Article Text


Review: evidence on the comparative effectiveness and adverse effects of antipsychotics in young people is limited
  1. Pratibha N Reebye1,
  2. Dean Elbe2
  1. 1Children's & Women's Mental Health Programs, BC Children's Hospital, Vancouver, Canada
  2. 2Department of Pharmacy, BC Children's Hospital, Vancouver, Canada

Statistics from


Question: What is the evidence on the comparative effectiveness and safety of the antipsychotics for children and young adults with psychiatric and behavioural conditions?

Outcomes: Efficacy outcomes (disorder specific and non-specific symptoms; health-related quality of life; legal involvement; medication adherence; patient, parent- or care provider-reported outcomes; suicide-related behaviour); adverse events (dyslipidaemia, extrapyramidal symptoms, insulin resistance, prolactin-related and sexual side effects, sedation, weight and body composition).


Design: Systematic review and meta analysis.

Data sources: MEDLINE, Embase, CENTRAL, PsycINFO, IPA, CINAHL, Scopus, ProQuest Dissertations International, MedEffect Canada and TOXLINE, were searched from 1987 to February 2011. In addition, the authors hand searched clinical trial registers, reference lists of relevant reviews, proceedings from scientific meetings and Food and Drug Administration (FDA) reports and contacted drug manufacturers for published and unpublished study data.

Study selection and analysis: Controlled clinical trials (randomised and non-randomised) or cohort studies (prospective or retrospective) assessing FDA-approved antipsychotics in children and young people were included. Population: aged ≤24 years with one or more psychiatric and behavioural conditions (pervasive developmental disorder (PDD), disruptive behaviour disorder (DBD), bipolar disorder, schizophrenia or related psychosis, Tourette syndrome, obsessive compulsive disorder (OCD), post-traumatic stress disorder (PTSD), anorexia nervosa or other behavioural symptoms). Intervention: FDA-approved first-generation antipsychotics (FGAs: chlorpromazine, droperidol, fluphenazine, haloperidol, loxapine, perphenazine, pimozide, prochlorperazine, thiothixene, thioridazine or trifluoperazine) or second-generation antipsychotics (SGAs: aripiprazole, asenapine, clozapine, iloperidone, olanzapine, paliperidone, quetiapine, risperidone or ziprasidone). Comparators: active comparators (FGA or SGA), placebo or different doses of the same antipsychotic. Data were meta-analysed using random effects models.

Main results

Overall, 81 studies were included (62 randomised controlled trials (RCTs), two non-RCTs and 17 cohort studies). Head-to-head drug comparisons were made in 38 studies, different doses of the same antipsychotic were compared in 17 studies and placebo comparisons in 26 studies. Median follow-up was 8 weeks. The most frequently examined condition was schizophrenia (31 studies), followed by PDD (12 studies), and bipolar disorder (11 studies). None of the studies examined OCD, PTSD or anorexia nervosa. Most of the clinical trials had a high risk of bias (90%), inadequate sequence generation, allocation concealment, blinding and incomplete data. Few showed evidence of selective outcome reporting. The majority of the trials (78%) received industry funding. The cohort studies were of moderate quality. All comparisons of FGAs versus each other, SGAs, or placebo had low or insufficient strength of evidence to draw conclusions. Moderate strength of evidence was shown for the following SGA comparisons: compared with placebo, SGAs improved clinical global impressions in schizophrenia (six RCTs), bipolar disorder (six RCTs) and DBD (seven RCTs), reduced positive and negative symptoms in schizophrenia (six RCTs), improved behavioural symptoms in DBD (seven RCTs) and improved tics in Tourettes (two RCTs). Olanzapine caused more dyslipidaemia and weight gain but fewer prolactin-related events than risperidone. Olanzapine caused more weight gain than quetiapine. There were also some differences between the SGAs and placebo for adverse events, mainly showing increased adverse events with SGAs.


The evidence on the comparative effectiveness and adverse effects of antipsychotics for children and young adults is generally limited. Some second-generation antipsychotics have a better side effect profile in children and young adults than others.

Abstracted from

OpenUrlAbstract/FREE Full Text


In the context of ongoing widespread expansion of antipsychotic drug usage, Seida et al present a meticulous review of antipsychotic trials in children, adolescents and the oft-forgotten age group of young adults. Including this group is important, since research shows brain development continues until approximately age 25.1 This age group is in transition from childhood to adulthood and medication response and adverse effects may more closely resemble observed paediatric patterns, than in adults (up to age 65) with whom they are typically lumped together in clinical trials.

The authors reviewed 30 comparisons across six therapeutic categories covering major areas of antipsychotic use. In the majority of comparisons, a low strength of evidence (SOE) rating was assigned, and none of the comparisons had high SOE ratings. This is dismaying but not unexpected. Until recently, very few randomised controlled trials were conducted beyond the realm of attention deficit/hyperactivity disorder in paediatric psychopharmacology.

Space limitations notwithstanding, adverse events are presented by drug class on the assumption that adverse events will occur regardless of indication. However, this ignores the dose-related nature of most adverse effects (eg, antipsychotic dosing is typically higher in schizophrenia compared with disruptive behaviour disorders), and complex drug-disease interactions (eg, the observed association of metabolic syndrome with bipolar disorder2 could modify a drug's adverse effect profile in this population, but these factors may not be present in patients with disruptive behaviour disorders). 8 week median study duration is not long enough to fully assess the metabolic effects of these medications. Researchers can reduce potential for bias in future research by providing better descriptions of allocation concealment, blinding procedures and funding sources. While risk of bias in the trials was high due to industry funding (resulting in lower SOE ratings), one should consider there would be very little data to discuss, if industry funded research was not permitted.


View Abstract


  • Sources of funding The Agency for Healthcare Research and Quality (AHRQ) and US Department of Health and Human Services.

  • ▸ References are available online at


  • Competing interests None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.