First- and Second-Generation Antipsychotics for Children and Young Adults

1
First- and Second-Generation Antipsychotics for
Children andYoung Adults

• Prepared for
• Agency for Healthcare Research and Quality (AHRQ)
  
• www.ahrq.gov

2
Outline of Material

• Introduction to pediatric use of first-generation
  and second-generation antipsychotics, including
  approved and off-label indications.
• Systematic review methods.
• The clinical questions addressed by the
  comparative effectiveness review.
• Results of studies and evidence-based conclusions
  about effectiveness and adverse effects of
  antipsychotics in pediatric use.
• Gaps in knowledge.
• What to discuss with patients and their
  caregivers.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

3
Introduction to Antipsychotics in Pediatric Use
(1 of 4)

• Antipsychotics can be classified into two
  categories based on the timeline of their
  development, their mechanisms of action, and
  their anticipated adverse effect profiles.
• First-generation antipsychotics (FGAs), also
  called conventional or typical antipsychotics
• Second-generation antipsychotics (SGAs), also
  called atypical antipsychotics
• FGAs were the first successful pharmacological
  treatments for primary psychotic disorders such
  as schizophrenia.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

4
Introduction to Antipsychotics in Pediatric
Use(2 of 4)

• First-generation antipsychotics are associated
  with side effects that are difficult to manage
  and in some cases are irreversible.
• Neurological side effects include extrapyramidal
  system movement disorders
• Tardive dyskinesia repetitive, involuntary
  muscle movements
• Akathisia restlessness
• Inability to initiate movement Parkinsons
  disease-like symptoms
• Development of the second-generation
  antipsychotics was driven by the need to limit
  neurological adverse effects.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

5
Introduction to Antipsychotics in Pediatric Use
(3 of 4)

• Many first-generation and second-generation
  antipsychotics are approved by the U.S. Food and
  Drug Administration (FDA) for use in children
  and/or adolescents.
• The FDA prohibits manufacturers from advertising
  or promoting the use of pharmaceuticals for
  indications that it has not approved. To do so is
  illegal.
• Off-label prescribing by physicians is permitted.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

6
Introduction to Antipsychotics in Pediatric Use
(4 of 4)

• Prescribing of antipsychotics for children with
  mental and behavioral disorders continues to
  increase and includes off-label use.
• However, the effects of both first-generation
  (FGAs) and second-generation (SGAs)
  antipsychotics on patient-centered outcomes such
  as growth, development, and quality of life are
  not well understood.
• Studies of efficacy, benefits, and adverse
  effects of FGAs and SGAs used in pediatric
  treatment are reported in the clinical
  literature.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

7
Antipsychotics Approved for Pediatric
UseFirst-Generation Antipsychotics
Antipsychotic Indication Age Group
Chlorpromazine Schizophrenia 112 years
Chlorpromazine Bipolar disorder (mania) 112 years
Chlorpromazine Severe behavioral problems 112 years
Chlorpromazine Hyperactivity 112 years
Droperidol Agitation children
Loxapine Schizophrenia 12 years
Perphenazine Schizophrenia 12 years
Pimozide Tourettes syndrome 12 years
Prochlorperazine Schizophrenia gt2 years and gt20 pounds
Thiothixene Schizophrenia 12 years
Thioridazine Schizophrenia children
Trifluoperazine Schizophrenia 6 years

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

8
Antipsychotics Approved for Pediatric
UseSecond-Generation Antipsychotics
Antipsychotic Indication Age Group
Aripiprazole Schizophrenia 1317 years
Aripiprazole Autism irritability 1017 years
Aripiprazole Bipolar disorder (manic/mixed) 617 years
Olanzapine Schizophrenia Adolescents (1317 years)
Olanzapine Bipolar disorder (manic/mixed) Adolescents (1317 years)
Olanzapine Bipolar disorder depressive episode Adolescents (1317 years)
Quetiapine Schizophrenia 1317 years
Quetiapine Bipolar disorder (acute manic) 1017 years
Risperidone Schizophrenia 1317 years
Risperidone Autism irritability 516 years
Risperidone Bipolar disorder (manic/mixed) 1017 years

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

9
Disorders Treated by Antipsychotics inPediatric
Use

• The clinical literature includes studies of
  pediatric use of first-generation and
  second-generation antipsychotics for treatment of
  these disorders
• Pervasive developmental disorders
• Attention deficit hyperactivity disorder and
  disruptive behavior disorders
• Bipolar disorder
• Schizophrenia and related psychosis
• Tourettes syndrome
• Behavioral issues
• No clinical studies were found for
  obsessive-compulsive disorder, post-traumatic
  stress disorder, or anorexia nervosa.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

10
Agency for Healthcare Research and Quality (AHRQ)
Comparative Effectiveness Review (CER) Development

• Topics are nominated through a public process,
  which includes submissions from health care
  professionals, professional organizations, the
  private sector, policymakers, members of the
  public, and others.
• A systematic review of all relevant clinical
  studies is conducted by independent researchers,
  funded by AHRQ, to synthesize the evidence in a
  report summarizing what is known and not known
  about the select clinical issue. The research
  questions and the results of the report are
  subject to expert input, peer review, and public
  comment.
• The results of these reviews are summarized into
  Clinician Research Summaries and Consumer
  Research Summaries for use in decisionmaking and
  in discussions with patients. The Summaries and
  the full report, with references for included and
  excluded studies, are available at
  www.effectivehealthcare.ahrq.gov/pedantipsych.cfm.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

11
Rating the Strength of Evidence From the
Comparative Effectiveness Review

• The strength of evidence was classified into four
  broad categories

High ??? Further research is very unlikely to change the confidence in the estimate of effect.
Moderate ??? Further research may change the confidence in the estimate of effect and may change the estimate.
Low ??? Further research is likely to change the confidence in the estimate of effect and is likely to change the estimate.
Insufficient ??? Evidence either is unavailable or does not permit estimation of an effect.
Seida J, Schouten J, Mousavi S, et al.
Comparative Effectiveness Review No. 39.
Available at www.effectivehealthcare.ahrq.gov/peda
ntipsych.cfm.
12
Clinical Questions Addressed by theComparative
Effectiveness Review (1 of 2)

• The reviewed literature was limited to studies
  performed in children, adolescents, and young
  adults from 1 to 24 years of age.
• What is the comparative efficacy or effectiveness
  of first-generation (FGAs) and second-generation
  (SGAs) antipsychotics for treating
  disorder-specific and nonspecific symptoms?
• Do FGAs and SGAs differ in short-term (within 6
  months) and long-term (after 6 months) outcomes,
  including
• Response rate and relationship to dosage speed
  of response duration of response remission
  relapse time to discontinuation and adherence
• Growth and maturation cognitive and emotional
  development
• Suicidal behaviors and ideation
• Work-related functional capacity school
  performance
• Patient insight into illness
• Patient- or caregiverreported outcomes
• Health-related quality of life
• Legal or justice system interactions
• Health care system utilization

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

13
Clinical Questions Addressed by theComparative
Effectiveness Review (2 of 2)

• Do first-generation (FGAs) and second-generation
  (SGAs) antipsychotics differ in the following
  medication-associated adverse events
• Overall adverse events?
• Specific adverse events?
• Withdrawals and time to withdrawal due to adverse
  events?
• Persistence and reversibility of adverse events?
• Do the efficacy and risks of FGAs and SGAs vary
  in differing subpopulations including
• Sex, age group, and race?
• Comorbidities?
• Cotreatment versus monotherapy
• First episode versus prior episodes (for
  schizophrenia)?
• Duration of illness?
• Treatment naïve versus history of antipsychotic
  use?

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

14
Clinically Significant Outcomes of Interest in
the Comparative Effectiveness Review

• Both core illness symptoms and nonspecific
  symptoms were of interest.
• A wide variety of psychiatric assessment
  instruments are used to evaluate symptoms in
  studies of pediatric use of antipsychotics. Lower
  scores indicate less severity. The assessments
  included

Instrument Application
Aberrant Behavior Checklist (ABC) Treatment effects on profoundly mentally retarded participants
Brief Psychiatric Rating ScaleChildren (BPRS-C) Childhood psychiatric disorders and response to treatment
Childrens Depression Rating Scale (CDRS) Severity of depression in children 612 years of age
Childrens Global Assessment Scale (CGAS) Overall measure of disturbance children 416 years of age
Clinical Global Impressions (CGI) Global rating of severity (-S), improvement (-I), or overall impression
Nisonger Child Behavior Rating Form (NCBRF) Childhood problems because of mental retardation
Overt Aggression Scale (OAS) Physical and verbal aggressive behaviors
Positive and Negative Syndrome Scale (PANSS) Positive and negative symptoms in schizophrenia
Young Mania Rating Scale (YMRS) Severity (not diagnosis) of mania
Seida J, Schouten J, Mousavi S, et al.
Comparative Effectiveness Review No. 39.
Available at www.effectivehealthcare.ahrq.gov/peda
ntipsych.cfm.
15
Adverse Events of Interest in theComparative
Effectiveness Review

• Adverse events data extracted from published
  studies included
• Mortality
• Cardiac and cerebrovascular events
• Weight and body composition
• Dyslipidemia
• Insulin resistance and diabetes
• Prolactin-related and sexual effects
• Neuromotor (extrapyramidal)
• Sedation

• Liver toxicity
• Neutropenia and agranulocytosis
• Thyroid dysfunction
• Seizures
• Neuroleptic malignant syndrome
• Constipation
• Exercise intolerance
• Precocious puberty
• Behavioral effects
• Dermatological effects

For adverse events reporting, standardized
instruments that evaluate extrapyramidal symptoms
are also used in clinical trials.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

16
Summary of Study Characteristics Evaluated in the
Effectiveness Review PICOTS

• Population Children to age 12 years, adolescents
  ages 1217 years, and young adults ages 1824
  years with these diagnosed mental disorders PDD,
  ADHD, DBD, bipolar disorder, schizophrenia or
  related psychosis, Tourettes syndrome, or other
  behavioral symptoms
• Interventions Any FGAs and SGAs approved by the
  FDA
• Comparators Other FGAs or SGAs, placebo
• Outcomes Core and nonspecific symptoms
  response remission growth and maturation
  cognitive and emotional development
  suicide-related behaviors adherence school and
  work capacity and performance patient insight
  patient-, parent-, or caregiver-reported outcome
  health-related quality of life legal system
  interaction health care system utilization
  adverse events (e.g., weight dyslipidemia
  insulin and blood glucose effects and diabetes
  extrapyramidal symptoms prolactin effects)
• Timing No minimum or maximum duration specified
• Setting Community and hospitalized care

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

17
Modes of Results Reporting and Statistical
Analysis in the Comparative Effectiveness Review
(1 of 2)

• 95-Percent Confidence Interval The range of
  statistically valid results that will include the
  true population mean in 95 of 100 repeated
  experiments.
• Mean Difference (MD) The difference between
  treatment and comparison group means.
• To determine a standardized mean difference
  (SMD), results from different scales are
  normalized to a common, standardized scale
  before calculating the mean difference.
• For MD and SMD, the result is statistically
  significant (p lt 0.05) when the 95-percent
  confidence interval does not include 0.0, which
  is the point of no difference between groups.
• Relative Risk (RR) The ratio of the rate
  (absolute risk, probability) of an event in the
  treatment group, to the rate of the event in the
  comparison group.
• For RR, the result is statistically significant
  at p lt 0.05 when the 95-percent confidence
  interval does not include 1.0, which is the point
  of equal risk for both groups.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

18
Modes of Results Reporting and Statistical
Analysis in the Comparative Effectiveness Review
(2 of 2)

• Absolute Risk Difference The absolute value of
  the mathematical difference between the rates
  (risk) of an event in the treatment and
  comparison groups.
• ARD ARC-ART
• Number Needed to Treat or Harm (NNT, NNH) The
  number of patients to be treated to observe
  benefit or harm in one patient more than seen in
  the comparison group. The number of patients to
  be treated in order to find a benefit or harm
  attributable to the intervention.
• NNT or NNH ARC-ART-1 for a benefit or adverse
  event, respectively
• Number of attributable events per 1000 1000 x
  ARC-ART
• The smaller the NNT or NNH, the greater the
  attributable effect.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

.
19
Effectiveness and Efficacy of Antipsychotics

• Efficacy studies
• Are randomized, controlled trials (RCTs) designed
  to demonstrate a clinical benefit
• Enroll participants based on restrictive
  inclusion and exclusion criteria
• The comparator is usually placebo
• Effectiveness studies
• May be RCTs but can include retrospective cohort
  studies and other study formats
• Are designed to examine clinical effects of
  treatments in typical medical practice
• Inclusion and exclusion criteria are not as
  restrictive as in efficacy trials
• The comparator is usually an active control or
  usual care, rather than placebo
• The treatment setting may vary widely among the
  participants

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

.
20
Summary of Efficacy and Effectiveness Studies of
Antipsychotics in Pediatric Use

• The included studies were limited to those that
  enrolled children (12 years of age), youth
  (1218 years of age), and young adults (1924
  years of age).
• Comparisons of second-generation antipsychotics
  (SGAs) with placebo were the most commonly
  reported studies.
• Too few placebo comparison studies of
  first-generation antipsychotics (FGAs) for
  treatment of psychiatric disorders of children
  were reported to permit conclusions about effect
  sizes and statistical significance.
• Head-to-head comparisons of FGAs with other FGAs
  included in this review (two studies) were
  studies about treatment of schizophrenia.
• Head-to-head comparisons of FGAs versus SGAs that
  met the criteria for inclusion in this review (17
  studies) were investigations for treatment of
  schizophrenia, pervasive developmental disorder
  (autism), and Tourettes syndrome.
• Head-to-head comparisons of SGAs with other SGAs
  (46 studies) were investigated for treatment of
  bipolar disorder, schizophrenia, and behavioral
  issues.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

21
Efficacy of Second-Generation Antipsychotics
Placebo Comparisons, by Indication (1 of 3)
Outcome (n studies, n participants) Meta-analytic Result (Mean Difference) and Statistically Valid Range (95 CI) Strength of Evidence
Bipolar Disorder Bipolar Disorder Bipolar Disorder
CGI (6, 978) Pooled SGA results (aripiprazole, olanzapine, quetiapine, and risperidone) CGI score is 0.67 points lower with SGAs when compared with placebo (0.51 to 0.84 points lower) ???
YMRS No meta-analysis of SGAs as a class due to statistical heterogeneity ???
YMRS (3, 328) Aripiprazole 7.22 points lower than placebo (5.17 to 9.28 points) ???
YMRS (1, 159) Olanzapine 7.66 points lower than placebo (5.68 to 9.74 points) ???
YMRS (3, 339) Quetiapine No pooled result due to statistical heterogeneity among reports ???
YMRS (1, 163) Risperidone 8.31 points lower than placebo (4.88 to 11.74 points) ???
YMRS (1, 218) Ziprasidone 5.22 points lower than placebo (2.36 to 8.08 points) ???
CDRS (4, 532) NSD (No pooled values) ???
Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ???

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

22
Efficacy of Second-Generation Antipsychotics
Placebo Comparisons, by Indication (2 of 3)
Outcome (n studies, n participants) Meta-analytic Result (Mean Difference) and Statistically Valid Range (95 CI) Strength of Evidence
Schizophrenia Schizophrenia Schizophrenia
CGAS (5, 880) Score is 4.56 points higher (2.55 to 6.57 points higher) with SGAs (aripiprazole, olanzapine, paliperidone, quetiapine, and risperidone) ???
CGI-I (3, 505) Score is 0.75 points lower (0.38 to 1.12 points lower) with SGAs (aripiprazole, olanzapine, and risperidone) ???
PANSS (6, 987) Score is 8.69 points lower (5.61 to 11.76 points lower) with SGAs (aripiprazole, olanzapine, paliperidone, quetiapine, and risperidone) ???
Tourettes Syndrome Tourettes Syndrome Tourettes Syndrome
YGTSS (2, 54) Score is 6.98 points lower (3.62 to 10.34 points lower) with SGAs (risperidone and ziprasidone) ???
Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ???

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

23
Efficacy of Second-Generation Antipsychotics
Placebo Comparisons, by Indication (3 of 3)
.
Outcome (n studies, n participants) Meta-analytic Result (Mean Difference) and Statistically Valid Range (95 CI) Strength of Evidence
Pervasive Developmental Disorder Pervasive Developmental Disorder Pervasive Developmental Disorder
ABC (3, 202) Score is 18.29 points lower (9.51 to 27.08 points lower) with SGAs (aripiprazole and risperidone) ???
CARS (2, 62) Score is 4.94 points lower (1.36 to 8.52 points lower) with SGAs (risperidone) ???
CYBOCS ( 3, 325) Score is 1.71 points lower (0.25 to 3.17 points lower) with SGAs (aripiprazole, risperidone, and olanzapine) ???
Attention Deficit Hyperactivity Disorder/Disruptive Behavior Disorders Attention Deficit Hyperactivity Disorder/Disruptive Behavior Disorders Attention Deficit Hyperactivity Disorder/Disruptive Behavior Disorders
ABC (4, 293) Score is 20.97 points lower (10.83 to 31.11 points lower) with risperidone ???
BPI (2, 225) Score is 3.77 points lower (1.39 to 6.16 points lower) with risperidone ???
CGI-I (3, 380) Score is 0.95 points lower (0.25 to 1.66 points lower) with risperidone ???
NCBRF ( 4, 590) Score is 6.93 points lower (3.49 to 10.38 points lower) with risperidone ???
Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ???

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

24
Summary of Results Efficacy of Antipsychotics in
Placebo Comparisons (1 of 2)

• When compared with placebo, second-generation
  antipsychotics (SGAs) result in greater
  improvement of disorder-specific symptoms.
• SGAs (aripiprazole, olanzapine, paliperidone,
  quetiapine, and risperidone) improve both
  clinical global impressions and positive and
  negative symptoms of schizophrenia (an approved
  indication).
• Strength of Evidence Moderate
• SGAs (aripiprazole, olanzapine, quetiapine, and
  risperidone) improve clinical global impressions
  (CGIs) of bipolar disorder and manic but not
  depressive symptoms of bipolar disorder.
• Strength of Evidence Moderate for CGIs
• Strength of Evidence Low for manic and
  depressive symptoms

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

25
Summary of Results Efficacy of Antipsychotics in
Placebo Comparisons (2 of 2)

• Risperidone and ziprasidone improve tics in
  Tourettes syndrome.
• Strength of Evidence Moderate
• Risperidone improves behavioral symptoms and
  clinical global impressions of attention deficit
  hyperactivity disorders/ disruptive behavior
  disorders.
• Strength of Evidence Moderate
• Second-generation antipsychotics (aripiprazole
  and risperidone) improve behavioral
  (irritability, an approved indication),
  obsessive-compulsive, and autistic symptoms of
  pervasive developmental disorders.
• Strength of Evidence Low

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

26
Results of Comparative Effectiveness
StudiesFGAs Versus SGAs

• The evidence from comparisons of first-generation
  (FGAs) and second-generation (SGAs)
  antipsychotics in Tourettes syndrome is
  insufficient to permit conclusions about effect
  size or statistical significance.

Comparison Outcome (n studies, n participants) Mean Difference and Statistically Valid Range Defined by the 95 CI Strength of Evidence
FGAs vs. SGAs for Pervasive Developmental Disorders FGAs vs. SGAs for Pervasive Developmental Disorders FGAs vs. SGAs for Pervasive Developmental Disorders FGAs vs. SGAs for Pervasive Developmental Disorders
Haloperidol vs. Olanzapine Autistic Symptoms (CPRS) (1, 12) Anger and hyperactivity factors reduced only in the olanzapine group (p 0.05) ???
Haloperidol vs. Risperidone Autistic Symptoms (ABC, CGI) (1, 30) Behavior, impulsivity, and language skills improved to a greater degree with risperidone (p 0.05) ???
FGAs vs. SGAs for Schizophrenia FGAs vs. SGAs for Schizophrenia FGAs vs. SGAs for Schizophrenia FGAs vs. SGAs for Schizophrenia
Haloperidol vs. SGAs (clozapine, olanzapine, and risperidone BPRS (3 , 71) Score is 11.44 points lower with SGAs (Statistically valid range is 3.52 to 19.35 points lower) ???
Haloperidol vs. SGAs (clozapine, olanzapine, and risperidone CGI-I (3, 87) Score is 0.76 points lower with SGAs (Statistically valid range is 0.26 to 1.25 points lower) ???
Haloperidol vs. SGAs (clozapine, olanzapine, and risperidone PANSS (2, 27) NSD (Statistically valid range is 33.85 points lower to 14.08 points greater) ???
Single study no meta-analysis. 95 CI 95-percent confidence interval the range of statistically valid results mean difference the difference between treatment and control group means NSD No statistically significant difference (p 0.05). Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Single study no meta-analysis. 95 CI 95-percent confidence interval the range of statistically valid results mean difference the difference between treatment and control group means NSD No statistically significant difference (p 0.05). Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Single study no meta-analysis. 95 CI 95-percent confidence interval the range of statistically valid results mean difference the difference between treatment and control group means NSD No statistically significant difference (p 0.05). Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Single study no meta-analysis. 95 CI 95-percent confidence interval the range of statistically valid results mean difference the difference between treatment and control group means NSD No statistically significant difference (p 0.05). Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ???

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

27
Results of Comparative Effectiveness Studies
SGAs Versus SGAs
Comparison Outcome (n studies, n participants) Mean Difference and Statistically Valid Range Defined by the 95 CI Strength of Evidence
SGAs vs. SGAs for Schizophrenia SGAs vs. SGAs for Schizophrenia SGAs vs. SGAs for Schizophrenia SGAs vs. SGAs for Schizophrenia
Olanzapine vs. Risperidone BPRS (3, 136) NSD ???
Olanzapine vs. Risperidone CGI-S (2, 111) NSD ???
Olanzapine vs. Risperidone PANSS (3, 143) NSD ???
Olanzapine vs. Clozapine BPRS (2, 27) NSD Statistically valid range from 10.82 points less to 0.64 points more ???
Olanzapine vs. Clozapine CGI-S (2, 64) NSD Statistically valid range from 1.01 points less to 0.08 points more ???
Olanzapine vs. Clozapine SANS (2, 64) NSD Statistically valid range from 15.45 points less to 5.36 points more ???
95 CI 95-percent confidence interval the range of statistically valid results mean difference the difference between treatment and control group means NSD no statistically significant difference (p 0.05) Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval the range of statistically valid results mean difference the difference between treatment and control group means NSD no statistically significant difference (p 0.05) Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval the range of statistically valid results mean difference the difference between treatment and control group means NSD no statistically significant difference (p 0.05) Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval the range of statistically valid results mean difference the difference between treatment and control group means NSD no statistically significant difference (p 0.05) Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ???

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at
• www.effectivehealthcare.ahrq.gov/pediantipsych.cfm
  .

28
Summary of Comparative Effectiveness Results (1
of 2)

• The evidence from within-class comparisons of
  FGAs with FGAs is insufficient to permit
  conclusions about effect sizes and statistical
  significance.
• In between-class comparisons of FGAs with SGAs
• Olanzapine and risperidone are more effective
  than haloperidol for reducing autistic symptoms
  (anger, hyperactivity, and Aberrant Behavior
  Checklist scores) in pervasive developmental
  disorders.
• Strength of Evidence Low
• In pooled analyses, SGAs (clozapine, olanzapine,
  and risperidone) are more effective than
  haloperidol in treating schizophrenia, as
  assessed by clinical global impressions, but not
  by effects on positive and negative symptoms.
• Strength of Evidence Low

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

29
Summary of Comparative Effectiveness Results (2
of 2)

• In within-class comparisons of second-generation
  antipsychotics with each other
• Olanzapine is not statistically different from
  risperidone or clozapine in treating
  schizophrenia, as assessed by clinical global
  impressions and positive and negative symptoms.
• Strength of Evidence Low

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

30
Adverse Effects of Antipsychotics In Children

• Adverse events were examined across all
  indications in total adverse events are not
  expected to be dependent on diagnosis.
• Several adverse events of interest had no data
  reported.
• The analysis focused on key adverse event
  categories reported in at least one study
• Weight and body composition
• Blood lipids
• Blood glucose, insulin, and diabetes
• Neuromotor effects (e.g., extrapyramidal
  symptoms, akathisia, tardive dyskinesia)
• Prolactin and related sexual adverse effects
• Sedative effects

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

31
Adverse Effects First-Generation Antipsychotics

• Placebo Comparisons
• There are few comparisons of first-generation
  antipsychotics (FGAs) with placebo that report
  differences in adverse events in children.
• The evidence is insufficient for all key adverse
  events.
• FGAs versus FGAs
• The evidence is insufficient for all head-to-head
  comparisons of FGAs.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

32
Adverse Effects FGAs Compared With SGAs

• FGAs versus SGAs
• The evidence from comparisons of FGAs as a class
  versus SGAs as a class, and for FGAs compared
  with the SGA clozapine, is insufficient to permit
  conclusions.
• Comparisons of haloperidol with two SGAs,
  olanzapine and risperidone, yielded the following
  results in a meta-analysis

Comparison Weight (95 CI) n studies, n participants Extrapyramidal Symptoms (95 CI) n studies, n participants Sedation (95 CI) n studies, n participants
Haloperidol vs. Olanzapine 5.79 kg greater with olanzapine (from 3.0 to 8.6 kg higher) 4, 132 ??? 3.53-fold greater risk with haloperidol (from 1.1-fold to 10.9-fold) 3, 62 ??? Not reported
Haloperidol vs. Risperidone NSD (statistically valid range is from 8.9 kg lower with haloperidol to 16.76 kg greater with haloperidol) 3, 130 ??? No meta-analysis. Two different scales used both show statistically significant, greater severity of extrapyramidal symptoms with haloperidol. 1, 34 ??? 6- fold greater risk of fatigue with haloperidol (from 1.5-fold to 24.2-fold) 1, 24
Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ???
Seida J, Schouten J, Mousavi S, et al.
Comparative Effectiveness Review No. 39.
Available at www.effectivehealthcare.ahrq.gov/peda
ntipsych.cfm.
33
Summary of Adverse Effects ofFirst-Generation
Antipsychotics

• For comparisons between FGAs or with placebo, the
  evidence is insufficient to permit conclusions
  about the degree of risk or severity of adverse
  effects.
• In direct comparisons of the FGA haloperidol with
  SGAs
• When compared with olanzapine
• Haloperidol is associated with a lower risk for
  adverse effects on weight and body composition
  but with a greater risk of extrapyramidal
  symptoms.
• No statistically significant differences are
  noted for prolactin-related measures or sedation.
• Strength of Evidence Low
• When compared with risperidone
• Haloperidol is associated with greater severity
  of extrapyramidal symptoms.
• No statistically significant differences in risk
  are noted for adverse effects on weight and body
  composition or prolactin-related effects.
• Strength of Evidence Low

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

34
Adverse Effects Second-Generation Antipsychotics
Versus Placebo (1 of 2)

• The risk and severity of key adverse events was
  examined in a meta-analysis of individual SGAs
  compared with placebo, in pediatric use. SGAs
  showed these effects

Adverse Effects Relative Risk or Mean Difference (95 CI) Relative Risk or Mean Difference (95 CI) Relative Risk or Mean Difference (95 CI) Relative Risk or Mean Difference (95 CI) Relative Risk or Mean Difference (95 CI)
Adverse Effects Aripiprazole Olanzapine Quetiapine Risperidone Ziprasidone
Dyslipidemia Elevated cholesterol Risk is elevated 2.5x (range from 1.4x to 4.4x ) NNH 4 ??? Elevated cholesterol Risk is elevated 10x (range from 1.4x to 73.2x) NNH 6 ??? Triglyceride level Triglycerides are 29.1 mg/dL higher (range from 7.3 to 50.9 mg/dL higher) ??? Insufficient ??? No Data
Weight Gain Weight is 0.77 kg greater (range from 0.4 to 1.2 kg) ??? Weight is 4.60 kg greater (range from 3.07 to 6.13 kg) ??? Weight is 1.8 kg greater (range from 1.1 to 2.5 kg) ??? Weight is 1.8 kg greater (range from 1.5 to 2.1 kg) ??? NSD ???
Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Most studies that measured changes in weight ranged from 4 to 8 weeks duration, but some were in the range of 6 months to 1 year. Dyslipidemia includes abnormalities in total cholesterol and triglycerides. Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Most studies that measured changes in weight ranged from 4 to 8 weeks duration, but some were in the range of 6 months to 1 year. Dyslipidemia includes abnormalities in total cholesterol and triglycerides. Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Most studies that measured changes in weight ranged from 4 to 8 weeks duration, but some were in the range of 6 months to 1 year. Dyslipidemia includes abnormalities in total cholesterol and triglycerides. Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Most studies that measured changes in weight ranged from 4 to 8 weeks duration, but some were in the range of 6 months to 1 year. Dyslipidemia includes abnormalities in total cholesterol and triglycerides. Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Most studies that measured changes in weight ranged from 4 to 8 weeks duration, but some were in the range of 6 months to 1 year. Dyslipidemia includes abnormalities in total cholesterol and triglycerides. Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Most studies that measured changes in weight ranged from 4 to 8 weeks duration, but some were in the range of 6 months to 1 year. Dyslipidemia includes abnormalities in total cholesterol and triglycerides.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

35
Adverse Effects Second-Generation Antipsychotics
Versus Placebo (2 of 2)
Adverse Events Relative Risk or Mean Difference (95 CI) Relative Risk or Mean Difference (95 CI) Relative Risk or Mean Difference (95 CI) Relative Risk or Mean Difference (95 CI) Relative Risk or Mean Difference (95 CI)
Adverse Events Aripiprazole Olanzapine Quetiapine Risperidone Ziprasidone
EPS Risk is elevated 4.2x (from 2.4x to 7.2x) NNH 4 ??? Not Reported NSD ??? Risk is elevated 2.7x (from 1.4x to 4.9x) NNH 15 ??? Risk is elevated 10.3x (from 1.4 x to 75x) NNH 9 ???
Somnolence Risk is elevated 2.7x (from 1.1x to 6.5x) NNH 10 ??? NSD ??? Risk is elevated 3.4x (from 2.0x to 5.8x) NNH 4 ??? Risk is elevated 2.9x (from 1.5x to 5.5x) NNH 4 ??? Risk is elevated 3.0x (from 1.7x to 5.2x) NNH 7 ???
Prolactin Levels 4.1 ng/mL lower (from 1.8 ng/mL to 6.3 ng/mL) ??? 11.5 ng/nL higher (from 8.8 ng/mL to 14.1 ng/mL) ??? NSD ??? 22.6 ng/mL higher (from 10.7 to 34.5 ng/mL) ??? Insufficient ???
Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) EPS extrapyramidal symptoms NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) EPS extrapyramidal symptoms NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) EPS extrapyramidal symptoms NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) EPS extrapyramidal symptoms NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) EPS extrapyramidal symptoms NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? 95 CI 95-percent confidence interval (the range of statistically valid results) EPS extrapyramidal symptoms NNH number needed to harm (the number of patients to be treated to affect one patient more than in the control group) NSD no statistically significant difference

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

36
Adverse Effects SGAs Versus SGAs (1 of 2)

• The risk and severity of key adverse events were
  examined in a meta-analysis of individual SGAs
  compared with other SGAs in head-to-head
  evaluations in pediatric use. The analysis found
  the following summary effects and statistically
  valid range of values for the effects, as defined
  by the 95-percent confidence interval (95 CI)

Comparison Dyslipidemia Weight
Olanzapine vs. Quetiapine The risk with olanzapine is 3.5 times that of quetiapine (from 1.1x to 11.2x). ??? No summary estimate.
Olanzapine vs. Risperidone Triglycerides are 17.3 mg/dL higher, on average, with olanzapine (from 3.5 to 31.1 mg/dL higher). ??? Weight is 2.39 kg greater, on average, with olanzapine (from 1.5 kg to 3.3 kg greater). ???
Olanzapine vs. Aripiprazole The risk with olanzapine is 4 times that of aripiprazole (from 1.25x to 12.5x). ??? Weight is 4.1 kg greater, on average, with olanzapine (from 2.7 kg to 5.5 kg greater). ???
Aripiprazole vs. Quetiapine Triglycerides are 39.4 mg/dL lower with aripiprazole (from 7.4 to 71.3 mg/dL lower). ??? Weight is 1.62 kg greater, on average, with quetiapine (from 0.3kg to 3.0 kg greater). ???
Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ??? Strength of Evidence Symbols High ??? Moderate ??? Low ??? Insufficient ???

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

37
Adverse Effects SGAs Versus SGAs (2 of 2)

• For other adverse events in direct comparisons,
  there is limited evidence of no statistically
  significant difference between second-generation
  antipsychotics (SGAs) for effects on insulin
  levels and glucose control, extrapyramidal
  symptoms, and sedative effects.
• Strength of Evidence Low
• The risk of elevated prolactin is 2.6 times
  greater with risperidone than with olanzapine
  (from 1.7x to 5x).
• Strength of Evidence Moderate
• The evidence from other head-to-head comparisons
  is insufficient to permit conclusions about
  differences in prolactin levels.
• Study durations were typically too short to
  evaluate adverse effects on some important
  outcomes such as insulin and glycemic control.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

38
Summary of Adverse Effects ofSecond-Generation
Antipsychotics

• Current evidence indicates that there are
  differences between individual second-generation
  antipsychotics (SGAs) in the risk and severity of
  key adverse effects.
• However, in the absence of direct comparisons,
  only subjective interpretations of relative risk
  and severity can be made using data from placebo
  comparisons.
• In the current evidence base of direct
  comparisons of SGAs, statistically significant
  differences are noted in rate or severity of
  dyslipidemia and in adverse changes in weight and
  body composition.
• Among the SGAs, olanzapine exhibits the most
  severe adverse effects on weight and blood
  lipids, and risperidone has the strongest effect
  for elevating prolactin levels, both in placebo
  comparisons and when compared with olanzapine.
• There is limited evidence of no statistically
  significant difference between SGAs in other
  direct comparisons of adverse effects (insulin
  and glucose control, extrapyramidal symptoms, and
  sedative effects).

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

39
Conclusions About the Benefits and Adverse
Effects of Antipsychotics (1 of 2)

• Evidence about the use of antipsychotics in
  children and adolescents is inadequate to support
  strong conclusions about their comparative
  effectiveness.
• Moderate-strength evidence indicates that
  second-generation antipsychotics (SGAs) as a
  class improve clinical global impressions in
  bipolar disorder, and low-strength evidence
  supports benefits for treating mania.
• Moderate-strength evidence shows that SGAs as a
  class improve both clinical global impressions
  and positive and negative symptoms of
  schizophrenia.
• Moderate-strength evidence shows that risperidone
  is effective for attention deficit hyperactivity
  disorder and disruptive behavior disorders, and
  that risperidone and ziprasidone can reduce tics
  in Tourettes syndrome.
• Limited evidence suggests that SGAs are more
  effective than first-generation antipsychotics
  for improving some autistic symptoms of pervasive
  developmental disorders.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

40
Conclusions About the Benefits and Adverse
Effects of Antipsychotics (2 of 2)

• Adverse effects of second-generation
  antipsychotics (SGAs) include extrapyramidal
  symptoms (EPS), somnolence, weight gain,
  dyslipidemia, and elevated prolactin levels.
• In head-to-head comparisons of SGAs, the risk and
  severity of abnormalities of weight and blood
  lipids are greatest with olanzapine.
• Risperidone raises prolactin levels more than
  olanzapine.
• There is low-strength evidence of no differences
  between SGAs in effects on insulin and glucose
  control, EPS, and sedation.
• The long-term safety of both first-generation
  antipsychotics (FGAs) and SGAs and their
  effectiveness for improving quality-of-life
  outcomes are not established.
• Although SGAs have been perceived as having fewer
  side effects than FGAs, data are very limited to
  compare the relative risks of adverse effects.
  The spectrum of adverse effects should be taken
  into account, along with possible alternatives,
  when considering use of these drugs.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

41
Gaps in Knowledge (1 of 2)

• The effectiveness review revealed areas where the
  evidence about the effectiveness of
  first-generation (FGAs) and second-generation
  (SGAs) in treating pediatric psychiatric
  disorders is limited or absent, including
• Few head-to-head comparisons of FGAs and SGAs
  exist, either within or between classes, to
  demonstrate their effectiveness, benefits, and
  adverse effects for use in pediatric and young
  adult populations.
• No studies were found that reported pediatric use
  of antipsychotics to treat obsessive-compulsive
  disorder, post-traumatic stress disorder, or
  anorexia nervosa.
• Studies of young adults (ages 1924) were rare.
• Few studies reported outcomes that are important
  to patients (e.g., health-related quality of
  life, school performance, and legal
  interactions), and there is no consensus on the
  minimal clinically important effects to be
  produced by treatments.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

42
Gaps in Knowledge (2 of 2)

• Evidence about efficacy and safety over several
  years is unavailable.
• Standardized scales and methods for
  systematically investigating adverse events are
  needed.
• How the characteristics of key patient
  subpopulations affect patient-centered outcomes
  is not understood.
• Large-scale effectiveness studies that apply few
  patient restrictions and closely match typical
  clinical practice are needed to inform clinical
  decisionmaking.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.

43
What To Discuss With Your Patients andTheir
Caregivers

• The role that antipsychotics may play as one
  component in the broader array of treatments for
  child and adolescent psychiatric disorders.
• The ability of both first-generation and
  second-generation antipsychotics to improve
  symptoms of psychiatric disorders in children,
  and the differences in strength of evidence for
  benefits in particular indications.
• The risks of extrapyramidal effects, weight gain,
  and blood lipid abnormalities, and the evidence
  about differences in risk among drugs in both
  classes.
• The limited evidence about long-term benefits and
  adverse effects on health and quality of life.

• Seida J, Schouten J, Mousavi S, et al.
  Comparative Effectiveness Review No. 39.
  Available at www.effectivehealthcare.ahrq.gov/peda
  ntipsych.cfm.