The relationship between borderline personality disorder and bipolar disorder has been a controversial topic ever since borderline personality disorder was incorporated into DSM in 1980. Borderline personality disorder was widely felt to be etiologically heterogeneous. Still, borderline patients were often divided into those whose disorder stemmed supposedly from biological factors and others whose condition was viewed as “purely” environmental.
Currently, many investigators have begun to adopt a more balanced view. Patients in whom borderline personality disorder and bipolar disorder co-occur are more apt to be understood as representing an interaction of biological and environmental forces. Sweeping statements like borderline personality disorder is a variant of affective disorder (especially of bipolar II type) are less in vogue. The same is true of the contrary assertion: that borderline personality disorder is not a genuine phenotype of bipolar disorder. As a former president famously said, it depends a lot on what “is” is: the moment one says borderline personality disorder “is” (or “is not”) a variant of bipolar disorder, one has launched an all-or-none argument. Using “is” in this way has become increasingly hazardous because with contemporary research, there is neither compelling evidence that borderline personality disorder and bipolar disorder are so indissolubly linked as to justify the “is,” nor that the two are “always” etiologically unrelated.
The strength of the article by Gunderson and his colleagues, published in this issue, rests on its more balanced position on the controversy. Theirs is a collaborative study derived from four centers in the Northeast (1), in which 629 patients were evaluated: 196 with borderline personality disorder and 433 with other personality disorders (schizotypal, avoidant, or obsessive-compulsive). Each group was subdivided further into those with a lifetime comorbid bipolar I or bipolar II disorder diagnosis versus those without such comorbidity: bipolarity was an accompaniment in 38 of the borderline personality disorder patients (bipolar I disorder: N=23; bipolar II disorder: N=15) and in 34 of the other personality disorders patients. The study embodied a 4-year prospective design with assessments at yearly intervals. The aims of the study were to determine what relationship may exist between borderline personality disorder and bipolar disorder, using as measures the rates of their co-occurrence, and also the effects of co-occurrence on the longitudinal course. The focus was on whether the presence of either disorder heightened the risk for new onsets of the other during the follow-up period. Guidelines were established for identifying a new onset of bipolar I disorder (1 week or more of mania or hypomania in a patient with no prior bipolar episodes) or of bipolar II disorder (1 week or more of hypomania and 2 weeks or more of depression in a patient with no prior hypomanic episodes).
Analysis of their data revealed that the combined bipolar disorder diagnoses were significantly more common in the borderline personality disorder group than in the other personality disorder group: 19.4% versus 7.9%. As to the effects on the course of borderline personality disorder, co-occurrence of bipolar disorder at the outset did not exert a negative impact: two-thirds of all the borderline personality disorder patients showed remission at 4 years whether or not they had a bipolar condition at the beginning.
If one looked at the total new cases of bipolar disorder emerging during the follow-up period, this phenomenon occurred more often, albeit to a modest degree, in the borderline personality disorder patients (13 of 158) than in those with a different personality disorder (12 of 399).
The reverse situation was also measured: borderline personality disorder occurring subsequently in bipolar disorder patients who at the outset had a personality disorder other than borderline personality disorder. Here the study showed only a somewhat higher tendency to manifest borderline personality disorder among patients with the combination “other personality disorder and bipolar disorder” than among “other personality disorders” patients who had not been diagnosed originally with bipolar disorder.
The study concluded that linkage between borderline personality disorder and bipolar disorder was discernible, but that this linkage was one of “modest association.” An interrelationship was more noticeable; furthermore, if one began with a group of borderline personality disorder patients, a proportion of them was more likely to develop a bipolar disorder eventually than was a group of bipolar patients likely to develop the attributes of borderline personality disorder. The authors cautioned that even the new bipolar onsets in the borderline personality disorder group did not represent an evolution from borderline psychopathology, given that the bipolar conditions most often appeared as sequelae of stressful neurobiological or life changes.
Some of the confusion in the literature stemmed from a tendency to conflate purely depressive and bipolar disorders under the heading of affective. The relationship between depression and borderline personality disorder, when looked at more closely, turned out to be “surprisingly weak and nonspecific”(2). When affective disorders occurring in borderline patients are divided according to categories, bipolar disorders are usually much less common than depressive disorders, yet significantly more common than are found in other personality disorders (3). Further large-scale studies with careful family history and long-term follow-up may help determine whether and to what extent genetic factors underlie the association between borderline personality and bipolar (as opposed to purely depressive) disorder.
From the standpoint of treatment, the authors argue against the growing tendency to diagnose patients who meet borderline personality disorder criteria as “really” examples of bipolar (usually, bipolar II) disorder, often omitting the personality disorder altogether. This tendency has, the authors mentioned, the decided disadvantages of relying excessively—at times, exclusively—on medication, while simultaneously neglecting the psychosocial interventions so important in the overall treatment of borderline personality disorder. It has not escaped the attention of the authors that, as a factor contributing to this problem, reimbursement for services is generally adequate for axis I conditions but meager when directed at “just” a personality disorder, such as borderline personality disorder.
In summary the Collaborative Study goes a fair way toward resolving the “is” dilemma: we should be equally reluctant to assert that borderline personality disorder is (as suggested by some [4–7]) or that it is not (as suggested by others [8, 9]) simply a variant of bipolar II disorder. Rather, there exists a subgroup within the borderline personality disorder domain where risk genes for bipolar illness may lead to a joint presentation of both illnesses (10). Hypomania (in the case of co-occurring borderline personality disorder and bipolar II disorder) may precede the emergence of borderline personality disorder in some patients, may surface a number of years later than when borderline personality disorder is first recognized, and in still others may become manifest at about the same time one is first diagnosed with borderline personality disorder. In valid cases of co-occurrence, both aspects of the condition may indeed arise from a common etiology and may respond well to a combined regimen of mood stabilizers and appropriate psychosocial interventions.