Author
SMITH M.; et al.;
Title
First- v. second-generation antipsychotics and risk for diabetes in schizophrenia: systematic review and meta-analysis.
Journal citation/publication details
British Journal of Psychiatry, 192(6), June 2008, pp.404-411.
Summary
Fourteen studies met the inclusion criteria for this review, including 11 suitable for meta-analysis. They indicate that second-generation psychotics are associated with a small additional risk for diabetes in people with schizophrenia. Although most of the studies were methodologically limited, and the apparent association between second-generation drugs and diabetes may be smaller than reported, clinicians are advised to screen schizophrenia patients for physical illnesses (including diabetes) and review medication in line with its likely impact on both psychiatric symptoms and the risk of diabetes.
Context
Diabetes has been reported in some studies to be two or three times more common among people with schizophrenia than in the general population. The reason is not well understood but is likely to include a number of factors, of which the effects of antipsychotic drugs may be one. To date, there has been no systematic review of studies of the association between diabetes and type of antipsychotic medication.
Methods
What sources were used?
The following databases were searched up to September 2006: CINAHL (Cumulative Index to Nursing and Allied Health Literature), from 1982; EMBASE (Excerpta Medica), from 1980; International Pharmaceutical Abstracts, from 1970; Medline, from 1966; PsycINFO, from 1967; and Web of Knowledge, from 1981. The proceedings of the 2000-2005 conferences of four diabetes organisations and two psychiatric organisations (all named) were also searched, as well as the reference lists of included studies and reviews. Study authors and unnamed experts in the field were contacted for further information on published and unpublished studies.
What search terms/strategies were used?
The search was conducted ‘using established Cochrane Collaboration search strategies and definitions for schizophrenia, antipsychotics and diabetes mellitus’ and details are available from the authors on request.
What criteria were used to decide on which studies to include?
The inclusion criteria were wide, covering cross-sectional, case-control, cohort or controlled trials of populations including children or adults with schizophrenia or related disorders. Studies were required to compare second-generation antipsychotics (defined as clozapine, olanzapine, risperidone and quetiapine) with first generation antipsychotics as listed in the British National Formulary (2005), and to measure diabetes as a primary or secondary outcome.
Who decided on their relevance and quality?
The searches delivered 1,974 studies, which were initially screened by the lead author. One foreign language paper was translated by a doctor who was a native speaker. Responsibility for further filtering is not specified but the overall process is summarised in a flow diagram, which includes reasons for the exclusion of studies in the later stages. Quality assessment was conducted using an adapted version of the Meta-analysis of Observational Studies in Epidemiology (MOOSE) and Strengthening of the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. This is specified in full and included eight components characterising a high quality study. Details of who carried out the assessments are not provided.
How many studies were included and where were they from?
Fourteen studies meeting the inclusion criteria and characterised as high quality were reviewed, of which 11 provided sufficient data for meta-analysis. The studies are referenced at the end of the paper, and also included as an online table for those with subscription access to the British Journal of Psychiatry. This shows that ten of the studies were conducted in the USA, and the remainder in Sweden, Korea, Turkey and the Netherlands.
How were the study findings combined?
Study data were coded using a standardised data extraction sheet, with at least two attempts made to retrieve missing data from study authors. For the 11 with sufficient data for statistical synthesis, information was pooled using random effects inverse variance weighted meta-analysis in the software package Stata (version 9 for Windows).
Findings of the review
The majority (10) of the 14 studies were hospital-based, and most are described as ‘retrospective or pharmacoepidemiological using large databases’ which are known to suffer from problems with the quality of clinical data recorded in them. Seven studies were funded by the pharmaceutical industry. Most were cross-sectional (4), retrospective cohort (4) or case control (2) designs, with only three using a prospective design. Convenience or consecutive samples were the norm, with only two using random selection, and methods of diagnosing schizophrenia and diabetes varied. Most studies did not adjust for confounding variables in methods or analysis, or adjust for all risk factors for diabetes.
The evidence base, therefore, is admittedly of limited quality: no study fulfilled all eight criteria of high quality and five fulfilled no more than four. The meta-analysis showed an overall relative risk of a diagnosis of diabetes in those prescribed a second-generation antipsychotic of 1.32, compared with being prescribed a first-generation psychotic. However, there was considerable variation between the studies. The relative risks for the particular drugs of interest were as follows: risperidone 1.16 (six studies); quetiapine 1.28 (three studies); olanzapine 1.28 (eight studies); and clozapine 1.39 (seven studies).
Because most of the studies were of cross-sectional or retrospective cohort design, it was not possible to determine the direction of association between second-generation antipsychotics and diabetes. Moreover, the methodological inadequacies of the studies mean that any detected association is ‘likely to be significantly biased’.
Authors' conclusions
‘Our review found that second-generation antipsychotics were associated with a 30% increased risk of diabetes compared with first-generation antipsychotics in people with schizophrenia – this is probably a biased observation with the true association being smaller.’
Implications for policy or practice
‘Our review emphasises that, to date, the evidence is very poor and should not be used alone as a guideline for, or switching, antipsychotic medications or implementing diabetes screening and management protocols for schizophrenia until further evidence comes to light.’ However, clinicians are advised to implement protocols for identifying physical illnesses, especially diabetes, in people with schizophrenic conditions, and for reviewing medication both in relation to the treatment of psychiatric symptoms and a possible heightened risk of diabetes.