Interview EXPERTS IN CHILD PSYCHIATRY

Choosing for children with : Evidence plus experience

Seeking effi cacy while managing adverse eff ects in early-onset psychosis

® Dowden“A patient Health I’ve seen Media for a number of years had been diag- nosed in the pervasive developmental disorder spectrum, but she was quite atypical. Her perseverative thinking CopyrightFor personalfocused use on aonly fantasy world, and she was so preoccupied that it was very diffi cult to pull her out of it. Now at age 12, she has a full-blown psychotic disorder, and the fantasy world is enveloping her. She hears people talking to her all day long.”

Jean A. Frazier, MD, who treats this patient and other children with psychotic disorders, was 1 of 4 principal investigators in the Treatment of Early-Onset Schizo- phrenia Spectrum Disorders (TEOSS) study, a random- ized, double-blind, multisite trial funded by the National Institute of Mental Health. The study, published in © 2009 MARK KAUFFMAN November 2008,1 compared the effi cacy and tolerabil- ity of 3 antipsychotics—, , and molindone—in pediatric patients with schizophrenia or schizoaffective disorder (Box 1, page 62). Dr. Frazier discusses the unexpected fi ndings of the TEOSS trial with Current Psychiatry Section Editor Robert A. Kowatch, MD, PhD. Based on the trial fi nd- ings and her experience, she tells how she makes deci- sions when prescribing antipsychotics for children and adolescents with schizophrenia and related disorders.

DR. KOWATCH: The TEOSS trial found no signifi cant diff er- ences in effi cacy between molindone and the atypical antipsychotics (olanzapine and risperidone) included in the study. You’ve prescribed both typical and atypi- Current Psychiatry 60 July 2009 cal antipsychotics in research and in your clinical prac-

For mass reproduction, content licensing and permissions contact Dowden Health Media.

60_CPSY0709 60 6/16/09 11:43:35 AM tice. Do you believe there’s any diff erence Jean A. Frazier, MD, is the between the 2 classes? Robert M. and Shirley S. Siff Chair DR. FRAZIER: There are some diff erences. and professor of psychiatry and pediatrics, and vice chair and For example, treatment-refractory patients, director, division of child and especially young children, sometimes adolescent psychiatry, University of Massachusetts Medical School, need more D2 blockade than some atypi- Worcester, MA. cal antipsychotics provide. I’ve seen more extrapyramidal side eff ects with the typical antipsychotics than the atypicals, although Robert A. Kowatch, MD, PhD, it’s not uncommon to see some akathisia a Section Editor for CURRENT PSYCHIATRY, is professor with ari piprazole or some dystonia and dys- of psychiatry and pediatrics, kinesia with risperidone. Cincinnati Children’s Hospital Medical Center, Cincinnati, OH. DR. KOWATCH: What are the benefi ts and risks of using antipsychotics in young children? DR. FRAZIER: The benefi t is that antipsychot- ics can decrease children’s suff ering and get them more centered in reality so they can Clinical Point enjoy their friends and progress in school. fi nding of this study to me was the side ef- The 3 antipsychotics And when that happens, it’s wonderful. What fect profi le of these agents. in TEOSS showed no are the risks? With the atypicals my greatest DR. KOWATCH: You mean weight gain with concern is weight gain, and with the typical olanzapine and extrapyramidal symptoms diff erence in effi cacy, agents it’s tardive dyskinesia. with molindone? but the meaningful DR. KOWATCH: Have you changed the way DR. FRAZIER: Yes. fi nding to me was you prescribe antipsychotics as a result of the side eff ect profi le the TEOSS study? DR. FRAZIER: Actually, I have. Clinicians have Managing side eff ects of these agents to be very careful about selecting psycho- DR. KOWATCH: How do you manage antipsy- tropic agents that can worsen pediatric- chotic side eff ects? onset obesity. Olanzapine is an eff ective DR. FRAZIER: For any of the antipsychotics’ agent for targeting psychosis and mood side eff ects, you have to decide whether symptoms, but the weight gain associ- to continue the agent or switch to anoth- ated with it is a concern. I do not prescribe er . For example, I’ve had a olanzapine as much as I have in the past, number of children—many with signifi cant although I keep it in my armamentarium weight problems—whose psychotic symp- and tend to reserve it for third- or fourth- toms have responded only to risperidone. line therapy. So we put them back on risperidone, and I have found molindone to be quite ef- the decision then becomes what can we do fective in children with schizophrenia or to help with the weight gain while continu- schizoaff ective disorder, especially in those ing that agent. who have gained a lot of weight on atypi- For weight gain, I think the best inter- cal antipsychotics. They usually lose weight vention is diet, exercise, and drinking a lot on molindone. of water, but that can be eff ective only if DR. KOWATCH: Do you think the TEOSS you engage the patient’s entire family in study had adequate power to demonstrate the intervention as well. Short of that, a diff erences among molindone, olanzapine, number of pharmacologic interventions and risperidone? have been studied, although not specifi - DR. FRAZIER: We enrolled 119 patients— cally in children. which is large for a study such as this—but In an open-label trial our group con- we did not reach our target of 168 patients, ducted with 11 children age 10 to 18 years which might have increased our power to who had gained weight while taking atyp- detect diff erences. Among the children we ical antipsychotics, metformin decreased did enroll, the 3 antipsychotics showed no lipid levels and body mass index but not Current Psychiatry diff erence in effi cacy, but the meaningful signifi cantly. I’ve followed these children Vol. 8, No. 7 61

61_CPSY0709 61 6/16/09 11:43:40 AM Box 1 Antipsychotics in children with schizophrenia: TEOSS study adds to debate about effi cacy and tolerability

he 5-year National Institute Tof Mental Health-funded Observed PANSS total score by week Treatment of Early-Onset of treatment Children and Schizophrenia Spectrum Disorders antipsychotics (TEOSS) trial began with an ambitious goal: to compare the Molindone Olanzapine Risperidone effi cacy and safety of 1 typical 110 and 2 atypical antipsychotics 100 in children age 8 to 19 with 90 schizophrenia. The primary 80 hypothesis was that atypical 70 agents would show greater 60 Clinical Point effi cacy and tolerability when given 50 for 8 weeks. Instead, the atypical total score PANSS 40 If a child has not agents showed no greater effi cacy, 30 and adverse effects occurred with 012345678 responded to 2 trials all 3 antipsychotics. Because the Weeks with atypicals, trial was designed for 168 subjects but enrolled 119, it may not have Mean Positive and Negative Symptom Scale (PANSS) total scores of I might start thinking observed cases during each week of the TEOSS trial. Minimum possible been adequately powered to PANSS score is 30; scores >60 typically are viewed as problematic. about a typical agent detect differences among the 3 or agents. Medications: Most of the 116 children who treatment discontinuation. Extrapyramidal received medications were severely ill with symptoms were monitored with involuntary psychotic symptoms when randomly assigned movement and akathisia scales. to 1 of the 3 antipsychotics for 8 weeks of Among the 70 patients who completed double-blind treatment. Administration began at treatment (25 of 40 with molindone, 17 of 35 the lowest dose in a set range and usually was olanzapine, and 28 of 41 with risperidone), increased to midrange within 10 to 14 days. more than one-half failed to achieve an Dosing remained fl exible within these ranges: adequate response. Response rates were • molindone, 10 to 140 mg/d (mean endpoint 50% with molindone, 34% with olanzapine, dose 59.9 mg/d) and 46% with risperidone. The atypical • olanzapine, 2.5 to 20 mg/d (mean endpoint antipsychotics did not show greater effi cacy dose 11.4 mg/d) than molindone, and mean reductions in psychotic symptoms were modest (20% to • risperidone, 0.5 to 6 mg/d (mean endpoint 34% on the PANSS). Mean medication doses dose 2.8 mg/d). were midrange and considered moderate. Benztropine, ≥1 mg/d, was given to all patients Tolerability: Sedation, irritability, and anxiety treated with molindone, 14% of those treated were frequent adverse events. Patients with olanzapine, and 34% of those treated with receiving molindone reported signifi cantly risperidone to prevent or manage akathisia. higher rates of akathisia (P < .0008). Those Effi cacy: Two criteria defi ned treatment receiving olanzapine reported signifi cantly response: a Clinical Global Impression higher rates of weight gain (P < .0001) and improvement score of 1 or 2 and a ≥20% were the only group with increased lipid and reduction in baseline Positive and Negative insulin serum levels and liver function tests. Syndrome Scale (PANSS) score. Tolerability Patients in the risperidone group reported outcomes included neurologic side effects, signifi cantly higher rates of constipation P( < weight changes, laboratory analyses, vital .021) and were the only group that experienced signs, ECG, serious adverse events, and elevated serum prolactin.

Source: Sikich L, Frazier JA, McClellan J, et al. Double-blind comparison of fi rst- and second-generation antipsychotics in early-onset schizophrenia and schizoaffective disorder: fi ndings from the Treatment of Early-Onset Schizophrenia Spectrum disorders (TEOSS) Current Psychiatry study. Am J Psychiatry. 2008;165:1420-1431 62 July 2009 continued on page 66

62_CPSY0709 62 6/16/09 11:43:46 AM continued from page 62 Box 2 Prodromal symptoms of early-onset psychotic disorder

became interested in the complicated clearly differentiate the prodromal signs of I overlap between pervasive developmental psychosis from normal childhood behaviors. disorders spectrum and psychotic disorders Children who are psychotic often don’t make early in my training. More is known about good eye contact. When you try to engage Children and prodromal symptoms in adolescents and them in discussion about hearing voices, antipsychotics adults than in children. they’re inattentive and internally preoccupied. A group in the Netherlands5 compared 32 adolescents with severe early defi cits Normal vs psychotic children. You want a in affect regulation, anxiety, disturbed child in the latency age to have a rich fantasy social relationships, and thought disorder life. If they do not, that raises concerns. Both (characterized as “multiple complex normal and psychotic children sometimes developmental disorder” [MCDD]) with say an imaginary friend told them something. 80 adolescents with prodromal psychotic Normal children eventually will admit this friend symptoms who met criteria for “at-risk is imaginary. When children are psychotic, mental state” (ARMS). Three-quarters especially at an early age, you can’t pull them Clinical Point of the children with MCDD (78%) were out of thinking about the imaginary friend, When you try to found to meet criteria for ARMS, and the 2 and they can’t distinguish fantasy from reality. groups showed similar schizotypal traits, Psychotic children also hear imaginary friends engage psychotic disorganization, and prodromal symptoms. talking to them much more often. Normal children usually are not afraid of children to discuss Signs of progression to psychosis and their imaginary friends, whereas psychotic schizophrenia in children typically include: hearing voices, children—particularly adolescents—often are • change in personality afraid of the voices they hear. However, if a they are inattentive • decrease in functioning or decline psychotic child has heard voices from a young in ability to perform at school and internally age, the voices aren’t always ego-dystonic. The • unusual thoughts or behaviors preoccupied girl I mentioned at the beginning of this article • crippling anxiety likes having the voices around. In fact, she gets • supersensitivity to stress. uncomfortable when the voices are quiet. With experience, the clinician can more Jean A. Frazier, MD

in my practice, however, and all those who DR. FRAZIER: doesn’t tend to continued taking metformin over a period be associated with as much weight gain of months lost weight. as olanzapine or risperidone, although I’ve had children—especially in the autism spectrum—who have gained quite a bit of Choosing antipsychotics weight on aripiprazole. Clinically, I’ve no- DR. KOWATCH: Let’s say you’re seeing psy- ticed that aripiprazole seems to brighten up chosis in a 12-year-old whom you think is children’s aff ect. It also seems to help many schizophrenic, and he or she has not yet re- children in my practice with attentional ceived an antipsychotic. What are your top symptoms, although that’s anecdotal. 3 treatment choices? Although we don’t have a lot of data DR. FRAZIER: The fi rst agent I usually select to inform this discussion about aripipra- is risperidone. We have the most data on zole, a placebo-controlled study of 302 the use of this in chil- adolescents diagnosed with schizophrenia dren and adolescents, and most psychotic showed that aripiprazole, 10 mg/d, target- children I see do better with a bit more D2 ed negative symptoms fairly well, based on blockade than some of the other atypicals changes from baseline in PANSS (Positive provide. That said, I remain concerned about and Negative Syndrome Scale) total scores. risperidone’s side eff ects—such as weight This was a 6-week multicenter, double- gain and increased serum prolactin—so my blind, randomized, trial.2 usual second-line agent is aripiprazole. Ultimately, cognition in patients with DR. KOWATCH: Why do you like aripiprazole schizophrenia is the strongest predictor of Current Psychiatry 66 July 2009 for this patient population? success in the workplace and in school. We

66_CPSY0709 66 6/17/09 10:50:29 AM need data on what happens to neurocog- nitive functioning with aripiprazole—and Related Resources all the other atypical agents. • Longitudinal assessment and monitoring of clinical status and brain function in adolescents and adults. Boston Center DR. KOWATCH: What would be your third- for Intervention Development and Applied Research (CIDAR) line agent? study. www.bostoncidar.org. DR. FRAZIER: Well, that varies for me. I’m try- • Frazier JA, Hodge S, Breeze JL, et al. Diagnostic and sex eff ects on limbic volumes in early-onset bipolar disorder and ing to match the medication I use with the schizophrenia. Schizophr Bull. 2008;34(1):37-46. individual patient, and at this point I pre- • Frazier JA, McClellan J, Findling RL, et al. Treatment of scribe based on the side-eff ect profi le more Early-Onset Schizophrenia Spectrum disorders (TEOSS): demographic and clinical characteristics. J Am Acad Child than anything else. I also consider if the child Adolesc Psychiatry. 2007;46:979-988. has a family member who has suff ered from Drug Brand Names a similar condition and what agents the fam- Aripiprazole • Abilify Molindone • Moban ily member responded to. Benztropine • Cogentin Olanzapine • Zyprexa Let’s say I have a child who has tried 1 or Clozapine • Clozaril • Trilafon • Haldol • Seroquel 2 atypical antipsychotics and has not had a Metformin • Glucophage Risperidone • Risperdal good response. Many times I decide to try Disclosures Clinical Point yet another atypical, and often I will try que- Dr. Kowatch receives grant/research support from the tiapine. But after a patient has not respond- Stanley Foundation, National Institute of Mental Health, Sometimes I use a ed to 2 atypicals, I might start thinking about National Institute of Child Health and Human Development, and the National Alliance for Research on Schizophrenia and very low dose of a typical agent or clozapine. I use clozapine Depression. He is a consultant to AstraZeneca and Forest quite a bit. I fi nd it is the most effi cacious Pharmaceuticals and a speaker for AstraZeneca. haloperidol (0.5 to agent available, and the data speak to this as Dr. Frazier receives grant/research support from Bristol-Myers 2 mg/d) with an Squibb, Eli Lilly and Company, GlaxoSmithKline, Johnson & well.3,4 It has been truly remarkable for some Johnson, Neuropharm, Otsuka America Pharmaceuticals, atypical agent, and and Pfi zer Inc. children in my practice. the combination can

an optimal treatment plan for a child or ad- be quite eff ective Less than 50% chance of effi cacy? olescent with psychosis? DR. KOWATCH: The TEOSS study found 50% DR. FRAZIER: These children need a multi- or lower response rates across 8 weeks of modal approach. Pharmacotherapy is the antipsychotic treatment. Clinically, what cornerstone because you want to decrease kind of response rates do you see with anti- positive symptoms of psychosis, but often psychotics in children and adolescents? these children require therapeutic school DR. FRAZIER: I probably see about a 50% placements or residential programs. If they’re response rate in my practice as well. It’s old enough, cognitive-behavioral therapy variable, and the earlier the onset of the ill- can help by teaching them skills to manage ness, the harder it is to treat. ongoing psychotic symptoms. Older teens DR. KOWATCH: Do you ever combine a typi- often have comorbid substance abuse and cal antipsychotic with an atypical? may require substance abuse intervention. DR. FRAZIER: I try not to, but a number of chil- dren in the schizophrenia spectrum have en- during positive symptoms after 2 or 3 trials of Are antipsychotics overused? atypical antipsychotics. Sometimes adding a DR. KOWATCH: Do you think antipsychot- touch of a typical agent can improve the sit- ics are overused in pediatric patients with uation. The typicals I usually try are perphen- psychosis? azine (around 8 to 16 mg/d) or molindone DR. FRAZIER: In pediatric patients with psy- (around 20 to 60 mg/d). Sometimes I use a chosis? No. very low dose of haloperidol (such as 0.5 to DR. KOWATCH: What about in pediatric pa- 2 mg/d) with an atypical agent, and it can be tients with behavioral disorders? quite eff ective. DR. FRAZIER: We need more studies to in- DR. KOWATCH: That has been our experi- form our practice and to be mindful of the ence as well; sometimes combining typi- evidence. Most children with schizophre- cal and atypical agents improves response. nia have substantial developmental chal- Current Psychiatry Besides medications, what do you consider lenges (Box 2).5 In the autism spectrum, Vol. 8, No. 7 67

67_CPSY0709 67 6/16/09 11:43:55 AM often an atypical antipsychotic is the only that stimulants can be quite helpful for the agent that can help a patient who is ag- aggressive child with ADHD.7 gressive, self-injurious, or agitated. DR. KOWATCH: I don’t see any child and ado- In terms of bipolar disorder in children lescent psychiatrist in the United States using and adolescents, it would be ideal if we antipsychotics to treat uncomplicated ADHD. had more head-to-head comparator stud- The kids we see [at specialty clinics] have co- ies to inform our prescriptive practice. For morbid problems such as conduct disorder, Children and example, we need more studies comparing oppositional-defi ant disorder, mood instabil- antipsychotics traditional mood stabilizers such as lithium ity—whatever you want to call it. And we’re with the atypical agents. seeing these patients because they haven’t Of course it would be ideal if we could done well on other medications, such as use monotherapy in children who suff er stimulants. Usually the parents are desperate from bipolar disorder and schizophrenia. because these children are moody and ag- But early-onset bipolar disorder—like early- gressive. I don’t think anybody wants to treat onset schizophrenia—can be very diffi cult children with antipsychotics or mood stabiliz- to treat and often requires more than 1 ers, but it’s what keeps these children well. Clinical Point agent. DR. FRAZIER: Yes, I agree. Early-onset bipolar In a recent study of a pharmacotherapy algorithm for treating pediatric bipolar dis- References disorder and 1. Sikich L, Frazier JA, McClellan J, et al. Double-blind comparison order,3 the children who did the best were of fi rst- and second-generation antipsychotics in early-onset schizophrenia can be schizophrenia and schizoaffective disorder: fi ndings from the on a combination of a mood stabilizer and Treatment of Early-Onset Schizophrenia Spectrum Disorders very diffi cult to treat an atypical antipsychotic. That has been (TEOSS) study. Am J Psychiatry. 2008;165:1420-1431. 2. Findling RL, Robb A, Nyilas M, et al. A multiple-center, and often requires my experience, too. I do my best to man- randomized, double-blind, placebo-controlled study of oral age children on a mood stabilizer alone, aripiprazole for treatment of adolescents with schizophrenia. more than one agent Am J Psychiatry. 2008;165(11):1432-1441. but I rarely have been able to do that. 3. Findling RL, Frazier JA, Gerbino-Rosen G, et al. Is there a role for clozapine in the treatment of children and adolescents? J In terms of attention-defi cit/hyperactiv- Am Acad Child Adolesc Psychiatry. 2007;46(3):423-428. ity disorder (ADHD), it depends on what’s 4. Kim Y, Kim BN, Cho SC, et al. Long-term sustained benefi ts of clozapine treatment in refractory early onset schizophrenia: a going on with the child. Certain children retrospective study in Korean children and adolescents. Hum with an ADHD diagnosis have complicated Psychopharmacol. 2008;23(8):715-722. 5. Sprong M, Becker HE, Schothorst PF, et al. Pathways to behavioral issues. First I would wonder if psychosis: a comparison of the pervasive developmental disorder subtype multiple complex developmental disorder they have a diff erent diagnosis, particularly and the “at risk mental state.” Schizophr Res. 2008;99:38-47. if it gets to the point that an atypical agent 6. Pavuluri MN, Henry DB, Devineni B, et al. A pharmacotherapy algorithm for stabilization and maintenance of pediatric is being considered. But sometimes it be- bipolar disorder. J Am Acad Child Adolesc Psychiatry. comes a question of treating pronounced 2004;43(7):859-867. 7. Sinzig J, Döpfner M, Lehmkuhl G, et al. Long-acting aggression. We need more studies to in- has an effect on aggressive behavior in form what we do. Some studies indicate children with attention-defi cit/hyperactivity disorder. J Child Adolesc Psychopharmacol. 2007;17(4):421-432.

Bottom Line Early-onset schizophrenia can be very diffi cult to treat. Antipsychotics are the cornerstone of therapy, but they have serious side eff ects and adequately control symptoms in only 50% of psychotic children. The TEOSS study reported antipsychotics’ benefi ts and risks for children with schizophrenia. These patients need a multimodal approach, with medication, therapeutic schooling, and other interventions such as cognitive-behavioral therapy.

Current Psychiatry 68 July 2009

68_CPSY0709 68 6/17/09 10:50:34 AM