Author
RAJJI T.K.; ISMAIL Z.; MULSANT B.H.;
Title
Age at onset and cognition in schizophrenia: meta-analysis.
Journal citation/publication details
British Journal of Psychiatry, 195(4), October 2009, pp.286-293.
Summary
This meta-analysis of data from 109 studies revealed different patterns of cognitive deficits between patients with youth-onset, late-onset and adult first-episode schizophrenia. Those with youth-onset and adult first-episode schizophrenia displayed large deficits in most cognitive measures while those with late-onset schizophrenia showed preserved function for many measures but greater deficits on others. There were a number of limitations to the meta-analysis including the variability in the definitions of the age of onset used in the studies and the aggregation of some cognitive tests into a single measure.
Context
In schizophrenia an earlier age of onset has been associated with more severe symptoms of the disorder, including cognitive deficits. The aim of the study was to examine, using meta-analysis, whether such associations exist across the life-span by estimating and comparing the severity of cognitive deficits in patients with adult-onset, youth-onset, and late-onset schizophrenia
Methods
What sources were searched?
A total of 29 electronic databases were searched including CINAHL, Embase, Medline, PsycINFO (see appendix for details).
What search terms/strategies were used?
Search terms to describe schizophrenia, age of onset and cognition are listed in the appendix. Databases were searched from 1980; searches were performed on 30 September 2008.
What criteria were used to decide on which studies to include?
Studies on schizophrenia and related disorders (schizoaffective, schizophreniform, delusional disorder, very-late-onset schizophrenia-like psychosis) which included a healthy control group and data on age of onset and cognitive measures were eligible for inclusion; drug trials with baseline cognitive data were also eligible. Studies in which participants with schizophrenia were included in the same sample as those with a comorbid diagnosis (such as bipolar disorder), or where samples were of mixed age of onset, were excluded. The definition of age of onset used was as described in each study.
Who decided on their relevance and quality?
A single named author reviewed the retrieved articles and their reference lists. There is no indication that the quality of the studies was formally assessed.
How many studies were included and where were they from?
Of the 913 publications identified from the searches 605 were reviewed in full and 109 were included in the analysis. Reasons for exclusion are presented in Fig.1 and the full list of publications included in the analysis is available online to subscribers at http://bjp.rcpsych.org/content/vol195/issue4/.
How were the study findings combined?
Participants were classified as youth-onset (maximum age at onset 19 years), late-onset (minimum age at onset 40 years), or first-episode schizophrenia (any other age at onset). Cognitive data was classified according to one of 22 measures identified by Heinrichs et al (1998). Effect sizes for each cognitive test were calculated using Cohen's d and effect size weighted means were generated. The three groups' weighted values were compared using the Q test. Nine late-onset, 23 youth-onset, and 78 first-episode onset studies were included in the meta-analysis giving a total of 5010 participants: 4057 first episode, mean age at onset 23.7 years; 692 youth-onset, mean age at onset 13.6 years; 261 late-onset, mean age at onset 60.7 years.
Findings of the review
Mean effect sizes (95% CI) of cognitive deficits are presented in Fig. 2 for each of the 22 measures and the three participant groups. Adults with first-episode schizophrenia and participants with youth-onset showed large deficits on most of the cognitive measures especially performance IQ, digit symbol coding, Tower of London and related tests, visual and verbal memory, vocabulary, arithmetic and fluency. Youth-onset participants demonstrated larger deficits on visuospatial construction, Stroop test and Trail Making Test B compared to those with adult first-episode. The late-onset group were more impaired on auditory and visual attention, fluency and visuospatial construction.
There were significant differences in the weighted mean effect sizes of the three groups for 19 out of the 22 measures (Table 2, Fig.3): the youth-onset group had greater deficits compared to the first-episode group on full-scale IQ, psychomotor speed of processing, Trail Making Test A and B, verbal spatial memory, and Wisconsin card tests; and were more impaired than the late-onset group on arithmetic, digit symbol coding, vocabulary and Wisconsin card tests but less impaired on auditory and visual attention. Adult first-episode participants were more impaired than late-onset on digit symbol coding, and less impaired on auditory attention, fluency, full scale IQ, global measure of cognition and visual attention. The differential pattern of cognitive impairments is also reflected in the comparative magnitudes of the weighted d-values for the three participant groups.
Authors' conclusions
'Strikingly different patterns of cognitive deficits' were demonstrated for the three groups of schizophrenia patients based on their age at onset.
Implications for policy or practice
Longitudinal and controlled studies are needed to distinguish the cognitive changes that are a result of the illness from those that are due to normal ageing or allostatic factors.