No abstract

ObjectiveA premorbid IQ deficit supports a developmental dimension to schizophrenia and its cognitive aspects that are crucial to functional outcome. Better characterisation of the association between premorbid IQ and the disorder may provide further insight into its origin and etiology. We aimed to quantify premorbid cognitive function in schizophrenia through systematic review and meta-analysis of longitudinal, population-based studies, and to characterize the risk of schizophrenia across the entire range of premorbid IQ.MethodElectronic and manual searches identified general population-based cohort or nested case–control studies that measured intelligence before onset of schizophrenic psychosis using standard psychometric tests, and that defined cases using contemporaneous ICD or DSM. Meta-analyses explored dose–response relationships between premorbid cognitive deficit (using full-scale, verbal and performance IQ) and risk of schizophrenia. Meta-regression analyses explored relationships with age of illness onset, change in premorbid intelligence over time and gender differences.ResultsMeta-analysis of 4396 cases and over 745 000 controls from 12 independent studies confirmed significant decrements in premorbid IQ (effect size − 0.43) among future cases. Risk of schizophrenia operated as a consistent dose–response effect, increasing by 3.7% for every point decrease in IQ (p < 0.0001). Verbal and nonverbal measures were equally affected. Greater premorbid IQ decrement was associated with earlier illness onset (p < 0.0001). There was no evidence of a progressively increasing deficit during the premorbid period toward illness onset.ConclusionsStrong associations between premorbid IQ and risk for schizophrenia, and age of illness onset argue for a widespread neurodevelopmental contribution to schizophrenia that operates across the entire range of intellectual ability. This also suggests higher IQ may be protective in schizophrenia, perhaps by increasing active cognitive reserve.

The catechol-O-methyltransferase (COMT) Val 158 Met polymorphism is hypothesized to affect executive function in patient and control populations. Studies inconsistently report better performance on the Wisconsin Card Sort Test (WCST) in individuals with one or more Met alleles. We conducted a meta-analysis of studies published until August 2006 that reported WCST perseverative errors from healthy volunteers or patients with schizophrenia-spectrum disorders. Twelve studies met inclusion criteria (total n = 1910) providing 10 samples each of patients and controls. In healthy controls, individuals with the Met/Met genotype performed better than those with the Val/Val genotype (d = 0.29; 95% confidence interval (CI) 0.02-0.55; P = 0.03), but this was not supported in the patient sample (d = À0.07; 95% CI À0.40 to 0.26; P = 0.68). Post hoc analyses suggested that Val and Met alleles are codominant in their effects on cognition. Effect size was greater in studies published at an earlier date and may also be larger in non-Caucasian samples. Gender did not affect the results. There was no evidence of publication bias. We conclude that there is small but significant relationship between Val 158 Met genotype and executive function in healthy individuals but not in schizophrenia.