Letters to the Editor


Principal-Component Analysis of the Positive and Negative Syndrome Scale in Patients With Schizophrenia: Does a 5-Factor Model Apply to Published Data?

Dear Editor:

The simple 2-dimensional model of schizophrenia can no longer account for the clinical heterogeneity observed in schizophrenia symptoms; more satisfactory solutions would retain at least 3 factors. We diagnosed 72 outpatients, using the Composite International Diagnostic Interview (CIDI), according to DSM-III-R and ICD-10 criteria (n = 58, schizophrenia; n = 14, schizoaffective disorders). Of these, 68.1% were men, and 31.9% were women. The subjects’ mean age was 40.1 years (SD 10.37), and mean duration of illness was 16.1 years (SD 9.82).

We performed a principal component analysis over correlation matrix of the Positive and Negative Syndrome Scale (PANSS), and we rotated the initial factor solution, using the Varimax procedure. This analysis led us to a 5-factor solution that defines 5 independent dimensions: negative, disorganization-excitement, reality distortion, anxiety-depression, and hostility. Our results are similar to those of 9 recent principal-component analysis studies that find a 5-dimension solution to factor analyses of PANSS in schizophrenia patients (1–8).

Regarding the negative dimension, comparison has confirmed that 5 of our first-factor items are grouped together in the 9 studies. Conversely, in the 9 studies, conceptual disorganization (P2), poor attention (G11), and difficulty in abstract thinking (N5) constitute the disorganization dimension, whereas in our study N5 belongs to the negative dimension, and P2 and G11 are grouped together with excitement (P4), disturbance of volition (G13), and grandiosity (P5).

For the reality distortion dimension, our results are for the most part similar to the compared studies, with P1 and P6 found in 8 out of 9 studies, and P3 in 6 out of 9. However, we do not locate unusual thought content (G9) in this dimension, while 5 out of 9 studies do.

We locate anxiety (G2), tension (G4), and depression (G6) in an anxiety-depression dimension, as do, respectively, 7 out of 9, 5 out of 9 and 6 out of 9 studies.

To us, hostility (P7) and tension (G4) represent the hostility dimension. Most studies include excitement (P4) (8 out of 9) and uncooperativeness (G8) (8 out of 9). Item differences in this dimension are not surprising, because our patients were free and consented to be rated over 3 half-days.

The negative dimension is obviously the most robust dimension examined throughout the studies. Conversely, analysis of the schizophrenia-related reality distortion dimension seems more complex. Conceptual disorganization is the main symptom in the disorganization-excitement dimension (P2); cognitive disorders might be linked to this dimension, depending on the chronicity and severity of schizophrenia (4). Anxiety-depression and hostility dimensions shed new light on schizophrenia psychopathology. They seem to be complementary to the 3 main dimensions. Anxiety and depression could be 2 distinct levels in 1 dimension in the same way as are excitement and hostility.

Acknowledgement

We thank Lilly France for its financial support.

References

1. Kay SR, Sevy S. Pyramidal model of schizophrenia. Schizophr Bull 1990;16:537–45.
2. Dollfus S, Petit M, Lesieur P, menard JF. Principal-component analyses of PANSS and SANS–SAPS global ratings in schizophrenic patients. European Psychiatry 1991;6:251–9.
3. Lépine JP, Piron JJ, Chapatot E. Factor analysis of the PANSS in schizophrenia patients. In: Stefans CN, Soldatos CR, Rabavillas AD, Editors. Psychiatry today: accomplishments and promises. Amsterdam (NE) : Excerpta medica; 1989.
4. Bell MD, Lysaker PH, Milstein RM, Beam-Goulet JL. Concurrent validity of the cognitive component of schizophrenia: relationship of PANSS scores and neuropsychological assessments. Psychiatry Res 1994;54:51–8.
5. Kawasaki Y, Maeda Y, Sakai N, Higashima M, Urata K, Yamaguchi N, Kurachi M. Evaluation and interpretation of symptom structures in patients with schizophrenia. Acta Psychiatr Scand 1994;89:399–404.
6. Lindenmayer JP, Bernstein-Hyman R, Grochowski S. A new five-factor model of schizophrenia. Psychiatr Q 1994;65:299–322.
7. Dollfus S, Petit M. Principal-component analyses of PANSS and SANS-SAPS in schizophrenia: their stability in an acute phase. European Psychiatry 1995;10:97–106.
8. Lançon C, Auquier P, Llorca PM, Martinez JL, Bougerol T, Scotto JM. Psychometric properties of the PANSS in a sample of French schizophrenic patients. Encéphale 1997;XXIII:1–9.

Olivier Canceil, MD
Marcos Sampaio-Meireles, MD
Marie- France Poirier-Littre, MD, PhD
Marie-Chantal Bourdel, IE
Jean-Pierre Olie, MD
Dominique Attar-Levy, MD
Paris, France
Véronique Olivier, MD
Saint-Cloud, France