Randomised controlled trials of conventional antipsychotic versus new atypical drugs, and new atypical drugs versus clozapine, in people with schizophrenia responding poorly to, or intolerant of, current drug treatment

Research output: Contribution to journalArticle

  • External authors:
  • P. B. Jones
  • T. R E Barnes
  • R. M. Murray
  • R. Kerwin
  • D. Taylor
  • A. Markwick
  • H. Lloyd


Objectives: To determine the clinical and cost-effectiveness of different classes of antipsychotic drug treatment in people with schizophrenia responding inadequately to, or having unacceptable side-effects from, their current medication. Design: Two pragmatic, randomised controlled trials (RCTs) were undertaken. The first RCT (band 1) compared the class of older, inexpensive conventional drugs with the class of new atypical drugs in people with schizophrenic disorders, whose current antipsychotic drug treatment was being changed either because of inadequate clinical response or owing to side-effects. The second RCT (band 2) compared the new (non-clozapine) atypical drugs with clozapine in people whose medication was being changed because of poor clinical response to two or more antipsychotic drugs. Both RCTs were four-centre trials with concealed randomisation and three follow-up assessments over 1 year, blind to treatment. Setting: Adult mental health settings in England. Participants: In total, 227 participants aged 18-65 years (40% of the planned sample) were randomised to band 1 and 136 (98% of the planned sample) to band 2. Interventions: Participants were randomised to a class of drug. The man aging clinician selected the individual drug within that class, except for the clozapine arm in band 2. The new atypical drugs included risperidone, olanzapine, quetiapine and amisulpride. The conventional drugs included older drugs, including depot preparations. As in routine practice, clinicians and participants were aware of the identity of the prescribed drug, but clinicians were asked to keep their participating patient on the randomised medication for at least the first 12 weeks. If the medication needed to be changed, the clinician was asked to prescribe another drug within the same class, if possible. Main outcome measures: The primary outcome was the Quality of Life Scale (QLS). Secondary clinical outcomes included symptoms [Positive and Negative Syndrome Scale (PANSS)], side-effects and participant satisfaction. Economic outcomes were costs of health and social care and a utility measure. Results: Recruitment to band 1 was less than anticipated (40%) and diminished over the trial. This appeared largely due to loss of perceived clinical equipoise (clinicians progressively becoming more convinced of the superiority of new atypicals). Good follow-up rates and a higher than expected correlation between QLS score at baseline and at follow-up meant that the sample as recruited had 75% power to detect a difference in QLS score of 5 points between the two treatment arms at 52 weeks. The recruitment to band 2 was approximately as planned. Follow-up assessments were completed at week 52 in 81% of band 1 and 87% of band 2 participants. Band 1 data showed that, on the QLS and symptom measures, those participants in the conventional arm tended towards greater improvements. This suggests that the failure to find the predicted advantage for new atypicals was not due to inadequate recruitment and statistical power in this sample. Participants reported no clear preference for either class of drug. There were no statistically significant differential outcomes for participants entering band 1 for reasons of treatment intolerance to those entering because of broadly defined treatment resistance. Net costs over the year varied widely, with a mean of £18,850 in the conventional drug group and £20,123 in the new atypical group, not a statistically significant difference. Of these costs, 2.1% and 3.8% were due to antipsychotic drug costs in the conventional and atypical group, respectively. There was a trend towards participants in the conventional drug group scoring more highly on the utility measure at 1 year. The results for band 2 showed an advantage for commencing clozapine in quality of life (QLS) at trend level (p = 0.08) and in symptoms (PANSS), which was statistically significant (p = 0.01), at 1 year. Clozapine showed approximately a 5-point advantage on PANSS total score and a trend towards having fewer total extrapyramidal side-effects. Participants reported at 12 weeks that their mental health was significantly better with clozapine than with new atypicals (p <0.05). Net costs of care varied widely, but were higher than in band 1, with a mean of £33,800 in the clozapine group and £28,400 in the new atypical group. Of these costs, 4.0% and 3.3%, respectively, were due to antipsychotic drug costs. The increased costs in the clozapine group appeared to reflect the licensing requirement for inpatient admission for commencing the drug. There was a trend towards higher mean participant utility scores in the clozapine group. Conclusions: For band 1, there is no disadvantage in terms of quality of life and symptoms, or associated costs of care, over 1 year in commencing conventional antipsychotic drugs rather than new atypical drugs. Conventional drugs were associated with non-significantly better outcomes and lower costs. Drug costs represented a small proportion of the overall costs of care (

Bibliographical metadata

Original languageEnglish
Pages (from-to)1-94
Number of pages93
JournalHealth Technology Assessment
Issue number17
Publication statusPublished - May 2006

Related information