Objective: Anorexia nervosa (AN) is a highly debilitating disease which frequently results in chronification and often originates in adolescence. Personality traits have been associated with the onset and maintenance of AN; moreover, study results indicated a worse treatment outcome in patients with AN and comorbid personality disorder (PD). However, research on PD in adolescent AN is scarce. Methods: The sample consists of 73 female adolescent patients with AN. We investigated comorbid PD and AN symptom severity performing the International Personality Disorder Examination (IPDE) and the Eating Disorder Inventory 2 (EDI-2). Results: Almost a third (27.4%) of all participants were diagnosed with comorbid PD. They had significantly higher EDI-2 total scores reflecting overall stronger symptom severity, as well as significantly higher scores in the subscales “ineffectiveness,” “interpersonal distrust,” “interoceptive awareness,” “asceticism,” “impulse regulation,” and “social insecurity.” Conclusion: PD is an important and frequent comorbid condition in adolescent AN and should be addressed in diagnostic and treatment planning. Early diagnosis of comorbidity could have an impact on choosing specialized treatment for adolescents with AN and PD in order to enhance the outcome.

© 2022 The Author(s). Published by S. Karger AG, Basel

Introduction

Anorexia nervosa (AN) is a debilitating and severe psychiatric disorder with a mean crude mortality rate of 5% [1, 2]. It typically develops during adolescence with the highest incidence reported for females between 15 and 19 years [3]. The pathogenesis is postulated to be multifactorial including a combination of genetic, biological, psychological, and sociocultural factors [4-6]. Personality traits are discussed as influential for treatment outcome in adults [7, 8].

From a psychodynamic and personality-focused viewpoint, AN can be defined as a struggle for control, for a sense of identity, competence, and effectiveness [9]. This accounts for the onset of AN in adolescence, when identity formation and individuation play an essential role (Krischer, 2020). The problem involves a wide range of deficits in conceptual development, body image and awareness, individuation, and emotional openness [10, 11]. The symptoms function to maintain the cohesion and stability of a sense of self [12]. The body is used as a metaphor as there is a striking closeness between emotions and bodily experiences [13]. The heightened sensitivity to external visual information of the body overrides internal information such as hunger cues, leading to a representation of the body by external visual and less by internal somatosensory information [14]. Social integration and emotional maturity of these patients are often stunted [15].

Personality traits have been discussed as being connected to the onset and maintenance of AN [16], as AN is characterized by high levels of perfectionism, obsessive-compulsiveness, negative emotionality, harm avoidance, anxiety, problems in social cognition, and body image distortion. Recent studies stress the symptomatic overlap of autism spectrum disorder (ASD) and AN, such as the impairment of social interaction, difficulties with emotion regulation, and restricted behaviors [15, 17], as well as an overlap of symptoms of ASD and personality disorder (PD) [18].

While literature regarding comorbid PD is already scarce for adults, there is a great lack of literature examining adolescents with AN [19]. Over the last decade, it has been increasingly acknowledged that PD does occur and can be validly assessed in children and adolescents [20, 21]. Experts recommend a timelier diagnosis of PD in affected adolescents in order to improve effective treatment [22, 23]. As of today, PD is considered a highly prevalent and increasingly valid diagnosis in adolescence with high clinical implications [21, 24]. Therefore, the classification systems DSM-5 and ICD-11 no longer define age limits and recommend diagnosing and studying PDs as early as the age of 13 years.

The few studies investigating comorbid PD in patients with eating disorders show controversial results with reported rates of comorbid PD in adult patients ranging between 27% and 77% [25]. A study investigating the prevalence of PD in adolescents with AN treated in an outpatient setting showed that almost a quarter of patients had one or more PDs compared with 4% of the healthy control group [26]; the study by Magallon-Neri et al. [27] reported 22–28% of comorbid PD in adolescents with AN. Moreover, comorbid PD traits have been associated with a greater severity of AN in adolescence [26, 28].

A recently published long-term follow-up study reflected on the course of adolescent-onset AN [29]. It has shown a favorable outcome regarding mortality and full symptom recovery. However, one in five had a chronic eating disorder. Researchers tried to define risk factors for poor outcome. The role of comorbid PD in the chronification of eating disorders has not yet been consistently characterized. Although there are various findings, the diagnosis of PD contributed to a greater severity of AN in adults and poorer outcomes following treatment as usual [25, 30, 31]. Current research studies have found a possible connection between treatment response and chronification of symptoms among adult patients with AN and comorbid personality pathology, such as borderline PD (BPD) or obsessive-compulsive PD (OCPD) [8, 32, 33]. A recent meta-analysis showed that the diagnosis of PD had a negative impact on treatment outcome for eating disorders in adults. The authors identified a need to examine this association more thoroughly and for psychotherapeutic treatments to be tailored accordingly [34]. Similar outcome studies with adolescent AN patients are still missing.

Adolescence is a period of rapid change and a crucial period for the development and consolidation of identity. Eating disorders are highly prevalent in adolescence, and younger incidence has increased [35]. Diagnosing comorbid PD is important since patients with PD present with severe consequences in terms of psychosocial functioning and personal suffering [36], and therefore, early intervention is needed. Looking at the scarcity of existing data, the objective of this study was to extend the knowledge about PD in adolescent AN patients. We assessed PD in a sample of adolescent AN patients admitted to inpatient treatment. We hypothesized that we can identify a subgroup of patients who present with PD, and that there is an association between PD and symptom severity.

MethodsParticipants

Recruitment of patients took place at the inpatient treatment facilities of the Department of Child and Adolescent Psychiatry at the University Hospital of Cologne. The criteria for inclusion in this study were as follows: main diagnosis of AN, either typical (ICD-10: F50.0) or atypical (ICD-10: F50.1), female, 14–18 years old and an IQ >80 to ensure sufficient understanding of a PD questionnaire. Adolescents were consecutively admitted during a period from January 2010 to December 2016. Data on AN diagnosis, age, height, and weight including body mass index and body mass index percentile were collected. Patients in this sample were excluded from participation if they had an IQ below 80 which concerned two cases. During the selected period, a total of n = 73 patients were included based on the criteria set out.

InstrumentsAssessment of Axis I Disorders

AN diagnosis was defined by the International Classification of Diseases and Related Health Problems, 10th edition (ICD-10) and assessed in a clinical interview.

Assessment of Severity of Eating Disorder Symptoms

Eating Disorder Inventory 2. The original English version of the Eating Disorder Inventory 2 (EDI-2) [37] is a self-assessment inventory translated into German by G. Rathner [38]. It was standardized based on a sample of healthy adolescents and based on a sample of adolescent patients with eating disorders. The EDI-2 consists of 91 items with a 6-point rating scale. Its long version comprises 11 scales (drive for thinness, bulimia, body dissatisfaction, ineffectiveness, perfectionism, interpersonal distrust, interoceptive awareness, maturity fears, asceticism, impulse regulation, social insecurity); a short version with 64 items consists of the first 8 scales. The scales obtain psychopathological problem areas that are often relevant in patients with eating disorders. The EDI-2 total score measures the severity of eating disorder symptoms. It was examined in different populations and shows good internal consistency with a Cronbach’s α ≥ 0.70 for subscales and between 0.90 and 0.98 for total score and test-retest reliabilities [38] as well as good ability to discriminate between patients with eating disorders and healthy comparative samples [39].

Assessment of PDs

International Personality Disorder Examination. The International Personality Disorder Examination (IPDE) is a structured interview for the diagnosis of PDs according to DSM-IV and ICD-10 [40, 41] and has been validated for use from the age of 15 years onward. During the first part of the interview, the patient’s personal and clinical history is assessed. In the second part, the interview is structured and begins with questioning areas such as work and school and then goes on to address the patient’s sexuality and possible delinquency. The IPDE used in this study consists of 99 semi-structured questions assessing the 11 PDs defined by the DSM-IV, based on dimensional as well as categorical ratings. There are three possible outcomes for PD according to the IPDE: positive, probable, or negative. For adolescents, subthreshold (probable) diagnoses are recommended when using interviews that aim on adult psychopathology, taking into account that PDs are not fully developed at the age of 13–15 years [42].

The IPDE was examined in a worldwide field study. It turned out to be useful in all of the countries in which the study was executed. The inter-rater reliability showed excellent agreement among the examiners; also the test-retest reliability was high.

Procedure

Axis I disorders were assessed using the ICD-10 diagnostic criteria independently by child and adolescent psychiatrists from the department. All included patients were asked to fill out the EDI-2 and were interviewed using the IPDE by the research evaluation team. All interviewers received an internal training program for administering the IPDE and held a university degree in clinical psychology or clinical education, at the same time being in training for licensure as child and adolescent psychotherapist.

Statistical Analysis

Relevant parameters were extracted and inserted under pseudonymized patient IDs into the statistical software “IBM Statistics SPSS, version 27.” Descriptive statistics (percentages, means, standard deviations) relating to clinical variables were performed to describe the sample and to present the prevalence of PDs (including 95% confidence intervals) among adolescents with AN. In addition to descriptive analyses, we used Student t tests (including effect sizes in terms of Cohen’s d) to examine differences in EDI-2 total scores and subscales between individuals with positive and negative IPDE, respectively, and patients with comorbid versus no comorbid PD. For the purpose of the present study, the IPDE outcomes were categorized into positive (representing a definite PD diagnosis) and negative (below cutoff for definite PD diagnosis). The level of significance was set at p < 0.05. According to Cohen [43], effect sizes of d = 0.2, d = 0.5, and d = 0.8 can be interpreted as small, medium, and large. We performed a post hoc power analysis with G*Power to compute the achieved power of the t test to detect differences between patients with and without comorbid PD regarding the EDI-2 total score. Given the obtained sample size of the two groups, the detected effect size of d = 0.92 (see results section) and a significance level of α = 0.05, the two-tailed t test achieved a power of 84%. However, due to the limited sample size which may reduce the power of statistical tests for detecting differences in EDI-2 subscales, the focus is also on the interpretation of effect sizes and not only on the results of statistical tests. The statistical analysis was performed with the statistics program IBM Statistics SPSS Version 27 Apple Macintosh OSX.

ResultsSociodemographic and Clinical Variables

All 73 participants fulfilled the diagnostic criteria of AN (83.6% typical [F50.0] vs. 16.4% atypical [F50.1]). The sample’s sociodemographic and clinical characteristics are shown in Table 1.

Table 1.

Sociodemographic and clinical characteristics

Prevalence and Distribution of PD

Table 2 reports the prevalence of definite PD including percentages and a 95% confidence interval measured by the IPDE. A total of 20 (27.4%) patients reached cutoffs for any PD. The most frequently present PDs in the total sample were, in order of prevalence, the following: 11.0% obsessive-compulsive, 11.0% avoidant, 5.5% borderline, 5.5% unspecified, and 2.7% dependent. When using cutoffs for probable PD, 52 patients (71.2%; 95% CI [60.7; 81.7]) met the criteria for any PD. Distribution of probable PD was as follows: 17.8% obsessive-compulsive; 11.0% avoidant; 4.1% borderline; 2.7% dependent; and schizoid, histrionic, and unspecified each 1.4%.

Table 2.

Prevalence and distribution of definite PD diagnoses assessed by the IPDE

Prevalence of PD and Severity of Eating Disorder Symptoms

EDI-2 long version (11 subscales and total score) was available for 47 individuals and EDI-2 short version (8 subscales) for 53 individuals. Table 3 shows differences in the severity of eating disorder symptoms between patients with and without PD. The T test analyses revealed significant differences between individuals with comorbid PD versus without comorbid PD regarding the severity of the eating disorder symptoms in terms of the subscales “ineffectiveness,” “interpersonal distrust,” “interoceptive awareness,” “asceticism,” “impulse regulation,” “social insecurity” and in the total score of the EDI-2 with patients having comorbid PD showing higher eating disorder severity scores compared with patients having no comorbid PD. Effect sizes for these differences can be regarded as medium to large (0.65 < d < 1.20). Although we observed no statistically significant group differences for the other EDI-2 subscales, effect size was also in the medium range for the “maturity fears” subscale.

Table 3.

Difference in EDI-2 total score and subscales between AN patients with and without PD according to the IPDE

Discussion

The present study investigated the prevalence of PD in a highly selected sample of adolescents suffering from AN who were admitted to inpatient treatment at a child and adolescent psychiatric unit. We studied a total sample of 73 patients with a mean age of 15.7 years. For PD diagnosis, we conducted an IPDE interview, the official instrument of the World Health Organization for diagnosing PD according to ICD-10 and DSM-IV. By investigating this large sample of adolescents, we expand our knowledge of comorbid PD in AN. Previous research on the prevalence of PD in adolescent patients with AN is rare and yielded inconsistent findings.

Müller et al. [44] investigated 32 former adolescent patients 10 years after discharge from inpatient treatment and diagnosed PD in 28%. Another study [31] reported 22% of comorbid PD in a large sample of AN patients, but the mean age of the youngest group was 20 years. A third study examining 57 adolescent patients with AN, restrictive type [45], found a lower proportion of 14% having a comorbid axis II diagnosis. In 57 outpatient AN adolescents, Gaudio and DiCiommo found 22.8% to have at least one PD as a comorbidity [26]. We found only one study using the IPDE for diagnosing PD in adolescents [27]. The researchers included patients with various eating disorders; 32 out of a total sample of 100 had AN, 8 of the restrictive type, and 24 of the binge-purging type. In the AN total group of 32 patients, they found a higher prevalence of PD (28%) than reported in the majority of previously published studies. However, this is in line with the results of the present study. In our sample, almost one third had at least one comorbid PD (27.9%). Using a more broad definition of PD, we found that 71.2% of the patients had either a definite or a probable diagnosis of PD, confirming the upper prevalence level of PD in AN seen in the literature [25, 27].This suggests that in the vast majority of adolescent patients suffering from AN, problems in the development of personality might be underlying to core symptomatology. Since poor treatment outcome was found in some (adult and adolescent) patients with AN and PD [32, 34, 46]; this is an important finding for treatment planning and choosing treatment options in order to make psychotherapy more effective for such patients.

We found, concordant with other studies, mostly cluster C PDs, i.e., OCPD and avoidant PD (APD). These PDs are most commonly associated with AN [26]. Other PDs found in our sample were BPD and PD unspecified. We did not find any cluster A PDs.

A thorough review highlights the wealth of literature describing the “typical” anorectic personality as rigid, perfectionist, and inflexible. Also, anorectic behaviors and thoughts overlap with essential diagnostic features of OCPD, namely, preoccupation with orderliness, perfectionism, and mental and interpersonal control at the expense of flexibility, openness, and efficiency [47]. There is evidence suggesting a shared familial transmission of AN and OCPD [48]. It might be valuable to consider these personality features for modifying treatment options in adolescence.

In line with a current meta-analysis in adults, BPD was the only cluster B-specific PD [49]. Impulsivity, instability, and anger in BPD have a larger association with Bulimia Nervosa compared to AN [50]. In our sample of AN adolescents, 5.5% presented with a BPD, being the third most common PD after OCPD (11%) and APD (11%). This might be a link to the instability of and cross-over between eating disorder diagnoses since the course of AN oftentimes includes the emergence of bulimic symptoms [51-53]. Research on prognostic factors for this cross-over is scarce [54], and since our study has a cross-sectional design, the role comorbid BPD might play is hypothetical.

Even though ASD was not diagnosed in our sample that was collected until the end of 2016, an overlap with autistic spectrum disorder, ASD, BPD, and AN is also an interesting topic for future research [15]. Neuropsychological similarities between AN and ASD have been reported [55-57], including superior attention to detail, difficulties in theory of mind, emotion recognition, and empathy. Difficulties in theory of mind and empathy were also considered when investigating overlaps between ASD and PD. Empathy is a construct composed of two components: one is cognitive, involving the process of understanding another person’s perspective; and one is affective, involving sharing the emotional experiences of others [58]. BPD subjects had higher average levels of affective empathy and lower levels of cognitive empathy compared with ASD [58, 59]. These considerations are relevant for BPD and might play a role for the comorbidity with AN, but research regarding an overlap of ASD and the predominant PDs in our sample, OCPD and APD, is scarce. There is still a diagnostic gray zone, and further research is needed to address these questions.

The EDI-2 total score reflecting AN severity was significantly higher in patients with AN and comorbid PD, confirming our main hypothesis. Significantly, higher severity levels of ineffectiveness, interpersonal distrust, interoceptive awareness, asceticism, impulse regulation, and social insecurity were observed in patients with comorbid PD. Specialized treatment options for adolescents with PD and AN might improve treatment outcome in some patients and should be the focus of future research.

The effect sizes of differences between patients with versus without comorbid PD regarding other symptoms assessed in the EDI-2, namely, drive for thinness, bulimia, body dissatisfaction, perfectionism, and maturity fears, were of small to medium, though these effects did not reach statistical significance. One could speculate that these latter symptoms are key symptoms in AN and are more often addressed in cognitive psychotherapeutic approaches, whereas the areas significantly associated with PD (ineffectiveness, interpersonal distrust, interoceptive awareness, asceticism, impulse regulation, and social insecurity) tend to be less addressed or at a later stage of therapy. However, these symptoms could have a major impact on the maintenance of the illness and might need specialized treatment.

Strengths and Limitations

The strength of this study is the rather large sample size compared to other studies. The sample is homogenous as we included AN patients that were admitted to inpatient treatment. Still, the sample size, particularly of the group with comorbid PD, was low. Thus, the statistical tests were not sufficiently powered to detect differences between patients with versus without comorbid PD regarding specific subdomains of eating disorder pathology. However, the present study has clearly pointed out that the effect sizes of these differences are profound and, thus, should gain more attention in future research. A further strength is the diagnostic procedure, as we did not use self-rating instruments for diagnosing PD. The IPDE is a standardized interview for diagnosing PD in adults and adolescents. Our results might have implications on treatment planning and outcome but are limited due to the cross-sectional design; these are hypotheses that need further research.

Conclusion and Implications

In a sample of AN adolescents, almost one third presented with comorbid PD. These patients had significantly higher EDI-2 total scores, reflecting greater symptom severity. This might indicate that patients with comorbid PD and AN are more difficult to treat and are less responsive to regular eating disorder treatment. The outcome has potential implications for treatment planning as an intensive form of psychotherapy for PD in adolescents exists: transference-focused psychotherapy for adolescents (TFP-A) [60-65], dialectic behavioral psychotherapy for adolescents (DBT-A) [66], and mentalization-based psychotherapy for adolescents (MBT-A) [67]. These forms of psychotherapy should be considered as a treatment option for adolescent patients suffering from severe AN and comorbid PD in order to improve treatment outcome. It might be interesting for future research to follow patients with AN and comorbid PD to further investigate whether patients with PD have poorer prognosis and which PD-specific treatment is effective for long-term outcome.

Acknowledgments

The authors would like to thank MA Cecily Jahn for proofreading the manuscript.

Statement of Ethics

This study received a positive vote from the Ethics Committee of the University of Cologne (reference number: 10-104). After a complete explanation of the study, written informed consent was obtained from all eligible adolescents and their parents or legal guardians.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

The Marga und Walter Boll-Stiftung financed part of the study. The funder had no role in study design, collection, analysis, and interpretation of data, or writing of the report.

Author Contributions

Conception and design of the study: Maya Krischer and Gerd Lehmkuhl; acquisition and analysis of data: Clarissa Laczkovics, Klara Czernin, Jessica Carlitscheck, Michael Zeiler, Pauline Schlund, and Heidrun Wunram; drafting of the manuscript or tables: Clarissa Laczkovics, Klara Czernin, and Maya Krischer. All the authors have reviewed and approved the final manuscript.

Data Availability Statement

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.

This article is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC). Usage and distribution for commercial purposes requires written permission. Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug. Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.