 1.373 4.454 0.000

Age 0.051 0.000 0.103 0.051

Nonsmoker j1.233 j2.497 0.031 0.056

Nondrinker j0.345 j1.970 1.280 0.678

No family history of diabetes 0.596 j0.233 1.426 0.159

Olanzapine j0.099 j1.119 0.922 0.850

Mean dosagey 0.064 0.015 0.113 0.010

Duration of treatment 0.014 j0.008 0.035 0.209

Total Cholesterol* 0.547 0.161 0.934 0.006

HDL Cholesterol* j2.515 j4.019 j1.011 0.001

Triglycerides* 0.800 0.247 1.352 0.005

Glucose 0.289 j0.457 1.035 0.448

Mean blood pressure* 0.096 0.049 0.143 0.000

Body height j2.810 j12.675 7.056 0.557

Posttreatment body weighty 0.137 0.013 0.261 0.030

Weight change 0.001 j0.072 0.074 0.983

Waist circumference 0.013 j0.081 0.107 0.784

Hip circumference 0.062 j0.171 0.046 0.261

*P < 0.01.yP < 0.05.

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins708

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

sample size. The higher mean total cho-lesterol, triglycerides, and weight gainin olanzapine require special attentionin physical monitoring of patients withschizophrenia.

There are several limitations of thestudy requiring attention. First, this is across-sectional study, and the pretreat-ment proportion of metabolic syndromeis not known. Second, the small sam-ple size and specific subgroup of psy-chiatric patients need precaution beforegeneralization.

This study represents prelimi-nary efforts to find out the relationbetween metabolic outcomes and atyp-ical antipsychotics in Chinese psychi-atric patients. The findings providevaluable information on the extent ofantipsychotics-induced metabolic ef-fects in our locality. This could increaseawareness about the importance ofphysical health monitoring in Chinesepsychiatric patients.

AUTHOR DISCLOSUREINFORMATION

The authors declare no conflict ofinterest.

Edwin Lee, MSc, MBChB, MRCPsychChi-Ming Leung, MBBS, MRCPsych

Department of PsychiatryThe Chinese University of Hong Kong

Shatin, Hong KongPeople’s Republic of China

elee@cuhk.edu.hk

REFERENCES1. Henderson DC. Schizophrenia and comorbid

metabolic disorders. J Clin Psychiatry. 2005;66(suppl 6):11–20.

2. Melkersson K, Dahl ML. Adverse metaboliceffects associated with atypical antipsychotics:literature review and clinical implications.Drugs. 2004;64:701–723.

3. Koro CE, Fedder DO, L’Italien GJ, et al. Anassessment of the independent effects of olan-zapine and risperidone exposure on the risk ofhyperlipidemia in schizophrenic patients. ArchGen Psychiatry. 2002;59:1021–1026.

4. Meyer JM. A retrospective comparison ofweight, lipid, and glucose changes betweenrisperidone- and olanzapine-treated inpatients:metabolic outcomes after 1 year. J Clin Psychia-try. 2002;63:425–433.

5. National Institutes of Health. 2002 NationalInstitutes of Health, Third Report of the Na-tional Cholesterol Education Program (NCEP)expert panel on detection, evaluation, and treat-ment of high blood cholesterol in adults (AdultTreatment Panel III) final report. Circulation.2002;106:3143–3421.

6. Chan NN, Kong AP, Chan JC. Metabolic syn-drome and type 2 diabetes: the Hong Kong per-spective. Clin Biochem Rev. 2005;26(3):51–57.

7. Nasrallah HA, Meyer JM, Goff DC, et al. Lowrates of treatment for hypertension, dyslipidae-mia and diabetes in schizophrenia: data fromthe CATIE schizophrenia trial sample at base-line. Schizophr Res. 2006;86(1–3):15–22.

8. Newcomer JW. Antipsychotic medications:metabolic and cardiovascular risk. J ClinPsychiatry. 2007;68(suppl 4):8–13.

Long-ActingRisperidone–InducedPeriorbital Edema

To the Editors:Risperidone is a benzisoxazole an-

tipsychotic agent with a therapeuticactivity mediated through a combinationof D2 and 5-HT2 receptor antagonism.1

To my knowledge, there have been 9cases2–10 of edema associated with ris-peridone. The first 6 cases involved oralrisperidone combined with other psy-chotropic medications. The last 3 casesreported edema secondary to oral ris-peridone alone. Here, a case of perior-bital edema occurring during treatmentwith risperidone long-acting injectionis reported.

CASE REPORTMs V. was a 39-year-old woman with a

5-year history of paranoid schizophrenia andpoor medication compliance. She was ad-mitted with persecutory and referential de-lusions and auditory hallucinations. She hadunilaterally discontinued her antipsychotictherapy 1 month before hospitalization.

In the last 7 months, her psychopa-thological condition was stable and welltolerated with 4 mg of oral risperidone alone.No relevant adverse side effects were notedduring risperidone treatment. Her medicalhistory was also unremarkable; she had ne-ver shown signs or symptoms of conges-tive heart failure, hepatic disease, nephrosis,hyponatremia, or vascular occlusive disorder.

At the admission, Ms V. was readminis-tered oral risperidone 4 mg at bedtime. Onday 3, 50 mg risperidone long-acting injec-tion was administered because of her poormedication compliance when she is athome. After some days, the patient graduallyimproved, agitation and psychotic symp-toms decreased, and no relevant adversereactions were emphasized. Nevertheless,on day 14, she developed a marked perior-bital edema that gave the face a bloated ap-pearance, and the eyes became partially

closed by the swollen eyelids. All laboratoryfindings (including immunoglobulin [Ig] M,IgG, IgE, C3 and C4, thyroid-stimulatinghormone, FT3 and FT4, creatinine, urea,electrolytes, and liver enzyme serum levels)were within the normal limits. Her oralrisperidone was then discontinued, but herperiorbital edema still persisted for another4 weeks after the intramuscular injection.When the edema was completely resolved,she was subsequently administered olanza-pine and tolerated its titration well.

DISCUSSIONMs V. was not on any other medi-

cation, and as such, a drug interactioncould be ruled out. Given her normal IgE,C3, and C4 levels, the edemawas unlikelyto have been secondary to an allergicreaction. Given her normal thyroid andliver and kidney function values, hepaticdisease, hypothyroidism, and nephroticsyndrome were excluded. Because shealready had taken oral risperidone for7 months without showing any relevantadverse reactions, risperidone long-actinginjection was thought to be the culprit ofher periorbital edema.

However, the etiology of the patient’speriorbital edema remains unclear, as wellas the possible pathogenetic role of thenontherapeutic components of risperi-done long-acting injection, a combinationof polylactide-co-glycolide extended-release microspheres and a diluent forparental use (which includes polysorbate20, sodium carboxymethyl cellulose,disodium hydrogen phosphate dihydrate,citric acid anhydrous, sodium chloride,and sodium hydroxide). Perhaps, as ourexperience with risperidone long-actinginjection expands, the relevance of edemaas a side effect will be further elucidated.

AUTHOR DISCLOSUREINFORMATION

The author declares no conflicts ofinterest or sources of support.

Lorenzo Pelizza, MD, PhDReggio Emilia Mental Health Department

Guastalla Psychiatric Operative UnitGuastalla, Italy

anolino@yahoo.it

REFERENCES1. Stahl S. Essential Psychopharmacology. Cam-

bridge: Cambridge University Press; 1996.2. Cooney C, Nagy A. Angio-edema associated

with risperidone. BMJ. 1995;311:1204.

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 709

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

3. Baldassano C, Ghaemi SN. Generalized edemawith risperidone, divalproex sodium treatment.J Clin Psychiatry. 1996;57:422.

4. Terao T, Kojima E, Eto A. Risperidone and aller-gic reactions. J Clin Psychiatry. 1998;59:82–83.

5. Sanders R, Lehrer D. Edema associated withaddition of risperidone to valproate treatment.J Clin Psychiatry. 1998;59:689–690.

6. Shuster J. Risperidone, edema, and more.Nursing. 1999;29:75.

7. Masson M, Elayli R, Verdoux H. Risperidoneand edema, apropos of a case. Encephale.2000;26:91–92.

8. Ravasia S. Risperidone-induced edema. CanJ Psychiatry. 2001;46:453–454.

9. Kores B, Vitorovic S, Zalar B. Three chal-lenges and a rechallenge episode of angio-edema occurring in treatment with risperidone.Eur Psychiatry. 2001;16:506–507.

10. Feroz C, Selvaraj P, Roy M. Risperidone-induced edema in a child with learning dis-ability and autism. Autism. 2006;10:308–310.

Antidepressant-InducedSweating Alleviatedby Aripiprazole

To the Editors:As a major cause of medication non-

compliance, excessive sweating, or hy-perhidrosis, is reported in 5% to 20% ofindividuals receiving antidepressant ther-apy.1 Because switching antidepressantsoften does not resolve sweating, anticho-linergic medications are commonly givento reduce sweating, despite their numerousside effects. As one of the newest atypicalantipsychotics, aripiprazole is increasinglyprescribed together with antidepressants,and there have been no reports whether itmodulates antidepressant-related sweat-ing. In this letter, we present 2 cases ofantidepressant-induced sweating alleviatedby aripiprazole. We also discuss possiblemechanisms to account for this phenom-enon, in the context of the unique mo-lecular property of aripiprazole.

CASE REPORTS

Case 1Ms A is a 57-year-old white woman

with a 14-year history of bipolar II disorder.Having discontinued buproprion and sertra-line because of sweating, she began a trial offluoxetine (60 mg daily). At this dose, shenoted significant resolution of her depressivesymptoms but again complained of dose-related sweating, which she described asconstant and generalized, leading to embar-rassment at work. Her fluoxetine was tapered

to 20 mg, resulting in less sweating but areturn of her depressive symptoms. At thistime, she was prescribed aripiprazole (10 mgdaily) to provide mood stabilization and totreat her excessive ruminations. During hernext visit 6 months later, she noted completeresolution of her sweating and agreed to in-crease her fluoxetine to 40 mg per day. Thedose-related sweating due to fluoxetinereturned, but the degree was much milderand tolerable, no longer causing distress atwork. Ms A is postmenopausal and has ahistory of hypothyroidism, for which she hasbeen receiving replacement therapy from herprimary care provider before her fluoxetinetreatment. She denied sweating changes re-lated to her thyroid treatment and had awithin-normal level of thyroid-stimulatinghormone (0.868 2IU/mL; reference range,0.3–3.0 2IU/mL) at the time of her aripipra-zole treatment. She did not exhibit otherendocrine symptoms, and a review of hermedical history and other laboratory studieswas unremarkable.

Case 2Ms B is a 59-year-old white woman with

a 13-year history of major depressive disorderwith psychotic features. She developed sweat-ing shortly after starting duloxetine (60 mgdaily) 1 year ago. The sweating was general-ized and constant, requiring her to change allof her clothing several times a day. A fewmonths after starting duloxetine, ziprasidone(60 mg daily) was added to treat occasionalvisual hallucinations. Although ziprasidoneresolved her hallucinations, it had little effecton the severity of her sweating. Because wewere able to lessen sweating with aripiprazolein Ms A, we discontinued Ms B’s ziprasidoneand placed her on 10 mg of aripiprazole perday. Two weeks later, Ms B reported signifi-cant decrease of her sweating, having tochange her clothes only once or twice a day.Her aripiprazole was then increased to 20 mgdaily, at which she noted complete resolutionof her sweating complaints. The combinationof duloxetine and aripiprazole had also kepther depression and psychosis in remission.Ms B is postmenopausal, with no knownmedical comorbidity or abnormal routinelaboratory workup. Previous visits to herother care providers, including an endocrinol-ogist and a gynecologist, yielded no apparentcause to her sweating, apart from her anti-depressant use.

DISCUSSIONSweating is mediated by the sympa-

thetic system and under the control of athermoregulatory center in the hypothal-amus. Hypothalamic thermoregulation isbelieved to require dopaminergic activ-

ities,2,3 which can be inhibited by sero-tonergic surges.4 Reduced dopaminergicactivities in the central nervous system,such as in Parkinson disease, frequentlyresult in increased sweating. Althoughnot yet fully understood, antidepressant-induced sweating is attributed to eitheraltered serotonergic or dopaminergic tonesin the hypothalamus or a direct adrenergicstimulation of the sympathetic system, in-cluding the cholinergic-innervated sweatglands.1 In support of these explanations,serotonin antagonist such as cyprohepta-dine, antiadrenergic agents such as cloni-dine, and anticholinergic agents such asbenztropine have been found effective intreating antidepressant-related sweating.1

There are also a few case reports of lesssweating with certain antidepressants; boththe anticholinergic property of paroxetine5

and the antiserotonergic effect of mirtaze-pine6 had been proposed as potential an-tisweating mechanisms.

There exist several possibilities toaccount for aripiprazole minimizingantidepressant-induced sweating. Aripi-prazole has weak affinity for the adre-nergic !1 receptor, through which, itmay counter the sympathetic system.The strong serotonin 2A antagonismcould block the serotonergic effect ofantidepressants on the hypothalamus,indirectly restoring the dopaminergictone and thermoregulation. As a partialdopamine agonist,7 aripiprazole maycontribute to temperature homeostasisby directly stabilizing the dopamine-sensitive thermoregulatory center. Todetermine which of these mechanismsaccounted for the 2 cases above, we notethat ziprasidone has similar if notstronger affinity for the !1 and theserotonin 2A receptors than aripipra-zole,7,8 but the use of ziprasidone inMs B did not relieve sweating. Thus, itis likely that decreased sweating islinked to the unique property of aripi-prazole as a partial dopamine agonist.

As a partial dopamine agonist,aripiprazole may confer a homeostaticeffect on the dopamine-dependent func-tions of the hypothalamus, such as ther-moregulation and control of prolactinsecretion. For example, among the fewreported cases of neuroleptic malignantsyndrome–like symptoms due to aripi-prazole,9,10 no fever had been observed,and this may in part be attributed tothe putative thermoregulatory effect of

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins710

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

aripiprazole. In addition, whereas an-tipsychotics and selective serotoninreuptake inhibitors are known to in-crease prolactin secretion via inhibitingdopaminergic activity in the hypothal-amus, aripiprazole is not associated withhyperprolactinemia11 and can even nor-malize antipsychotic-induced prolactinelevations.12,13 In summary, we reportthat aripiprazole, especially when clin-ically warranted as an augmentationagent,14 can be effective in reducingantidepressant-related sweating. This notonly improves treatment compliance butalso lessens the need for traditional anti-sweating interventions.

ACKNOWLEDGMENTDr Lu is supported in part by a

Young Investigator Award from NARSAD.

Brett Y. Lu, MD, PhD*Chandra E. Cullen, MDy

Carla E. Eide, MDyCynthia C. Williams, MDy

William J. Apfeldorf, MD, PhDy*Department of Psychiatry

University of HawaiiHonolulu, HI

and yDepartment of PsychiatryUniversity of New Mexico

Albuquerque, NMlub@dop.hawaii.edu

REFERENCES1. Marcy TR, Britton ML. Antidepressant-

induced sweating. Ann Pharmacother. 2005;39:748–752.

2. Gurrera RJ. Sympathoadrenal hyperactivityand the etiology of neuroleptic malignant syn-drome. Am J Psychiatry. 1999;156:169–180.

3. Lee TF, Mora F, Myers RD. Dopamine andthermoregulation: an evaluation with specialreference to dopaminergic pathways. NeurosciBiobehav Rev. 1985;9:589–598.

4. Damsa C, Bumb A, Bianchi-Demicheli FB,et al. BDopamine-dependent[ side effectsof selective serotonin reuptake inhibitors: aclinical review. J Clin Psychiatry. 2004;65:1064–1068.

5. Praharaj SK, Arora M. Paroxetine useful forpalmar-plantar hyperhidrosis. Ann Pharma-cother. 2006;40:1884–1886.

6. Buecking A, Vandeleur CL, Khazaal Y, et al.Mirtazapine in drug-induced excessive sweat-ing. Eur J Clin Pharmacol. 2005;61:543–544.

7. Goodnick PJ, Jerry JM. Aripiprazole: profileon efficacy and safety. Expert Opin Pharma-cother. 2002;3:1773–1781.

8. Stahl SM, Shayegan DK. The psychophar-macology of ziprasidone: receptor-bindingproperties and real-world psychiatric practice. JClin Psychiatry. 2003;64(suppl 19):6–12.

9. Chakraborty N, Johnston T. Aripiprazole andneuroleptic malignant syndrome. Int ClinPsychopharmacol. 2004;19:351–353.

10. Rodrigues OP, Dowell MS. A case report ofneuroleptic malignant syndrome without feverin a patient given aripiprazole. J Okla StateMed Assoc. 2006;99:435–438.

11. Bowles TM, Levin GM. Aripiprazole: a newatypical antipsychotic drug. Ann Pharma-cother. 2003;37:687–694.

12. Lorenz RA, Weinstein B. Resolution ofhaloperidol-induced hyperprolactinemia witharipiprazole. J Clin Psychopharmacol. 2007;27:524–525.

13. Shim JC, Shin JG, Kelly DL, et al. Adjunctivetreatment with a dopamine partial agonist,aripiprazole, for antipsychotic-induced hyper-prolactinemia: a placebo-controlled trial. Am JPsychiatry. 2007;164:1404–1410.

14. Philip NS, Carpenter LL, Tyrka AR, et al.Augmentation of antidepressants with atypicalantipsychotics: a review of the current liter-ature. J Psychiatr Pract. 2008;14:34–44.

Ziprasidone-AssociatedMania in Korean

Schizophrenic Patient

To the Editors:Ziprasidone, one of a series of atyp-

ical antipsychotics, was released in Koreain 2005 as an alternative treatment forschizophrenia, schizoaffective disorder,and bipolar disorder and has been ob-served being accompanied by a uniqueside-effect profile.1 Like other atypicalantipsychotics, ziprasidone has a highaffinity for blocking 5-HT2 and D2 re-ceptors. However, it also shows variousinteractions with other 5-HT receptorsand inhibits serotonergic and noradren-ergic reuptake sites.1 It is possible thatthese additional pharmacological charac-teristics of ziprasidone may be associatedwith mania.2 We report what we believeto be the first case of mania associatedwith the use of ziprasidone in a schizo-phrenic patient in Asia.

CASE REPORTMr K., a 28-year-old male Korean with a

2-year history of schizophrenia, was admittedto our hospital in 2005. The presenting symp-toms were persecutory delusion, auditoryhallucinations, and negative symptoms in-cluding social withdrawal and anhedonia. Hewas initially treated with olanzapine. After2 months of hospital treatment, acute symp-toms had largely remitted. Therefore, he wasdischarged to a day-clinic treatment. After2 years of relative stability, he discontinuedthe treatment with 10 mg of olanzapine in

September 2007. Three months later, hepresented with disorganized speech andbehavior, social withdrawal, and auditoryhallucination, but no mood symptoms. Hewas readmitted for exacerbation of symp-toms. At this time, he had a Positive andNegative Syndrome Scale score of 109. Hiscomplete blood count, serum metabolicpanel, thyroid panel, and urinalysis all hadbeen within the reference range, and noabnormalities were seen in his electro-cardiogram and brain magnetic resonanceimaging. Because of his weight gain by takingolanzapine, we began to administer ziprasi-done, which was titrated to 160 mg/d by day 4of hospitalization. He also received 0.5 mg ofclonazepam daily. On day 18, his Positive andNegative Syndrome Scale score had de-creased to 78. He had no side effects ofziprasidone. However, on day 19, Mr K. be-gan to show features suggesting mania,including elated mood, talkativeness, grandi-osity, and physical aggression. By day 22, hehad a Young Mania Rating Scale (YMRS)score of 32. Ziprasidone treatment was thendiscontinued; lithium and quetiapine wereinitiated. By day 29, at 400 mg/d of que-tiapine and 1200 mg of lithium, his manicsymptoms improved substantially. He had aYMRS score of 16. On day 36, his YMRSscore had decreased to zero. He did not reportany other manic symptoms. On day 39, hewasdischarged to a day clinic. Afterward, hewas maintained on 300 mg of quetiapine and1200 mg of lithium daily. The follow-up after2 months revealed no further symptomsof mania.

DISCUSSIONThis might be another case to sup-

port the hypothesis that daily use ofziprasidone is related to an onset of ma-nia. The onset of mania in other casereports3–7 of schizophrenic or schizoaf-fective patients was earlier than in ourcase, ranging from 7 to 10 days. Our casecould be considered an outlier in thisview, because it took 18 days until theonset, like other 2 case reports,8 whereasother cases showed smaller numbers.However, it still yielded a rapid changein the symptom picture. Mr K. hadreceived only ziprasidone and clonaze-pam. In addition, a review of Mr K.’sprevious records also revealed no manicepisodes. However, the limitations of asingle case report must be acknowledged.

In addition to ziprasidone, many atyp-ical antipsychotics seem to be asso-ciated with mania/hypomania, becausethere are, at present, considerable numberof case reports. Atypical antipsychotics

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 711

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

have common pharmacological feature ofcombined 5-HT2A and D2 antagonism.Especially the relatively higher affinityfor 5-HT2A than D2 receptors is charac-teristic for atypical antipsychotics.9

It has been taken into considerationthat risperidone-induced mania/hypo-mania is mainly related to dosage. Lowdoses possibly result in blockade of5-HT2A but not D2 receptors, causingdisinhibition of frontal dopamine releaseand induction of manic symptoms.10 Athigh doses, however, the saturation of5-HT2A receptors and the increasing D2

occupancy counteract this action of the5-HT2A activity.11 It seems that olanza-pine is similar to risperidone with dosesin its mechanism.12 However, mania/hy-pomania was induced by moderate tohigh daily doses of atypical antipsy-chotics in many case reports.13 Thus, itmay suggest that the enhancement ofthe frontal dopaminergic transmissionand the subsequent mood alteration donot necessarily associate with a dose-dependent function of the 5-HT2A/D2

occupancy ratio. Another view has beensuggested that the combined 5-HT2A

and D2 receptors blockade of atypicalantipsychotics enhances the ability of5-HT1A receptors to cause frontal do-pamine release, regardless of intrinsic5-HT1A affinity.14

Regarding quetiapine, it has thelowest 5-HT2A affinity and 5-HT2A/D2

occupancy ratio among serotonin do-pamine antagonist atypical antipsy-chotics.15,16 It has a relatively low D2

antagonism as well. However, it pos-sesses a higher affinity for 5-HT2A

than D2 receptors at low doses.17 It isalso a partial agonist for 5-HT1A recep-tors.18 In addition, it has been taken intoconsideration that the low level of D2

antagonism may reflect the capacity ofthese compounds to bind loosely to theD2 receptor and to dissociate fast fromit, causing attenuated physiological dop-amine transmission to continue, despitethe occupancy of D2 receptors.19 Thus,these seem to be involved in theinduction of mania/hypomania.

Aripiprazole can obviously both beefficacious in the treatment of maniaand induce mania because of its partialD2 agonism, depending on the D2 recep-tors’ state and occupancy.20,21 For ex-ample, in 1 case of aripiprazole-inducedmania,20 it may have displayed D2

agonistic rather than antagonistic effectsbecause of chronic schizophrenia with apresumably frontal hypodopaminergicstate. Although it has only a low affinityfor the 5-HT2A receptor,22 this affinitymay still be of importance. It is also apartial agonist of 5-HT1A receptors.Thus, these also seem to be involved inthe induction of mania/hypomania.

Amisulpride—a nonserotonin dopa-mine antagonist atypical antipsychotic—has presynaptic D2 and D3 autoreceptor-blocking activities with low doses. Thus,it enhances dopamine transmission in thefrontal cortex, resulting in potential maniainduction despite the lack of 5-HT2Areceptor affinity.23 However, the patientreported in Australia was not on a low-dose, daily amisulpride (400 mg/d).23

In accordance to the above, it seemsthat many atypical antipsychotics areinvolved in the induction of mania/hy-pomania in susceptible individuals viavarious mechanisms. Especially, ziprasi-done has another mechanism that is thepartial agonist of 5-HT1A receptors andinhibits the neuronal uptake of serotoninand norepinephrine comparable to imip-ramine, a feature that is unique amongatypical antipsychotics.24

In conclusion, although ziprasidonehas been shown to be effective in treat-ing mania,25 it may possibly be associ-ated with the induction of mania.Although the clear cause of mania inthis case may remain unknown, a clini-cian should carefully and vigilantlymonitor for the potential induction ofhypomania or mania associated withziprasidone treatment.

Young-Min Park, MD, PhD*Seung-Hwan Lee, MD, PhD*Heon-Jeong Lee, MD, PhDy

Hyun Kim, MD, PhD*Kang-Joon Lee, MD, PhD*

Seung-Gul Kang, MD, PhDy*Department of Neuropsychiatry

Inje UniversityCollege of Medicine

Goyangand yDepartment of Psychiatry

Korea UniversityCollege of MedicineSeoul, South Korea

medipark@hanmail.net

REFERENCES1. Daniel DG, Zimbroff DL, Potkin SG, et al.

Ziprasidone 80 mg/day and 160 mg/day in

the acute exacerbation of schizophrenia andschizoaffective disorder: a 6-week placebo-controlled trial. Neuropsychopharmacology.1999;20:491–505.

2. Rachid F, Bertschy G, Bondolfi G, et al.Possible induction of mania or hypomania byatypical antipsychotics: an updated review ofreported cases. J Clin Psychiatry. 2004;65(11):1537–1545.

3. Lu BY, Lundgren R, Escalona PR, et al. Acase of ziprasidone-induced mania and therole of 5-HT2A in mood changes induced byatypical antipsychotics. J Clin Psychiatry.2002;63(12):1185–1186.

4. Larson MF, Hauser A. Possible ziprasidone-induced mania. J Am Acad Child AdolescPsychiatry. 2003;42(9):1012–1013.

5. Nolan BP, Schulte JJ Jr. Mania associated withinitiation of ziprasidone. J Clin Psychiatry.2003;64(3):336.

6. Brieger P. Hypomanic episodes after receivingziprasidone: an unintended Bon-off-on[ courseof treatment. J Clin Psychiatry. 2004;65(1):132–135.

7. Baldassano CF, Ballas C, Datto SM, et al.Ziprasidone-associated mania: a case seriesand review of the mechanism. Bipolar Disord.2003;5(1):72–75.

8. Keating AM, Aoun SL, Dean CE. Ziprasidone-associated mania: a review and report of 2additional cases. Clin Neuropharmacol. 2005;28(2):83–86.

9. Meltzer HY, Matsurbara S, Lee JC. The ratiosof serotonin2 and dopamine2 aff initiesdifferentiate atypical and typical antipsychoticdrugs. Psychopharmacol Bull. 1989;25(3):390–392.

10. Lane HY, Lin YC, Chang WH. Mania inducedby risperidone: dose related? J Clin Psychia-try. 1998;59(2):85–86.

11. Remington G, Kapur S. D2 and 5-HT2

receptor effects of antipsychotics: bridgingbasic and clinical findings using PET. J ClinPsychiatry. 1999;60(suppl 10):15–19.

12. Bymaster F, Perry KW, Nelson DL, et al.Olanzapine: a basic science update. Br JPsychiatry. 1999;176(suppl 37):36–40.

13. Michalopoulou PG, Lykouras L. Manic/hypomanic symptoms induced by atypicalantipsychotics: a review of the reported cases.Prog Neuropsychopharmacol Biol Psychiatry.2006;30(4):549–564.

14. Ichikawa J, Ishii H, Bonaccorso S, et al.5-HT(2A) and D(2) receptor blockade in-creases cortical DA release via 5-HT(1A)receptor activation: a possible mechanismof atypical antipsychotic-induced cortical do-pamine release. J Neurochem. 2001;76(5):1521–1531.

15. Richelson E. Receptor pharmacology of neu-roleptics: relation to clinical effects. J ClinPsychiatry. 1999;60(suppl 10):5–14.

16. Scotte A, Janssen PF, Gommeren W, et al.Risperidone compared with new and referenceantipsychotic drugs: in vitro and in vivoreceptor binding. Psychopharmacology (Berl).1996;124(1–2):57–73.

17. Gefvert O, Lundberg T, Wieselgren IM, et al.D(2) and 5HT(2A) receptor occupancy ofdifferent doses of quetiapine in schizophrenia:a PET study. Eur Neuropsychopharmacol.2001;11(2):105–110.

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins712

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

18. Nemeroff CB, Kinkead B, Goldstein I.Quetiapine: preclinical studies, pharmacoki-netic, drug interactions, and dosing. J ClinPsychiatry. 2002;63(suppl 13):5–11.

19. Kapur S, Seeman P. Does fast dissociationfrom the dopamine D(2) receptor explain theaction of atypical antipsychotics?: a newhypothesis. Am J Psychiatry. 2001;158(3):360–369.

20. Traber R, Schneiter R, Modestin J. A case ofaripiprazole-induced mania. Pharmacopsy-chiatry. 2007;40(1):37–38.

21. Keck PE, Marcus R, Tourkodimitris S, et al. Aplacebo-controlled, double-blind study of theefficacy and safety of aripiprazole in patientswith acute bipolar mania. Am J Psychiatry.2003;160(9):1651–1658.

22. Hirose T, Uwahodo Y, Yamada S, et al. Mecha-nism of action of aripiprazole predicts clinicalefficacy and a favourable side-effect profile.J Psychopharmacol. 2004;18(3):375–383.

23. Murphy BP. Amisulpride-induced mania in apatient with schizophrenia. Br J Psychiatry.2003;183:172.

24. Schmidt AW, Lebel AL, Howard Jr HR, et al.Ziprasidone: a novel antipsychotic agent witha unique human receptor binding profile. EurJ Pharmacol. 2001;425(3):197–201.

25. Potkin SG, Keck PE Jr, et al. Ziprasidone inacute bipolar mania: a 21-day randomized,double-blind, placebo-controlled replicationtrial. J Clin Psychopharmacol. 2005;25(4):301–310.

Leukopenia DuringTherapy With

Risperidone Long-ActingInjectable

Two Case Reports

To the Editors:Hematologic abnormalities are fre-

quently encountered during treatmentwith antipsychotic drugs. Most of theseare mild and of no clinical significance,but in a small minority of patients, haz-ardous, potentially life-threatening hem-atologic effects, including leukopeniaand agranulocytosis, can occur due to acombination of pharmacological andhost factors.1 Clozapine is associatedwith several well-known abnormalitiesof blood cell count,2 but some casereports associate novel antipsychotics,such as risperidone, with leukopenia andagranulocytosis.3–5 This report describes2 cases of leukopenia under treatmentwith risperidone long-acting inject-able, suggesting the necessity of earlyrecognition of leukopenia to prevent oc-

currence of potentially life-threateningagranulocytosis.

CASE REPORT 1Patient M.M. (male, aged 29 years, an

administrativeofficer), diagnosedwith schizo-phrenia, according to the Diagnostic and

Statistical Manual of Mental Disorders,Fourth Edition, Text Revision, was hospital-ized because of psychotic symptoms in theUniversity Department, Vrapce PsychiatricHospital, Zagreb, in May 2005.

After admission, as a part of routinepractice, laboratory examination was per-formed with the following results: whiteblood cell (WBC) count, 2.2 � 103/2L(reference range, 4.5–11.0 � 103/2L); redblood cell (RBC) count, 4.35 � 106/2L(reference range, 4.4–5.8 � 106/2L); hemo-globin, 13.9 g/dL (reference range, 13.0–18.0 g/dL); hematocrit, 38.9% (referencerange, 37.0%–49.0% of RBCs); mean cor-puscular volume, 89.3 2m3 (reference range,78–100 2m3); mean corpuscular hemoglobin(MCH), 31.9 pg/cell (reference range, 25.0–35.0 pg/cell); MCH concentration, 35.8 g/dL(reference range, 31.0–37.0 g/dL); platelets,239 � 103/2L (reference range, 130–400 �103/2L).

The values of biochemistry and urinal-ysis were within reference ranges. Uponadmission, his somatic and neurologicalstatus was normal. There was no infectionor any other symptoms indicating leukope-nia. Before hospitalization, the patient re-ceived 5 risperidone long-acting injectablesduring outpatient treatment. The dosage ofrisperidone long-acting injectable was 50 mgevery 2 weeks. Before treatment with risper-idone long-acting injectables was initiated,the patient was taking risperidone tablets indaily dosage of 4 mg for a period of 1 year,and during that time, the patient’s conditionwas stable. During that period, laboratoryexaminations were performed every 2 weeksduring the first month and, after that, every2 months, and values of laboratory parame-ters were within reference ranges. Thetherapy with risperidone long-acting inject-ables was initiated at the patient’s request—itwas more convenient for the patient to takethe medication every 2 weeks. The patientdenied taking any other medication besidesrisperidone long-acting injectable, whichsuggested that low values of leukocytes wereassociated with this therapy. Admission ofrisperidone long-acting injectable was termi-nated. The patient was transferred to theHematological Department of Internal Med-icine, where the laboratory parameters wereobtained showing improvement in valuesduring 10 days. After 10 days of hospi-talization in the Hematological Depart-

ment of Internal Medicine, the patient wastransferred back to the University Depart-ment. After returning to the UniversityDepartment, therapy with olanzapine orallydisintegrating tablets was initiated. Olanza-pine was administered in daily dosage of 10mg in the evening. The laboratory examina-tion was repeated after 2 days, and WBCcount was within the reference range. Thefollowing laboratory examinations were per-formed on weekly basis, and WBC countwas within the reference range, as well as theother values of complete blood count. Thepatient was discharged from the hospitalafter 30 days of hospital treatment. Afterthe patient was discharged from the hospital,the laboratory examinations were performedon monthly basis, and the values remainedwithin reference ranges.

CASE REPORT 2Patient D.B. (female, aged 38 years, an

economist), diagnosed with schizoaffectivedisorder, according to the Diagnostic andStatistical Manual of Mental Disorders,Fourth Edition, Text Revision, was hospital-ized in the University Department, VrapcePsychiatric Hospital, Zagreb, because ofpsychotic symptoms and aggressive behav-ior, in January 2007.

After admission, as a part of routinepractice, laboratory examination was perform-ed with the following results: WBCs, 2.5 �103/2L (reference range, 4.5–11.0 � 103/2L);RBC count, 4.51 � 106/2L3 (reference range,3.9–5.2 � 106/2L3); hemoglobin, 14.6 g/dL(reference range, 12.0–16.0 g/dL); hematocrit,39.7% (reference range, 36.0%–46.0% ofRBCs); mean corpuscular volume, 83.7 2m3

(reference range, 78–102 2m3); MCH, 32.1pg/cell (reference range, 25.0–35.0 pg/cell);MCH concentration, 33.7 g/dL (refer-ence range, 31.0–37.0 g/dL); platelets, 341 �103/2L (reference range, 130–400 � 103/2L).

The values of biochemistry and urinal-ysis were within reference ranges. Uponadmission, her somatic and neurologicalstatus was normal. There was no infectionor any other symptoms indicating leukope-nia. The patient was examined by thespecialist of internal medicine, who advisedfurther monitoring of laboratory values.Before hospitalization, the patient received7 risperidone long-acting injectables duringoutpatient treatment. The dosage of risper-idone long-acting injectable was 25 mg every2 weeks. Before treatment with risperidonelong-acting injectables was initiated, thepatient was taking risperidone tablets in dailydosage of 2 mg for a period of 2 years, andduring that time, the patient’s condition wasstable. During that period, laboratory exami-nations were performed every 2 weeksduring the first month and, after that, every

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 713

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

2 months, and values of laboratory parame-ters were within reference ranges. As in thefirst case, the therapy with risperidone long-acting injectables was initiated at thepatient’s request, because it was more con-venient for the patient to take medicationevery 2 weeks. The patient denied taking anyother medication besides risperidone long-acting injectable, which suggested that lowvalues of leukocytes were associated withthis therapy. Admission of risperidone long-acting injectable was terminated. Therapywith quetiapine was initiated in daily dosageof 50 mg. The daily dosage was increased to100 mg on the third day of therapy withquetiapine. The laboratory examination wasrepeated after 2 days, showing improvementin WBC count. The daily dosage of quetia-pine was increased to 200 mg on the fourthday and to 400 mg on the sixth day. Onthe eighth day, the dosage was increased to600 mg, and this daily dosage remainedunchanged until the end of hospital treat-ment. The following laboratory examinationswere performed on weekly basis, and WBCcount was within the reference range, as wellas the other values of complete blood count.The patient was discharged from the hospitalafter 32 days of hospital treatment. After thepatient was discharged from the hospital,laboratory examinations were performed onmonthly basis, and the values remainedwithin reference ranges.

DISCUSSIONPsychotropic medications have been

known to cause blood dyscrasias, includ-ing neutropenia, and since the adventof clozapine, this side effect is now in-creasingly being recognized. Almost allthe major classes of psychotropic medi-cations have been associated with neu-tropenia.6 In the described 2 cases ofpatients receiving therapy with risperi-done long-acting injectable, leukopeniaoccurred during treatment with thismedication. In both cases leukopeniawas discovered by serendipity, afterroutine examination of laboratoryparameters. In 1 case, the dosage was50 mg every 2 weeks, and in the othercase, it was 25 mg every 2 weeks,suggesting that leukopenia can occurwhen either of these dosages is adminis-tered. In both described cases, therapywith risperidone long-acting injectablewas initiated after therapy with risper-idone tablets, during which no signifi-cant decrease in WBC count wasconfirmed. The fact that leukopeniaoccurred during therapy with risperidone

long-acting injectable but not duringtherapy with oral risperidone can beexplained by the cumulative period ofexposition of patients to risperidone(application of risperidone long-actinginjectable followed after oral application)and with different absorption and metab-olism of the medication depending on theroute of application of risperidone (oralvs intramuscular). On the other hand,there are cases suggesting a differentperspective. Two cases of patients whodeveloped blood dyscrasias, involvingneutropenia and thrombocytopenia,respectively, after treatment with pheno-thiazine and butyrophenone antipsy-chotics, were reported. A subsequentgood therapeutic response was obtainedwith risperidone. It was suggested thatrisperidone could be considered as analternative when blood dyscrasias occurwith classic antipsychotics.7 Anotherissue is the period of action of risper-idone long-acting injectable, which issignificantly longer compared with tab-lets, which makes it more difficult todeal with leukopenia. These case reportssuggest the necessity of follow-up ofWBC count in the patients takingrisperidone long-acting injectable, evenafter a period of treatment with risper-idone tablets, to recognize leukopeniaand prevent occurrence of potentiallylife-threatening agranulocytosis. Furtherobservation is necessary for more pre-cise understanding of this condition,particularly with respect that both ofthese patients were taking risperidonetablets before administration of risper-idone long-acting injectable, with nodecrease in the values of leukocytes.

Suzana Uzun, MD, PhD*Oliver Kozumplik, MD, PhD*Miro Jakovljevi(, MD, PhDyz

Vera Folnegovi(-Smalc, MD, PhD*z*University Department

Vrapce Psychiatric HospitalyUniversity DepartmentClinical Hospital Zagreband zSchool of Medicine

University of ZagrebZagreb, Croatia

suzana.uzun@bolnica-vrapce.hr

REFERENCES1. Hall RL, Smith AG, Edwards JG. Haematolog-

ical safety of antipsychotic drugs. Expert OpinDrug Saf. 2003;2(4):395–399.

2. Mihaljevic-Peles A, Jakovljevic M, Mrsic M,et al. Thrombocytopenia associated with cloza-pine and fluphenazine. Nord J Psychiatry.2001;55(6):449–450.

3. Sluys M, Guzelcan Y, Casteelen G, et al.Risperidone-induced leucopenia and neutrope-nia: a case report. Eur Psychiatry. 2004;19(2):117.

4. Biswas A, Mittal P, Chaturvedi S, et al.Risperidone induced cytopenias. J Assoc Physi-cians India. 2000;48(11):1122–1123.

5. Finkel B, Lerner AG, Oyffe I, et al. Risper-idone-associated agranulocytosis. Am J Psychi-atry. 1998;155(6):855–856.

6. Duggal HS, Singh I. Psychotropic drug-induced neutropenia. Drugs Today (Barc).2005;41(8):517–526.

7. Mahmood T, Silverstone T, Spittle B. Risper-idone appears safe in patients with antipsy-chotic-induced blood dyscrasias. Int ClinPsychopharmacol. 1996;11(1):53–54.

Recent Trends inSelective Serotonin

Reuptake Inhibitor Usein Pregnancy

To the Editors:Recent data indicate that approx-

imately 10% to 15% of women willhave depression at any point duringpregnancy or the postpartum period.1,2

Selective serotonin reuptake inhibitors(SSRIs) are the first-line pharmaco-therapy to treat depression, and SSRIuse in pregnancy has been well docu-mented. Recent studies have reportedan increased risk of adverse neonataleffects to both early and late fetalexposure to SSRIs.3–12 In addition, inDecember 2005, the Food and DrugAdministration (FDA) issued a publichealth advisory regarding the potentialrisk of ventricular septal defects (VSDs)in fetuses exposed to paroxetine in thefirst trimester.13 To understand thepotential public health consequences ofSSRI use in pregnancy, it is imperativeto know how many fetuses are exposedto SSRIs in utero.

There have been limited data on theprevalence of use of antidepressants inpregnancy.14 A previous study reportedthat approximately 2% to 3% of preg-nant women in the United States15 wereexposed to antidepressants in general,whereas a more recent study demon-strates 13.3% of pregnant women wereexposed to SSRIs.16 The objective of

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins714

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

this study was to determine the preva-lence of use of SSRIs in pregnancy ofthe past 14 years at our medical center,as well as determine any changes inprescribing trends of SSRIs duringpregnancy since the FDA advisoryregarding the potential risk of VSDs infetuses exposed to paroxetine.

With Mayo Foundation InstitutionalReview Board approval, we conducteda review of all pregnant women with de-livery dates between January 1, 1993 toMay 1, 2007, who presented to ourmedical center. The Department of Ob-stetrics and Gynecology maintains aprospective database on all pregnan-cies, and this was used to identify pa-tients who had been exposed to SSRIs.Cases identified from this databasewere then confirmed using individualchart review.

We identified 27,908 total deliveriesduring that period. Of the total number ofdeliveries, 1039 mothers were prescribedSSRIs at some point during their preg-nancy. Selective serotonin reuptake in-hibitors included fluoxetine, paroxetine,sertraline, citalopram, escitalopram, andvenlafaxine. Venlafaxine, a serotonin-norepinephrine reuptake inhibitor, wasincluded in this analysis because itsmechanism of inhibition of the reuptakeiof serotonin is similar to the SSRIs.Pregnant women with a documentedmedication list during their pregnancythat included an SSRI or who weregiven at least 1 SSRI prescription duringtheir pregnancy were selected as theexposed group.

To determine the prevalence ofSSRI use over the course of the past

several years, we reviewed the totalnumber of deliveries and number ofwomen exposed to SSRIs at any pointduring their pregnancy. Prevalence datafor all SSRIs from January 1, 1993through May 1, 2007, were analyzedusing a #2 analysis and JMP 7.0.1software (SAS Institute, Cary, NC).JMP is a commercial computer programthat performs simple and complex sta-tistical analyses.

Of a total number of 27,908 deliv-eries, 1039 pregnant women were iden-tified as having been prescribed an SSRIduring pregnancy. Over the course of14 years at our medical center, theprevalence of use of SSRIs in pregnancyhas steadily increased from 0.44% in1993 to 6.61% in 2007 (Fig. 1). Thisincrease was statistically significant,with P < 0.001.

There was no discernable decreasein overall prevalence of SSRI use duringpregnancy in 2006 and 2007. In 2005,the most prescribed SSRI was sertraline,followed by citalopram, paroxetine,fluoxetine, escitalopram, and venlafax-ine. By 2007, there was a shift in pres-cribing patterns. Sertraline continued tobe the most prescribed SSRI in preg-nancy; however, this was followed byvenlafaxine, fluoxetine, citalopram, andescitalopram. Paroxetine was not pre-scribed to any pregnant patients in themonths we reviewed in 2007.

Paroxetine use averaged approx-imately 14% of all SSRIs prescribedbefore December 2005 and decreased to0% in the months reviewed in 2007. Therewas a 87.5% increase in the use of ven-lafaxine during the period from 2005 to

2007, whereas other SSRIs used in preg-nancy declined between 25% and 50%.

DISCUSSIONOur study had several strengths. We

were able to review data over the courseof several years at a single medicalcenter. Although we are a tertiary med-ical facility, the majority of our obstet-rics population is community based. Wewere able to review all pregnanciesduring this period, which certainly de-creases our selection bias and allowed usto look at prescription practices. Drugexposure information in our data wasnot based on a patient’s memory; hence,there was no possibility of recall bias.

Limitations of our study shouldnot be overlooked. The design of thestudy was retrospective and observa-tional in nature. Although our studysample was quite large, numbers be-came smaller as SSRIs were stratified.In addition, only pregnant women en-rolled at our medical center were in-cluded; thus, results may not be gener-alizable to other populations.

Determining compliance and adher-ence to prescribed SSRI therapy duringpregnancy is problematic. A patient whohad an SSRI documented or prescribedduring pregnancy, but who did notingest the SSRI would be misclassifiedas an exposure. The only way to deter-mine data of actual drug consumptionduring pregnancy would be by patientreport. Our data were taken from physi-cian prescriptions, rather than pharmacyrecords, which may have an increasedrisk of noncompliance.

Furthermore, we cannot determinecausal relationship that the prescribingtrend changes of paroxetine in pregnantwomen are due to the FDA warning re-garding the increased risk of VSDs withfirst-trimester exposure of paroxetine.

In conclusion, SSRI use in preg-nancy has been increasing over time andis nearing 7% in our population. Pre-scribing trends of SSRIs in pregnancyhave changed; paroxetine is no longerused in our pregnant population,whereas venlafaxine is now used ingreater rates than previously prescribed.This may be in part due to recent FDAwarnings regarding the potential risk ofVSDs in fetuses exposed to paroxetinein the first trimester.

FIGURE 1. Prevalence of SSRI use in pregnancy.

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 715

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

AUTHOR DISCLOSUREINFORMATION

The authors declare no conflicts ofinterest.

Christina L. Wichman, DO*Amy Fothergill, BSy

Katherine M. Moore, MD*Tara R. Lang, MDz

Robert H. Heise Jr, MDxWilliam J. Watson, MDx

Department of *Psychiatry and PsychologyMayo Clinic

yMayo Medical SchoolDepartments of zPediatrics and

xObstetrics and GynecologyMayo Clinic

Rochester, MNwichman.christina@mayo.edu

REFERENCES1. Burt VK, Hendrick VC. Clinical Manual

of Women’s Mental Health. Washington,DC: American Psychiatric Publishing, Inc;2005.

2. Kornstein SG, Clayton AH. Women’s MentalHealth. A Comprehensive Textbook. NewYork, NY: Guilford Press; 2002.

3. Alwan S, Reefhuis J, Rasmussen SA, et al.Use of selective serotonin-reuptake inhibitorsin pregnancy and the risk of birth defects.NEJM. 2007;356:2684–2692.

4. Bar-Oz B, Einarson T, Einarson A, et al.Paroxetine and congenital malformations:meta-analysis and consideration of poten-tial confounding factors. Clin Ther. 2007;29:918–926.

5. Berard A, Ramos E, Rey E, et al. First trimes-ter exposure to paroxetine and risk of cardiacmalformations in infants: the importance ofdosage. Birth Defects Res B Dev Reprod Toxicol.2006;80:18–27.

6. Chambers CD, Johnson KA, Dick LM, et al.Birth outcomes in pregnant women takingfluoxetine. NEJM. 1996;335:1010–1015.

7. Chambers CD, Hernandez-Diaz S, Van MarterLJ, et al. Selective serotonin-reuptake inhib-itors and risk of persistent pulmonary hyper-tension of the newborn. NEJM. 2006;354:579–587.

8. Louik C, Lin AE, Werler MM, et al. Firsttrimester use of selective serotonin-reuptakeinhibitors and the risk of birth defects. NEJM.2007;356:2675–2683.

9. Malm H, Klaukka T, Neuvonen PJ. Risksassociated with selective serotonin reuptakeinhibitors in pregnancy. Obstet Gynecol. 2005;106:1289–1296.

10. Oberlander T, Warburton W, Misri S, et al.Neonatal outcomes after prenatal exposure toselective serotonin reuptake inhibitor antide-pressants and maternal depression using pop-ulation-based linked health data. Arch GenPsychiatry. 2006;63:898–906.

11. Pastuszak A, Schick-Boschetto B, Zuber C,et al. Pregnancy outcome following first-

trimester exposure to fluoxetine (Prozac).JAMA. 1993;269:2246–2248.

12. Sanz EJ, De-las-Cuevas C, Kiuru A, et al.Selective serotonin reuptake inhibitors inpregnant women and neonatal withdrawal syn-drome: a database analysis. Lancet. 2005;365:482–487.

13. Food and Drug Administration. FDA publichealth advisory: paroxetine [electronic citation].Available at: http://www.fda.gov/cder/drug/advisory/paroxetine200512.htm. AccessedDecember 17, 2007.

14. Mills JL. Depressing observations on theuse of selective serotonin-reuptake inhi-bitors during pregnancy. NEJM. 2006;354:636–638.

15. Andrade S, Gurwitz J, Davis R, et al.Prescriptiondruguse inpregnancy.Am J ObstetGynecol. 2004;191:398–407.

16. Cooper WO, Willy ME, Pont SJ, et al.Increasing use of antidepressants in preg-nancy. Am J Obstet Gynecol. 2007;196:544e–544e5.

Comments on Articleby Dr Birkenhager et al

’’Efficacy of Imipramine inPsychotic Versus

Nonpsychotic Depression’’

To the Editors:In their recent article in this Jour-

nal,1 Dr Birkenhager and his colleaguesfrom the Erasmus Medical Center inRotterdam, contrary to expectation, con-cluded that 40 patients with major de-pression with psychotic features showedgreater improvement in HamiltonDepression Rating (HAM-D) scoresthan 82 nonpsychotic depressed pa-tients, with mean decreases of 13.7versus 8.9 points, or 4.8 points moreamong psychotic patients. In addition,respective responder rates (improvingby Q50%) were 55% versus 39%. Theyconsidered potential confounding fac-tors in their analyses but did not covaryby baseline HAM-D ratings, which were4.5 points higher among the psychoticpatients (mean, 28.3 vs 23.8), or nearlythe same as the difference in HAM-Ddecrease of 4.8 points (Table 4).

Because higher initial scores oftenyield greater changes, might not relianceon decrease in HAM-D scores poten-tially be misleading? Of note, the pro-portion of patients achieving remission(final HAM-D score e7) was nearly

identical in the 2 groups (25% vs 24%,respectively), suggesting similar treat-ment effects in both groups.

It would be of interest to considerfurther analyses of various outcomessimilar to that provided by Kocsisand colleagues (1990; ref. 22 citedby Birkenhager et al1), with 3 groups(with the first 2 groups defined on acase-control basis): (a) psychotic (psy-chotic and relatively severely depressed),(b) nonpsychotic but matched for initialHAM-D score (relatively severe but non-psychotic), and (c) other nonpsychoticcases (nonpsychotic and less severe).

If the reported findings and con-clusions are valid and replicable, theywould be clinically reassuring, althoughat variance with a long-standing liter-ature that has quite consistently foundinferior responses to antidepressanttreatment among depressed patients withpsychotic features, even with matchingfor initial depression severity, and so inneed of explanation.

AUTHOR DISCLOSUREINFORMATION

Dr Baldessarini has served as aconsultant to, or engaged in collabora-tive, investigator-initiated research withAstraZeneca, Auritec (2008–2009), Bio-trofix, Eli Lilly (2008–2009), IFI, JanssenPharmaceutical, JDS-Noven, Luitpold(2008–2009), Merck, NeuroHealing,Novartis, SK-BioPharmaceutical, andSolvay Corporations, but is not a memberof pharmaceutical speakers’ panels, nordoes he or any family member hold eq-uity positions in pharmaceutical or bio-technology companies. He occasionallyteaches for the New England EducationalInstitute and PriMed Continuing MedicalEducation Corporations.

Ross J. Baldessarini, MDDepartment of PsychiatryHarvard Medical School

and Psychopharmacology ProgramMcLean Division of

Massachusetts General HospitalBoston, MA

rjb@mclean.org

REFERENCE1. Birkenhager TK, van den Broek WW, Mulder

PG, et al. Efficacy of imipramine in psychoticversus nonpsychotic depression. J Clin Psycho-pharmacol. 2008;28:166–170.

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins716

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Reply to DrBaldessarini’sComments on

’’Efficacy of Imipraminein Psychotic Versus

NonpsychoticDepression’’

To the Editors:

First of all, we would like to thankDr Baldessarini for his comments onour recent publication entitled BEfficacyof Imipramine in Psychotic Versus Non-psychotic Depression.[1

We agree that the superior effica-cy of imipramine in psychotic depres-sion could be influenced by the greaterlevel of severity in our particular sam-ple. Indeed, we did perform a secondregression analysis (presented in theResults section of our article)—whichincluded baseline Hamilton DepressionRating Scale (HAM-D) score as anextra covariable—in which no signifi-cant difference was found between our2 groups. We consider mean change inthe HAM-D score to be the appropriateoutcome measure, because both re-sponse (HAM-D ,, Q50%) and remis-sion (final HAM-D score, e7) requiredichotomization of results, which im-plies that a proportion of the informa-tion is lost.

Dr Baldessarini suggests perform-ing an analysis similar to that done byKocsis et al,2 which involved 3 groups:psychotic depressed, nonpsychotic se-verely depressed, and nonpsychotic lessseverely depressed. Such a comparisonis not possible with our data because theproportion of severely depressed non-psychotic patients is too small. How-ever, even if our sample of severelydepressed nonpsychotic patients hadbeen larger, comparing our resultswith those of Kocsis et al2 would havebeen difficult. Similar to our study,that by Kocsis et al2 included inpa-tients who were diagnosed with majordepressive disorder according to theSchedule for Affective Disorders andSchizophrenia,3 after a drug-free base-line period. Patients were treated witheither imipramine (n = 42) or amitripty-line (n = 53) for 4 weeks; for both

antidepressants, the maximum dosagewas 250 mg/d.

There are several important differ-ences in methodology between ourstudy and that of Kocsis et al,2 whichmay explain the difference in efficacy ofimipramine in the sample with psychoticdepression.

The most obvious difference be-tween both studies is that the study ofKocsis et al2 included a considerablenumber of patients with bipolar depres-sion in their study (47 of 132 [36%]);moreover, in their sample, 12 (48%) of25 patients with psychotic depressionactually had bipolar depression. Becausethe efficacy of antidepressants in bipolardepression seems to be disappointing,4

the inclusion of these patients in theirstudy has probably decreased the res-ponse to imipramine in the sample withpsychotic depression.

Differences in dosing strategiesare another possible explanation for thevariance in treatment outcome. Kocsiset al2 aimed for a maximum dose of250 mg of amitriptyline/imipramine.They measured plasma drug concen-trations of both antidepressants, butthey did not adjust the dose of the anti-depressant to obtain a predeterminedplasma level—as we did in our study.This difference may be relevant be-cause 250 mg is probably an appropriatemean daily dose (at least for imipra-mine), that is, the dose necessary toobtain a therapeutic plasma level rangesfrom 50 to 450 mg/d.5 Using a fixeddaily dose of 250 mg, some patients willhave attained a therapeutic antidepres-sant plasma level, whereas others wouldstill have subtherapeutic levels, and athird group would have high plasmalevels. In a meta-analysis,6 evidence wasfound for a curvilinear relationship be-tween plasma level and outcome forboth imipramine (therapeutic window,175–350 ng/mL) and amitriptyline (ther-apeutic window, 93–140 ng/mL). There-fore, antidepressant efficacy in patientswith antidepressant plasma levels out-side the therapeutic window may havebeen reduced. In addition, the durationof antidepressant treatment was signifi-cantly shorter in the study of Kocsiset al2 (4 weeks) compared with that inour study (6 weeks). This is likely tohave also reduced antidepressant efficacy.Because psychotic depressed patients

show a lower placebo response than non-psychotic depressed patients,7 suboptimalantidepressant treatment could reduce theresponse in psychotic depressives to alarger extent than in depressed patientswithout psychotic features.

Dr Baldessarini’s final comment isthat the long-standing literature hasquite consistently found inferior antide-pressant response in patients with psy-chotic depression.

However, several studies8,9 haveshown that monotherapy with antide-pressants seemed to be effective in pa-tients with psychotic depression. Similarto our study, these latter studies in-cluded almost exclusively depressed pa-tients with mood-congruent psychoticfeatures.

The superior response to imipra-mine in our sample with psychotic de-pression might be due to the greaterseverity; nevertheless, the severely de-pressed sample showed a good antide-pressant response.

We agree with Dr Baldessarinithat replication of our findings wouldbe desirable; for this, double-blind ran-domized studies comparing antidepres-sant monotherapy with combinationtherapy in patients with psychotic de-pression are warranted. Such studiesshould take into account the methodo-logical aspects previously mentioned.Recently, an 8-center randomized con-trolled trial investigating the phar-macological treatment of psychoticdepression was conducted in the Nether-lands. The results have not yet beenpublished, but the imipramine responserate (52%) was almost identical to thatreported in our present study.

Tom K. Birkenhager, MD, PhD*Walter W. van den Broek, MD, PhD*

Peter Moleman, PhDyzJan A. Bruijn, MD, PhD*x

*Department of PsychiatryErasmus Medical Centre

RotterdamyMoleman Psychopharmacology

AmerongenzBehavioural Science Institute

Faculty of Social SciencesRadboud University

Nijmegen

and xGGZ RegioBredaMuiderslotstraat, Breda

The Netherlandst.birkenhager@erasmusmc.nl

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 717

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

REFERENCES1. Birkenhager TK, van den Broek WW, Mulder

PG, et al. Efficacy of imipramine in psychoticversus nonpsychotic depression. J Clin Psycho-pharmacol. 2008;28:166–170.

2. Kocsis JH, Croughan JL, Katz MM, et al.Response to treatment with antidepressantsof patients with severe or moderate non-psychotic depression and of patients withpsychotic depression. Am J Psychiatry. 1990;147:621–624.

3. Spitzer RL, Endicott J. Schedule for AffectiveDisorders and Schizophrenia–Lifetime Version.3rd ed. New York, NY: Psychiatric State Ins-titute, Biometrics Research; 1979.

4. Nemeroff CB, Evans DL, Gyulai L, et al.Double-blind, placebo-controlled comparisonof imipramine and paroxetine in the treatmentof bipolar depression. Am J Psychiatry. 2001;158:906–912.

5. Birkenhager TK, van den Broek WW, MolemanP, et al. Imipramine dose in relation to ther-apeutic plasma level: are clinical trials usingimipramine as a positive control flawed? Psy-chopharmacology. 2005;181:595–599.

6. Perry PJ, Zeilmann C, Arndt S. Tricyclicantidepressant concentrations in plasma: anestimate of their sensitivity and specificity asa predictor of response. J Clin Psychopharma-col. 1994;14:230–240.

7. Spiker DG, Kupfer DJ. Placebo response ratesin psychotic and nonpsychotic depression. JAffect Disord. 1988;14:21–23.

8. Zanardi R, Franchini L, Gasperini M, et al.Double-blind controlled trial of sertralineversus paroxetine in the treatment of delu-sional depression. Am J Psychiatry. 1996;153:1631–1633.

9. Zanardi R, Franchini L, Serretti A, et al.Venlafaxine versus fluvoxamine in the treat-ment of delusional depression: a pilot double-blind controlled study. J Clin Psychiatry. 2000;61:26–29.

Comparison BetweenLithium and Valproatein the Treatment of

Acute Mania

To the Editors:Bipolar disorder is a common de-

bilitating illness, characterized by acuteaffective episodes with full or partialinterepisode remission.1–3 Effective andacceptable treatment of acute episodes isrequired. Valproate (valproic acid [Depa-kene] or divalproex sodium [Depakote])in recent years has surpassed lithium inuse for acute mania.2–4 The aim of thisstudy was to compare the effectivenessof valproate sodium with that of lithiumcarbonate in patients with acute maniain a randomized, double-blind, parallel-

group study of treatment outcomes inpatients with bipolar I disorder.

Thirty female inpatients with a di-agnosis of bipolar I disorder (accordingto the Diagnostic and Statistical Manualof Mental Disorders, Fourth Edition, TextRevision diagnostic criteria) who hadbeen admitted in the hospital because ofrecent appearance of a new episode ofmanic symptoms (code F31.x), after aminimum 3-day washout period, wereentered into a 3-week parallel-group,double-blind study for random assign-ment to valproate sodium or lithium car-bonate in a 1:1 ratio. All of them signedthe informed consent, and the whole pro-cedure was approved by the related eth-ical committee of the university. The firstgroup (n = 15) was designated to lithiumcarbonate (300-mg uncoated tablets) asthe solitary antimanic agent, whereas thesecond one (n = 15) was chosen for pre-scription of valproate sodium (200-mgcoated tablets) for the same aim. Bothdrugs were prescribed according to prac-tice guidelines and standard-titration pro-tocols. Although using benzodiazepine(lorazepam) and typical antipsychotic(haloperidol) as adjunctive agents werepermissible during trial, neither combin-ing anticonvulsant nor atypical antipsy-chotic was prescribed throughout theaforesaid assessment. In addition, nopsychosocial intervention other than or-dinary care was used through this phase.The tablets were prescribed while previ-ously inserted into empty and similar cap-sules, which were prepared in this regardto make patients blind with respect to theprocedure. The evaluators were also un-aware concerning the aforesaid partitionand the type of medications arranged foreach group. Primary outcome measureswere changes in the Manic-State RatingScale (MSRS) at baseline and weekly in-tervals up to the third week. Similarly,overall illness severity was rated using theClinical Global Impressions–Severity ofIllness Scale (CGI-S) at baseline andat the end of the third week. Mean do-sage of valproate prescribed in this trialwas 1026 T 148.64 mg, and for lithium,this amount was 1240 T 222.96 mg/d.Mean serum level of lithium also was0.873 T 01486 mEq/L. Also, meandosage of haloperidol used in this trialwas 5.87 T 1.09 mg in the lithium groupand 6.02 T 0.84 mg in the valproategroup. Mean dosage of adjunctive lor-

azepam also was 3.38 T 1.06 in the firstgroup and 3.46 T 0.94 in the second one.There was no significant difference inthis regard.

Treatment group comparisons ofpatient demographic characteristics andbaseline severity measurements weredone using the Student t test. Statisticalsignificance was defined as a 2-sidedP e 0.05. Mean MSRS and CGI-S totalscore changes by time were analyzedwithin each group by means of the pairedt test and compared across the 2 groupsvia the unpaired t test. Analysis of effectsize (ES) by calculation of the value ofCohen d and the ES correlation, r, andcomputation of required confidence in-tervals (CIs) had been performed.

Intent-to-treat analysis for efficacywas based on data from equal number ofpatients in valproate and lithium groups(n = 15 in each). Groups were initiallycomparable, and demographic variables,diagnostic variables, and dosages of neu-roleptics and benzodiazepines used ad-junctively were similar as well (Table 1).There was no premature discontinuationin neither of groups. According to thefindings, both lithium and valproatecaused significant reduction in the se-verity of the manic symptoms at the endof the third week (95% CI, j33.179 toj46.781 [P < 0.0001] along with 95%CI, j25.921 to j41.199 [P < 0.0001],for intensity of symptoms, respectively,and also 95% CI, j38.67 to j55.17[P < 0.0001] and 95% CI, j33.64 toj44.47 [P < 0.001], for the frequencyof symptoms, respectively), which wasnoteworthy from the end of the secondweek. Intensity and frequency of symp-toms decreased correspondingly 10.71%,24.16%, and 50.36% and 12%, 31.09%,and 58.07% from baseline at the end ofthe first, second, and third week by lith-ium. Similar decrement by valproate re-garding the intensity and frequency ofthe symptoms was in that order 8.52%,21.40%, and 41.86% and 6.99%,27.13%, and 49.80% at the end of thefirst, second, and third week. It is in-teresting that there was no main differ-ence between lithium and valproateconcerning their potency in decrementof the intensity and frequency of manicsymptoms in the beginning of the treat-ment during the first 2 weeks, but atthe end of the third week, lithium wasshown significantly more effective in

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins718

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

this regard (95% CI, 0.695–13.705 [P <0.03] as well as 95% CI, 0.121–13.879[P < 0.04], respectively). The intensityof almost all of the items (except items16 and 21 [argumentative and verbal-izes feelings of well-being, respectively])were significantly improved (P < 0.05)by lithium. The matching theme also wasright concerning valproate except item 6(dresses inappropriately) along with item21 (verbalizes feelings of well-being).Moreover, because the sample size wassmall, the ES was analyzed for changeon the MSRS at the end of treatment,which indicated a large d (Q0.8), readilyobservable improvement with both val-proate (Cohen d = 3.47, ES r = 0.866;and Cohen d = 2.840, ES r = 0.817 forfrequency and intensity of symptoms, re-spectively) and lithium (Cohen d = 2.836,ES r = 0.817; and Cohen d = 2.786,ES r = 0.812 for frequency and in-tensity of symptoms, respectively). Onthe other hand, according to CGI-S,the severity of illness decreased from5.2 T 1.08 to 4.12 T 1.33 (P < 0.02)along with 5.06 T 1.36 to 4.31 T 1.16(P < 0.05) in the lithium and valproategroups, respectively.

DISCUSSIONThere are some claims that valpro-

ate is as effective as lithium in acutemania5–7 and about two thirds of personswith acute mania respond to valproate,2

and finally, especially with regard to dys-phoric mania and rapid cycling, it ismore effective than lithium.3,8,9 Becausein this trial only bipolar patients withdistinct manic episode (code F31.x) wererecruited, we cannot infer any outcome

regarding the comparative effectivenessof these mood stabilizers with respectto dysphoric mania. But in this study,although both of them were signifi-cantly effective in amelioration of manicsymptoms, nevertheless, according to theresults, lithium was remarkably more ef-fective in this regard after 2 weeks andat the end of trial. The elapsed time be-tween commencement of treatment andsignificant launching of clinical improve-ment was similar in both groups, and itwas in progress at the end of the secondweek. Our findings are not comparableto the results of Swann et al8 becausewe did not consider the relationship be-tween the number of previous episodes,especially more ones, and the antima-nic response to lithium and divalproex.Moreover, because we did not choosean oral loading dose of valproate, in op-position to Hirschfeld et al,10 we did notsee any earlier efficacy of valproate incomparison to lithium with a standardtitration. Maybe with the use of a rapidloading dose of valproate, we couldexpect an earlier response in the val-proate group.

Our findings were somewhat differ-ent from those of Bowden et al5,6 andMacritchie,1 who found divalproex moreeffective or at least equal to lithium inameliorating mania. In Macritchie’s ar-ticle in Cochrane Database of SystematicReviews in 2003, according to 3 random-ized controlled trials, there was no sig-nificant difference between valproate andlithium. Data were extracted on the mainoutcome Bfailure to respond by the endof the study[ assessed by less than 50%reduction in the Young Mania RatingScale or the Kiddie Schedule for Affec-

tive Disorders and Schizophrenia (K-SADS) mania rating scale. Also, therewere no significant differences in accept-ability as measured by the total numberof subjects withdrawing from the study.In this trial, similar to the findings ofSwann et al8 regarding divalproex incomparison with lithium and placebo,valproate sodium was more efficient thanlithium carbonate on irritability (P <0.007), combativeness (P < 0.014), andimpulsivity (P < 0.038). But this was inconverse with respect to making threats(P < 0.025). This controversy necessi-tates more care with respect to any sharpdeduction based on such small samplesizes. Our findings were very similar tothe findings of Freeman et al,11 partic-ularly because their sample size wasespecially comparable to ours, exceptthat because of different scales used inthe trials, any deduction with respect tocomplementary response to valproatesodium anchored in increased pretreat-ment depression episodes is not possiblein our study. There is no doubt that usingconventional antipsychotics and benzo-diazepines as adjunctive therapies wouldinfluence the whole process, but therandomization used and similar averageprescribed adjuvant agents may havemoderated the influence of these con-founding variables. Small size of thesamples, short duration of the trial, andexclusion of mixed episodes were amongthe weaknesses of this trial. Accessiblecareful monitoring of serum levels, prac-ticable employment of regular physicaland laboratory examinations, and most ofall the requisite of ensuring enhancedrecovery from manic episode are amongthe reasons that can persuade clinicians

TABLE 1. Comparison Between Baseline and Weekly Scores of Patients Receiving Lithium and Valproate

MSRS (Frequency) t P MSRS (Intensity) t P CGI-S t P

Baseline lithium 80.79 T 11.68 79.83 T 9.64 5.2 T 1.08

Baseline valproate 81.43 T 10.73 0.156 (L-V) 0.8769 80.16 T 11.39 0.202 (L-V) 0.8410 5/06 T 1.36 0.319 (L-V) 0.7572

1st wk lithium 72.13 T 16.03 1.691 (L-L) 0.1019 73.93 T 16.02 1.129 (L-L) 0.2615

1st wk valproate 75.73 T 12.37 1.348 (V-V) 0.1884 73.33 T 10.53 1.705 (V-V) 0.0992

2nd wk lithium 55.67 T 11.50 5.935 (L-L) 0.0001 60.20 T 10.42 5.233 (L-L) 0.0001

2nd wk valproate 59.33 T 9.38 6.006 (V-V) 0.0001 63 T 5.03 5.338 (V-V) 0.0001

3rd wk lithium 33.87 T 10.34 11.649 (L-L) 0.0001 39.40 T 8.51 12.042 (L-L) 0.0001 4.13 T 1.33 2.441 (L-L) 0.0212

3rd wk valproate 40.87 T 7.89 11.795 (V-V) 0.0001 46.60 T 8.88 9 (V-V) 0.0001 4.26 T 1.16 1.733 (V-V) 0.0940

2.084 (L-V) 0.0464 2.267 (L-V) 0.03

L-L indicates comparison between baseline score and 1st, 2nd, or 3rd week’s score in the lithium group; V-V, comparison between baseline score and 1st, 2nd, or 3rd week’s score

in the valproate group; L-V, comparison between lithium and valproate groups at baseline or 3rd week.

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 719

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

of an optimistic review with respect tolithium.

Lithium and valproate were bothvaluable in improving manic symptoms,but lithium illustrated more efficaciousincidentally.

AUTHOR DISCLOSUREINFORMATION

The authors declare no funding orconflicts of interest.

Saeed Shoja Shafti, MD*yBahman Shahveisi, MDy*Department of Psychiatry

University of Social Welfareand Rehabilitation Sciences

and yRazi Psychiatric HospitalAmin Abad, Share Rey

Tehran, Iranssshafti@yahoo.com

REFERENCES1. Macritchie K, Geddes JR, Scott J, et al.

Valproate for acute mood episodes in bipolardisorder. Cochrane Database Syst Rev. 2003:CD004052.

2. Sadock BJ, Sadock VA. Depression andBipolar Disorder. Kaplan & Sadock’s Syn-opsis of Psychiatry. 10th ed. Philadelphia,PA: Lippincott, Williams & Wilkins; 2007:536–562.

3. Post RM, Altshuler LL. Mood Disorders:Treatment of Bipolar Disorders; Kaplan &Sadock’s Comprehensive Text Book of Psy-chiatry. 8th ed. Philadelphia, PA: Lippincott,Williams & Wilkins; 2005;1661–1707.

4. Ginsberg DL. Anticonvulsants may reducesuicidality in bipolar disorder. Primary Psy-chiatry. 2003;10:19.

5. Bowden CL, Davis J, Morris D, et al. Effectsize of efficacy measures comparing dival-proex, lithium and placebo in acute mania.Depress Anxiety. 1997;6:26–30.

6. Bowden CL, Brugger AM, Swann AC, et al.Efficacy of divalproex vs lithium and pla-cebo in the treatment of mania. The Depa-kote Mania Study Group. JAMA. 1994;271;918–924.

7. Pope HG Jr, McElroy SL, Keck PE Jr, et al.Valproate in the treatment of acute mania. Aplacebo-controlled study. Arch Gen Psychia-try. 1991;48:62–68.

8. Swann AC, Bowden CL, Calabrese JR, et al.Differential effect of number of previousepisodes of affective disorder on response tolithium or divalproex in acute mania. Am JPsychiatry. 1999;156;1264–1266.

9. Swann AC, Bowden CL, Calabrese JR, et al.Pattern of response to divalproex, lithium, orplacebo in four naturalistic subtypes of mania.Neuropsychopharmacology. 2002;26:530–536.

10. Hirschfeld RM, Baker JD, Wozniak P, et al. Thesafety and early efficacy of oral-loaded dival-proex versus standard-titration divalproex, lith-ium, olanzapine, and placebo in the treatment

of acute mania associated with bipolar disorder.J Clin Psychiatry. 2003;64:841–846.

11. Freeman TW, Clothier JL, Pazzaglia P, et al. Adouble-blind comparison of valproate andlithium in the treatment of acute mania. AmJ Psychiatry. 1992;149:108–111.

Possible Varenicline-Induced Paranoiaand Irritability in aPatient With MajorDepressive Disorder,Borderline Personality

Disorder, andMethamphetamineAbuse in Remission

To the Editors:

Varenicline was introduced to themarket in 2006 for smoking cessation.In November 2007, the Food and DrugAdministration announced that it hadreceived postmarketing reports thatpatients using varenicline for smokingcessation had experienced several ad-verse psychiatric events, including sui-cidal ideation and erratic behavior.There are 3 case reports in the literatureof exacerbated psychotic or hypomanicsymptoms in 3 patients with schizophre-nia or bipolar disorder,1–3 but no othercase reports in any other psychiatricpatients have been reported to myknowledge. I report here the case of apatient who received varenicline twiceand became quite paranoid on bothoccasions.

CASE REPORTA 35-year-old woman with major depres-

sive disorder (in partial remission), generalizedanxiety disorder, borderline personality disorder(currently stable), and crystal methamphet-amine abuse (now in full remission) was beingtreated with weekly therapy and a medicationregimen (with no changes for 2 months) oftopiramate 200 mg PO daily, duloxetine 120 mgPO daily, modafinil 400 mg PO daily, andclonazepam 2 mg PO at night. She had beenabstinent from crystal methamphetamine for6 months, as confirmed by monthly randomurine toxicology screening. She smoked ap-proximately 1 pack per day of cigarettes, andalso smoked 1 joint of marijuana daily, butotherwise she had no other drug use. Shewanted to stop smoking and was prescribed va-

renicline according to the manufacturer’sinstructions, to a final dose of 1 mg PO BID.While taking the varenicline, she be-came increasingly paranoid, believing thatpeople were talking about her and looking ather on the street and at work. This was verysimilar to paranoia that she had experienced 1year prior while abusing crystal methamphet-amine, but she convincingly denied currentmethamphetamine abuse, which was confirmedby negative urine toxicology screens over a 1-month period, and she also denied any changein the quantity or quality of the marijuana thatshe smoked. The varenicline was stopped, andthe paranoia resolved gradually over 2 weeks.The patient had found varenicline to be veryhelpful for smoking cessation, so it wasdecided to retry this medication at a lowerdose. However, at 0.5 mg PO BID, she againbecame quite paranoid, thinking that peopleseveral rows back in an airplane were talkingabout her, and she also became much moreirritable, so much so that she had an alter-cation at her workplace and was subsequentlyfired. Varenicline was again stopped, withresolution of the paranoia and decreasedirritability. Since the discontinuation of vare-nicline (approximately 6 months ago), thisparanoia has not recurred, and she hasremained abstinent from crystal methamphet-amine for more than 1 year now. Unfortu-nately, she has begun smoking cigarettesagain, but now to a lesser extent.

DISCUSSIONNicotine is known to activate several

classes of brain cholinergic receptors, in-cluding many high-affinity receptorscomposed primarily of !4 and "2 sub-units. These receptors are presynaptic,and their activation leads to the releaseof dopamine and other neurotransmit-ters, accounting for the pleasurableeffects of smoking. Varenicline is princi-pally a partial agonist at the !4-"2nicotinic acetylcholine receptor, activat-ing the receptor to cause release ofdopamine at a low level while alsoblocking nicotine binding. During absti-nence from cigarette smoking, this canlead to more sustained release of dopa-mine in comparison to the briefer pulsesof dopamine with each cigarette smoked.It is possible that the increased psychosesin this susceptible patient was triggeredby this elevated and sustained level ofdopamine release. Of course, given hercomplicated history and requirement forseveral concomitant medications, it is notpossible to conclude with certainty thatthe paranoia was directly related to thevarenicline use. However, it is notable

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins720

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

that the patient did not alter her medi-cation regimen or illicit drug use in anyway, other than starting varenicline. Inaddition, the paranoia temporally coin-cided with varenicline use and recurredagain with varenicline rechallenge.

The above case highlights furtherconcerns with possible neuropsychiatricside effects associated with vareniclineuse in patients with psychiatric issues.Further safety assessment in this patientpopulation will be needed, and physi-cians should consider and monitor forpossible psychiatric side effects whenprescribing this medication to patientswith preexisting psychiatric disorders orvulnerability to psychoses.

AUTHOR DISCLOSUREINFORMATION

This author has no conflicts of interestto report.

Gholson J. Lyon, MD, PhDNYU Child Study Center

New York, NYGholson.Lyon@nyumc.org

REFERENCES1. Freedman R. Exacerbation of schizophrenia by

Varenicline. Am J Psychiatry. 2007;164:1269.2. Kohen I, Kremen N. Varenicline-induced manic

episode in a patient with bipolar disorder. Am JPsychiatry. 2007;164:1269–1270.

3. Morstad AE, Kutscher EC, Kennedy WK, et al.Hypomania with agitation associated with vare-nicline use in bipolar II disorder. Ann Pharmac-other. 2008;42:288–289; Epub 2008, Jan 15.

Five Female Cases ofProlonged Depressionin Chronic AnorexiaNervosa Treated WithSelective EstrogenReceptor Modulator

Raloxifene-AugmentedTherapy

To the Editors:Depression that reaches levels re-

quiring clinical attention can frequentlybe observed in female patients with an-orexia nervosa (AN).1–3 The lifetimeprevalence of major depression for fe-males with AN ranges from 20% to 80%

and dysthymia from 19% to 93%.4,5

Anorexia nervosa patients presentingwith depression may not necessarilyshow ordinary, typical depressive moodcharacteristics. In addition to a depres-sive mood and lability, these patientsoccasionally have a poor interoceptiveawareness to identify their own moodstates, resulting in the so-called alexi-thymia. Frustration and social isolationmarked by chronic disorders, as well asthe psychopathology of AN, which in-cludes low self-esteem and body imagedisturbances, exert a further influenceon the patients’ mood. Malnutrition, asecondary condition of anorexia, alongwith metabolic and endocrine disorders,is reported to influence the mood statusof the patients.6,7 Until recently, severalantidepressants have been used for treat-ing AN not only to improve obsessivesymptoms in relation to the patient’s ab-normal eating behavior but also to treatdepression. However, placebo-controlledstudies on conventional antidepressantsfor AN show disappointing results interms of efficacy.8–10

Selective estrogen receptor modula-tors (SERMs), like raloxifene, functionas agonists for estrogen receptors of theskeletal system, central nervous system(CNS), and the cardiovascular systemand function as antagonists for mam-mary glands and the female reproductivesystem.11,12 These pharmacologicalcharacteristics make SERMs generallysuitable to be prescribed for treatingosteoporosis in postmenopausal women.Recently, there have also been reports ofSERMs being used as an antidepressantaugmentation.13,14 This function is pos-sibly caused by the stimulation of estro-gen receptors in the CNS.14,15

We used raloxifene for treating per-sistent depression in female patients withchronic AN who had a resistance to anti-depressants and present 5 cases where animprovement in depression was observed.This project was approved by the medicalethical committee of the Faculty of Med-icine at Shinshu University, and weobtained an informed consent from allparticipants before the study. All 5patients out of 5 entries were diagnosedwith AN based on the Diagnostic andStatistical Manual of Mental Disorders,Fourth Edition and as described in Table1. All patients were females who had ANfor more than 10 years, as well as bouts

of prolonged depression, and had ahistory of more than 3 failed pharmaco-therapies for depression. The average ageof these women was 39.6 years (SD, 9.4years), and the average duration of theirdisorder was 18.8 years (SD, 6.1 years).The average body mass index of thepatients at the start of therapy was13.5kg/m2 (SD, 0.5 kg/m2), which wasabout 60% of their medically recommen-ded body weight. In regard to thesepatients’ depression, the precise durationof their depressive symptoms wasunclear but continued for at least aminimum of 5 years. Pharmacotherapyusing antidepressants continued for morethan 3 years, which was not observed tobe effective. No change in the pharma-cotherapy for each patient occurred in, ata minimum, the 6 months before start-ing this treatment intervention.

A continual dosage of 60 mg/d ofraloxifene (average, 1.75 mg/kg per day;SD, 0.21 mg/kg per day) was adminis-tered to each patient. The extent of thepatients’ depression was assessed usingthe 21-item version of Hamilton Depres-sion Rating Scale (HDRS),16 whichreturned to an average baseline of 33.8(SD, 3.0). Four weeks after commencingraloxifene administration, the averagescore reduced to 29.8 (SD, 5.1); no sig-nificant difference was observed relativeto the levels before beginning the treat-ment (P = 0.2980). However, 8 weeksafter commencing administration, theaverage HDRS score was 19.8 (SD, 1.6);a significant decrease was observedrelative to both the score before treatment(P = 0.0001) and the score obtained at4 weeks (P = 0.0023). Score analysisshowed that improvement occurredprimarily in symptoms associated withmood and anxiety, whereas almost nochange was observed in physical symp-toms. Furthermore, psychological symp-toms of AN other than depressionshowed almost no change throughoutthe course of treatment. We also ensuredthat, through the duration of this study,there was no change in pharmacother-apy other than raloxifene, including anti-depressants and vitamin supplements, orstrategies of psychotherapy.

In the HDRS scores, 3 items (as-signed 6 points) are related to the un-avoidable general physical symptoms,gastrointestinal symptoms, and decreasedlibido. No change in these symptoms was

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 721

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

observed in all patients throughout thetreatment. It is believed that these itemswill not change until the patients recoverfrom chronic AN; therefore, except fromthese 6 points, we believe that the HDRSscores at 2 months subsequent to com-mencing raloxifene therapy showed asubstantial improvement in quality of lifeof the patients.

The limitation is as follows: thenumber of entries was small, and thestudy design was open, uncontrolled,and single armed. The results have to bejudged with caution. Nevertheless, ourresults led us to consider that the im-paired function of estrogen in the CNSmay contribute to the prolonged depres-sion of AN and that SERMs may be atreatment option for patients with chronicAN with persistent depression.

Shin Yokoyama, MD, PhD*yNobuhiro Sugiyama, MD, PhD*

Eiko Sugiyama, PhDyNaoji Amano, MD, PhD*

*Department of PsychiatryShinshu University School of Medicine

Matsumotoand yDepartment of Nutritional Science

Nagano Prefectural CollegeNagano, Japan

syokoyama-psy@umin.net

REFERENCES1. Cantwell D, Sturzenberger S, Burroughs J,

et al. Anorexia nervosa: an affective disorder?Arch Gen Psychiatry. 1977;34:1087–1093.

2. Piran N, Kennedy S, Garfinkel P, et al.Affective disturbances in eating disorders.J Nerv Ment Dis. 1985;173:395–400.

3. Wade TD, Bulic CM, Neale M, et al. Anorexianervosa and major depression: shared geneticand environmental risk factors. Am J Psychia-try. 2000;157:269–271.

4. Herzog DB, Keller MB, Sacks NR, et al.Psychiatric comorbidity in treatment-seekinganorexics and bulimics. J Am Acad ChildAdolesc Psychiatry. 1992;31:810–818.

5. Casper RC. Depression and eating disorders.Depress Anxiety. 1998;8(suppl 1):96–104.

6. Keys A, Brozek J, Henschel A. The Biology ofHuman Starvation. Minneapolis, MN: Univer-sity of Minnesota Press; 1950.

7. Woodside BD, Staab R. Management of psy-chiatric comorbidity in anorexia nervosa andbulimianervosa.CNS Drugs. 2006;20:655–663.

8. Biederman J, Herzog DB, Rivinus TM, et al.Amitriptyline in the treatment of anorexianervosa: a double-blind, placebo-controlledstudy. J Clin Psychopharmacol. 1985;5:10–16.

9. Attia E, Haiman C, Walsh BT, et al. Doesfluoxetine augment the inpatient treatment ofanorexia nervosa? Am J Psychiatry. 1998;155:548–551.

10. Ruggiero GM, Laini V, Mauri MC, et al.A single blind comparison of amisulpride,fluoxetine and clomipramine in the treat-ment of restricting anorectics. Prog Neuro-psychopharmacol Biol Psychiatry. 2001;25:1049–1059.

11. Grese TA, Sluka JP, Bryant HU, et al.Molecular determinants of tissue selectivityin estrogen receptor modulators. Proc NatlAcad Sci U S A. 1997;94:14105–14110.

12. Riggs BL, Hartmann LC. Selective estrogen-receptor modulators—mechanisms of actionand application to clinical practice. N EnglJ Med. 2003;348:618–629.

13. Grigoriadis S, Kennedy SH, Srinivisan J, et al.Antidepressant augmentation with raloxifene.J Clin Psychopharmacol. 2005;25:96–68.

14. Sugiyama N, Sasayama D, Amano N.Remarkable antidepressant augmentationeffect of raloxifene, a selective estrogenreceptor modulator, in a partial responder tofluvoxamine: a case report. J Clin Psychiatry.2007;68:636–637.

15. Jarkova NB, Martenyi F, Masanauskaite D,et al. Mood effect of raloxifene in postmen-opausal women. Maturitas. 2002;42:71–75.

16. Hamilton M. A rating scale for depression.J Neurol Neurosurg Psychiatry. 1960;23:56–62.

Alleviation of BothBinge Eating andSexual DysfunctionWith Naltrexone

To the Editors:Naltrexone is a long-acting opioid

receptor antagonist used clinically in al-coholism, opiate abuse, bulimia nervosa,obsessive-compulsive disorder, andimpulse control disorders. Its effects onsexual function are controversial be-cause it has been shown both to increasethe number of spontaneous erections inmale patients with idiopathic erectiledysfunction1–3 and to decrease overallarousal, compulsive sexual fantasies, andmasturbation in sex offenders or inpatients with other forms of compulsivesexual behavior.4,5 We present a casereport of a patient with depression andposttraumatic stress disorder whosesexual dysfunction remitted with nal-trexone. Informed consent was obtainedfrom the patient.

CASE REPORT

Ms N was a 63-year-old white womanwith major depressive disorder, posttrau-matic stress disorder, and a history of heroindependence (in remission for 30 years) andalcohol abuse (in remission for 15 years), aswell as well-controlled hypertension andosteopenia. She had a history of childhoodphysical abuse and involvement in prosti-tution starting at the age of 11; she also hada history of severe domestic violence. Shehad a hysterectomy and bilateral salpingo-oophorectomy at the age of 22 for endo-metriosis; she took estrogen replacementtreatment during most of the subsequenttime. For her opiate addiction, she had ini-tially undergone methadone maintenance,tapering off methadone after a 5-year treat-ment. Her major depression was in full re-mission on fluoxetine 80 mg daily. Othermedications included atorvastatin 40 mgdaily, lisinopril 2.5 mg daily, alendronate70 mg weekly, pantoprazole 40 mg daily, andpremarin 0.625 mg daily. Approximately ayear after her depression remitted, Ms Nreported episodes of binge eating followed

TABLE 1. Patient Characteristics and Course of Raloxifene-Augmented Therapy

Case Age, yrs SexDuration ofAN, yrs

BMI,kg/m2

Duration ofDepression, yrs

Antidepressant, mg/kgper day

Raloxifene, mg/kgper day

HDRS Score

Baseline 4 wk 8 wk

1 51 F 26 14.0 >10 Milnacipran, 2.33 1.40 33 31 18

2 48 F 24 13.5 >5 Paroxetine, 0.63 1.88 35 33 19

3 37 F 18 12.7 >10 Fluvoxamine, 1.56 1.88 35 35 21

4 32 F 14 13.6 >5 Milnacipran, 1.56 1.88 29 28 19

5 30 F 12 13.6 >5 Fluvoxamine, 4.29 1.71 37 22 22

F indicates female; BMI, body mass index.

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins722

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

by spontaneous vomiting, occurring at least3 times weekly and resulting in significantweight gain (50 lb). Ms N was treated withcognitive-behavioral tactics for severalmonths, with little effect on the frequencyof binging. Naltrexone, 50 mg, was thenstarted to target compulsive eating. She con-comitantly discontinued use of food logs andfood delaying tactics Bto see if the medica-tion would really work on its own.[ Withinseveral weeks, she noticed a significantreduction in the urge to binge. She alsoincidentally reported the new experience ofspontaneous sexual desire and function,which she experienced as pleasurable. Ms Nbelieved that she had actually never experi-enced such desires or sexual pleasure previ-ously. There had been no other medicationchanges for several months before the nal-trexone trial, and there was no clinicalevidence of mania. No significant drug in-teractions were identified. She stopped themedication for 1 week when a prescriptionran out, and within several days, her sexualfantasies remitted, and the urge to bingereturned. Ms N continued to take naltrexone,with a dose increase to 100 mg when sheobserved some waning in efficacy for sup-pressing binge eating, and lost a total of25 lb over 3 months of the treatment course.The increased sexual desire was sustainedover 3 months as well.

DISCUSSIONAdmittedly, Ms N’s pretreatment

sexual dysfunction was likely multifac-torial, with contributions from thetrauma history, selective serotonin reup-take inhibitor (SSRI), essential hyper-tension, antihypertensive medication,and postmenopausal hormonal factors.However, her blood pressure was verywell controlled, and she was taking alow dose of a single antihypertensive.She was receiving estrogen replacementtherapy. It is possible that her long-termhistory of opiate addiction had nega-tively impacted her sexual functionbecause opiates are known to causedecreased libido, erectile dysfunction,and delayed ejaculation. However, aftersome time of complete cessation ofheroin, most addicts return to the levelof sexual performance experiencedbefore the use of opioids.6 Ms N hadbeen free of opiate use for more than 30years. Thus, in this patient, there is ahigh probability of combined SSRI-related and trauma-associated sexualdysfunction. Without more informationabout naltrexone’s mechanism of actionin this case, it is premature to propose

naltrexone as an antidote for SSRI-induced sexual side effects. However,the case does illustrate that opioids andopioid antagonists have different effectson sexual function and that naltrexonemay produce increased sexual desire. Itpreliminarily suggests the need for fur-ther study of naltrexone as an agent thatmay enhance the physiological sexualresponse in patients with histories ofaddiction and trauma.

ACKNOWLEDGMENTThe author thanks Mark Albanese,

MD, for comments on the manuscript.

AUTHOR DISCLOSUREINFORMATION

The author declares no conflicts ofinterest.

Fremonta Meyer, MDDivision of Psychosocial Oncology

and Palliative CareDana-Farber Cancer Institute

Boston, MAflm@alum.dartmouth.org

REFERENCES1. Fabbri A, Jannini EA, Gnessi L, et al. Endor-

phins in male impotence: evidence for naltrex-one stimulation of erectile activity in patienttherapy. Psychoneuroendocrinology. 1989;14:103–111.

2. Brennemann W, Stitz B, Van Ahlen H, et al.Treatment of idiopathic erectile dysfunction inmen with the opiate antagonist naltrexone—adouble-blind study. J Androl. 1993;14:407–410.

3. van Ahlen H, Piechota HJ, Kias HJ, et al.Opiate antagonists in erectile dysfunction: apossible new treatment option? Results of apilot study with naltrexone. Eur Urol. 1995;28:246–250.

4. Ryback R. Naltrexone in the treatment ofadolescent sexual offenders. J Clin Psychiatry.2004;65:982–986.

5. Grant JE, Kim SW. A case of kleptomania andcompulsive sexual behavior treated with nal-trexone. Ann Clin Psychiatry. 2001;13:229–231.

6. Palha AP, Esteves M. A study of the sexualityof opiate addicts. J Sex Marital Ther. 2002;28:427–437.

Tactile and VisualHallucinationsin a Child WithMethylphenidateand FluoxetineCombination

To the Editors:Methylphenidate has been reported

to produce tactile, visual, or somatic hal-lucinations1,2 at therapeutic doses. Theseare rare, usually of acute onset, relativelyeasy to recognize, and resolve within afew days after discontinuation of medi-cation.1,3 Similarly, there are reports ofvisual or auditory hallucinations relatedto fluoxetine treatment.4,5 Besides, bothmethylphenidate and fluoxetine havebeen reported to cause hallucinationseither in combination with each other orin combination with other med-ications.6–8 Here we report a pediatricpatient who developed tactile and visualhallucinations with the combination ofOROS methylphenidate and fluoxetineand plan to discuss possible pathophys-iological mechanisms and clinical im-plications of this unusual side effect.

CASEThe patient was a 10-year-old boy

with a weight of 30.5 kg. He was brought inby his biological mother with the complaintsof aggressive and oppositional behaviors,hyperactivity, irritability, frequent tempertantrums, excessive worries, and fears of los-ing or being separated from his parents. Afterdetailed psychiatric evaluation, he was diag-nosed with attention-deficit/hyperactivity dis-order (ADHD), oppositional defiant disorder,and generalized and separation anxiety dis-orders, according to Diagnostic and Statisti-cal Manual of Mental Disorders, FourthEdition criteria. His developmental historyand intellectual capacity were within normallimits. He has never been treated with anypsychopharmacological agents so far and didnot have past or current diagnosis of any neu-rological diseases. He was started on OROSmethylphenidate 18 mg/day and fluoxetineliquid 10 mg/day in the morning. On thefourth day of treatment, he experienced a rel-atively acute-onset episode of intense halluci-nations 3 hours after taking the medicationswhile he was at home. His mother reportedthat he was frightened and shouted, BThereare people over there. There is a car comingtoward me. It will strike me! Do you not seethat?[ He also claimed, BThere are bugsmoving around on my face[, and he insistedhis mother remove the bugs. The motherreported that visual hallucinations lastedaround 1 hour, but tactile hallucinationslasted more than 2 hours. He subsequentlywent to sleep. Upon awaking, 2 hours later,he remembered what happened to him butdid not report any of these symptoms again.The next day, a similar episode reoccurredwhile he was outside with his mother. But

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 723

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

this time, it was milder and lasted less than 1hour. The mother discontinued both medi-cations, and he was re-examined after 10-daymedication-free period. The mother and thechild denied any history of substance use orexposure before these episodes, and he didnot have these symptoms during themedication-free period. Neurology and gen-eral pediatric consultations revealed no sig-nificant problems. Because his behaviorproblems continued; he was first readminis-tered with OROS methylphenidate 18 mg/day monotherapy with close monitoring.During the next 2 weeks, his behaviorproblems showed moderate improvement,and he did not report any hallucinations,but reported worsening of previous sleepdisturbance and decreased appetite. Mirtaza-pine 15 mg/day at bedtime was added for hisanxiety symptoms, sleep disturbance, anddecreased appetite. During the next 2months, he tolerated both medications gen-erally well, his condition improved signifi-cantly, and no further tactile or visualhallucinations were reported.

DISCUSSIONAny possible psychiatric disorders

or medical conditions that can causehallucinations were excluded by clinicalinterview and neurological and pediatricexaminations and investigations. Anxi-ety or depressive disorders are frequentlyreported comorbid conditions in childrenwith ADHD. However, studies con-ducted with combination of methylphe-nidate with selective serotonin reuptakeinhibitors did not report hallucina-tions.9,10 The timely occurrence withthe treatment and exclusion of otherpossible psychiatric or medical causessuggest the role of medications in theemergence of these hallucinations. How-ever, it is not clear whether the causativeagent was the combination of bothmedications or fluoxetine alone. Restart-ing OROS methylphenidate monother-apy did not cause any hallucinations.So, it could be assumed that OROSmethylphenidate was not the causativeagent alone. However, we did notrechallange with fluoxetine because ofboth ethical considerations and parents’refusal. Therefore, given the fact thathallucinations were not reproduced withfluoxetine alone or in combination, thequestion whether the causative agentwas fluoxetine itself or combination ofboth medications remains unclear at thispoint. A review of literature revealed

that there are reports of visual, tactile,auditory, or somatic hallucinations re-lated to either medications alone or incombination,1–8,11 and the presence ofunderlying psychotic, substance-related,or neurological disorders is consideredas a risk factor for the developmentof hallucinations.3,4,11 However, thereis no report on the co-occurrence oftactile and visual hallucinations witheither medications alone or with combi-nation of both. To our knowledge,there are 2 reports of treatment-relatedhallucinations with combination offluoxetine with either methylphenidate6

or dextromethorphan.8

Two hypotheses have been proposedthat may clarify underlying mechanismsof action of methylphenidate-related hal-lucinations: (1) noradrenergic facilitationtransmission of information in the visualpathway, and (2) interactions among brainmonoaminesystems.2 However, the mech-anisms of fluoxetine-methylphenidatecombination–related hallucinations remainunclear at this point.

In conclusion, we think that thecausative agent was the combinationof both medications rather than eithermedication alone. However, in either sit-uation, it is important to note that thisdistressing side effect may occur evenin the absence of underlying psychotic orsubstance-related disorders, and clini-cians’ awareness is important in this issue,particularly in cases where polypharmacyis considered. In some reports of treat-ment-related unusual side effects (such ashallucinations as in this case), it could bepossible to see that the offending agentwas reintroduced without a crucial needof the patient for this particular medica-tion, possibly to be sure of causality.However, fluoxetine was not crucial inthe treatment of this child as we hadalternative medications. We did not con-sider as ethical to attempt to reproducethese distressing hallucinations in a childby restarting fluoxetine. We consider thatclinicians should try to avoid in suchcases, particularly in children, to repro-duce symptoms by rechallenging, if themedication is not crucial or there arealternatives, without informed consent ofthe patient or family. However, we addedmirtazapine 15 mg/day at bedtime tocounteract methylphenidate’s side effects(such as sleep disturbance and decreasedappetite) and treat his anxiety symptoms.

His behavior problems, anxiety symp-toms, sleep disturbance, and decreasedappetite improved significantly with thiscombination, and no significant sideeffects, including hallucinations, were re-ported. The combination of methylpheni-datewithmirtazapinemay be a reasonablestrategy in the presence of comorbidanxiety or depressive symptoms andmethylphenidate-related side effects inindividuals with ADHD. However, moreresearch is warranted on the efficacy andsafety of this combination.

AUTHOR DISCLOSUREINFORMATION

The authors have no conflict ofinterest with any commercial or otherassociations, and no financial ties todisclose in connection with the submit-ted article.

Murat Coskun, MD

Salih Zoroglu, MDChild and AdolescentPsychiatry Department

Istanbul Medical FacultyIstanbul University

Istanbul, Turkeymuratcoskun78@yahoo.com

REFERENCES1. Rashid J, Mitelman S. Methylphenidate and

somatic hallucinations, J Am Acad ChildAdolesc Psychiatry. 2007;46(8):945–946.

2. Young JG. Methylphenidate-induced hal-lucinosis: case histories and possible mecha-nisms of action. J Dev Behav Pediatr. 1981;2:35–38.

3. Ross R. Psychotic and manic-like symptomsduring stimulant treatment of attention deficithyperactivity disorder. Am J Psychiatry. 2006;163(7):1149–1152.

4. Omar SJ, Robinson D, Davies HD, et al.Fluoxetine and visual hallucinations indementia. Biol Psychiatry. 1995;38(8):556–558.

5. Webb A. Fluoxetine induced auditory halluci-nations in an adolescent, J Paediatr ChildHealth. 2003;39:637–639.

6. Abali O, Mukaddes NM. Methylphenidateinduced hallucinations: case report [in Turk-ish]. Bull Clin Psychopharmacol. 2007;17:195–197.

7. Bourgeois JA, Thomas D, Johansen T, et al.Visual hallucinations associated with fluoxe-tine and sertraline. J Clin Psychopharmacol.1998;18(6):482–483.

8. Achamallah NS. Visual hallucinations aftercombining fluoxetine and dextromethorphan.Am J Psychiatry. 1992;149(10):1406.

9. Findling R. Open-label treatment of comorbiddepression and attentional disorder with co-administration of SRI’s and psychostimulants

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins724

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

in children, adolescents, and adults: a caseseries. J Child Adolesc Psychopharmacol.1996;6:165–175.

10. Gammon KD, Nolan EE, Sverd J, et al.Methylphenidate in combination for treat-ment of attention-deficit disorder and comor-bid depressive disorder. J Child AdolescPsychopharmacol. 1993;3:1–10.

11. Lauterbach EC. Dopaminergic hallucinosiswith fluoxetine in Parkinson’s disease. Am JPsychiatry. 1993;150(11):1750.

Insomnia, NightTerror, and DepressionRelated to Clonidinein Attention-Deficit/Hyperactivity Disorder

To the Editors:

Clonidine is an !-adrenergic agonistthat has both central and peripheral ac-tions.1 This report is about a child withattention-deficit/hyperactivity disorder(ADHD) who experienced insomnia, nightterrors, and depression associated with thelong-term use of clonidine. He revealedresolution of insomnia and night terrorwhen clonidine was removed. Informedconsent was obtained for this case report.

CASE REPORTThe patient is an 8.5-year-old student,

referred by his family about 2.5 years agofor problems including inattention to details orBlack of attention to schoolwork,[ Bdifficultysustaining attention in tasks or play activities,[Boften seemingly not listening when spoken todirectly,[ not following through instructions andfailing to finish schoolwork, and difficultyorganizing tasks and activities. Other problemsincluded Bfidgeting with hands or feet orsquirming in seat,[ Boften leaving seat in class-room, running about or climbing excessively insituations in which it is inappropriate, and difficulty playing or engaging in leisure activitiesquietly.[ He was diagnosed as a case of ADHDand comorbid oppositional defiant disorderaccording to Diagnostic and Statistical Manualof Mental Disorders, Fourth Edition criteria.Methylphenidate (Mph) was started and titratedto 10 mg/d for about 9 months, and then it wasdiscontinued in the summer school holiday.Three months later, the medication was startedand titrated to 20 mg/d. After 1 month, clo-nidine 0.1 mg/d was prescribed since thefamily did not agree with the increase in thedosage of Mph for controlling the child’s hy-peractivity because of the significant decreasein appetite. His hyperactivity and impulsivity

improved significantly. On some days, hermother gave him 0.2 mg/d for better controlof his symptoms. About 1 month after admin-istering clonidine, he had night terrors evenduring day sleep for about 2 to 3 times perweek. The family refused to discontinue themedication because they were worried abouthis problems in school. Clonidine was con-tinued for about 9 months. In these 9 months,he had insomnia after discontinuation ofclonidine. Sometimes he was insomniacthroughout the night. He experienced a severeinsomnia in about 5 trials after clonidine wasdiscontinued. After about 9months, he graduallycooled down, and his separation from his motherbecame very difficult. Clonidine was tapered anddiscontinued while he took Mph for 1 monthlater. Then, he became very sad, anxious, ir-ritable, and depressed. He usually cried in schooland at home. He was reluctant to be separatedfrom his mother. Even his mother had to stay atschool to soothe him. There were no hyper-activity symptoms. The nightmares disappearedabout 2 weeks after discontinuing medication,and the insomnia was resolved about 1 weeklater.Methylphenidatewas discontinued about 1month after discontinuation of clonidine. Therewas no reported history of depressive andanxiety disorders in the child and his family.No pathological finding was found in physicalexamination.

DISCUSSIONAugmentation ofMphwith clonidine

is a strategy for reducing aggression inchildren with ADHD. In addition, cloni-dine alone has been effective in reducingthe core symptoms of ADHD.2 There areonly a few controlled studies on theefficacy of this treatment combination.3,4

Two studies suggested the safety andefficacy of clonidine alone or in com-bination with Mph for the treatmentof ADHD. All of these included a rel-atively very short duration of takingclonidine. However, some concerns havebeen raised about the relative safety ofclonidine even in a short period such astransient sedation, cardiac event, and iso-lated cases of sudden death in children onclonidine in combination with psychosti-mulants.5 A study reported no significantdifference between the 2 groups of cloni-dine and placebo for the side effects ofinsomnia, sadness, and nightmares. Thenumber of subjects taking clonidine was39, and the duration of the study was6 weeks.4

The lack of reports on long-termstudies is surprising. Recently, there hasbeen a report of a patient experiencinginsomnia, night terrors, and depression

associated with long-term subarachnoidinfusion that included clonidine. Symp-toms were resolved when clonidine wasremoved from the infusion.6

A possible evidence to support therole of clonidine in the patient’s com-plaints was the onset time of the nightterror and insomnia. Also, improvementhappened when clonidine was discontin-ued. The patient has taken Mph for about9 months in the previous year and hasnever experienced these complaints.Another possible explanation is that thispatient may have been at increased risk ofcomplications because of his underlyingdisorders and taking Mph.

This case report is rather descriptiveand concerns only 1 patient. Further re-search is necessary to determine if thereis any possible association. Although thespecific contribution of clonidine to thecomplaints remains controversial, it ispossible that clonidine will be associatedwith insomnia, night terror, and severedepression for a long time. However, itmight not be the cause of the problems onits own. The concerns about safety add tothe need for controlled trial evidence inlong-term studies. Then, the risk-to-benefit ratio of this treatment strategymight be considered as controversial.

The complaints induced by clonidinewill certainly remain controversial untilfurther evidence of a cause-and-effectrelationship emerges in a greater num-ber of patients. However, cliniciansshould be aware of the potential side ef-fects of clonidine.

AUTHOR DISCLOSUREINFORMATION

The author declares no conflicts ofinterest.

Ahmad Ghanizadeh, MDChild and Adolescent Psychiatrist

Director of Research Center of Psychiatryand Behavioral Sciences

Shiraz University of Medical SciencesHafez Hospital

Shiraz, Iranghanizad@sina.tums.ac.ir

ghanizadeha@hotmail.com

REFERENCES

1. Lowenstein J. Drugs five years later: clonidine.Ann Intern Med. 1980;92:74–77.

2. Connor DF, Fletcher KE, Swanson JM. Ameta-analysis of clonidine for symptomsof attention-deficit hyperactivity disorder.

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 725

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

J Am Acad Child Adolesc Psychiatry. 1999;38:1551–1559.

3. Connor DF, Barkley RA, Davis HT. A pilotstudy of methylphenidate, clonidine, or thecombination in ADHD comorbid with aggres-sive oppositional defiant or conduct disorder.Clin Pediatr (Phila). 2000;39:15–25.

4. Hazell PL, Stuart JE. A randomized controlledtrial of clonidine added to psychostimulant med-ication for hyperactive and aggressive children.J Am Acad Child Adolesc Psychiatry. 2003;42:886–894.

5. Popper CW. Combining methylphenidate andclonidine: pharmacologic questions and newsreports about sudden death. J Child AdolescPsychopharm. 1995;5:157–166.

6. Bevacqua BK, Fattouh M, Backonja M. Depres-sion, night terrors, and insomnia associated withlong-term intrathecal clonidine therapy. PainPract. 2007;7:36–38.

The Disparity ofPharmacokinetics and

Prolactin Studyfor Risperidone

Long-Acting Injection

To the Editors:We read with great interest a recent

report by Bai et al,1 which suggestedthat shifting oral risperidone to risper-idone long-acting injection (RLAI) inpatients with schizophrenia may de-crease the level of prolactin and thatsignificant correlation between prolactinand 9-hydroxy (OH) risperidone wasobserved at baseline when patients tookoral risperidone. However, our similarstudy did not replicate the results. Someissues regarding study design should beconsidered before making a clinicalimplication.

Our study had been approved by theappropriate institutional review board,and written informed consent wasobtained from all participants beforeenrollment. Twenty-two patients withDiagnostic and Statistical Manual ofMental Disorders, Fourth Edition, diag-nosis of schizophrenia were recruited. Allsubjects were symptomatically stable,which is defined by Positive and Neg-ative Symptoms Scale (PANSS) scores ofless than 80, and had received oral ris-peridone for at least 8 weeks before thescreening visit. The medication wasswitched to RLAI, starting from 25 mg

and flexibly adjusted every 2 weeks after4 weeks by clinical judgment, whilemaintaining oral risperidone for the ini-tial 3 weeks. Clinical efficacy, side ef-fects, and serum concentrations of activemoiety (risperidone plus 9-OH risperi-done) were assessed at baseline andWeeks 2, 4, 8, and 12. Prolactin levelswere measured at baseline and Week12. Plasma concentrations of risperidoneand 9-hydroxyrisperidone were measuredby high-performance liquid chromatog-raphy using ultraviolet detection.2

Eighteen patients completed the12-week trial. The data from 19 patientswho had received at least 1 postassess-ment injection (9 men and 10 women;mean age, 41.4 T 10.2 years) were an-alyzed. The mean dosage of oral risper-idone at baseline was 4.8 T 1.2 mg/d. AtWeek 12, the mean dosage of RLAIwas 34.0 T 8.4 mg every 2 weeks. Theplasma concentrations of active moietyin patients receiving 25 (n = 7), 37.5 (n =9), and 50 mg (n = 2) every 2 weeks ofRLAI were 30.9 T 45.3, 42.2 T 20.6, and71.0 T 24.0 ng/mL, respectively. Theprolactin levels were changed from72.7 T 58.8 to 76.3 T 60.6 ng/mL (P =0.44). In addition, PANSS total scoressignificantly improved from 68.26 T11.78 to 57.42 T 8.75 (P = 0.001). Ourdata revealed no significant correlationbetween active moiety and age, weight,oral risperidone dosage, initial PANSSscores, or prolactin levels. Besides, theconcentrations of 9-OH risperidone ortotal active moiety at baseline were notassociated with oral risperidone dosage.An intriguing trend was found that theactive moiety correlated with changes ofPANSS total scores (P = 0.08) andchanges of positive subscores (P =0.01). No correlation between activemoiety and other clinical variables oradverse effects was noted.

Compared with those of the study ofBai et al,1 our results did not show im-provement of hyperprolactinemia aftershifting oral risperidone to RLAI, andthere was no correlation between oralrisperidone dosage and baseline 9-OHrisperidone levels. Although the dosageof baseline oral risperidone (4.8 T 1.2 inour study vs 4.7 T 1.7 mg/d in the studyby Bai et al1) and RLAI at Week 12(34.0 T 8.4 vs 32.0 T 8.9 mg/2 wk) wassimilar in both studies, the concentra-tions of active moiety at baseline

(57.05 T 33.33 vs 16.27 T 8.18 ng/mL)and after RLAI administration (41.00 T33.26 vs 7.84 T 3.52 ng/mL) wereapparently lower in their study. How-ever, the active moiety levels for oralrisperidone in our study were consistentwith the previous findings,2 whichrevealed 40.4 T 31.1 ng/mL in low-dosegroup (3.6 T 0.9 mg/d; range, 2–5 mg)and 49.7 T 13.4 ng/mL in high-dosegroup (6 mg/d), respectively, in Chinesepatients. Some possible explanationsmight account for disparity. Wide inter-individual variability, different agegroups, and divergent CYP2D6 geno-types may all contribute to the signifi-cant differences in these 2 studies withsmall sample size.3

In addition, their long-term follow-up study showed improvement in PANSSscores only in 50 mg group (j11.67 T11.55) but not in 25 mg (3.36 T 7.09) and37.5 mg (5.75 T 14.32) groups.4 In 50 mggroup, the plasma concentrations of ac-tive moiety are nearly the same (j0.42 T9.32 ng/mL), and prolactin levels wereincreased (11.51 T 39.60 ng/mL) afteroral risperidone switched to RLAI. Incontrast, our study, as well as anotherEuropean study (Switch to RisperidoneMicrospheres or STORMI) with largesample size showed improvement inPANSS scores after the switch.5 Accord-ing to our speculation, the decreased pro-lactin levels after the 12-week shift in thestudy by Bai et al1 might be due to con-servative dosing strategies of RLAI.Thus, PANSS improvement would be ex-pected when the dosing method wasadjusted.

Although some studies with smallsample size (N = 20–25), paralleled withthe observation of Bai et al,1 have shownsignificant correlation between 9-OHrisperidone level and prolactin elevationrecently,6,7 nevertheless, a study withlarger sample size (N = 118) in Chinesepopulation did not yield the samecorrelation.8 The issue of timing forpharmacokinetic sampling, the smallsample size with unknown CYP2D6metabolizer status, and lack of controlfor the variables, such as age, sex,menopausal status, diurnal variation,and time course of treatment,3 willconfound the results of correlationbetween prolactin levels and the phar-macokinetic parameters of antipsy-chotics. Because shifting conventional

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins726

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

depot to RLAI in patients with schizo-phrenia may still cause prolactin eleva-tion,9 whether the shifting from oralrisperidone to RLAI would improvehyperprolactinemia needs more investi-gation. Further studies to evaluate thealterations of prolactin levels duringRLAI treatment should consider moreadequate dosage and the possible con-founding factors in the study design.

ACKNOWLEDGMENTSThis study was supported by grants

from Janssen-Cilag Taiwan, Taipei, andthe Johnson & Johnson Corporation,Taipei. They were not involved in any ofthe processes of case enrollment, dataanalysis, and manuscript preparation.

AUTHOR DISCLOSUREINFORMATION

The authors declare no conflicts ofinterest.

Po-Wei Peng, MD*

Ming-Chyi Huang, MD*yChang-Jer Tsai, MD*Chun-Hung Pan, MD*

Chiao-Chicy Chen, MD, PhD*yChih-Chiang Chiu, MD*yz

*Department of PsychiatryTaipei City Psychiatric Center

Taipei City Hospitaland yDepartment of Psychiatry

College of MedicineTaipei Medical University

Taipei, Taiwanand zDepartment of Epidemiology

King’s College London(Institute of Psychiatry)

London, United Kingdomeric.ccchiu@gmail.com

REFERENCES1. Bai YM, Chen TT, Lin WK, et al. Pharmaco-

kinetics study for hyperprolactinemia amongschizophrenics switched from risperidone torisperidone long-acting injection. J Clin Psy-chopharmacol. 2007;27:306–308.

2. Lane HY, Chiu WC, Chou JC, et al. Risper-idone in acutely exacerbated schizophrenia:dosing strategies and plasma levels. J ClinPsychiatry. 2000;61:209–214.

3. Pandina GJ, Canuso CM, Youssef E, et al.Limitations in discerning the effects of risper-idone and its 9-hydroxy metabolite on prolac-tin levels in a small study of patients withschizophrenia. Hum Psychopharmacol. 2007;22:326–327.

4. Bai YM, Ting TC, Chen JY, et al. Equivalentswitching dose from oral risperidone to ris-peridone long-acting injection: a 48-week

randomized, prospective, single-blind phar-macokinetic study. J Clin Psychiatry. 2007;68:1218–1225.

5. Schmauss M, Sacchetti E, Kahn JP, et al.Efficacy and safety of risperidone long-actinginjectable in stable psychotic patients previ-ously treated with oral risperidone. Int ClinPsychopharmacol. 2007;22:312.

6. Knegtering R, Baselmans P, Castelein S,et al. Predominant role of the 9-hydroxymetabolite of risperidone in elevating bloodprolactin levels. Am J Psychiatry. 2005;162:1010–1012.

7. Melkersson KI. Prolactin elevation of theantipsychotic risperidone is predominantlyrelated to its 9-hydroxy metabolite. HumPsychopharmacol. 2006;21:529–532.

8. Wang L, Yu L, Zhang AP, et al. Serumprolactin levels, plasma risperidone level, poly-morphism of cytochrome P450 2D6 and clin-ical response in patients with schizophrenia.J Psychopharmacol. 2007;21;837–842.

9. Lai YC, Chiu CC, Chen CH, et al. Significantelevations of prolactin levels in patients whoshifted from conventional depot antipsychoticsto long-acting risperidone. J Clin Psychophar-macol. 2007;27:523–524.

Reply to Comments byDr Peng et al for’’Pharmacokinetics

Study forHyperprolactinemia

Among SchizophrenicsSwitched fromRisperidone

to RisperidoneLong-Acting Injection’’To the Editors:

Dr Peng et al1 conducted a similarstudy as ours to switch the antipsychot-ics in symptomatically stable patientswith schizophrenia from oral risperi-done to risperidone long-acting injection(RLAI) for 12 weeks. They reported 2different findings from ours: (1) theirdata revealed no significant correlationbetween the concentrations of 9-OH ris-peridone with oral risperidone dosage,(2) their results did not show improve-ment of hyperprolactinemia after shift-ing oral risperidone to RLAI.

The characteristics of the 2 studieswere similar in sample size (19 in thestudy by Peng et al1 vs 25 in ours),baseline oral risperidone dose (4.8 T 1.2

mg/d in the study by Peng et al1 vs 4.7 T1.7 mg/d in ours), and final averageRLAI dose (34.0 T 8.4/2 wk in the studyby Peng et al1 vs 32.0 T 8.9 mg/2 wk inours). However, our subjects were hos-pitalized patients; during the whole studyperiod, all the subjects were admitted inthe psychiatric hospital,2 that we couldhave good control for the adherence ofbaseline oral risperidone. It has been es-timated that the adherence rates of an-tipsychotics in outpatients fall from 60%to 85% during the first month of treat-ment to as low as 50% by Month 6.3,4

Similar finding was noted in Clinical An-tipsychotic Trial of Intervention Effec-tiveness Study, in which 74% of patientsdiscontinued medication within a fewmonths.5 In previous studies of RLAI,more than 30% of study subjects wereenrolled because of history of poor drugadherence.6,7 If the study by Peng et al1

was also outpatient study as others, thepoor adherence of baseline oral risperi-done may confound the correlation be-tween concentration of 9-OH risperidoneand baseline oral risperidone dose. Sec-ond, in our inpatient study, the bloodsamples were taken at 7 A.M., exactly12 hours after the evening dose ofrisperidone; the standardized proceduremay reduce the possible confounding ofdiurnal variation in concentrations of9-OH risperidone and prolactin. Previ-ous reports had shown diurnal fluctua-tion of risperidone active metabolites,8

and a surge in prolactin concentrationwas noted within 1 to 2 hours after oralrisperidone administration.9 These 2 fac-tors should be considered while address-ing the inconsistent results between ourstudy and that of Peng et al.1 Their base-line prolactin level may be possibly un-derestimated if the enrolled outpatientswith problem of drug adherence. How-ever, after they were switched to RLAIwith improved drug adherence, althoughthe prolactin level may actually decrease,no significant difference will be noted.The phenomenon of significantly de-creased level of prolactin from oral ris-peridone to RLAI was also observed inthe 8-week double-blind study of 640patients with schizophrenia by Chueet al.10 Although the subjects were mixedwith outpatients and inpatients, with ran-domization, the possible confounding ofdrug adherence and timing for bloodsampling could be balanced in each

Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008 Letters to the Editors

* 2008 Lippincott Williams & Wilkins 727

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

group, so that the improvement of hy-perprolactinemia could be observed. Fi-nally and importantly, we agree with thecomment by Dr Peng et al1 about thedosing. The change of prolactin level isdoomed, related to the dose before andafter the switching. In this 12-weekstudy, we designed the switching meth-od as the following: patients who origi-nally on oral risperidone dose of 4 mg/dor less received 25 mg RLAI; thosetaking oral dose of more than 4 mg/dand of 6 mg/d or less received 37.5 mgRLAI; and those taking more than6 mg/d received 50 mg RLAI. The12-week result showed no significantdifference of Positive and Negative Syn-drome Scale (PANSS) scores, includingtotal, negative, and general psychopathol-ogy before and after the switching.11

However, while the study period wasextended to 48 weeks, exacerbation ofPANSS positive score was noted, al-though no significant difference wasnoted for the PANSS scores for posi-tive and general psychopathology.2 Sowe had suggested a mildly adjustedswitching equivalent dose in our report:those originally on oral risperidone doseof 3 mg/d or less received 25 mg RLAI;those taking oral dose of more than 3mg/d, and of 5 mg/d or less, received37.5 mg; and those taking oral dose ofmore than 5 mg/d received 50 mgRLAI.2 Our article is still the only oneto suggest the equivalent switchingdose from oral risperidone to RLAI un-der the unique design of inpatient sub-

jects, and we do not know whether theimprovement of hyperprolactinemia willbe still noted with the minor adjustmentof equivalent dose. We thank Dr Penget al1 for their comments to raise someimportant issues about RLAI. Morestudies are required to resolve the debate,but the factor of drug adherence and timefor blood sampling should be optimallycontrolled.

Ya Mei Bai, MD*yTzu Ting Chen, MDzx

Wen Kuo Lin, MDzWen-Ho Chang, MDk

Chih Hung Hung, MDzJen-Yeu Chen, MDz

*Department of PsychiatryCollege of Medicine

National Yang-Ming UniversityyDepartment of Psychiatry

Taipei Veterans General HospitalzDepartment of Psychiatry

Yu-Li Veterans HospitalxGraduate Institute of Behavior Sciences

Kaohsiung Medical Universityand kDepartment of Psychiatry

Buddhist Tzu Chi General HospitalDa Lin,Taiwan, Republic of China

sms0604@ms1.hinet.net

REFERENCES1. Peng P-W, Huang M-C, Tsai C-J, et al. The

disparity of pharmacokinetics and prolactinstudy for risperidone long-acting injection.J Clin Psychopharmacol. 2008;28:726–727.

2. Bai YM, Ting CT, Chen JY, et al. Equiva-lent switching dose from oral risperidone to

risperidone long-acting injection: a 48-weekrandomized, prospective, single-blind pharma-cokinetic study. J Clin Psychiatry. 2007;68:1218–1225.

3. Byerly MJ, Fisher R, Carmody T, et al. A trialof compliance therapy in outpatients withschizophrenia or schizoaffective disorder. JClin Psychiatry. 2005;66:997–1001.

4. Remington G, Kwon J, Collins A, et al. Theuse of electronic monitoring (MEMS) toevaluate antipsychotic compliance in out-patients with schizophrenia. Schizophr Res.2007;90:229–237.

5. Lieberman JA, Stroup TS, McEvoy JP, et al.Effectiveness of antipsychotic drugs in patientswith chronic schizophrenia. N Engl J Med.2005;353:1209–1223.

6. Moller HJ, Llorca PM, Sacchetti E, et al.Efficacy and safety of direct transition torisperidone long-acting injectable in patientstreated with various antipsychotic thera-pies. Int Clin Psychopharmacol. 2005;20:121–130.

7. Fleischhacker WW, Eerdekens M, Karcher K,et al. Treatment of schizophrenia with long-acting injectable risperidone: a 12-monthopen-label trial of the first long-acting second-generation antipsychotic. J Clin Psychiatry.2003;64:1250–1257.

8. Eerdekens M, Van Hove I, Remmerie B, et al.Pharmacokinetics and tolerability of long-actingrisperidone in schizophrenia. Schizophr Res.2004;70:91–100.

9. Turrone P, Kapur S, Seeman MV, et al. Ele-vation of prolactin levels by atypical antipsy-chotics. Am J Psychiatry. 2002;159:133–135.

10. Chue P, Eerdekens M, Augustyns I, et al.Comparative efficacy and safety of long-actingrisperidone and risperidone oral tablets. EurNeuropsychopharmacol. 2005;15:111–117.

11. Bai YM, Wu B, Chen TT, et al. A comparativeefficacy and safety study of long-actingrisperidone injection and risperidone oraltablets among hospitalized schizophreniapatients: 12-week randomized, single-blindstudy. Pharmacopsychiatry. 2006;39:135–141.

Letters to the Editors Journal of Clinical Psychopharmacology � Volume 28, Number 6, December 2008

* 2008 Lippincott Williams & Wilkins728

Copyright @ 2008 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.