ویژگی های شخصیتی در اسکیزوفرنی و اختلالات شخصیتی مرتبط

Abstract We investigated whether schizophrenia spectrum disorders share common personality characteristics or traits. Participants with a diagnosis of schizophrenia or schizoaffective disorder (SZ) or with a schizophrenia spectrum personality disorder (schizophrenia spectrum PD: schizoid, paranoid, and schizotypal personality disorder) were compared with non-psychiatric control subjects on the five-factor model of personality and the psychosis-proneness scales. On the five-factor personality scales, SZ subjects showed higher levels of neuroticism, and lower levels of openness, agreeableness, extraversion, and conscientiousness than control subjects. Higher scores on openness and lower scores on neuroticism distinguished schizophrenia spectrum PD from SZ. On the psychosis-proneness scales, both PD and SZ participants scored high relative to non-psychiatric control participants on magical ideation and perceptual aberration, while PD participants scored intermediate between non-psychiatric control participants and SZ on social anhedonia. Discriminant analysis indicated that schizophrenia spectrum patients could be distinguished from PDs by more severe social withdrawal and maladjustment, while subjects with PDs could be best distinguished from control subjects on the basis of odd or novel ideation and decreased conscientiousness.

Introduction Patients with schizophrenia show abnormalities on basic dimensions of personality (Berenbaum and Fujita, 1994, Kentros et al., 1997, Bagby et al., 1997, Bagby et al., 1999 and Gurrera et al., 2000). These personality disturbances may be a manifestation of liability to schizophrenia (Chapman et al., 1994, Claridge, 1997, Kendler et al., 1993, Lenzenweger and Loranger, 1989 and Meehl, 1989). Meehl (1989) proposed that personality disturbance was the result of the interaction of a neural integrative deficit, termed schizotaxia, with social learning during development. Meehl suggested that while schizotaxia usually resulted in schizotypal personality, only a fraction of such individuals subsequently developed schizophrenia. This theory raises the issue of whether individuals with personality disorders linked to schizophrenia show similar personality characteristics. The schizophrenia spectrum personality disorders (PDs) include schizotypal personality disorder, paranoid personality disorder, and schizoid personality disorder, termed “Cluster A” in the DSM-IV diagnostic schema (DSM-IV, American Psychiatric Association, 1994). While these personality disorders are not associated with the severe psychosocial disturbance characteristic of schizophrenia, evidence from familial studies suggests that they reflect the phenotypic expression of a liability for schizophrenia (Battaglia et al., 1995, Baron et al., 1985, Frangos et al., 1985, Kendler et al., 1984 and Lowing et al., 1983). Consequently, comparing schizophrenia and schizophrenia spectrum PDs may yield insights into which personality disturbances are common to schizophrenia spectrum personality disorders varying in clinical severity, and which only occur with psychosis. The relationship between diagnostic categories and personality traits can be characterized by comparing dimensional traits between diagnostic groups. This approach has been frequently used to evaluate personality traits in schizophrenia. Based on a meta-analysis of research on personality and schizophrenia, Berenbaum and Fujita (1994) concluded that patients with schizophrenia tended to be more introverted (low E) and more neurotic (high N) than control participants. In an attempt to standardize these dimensions, a number of investigators have characterized schizophrenia in terms of the Five-Factor model (FFM) of personality. This model comprises the following five dimensions: Neuroticism (N), Extraversion (E), Openness to Experience (O), Agreeableness (A), and Conscientiousness (C) (Costa and McCrae, 1990 and McCrae and Costa, 1997). Several recent studies have used versions of the NEO Five-Factor Inventory (Costa and McCrae, 1989 and Costa and McCrae, 1992) to evaluate the FFM in schizophrenia. On these scales, patients with schizophrenia and schizoaffective disorder showed elevated N, low E and C, and sometimes low A and O scores relative to normative data (Bagby et al., 1997 and Bagby et al., 1999) or a control group (Gurrera et al., 2000 and Kentros et al., 1997). These findings lead to the question of whether individuals with schizophrenia spectrum PDs show the same pattern of personality characteristics as patients with schizophrenia. No studies to our knowledge have contrasted FFM scores among schizophrenia spectrum PD (diagnosed with DSM-III or DSM-IV criteria), schizophrenic, and non-psychiatric control subjects. However, a variety of studies have examined the relationship of measures of schizophrenia spectrum PD symptoms and measures of the FFM. These studies can be divided into those that use clinical samples, which often include a number of subjects with Cluster A disorders; and those that use non-clinical samples, such as university undergraduates. Because the base rates of schizophrenia spectrum PDs in the population are low, it is unlikely that community or university samples will include more than a small percentage of subjects who meet DSM diagnoses for Cluster A personality disorders. We therefore review non-clinical and clinical samples separately below. 1.1. FFM in non-clinical samples Studies of community or university samples have used a variety of measures to characterize traits associated with schizotypy or “psychosis proneness,” such as the self-report scales developed by Chapman and his associates (Chapman et al., 1978, Eckblad et al., 1982 and Eckblad and Chapman, 1983). The Chapman Scales include the Revised-Social Anhedonia Scale (R-SAS), the Magical Ideation Scale (MIS), and the Perceptual Aberration Scale (PAS; Kwapil, 1998, Eckblad and Chapman, 1983 and Chapman et al., 1978). The R-SAS indexes social withdrawal due to a lack of interest in intimacy and interaction. The MIS measures magical beliefs and ideas of reference. The PAS assesses perceptual distortions, especially disturbances of body image. High scores reflect increased liability for psychosis and substance abuse at long-term follow-up (Chapman et al., 1994 and Kwapil, 1998). The reliability of the Chapman Scales has been demonstrated by internal consistency coefficient alphas in the upper 0.70s to the lower 0.90s in several studies (Chapman et al., 1978, Chapman et al., 1994 and Mishlove and Chapman, 1985), and test–retest reliabilities have ranged between 0.75 and 0.85 (Chapman et al., 1994). Other researchers have used scales derived from the Minnesota Multiphasic Personality Inventory to measure dimensions related to schizophrenia spectrum PDs (e.g. Costa and McCrae, 1990) or subsets of items from other personality measures (e.g. Wiggins and Pincus, 1989). In undergraduate or community samples, high scores on scales measuring schizophrenia spectrum PDs are consistently associated with increased neuroticism (Coolidge et al., 1994, Dyce and O'Connor, 1998, Costa and McCrae, 1990 and Ross et al., 2002). Additionally, specific traits have shown correlations with FFM measures. High scores on the dimensions of schizotypal PD or positive schizotypy have been most consistently associated with low extraversion (Coolidge et al., 1994, Dyce and O'Connor, 1998 and Costa and McCrae, 1990) and increased openness (Coolidge et al., 1994, Dyce and O'Connor, 1998 and Ross et al., 2002). Less uniformly, schizotypal PD traits have also been associated with decreased agreeableness (Coolidge et al., 1994 and Dyce and O'Connor, 1998) and decreased conscientiousness (Coolidge et al., 1994). Elevations on scales measuring the schizoid or negative schizotypy dimension have been associated with decreased agreeableness (Coolidge et al., 1994, Dyce and O'Connor, 1998, Costa and McCrae, 1990 and Ross et al., 2002) and, less consistently, decreased extraversion (Dyce and O'Connor, 1998 and Ross et al., 2002), decreased openness (Ross et al., 2002), and decreased conscientiousness (Costa and McCrae, 1990). Elevated scores on the paranoid dimension have been associated with decreased agreeableness (Coolidge et al., 1994, Dyce and O'Connor, 1998 and Costa and McCrae, 1990) and decreased extraversion (Dyce and O'Connor, 1998 and Ross et al., 2002). 1.2. FFM in clinical or relative studies Most clinical studies have examined FFM scores among diagnostically heterogeneous inpatient and outpatient samples, although these samples often had a significant number of patients diagnosed with a schizophrenia spectrum PD. Blais (1997) studied patients with personality disorders, and used interview ratings of the FFM. Thirty percent of the patients had paranoid personality disorder, and 7% had schizoid personality disorder. N was positively related to all three personality traits, A was negatively correlated with all three scores, and E was negatively correlated with SPD and schizoid scores. Morey et al. (2000) evaluated patients with personality disorders, 47% of whom had a schizophrenia spectrum PD diagnosis. Most patients also had Axis I disorders, primarily major depression or anxiety disorders. Profiles indicated that schizotypal, schizoid, and paranoid personality disorder were associated with high N. Schizoid PD was associated with low E, while paranoid PD was associated with low A. Trull (1992) evaluated outpatients, of whom 15% received a diagnosis of a schizophrenia spectrum PD. Schizoid and schizotypal PD symptoms were negatively correlated with E, schizoid scores were negatively correlated with O, and all three types of symptoms were negatively correlated with A. Yeung et al. (1993) evaluated relatives of patients with psychotic disorders. Of these, five percent had a diagnosis of a schizophrenia spectrum PD. Among the entire sample, schizotypal characteristics were positively correlated with N and negatively correlated with A and C; schizoid features were negatively correlated with E and A; and paranoid features were negatively correlated with A. Tien et al. (1992) evaluated 38 subjects from an epidemiologic sample who had symptoms of psychosis or panic. Of these, two had diagnoses of a schizophrenia spectrum PD. Among the entire sample, SPD traits were positively correlated with N; schizoid traits were positively correlated with N and negatively correlated with A; and paranoid traits were positively correlated with N and negatively with A. In summary, studies of clinical samples suggest that symptoms of schizophrenia spectrum PD are frequently associated with increased neuroticism and decreased extraversion. Reduced agreeableness is more consistently associated with schizoid and paranoid symptoms than with schizotypal symptoms. Reduced extraversion is most characteristic of schizoid disorders. As noted by Ross et al. (2002), clinical samples do not appear to show the elevation of openness that has frequently been reported in non-clinical samples. 1.3. Objectives of the present study The present study compares the FFM and Chapman Scales in schizophrenia, schizophrenia spectrum personality disorders, and non-psychiatric control participants. Unlike previous studies, the schizophrenia spectrum PD group was composed only of subjects who met DSM-IV criteria for a Cluster A personality disorder, were not comorbid for Axis I disorders, and were not receiving psychiatric treatment. The relationship between DSM-IV diagnosis and personality measures remains a critical question, since most familial studies of schizophrenia and spectrum personality disorders use these diagnostic criteria. Based on previous studies, we predicted that schizophrenia spectrum PD participants would score similarly to schizophrenia subjects on neuroticism, intermediate between schizophrenia and non-psychiatric participants on extraversion, agreeableness, and conscientiousness, and show similar or higher scores on openness relative to control subjects. We also investigated the relationships among scores on the FFM dimensions and the Chapman psychosis proneness scales using correlational and factor analysis. Discriminant analysis was used to determine which personality dimensions differentiated between schizophrenia spectrum PD and schizophrenia.

Results 3.1. Five-factor personality scores Table 1 shows the means and standard deviations across groups on the five NEO-FFI traits. There was a main effect of group (F[2, 139]=19.02, p<0.001), and a significant interaction of group by personality dimension (Wilks' Lambda value=0.42, F[8, 272]=18.20, P<0.001). There was also a main effect of dimension (lambda=0.60, F[4, 136]=22.96, P<0.001), indicating that NEO dimensions of E, A, and C had lower scores than N and O. Table 1. Means and standard deviations of the five NEO Five-Factor Inventory across groups NC SSPD SZ Mean S.D. Mean S.D. Mean S.D. N 46.7 8.9 56.2 12.0 63.9 11.8 E 54.3 11.8 49.6 11.4 42.5 12.2 O 55.9 10.7 65.4 8.6 46.7 9.6 A 53.8 10.3 48.7 12.8 42.0 10.8 C 50.0 10.6 40.5 11.7 40.0 12.0 N: Neuroticism; E: Extraversion; O: Openness; A: Agreeableness; C: Conscientiousness. Table options Between-group ANOVAs for each personality dimension revealed a significant effect for group on all measures (F>12.6; P<0.001). Tukey's Honestly Significant Difference (HSD) test was used to compare scores among the SZ, schizophrenia spectrum PD, and control groups on each FFI measure. Schizophrenia spectrum PD participants were more neurotic (P<0.001), more open (P<0.001), and less conscientious (P<0.002) than control participants. Schizophrenics were more neurotic (P<0.001), less extroverted (P<0.001), less open (P<0.001), less agreeable (P<0.001), and less conscientious (P<0.001) than control participants. Schizophrenic patients were more neurotic (P<0.01), less extroverted (P<0.05), less open (P<0.001), and less agreeable (P<0.05) than schizophrenia spectrum PD participants. 3.2. Chapman Scales Table 2 presents the means and standard deviations for the three Chapman Scales for each group. There was a main effect of group, (F[2, 105]=51.72, P<0.001), and a significant interaction of group by test (Wilks' Lambda=0.90, F[4, 208]=2.86, P<0.05). There was also a main effect of test (Wilks' Lambda=0.54, F[2, 104]=43.57, P<0.001), indicating that scores on the MIS and R-SAS were significantly higher than scores on the PAS. Table 2. Means and standard deviations for the Chapman Scales Control SSPD Schizophrenia Mean S.D. Mean S.D. Mean S.D. PAS 1.9 2.8 8.3 8.1 10.2 7.0 MIS 4.6 4.6 13.6 6.7 14.2 7.5 R-SAS 6.5 5.1 11.3 7.1 18.5 9.9 PAS: Perceptual Aberration Scale; MIS: Magical Ideation Scale; R-SAS: Revised Social Anhedonia Scale; SSPD: Schizophrenia Spectrum Personality Disorder. Table options Between-group ANOVAs for each personality measure revealed a significant effect for group on all measures (F>24.1; P<0.001). Tukey's HSD test was then used to compare schizophrenia and schizophrenia spectrum PD participants' scores with control participant scores on each measure. Schizophrenia spectrum PD participants had higher levels of perceptual distortions (P<0.001), magical ideation (p<0.001), and social anhedonia (P<0.02) than control participants. Schizophrenics had higher levels of perceptual distortions (P<0.01), magical ideation (P<0.001), and social anhedonia (P<0.001) than control participants. Patients with schizophrenia had more severe social anhedonia (P<0.001) than schizophrenia spectrum PD participants. There were no significant differences between schizophrenia spectrum PD and SZ participants for perceptual distortions or magical ideation. 3.3. Correlation and factor analyses Spearman correlation coefficients were evaluated between the five component NEO-FFI and the three Chapman Scales in each group. Coefficients with a two-tailed P-value <0.05 are reported. Table 3 summarizes the results of the correlational analyses examining the relationship between the five components of the NEO-FFI and the three Chapman Scales for each of the three participant groups. Table 3. Correlation among Five-Factor Inventory traits and Chapman Scales for each group Measure N E O A C PAS Control 0.13 0.06 −0.07 −0.38** −0.12 SSPD 0.34 0.03 0.29 0.39 −0.23 SZ 0.38* 0.09 0.20 −0.10 −0.21 All 0.53** −0.14 0.00 −0.22* −0.36** MIS Control 0.17 0.08 0.19 −0.22 −0.16 SSPD 0.28 0.38 0.34 0.33 −0.29 SZ 0.21 0.18 0.30 −0.11 0.02 All 0.50** −0.08 0.11 −0.26** −0.33** R-SAS Control 0.29* −0.45** −0.21 −0.38** −0.24 SSPD −0.39 −0.51* 0.01 −0.43* 0.26 SZ 0.44* −0.29 0.36* −0.05 −0.20 All 0.48** −0.51** −0.15 −0.43** −0.25** PAS: Perceptual Aberration Scale; MIS: Magical Ideation Scale; R-SAS: Revised Social Anhedonia Scale. * P<0.05. ** P<0.01. Table options A factor analysis using both the NEO-FFI scores and the Chapman Scale scores was carried out to determine the relationship among these measures within the entire pool of participants. A varimax rotation with Kaiser normalization was used with a minimum eigenvalue of 1 for factor retention. Two factors were obtained. Factor 1 accounted for 42.5% of the variance of the solution, and Factor 2 accounted for 17.5% of the solution. Table 4 summarizes the test loadings of the solution for each factor. Factor 1 had loadings >0.6 for N, E, A, C and Social Anhedonia. This factor appeared to represent social and personality adjustment, with high factor scores for a well-adjusted, extraverted individual; and low scores for a poorly adjusted, socially withdrawn individual. The second factor had high loadings for magical ideation, perceptual aberration, and O. This factor appeared to represent the degree to which a person endorsed novel or unconventional ideas, superstitious beliefs, and odd perceptual experiences. Table 4. Factor analysis using the NEO-FFI and the Chapman Scale scores Scale Factor 1 2 Neuroticism −0.79 0.31 Extraversion 0.77 0.21 R-SAS −0.72 0.19 Agreeableness 0.65 0.02 Conscientiousness 0.64 −0.19 MIS −0.37 0.84 PAS −0.44 0.78 Openness 0.40 0.48 Factor loadings greater than 0.50 are indicated by bold font. Table options 3.4. Discriminant analyses Discriminant analysis was used to determine which linear combination of psychometric measures best distinguished among control, SZ and schizophrenia spectrum PD participants. Fifty-four control, 30 SZ and 24 schizophrenia spectrum PD participants had complete data on the set of NEO-FFI and Chapman Scales. The discriminant analysis yielded two canonical variables or functions. Each canonical variable is a linear combination of the psychometric variables that maximize the differences between group means in one dimension. Canonical variable 1 accounted for 80.2% of the variance of the analysis (eigenvalue=1.78), while canonical variable 2 accounted for 19.8% of the variance (eigenvalue=0.44). The scores of each subject on the two canonical variables are shown in Fig. 1. In order to interpret this graph, the standardized canonical discriminant function coefficients are provided for each psychometric measure on each canonical variable in Table 5. The coefficient for a psychometric measure indicates the relative contribution of the measure to the canonical variable. These scores indicate that N (0.52), magical ideation (0.49), and social anhedonia (0.38) were positively associated with variable 1, whereas O (−0.53) was negatively associated with variable 1. Consequently, a subject with a high score on canonical variable 1 would tend to have high scores on N, magical ideation, and social anhedonia, and a low score on openness. Canonical variable 2 was positively associated with openness (0.75) and magical ideation (0.43), and negatively associated with conscientiousness (−0.41). The graph of individual scores in Fig. 1 shows that SZ and schizophrenia spectrum PD are distinguished from control subjects on different dimensions. SZ subjects usually had high scores on Function 1, indicating high levels of neuroticism, social anhedonia, and magical ideation. However, they did not differ from control subjects on Function 2. Schizophrenia spectrum PD subjects, on the other hand, were best distinguished from control subjects on Function 2, indicating that schizophrenia spectrum PD subjects had higher levels of O and magical ideation, but lower levels of C compared with control subjects. Subjects with schizophrenia spectrum PD are slightly higher, as a group, on Function 2 as well. These results concur with the univariate analysis, which indicates that participants with schizophrenia spectrum PD have higher scores on O, and lower scores on N and R-SAS than SZ participants. The discrimination function based on these canonical variables correctly classified 88.9% of the SZ subjects, 87% of the SZ subjects, and 71% of the schizophrenia spectrum PD subjects. Plot of discriminant scores for control subjects, patients with schizophrenia, ... Fig. 1. Plot of discriminant scores for control subjects, patients with schizophrenia, and schizophrenia spectrum personality disorder subjects. Group centroids indicated by rectangles. NC: non-psychiatric control subjects; SPD: schizophrenia spectrum personality disorder; SZ: schizophrenia. Figure options Table 5. Standardized canonical discriminant function coefficients Scale Function 1 2 Neuroticism 0.52 −0.23 Extroversion −0.09 −0.09 Openness −0.53 0.75 Agreeableness −0.12 −0.11 Conscientiousness 0.16 −0.41 Perceptual aberration 0.03 0.06 Magical ideation 0.59 0.43 Social anhedonia 0.38 −0.04 Coefficients greater than 0.30 are indicated by bold font.