5th Congress Autism-Europe
Articulos / Proceeding
Autism-Spain

A Comparison between Diagnostic-Classificatory Systems for Pervasive Developmental Disorders: A study of twenty cases

G. Lanzi, U. Balottin, C. A. Zambrino, A. Gerardo, E. Bettaglio, P. Manfredi.

Child Neuropsychiatry Division, C. Mondino Foundation, IRCCS, University of Pavia, Italy.

INTRODUCTION

In the area of Pervasive developmental disorders (PDD), as in other neuropsychiatric clinical fields, diagnosis should be based on a systematic assessment of clinical findings, pattern of onset and course, etiology and pathogenesis. General agreement among workers on these aspects would provide the basis for a more global classification system. However, within the field of these disorders, there is in reality general disagreement on the models that should be applied, since there is a lack of well-substantiated experimental evidence on which they can be based (Rutter & Schopler, 1992). lf there is agreement for autism, which most workers accept to be a syndrome with an organic basis about which however we lack specific knowledge, it is all the more true as regards the other PDDS.

As a result of this situation classificatory systems in which the descriptions are based only on symptomatological and behavioural aspects have had great success: DSM-III (APA, 1980), DSM-III-R (APA, 1987) and ICD-10 (WHO, 1992). This type of approach is principally characteristic of the psychiatry of the English-speaking world, but these classifications are now more and more generally accepted.

Among those authors who recognise the use of such classificatory systems as being the only possible model in view of the present state of our knowledge, there is however hot debate aimed at their improvement: DSM is in fact an on-going project with an Advisory Commitee which still meets. The task facing those who work on the problem is far from being easy, since a classificatory system must satisfy a large number of objectives which are often difficult to reconcile, because clinical and research needs are often in contrast.

It is very important that there should be a rigorous and consistent basis for research and epidemiology so that syndromes may be grouped together or differentiated on the basis of similar observational categories. This is the point of view, for instance, that underlies the criticisms made of ICD-10 by Waterhouse et al. (1992) and Giliberg (1994) who note that the WHO manual subsumes categories like Childhood autism (CA) or Atypical autism (AA) under PDD on the basis of simple behavioural aspects, and at the same time creates sub-categories for Rett syndrome and Childhood disintegrative disorder. As regards these latter, Rett syndrome can in some cases correspond to CA from a symptomatological point of view, while Childhood disintegrative disorder implies aspects which can be ascribed to neurological deficfts. According to Waterhouse et al. (1992) the inclusion of these two subgroups should suggest an additional internal axis for diagnosis within the PDD. Rutter and Schopler (1992) replied that from a clinical point of view these subgroups can be differentiated on the basis of their course, and that it would be "absurdly restrictive" to base a diagnostic system on a purely behavioural approach, a cross-sectional clinical picture. They maintain that it would be wrong to exclude Rett syndrome and Childhood disintegrative disorder from PDD: these disorders present with a similar pattern and the patients have similar needs.

Another topic of fierce debate is the sensitivity and the specificity of the different systems. It is now well known that DSM-III-R tends to overdiagnose autism, since its inclusion criteria are rather wide, while DSM-III is more restrictive and ICD-10 intermediate (Factor et al. 1989; Hertzig et al. 1990; Spitzer & Siegel 1990; Volkmar et al. 1992); since a perfect system is not possible, it should be decided a priori whether a tendency to produce false negatives or false positives is preferable, particularly in those countries where access to treatment is based on the label given to the syndrome (Szatmary 1992).

The very term PDD has been questioned (Baird et al. 1991; Gillberg 1991; Happé & Frith 1991) along with the necessity of having further subgroups within the heterogeneous category of PDD (Tsai 1992; Volkmar 1992, Waterhouse et al. 1992; Rutter & Schopler 1992; Castelloe & Dawson 1993; Gillberg 1994; Rutter et al. 1994).

The interminable debate about symptomatological classifications - and we have reported only some aspects which seem important to us - has led to the preparation of DSM-IV (Draft Version, APA 1993). As well as a different choice of items compared to the previous version, this includes the subgroups Rett syndrome and Childhood disintegrative disorder within PDD, and, provisionally, Asperger's disorder, thus moving into line with ICD-10. Furthermore the manual specifically indicates that atypical forms of autism should be classified as Pervasive Developmental Disorder not otherwise specified (PDDNOS), thus supporting the possible continuity between autism and PDDNOS.

Child psychiatrists of the French schools are not involved in this debate, and criticise the premises which inspired the DSM and ICD10. Misès et al. (1988) maintain that DSM-III is not at all

atheoretical, but rather that it is in line with the dominant paradigms of North American psychiatry, which is cognitivist and behaviourist. They believe that is is much more useful to try to produce a classificatory system which takes into consideration the mobility or the fixedness of the disturbances, the nature of the changes that underlie the symptom and which develop through time, the risk of sliding into an adult mental pathology, all of which based on precise psychopathological criteria, the psychoanalytical framework which is traditional in the French child psychiatry. Other workers have supported this position of Misès et al. (Lebovici 1988; Bursztein et al. 1990; Touati et al. 1990), and have introduced a number of themes which are in clear opposition to the premises which led to the APA and WHO classifications: the analysis of the structural organisation and its disturbance which lie behind the symptom, the importance of the family context, the revaluation of subjectivity and intuition as valid clinical tools, the concept of psychosis as a psychopathological continuum with its own course, in contraposition to the definition of PDD and to the "tendency to place autism and childhood psychoses within a deficitary dimension" (Touati et al. 1990).

The French classification of the mental disorders of children and adolescents (CFTMEA, Misès et al. 1988) is offered as an alternative proposal to the DSMs and ICD-10 in the light of the considerations reported above; it reproposes the term "Childhood psychoses" in the place of PDD and leaves the psychiatrist free to make a diagnosis on the basis of his experience, without the limit of a checklist.

We believe that an approach that is as open as possible to the different schools of thought as regards diagnosis and classification is important for our clinical and diagnostic practice. The combined use of the different classificatory systems which we have discussed in this introduction and in a previous paper (Bettaglio et al., in press) could give additional information about individual cases. For us DSM-III-R and ICD-10 are an agile tool for initial clinical screening, and allow comparisons to be made with data reported in the international bibliography, while CFTMEA and the Operational Classification proposed by Manzano and Palacio-Espasa (1983) are indispensable because of their psychopathological orientation, which is often close to that of the Services to which our patients are referred for treatment.

This study compares the diagnoses obtained using the above-mentioned classificatory systems (DSM-III-R, ICD-10, CFTMEA and the Operational Classification of Manzano and Palacio-Espasa) on a population of 20 paediatric subjects; the recently published draft version of DSM-IV is also used and the qualitative differences compared to the previous version are tested.

MATERIALS AND METHODS

We examined 20 children, 17 males and 3 females aged between 2 years 9 months and 7 years 7 months (average age 5 years 2 months), admitted consecutively to the Child Neuropsychiatry Division of the Universfty of Pavia for a diagnostic and therapeutic assessment. The protocol used for the diagnostic assessment of the patient inciudes: anamnestic investigation, neuropsychiatric examination, a battery of instrumental and laboratory tests (routine hematochemical tests, lactic acid dosage, thyroid hormones, immunogiobulins for the TORCH complex, plasma and urine amino acids, caryotype for fragile X, EEG awake and sleeping, visual and brainstem auditory evoked potentials, cranial CT or MR). On the basis of our findings we applied the diagnostic criteria proposed by DSM-III-R and assumed as the criterion for inclusion in our study a diagnosis of PDD, either of the subtype AD or PDDNOS. We then compiled the BSE scale (Barthélémy et al. 1990) on the basis of at least one week's observation during hospitalisation. We also assessed cognitive development using Griffiths' Mental Development Scales (Griffiths 1986). We also performed a standardised psychodynamic observation (SPO) according to the following protocol:

- video recorded observation of a semi-structured interview with the mother (or mother and father) and the child;

- video recorded observation of a series of sessions of the child whith the neuropsychiatrist;

- team viewing of the films with interpretation of the behaviour according to fixed parameters.

On this basis we elaborated a structural diagnostic matrix according to Palacio-Espasa, in function of the SPO and with reference to the findings of each of the parameters.

In the light of this work-up we first verified the diagnoses that had previously been made according to DSM-III-R, and then we applied four other diagnostic classifications: the recent DSM-IV, ICD-10, CFTMEA, and the classification proposed by Manzano and Palacio-Espasa.

RESULTS

According to DSM-III-R, 12 of the 20 subjects examined were diagnosed as AD and the remaining 8 as PDDNOS. According to DSM-IV only 7 were DA and 13 were PDDNOS, of which 5 were diagnosed as AA; DSM-III-R had classified these five cases as AD. ICD-10 found 8 cases of CA, 4 AA, and 8 Other syndromes of global alteration of psychological development. CFTMEA found only 2 cases of Kanner's Precocious CA, 6 as Other forms of CA, 6 as Precocious deficitary psychosis, and 6 as Psychotic disharmony. According to Manzano and Palacio-Espasa's classification there were 2 cases of Primary autism, 2 of Secondary autism, 5 of Disorganising psychosis (typical form), and 6 of Precociously deficitary psychosis, while for 5 cases it was not possible to formulate a diagnosis since according to the criteria they presented as transitional forms, and were therefore without the necessary and sufficient characteristics for insertion into a precise diagnostic category.

DISCUSSION

The criterion for inclusion of subjects in our sample was a diagnosis of PDD following DSM-III-R. This classification subdivides PDD into AD and PDDNOS. Our experience in evaluating our cases was that the two diagnostic criteria are well differentiated. In particular the description of the single items with regard to AD and the criteria that are specified for assignment to this category allowed us to differentiate this group quite clearly; all the same there are children in this group who are classified differently by the other classifications we considered, and these are above all the cases that Misès et al. and Manzano and Palacio-Espasa call Precocious deficitary psychoses. Only in two cases did AD according to DSM-III-R correspond to the same diagnosis in all the other systems (AD for DSM-IV, CA for ICD-10, Kanner's Precocious CA for CFTMEA, and Primary autism for Manzano and Palacio-Espasa). Together these describe a serious picture of autism which is deeply rooted. Our quantitative (importance of the disorder) and qualitative (underlying intrapsychic and relational functioning) work-up showed us that the cases defined as AD within DSM-III-R presented, beyond the symptomatological aspects, deeply differing characteristics. Indeed it was possible to show different profiles and residual areas of cognitive and relational functioning which were not equivalent. This finding seems to us not only to be fundamental in a diagnostic sense, but also in prospect of treatment program which will be different in relation to the potentialities and residual areas of functioning of each child.

The category PDDNOS is substantially defined by a criterion of exclusion. The qualitative impairment in the development of reciprocal social interaction and verbal and non-verbal communication skills which marks PDDNOS remains vague. The inclusion of all those forms which cannot be diagnosed as AD within PDDNOS means that within this subgroup there are clinical situations that differ widely also from the behavioural point of view, exactly because the qualitative characteristics of the impairment are vaguely defined. The behavioural differences are confirmed by further investigation of the cognitive, affective and relational aspects. The Psychotic disharmony described by Misès et al. and Manzano and Palacio-Espasa's Disorganising Psychosis (typical form) both fall within this category.

Similar considerations may be made concerning DSM-IV, in as much as the parameters for inclusion in AD and PDDNOS do not seem to have been formally changed. However AA is included within the latter category, which covers the cases which do not fulfil the criteria for AD because of a later age at onset of the symptomatology, an atypical or less serious symptomatology or for all of these reasons. In this study we therefore used this diagnostic category to differentiate those cases in which the quantitative and qualitative impairment of the social interaction appeared to be clinically less serious compared to those which were classified as AD or in whom the age of onset was of over three years. For this reason, in order to make our observational criteria more objective, as well as using clinical assessment we also based ourselves on the scores in the BSE scale, particularly with regard to those items that Barthélémy et al. (1990) consider to be specific indicators of the autistic symptomatology. Following these criteria only 7 of the 12 children who were diagnosed as AD according to DSM-III-R were assigned the same diagnosis by DSM-IV, while the other 5 were defined as AA and therefore assigned to PDDNOS. This use of DSM-IV thus leads to an underestimate of the autistic disorder compared to both DSM-III-R and ICD-10. We therefore believe that a is necessary that the definition and way that this category is to be used should be specified better, as is the case in ICD-10, which differentiates AD qualitatively from AA (in which the deficitary aspects prevail) and does not leave the diagnostic classification as arbitrary (atypical symptomatology, subthreshold symptomatology). In other words we believe that it would be more useful if AA was a further subgroup of PDD than to further widen the heterogeneity of the patterns that are classifiable as PDDNOS.

For the cases we studied PDDNOS/AA always corresponds to CA according to ICD-10; it seems to be comparable to Other forms of CA in the CFTMEA classification and covers most of the cases which could not be defined by the Manzano and Palacio-Espasa system in as much as they presented as transitional forms. PDDNOS is however a category defined by exclusion, which subsumes situations which are very different also behaviourally. The cases diagnosed as PDDNOS by DSM-III-R were assigned the same diagnosis by DSM-IV.

As regards DSM-IV our quantitative and qualitative work-up showed that AD itself inciuded cases that differed in their mental and relational functioning. This differentiation was even more obvious when we considered those cases which were classified as PDDNOS.

In general our experience shows that DSM-IV allows a more precise clinical definition compared to DSM-III-R, in as much as it introduces new diagnostic categories which it is perhaps worth getting to know better through more long-term use and thus greater experience.

In the same way ICD-10 operates a distinction between Autistic disorder and Other syndromes of global alteration of psychological development. The criteria for the definition of CA are substantially similar to those proposed by the APA classifications, even if they are understood as reference points for a more general diagnosis. In the same way the definftion of Other syndromes of global alteration of psychological development could be assimilated to that of PDDNOS in DSM-III-R and IV, in as much as their inclusion criteria are not specified and they are differentiated by exclusion from the other forms described. In our cases PDDNOS always corresponded to this category in ICD-10. AD corresponds to CA and AA as defined by ICD-10, this latter characterised by a prevalence of cognitive deficits or by a specific language impairment, so that it is thus more similar to the Precocious deficitary psychosis describes by Misès and to Precociously deficitary psychosis in Manzano and Palacio-Espasa's system. None of the cases we considered could be classed in the other categories described by ICD-10. Our work-up as regards mental and affective relational functioning substantially confirmed these considerations, contributing to further differentiate the children.

CFTMEA seems to be very accurate in identifying the different types of childhood psychosis, and as well as the behavioural aspects it also concentrates on certain aspects of the mental functioning of the subject, so that its definftion of the category goes beyond purely symptomatological aspects. As we have already said, in our cases Precocious childhood autism corresponds to AD, the Other forms of childhood autism to AA in DSM-IV, and Psychotic disharmony to PDDNOS, while our cases of Precocious deficitary psychosis go under AD, apart from n.7 and n.17 who did not present autistic symptoms. This classification, unlike the others, includes a second axis which covers associated or previous, and perhaps etiologic, factors. It therefore allows cases to be classified on this basis, that is also to show up particular family or development situations, and the presence of organic aspects, aspects which we considered when working-up our clinical cases.

We adopted Manzano and Palacio-Espasa's method of assessing intrapsychic aspects: the classification which they propose is specifically designed to define each case dynamically. However it was often difficult to classify individual patients according to their diagnostic categories. For five patients a diagnosis was not made using the Manzano and Palacio-Espasa system since they represented situations which, in terms of their approach, were characterised and differentiated by the presence of elements which were developing, that is they were transaional between forms. Highly specific definitions are difficult to adapt to clinical situations which are often not so clear. In our cases Primary autism corresponded to Precocious lnfantile Autism (CFTMEA) and AD; Secondary autism and Disorganising psychosis are very specific definitions although for some aspects they can be related to Other forms of childhood autism and Psychotic disharmony (CFTMEA), respectively.

CONCLUSIONS

Our data lead us to make a number of considerations, and these are primarily methodological: it is useful to to adopt tools which allow different levels of assessment, so as to be able to generate an overall picture which is different for each patient and constitutes not only a diagnostic assessment, but also gives indications for treatment and prognosis. We have tried to show up the different characteristics of the classification systems which we used, differences which also depend on the different theoretical premises on which they are based. Our clinical experience shows the usefulness of DSM-III-R, DSM-IV and ICD-10 for initial screening, in view of their high sensitivity. Further study on the use of DSM-IV is necessary; at this point the possibility of greater differentiation, not only between different situations but also in relation to the degree of impairment, seems to us to be clinically useful. The other two systems were necessary for the specific classification of all those forms of childhood psychosis which are not described by the former classifications and to understand the functioning of each patient: CFTMEA meets these needs and also allows the enhancement of the associated or anamnestic factors that are present, while clinical use of the system proposed by Manzano and Palacio-Espasa seems to be complex, as it presupposes a long and in-depth psychodynamic training.