INTRODUCTION

A number of recent reviews, including one by Bernstein et al.^[1▪] and another by the author of the present review ^[2▪], have found widespread dissatisfaction with the classification of personality disorders in the previous editions of the Diagnostic and Statistical Manual of Mental Disorders (DSM) up to DSM-IV-TR. Major shortcomings of the model include poor inter-rater reliability of personality disorder diagnoses, poor stability over time, poor discriminant validity and poor general coverage of personality disorder as well as poor clinical utility.

As a consequence, the DSM-5 Personality and Personality Disorders Work Group has decided to move away from the classification of personality disorders adopted in the previous editions of the DSM and to investigate new ways of classifying personality disorder. Dimensional ^[3▪▪], hybrid (dimensional–categorical) ^[4▪▪] and severity models ^[5] are prominent among alternate models that have been considered to replace the existing categorical system, resulting in a major reconceptualization of personality psychopathology. In the end, the Work Group opted in favour of a hybrid model.

The first draft of a new model was posted on the DSM-5 Website (http://www.dsm5.org) on 10 February 2010. This version was updated on 21 January 2011. A second version, which differed from the first on several important points, was posted on the Website on 21 June 2011. This version was updated again on 1 May 2012. A final version is scheduled to be posted on the Website by the end of 2012, following the results of the field trials, and in advance of the publication of DSM-5, which is scheduled for May 2013.

The focus of the present review is on the development of a particular aspect of the new model for classifying personality disorders in DSM-5, namely the decision to retain or to remove specific personality disorder types from the classification. It reviews the arguments presented by the DSM-5 Work Group to remove half of the personality disorders included in DSM-IV-TR and presents reactions elicited by those decisions and published between February 2010 and October 2012.

MATERIALS AND METHOD

The DSM-5 Website was reviewed for presentations and discussions of the new model that have been posted on the site between 10 February 2010 and 30 September 2012. In addition, PubMed was systematically searched for presentations of – and reactions to – successive drafts of the new model during the same period, using the search words ‘DSM-5’, ‘personality disorders’ and ‘personality’.

Personality disorder types retained in – or removed from – DSM-5

The limitations of the DSM-IV-TR classification of personality disorders, including in particular poor validity and poor clinical utility of the individual personality disorders, has led the DSM-5 Personality and Personality Disorders Work Group to remove several of the DSM-IV-TR personality disorders from the classification. To decide which personality disorders to retain and which to remove, the Working Group followed the ‘Guidelines for making changes to DSM-V’ as proposed by Kendler et al.^[6▪] and posted on the DSM-5 Website in 2009.

According to the criteria proposed by Kendler et al.^[6▪], changes in DSM-5 should be made on the basis of literature reviews and secondary data analyses that document the clinical utility and validity of such changes. In particular, addition of a new diagnosis or the decision to delete a diagnosis should be based on evidence and clear rationales. The prime candidates for deletion from the DSM-5 should be those categories that have low clinical utility and minimal evidence for validity.

The first criterion, clinical utility, can be broken down into several components, including the frequency of use of the diagnosis, its importance in making treatment decisions, its role in stimulating the development of clinical programmes and increasing attention to the diagnosis in professional and lay groups, and the magnitude of adverse effects for patients that would arise from the deletion of the syndrome.

The second criterion to be considered is the overall quality of information about the validity of the category. Although it is unrealistic to expect all extant diagnoses in DSM-IV to meet the criteria that would be required for a new diagnosis in DSM-5, these criteria nonetheless provide a helpful goal towards which to move with subsequent DSM editions.

As stated by the chair of the DSM-5 working group ^[4▪▪], each personality disorder in DSM-IV-TR was the subject of a literature review performed by Work Group members and advisors. The reviews revealed that antisocial/psychopathic, borderline and schizotypal personality disorders had the most extensive empirical evidence of validity and clinical utility, whereas there were almost no empirical studies focused explicitly on paranoid, schizoid or histrionic personality disorders. In his review, Skodol ^[4▪▪] highlights literature relevant to the retention versus deletion of each personality disorder in DSM-5. According to the author, most DSM-IV-TR personality disorders suffer from the problem of excessive co-occurrence with other personality disorders (i.e. poor discriminative validity), but the relative weight of evidence of clinical utility and external validity favours retention of some of these disorders more than others.

With regard to the guidelines set up by Kendler et al.^[6▪], the DSM-5 work group on Personality and Personality Disorders decided to retain only half of the personality disorder types listed in DSM-IV-TR. Schizotypal, borderline, antisocial, narcissistic, avoidant and obsessive–compulsive personality disorders are slated for inclusion in DSM-5. Narcissistic personality disorder, first slated for removal, was reinstated later on. Paranoid, schizoid, histrionic and dependent are no longer considered for inclusion as specific personality disorders in DSM-5. Depressive and passive-aggressive (negativistic) personality disorders – listed in Appendix B of DSM-IV-TR among criteria sets and axes provided for further studies – will not be introduced in the official part of DSM-5.

DSM-IV-TR personality disorders not retained as specific personality disorder types (paranoid, schizoid, histrionic and dependent personality disorders), the depressive and passive-aggressive (negativistic) personality disorders listed in Appendix B of DSM-IV-TR and the residual category of Personality Disorder not Otherwise Specified (PDNOS) will be diagnosed as personality disorder trait-specified (PDTS) in DSM-5. Disorders diagnosed as PDTS will be characterized by mild impairment or greater on the Levels of Personality Functioning Scale combined with descriptive specifications of patients’ personality trait profiles.

On the whole, the types that are retained as specific personality disorder types represent types that are considered to represent particularly salient configurations or interactions of traits, in contrast to the remaining disorders, which can be largely modelled using fewer traits, often from a single, specific trait domain.

General reactions to the decisions of the DSM-5 work group

A number of authors favour the retention of all or most of the personality disorders included in DSM-IV-TR. Reasons for opposing the deletion of some of the current categories include inadequate literature research, high frequency of use and significant clinical validity of personality disorders slated for removal, the fact that the remaining personality disorders would not be able to encompass the spectrum of personality pathology seen in the community, the fact that the removal of some personality disorders would not significantly reduce the comorbidity of the remaining personality disorders and the frequency of psychosocial morbidity associated with personality disorders slated for removal.

Mullins-Sweatt et al.^[7▪] surveyed members of two personality disorder associations, the Association for Research on Personality Disorders (ARPD) and the International Society for the Study of Personality Disorders (ISSPD), with respect to the utility, validity and status of each DSM-IV-TR personality disorder diagnosis. Findings indicate that the proposal to delete five of the personality disorders lacks consensus support within the personality disorder community. The personality disorders used frequently to very frequently by a majority of the respondents were the borderline (92%), antisocial (61%), narcissistic (57%) and avoidant (51%). The personality disorder diagnoses used infrequently to very infrequently by a majority of the respondents were the schizotypal (69%) and schizoid (69%). A majority of the respondents felt that all but one of the personality disorder diagnoses probably or definitely should not be deleted. Support for retention was particularly strong for the borderline (94%), antisocial (89%), dependent (75%), paranoid (74%), narcissistic (73%) and avoidant (72%) diagnoses.

In a commentary on the first draft of the proposal to remove five of the DSM-IV-TR personality disorders in DSM-5, Shedler et al.^[8] object that the five personality types proposed by the Work Group are insufficient to encompass the spectrum of personality disorder seen in the community and that combinations of the proposed dimensional trait rating scales will not easily yield the omitted syndromes, for which a significant amount of empirical data and an even more significant body of clinical wisdom have accumulated over the past decade. As such, the authors strongly advocate that the system be expanded to encompass the range of personality syndromes seen in the community and identified empirically.

The main goal of the Work Group to limit the number of specific personality disorders was to reduce comorbidity. In a report from the Rhode Island Methods to Improve Diagnostic Assessment and Services (MIDAS) project, Zimmerman et al.^[9] examined the impact of eliminating five of the DSM-IV personality disorders on the prevalence of personality disorders in a large sample of psychiatric outpatients presenting for treatment. The comorbidity rate dropped from 29.8 to 21.3%. Compared with patients without a personality disorder, the patients with either a retained or an excluded personality disorder had greater psychosocial morbidity. There was little difference in psychosocial morbidity between patients with a retained and an excluded personality disorder. On the whole, comorbidity would not be eliminated and the reduction of comorbidity could come with a cost of false-negative diagnoses. As such, the results do not provide unambiguous support for deleting any of the personality disorders included in DSM-IV-TR.

Zimmerman et al.^[10] have examined the association of the various DSM-IV-TR personality disorders with indices of psychosocial morbidity. The authors tested the hypothesis that the disorders recommended for retention in DSM-5 would be associated with more severe morbidity than the disorders recommended for deletion. A total of 2150 psychiatric outpatients were evaluated with semi-structured diagnostic interviews for DSM-IV Axes I and II disorders and seven measures of psychosocial morbidity. The results of the study do not provide clear evidence for the preferential retention of some personality disorders over others on the basis of their association with indices of psychosocial morbidity.

Bornstein ^[11] has objected that scrutiny of studies cited in support of retaining versus deleting specific personality disorders from DSM-5 has revealed difficulties in three areas: (1) inadequate information regarding parameters of the literature search; (2) mixed empirical support for proposed changes; and (3) selective attention to certain disorders and not others. According to the author, limitations in the research base cited by the DSM-5 Work Group illuminates gaps in the personality disorder literature and may serve as a starting point for systematic research on personality disorder so that adequate empirical data are available to decide which syndromes to retain, revise or remove in future versions of the diagnostic manual.

Pilkonis et al.^[12] believe, despite many reservations about the current format, that there is a persuasive argument for the retention of types. They propose to retain all the categorical types proposed since DSM-III. The DSM-5 Work Group should not rush to supplant the existing categories with a new personality type system, but develop a rationale and tools for effective translations from traits to types. According to the authors, the justification for many of the current and earlier personality disorder types was often made on clinical grounds, and it would be important to acknowledge the potential value of those contributions while developing new dimensional models. In this regard, greater inclusiveness could be a virtue leading to greater eventual consensus.

Reactions to deletion of specific personality disorders

The decision to remove several personality disorder types from DSM-5 has not gone unnoticed. In particular, experts in a given specific personality disorder have reacted to the decision not to include this personality disorder in DSM-5.

Narcissistic personality disorder

In the first drafts of the Chapter on Personality Disorders, narcissistic personality disorder had not been retained among the specific personality disorder types to be included in DSM-5. The proposal met with considerable controversy that ultimately led to the reinstatement of the disorder.

Ronningstam ^[13] has reviewed evidence in favour of keeping narcissistic personality disorder as a specific personality disorder type with a set of separate diagnostic criteria in DSM-5. The evidence includes significant prevalence rates of narcissistic personality disorder, extensive clinical and empirical reports, and psychiatric, social and societal significance of the disorder, especially when associated with vocational and interpersonal impairment, social and moral adaptation, and acute suicidality. As a consequence, the author recommends retaining narcissistic personality disorder as a distinct personality disorder type in DSM-5. She does, however, recommend also promoting an informative and guiding conceptualization of the disorder and including in its definition traits that capture basic enduring indicators.

According to the same author (E. Ronningstam, this issue), the new diagnostic approach to narcissistic personality disorder taken by the DSM 5 Working Group captures clinically relevant psychological features of pathological narcissism that go beyond grandiosity and the immediate external functioning. As such, it takes into consideration that some aspects of narcissistic disorder are more externally noticeable and less readily identified by the patient himself/herself, whereas others are more readily identified by the patient himself/herself and maybe less so by others.

On the basis of a review of the pertinent literature of the last 12 years, Alarcón and Sarabia ^[14] take a different stance from the preceding author. According to these authors, narcissistic personality disorder is rare when compared with other personality disorders, clinical descriptions of the condition vary within a wide range of descriptors, and comorbidity with many Axes I and II conditions is frequent. They conclude that narcissistic personality disorder shows nosological inconsistency and that it should be considered as a trait domain instead of as a type or disorder.

A similar position has been taken by Karteruda ^[15]. The results of a study involving 2277 patients, 80% of whom had a personality disorder, have led the authors to challenge the notion of narcissistic personality disorder as a distinct diagnostic category. They conclude that their data justify the deletion of narcissistic personality disorder as a distinct category from DSM-5.

Dependent personality disorder

Objections to deleting dependent personality disorder have been reviewed in detail by Bornstein ^[16]. The author argues that dependent personality disorder has adequate clinical utility and frequency of use to merit inclusion as a distinct personality disorder category in DSM-5. In particular, dependent personality disorder has proved useful in predicting risk of parasuicide, perpetration of child abuse, treatment compliance and health service use in children, adolescents, young adults and old adults. In addition, the base rate of dependent personality disorder diagnoses in inpatient and outpatient settings is comparable to that of the majority of personality disorders proposed for inclusion.

The arguments presented by Bornstein did not, however, succeed in getting the disorder back into the latest draft of DSM-5.

Histrionic personality disorder

Bakkevig and Karterud ^[17] have investigated crucial aspects of histrionic personality disorder as defined in DSM-IV-TR, including its prevalence, comorbidity with other personality disorders, internal consistency, severity indices and factor analyses. Prevalence of the disorder was very low, comorbidity was high, especially with borderline, narcissistic and dependent personality disorders, and internal consistency was low. As a consequence, the authors propose not to retain the histrionic personality disorder category in DSM-5, but to redefine the essence of the disorder and to include it as an exhibitionistic attention-seeking subtype of narcissistic personality disorder. This would allow what they call the current ‘grandiose male-dominated’ narcissistic personality disorder to be supplemented by a more female-dominated attention-seeking subtype, described by a behaviour pattern of (1) explicit attention-seeking; (2) outer-directedness, attentiveness and directedness towards others; (3) self-dramatization; (4) dependency on others’ attention, approval and affiliation; (5) presentation of self through sexualized means; and (6) easily emotionally triggered.

Paranoid and schizoid personality disorder

No published objections have been found to removing paranoid and schizoid personality disorders from DSM-5.

Reactions to not including depressive personality disorder and passive-aggressive (negativistic) personality disorder

Huprich ^[18] has reviewed the guidelines proposed by Kendler et al.^[6▪] for the inclusion or exclusion of a diagnostic category in DSM-5 as they relate to depressive personality disorder. The author concludes that there is considerable support for depressive personality disorder as a specific diagnostic category. The conclusions emphasize the clinical utility that depressive personality disorder appears to hold among clinicians and highlight empirical findings justifying continued interest in the disorder, including further investigations of its biogenetic origins, its phenotypic manifestations (including its trait profile) and possible characterization as an endophenotype.

No published objections have been found to not introducing passive-aggressive personality disorder in DSM-5.

Reactions to keeping any of the specific personality disorders

The WHO's ICD-11 is developing a proposal for classifying personality disturbance that differs significantly from the one taken by the developers of DSM-5. Peter Tyrer ^[19▪▪], who chairs the Working Group for the Revision of Personality Disorders in the World Health Organization's ICD-11, advocates a proposal that he himself describes as ‘radical’. Tyrer objects to the retention of any specific personality disorder type, including the six specific personality disorders currently slated for inclusion in DSM-5.

According to the author, there is a lack of support from empirical research for any one of these conditions. While the DSM-5 Personality and Personality Disorder Work Group is developing a complex hybrid categorical–dimensional model that retains six specific personality disorder types, the ICD-11 Working Group is promoting a simple dimensional model that rests upon essentially two components: the measurement of severity and the identification of trait domains. There are five levels of severity: (1) no personality disturbance, (2) personality difficulty not qualifying as a disorder, (3) personality disorder, (4) moderately severe personality disorder and (5) severe personality disorder. Trait domains qualify the nature of the disturbance but are not essential for the diagnosis. Four or five trait domains are currently considered: social/schizoid, dissocial/externalizing, anxious dependent/internalizing, obsessional/anankastic and emotional distress/instability. According to Tyrer ^[19▪▪], the ICD-11 proposal has great clinical utility, reflects the dimensional nature of personality disturbance, eliminates the comorbidity of personality disorder and makes personality disorder a less stigmatized condition.

Major objections to retaining any personality disorder types at all have also been raised by Livesley ^[20], who, in a recent commentary, wonders whether the retention of discontinuous types ‘is merely a reflection of the personality disorder establishment doing what establishments do - maintaining the status quo’.

DISCUSSION AND CONCLUSION

The proposal to retain six of the DSM-IV-TR personality disorders in the forthcoming DSM-5 and to remove the four remaining ones from the future classification has met with considerable controversy. This has led the DSM-5 Personality and Personality Disorders Work Group to reconsider several of its initial decisions. In particular, the Work Group has reinstated narcissistic personality disorder among the specific personality disorder types slated for inclusion in the new classification.

In fact, the DSM-5 revision of the Chapter on Personality Disorders goes far beyond retaining or removing some of the DSM-IV-TR personality disorders. It introduces a new model for making personality disorder diagnoses that involves a radically new approach to diagnosing personality disorder types that departs in many ways significantly from the one that prevailed in all previous DSMs, including the most recent revision, DSM-IV-R. The new model introduces a hybrid, categorical–dimensional approach to the diagnosis of personality disorder. In this context, the retention or removal of several specific personality disorder types is just one of several major new elements included in this approach and not the main issue. The decision about what personality disorder types to retain and what types to remove has nonetheless led to a controversy that is not likely to abate in the near future.

Editor's comment

This article was written before the APA's final last minute decision concerning the diagnosis and classification of Personality disorders in DSM-5.

Acknowledgements

None.

Conflicts of interest

There are no conflicts of interest.

REFERENCES AND RECOMMENDED READING

Papers of particular interest, published within the annual period of review, have been highlighted as:

• ▪ of special interest
• ▪▪ of outstanding interest