Thomas A. Ban: Peralta, Cuesta and their associates’ findings on the higher familiality of Leonhard’s classification in polynosologic study


Victor  Peralta’s response to Thomas A. Ban’s essay and Marcelo Cetkovich, Ernst Franzek, Carlos Morra and Larry Stein’s comments


I congratulate Thomas A. Ban and all the commentators of our paper for their fine and insightful comments about the role of Leonhard’s classification in research in genetics and psychopharmacology. I think that the historical perspective presented by Thomas A. Ban provides a rare and precious opportunity for revisiting Leonhard’s classification and its far-reaching implications for research in endogenous psychosis.   

I am very grateful to all commentators for their comments about our manuscript. In my response, I will focus first on the difference between our findings and those of Leonhard and his followers regarding the familial load of systematic and unsystematic schizophrenias. Second, I’ll present new (unpublished) data from our study on the differential response to antipsychotic treatment in the systematic and unsystematic schizophrenias.   

Our findings that systematic schizophrenias have more familial aggregation than unsystematic schizophrenias are at odds with the findings of Leonhard and his followers. Although there are numerous factors that could explain the disagreement, I will discuss only four.


1. Family methodology

Ours was a study with multiple affected families and such a study is different from the common family and twin studies. Therefore, it is not known to what extent our findings are comparable to the common studies of familial aggregation.  High density family studies are more powerful means than common family studies, but less powerful than twin studies, for detecting familial/genetic underpinnings. Another advantage of examining multiplex families is the higher statistical power they have for detecting meaningful effects than common family and twin studies. Considering the high number of multiplex families in our study, it might be stated that our study had more statistical power than the studies of Leonhard and his followers.


2. Sample representativeness

Leonhard’s twin registry included heterogeneous populations (as acknowledged by Franzek & Beckmann, 1988) and consequently, the reported morbidity risks for systematic schizophrenias were very different in different samples (as acknowledged by Ungvári, 1985). Furthermore, the twin study by Franzek and Beckmann (1988), although much more rigorous than the one by Leonhard, might have been somewhat biased because it was based on hospital registers. I believe that the higher familial aggregation of unsystematic schizophrenias than systematic schizophrenias has been consistently demonstrated only for the rare condition of periodic catatonia. 

It might be also important that the samples of Leonhard and Franzek & Beckmann were hospital-based while our sample was mental health service-based. Therefore our results are more representative of the population of psychotic disorders in the general population. In fact, in our sample, of the 1094 affected subjects, 216 (19.7%) had never been hospitalized and 65 (6%) had never been under psychiatric care. We had a couple of families in which most of the affected members with systematic schizophrenia had never been admitted to psychiatric facilities.


3. Blindness to proband diagnosis

While Leonhard and his followers did not diagnose blindly, in our study, the diagnoses were blindly performed regarding subject identity and group status of both probands and relatives in 75% of the pedigrees.


4. Diagnostic procedure

Ernst Franzek and Marcelo Cetkovich raised some doubts about our diagnostic procedure related to Leonhard’s system. In this respect it should be clearly understood that in our study all subjects were personally interviewed by trained clinicians and, as stated in our paper, Leonhard’s check-list was supplemented by other clinical information provided by the CASH in order to achieve a diagnosis based on Leonhard’s classification. Course of the disorder was a central criterion for diagnosing all types of psychoses and, for diagnosing systematic schizophrenia, particular attention was paid to mode of onset, continuous course and presence of enduring symptoms. Furthermore, all available information was reviewed by Dr. Cuesta and me to diagnose patients according to the specifications of the different diagnostic systems. Considering the mean time of illness duration in our sample (16.5 years), it is rather unlikely that course of the disorders had not been adequately considered. Indeed, the course pattern of (a) symptoms, (b) syndromes, (c) disorders and (d) other illness-related features was carefully assessed over the illness duration.     

Additional support for our adequate diagnostic procedure for Leonhard’s system is provided by the following:


1)      The close correspondence between our comparative prevalence rates of systematic schizophrenia, unsystematic schizophrenia and cycloid psychosis and those reported by Leonhard and others (affective psychoses do not apply because we only included psychotic forms). 

2)      The good fit between the associated clinical characteristics of Leonhard’s classes of psychotic disorders and those reported by Leonhard and others. 

3)      The similarity between our empirically derived “core schizophrenia” class and Leonhard’s systematic schizophrenia in terms of external validators, such as familiality estimates, clinical characteristics and diagnostic concordance. The same also applies to the diagnosis of deficit schizophrenia, a diagnosis that fits well within Leonhard’s systematic schizophrenias.  

4)      The similarity of our data to those of Franzek & Beckmann (1998) regarding the correspondence between Leonhard’s and DSM diagnoses.


In view of these considerations, I firmly believe that our diagnostic procedure was in line with that of Leonhard’s. I must acknowledge, however, that in about 10% of the subjects there were some doubts about their allocation within the Leonhard system, and given that Leonhard’s classification does not provide a residual category of “unspecified psychosis,” they were allocated by consensus to  the “best fitting diagnostic class” in the system. It should be noted that even higher rates of doubtful diagnoses (15%) were encountered in the DSM and ICD classifications. In summary, I believe that the difference between our findings and the findings of Leonhard and his followers is more related to sample representativeness and family study methodology than to diagnosing. Regardless, our findings will need to be corroborated by other investigators using representative samples of psychotic disorders in the general population. 

A further and strictly personal caveat concerns as to the nature of Leonhard’s systematic catatonias. Reading Leonhard’s description of his patients, one wonders whether some of his patients with these diagnoses had a “functional” psychotic disorder at all or suffered  from an unknown or non-demonstrable brain disorder (for example: encephalitis, brain anoxia during gestation, etc.). But this is an issue that could be the subject matter of another exchange.               

To examine the differential antipsychotic treatment response in the systematic and unsystematic schizophrenias, we analyzed the data of patients with systematic (n=242) and unsystematic schizophrenia (n=210) from their last admission, using the CGI efficacy index. The level of treatment response is shown below in the Table:



Treatment effect (%)






Systematic Schizophrenias





Unsystematic Schizophrenias





                                                           Chi-squared=117.617, df=3, p < 0.0001




This Table shows that about two thirds of systematic schizophrenias are unresponsive to antipsychotic medication and that the reverse is the case for unsystematic schizophrenias, i.e., two thirds of unsystematic schizophrenias are responsive to antipsychotic drugs.

When we examined the efficacy index dimensionally in the two populations we found that the mean scores in systematic and unsystematic schizophrenias were 2.98 (SD=0.80) and 2.07 (SD=0.74), respectively (t=12.448, df=450, p<0.0001).  

These findings fully corroborate those of Fish and Astrup as reviewed by Thomas Ban. Furthermore they demonstrate that Leonhard’s systematic and unsystematic schizophrenias are different homogeneous classes of disorders in terms of familial-genetic factors and response to antipsychotic drugs.  





Franzek E, Beckmann H. Different genetic background of schizophrenia spectrum psychoses: a twin study. Am J Psychiatry. 1998;155:76-83.


Jablensky A. Classification of nonschizophrenic psychotic disorders: A historical perspective. Curr Psychiat Rep. 2001;3(4):326-331.


Leonhard K. The Classification of Endogenous Psychoses. 5th ed. Ed E. Robins. New York: Irvington Publishers; 1979.


Leonhard K. Differentierte Diagnostik der Endogenen Psychosen unter Anlehnung an einem Symptommenkatalog. Psychiat Neurol Med Psychol 1990;42(3):136-145.


Ungvári G. A contribution to the validity of Leonhard's contribution of endogenous psychoses. Acta Psychiatr Scand. 1985;72(2):144-149.


Victor Peralta

November 17, 2016