Executive dysfunctions as potential markers of familial vulnerability to bipolar disorder and schizophrenia

Abstract

Attentional and executive impairments have been found both in patients with schizophrenia and in their unaffected first-degree relatives, suggesting that they might be considered as familial vulnerability markers. Several studies have shown that the performance of bipolar patients does not significantly differ from that of schizophrenic patients, so that executive and attentional deficits might not be specific to schizophrenia. In the present study, we aimed to identify executive dysfunctions in schizophrenia and bipolar disorder that might be vulnerability trait markers specific to one or common to both of these diseases. We assessed cognitive performance of euthymic bipolar and schizophrenic patients, their unaffected first-degree relatives and a healthy control group, using neuropsychological tasks to test different components of executive function: the Verbal Fluency Test, the Stroop Word Colour Test, the Wisconsin Card Sorting Test and the Trail Making Test. The two groups of patients and their unaffected relatives demonstrated disproportionately increased slowness on the Stroop test in comparison to the normal healthy group. Patients with schizophrenia performed poorly on all the tests in comparison to the normal healthy subjects, while no other impairment was observed in the bipolar patients and in the relatives of schizophrenic and bipolar patients. Enhanced susceptibility to interference and reduced inhibition could be transnosographical markers for a shared familial vulnerability common to schizophrenia and bipolar disorders.

Introduction

Although classical genetic studies have consistently shown that there is a strong genetic contribution to the risk of schizophrenia and bipolar disorders, the search for the predisposing genes has so far been unsuccessful. One obstacle is our inability to define the heritable phenotype (Leboyer et al., 1998). Various types of phenotypic misclassification reduce the power of linkage studies because of phenocopies or genetic heterogeneity within apparently affected subjects and within apparently unaffected subjects or controls, because of our inability to identify carriers of vulnerability genes (incomplete penetrance). One possible approach to minimise the arbitrariness of categorical diagnoses is to use subclinical quantitative traits, called endophenotypes, as phenotypes, as they are closer to the pathogenic genotype than the disease phenotype itself. The ‘endophenotype’ strategy assumes that complex psychiatric disorders can be conceptualised as the interaction of several elementary neurobiological phenotypes, each caused by a specific defect in a given candidate gene and each increasing the risk of developing complex psychiatric disorders (Egan et al., 2000, Tsuang, 1993). Endophenotypes might be affected by Mendelian inheritance patterns, which would considerably diminish the sample size required to detect the responsible genetic mutation (Freedman et al., 1999). Thus, this alternative phenotypic strategy would greatly enhance the power of the genetic analysis of complex psychiatric disorders. A behavioural characteristic should meet the following criteria if it is to be considered as an endophenotype: firstly, it should be specific, i.e. more common in patients than in controls; secondly, it should be independent of the clinical state; and thirdly, it should be familial and heritable, i.e. more common in the non-psychotic, first-degree relatives of patients than in the general population. According to this conceptual framework, cognitive endophenotypes are traits that are closely linked to gene expression and, when present in unaffected subjects, they are believed to indicate the genetic liability of the disorder (Leboyer et al., 1998). These indicators can be disease-specific, as the endophenotype might be the marker of vulnerability to a particular disorder or a marker of common vulnerability to more than one psychiatric disorder.

Crow (1997) argues that ‘schizophrenia’ is not a natural category and that psychosis can be better understood in terms of a continuum of phenomenological symptoms stretching from schizophrenic psychoses to manic-depressive disease. Within the perspective of finding genetically mediated vulnerability risks, the comparison of unaffected relatives of individuals with schizophrenia and those with manic-depressive disorder may provide relevant information. Genetic factors play an important role in these two diseases, with a comparable prevalence of the same type of disorder in first-degree relatives of schizophrenic and bipolar patients (respectively, 10 and 8%), and a similar concordance rate for monozygotic twins in bipolar disorder (58–74%, Faraone and Tsuang, 1985) and schizophrenia (50%, Gottesman and Shields, 1982). Evidence from family and linkage studies suggested that the two disorders share some aetiological and pathophysiological features (Berrettini, 2000, Potash et al., 2001, Valles et al., 2000), although no reliable results about candidate genes are available yet.

Numerous studies (Conklin et al., 2000, Chen and Faraone, 2000, Kremen et al., 1994) have revealed that the unaffected relatives of schizophrenic patients may share subtle neuropsychological abnormalities with their affected relatives. These studies provided several examples of neurocognitive risk indicators or endophenotypes in domains of working memory, attention and executive functions (Goldberg et al., 1990, Keefe et al., 1994, Mirsky et al., 1995, Dollfus et al., 2002, Gilvarry et al., 2001). Conversely, few neuropsychological investigations have assessed neurocognitive functioning in bipolar patients and their unaffected relatives (Murphy and Sahakian, 2001) and even fewer studies directly compared the two groups of patients and their unaffected first-degree relatives (Gilvarry et al., 2001).

Although bipolar disorder often has a better outcome and a more episodic course than schizophrenia, remitted bipolar patients manifest neuropsychological deficits that are sometimes similar to those of schizophrenic patients. Cognitive disorders in schizophrenia and manic-depressive disease have often been associated with frontal lobe dysfunction. Keri et al. (2001) showed that the siblings of bipolar patients could be distinguished from the siblings of schizophrenic patients due to impaired spatial working memory, although both groups showed verbal recall deficit. This suggests a common trait marker of vulnerability in the fronto-hippocampal system.

Both schizophrenic and remitted bipolar patients have deficits in attention tasks such as the Continuous Performance Task and the Digit Span Task, regardless of their clinical state and in working memory and executive functions such as the Wisconsin Card Sorting Test (WCST), the Trail Making Test (TMT), the Verbal Fluency Test (VFT) and the Stroop Word Colour Test (SWCT) (Morice, 1990, Rossi et al., 1997, Addington and Addington, 1997, Ferrier et al., 1999). Abnormalities in backward masking have also been described in patients with schizophrenia and bipolar disorder (Rund, 1993, Tam et al., 1998). McGrath et al. (1997) reported impairments in strategy generation, cognitive flexibility, and initiation and suppression abilities in schizophrenia and bipolar disorder by using the TMT, the WCST and the SWCT. Furthermore, no significant improvements were observed over time in the performance of the manic group on the Stroop task when controlled for practice effects (McGrath et al., 1997) or in the performance of depressed patients after clinical improvements (Trichard et al., 1995).

Neuroimaging studies have shown similar structural and functional brain abnormalities in bipolar and schizophrenic patients, such as decreased size or hypoactivation in the prefrontal cortex (Rieder et al., 1983, Carter et al., 1998, Bearden et al., 2001). Overall, these findings suggest that executive dysfunctions, associated with frontal lobe abnormalities, are not specific to schizophrenia.

In an analysis of previous studies, Goldberg (1999) stated that despite an overlap in their phenomenological symptoms, cognitive functions in bipolar individuals are generally more preserved than in schizophrenia, in particular, in the domain of working memory, even when IQ was controlled. However, several discrepancies and methodological pitfalls exist in the literature, such as unclear remission criteria, diagnostic heterogeneity and small sample sizes (Martinez-Aran et al., 2000). Moreover, the results obtained often do not use comprehensive neuropsychological batteries and some do not include normal controls.

In the present study, we aimed to identify putative executive dysfunctions as vulnerability indicators specific or common to schizophrenia and bipolar disorders. Therefore, we compared the performance of bipolar and schizophrenic probands, their unaffected first-degree relatives and normal controls by use of a series of neuropsychological tests sensitive to prefrontal cortical dysfunction: the VFT, the SWCT, the WCST and the TMT. We predicted that relatives of schizophrenic and bipolar patients would mirror the pattern of deficits observed in their affected relatives and that the performance of bipolar patients and their relatives would be globally better preserved than that of schizophrenic patients and their relatives, respectively. The presence of similar cognitive abnormalities in probands and their unaffected relatives provides relevant information about the familial, presumed genetic, neurobiological characteristics associated with these diseases that are independent of severity, ongoing symptoms and treatments. These endophenotypic markers could subsequently be useful for genetic linkage analyses of these complex psychiatric disorders.