Neuropsychological Assessment Monozygotic Twins Discordant for Schizophrenia
of
Terry E. Goldberg, PhD; J. Daniel Ragland, MA; E. Fuller Torrey, MD; James M. Gold, PhD; Llewellyn B. Bigelow, MD; Daniel R. Weinberger, MD \s=b\ A comparison of monozygotic twins discordant for schizophrenia controls for genetic variance and reduces variance due to environmental circumstances, thus serving to highlight differences due to phenotypic-related variables. In this study, we assessed 16 such twin pairs on a wide range of neuropsychological tests. The affected twins tended to perform worse
counterparts on most of the tests. Deficits were especially severe on tests of vigilance, memory, and concept formation, suggesting that dysfunction is greatest in the frontotemporal cortex. While manifest symptoms were not highly associated with neuropsychological scores, global level of functioning was. To address the issue of genetic liability, we also compared the sample of discordant unaffected twins with a samthan their unaffected
ple of seven pairs of normal monozygotic twins. No significant differences between the groups were found for any neuropsychological test. In fact, the results suggest that neuropsychological dysfunction is a consistent feature of schizophrenia and that it is related primarily to the clinical disease process and not to genetic or nonspecific environmental factors. (Arch Gen Psychiatry. 1990;47:1066-1072)
schizophrenia generated Recent accompaniesimportance cognitive impairment disorder.1,2 proposed
research in has renewed interest in the of the that often the It has been that abnormal cognition may be especially important in ac¬ counting for the long-term social and vocational cost of the disorder, as deficits tend to remain stable over time, even as other symptoms wax and wane.3 Although a variety of ap¬ proaches have been used to assess cognitive deficits, stanAccepted for publication March 8,1990. From the Clinical Brain Disorders Branch, Intramural Research Program, National Institute of Mental Health, Neuroscience Center at St Elizabeths, Washington, DC. Reprint requests to Clinical Brain Disorders Branch, Intramural Research Program, National Institute of Mental Health, Neuroscience Center at St Elizabeths, Washington, DC 20032 (Dr Goldberg).
dardized neuropsychological tests are particularly advanta¬ geous because they are reliable and they have been validated on groups of patients with known neuropathological insult, making it possible to infer patterns of regional cerebral
dysfunction.
There are, however, fundamental problems in the neuro¬ psychological approach to schizophrenia. One problem in¬ volves the difficulty inherent in assessing what is "normal" cognitive performance for a given individual. While signifi¬ cant differences in the performance means have been consis¬ tently found between schizophrenic and control groups, the degree to which the groups overlap in performance is also often high. That is, many schizophrenic patients may perform in the normal range on any given test. Because it is difficult to know whether a patient who performs within the normal range has fallen from his or her premorbid level, it is impossi¬ ble to be sure that even normally performing subjects have not suffered cognitive impairment. Thus, it is unclear wheth¬ er cognitive dysfunction characterizes a subgroup of patients (outside the range of normal) or the majority of patients, each of whom differs from some individualized, ideal control. A second problem with the neuropsychological approach to schizophrenia concerns the possible impact of environmental factors in the etiology of schizophrenia and cognitive impair¬ ment." It has been argued that socioeconomic status has a role in the genesis of schizophrenia. Moreover, socioeconomic dif¬
as well as educational differences, have often been associated with differences in cognitive level.5 Therefore, evidence of cognitive impairment in schizophrenia may be an artifact of phenomena that have not been adequately controlled. Another problem concerns the very nature of the cognitive impairments themselves. Some have argued that the impair¬ ments are general,6 and thus they implicitly suggest that many cerebral regions are dysfunctional in persons with schizophrenia. Others believe deficits to be more specific, possibly restricted to a single brain region or system. For
ferences,
Downloaded From: http://archpsyc.jamanetwork.com/ by a Cornell University User on 09/28/2016
example, poor performance on tests of executive function710 have implicated prefrontal cortex. Impaired performances on selected memory tasks have implicated medial temporal and frontal regions.11"13 Attentional deficits have been thought to reflect midbrain or frontal dysfunction.14,15 Moreover, pat¬ terns of deficits have sometimes been thought to reflect left hemispheric or right hemispheric dysfunction.16"18 The degree to which sampling biases in the selection of patients and controls contributes to the variability of these findings is
uncertain. The study of monozygotic twins who are discordant for schizophrenia is a research strategy used in the present study that should control for most of these problems. Because each pair of twins shares a common genome and similar economic, educational, and family backgrounds, the degree to which these factors contribute to neuropsychological variance is minimized. Differences that distinguish the affected from the unaffected group thus represent a process that is probably independent of shared genetic and psychosoeial circum¬ stances and, thus, related to the disease phenotype. Howev¬ er, similarities in cognitive function may be related to the shared characteristics of the twins. To further examine the question of genetic liability, use of a normal control group of monozygotic twins genetically unrelated to the discordant twins is desirable. Such a group was also included in this study. To our knowledge, no study yet has examined a group of monozygotic twins affected with schizophrenia, their matched monozygotic, nonschizophrenic co-twins, and an un¬ related group of normal monozygotic twins from a neuropsy¬
chological standpoint.
SUBJECTS AND METHODS Two sets of twins were studied. In the first set, subjects were members of 16 pairs of identical twins discordant for schizophrenia. Of the pairs, eight were male and eight were female. In the second set, subjects were members of seven pairs of normal monozygotic
twins. Three pairs were male and four pairs were female. Monozygosity was determined by appearance19 and analysis of 19 red blood cell antigens.20 All discordant and concordant twins underwent a stan¬ dardized structured diagnostic interview, the Structured Clinical Interview for DSM-III-R, to establish DSM-III-R diagnoses.21 Unaf¬ fected and normal twins also completed a Structured Clinical Inter¬ view for DSM-III-R for personality disorders. Of the twins affected with schizophrenia, 11 were considered to be undifferentiated, two were considered to be paranoid, two were considered to be residual, and one was considered to be schizoaffective. Eight of the unaffected twins did not receive a psychiatric diagnosis. Five subjects met criteria for a past history of substance abuse. Of the remaining unaffected twins, one met criteria for major depression in remission, one met criteria for atypical specific developmental disorder, and one met criteria for antisocial personality disorder. The discordant group was recruited through the network of the National Alliance for the Mentally 111. The normal sample was recruited by "word of mouth" and was carefully screened to exclude subjects who had suffered prenatal, perinatal, or postnatal neurological insult or who had diagnosable psychiatric symptomatology. No twin in the concordant group carried a DSM-III-R Axis I or II diagnosis. All sets of twins were reared together. The mean age of the discordant twins was 32 years (range, 24 to 44 years). The mean age of the normal concordant group was 31 years (range, 19 to 44 years). Ages between the groups did not differ significantly (t .29, =. 75). The group of normal twins had a mean educational level of 16.6 years; the discordant unaffected group had a mean educational level of 15.6 years; and the discordant affected group had a mean educational level of 13.8 years. Between-group differences were significant (F 7.31, .002); the affected group differed significantly from the other two groups (i>2.02, P2.02, P
of
Terry E. Goldberg, PhD; J. Daniel Ragland, MA; E. Fuller Torrey, MD; James M. Gold, PhD; Llewellyn B. Bigelow, MD; Daniel R. Weinberger, MD \s=b\ A comparison of monozygotic twins discordant for schizophrenia controls for genetic variance and reduces variance due to environmental circumstances, thus serving to highlight differences due to phenotypic-related variables. In this study, we assessed 16 such twin pairs on a wide range of neuropsychological tests. The affected twins tended to perform worse
counterparts on most of the tests. Deficits were especially severe on tests of vigilance, memory, and concept formation, suggesting that dysfunction is greatest in the frontotemporal cortex. While manifest symptoms were not highly associated with neuropsychological scores, global level of functioning was. To address the issue of genetic liability, we also compared the sample of discordant unaffected twins with a samthan their unaffected
ple of seven pairs of normal monozygotic twins. No significant differences between the groups were found for any neuropsychological test. In fact, the results suggest that neuropsychological dysfunction is a consistent feature of schizophrenia and that it is related primarily to the clinical disease process and not to genetic or nonspecific environmental factors. (Arch Gen Psychiatry. 1990;47:1066-1072)
schizophrenia generated Recent accompaniesimportance cognitive impairment disorder.1,2 proposed
research in has renewed interest in the of the that often the It has been that abnormal cognition may be especially important in ac¬ counting for the long-term social and vocational cost of the disorder, as deficits tend to remain stable over time, even as other symptoms wax and wane.3 Although a variety of ap¬ proaches have been used to assess cognitive deficits, stanAccepted for publication March 8,1990. From the Clinical Brain Disorders Branch, Intramural Research Program, National Institute of Mental Health, Neuroscience Center at St Elizabeths, Washington, DC. Reprint requests to Clinical Brain Disorders Branch, Intramural Research Program, National Institute of Mental Health, Neuroscience Center at St Elizabeths, Washington, DC 20032 (Dr Goldberg).
dardized neuropsychological tests are particularly advanta¬ geous because they are reliable and they have been validated on groups of patients with known neuropathological insult, making it possible to infer patterns of regional cerebral
dysfunction.
There are, however, fundamental problems in the neuro¬ psychological approach to schizophrenia. One problem in¬ volves the difficulty inherent in assessing what is "normal" cognitive performance for a given individual. While signifi¬ cant differences in the performance means have been consis¬ tently found between schizophrenic and control groups, the degree to which the groups overlap in performance is also often high. That is, many schizophrenic patients may perform in the normal range on any given test. Because it is difficult to know whether a patient who performs within the normal range has fallen from his or her premorbid level, it is impossi¬ ble to be sure that even normally performing subjects have not suffered cognitive impairment. Thus, it is unclear wheth¬ er cognitive dysfunction characterizes a subgroup of patients (outside the range of normal) or the majority of patients, each of whom differs from some individualized, ideal control. A second problem with the neuropsychological approach to schizophrenia concerns the possible impact of environmental factors in the etiology of schizophrenia and cognitive impair¬ ment." It has been argued that socioeconomic status has a role in the genesis of schizophrenia. Moreover, socioeconomic dif¬
as well as educational differences, have often been associated with differences in cognitive level.5 Therefore, evidence of cognitive impairment in schizophrenia may be an artifact of phenomena that have not been adequately controlled. Another problem concerns the very nature of the cognitive impairments themselves. Some have argued that the impair¬ ments are general,6 and thus they implicitly suggest that many cerebral regions are dysfunctional in persons with schizophrenia. Others believe deficits to be more specific, possibly restricted to a single brain region or system. For
ferences,
Downloaded From: http://archpsyc.jamanetwork.com/ by a Cornell University User on 09/28/2016
example, poor performance on tests of executive function710 have implicated prefrontal cortex. Impaired performances on selected memory tasks have implicated medial temporal and frontal regions.11"13 Attentional deficits have been thought to reflect midbrain or frontal dysfunction.14,15 Moreover, pat¬ terns of deficits have sometimes been thought to reflect left hemispheric or right hemispheric dysfunction.16"18 The degree to which sampling biases in the selection of patients and controls contributes to the variability of these findings is
uncertain. The study of monozygotic twins who are discordant for schizophrenia is a research strategy used in the present study that should control for most of these problems. Because each pair of twins shares a common genome and similar economic, educational, and family backgrounds, the degree to which these factors contribute to neuropsychological variance is minimized. Differences that distinguish the affected from the unaffected group thus represent a process that is probably independent of shared genetic and psychosoeial circum¬ stances and, thus, related to the disease phenotype. Howev¬ er, similarities in cognitive function may be related to the shared characteristics of the twins. To further examine the question of genetic liability, use of a normal control group of monozygotic twins genetically unrelated to the discordant twins is desirable. Such a group was also included in this study. To our knowledge, no study yet has examined a group of monozygotic twins affected with schizophrenia, their matched monozygotic, nonschizophrenic co-twins, and an un¬ related group of normal monozygotic twins from a neuropsy¬
chological standpoint.
SUBJECTS AND METHODS Two sets of twins were studied. In the first set, subjects were members of 16 pairs of identical twins discordant for schizophrenia. Of the pairs, eight were male and eight were female. In the second set, subjects were members of seven pairs of normal monozygotic
twins. Three pairs were male and four pairs were female. Monozygosity was determined by appearance19 and analysis of 19 red blood cell antigens.20 All discordant and concordant twins underwent a stan¬ dardized structured diagnostic interview, the Structured Clinical Interview for DSM-III-R, to establish DSM-III-R diagnoses.21 Unaf¬ fected and normal twins also completed a Structured Clinical Inter¬ view for DSM-III-R for personality disorders. Of the twins affected with schizophrenia, 11 were considered to be undifferentiated, two were considered to be paranoid, two were considered to be residual, and one was considered to be schizoaffective. Eight of the unaffected twins did not receive a psychiatric diagnosis. Five subjects met criteria for a past history of substance abuse. Of the remaining unaffected twins, one met criteria for major depression in remission, one met criteria for atypical specific developmental disorder, and one met criteria for antisocial personality disorder. The discordant group was recruited through the network of the National Alliance for the Mentally 111. The normal sample was recruited by "word of mouth" and was carefully screened to exclude subjects who had suffered prenatal, perinatal, or postnatal neurological insult or who had diagnosable psychiatric symptomatology. No twin in the concordant group carried a DSM-III-R Axis I or II diagnosis. All sets of twins were reared together. The mean age of the discordant twins was 32 years (range, 24 to 44 years). The mean age of the normal concordant group was 31 years (range, 19 to 44 years). Ages between the groups did not differ significantly (t .29, =. 75). The group of normal twins had a mean educational level of 16.6 years; the discordant unaffected group had a mean educational level of 15.6 years; and the discordant affected group had a mean educational level of 13.8 years. Between-group differences were significant (F 7.31, .002); the affected group differed significantly from the other two groups (i>2.02, P2.02, P
