M Taylor, M Turner, L Watt, D Brown, M Martin, K Fraser
Correspondence: Dr Mark Taylor, Springpark Centre, Glasgow G22 5EU. Email: mark.taylor@glacomen.scot.nhs.uk
SMJ 2005 50(3): 102-106
Abstract
Background: There are few real life independent comparative studies of atypical antipsychotics. We prospectively examined five commonly used atypical antipychotics in the UK, without support from the pharmaceutical industry. Method: Prospective naturalistic systematic clinical evaluation. Patients being newly prescribed atypical anti-psychotics over a one year period were assessed by psychiatrists at initiation and after six months treatment using five outcome measures: clinical global impression; positive and negative psychotic symptoms; drug related side effects; and quality of life. Results: 373 patients participated in total. Olanzapine and risperidone produced statistically significant reductions in all ratings at six months. Amisulpride, clozapine, and quetiapine were also studied. There was limited variance between the different drugs, although some sample sizes were small. Conclusion: Atypical anti-psychotics were found to be clinically effective, and produced similar outcomes. Routine monitoring of outcomes in psychiatry is feasible.
Key words: Atypical; anti-psychotics
Introduction
Newer or ‘atypical’ anti-psychotic drugs, modelled on clozapine, represented an evolution in the treatment of schizophrenia. Equal or greater efficacy compared to older anti-psychotics, and reduced extrapyramidal sideeffects are advantages balanced by a greater financial cost. Atypical anti-psychotics (‘atypicals’) now command widespread clinician confidence and an expert consensus on the use of atypicals in the US has recently been published.1 However there are few long term head-to-head studies of comparative efficacy and safety, and fewer still that are independent of industry support. Perhaps unsurprisingly commercially supported comparative studies2,3 tend to be outcome-neutral or favour the drug produced by the sponsor.
Demonstrating efficacy in randomised controlled trials (RCTs) is not the same as showing effectiveness in routine clinical practice, where dual diagnosis or comorbidity, and lack of adherence are often the norm. Systematic allinclusive open label studies can complement RCTs, and a local demonstration of effectiveness can be used to help secure funding for new treatments. A validated system with a scale based on the Clinical Global Impression had previously been developed4 and the same methodology was used to compare outcomes in five major atypical antipsychotics throughout the city of Glasgow, UK .
Method
Subjects
All patients from adolescent, adult, and old age psychiatry in the Greater Glasgow area (population ~ 1.0 million ) with a clinical diagnosis (from a senior psychiatrist) of schizophrenia or schizophreniform disorder, and who were prescribed either amisulpride, clozapine, olanzapine, quetiapine, or risperidone were prospectively recruited into the study. The study recruitment ran for one year starting in 2002, with a further six months for follow-up assessments. As this study was viewed as part of normal clinical practice, ethical approval and informed consent were not sought, and all personal data was anonymised. Participation in the trial did not affect treatment choice or delivery in any way.
Measures
Demographic and clinical information (including one or more of five pre-specified reasons for initiating the antipsychotic) was documented by the prescribing clinician. A standardised assessment form with five linear analogue scales and referenced anchor points was also completed. The scales comprised the Clinical Global Impression (CGI) scale (score range 0 to 7), as well as an assessment of the positive and the negative symptoms of schizophrenia, drug related adverse effects and impairment of quality of life (all with score range 0 to 4). For all assessments the score increases with severity of symptoms. Where possible, the same clinicians re-assessed the patients still on their medication at six months. Previously ascertained 4 inter-rater reliability weighted kappa scores for the five outcome scales were 0.60 for the CGI, 0.44 for positive symptoms, 0.39 for negative symptoms, 0.69 for drug-related side effects, and 0.75 for quality-of-life impairment.
Statistics
Scores at baseline and six months were compared with the non-parametric Wilcoxon matched-pairs signed-ranks test, but mean scores are presented here for ease of comprehension. Improvement after six months treatment is presented as a percentage change from mean baseline rating. Change across the five treatment groups was analysed using the Kruskal-Wallis test.
Results
Three hundred and seventy three patients (192 men, 181 women) were enrolled with a mean age of 45.9 years (range = 14 - 99). These were individuals being treated in public sector in-patient and out-patient settings. The enrolment represented an uptake rate of 51% by psychiatrists (35 in total) who were newly prescribing atypical antipsychotic medication within a one year period in Glasgow. This 51% of subjects were evenly distributed across geographic and specialty areas within Glasgow, and hence is felt to be representative of the whole population.
There was a 64% completion rate of returns at the six month review point (n=157) or at discontinuation (n=81), with the remaining 135 subjects being lost to follow up. Those subjects lost to follow up were evenly represented among the different treatment groups.
Criteria for patient selection as indicated by clinician (n=373): The clinical indications for specific drug selection allowed were intolerable side effects (including extra-pyramidal symptoms or EPS) from traditional antipsychotics, minimisation of side effects (SE) from the outset of treatment as a priority, marked negative symptoms, refractory schizophrenia, being a first episode case, and other. Indicating more than one of these criteria was allowed. Table I details the results.
Baseline assessments (n=373). The mean scores for all baseline ratings of impairment or disability are given in Table II. Mean CGI scores ranged from 4.0 to 5.3 (moderate to markedly ill), although there was the expected difference between clozapine (the only drug licensed in the UK for treatment resistant schizophrenia) and the other four medications.
The other ratings were: positive symptoms mean scores between 1.9 and 2.8 (2 = moderate and 3 = marked pathology); negative symptoms mean scores were 0.8 to 1.7 (1 = mild to 2 = moderate severity); drug-induced side effects had mean scores between 1.0 and 2.0 (1= mild and 2= moderate disability); impairment in quality of life had mean scores from 2.9 to 3.6 (3 = moderate, 4 = severe).
Dose:
The mean daily dose at six month review for amisulpride (n=16) was 487.5mg, for clozapine (n=12) was 429 mg, for olanzapine (n=65) was 13.7 mg, for quetiapine (n=8) was 350 mg, and for risperidone (n=56) was 3.4 mg.
Comparison of baseline and follow-up assessments (total n=157).
After six months treatment, individual ratings by the treating psychiatrist ranged from deterioration through unchanged to improvement. Mean ratings for each the five treatment groups were improved at the six month review point, and the comparative mean improvement in the five rating scales for the five treatment groups are presented in Table III, along with percentage change from baseline. There was no significant difference (by Kruskal-Wallis) between the five groups of medication across the five different rating measures, after six months of treatment. For example, differential improvement in CGI between the five treatments, p=0.095.
Dropouts.
Eighty-one patients were documented as discontinuing treatment within six months. More than one reason for discontinuation could be specified by the treating psychiatrist. The specified reasons are listed in Table IV.
Discussion
Main findings
Most clinicians in the study were choosing atypical antipsychotics (or ‘atypicals’) due to concern over side effect profile. In one Glasgow catchment or sector (ref. personal communication, Dr Taylor), 52% of all individuals with schizophrenia receive atypicals. Olanzapine and risperidone are the two most favoured atypicals in this study, perhaps due to their longer time since licensed release in the UK compared to amisulpride and quetiapine. Individuals prescribed clozapine had noticeably higher levels of psychopathology at baseline and less improvement in side effects after six months treatment compared to the four other groups. Patients being initiated on amisulpride, olanzapine, quetiapine, and risperidone had remarkably similar mean levels of observer rated psychopathology at baseline assessment.
The discontinuation rate (81/238 or 34%) of atypical therapy is not dissimilar to dropout rates of 22% to 47% reported by Cochrane in Leucht et al.5 The reasons for discontinuation in everyday practice are usually several and medication non compliance or adherence is not necessarily linked to experience of adverse side-effects, as illustrated in Table IV.
Some of the mean improvement scores at six month review do not achieve statistical significance, although for amisulpride, clozapine, and quetiapine this could be due to a lack of power secondary to the low numbers of patients still on these treatments at six month review. It is clear that all five atypicals studied produce clinically observable improvements in the five global ratings of pathology. Improvements in the clinical global impression (CGI) are mirrored by improvements in positive symptoms and quality of life, rather than the other two measures. Quality of life assessment involved consideration of time utilisation and activities of daily living, and occupational or social role fulfilment. Differences (trends) in outcome between the five treatments after six months are evident from Table III, but no statistically significant difference between the drugs for any of the five outcome measures was detected. This is interesting in view of the chemically disparate nature of the five treatments studied. It should also be remembered that by definition, clozapine was being used in patients who were already treatment resistant and thus the therapeutic gains documented might have been harder to achieve.
Related studies
Some naturalistic outcome data are available for clozapine and risperidone, and olanzapine.6,7 Ho et al8 found statistically significant reductions on ratings of positive psychotic symptoms, total symptoms, a global assessment and some measures of quality-of-life for in-patients who remained on olanzapine and risperidone at six month follow-up. Gilchrist et al4 used the same methodology as the present study to demonstrate the clinical effectiveness of olanzapine in general and forensic settings, at a mean dose of 15mg and 17mg respectively.
In a meta-analysis, Leucht et al9 concluded that risperidone and olanzapine were slightly more effective, and quetiapine as effective as haloperidol in global impression (CGI). This mirrors our comparative mean improvement scores in CGI at six months. Furthermore, Leucht et al found that quetiapine was comparatively less effective in treating negative symptoms than olanzapine or risperidone, and he later5 highlighted the role of amisulpride in treating negative symptoms.
A recent systematic review10 compared clozapine to other atypicals, and concluded that outcomes with clozapine were similar to other atypicals using a CGI-like measure, but better for positive and worse for negative symptoms. Our results are not dissimilar but again bear the caveat of low numbers within the clozapine group. Karow and Naber11 emphasise quality of life as being of major clinical relevance, and in a review of open and controlled studies found that atypicals significantly increase quality of life for the patient. The findings of the present study add weight to that conclusion.
Strength and limitations of this report
A naturalistic design was adopted for this study. Data were collected prospectively on a consecutively recruited sample of patients with a clinical diagnosis of schizophrenia or schizophreniform psychosis. Symptom severity was measured with a simple compound scale based on standardised psychiatric rating scales with proven reliability. Importantly, all patients with schizophrenia-type psychosis were included regardless of any additional diagnoses or problems. The findings, therefore, complement those reported from randomised control trials, where the necessary use of inclusion and exclusion criteria mean the results may not be applicable to all patients, especially those who are involuntarily detained, liable to pregnancy, or with substance misuse co-morbidity. We also feel our study is not liable to sponsorship bias.
We must acknowledge the limitations of the study. We did not conduct standardised psychiatric interviews in all subjects and we cannot therefore be sure that they meet diagnostic criteria for schizophrenia, although the subjects are likely to be representative of patients given that clinical diagnosis by experienced psychiatrists. We previously demonstrated 4 that the scales we used had satisfactory interrater reliability at one point in time, but we cannot be sure that all the raters used in the study would all have agreed with each other to an acceptable extent. Although the raters were not specifically trained in the use of the scales, explicit anchor descriptions were provided for each point of each scale, and the CGI scale is widely used in research and clinic settings. Our kappa values of 0.4 – 0.75 are similar to the typical values of 0.5 for most of the components of the physical examination in general medical practice 12. Indeed, in studies such as this one has to balance what is desirable with what is practical, and the use of potentially different raters for particular patients is analogous to clinical situations where one doctor has to assess the effect of a drug prescribed by another. The use of some form of routine outcome rating is likely to provide more reliable information than a variable standard of case note documentation based on components of a typical mental state examination, which may itself be even less reliable than a typical physical examination.13
Conclusions
We have found that in an open label real-world study, five of the most commonly prescribed atypical antipsychotics are clinically effective and generally well tolerated. Minor differences in effect appear to exist between these chemically disparate compounds, but our results highlight the similarities rather than the differences in outcome between the five atypicals studied (cf. Tandon and Jibson14). Secondly, we have demonstrated the practical feasibility of routinely measuring patient outcomes in psychiatric settings. The impact of new treatments can be systematically evaluated in this manner, and influence medication budgets. The information generated can also inform the development of local treatment guidelines. Clinical effectiveness, and naturalistic outcome studies cannot replace randomised controlled trials, but nor are they obviated by them. They are complementary and both are necessary.
Declaration of Interest:
This study was not commercially funded. All authors have attended meetings as guests of various pharmaceutical companies. MT, MT, and DB have received honoraria from Astra Zeneca; Janssen; Lilly, Novartis, and Sanofi.
ACKNOWLEDGEMENTS: We thank all medical, nursing, pharmacy and other staff involved in this study. In particular, we thank Aileen Smith and Michele Robertson.
REFERENCES
1 Kane JM, Leucht S, Carpenter D, Docherty JP (eds.) The Expert Consensus Guideline Series: Optimising the pharmacologic treatment of psychotic disorders. J Clin Psychiatry 2003; 64 (suppl 12) 1-98.
2 Conley RR and Mahmoud R. A randomised double blind study of risperidone and olanzapine in the treatment of schizophrenia or schizoaffective disorder. Am J Psychiatry 2001; 158: 765-774.
3 Tran PV, Hamilton SH, Kuntz AJ, et al. Double blind comparison of olanzapine versus risperidone in the treatment of schizophrenia and other psychotic disorders. J Clin Psychopharmacology 1997; 17:407-18.
4 Gilchrist A, Taylor M, Wright M, Drury K, Short A, Steel R, Murray P and Lawrie S. Is olanzapine clinically effective? A naturalistic outcome survey in two hospital settings. Pharmaceutical Journal 2002. 269, 222-5.
5 Leucht S, Pitschel-Walz G, Engel RR, Kissling W. Amisulpride, an unusual “atypical” antipsychotic: a meta-analysis of randomised controlled trials. Am J Psychiatry 2002; 159: 180-90.
6 Kasper S, Rosillon D and Duchesne I on behalf of the RODOS Investigator group. Risperidone Olanzapine Drug Outcomes studies in Schizophrenia (RODOS): efficacy and tolerability results of an international naturalistic study. Int Clin Psychopharmacology 2001; 16: 179-87.
7 Jones PB, Lepine JP, and Pancheri P, eds. European Schizophrenia Outpatient Health Outcomes (SOHO) study. Acta Psychiatr Scand. 2003; 107 (Suppl 416).
8 Ho BC, Miller D, Nopoulos P, et al. A comparative effectiveness study of risperidone and olanzapine in the treatment of schizophrenia. J Clin Psychiatry 1999; 60:658-63.
9 Leucht S, Pitschel-Walz G, Abraham D, and Kissling W. Efficacy and extrapyramidal side-effects of the new antipsychotics olanzapine, quetiapine, risperidone, and sertindole compared to conventional antipsychotics and placebo. A meta-analysis of randomised controlled trials. Schiz Research 1999; 35 51-68.
10 Tuunainen A, Wahlbeck K, and Gilbody S. Newer atypical antipsychotic medication incomparison to clozapine: a systematic review of randomised trials. Schiz Research 2002; 56, 1-10.
11 Karow A and Naber D. Subjective well being and quality of life under atypical antipsychotic treatment. Psychopharmacology 2002; 162 (1): 3-10.
12 Sackett DL, Haynes RB, Guyatt GH, et al. Clinical Epidemiology. A basic science forclinical medicine. eds Boston: Little, 1991. Brown & Co.
13 Mojtabai R, Nicholson RA. (1995). Interrater reliability of ratings of delusions and bizarre delusions. Am J Psychiatry 1995; 152:1804-6.
14 Tandon R, and Jibson MD. Efficacy of newer generation antipsychotics in the treatment of schizophrenia. Psychoneuroendocrinology 2003; 28 9-26.