Objectives
Clinical studies suggest a high co-morbidity rate of borderline personality disorder (BPD) with bipolar disorder (BD). This study examines the prevalence and correlates of BPD in BD (I and II) in a longitudinal population-based survey.
Methods
Data came from waves 1 and 2 (wave 2: N = 34,653, 70.2% cumulative response rate; age ≥20 years) of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). Lay interviewers conducted in person interviews using the Alcohol Use Disorders and Associated Disabilities Interview (AUDADIS-IV), a reliable diagnostic tool of psychiatric disorders based on DSM-IV criteria. Subjects with BD I (n = 812), BD I/BPD (n = 360), BD II (n = 327) and BD II/BPD (n = 101) were examined in terms of sociodemographics, mood, anxiety, substance use and personality disorder co-morbidities and history of childhood traumatic experiences.
Results
Lifetime prevalence of BPD was 29.0% in BD I and 24.0% in BD II. Significant differences were observed between co-morbid BD I/II and BPD versus BD I/II without BPD in terms of number of depressive episodes and age of onset, co-morbidity, and childhood trauma. BPD was strongly and positively associated with incident BD I (AOR = 16.9; 95% CI: 13.88–20.55) and BD II (AOR = 9.5; 95% CI: 6.44–13.97).
Conclusions
BD with BPD has a more severe presentation of illness than BD alone. The results suggest that BPD is highly predictive of a future diagnosis of BD. Childhood traumatic experiences may have a role in understanding this relationship.
Bipolar disorder (BD) has frequently been linked in clinical literature with personality disorders [1]. High rates of borderline personality disorder (BPD) in patients with BD have been reported in clinical studies and interestingly, this co-occurrence seems to be bidirectional in nature, regardless of the index population studied (BPD with BD or vice versa) [2] and [3].
The underlying nature of the relationship between BD and BPD remains unclear and continues to foster debate. While BPD can frequently co-occur with other disorders such as major depressive disorder, the association between BPD and BD seems to be particularly robust [4]. High heritability rates have been reported for personality traits [5], [6] and [7] and it is possible that certain personality features influence or indicate vulnerability toward the development of a mood disorder such as BD [8], [9], [10], [11] and [12]. It has been suggested that the development of personality structure may be affected by the early onset of a mood disturbance [13], [14], [15], [16], [17] and [18] and that environmental determinants, such as early childhood adversity, may have implications in the development and clinical course of psychiatric disorders such as BD and BPD [16], [19], [20] and [21]. Childhood adversity is understood to have a broad and likely non-specific impact on vulnerability to adult psychiatric disorders, but the importance of traumatic experience is particularly prominent in studies of BPD [22], [23], [24], [25] and [26].
Phenomenological overlap between BPD and BD, especially in terms of affective instability and impulsivity, has long been recognized, suggesting that this co-morbidity could represent an artifact of the definition [2], [27], [28], [29], [30] and [31]. Diagnostic challenges can emerge, particularly in the context of BD II and BPD [32]. Specifically, some authors have proposed that BPD may be a misdiagnosis in patients with bipolar spectrum conditions [33], [34] and [35], while others argue that BPD would be better conceptualized as an atypical variant of a mood disorder [36], [37], [38], [39], [40] and [41]. Contemporary diagnostic manuals including DSM-5 describe these disorders as distinct entities in separate sections of the manual, with clear differences in terms of prevalence, outcomes and course of illness [2], [42], [43] and [44].
While the existing literature regarding the course of illness of co-occurring BD and BPD is limited, clinical literature highlights the debilitating nature of this co-morbidity, which may be characterized by an earlier onset, more numerous and severe mood episodes, worse inter-episode functioning, reduced treatment adherence and worse outcomes with medication treatment [45], [46], [47], [48], [49] and [50]. The present study examines this co-morbid condition in regard to its development, presentation and scope by exploring the co-morbidity of BD (type I and II) and BPD using a large, comprehensive, and longitudinal nationally representative sample.
The study addressed four main aims: (1) to determine the prevalence of BPD in BD, (2) to examine the correlates and impact of BPD on the presentation of BD illness, (3) to compare the frequency of childhood traumatic events in BD with and without BPD and (4) to evaluate BPD versus other personality disorders as a predictor of a future diagnosis of BD. We hypothesized that BPD would be highly prevalent among subjects with BD, particularly BD type II. We expected BPD would have pervasive negative effects on the presentation of BD and that subjects with co-morbid BD and BPD would report a higher frequency of childhood adversities than both BD without BPD and non-BD. Finally, we predicted that effect sizes would be the largest for BPD predicting new-onset BD compared to all other personality disorders in longitudinal analyses.
BPD is a common co-morbidity in both BD type I and BD type II. Childhood traumatic experiences are potentially implicated in the development of this co-morbid condition. A more severe course of illness can be anticipated in patients with the co-morbid condition compared to BD alone. Patients with BPD appear to be at high risk of developing BD during longitudinal follow up, and an inclusive approach toward screening for BD in individuals with BPD, particularly if there is a history of childhood adversity, is necessary. Studies to date have not yet delineated the most appropriate strategies for treatment of this complex patient population. Future studies of treatment strategies are clearly indicated.
