The authors have declared that no competing interests exist.
Conceived and designed the experiments: HNS JE. Analyzed the data: HNS JE. Contributed reagents/materials/analysis tools: RWG. Wrote the paper: HNS JE RWG.
The high co-occurrence between borderline personality disorder and affective disorders has led many to believe that borderline personality disorder should be considered as part of an affective spectrum. The aim of the present study was to examine whether the prevalence of affective disorders are higher for patients with borderline personality disorder than for patients with other personality disorders.
In a national cross-sectional study of patients receiving mental health treatment in Norway (N = 36 773), we determined whether psychiatric outpatients with borderline personality disorder (N = 1 043) had a higher prevalence of affective disorder in general, and whether they had an increased prevalence of depression, bipolar disorder or dysthymia specifically. They were compared to patients with paranoid, schizoid, dissocial, histrionic, obsessive-compulsive, avoidant, dependent, or unspecified personality disorder, as well as an aggregated group of patients with personality disorders other than the borderline type (N = 2 636). Odds ratios were computed for the borderline personality disorder group comparing it to the mixed sample of other personality disorders. Diagnostic assessments were conducted in routine clinical practice.
More subjects with borderline personality disorder suffered from unipolar than bipolar disorders. Nevertheless, borderline personality disorder had a lower rate of depression and dysthymia than several other personality disorder groups, whereas the rate of bipolar disorder tended to be higher. Odds ratios showed 34% lower risk for unipolar depression, 70% lower risk for dysthymia and 66% higher risk for bipolar disorder in patients with borderline personality disorder compared to the aggregated group of other personality disorders.
The results suggest that borderline personality disorder has a stronger association with affective disorders in the bipolar spectrum than disorders in the unipolar spectrum. This association may reflect an etiological relationship or diagnostic overlapping criteria.
Traditionally, borderline personality disorder (BPD) has been considered as a threshold psychotic disorder
There is a huge variation in prevalence figures between studies of BPD among patients with major depression. Research findings vary between 7% and 63%
McGlashan et al.
Even though there seems to be a high comorbidity between affective disorders and BPD, it is no current consensus as to how this should be understood. Some researchers suggest that BPD should be viewed as an affective spectrum disorder, and are consequently reluctant to classify BPD as a personality disorder
How we understand the interaction between BPD and affective disorders has implications for how affective symptoms among patients with BPD are contextualized as either state or trait dependent, biologically or situationally determined. This may in turn influence the treatment offered to patients with BPD symptoms. On this basis, comparing the prevalence of affective disorders in BPD with other personality disorders may contribute further to our understanding of the association between affective spectrum disorders and BPD. To our knowledge, only a few studies have examined this issue. Zanarini et al.
Further studies comparing affective disorders in BPD and specific other personality disorders are needed, because comparisons between BPD and a mixed selection of other personality disorders may be misleading. Clinical samples with different personality disorders may represent an average with high or low prevalence of affective disorders. It is also essential to differentiate affective disorders into unipolar depression, bipolar affective disorder and dysthymia, because factors that distinguish the different personality disorders from one another (e.g. biological, psychosocial and temperamental) may interact with the various affective disorders in different ways. It is possible that persons with BPD are vulnerable to unipolar depression. An increased number of stressors related to impaired psychosocial functioning may, on the one hand, increase the risk of depression. This may, however, also be relevant for other personality disorders. On the other hand, there may be a more direct relationship between bipolar disorder and BPD than between bipolar disorder and other personality disorders. Thus, the varying results from studies exploring the relationship between affective disorders and BPD may be due to differing mechanisms mediating the different affective disorders.
In order to provide a better understanding of the relationship between BPD and affective disorders, we wanted to compare the occurrence of affective disorders among patients with BPD and other types of personality disorders. The present study examines whether the prevalence of affective disorders in general, or unipolar depression, bipolar disorder, and dysthymia separately, are higher in persons with borderline personality disorder than in persons with other personality disorders.
Data was obtained from a cross-sectional study commissioned by the Norwegian Directorate of Health as part of the National Patient Registration in Norway in 2008. To ensure an accurate picture of the patient population, consent was not obtained during the registration process. No information that could identify the participants was recorded, therefor it has not been possible to obtain consent for the present study either. The study was approved by the Norwegian Regional Ethics Committee and the Norwegian Data Inspectorate.
Data was obtained from a cross-sectional study carried out by SINTEF Technology and Society, and commissioned by the Norwegian Directorate of Health as a part of the National Patient Registration in 2008. The data was collected from all Norwegian Community Mental Health Centres and all private practitioners (psychiatrists and clinical psychologists) with a licensed work agreement with the Norwegian Health Trusts. Demographic, clinical and treatment data were registered for all unique patients over age 18 receiving outpatient treatment during a 2-week period in 2008 (n = 36 773)
The therapists assessed demographic and clinical information for each patient, and included one primary and one secondary diagnosis listed in the 10th revision of the International Classification of Diseases (ICD-10)
All patients with personality disorder diagnosis (n = 3 752) were identified in the main study data set. In order to allow comparisons of affective and personality disorders, subjects with more than one personality disorder (n = 73) were excluded. Thus, the sample consisted of 3 679 outpatients with one registered personality disorder, wherein 70% had a personality disorder as their primary diagnosis. A total of 946 subjects had an unspecified personality disorder. A non BPD personality disorder group (n = 2 636) was computed in order to perform comparisons with other studies dichotomizing their samples into BPD versus other personality disorders
The following affective disorders were classified:
Diagnosis in the schizophrenia spectrum, schizotypal and delusional disorders were categorized as psychotic disorders (F20– F29 in ICD-10). All disorders due to psychoactive substance use were categorized as substance use disorders (F10– F19 in ICD-10).
Student t-tests and Chi-Square tests were performed to determine whether there were differences between the BPD group and each of the personality disorder groups on demographic, clinical and various affective disorder variables. Because BPD is compared to nine different personality disorders (multiple comparisons), we adopted a.0025 alpha level. Odds ratios were computed for each affective disorder comparing the BPD group to the non-BPD personality disorder group. The confidence interval was set to 95%.
Demographic and clinical characteristics of the different personality disorders are presented in
Personality disorder study groups | |||||||||||
Borderline | Paranoid | Schizoid | Dissocial | Histrionic | OCPD |
Avoidant | Dependent | Unspecified | Non-BPD PD |
||
(n = 1043) | (n = 194) | (n = 106) | (n = 64) | (n = 45) | (n = 131) | (n = 918) | (n = 232) | (n = 946) | (n = 2636) | ||
Gender,female (%) | 86 | 55 |
51 |
19 |
80 | 51 |
62 |
79 | 62 |
61 |
|
Age in years(Mean ± SD) | 35.5±10.8 | 41.0 |
38.5±12.3 | 35.5±10.4 | 47.2 |
43.0 |
40.3 |
42.0 |
40.0 |
40.4 |
|
Living alone/notliving alone |
70/30 | 70/30 | 71/29 | 71/29 | 70/30 | 47/53 |
59/41 |
57/43 |
62/38 |
61/39 |
|
Education |
2.8±0.8 | 2.9±0.9 | 3.1 |
2.3 |
3.3 |
3.5 |
3.0 |
3.2 |
3.2 |
3.1 |
|
Treatment length(mo)(Mean ± SD) | 25±34 | 27±33 | 35±43 | 18±24 | 32±30 | 36±47 | 28±34 | 36 |
28±36 | 29 |
|
Substance usedisorder (%) | 5.8 | 5.2 | 1.9 | 28.1 |
4.4 | .8 | 2.0 |
2.6 | 3.1 | 3.3 |
|
Psychotic disorder (%) | 1.8 | 2.1 | .9 | 7.8 |
4.4 | .0 | .4 | .0 | 1.9 | 1.3 |
Obsessive-compulsive personality disorder.
Non-BPD personality disorder.
Living alone = not married, separated/divorced, widowed/widower; Not living alone = married, cohabitant.
Education was rated on a five point scale: 1 = incomplete primary school, 2 = primary school, 3 = high school, 4 = university BA, 5 = university MA.
Differ significantly from BPD group (p<.0025).
The rates of affective disorders among patients with BPD and other personality disorders are shown in
Borderline | Paranoid | Schizoid | Dissocial | Histrionic | OCPD |
Avoidant | Dependent | Unspecified | Non-BPD PD |
||
(n = 1043) | (n = 194) | (n = 106) | (n = 64) | (n = 45) | (n = 131) | (n = 918) | (n = 232) | (n = 946) | (n = 2636) | ||
Affectivedisorder | % | % | % | % | % | % | % | % | % | % | OR |
Affectivedisorder |
22.0 | 26.8 | 32.1 | 12.5 | 22.2 | 32.8 | 32.4 |
30.6 | 27.6 | 29.4 |
.75 |
Depression | 14.9 | 22.2 | 22.6 | 6.2 | 17.8 | 21.4 | 24.9 |
26.7 |
20.8 |
22.6 |
.66 |
Bipolardisorder | 5.7 | 2.1 | 2.8 | 6.2 | 2.2 | 6.1 | 3.1 | 2.2 | 3.9 | 3.4 |
1.66 |
Dystymicdisorder | .9 | 2.6 | 6.6 |
.0 | .0 | 3.1 | 4.2 |
.9 | 2.3 | 3.0 |
.29 |
Obsessive-compulsive personality disorder.
Non-BPD personality disorder.
The percentages in the variable
Differ significantly from BPD group (p<.0025).
Odds ratio for the affective illnesses in the BPD group compared to non-BPD PD.
Odds ratios showed that patients with BPD had a 34% lower risk (OR = 0.66) for depression compared to the total sample of patients with non-BPD personality disorders. The risk for dysthymia among BPD subjects was one in three compared to the other personality disorders (OR = 0.29).With regard to the probability ratio for bipolar disorder, the direction of results was reversed, i.e. the BPD group had a 66% higher risk (OR = 1.66) for having a bipolar disorder than the total sample of non-BPD personality disorders.
The frequency of affective disorders was generally lower for BPD than in the aggregate sample of other personality disorders. This is in contrast to the findings of Zanarini et al.
The prevalence of affective disorders in the BPD group in our study was higher than among persons in the general population
In the total sample (n = 3 679), depression was five times more frequent than bipolar disorder, and almost nine times more frequent than dysthymia. This indicates that the affective disorders category primarily measures the incidence rate of depression. Variations in the incidence rates for specific affective disorders are therefore difficult to identify and the results may be somewhat misleading. This should be considered in future studies.
There was a lower rate of depression in the BPD group than among those with avoidant, dependent, unspecified personality disorders as well as among the total group of non-BPD personality disorders. This is in contrast to earlier studies showing no differences in the rate of comorbidity of depression between persons with BPD and avoidant personality disorder
The comparisons between BPD and paranoid, schizoid, dissocial, histrionic and obsessive-compulsive personality disorders yielded no significant differences in the rate of depression. These results verify earlier findings regarding the obsessive-compulsive
Compared to each personality disorder separately, the BPD group did not differ significantly in prevalence of bipolar disorder. BPD had a numerically higher rate, apart from the dissocial and obsessive-compulsive types. However, BPD was significantly more likely to have bipolar disorder compared to the aggregate sample of non-BPD personality disorders. The results indicate a tendency for BPD to be associated with bipolar disorder more often than other personality disorders.
BPD is a personality disorder in which the key characteristics are unpredictable and rapidly shifting mood, impulsivity and lack of self control
Several researchers are open to classify BPD as a mood disorder rather than a personality disorder
Although other research suggests that the relationship between BPD and bipolar disorder has a stronger biological component than for BPD and unipolar depression, it does not necessarily follow that BPD is a bipolar spectrum disorder. There may be several reasons for this. First, only 5.7% of the BPD patients in this study had bipolar disorder. Second, neurobiological studies propose that affective instability in BPD is characterized by reactivity to psychosocial cues, while affective instability in bipolar disorder primarily is internally driven
BPD is traditionally considered as chronic and a hard-to–treat disorder. Recent research has revealed that a significant number of patients with BPD respond to treatment
We did not find a higher rate of dysthymia among patients with BPD compared to the other personality disorder groups. This supports the findings of those few studies that earlier have examined this question
The frequency of coexisting substance use disorders among patients with BPD is lower than what has been found in other studies
In our sample of subjects with non-BPD personality disorders, as many as 80% had either an avoidant, dependent or unspecified personality disorder. The varied selection therefore primarily became a measure of the incidence of affective disorders among these three personality disorders. Certain types of personality disorders are quite rare. This is a common finding in both clinical
First, diagnostic assessment is not consistently based on systematic use of structural assessment methods. Since axis II-disorders are found to be under-diagnosed by routine clinical assessment
Second, we do not know how clinicians have assessed affective symptoms in various personality groups. Affective disorders may have been diagnosed to a lesser extent in BPD patients than the ones suffering from other personality disorders because BPD often presents with atypical affective symptoms
Third, whether other axis I or axis II disorders are acting as confounding variables has not been possible to determine, since a maximum of two diagnoses were recorded per patient.
At last, the study is based on the Norwegian population of adult patients receiving outpatient psychiatric care. The results may not apply to patient populations outside of Norway, to more seriously ill inpatients or persons with personality disorders who are not in a treatment setting.
This study found that subjects with BPD had a lower rate of affective disorders than an aggregated sample of subjects with other personality disorders. However, the rate varies somewhat among the groups of specific personality disorders. The findings regarding specific types of affective disorders revealed a disassociation, in that the rates of unipolar depression and dysthymia were lower among subjects with BPD than among subjects with several other personality disorders, whereas the rate of bipolar disorder tended to be higher. This suggests a stronger association between BPD and affective disorders in the bipolar spectrum than those in the unipolar spectrum. This association may reflect an etiological relationship or diagnostic overlapping criteria, or both.
To the authors’ knowledge this is the first study that compares the occurrence of different types of affective disorders in BPD to the entire spectrum of each individual personality disorder simultaneously. Further studies using structured diagnostic instruments are needed to confirm the results of this study. There is an overall need for studies addressing the relationships between the phenomenological, biological, pathogenetic and treatment-related aspects of BPD and other specific personality disorders, as well as patients with affective disorders without personality disorders. Such studies would help us identify which aspects of the interaction between BPD and affective disorders can be described as a general consequence of having a personality disorder, and which aspects of the interaction between BPD and affective disorders are type specific. This is clinically relevant since it may give mental health professionals the opportunity to offer a more precise and evidence based treatment to patients with BPD and coexisting affective disorders.
We thank SINTEF Technology and Society who collected the data used in this study.