C. Lord

Response to NIH Questions

Question 1: Is there a universally accepted definition of autism? Is there sufficient scientific evidence to support this current definition of autism spectrum disorders as separate from other developmental disorders? For the first time, there are consistent criteria for diagnosis of autism spectrum disorders in both DSM-IV (American Psychiatric Association, 1994) and ICD-10 (International Classification of Diseases, 10th ed., World Health Organization, 1993). The working group agreed that there is both national and international support for these newly published definitions. The precision of these definitions will continue to evolve as new research clarifies the phenotype (visible characteristics of autism). Identification of one or more biological markers for autism disorders is needed to diagnose definitely atypical cases. Strong empirical support in the DSM-IV international Field Trials and other NIH-funded research, however, indicates that the clinical diagnosis of autism remains one of the most reliable diagnoses in psychiatric or developmental research. Additional research is needed to establish the validity of the diagnosis in terms of criteria based on etiology, course, and response to treatment.

Definitions of Rett syndrome (RS) and childhood disintegrative disorder (CDD) also yielded clear, consistent differences from Autism (A) and from other disorders in the DSM-IV Field Trials (cf. Volkmar, following paper) and in other studies. Current definitions appear adequate for estimates of the incidence and prevalence of these disorders in the United States. The new definition of Asperger syndrome (AS) makes the distinction between A and AS clear and so provides the first opportunity to assess the incidence and prevalence of AS separate from A. A category such as Pervasive Developmental Disorder—Not Otherwise Specified (PDD-NOS) is still needed, but would be more appropriately entitled "Atypical Autism" for cases that meet some, but not all of the criteria for autism. Standard measures that yield these diagnoses across the age span from 3 years to adulthood are now available and used widely in research in North America and Europe.

Question 2: How do the characteristics of autism, Asperger syndrome, Rett syndrome, and childhood disintegrative disorder change with development? There are a number of studies in which characteristics of A, RS, and CDD are compared in different age groups of children, adolescents, and/or adults (i.e., cross-sectional studies). However, there are few studies following the same children as they age so we have little evidence of how individuals actually change across the life-span (i.e., longitudinal studies). Asperger syndrome is a term that has been used frequently with adolescents and adults rather than with young children, so we have few cross-sectional or longitudinal data on AS.

Several relatively large-scale, high quality, follow-up studies exist of psychometric (mental measurement test) data for autistic individuals. These studies have shown that the diagnosis of autism continues to apply as the children age and change, even after they have developed language. Longitudinal data are needed for all autism spectrum disorders in order to trace individuals’ paths of development. Such data would enable us to define the course of the disorders, to aid in projecting clinical outcomes, and to plan for clinical intervention at various ages.

Question 3: What other disorders that occur as separate disorders in conjunction with autism (comorbid conditions) must be taken into account in the diagnosis and assessment of autism, Asperger syndrome, Rett syndrome, and childhood disintegrative disorder? Most, but not all persons with A also have some degree of mental retardation. There is now a very clear expectation in behavioral research that degree of mental retardation and severity of language deficit must be considered in designing and interpreting studies of the autism spectrum disorders. These co-occurring factors (mental retardation and degree of language deficit) have received less consideration in biomedical research. Lack of control for mental retardation and language deficits may be one reason why the replication rate of findings (i.e., multiple investigators obtaining the same findings) within biomedical research has typically been much lower than within behavioral research. The importance of diagnosis of comorbid (co-occurring) conditions such as affective disorder (e.g., depression) or obsessive-compulsive disorder, particularly in adults with autism spectrum disorders is also recognized, but standard procedures to do so are not well established. These await further research. It is particularly important to clarify the difference between true comorbidity and other ways in which symptoms from autism and different disorders may overlap.

Question 4: What additional research in autism, its related conditions, or relevant normative behaviors is needed to clarify subgroups in this heterogeneous population? Many investigators have proposed different subgroups in autism and these differ depending upon the theory that the particular research group espouses about the cause and/or central deficit in autism (e.g., language, motor problems, immune and/or serotonergic system differences). However, often the neurobiological or other feature purported to differentiate between or among subgroups is itself of questionable reliability. Studies of subgroups are important, but are premature if they are conducted using approaches that are not reliable or well established. One or more biological markers (e.g., genetic or neurochemical) is urgently needed to distinguish autism from other disorders and to distinguish subgroups within the autism spectrum disorders.

Recommendations of the Working Group on Diagnosis

1. Use of standard diagnostic procedures that operationalize (specify for research DSM-IV/ICD-10 criteria is needed to promote communication among scientists, clinicians, and parents, and replicability across studies. Diagnoses are not being made consistently in clinical or research settings.
     
2. Identification of individuals with AS and the investigation of external factors that validate the distinction between AS and A, and between AS and related learning disabilities is a high priority as well as investigation of the distinctions between severe autism and severe/profound mental retardation.
     
3. A central registry of persons with CDD is needed to enable researchers to find enough well-defined cases for the scientific study of CDD and for clinical information purposes. A standard protocol for clinical investigations of children with disintegrative disorders is also needed.
     
4. Longitudinal studies are crucial in order to provide data concerning differences and similarities in the individual developmental patterns (trajectories) of children with autism, particularly for the highest functioning individuals whose accomplishments may have been underestimated.
     
5. Establishing diagnostic criteria for very young (under age 3) children with autism is an urgent priority. The national trend away from specific diagnoses for young children in favor of generic terms such as "developmentally delayed" will result in a lack of appropriate early intervention services for children with autism. Continued development of reliable screening as well as diagnostic instruments is a critical need.
     
6. Studies that allow more accurate description of adults with autism, particularly those that address issues in the transition from school to work, are a high priority.
     
7. To be useful, studies of subgroups must be hypothesis driven (i.e., address a specific research question) and must validate subgroups using reliable, well-established measures that are not part of the diagnostic features.
     
8. Studies of subgroups that have been replicated across independent centers and across time, and those that address significant aspects of diagnosis such as course, response to treatment and well-defined levels of etiology (causes), pathophysiology (mechanism of structural and functional changes), and behavioral repertoire are a high priority. Studies of sex differences in autism are also clearly needed.
     
9. Minimum standards for comparison groups in studies of autism include comparability on mental handicap, degree of language impairment, and comorbidity (co-occurrence) of other conditions.
     
10. In studies of comorbidity of other disorders such as depression or obsessive-compulsive disorder, there is an urgent need to develop standard, reliable procedures for diagnosis of such disorders in individuals with autism, particularly in those with insufficient verbal skills for typical methods to he employed.
      
11. Research is urgently needed to assess the following interrelated questions: (a) Does having autism and another disorder change the nature and particularly the response to treatment of the comorbid disorders? (b) Are persons with autism more at risk than other people for certain other disorders? (c) Are there symptoms or other disorders not currently included in the phenotype of autism that are so common in autism that they might better be considered part of autism, for example, certain movement disorders?

   first | previous | next | last