Relative risk for cognitive impairments in siblings of patients with schizophrenia.

BACKGROUND: Patients with schizophrenia have impairments in several domains of cognition, including working memory/executive function, verbal memory, language, oculomotor scanning/psychomotor speed, and general intelligence. Impairments have also been found in unaffected siblings, suggesting they could be heritable. To assess the suitability of cognitive dysfunction for use in genetic studies, we estimated relative risk (lambda) in a large cohort of siblings.

METHODS: One hundred forty-seven patients with schizophrenia, 193 of their siblings, and 47 control subjects were studied using a neuropsychological test battery, which included intelligence quotient (IQ), Wide Range Achievement Test, Wisconsin Card Sort, Wechsler Memory Scale (revised), California Verbal List Test, Trails A and B, and Letter and Category Fluency. Relative risk was estimated using a cutoff score of 1 SD below the control mean.

RESULTS: As expected, patients performed markedly worse than control subjects on all tests except the Wide Range Achievement Test. Siblings had impaired performance on the Wisconsin Card Sort and Trails B, with trends for reduction (p = .01-.05) on the California Verbal List Test and Letter Fluency. Relative risk to siblings was elevated on the Trails B (lambda = 4.0) and California Verbal List Test (lambda = 2.8). Trends (p = .01-.05) for increased lambda were also seen for Wisconsin Card Sort, Letter Fluency, Wechsler Memory Scale and decline in IQ (lambda = 1.74-2.4). Correlations between tests of different cognitive functions were weak, indicating they measure relatively independent processes.

CONCLUSION: Unselected siblings of patients with schizophrenia have impairments in several cognitive domains. Relative risk scores were in the moderate range, suggesting a significant genetic component. Impairments on one test only weakly predicted impairments on other tests. Thus, cognitive phenotypes identify distinct, familial traits associated with schizophrenia. Using this dimensional approach to subdividing schizophrenia may reduce the clinical and genetic heterogeneity of schizophrenia and improve the power of genetic studies.