Principal-component analyses of PANSS and SANS-SAPS in schizophrenia: their stability in an acute phase

Ear F’sychiafty (1995) 0 Elsevier. Paris

10.97-106

97

Original

article

Principal-component analyses of PANSS and SANS-SAPS in schizophrenia: their stability in an acute phase S Dollfus I, M Petit2 ‘Department 2Department

of psychiatry,

of psychiatry, Centre University of Rouen, (Received

Esquirol, CHU CBte de Nacre, Centre Hospitalier du Rouvray,

8 October

1993;

accepted

3 December

I4000 Caen, France; 76301 Sotteville-les-Rouen,

France

1993)

Summary -Whether studies agree or disagree on the positive-negative dichotomy in schizophrenia, the relevance of a third component, disorganization, remains a point of debate. Disorganization, as expressed by the scale for the assessment of negative symptoms and positive symptoms (SAN%SAPS) and the positive and negative syndrome scale (PANSS) principal-component analyses, could be considered as permanent and determinant a dimension as the positive and negative components. The aim of this study therefore was to determine whether this disorganization, with the negative and positive components, is stable and has the same composition in the acute and postacute phases of illness. This study was carried out in 57 patients, broadly defined by at least one of four diagnostic criteria (American Psychiatric Association, Langfeldt, Carpenter and Schneider), established with a computerized checklist, and evaluated with SANS-SAPS and PANSS. Principal component analyses (PCA) of these scales were performed at admission and discharge from hospital. The PCA of SANS-SAPS displayed a 3-factor solution, regardless of the phase of illness (acute or postacute), showing that the negative, positive and disorganization components were stable. The PCA of PANSS yielded negative and positive components perfectly stable over time and a disorganization component whose composition varied between admission and discharge. At admission, this component included the conceptual disorganization item negatively correlated with one of depression. At discharge, this disorganization component included two additional items, autistic preoccupation and mannerisms and one depression component appeared. The instability of the PCA of PANSS could express the role played by the phase of illness; in an acute phase, this disorganization component was constituted by more “positive” items such as grandiosity, unusual thought content and active social avoidance whereas in the postacute phase, it included items that reflected more the chronicity of the illness, such as mannerisms and autistic preoccupation. Moreover, the depressive item appeared, in the postacute phase, in a specific depressive component. This result could be due to the fact that depressive symptoms cannot be expressed when positive symptoms are very severe. which explains why no depressive components were shown during the acute phase. negative

and positive

symptoms

/ disorganization

dimension

INTRODUCTION Several authors have defined symptomatic dimensions in schizophrenia and have hypothesized that there are 2 forms of schizophrenia, based on distinct clusters of positive and negative symptoms (Strauss et al, 1974; Crow, 1980). The development of specific scales such as the scales for assessment of positive and negative symptoms (SANS-SAPS) of Andreasen and Olsen (1982) and the positive and negative syndrome scale (PANSS) of Kay et al (1987) has stimulated many clinical studies designed to test the validity of this dichot-

/ principal-component

analyses

I follow-up

/ schizophrenia

omy (Opler et al, 1984; Rosen et al, 1984; Bilder et al, 1985; Comblatt et al, 1985; Andreasen and Grove, 1986; Kay et al, 1986; Liddle, 1987; Kay and Singh, 1989; Kay and Sevy, 1990; Mortimer et al, 1990; Dollfus et al, 1991). Principal component analyses (PCA) of symptomatic scales tend to find unrelated negative and positive components, regardless of the scale used (Lewine et al, 1983; Bilder et al, 1985; Andreasen and Grove, 1986; Kulhara et al, 1986; Liddle, 1987; Lepine et al, 1989; Kay and Sevy, 1990; Mortimer et al, 1990; Dollfus et al, 1991; Bell et al, 1992; White et al, 1992; Peralta et al, 1992). But, according to

98

S Dolltirs, M Petit

Andreasen and Olsen (1982), positive and negative symptoms belong to the same dimension and are inversely linked together. Whether most studies agree on the absence of correlation between the two syndromes of the negative-positive dichotomy or not, one point still under debate is the relevance of a third component, disorganization. Indeed, several authors have emphasized the importance of this (Bilder et al, 1985; Gibbons et al, 1985; Liddle et al, 1989; Mortimer et al, 1990). This component had been displayed, in particular, with the principal-component analyses of SANS-SAPS (Andreasen and Olsen, 1982; Kulhara et al, 1986; Arndt et al, 1991; Dollfus et al, 1991; Gur et aE, 1991; Peralta et al, 1992) and PANSS (Dollfus et al, 1991). In these studies, the analyses were performed on patients in different phases, acute or stabilized but at times the phase of illness was not specified. On the whole, the data obtained with the SANSSAPS were homogeneous, showing at least three components: negative, positive and disorganization. This last component always included the positive formal thought disorder in six studies, four of them included bizarre behavior (Andreasen and Olsen, 1982; Kulhara et al, 1986; Dollfus et aE, 1991; Gur et al, 1991) and three of the studies included inattentiveness (Dollfus et al, 1991; Gur et al, 1991; Peralta et al, 1992). If in their first study, Andreasen and Olsen (1982) had obtained this particular component, they made no mention of it, suggesting only the hypothesis of a continuum of the positive and negative symptoms. In a later study (Andreasen and Grove, 1986), no disorganization dimension was shown. Concerning the PANSS, the principal-component analyses have always displayed two negative and positive components and a third one, ie, disorganization. This component has always included the item, conceptual disorganization, whereas the other items which constituted it differed from one study to another. In addition to conceptual disorganization, five of the six studies isolated a cognitive component composed of disorientation and/or difficulty in abstract thinking and/or mannerisms and posturing and/or poor attention (LCpine et al, 1989; Kay et Sevy, 1990; Bell et al, 1992; Risperidone study, 1992; Lindenmayer et al, 1994). The sixth study (Dollfus et al, 1991) yielded a disorganization component including conceptual disorganization, motor retardation, unusual thought content, disturbance of volition and autistic preoccupation. The discrepancies between the results of Dollfus et al (199 1) and the other 5 studies could be due to differences in the inclusion criteria. Indeed, for Doll-

fus et al’s study, a majority of patients were in an acute phase and the diagnostic criteria were broader than for the others. Thus, this disorganization component, regardless of the scales used (SANS-SAPS or PANSS) is reminiscent of Bleuler’s fundamental symptoms (loss of associations, autism). It could therefore be considered as an independent, permanent and determinant dimension associatedwith the positive and negative syndromes of schizophrenia. The aim of this study therefore was to determine whether this disorganization component, with the negative and positive components, is a stable one. We also addressedthe question of whether the disorganization component, which could be a characteristic of the chronicity of schizophrenia, as Kay and Sevy (1990) suggested, is found in patients in an acute phase as well as in patients in a postacute phase. This hypothesis was tested in schizophrenic patients during their hospitalization and in patients broadly defined by at least one out of four diagnostic criteria: two mainly symptomatic (Schneider, 1959; Carpenter et aE, 1973), the others characterized by duration of illness (Langfeld, 1960; American Psychiatric Association (DSMIII-R), 1987). Moreover, the hypothesis tested was based on the analysis of the two scales most used for positive and negative symptoms, the SANS-SAPS (Andreasen and Olsen, 1982) and the PANSS (Kay et al, 1987); the first one rates only specifically schizophrenic symptoms, the second one rates other general symptoms such as anxiety and depression. PATIENTS

AND METHODS

This prospective study involved a cohort of schizophrenic patients in an ongoing follow-up study. Only patients who were hospitalized for a relapse or for the first onset of the illness were included. The inclusion criteria were as follows: after a consensus between two senior psychiatrists, any patient presenting or having presented a psychotic state (ie, delusions or hallucinations and/or negative symptoms such as blunted affect, grossly disorganized behaviour or social withdrawal) not related to a physical illness or a major depressive or manic disorder (DSMIII-R) was selected. The patients participating in the study were then interviewed by one of the study team (SD), who had been previously trained in the use of different clinical instruments: A 183-item standardized computerized checklist was filled in (Dollfus et al, 1994). This checklist, inspired by the integrated lists of criteria for the taxonomic evaluation of non-affective psychoses (LICET-S) of Pull et al (1981), was extended to permit diagnosis of schizophre-

Stability of PANSS and CANS-SAPS in schizophrenia nia by 14 different diagnostic systems and by chronic hallucinatory psychosis, as defined by Pull et al (1987a,b). In this study, only the following four diagnostic systems for schizophrenia were used: American Psychiatric Association (DSMIII-R) (1987), Schneider’s (1959) with at least one current or past first rank symptom, the flexible system of Carpenter et al (1973) with a cut-off point of 5, and that of Langfeldt (1960). Clinical evaluations using different scales were performed at admission and discharge from hospital by the same examiner (SD). The scale for the assessment of negative symptoms (SANS) (Andreasen and Grove, 1986; French translation by Lecrubier and Boyer, 1987), a 25item scale subdivided into 5 subscales; the scale for the asessment of positive symptoms (SAPS) (Andreasen and Grove, 1986; French translation by Boyer and Lecrubier, 1987), a 34.item scale subdivided into 4 subscales. Each subscale includes a global rating which is highly correlated to each individual item on the subscale (Andreasen et al, 1982; 1986); The positive and negative syndrome scale (PANSS) (Kay et al, 1987; French translation by Lkpine 19891, a 30-item scale includes a 7-item positive scale, a 7-item negative scale and a 16-item general psychopathology scale. The doses of neuroleptics were converted into equivalent chlorpromazine according to Ban (1971) and Foster (1989). The four diagnoses, the clinical evaluations and the statistical analyses were computerized using PCSM (programme conversationnel de statistiques pour les sciences et le marketing, 1990) software only after the last pre-selected patient was interviewed: Student’s tests were made to compare the means of positive and negative scores at admission and discharge; a type R principal-component analysis (PCA) was used, based on correlations between symptoms (de Bonis et al, 1987; Reuchlin, 1987). Factor analyses were performed, twice, at admission and at discharge, on the full range of the 30 PANSS items, and separately on the nine global ratings of the SANS-SAPS, as usually carried out for PCA by Andreasen and Olsen (1982) and Kulhara et al (1986). Using the PCSM, Pearson’s product-moment correlation coefficients were computed and a principal-component procedure was used to extract the initial factors. The predetermined criterion chosen for the number of factors to be extracted was the commonly used eigenvalue criterion, whose factors are retained if they have an eigenvalue 2 1. Following the initial extraction of factors, orthogonal rotation via the varimax procedure was used to achieve the simplest and most meaningful factor structure. Orthogonal rotation assumes that the underlying factors are not correlated (Moorey et al,

Table

99

I. Main population characteristics (n = 57). mean f standard deviation

Age

Age at first hospitalization (years) Duration of illness (years) Number of previous hospitalizations Duration of previous hospitalizations (months) Duration of the current hospitalization (months)

40.3 29.5 9.2 3.4

-*- 13.9 f 13.9 + 9.4 * 4.8

9.8 + 23.6 2.1 + 2.1

1991). Only the items loading > 0.50 were retained. As for the 30 items of PANS& the orthogonal rotations were made on the first 5 factors after checking the eigenvalues, since Lindenmayer et al (1994) have shown the subdivision of PANSS in 5 subscales. A measure of the internal consistency of the PCA was computed with the BMDP software by Carmines’ theta, a special case of Cronbach’s alpha (Carmines and Zeller, 1979).

RESULTS Population characteristics Table I shows the main characteristics of the sample including 19 men (33.3%) and 38 women (66.7%) aged 40 years (m +- SD = 40.3 + 13.9). The mean duration of illness was 9.2 +- 9.4 years. The mean duration of current hospitalization was 2.1 2 2.1 months. All patients received neuroleptics during this period. For eighteen of the 57 patients (3 1.6%), it was the first onset of illness. Diagnosis The distribution of the sample according to diagnostic system was as follows: DSMIII-R, 47.4% (n = 27); Langfeldt, 50.8% (n = 29); Carpenter, 78.9% (n = 45); Schneider 87.7% (n = 50). Of the twenty-seven of the DSMIII-R patients, three (11.1%) were catatonic, twelve (44.4%) were disorganized, ten (37%) were undifferentiated and one (1.7%) was paranoid. Only one did not have any acute phase criteria,

but those of a residual

one.

Clinical evaluations Table II shows the mean of the different global ratings of SANS-SAPS and of the subscores of PANSS at admission (TO) and discharge (TI). All of the scores significantly decreased over time but the doses of neuroleptic

nificantly.

did not vary sig-

100 Table

S Dollfus, II. Symptomatic

evaluations

SAPS (global ratings) Hallucinations Delusions Bizarre behavior Positive formal thought

at admission

(TO) and at discharge

TI (m ic SD)

2.8 3.2 0.7 1.7

k 1.9 +_ 1.5 + 1.2 f 1.6

0.7 0.9 0.1 0.4

+ f f It

1.0 1.0 0.3 0.9

< lo-4

2.0 1.6 2.6 2.6 2.5

k * f + +

1.6 1.4 I.6 1.6 1.5

1.7 1.2 2.0 2.2 1.3

f * + + f

1.3 1.1 1.3 1.4 1.0

< 0.05 < 0.01

20.2 17.9 35.8 896.4

f + + f

6.2 8.2 6.4 1127.9

10.6 14.8 26.4 888.5

+ + + f

3.4 6.6 7.1 1081.02


SANS (global ratings) Affective flattening Alogia Avolition/apathy Anhedonia I asociality Attention

m k SD: mean 2 standard

deviation;

* Student’s

COMPONENT

P*


< I@3

< 0.01
test; NS: not significant.

Neuroleptic treatments Table III shows the different neuroleptics prescribed at admission. Only nineteen patients (33.3%) received one neuroleptic. Seven patients (12.3%) were treated by “stimulant” neuroleptics such as sulpiride, amisulpride, pimozide and pipotiazine whereas 34 patients (59.6%) received haloperidol and 20 (35%) received chlorpromazine or levomepromazine. PRINCIPAL

(Tl).

TO (m + SD)

disorder

PANSS (subscores) Positive Negative General psychopathology Equivalent-chlorpromazine

M Petit

ANALYSES

(PCA)

SANS-SAPS

Table

III.

Type of neuroleptics

Neuroleptics Amisulpride Chlorpromazine Cyamemazine Fluphenazine Fluphenazine oenanthate Haloperidol Haloperidol decanoate Levomepromazine Pimozide Pipamperone Pipotiazine Propericyazine Sulpiride Thioproperazine

prescribed

at admission

(TO).

N

%

1 10 13 5 2 34 2 10 2 1 3 1 1 3

I .75 17.54 22.8 1 8.77 3.51 59.65 3.51 17.54 3.51 I .75 5.26 1.75 1.75 5.26

Admission At admission, the PCA of the SANS-SAPS global ratings (table IV-a) yielded a 3-factor solution which explained 69.9% of the total variance (table V-a). Carmines’ theta was 0.77. The first factor (33.9% of the variance) embraced all the negative symptoms except attentional impairment. The second factor (19.7% of the variance) included only two positive items, hallucinations and delusions, with large positive loading. The third factor (16.3% of the variance), disorganization, was composed of one positive symptom, positive formal thought disorder, and one negative symptom, attentional impairment.

PANSS

Discharge At discharge, the PCA of the nine global-ratings of the SANS-SAPS also showed a 3-factor solution

Admission At admission, the PCA of the PANSS (table VI-a) disclosed a 5-factor solution which accounted for

explaining 67.5% of the total variance (table V-b) and the Carmines’ theta coefficient was 0.8. The negative and positive components had exactly the same composition as those observed at admission. The third factor could also be called disorganization, including positive formal thought disorder as at admission, but here associated with bizarre behaviour (table IV-b)

Stability Table Iv. Principal ratings at admission

component analyses and at discharge.

items a) Admission (TU) Afktive flattening Alogia Avolition / apathy Anhedonia / asociality Halhlcinations Delusions Positive formal thought disorder Attentional impairment b) Discharge (‘II) Affective flattening Alogia Avolition / apathy Anhedonia I asociality Hallucinations Delusions Positive formal thought disorder Bizarre behavior

FI

-

of PANSS

of SANS-SAPS

FII

and SANS-SAPS global

FM

0.88 0.70 0.88 0.92 0.82 0.88 0.83 0.72 0.92 0.78 0.86 0.89 - 0.77 - 0.82 0.87 0.77

59.5% of the total variance and included a negative component (18.9% of the variance), a disorganization component (11.4% of the variance), an excitation one (10.4% of the variance), a positive one (10.1% of the variance) and an anxiety component (8.7% of the variance). The different loadings of the items on these 5 components are shown in table VI-a and Carmines’ theta was 0.88. The negative component had five negative symptoms and 3 items of the general psychopathology subscale: motor retardation, disturbance of volition and autistic preoccupation. The disorganization component was negatively loaded by 4 items: conceptual disorganization, grandiosity, unusual thought content and poor attention. Two items were positively loaded: depression and active social avoidance. The third component could be called excitation because it contained excitement, hostility and uncooperativeness with one more item, stereotyped thinking. The fourth component was a positive one because it included delusions, hallucinatory behaviour, suspiciousness/persecution and lack of judgment and insight. The last component, an anxiety one, had somatic concern, anxiety and tension items. Discharge At discharge, the PCA of the PANSS disclosed a 5-factor solution explaining 58% of the total variance (table VI-b) with a Carmines’ theta at 0.88.

101

in schizophrenia

Table V. Explained variance by each component rotation) for the PCA of SANS-SAPS.

component

8 exphined variance

(after

varimax

% cumulative variance

a) Admission (TO) Negative Positive Disorganization

33.9 19.7 16.3

33.9 53.6 69.9

b) Discharge (Tl) Negative Positive Disorganization

35.7 16.1 15.7

35.7 51.8 67.5

The first component, a negative one (19.2% of the total variance) had exactly the same composition as that observed at admission. The second component, excitation (12.8% of the total variance) included 4 items (excitement, hostility, uncooperativeness and stereotyped thinking) which constituted the third component at admission. Three other items (somatic concern, anxiety and tension) were also loaded on this component. The third positive component (9.7% of the total variance) had the same composition at discharge as that observed at admission; it was characterized by delusions, hallucinatory behaviour, suspiciousness/persecution and lack of judgement and insight. The fourth disorganization component (9.02% of the total variance) included three items already found at admission (conceptual disorganization, unusual thought content and poor attention). Two other items were loaded on this component: autistic preoccupation and mannerisms. The fifth depressive component (7.3% of the variance) included guilt feelings, depression and disorientation.

DISCUSSION In this study, fifty-seven patients meeting at admission the criteria of at least one out of the four diagnostic systems for schizophrenia were included and were followed-up throughout their hospitalization. The relatively small size of this sample could have influenced the outcome. Nevertheless, the fact that Carmines’ theta coefficient was high (0.77 to O.SS), whatever the PCA, meant that the internal consistency of the components was good and thus gave good validity to our results. Among the four diagnostic systems used to

102 Table

S Dollfus, Via.

Principal

component

analyses

lrem Blunted affect Emotional withdrawal Poor rapport Passive/apathetic social withdrawal Lack of spontaneity and flow of conversation Motor retardation Disturbance of volition Autistic preoccupation

of PANSS

M Petit

(admissioin). FI

Fli

Fill

FV

FIV

- 0.85 - 0.91 - 0.65 - 0.82 -0.64 - 0.58 - 0.68 - 0.11

Conceptual disorganization Grandiosity Depression Unusual thought content Poor attention Active social avoidance

- 0.75 - 0.67 0.53 - 0.64 - 0.55 0.56

Excitement Hostility Stereotyped thinking Uncooperativeness

0.54 0.79 0.58 0.78

Delusions Hallucinatory behavior Suspiciousness/persecution Lack of judgment and insight

-

0.81 0.74 0.73 0.50

Somatic concern Anxiety Tension % explained

variance

-0.71 - 0.87 .- 0.81 18.9

define the schizophrenic sample, DSMIII-R was found to be the most restrictive one. This is in agreement with most studies comparing different sets of diagnostic criteria (Helzer et al, 1981; Stephens et al, 1982; McGlashan, 1984; Endicott et al, 1986; Dollfus et al, 1993a). Only about half of the fifty-seven patients were selected by the DSMIII-R, and about one third (31.6%) were suffering the first onset of their illness, which underscored the broad definition of this sample. All the global ratings of SANS-SAPS and of PANSS subscoresdecreased significantly, regardless of the negative, positive or general symptoms. Nevertheless, variation was weaker in 4 negative symptoms of SANS (affective flattening, alogia, avolition/apathy, anhedonia/asociality) than in the others. The absenceof a significant modification of the neuroleptic doses (in equivalent-chlorpromazine) shows that the treatment had been high since the first evaluation (admission). Our results about the stability of negative symptoms diverge from those of Kulhara et al (1990) and Mueser et al

1 I .4

10.4

10.1

8.7

(1991). Kulhara et al (1990) compared the scores of the SANS and SAPS in schizophrenic patients in an acute phase at admission and 18 to 30 months later; the scores of SAPS decreased significantly whereas no difference was observed for the scores of SANS. In Mueser et al’s study (1991), attentional impairment and avolition-apathy were improved by neuroleptics, in contrast to other SANS global ratings, such as affective flattening, alogia, anhedonia-asociality which were stable. In contrast, our results are close to those of Addington and Addington (1991) for whom all the global ratings of SANS-SAPS, except avolition decreased significantly under neuroleptic treatment between the acute phase and 6 months later. Philipps et al (199 1) also observed a highly significant decrease in the severity of all the types of symptoms of SANS-SAPS with standard treatment. These discrepancies could be due to the use of different therapeutic means, since small dosesof stimulant neuroleptics such as amisulpride, sulpiride, pipotiazine, are used by some authors becauseof their efficacy

Stability Table

VIb.

Principal

component

analyses

Items Blunted affect Emotional withdraxal Poor rapport Passive/apathetic social withdrawal Lack of spontaneity and flow of conversation Motor retardation Disturbance of volition Autistic preoccupation

of PANSS

of PANSS

and SANS-SAPS

103

in schizophrenia

(discharge). FII

FI

FIII

0.88 0.81 0.77 0.83 0.83 0.74 0.65 0.61

Excitement Hostility Stereotyped thinking Uncooperativeness Somatic concern Anxiety Tension

FIV

0.58 -

0.58 0.78 0.63 0.83 0.57 0.52 0.61

Delusions Hallucinatory behavior Suspiciousness/persecution Lack ofjudgment and insight

- 0.54 -0.81 - 0.79 - 0.73 - 0.64

Conceptual disorganization Mannerisms Unsual thought content Poor attention

0.69 0.66 0.65 0.54 - 0.69 - 0.54 - 0.58

Guilt feelings Depression Disorientation % explained

FV

variance

19.2

in the negative or deficit forms of schizophrenia (Petit et al, 1987; Pokier-Lit&k et al, 1989; Boyer et al, 1990). In our study, seven patients (12.3%) received such drugs at admission. Moreover, our results are close to those observed by several authors who have described an improvement of negative symptoms by various neuroleptics such as fluphenazine (Lapierre and LavalCe, 1975; Frangos et al, 1978), chlorpromazine or haloperidol (Kay and Singh, 1989), clozapine (Kane et al, 1988) and thiothixene (Tandon et al, 1990). The PCA of the SANS-SAPS global ratings and of the PANS& at admission, yielded a 3 and 5factor solution, with a total explained variance similar to that described in the literature: 69.9% for the SANS-SAPS (Andreasen and Olsen, 1982; Kulhara et af, 1986) and 59.5% for the PANSS (Kay and Sevy, 1990). As we have previously shown (Dollfus et al, 1991), the three components of the SANS-SAPS and three of the five components for the PANSS had a similar clinical meaning, evoking negative, positive and disorganization components.

12.8

9.7

9.02

7.3

When we compared the PCA at admission and at discharge, we observed some similarities and discrepancies for both scales: for the SANS-SAPS, whatever the phase of illness (acute or postacute), the PCA displayed three components: negative, positive and disorganization (tables IV and V). Therefore, our data show the excellent stability of these three components, although the symptomatology decreased significantly (table II). The frost two components (negative and positive) had exactly the same composition at both times of the evaluation and the third component, disorganization, always included an identical item, positive formal thought disorder. Therefore, according to the PCA of SANS-SAPS, the third component, like the negative and positive ones, appears to be determinant in characterizing the symptomatology of schizophrenia and independent of the acute or postacute phase of illness. With regards to the PANSS, three interesting points could be emphasized: i) First of all, the positive and negative compo-

104

S Dollfus,

nents were perfectly stable and had exactly the same items regardless of the time of evaluation: at admission (acute phase) and at discharge (post acute phase). ii) Secondly, the excitation component had the same 4 items at both evaluations, but it acquired anxiety symptoms (somatic concern, anxiety and tension) in the post-acute phase. iii) Thirdly, the disorganization component varied over time between admission and discharge with the emergence of a depressive component. The disorganization component existed at both times, at the acute and postacute phases, but its composition differed, reflecting modifications of the interrelationship between symptoms. At both times, three items were common: conceptual disorganization, unusual thought disorder and poor attention. These items were linked negatively with depression and active social avoidance in the acute phase, but positively linked with mannerism and autistic preoccupation in the postacute phase. The grandiosity and active social avoidance items found in the acute phase disappeared from this component in the post-acute phase, whereas two additonal items appeared, mannerisms and autistic preoccupation, which, along with conceptual disorganization, evoked the discordance described by Chaslin (1912) in France. In the acute phase, the disorganization component included the depression item. It is interesting to note that this item was positively loaded on this component, contrary to conceptual disorganization which was negatively loaded. In other words, there was a negative correlation between depression and conceptual disorganization. In the post-acute phase, this disorganization component did not include a depressive item. This appeared in a specific depressive component including guilt feelings and disorientation items. Therefore, these results could be due to the fact that depressive symptoms cannot be expressed when positive symptoms are very severe. This suggests that no depressive component was displayed in the acute phase. It is only when the positive symptoms have decreased that the depressive symptomatology can appear. These results concur with our earlier ones (Dollfus et al, 1993b) and evoke the revealed depression introduced by Knights and Hirsch (1981) who drew attention to the fact that although depressed symptoms are most prevalent in the acute phase of schizophrenia, they tend to go unnoticed or unappreciated because of florid psychotic symptoms. The modification of components other than the negative and positive ones displayed with the principal-components of the PANSS, could result from

M Petit

the instability of the scale structure, due to a too small sample size. However, the high level of Carmines’ theta (0.88 at admission and at discharge) is an argument against this hypothesis. This instability could also express the role played by the phase of illness, reinforcing the hypothesis, made in a previous study (Dollfus et al, 1991) that the variations of PANSS principal-component analysis results could be due to differences in the inclusion criteria from one study to another. In most studies (Ltpine et al, 1989; Kay and Sevy, 1990; Bell er al, 1992; Risperidone study, 1992; Lindenmayer et al, 1994), which used only the DSMIII-R criteria, inclusion criteria were more restrictive than in our study where only about half of the patients (47.4%) met the DSMIII-R criteria for schizophrenia. Moreover, stabilized patients or those in an acute phase are often mixed in these studies. These two points could explain the discrepancies regarding the composition of this component. In acute phase, this component could be composed of more positive items such as grandiosity and active social avoidance whereas in the post-acute phase, it could include more items reflecting the chronic&y of the illness such as mannerisms and autistic preoccupation, as Kay and Singh (1990) have shown. This instability of the factorial structure of the PANSS could also be due to the presence of general items such as anxiety and depression, which are absent from the SANS-SAPS. These results have highlighted to two essential points: on the one hand, SANS-SAPS and PANSS gathered the positive and negative components which were perfectly stable during the resolution of the acute phase. In contrast, the disorganization component, stable on the SANS-SAPS, was much less stable on the PANSS. This stability of the three components displayed with SANS-SAPS could have a practical involvement in assessing the efficacy of neuroleptic treatments. In most studies, this assessment is based on the modification of the total scores of the SANS and SAPS. However, it would be more interesting to establish this efficacy from the score variations of these three components. Whether the efficacy of neuroleptics on positive and negative symptoms is acknowledged or not (Goldberg, 1985; Kane and Mayerhoff, 1989), their activity on disorganization (characterized in particular by formal thought disorder) might justify further pharmacoclinical investigations. Concurrently, if the anxiety and depressive dimensions have to be evaluated, it would be more logical to use the PANSS; indeed the PANSS includes general symptoms such as anxiety and depression,

Stability

of PANSS

and SANS-SAPS

whereas the SANS and SAPS do not. The importance of the PANSS therefore should not be underrated. It is fundamental in the assessment of the relationship between general symptoms and specific symptoms of schizophrenia.

ACKNOWLEDGMENTS We thank S Noel and H Caputo for their technical We are also grateful to JF Menard for his

assistance. statistical

advice.

REFERENCES Addington J, Addington D. Positive and negative symptoms of schizophrenia. Their course and relationship over time. Schizophren Res 1991;5:51-9 American Psychiatric Association. Diagnostic and Statistic Manual of Mental Disorders (DSMIII-R), Washington 1987. Traduction francaise: Manuel diagnostique et statistique des troubles mentaux. Paris: Masson, 1989;624 Andreasen NC, Grove WN. Evaluation of positive and negative symptoms in schizophrenia. Psychiatry Psychobiol 1986;1,2:108-21 Andreasen NC, Olsen S. Negative versus positive schizophrenia. Arch Gen Psychiatry 1982;39:789-94 Arndt S, Alliger RJ, Andreasen NC. The distinction of positive and negative symptoms. The failure of a two-dimensional Model. BrJPsychiatry 1991;158:317-22 Ban TA. Drug treatment in schizophrenia. Canad Psychiatr Ass J 1971;16:473-85 Bell MD, Lysaker PH, Milstein RM, Beam-Goulet JL. Five factor model of schizophrenia: replication of PANSS factor structure and implications for subtyping. 145th Annual Meeting of the American Psychiatric Association. Washington, D.C., May 1992 Bilder RM, Mukhejee S, Rieder RO, Pandurangi AK. Symptomatic and neuropsychological components of defect states. Schizophren Bull 1985;11:40%19 Boyer P, Lecrubier Y. Fiche descriptive et traduction francaise de la SAPS. Psychiatr Psychobiol 1987$,6:425-37 Boyer P, Lecrubier Y, Puech AJ. Treatment of positive and negative symptoms: pharmacologic approaches. In: Andreasen N. ed. Schizophrenia: positive and negative symptoms and syndromes: Modern Problems of Pharmacopsychiutty. Basel: Karger, 1990;24: 152-74 Carmines EG, Zeller A. Reliability and validity assessment, In: Sullivan JL, ed. Quantitative applications in the social sciences. Beverly Hills: Sage publication, 1979 Carpenter WT. Strauss JS, Bartko JJ. Flexible system for diagnosis of schizophrenia: Report from the WHO international pilot study of schizophrenia. Science 1973; 182: 1275-78 Chaslin Ph. Groupe provisoire des folies discordantes. In: Asselin et Houzeau, eds. Jk%tents de semiologie et clinique mentales. Paris, 19 12 Comblatt BA, Lenzenweger MF, Dworkin RH, ErlenmeyerKimling L. Positive and negative schizophrenic symptoms, attention, and information processing. Schizophr Bull 1985;11,3:397-406 Crow TJ. Molecular pathology of schizophrenia: more than one disease process. Br J Psychiatry 1980;280:668 De Bonis M, Boek P, Lebeaux MO. Methodological problems in subtyping schizophrenics: correlations versus

in schizophrenia

105

hierarchical relationships. Psychiatry Psychobiol 1987; 11,6:425-37 Diagnostic and statistical manual of mental disorders. Third edition, revised (DSMIII-R). American Psychiatric Association, Washington D.C., 1987. Traduction francaise: Manuel diagnostique et statistique des troubles mentaux. Paris: Masson, 1989;624 Dollfus S, Petit M, Lesieur Ph. Menard JF. Principal components analyses of PANSS and SAPS-SANS in schizophrenia. Eur Psychiatry 1991;6:251-9 Dollfus S, Petit M, Menard JF, Brazo P, Besse-Assouly F, Preterre Ph, Lesieur Ph. Polydiagnostic approach of schizoohrenia: validitv of a comouter checklist (LJDE). Encep/i&?, 1994;xx:91-101 Dollfus S, Petit M, Menard JF, Lesieur Ph. Schizophrenia: comparison of 13 diagnostic systems in a cross-sectional study. Eur Psychiatry i993a;8:7-13 Dollfus S, Petit M. Menard JF. Relationship between deoressive and positive symptoms in schizophrenia. J Aflective Disord 1993b;28:61-9 Endicott J, Nee J. Cohen J, Fleiss JL, Simon R. Diagnosis of schizophrenia. Arch Gen Psychiatry 1986;43: 13-9 Foster P. Neurolentic equivalence. Pharm J 1989;30:431-2 Frangos H, Zissis NP, Ieontopoulos I, Diamantas N, Tsitouridis S, Gavril I. Tsolis K. Double blind therapeutic evaluation of fluspirilene compared with fluphenazine decanoate in chronic schizophrenic. Actu Psychiatr Stand 1978; 57~436-46 Gibbons RD. Lewine RRJ, Davis JM, Schooler NR, Cole JO. An empirical test of a Kraepelinian vs a Bleulerian view of negative symptoms. Schizophr Bull 1985;11:39@6 Golberg SC. Negative and deficit symptoms in schizophrenia do respond to neuroleptics? Schizophren Bull 1985; 11,3:4536 Gur RE, Mozley PD, Resnick SM, Levick S, Erwin R, Saykin AJ, Gur RC. Relations among clinical scale in schizophrenia. Am J Psychiatry 1991;148:472-8 Helzer JE, Brockington IF, Kendell RE. Predictive validity of DSMIII and Feighner definitions of schizophrenia. Arch Gen Psychiatry 1981;38:791-7 Kane JM, Honigfeld G, Singer J, Meltzer A. Clozapine in treatment resistant schizophrenia. Arch Gen Psychiatry 1988;45:789-95 Kayne JM, Mayerhoff D. Do negative symptoms respond to pharmacological treatment? Br J Psychiatry 1989;155 (suppl7): 115-8 Kay SR, Fiszbein A, Lindenmayer JP, Opler LA. Positive and -negative syndromes in schizophrenia -as a function of chronicitv. Acta Psvchiutr Scand 1986:74:507-18 Kay SR: Fiszbein A, Opler LA. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophren Bull 1987;13:261-76 Kay SR, Sevy S. Pyramidical model of schizophrenia. Schizophren Bull 1990;16:537-45 Kay SR, Singh MM. The positive-negative distinction in drugfree schizophrenic patients. Arch Gen Psychiatry 1989;46:711-8 Knights A, Hirsch SR. Revealed depression and drug treatment for schizophrenia. Arch Gen Psychiatry 1981;38:80611 Kulhara P, Chandiramani K. Positive and negative subtypes of schizophrenia; a follow-up study from India. Schizophr Res 1990;3:107-16 Kulhara P, Kota SK, Joseph S. Positive and negative subtypes of schizophrenia a study from India. Acta Psychiat Stand 1986;74:353-9 Langfeldt G. Diagnosis and prognosis of schizophrenia. Proc Roy Sot Med 1960;53: 1047-52

106

S Dollfus.

Lapierre YD, LavalICe J. Pimozide and the social behavior of schizophrenics. Curr Ther Res 1975:18.1:181-8 Lecrnbier -Y, Boyer P. Fiche descriptive et traduction fran$aise de la SANS. Psychiatr Psychobiol 1987;11,6:414-24 Lepine JP, Piron JJ, Chapotot E. Factor analysis of PANSS in schizophrenia patients. In: Stefanis CN, Soldatos CR, Rabavilas AD, eds. Psychiatry today: Accomplishments and promises. Amsterdam: Excerpta Medica, 1989 Lewine RRJ, Fogg L, Meltzer HY. Assessment of negative and positive symptoms in schizophrenia. Schizophren Bull 1983;9,3:368-76 Liddle PF. The symptoms of chronic schizophrenia. A re-examination of the positive negative dichotomy. Br J Psychiatry 1987;151:145-51 Liddle PF, Barnes TRE, Morris D, Haque S. Three syndromes in chonic schizophrenia. Br J Psychiatry 1989;155 (suppl7): I 19-22 Lindenmayer JP, Hyman RB, Grochowski S. Five factor model of schizophrenia: initial validation. J Nerv Merit Dis 1994; 182:631-8 McGlashan TH. Testing four diagnostic systems for schizophrenia. Arch Gen Psychiatry 1984;41: 141-4 Moorey S, Greer S, Watson M, Gorman C, Rowden L, Tummore R, Robertson B, Bliss J. The factor structure and factor stability of the hospital anxiety and depression scale in patients with cancer. Br J Psychiafry 1991;158:255-9 Mortimer AM, Lund CE, McKenna PJ. The positive: negative dichotomy in schizophrenia. Br J Psychiurry 1990; 157:41-9 Mueser KT, Douglas MS, Bellack AS, Morrison RL. Assessment of enduring deficit and negative symptom subtypes in schizophrenia. Schizophr Bull 1991;17,4:565-82 -. Opler LA, Katy SR, Rosado V. Lindenmayer JP. Positive and negative sindromes in chronic schizophrenic inpatients. JNervandMentDis 1984;172,6:317-25 Peralta V, de Leon J, Cuesta MJ. Are there more than two syndromes in schizophrenia? A critique of the positive-negative dichotomy. Br J Psychiutry 1992;161:335-43 Petit M, Zann M, Lesieur Ph. Colonna L. The effect of sulpiride on negative symptoms of schizophrenia. Br J Psychiatry 1987;150:270 Phillips MR, Zhao Z, Wxiong X, Cheng X, Sun G, Wu N. Changes in the positive and negative symptoms of schizo-

M Petit phrenic in-patients in China. Br J Psychiatry 1991; 159:226-3 1 Poirier-Littr6 MF, Galinowski A, Peron-Magnan P, Vanelle JM. Raffaitin F. Piron JJ, Piketti M. Loo H. Bipolar effect of pipotiazine according to dosage in negative &d positive forms of schizophrenia. VIII World Congress of Psychiatry, Athenes, 1989;abstract:661 Pull MC, Pull CM, Pichot P. Des critkres empiriques francais pour les psychoses. II. Consensus des psychiatres fran@s et definitions provisoires. Encephale 1987a;XIII:53-7 Pull CB, Pull CM, Pichot P. LICET-S: une liste inttgrke de crittres d’Cvaluation taxinomiques pour les psychoses nonaffectives. J Psy Biol Therap 1981;1,1:33-7 Pull MC, Pull CB, Pichot P. Des crittres empiriques frangais pour les psychoses. III. Algorythmes et arbre de dtcision. Encephule 1987b;5946 Reuchlin M. Prkis de stutistique. 4th ed. Paris: Presses Universitaires de France (PUF), 1987 Risperidone Study Group. Risperidone in the treatment of schizophrenia: a double-blind placebo and controlled active trial. Titusville, New Jersey: Janssen Research Foundation, I992 Rosen WG. Mohs RC, Johns CA, Small NS. Kendler KS, Horvath TB, Davis KL. Positive and negative symptoms in schizophrenia. Psychiatry Res 1984;13:277-84 Schneider K. Clinical psychopathology. New York: Grune and Stratton, 1959 Stephens JH, Astrup C, Carpenter WT. Shaffer JW, Goldberg JA. Comparison of nine systems to diagnose schizophrenia. Psychiatry Res 1982;6: 12743 Strauss JS, Carpenter WT, Bartko JJ. The diagnosis and understanding of schizophrenia. Part III. Speculations on the processes that underlie schizophrenic symptoms and signs. Schizophren Bull 1974;11:61-9 Tandon R, Goldman RS, Goodson J, Greden JF. Mutability and relationship between positive and negative symptoms during neuroleptic treatment in schizophrenia. Biol Psychiatry 1990;27:1323-6 White L, Parrella M, Harvey P, Powchik P, Davidson M. PANSS structure in adult and geriatric schizophrenia. New Research Presentation. Annual Conference of the American Psychiatric Association. Washington, D.C., May 1992