Borderline Personality

Abstract

BORDERLINE PERSONALITY DISORDER (BPD) IS A COMPLEX AXIS II Phenomenon that is typically described in a psychological or psychiatric context. In this article, we translate the various aspects of BPD to the primary care setting. Previous work in this area has explored specific relationships between BPD and individual medical disorders or between BPD and general somatic symptoms, but the synthesis of these findings and their augmentation with cogent psychological theory is new to the field. Specifically, we highlight the prevalence rate of BPD in the primary care setting, the effects on healthcare utilization, the themes of somatic preoccupation and somatization disorder, several medical syndromes that illustrate the dynamics of the disorder in the medical setting, and the relationship of BPD to disability. We believe that the BPD concept needs to extend beyond its traditional psychological/psychiatric borders to include the subset of BPD patients with somatic symptoms who are seen in primary care settings.

Borderline personality disorder (BPD) is an Axis II disturbance characterized by a transiently intact social façade and chronic impulsivity that manifests as self-regulation difficulties and self-destructive behavior. In the psychiatric setting, patients with BPD typically present for the treatment of self-regulation disturbances (e.g., eating disorders, substance abuse, mood lability, rage reactions), self-harm behavior (e.g., cutting or burning oneself), and/or relationship difficulties (e.g., perceived or impending abandonment). These same presentations may occur in the primary care setting as well, but often the focus of treatment is the consequence of behavior (e.g., hypokalemia in bulimia nervosa, unconsciousness due to an overdose). However, there may be other manifestations of BPD that are specific to the medical, rather than psychiatric, setting. These syndromes are the focus of this article.

Relevance of BPD Recognition in the Primary Care Setting

The recognition of BPD in the primary care setting is of particular importance for psychiatrists who provide consultation in these settings. These patients are often described as difficult rather than identified as borderline because the underlying personality pathology may be obscured by complex somatic presentations. The diagnosis of BPD is essential in consolidating and undertaking a management strategy in the primary care setting. Such strategies may include, for example, establishing clear and documented communication with the patient and firm limitations on types and quantities of medicines.

Prevalence of BPD in the Primary Care Setting

Using structured clinical interviews, Gross and colleagues¹ examined the prevalence of BPD in an urban primary care practice. Among 218 patients, 6.4 percent met the criteria for BPD. As expected, these patients evidenced high levels of psychiatric comorbidity, including mood disorders (36%), anxiety disorders (57%), and bipolar disorder (20%). Not surprisingly, approximately 20 percent reported suicidal ideation. Nearly half of these troubled patients were not recognized by their family physicians as having chronic mental health problems. Whether this data on prevalence rate translates to other primary care settings is unknown.

The Effect of BPD on Healthcare Utilization

BPD in the clinical setting appears to have an impact in the management of service delivery. For example, in mental health settings, BPD is a costly psychiatric disorder. Research indicates that those with BPD frequently utilize psychiatric emergency services² and various inpatient and outpatient mental health services³ and have relatively higher readmission rates for psychiatric hospitalization.⁴ In our own study in a psychotherapy outpatient clinic⁵ comparing BPD patients with non-BPD patients, those with BPD were prescribed significantly more psychotropic medications and attended more psychotherapy sessions, resulting in higher mental healthcare utilization.

These same patterns of high healthcare utilization also emerge in primary care settings. Using different patient samples and study variables, we have consistently found that, compared with non-BPD patients, those with BPD demonstrate higher utilization rates of primary care resources (e.g., greater number of office visits and prescriptions,^6,7 more contacts with the facility including telephone calls,⁷ and more frequent specialist referrals⁸). So, it appears that in both mental health and primary care settings, the diagnosis of BPD is consistently related to higher levels of healthcare utilization.

BPD and Somatic Preoccupation

One characteristic that may explain the observation of higher healthcare utilization in primary care settings may be the tendency for some BPD patients to cultivate somatic symptoms or preoccupation (e.g., multiple, diffuse, difficult-to-diagnose physical symptoms without significant verification through physical or laboratory examination). In this regard, a number of authors have referred to a relationship between BPD and somatic symptoms. In a group psychotherapy setting, Schreter⁹ observed a relationship between chronic somatic symptoms and borderline characteristics. Giovacchini^10,11 described a group of BPD patients with a psychosomatic focus and accompanying somatic symptoms. Bernstein¹² indicated that somatic pathology may actually mask an underlying BPD. Hull and colleagues¹³ described a borderline patient whose behavior of acting out was seemingly synchronized with exacerbations of physical illness. Finally, Janssen¹⁴ described two cases of BPD in which both patients presented with somatic problems.

The available research appears to support these clinical experiences and impressions. Using the Minnesota Multiphasic Personality Inventory, Lloyd and colleagues¹⁵ examined the response patterns of 27 outpatients with BPD. They determined that BPD patients were particularly prone to reporting somatic complaints. In examining patients with dissociative identity disorder, a probable variant of BPD, Ross and colleagues¹⁶ found an average of 15.2 reported somatic symptoms per participant. As expected, these participants also met a high number of BPD criteria (5.2 criteria, on average). Using the Diagnostic Interview Schedule, Prasad and colleagues¹⁷ found, among 21 patients with BPD, a small subsample with somatization disorder. Hudziak and colleagues¹⁸ examined BPD patients with and without major depression and found higher rates of somatization among the depressed subsample. The prevalence rate of somatization disorder in the entire sample of 75 BPD patients was 36 percent.

In our studies of patients in an outpatient internal medicine clinic, we also found relationships between BPD and somatic preoccupation. Specifically, in one study of 120 participants,¹⁹ we found that BPD demonstrated high correlations with somatic preoccupation. In another study,²⁰ we examined patients' perceptions of the health status of both parents. Using path analysis (a statistical approach that enables one to examine the probability of sequential events), we found that while the perceived health status of either parent did not demonstrate a direct effect on somatic preoccupation, the perception of a mother being in poor health demonstrated an indirect effect on somatic preoccupation via BPD.

Because of the accumulating clinical and empirical data, and our impressions from clinical work in primary care settings, we suggest that BPD may, in some individuals, predominantly manifest as somatic preoccupation and/or somatization disorder. If so, it would seem logical that this subset of BPD patients would focus their healthcare needs in medical rather than in psychiatric settings. Thus, as a group, they might be relatively unfamiliar to mental health professionals with little primary care exposure.

Given the existence of a subset of BPD patients with predominantly somatic symptoms, do they evidence the same general behavioral and clinical patterns as those BPD patients encountered in psychiatric settings? We believe so. We have previously described a number of interesting parallels among BPD patients in psychiatric and medical settings²¹ including high utilization of services, intense relationships with treating clinicians, boundary issues with staff, multiple diagnoses (psychiatric and/or medical), voluminous records, complex histories, infrequent resolution of symptoms, and multiple drug allergies.²²

Again, we wish to emphasize the importance of viewing somatic BPD patients as a relatively distinct subset of all BPD patients, although there is likely to be a continuum of somatic preoccupation among BPD individuals, in general. In support of this subset concept, Trappler and Backfield²³ indicate that this somatic subgroup may actually be a diagnostic subtype of BPD. As an example, they describe the guilt-inducing nature of somatic preoccupation among physically frail, older individuals. In this scenario, illness behavior functions to secure interpersonal contact and care through guilt. This subtype concept is hardly new. Back in 1985, Akiskal and colleagues²⁴ identified a BPD subgroup characterized by somatization, an observation that was based upon their research findings.²⁵ Surprisingly, in the Diagnostic and Statistical Manual of Mental Disorders,²⁶ there is not a single BPD diagnostic criterion related to multiple somatic symptoms, somatic preoccupation, or somatization disorder.

What variables might mediate a relationship between BPD and somatic preoccupation? We believe that early developmental trauma is one significant variable. In our own study,¹⁹ we found that among internal medicine outpatients, the summation of various types of childhood trauma (e.g., physical, sexual, emotional abuse; witnessing of violence; physical neglect) was associated with somatic preoccupation in adulthood. Dickinson and colleagues²⁷ encountered the same findings among female primary care patients exposed to childhood sexual abuse, as did Bendixon and colleagues² in their examination of a large sample of both male and female undergraduate students. According to several investigators, both childhood physical and sexual abuse may heighten the risk of somatic preoccupation in adulthood.^29,30 Likewise, earlier age of onset³¹ and duration of abuse³² may be contributory factors. Interestingly, Hexel and Sonneck³³ describe differences in somatic presentation as a function of the type of abuse experienced (e.g., sexual abuse: anal, urinary, or oral symptoms; physical abuse: gastric symptoms).

Using a complex schematic, Meares and colleagues³⁴ explain the connections between trauma and somatic preoccupation. Thakkar³⁵ simplifies this relationship by suggesting that traumatic life events in childhood potentially result in a weakening of the immune system, which causes physical symptoms and poor health in adulthood. To augment these theories, we have previously broached our Bifurcation Hypothesis.²¹ In the Bifurcation Hypothesis, we theorize that following the consolidation of BPD in childhood, two general BPD subpopulations arise—one characterized by psychiatric symptoms and the other by somatic symptoms. As a result, these two subsets of patients seek treatment in their respective settings. We currently are embarking on a study to explore several early developmental variables that may account for this divergence. We suspect that illness focus and behaviors in the early family environment may partially explain the divergence of these psychiatric and somatic BPD subgroups.

It is important to note that not all studies have found correlations between somatic symptoms and BPD patients. In this regard, Rogers and colleagues³⁶ found no association between BPD and somatic complaints among 50 psychiatric inpatients. Among 29 BPD psychiatric inpatients, Snyder and Pitts³⁷ found no elevations in the Hypochondriasis subscale of the Minnesota Multiphasic Personality Inventory. However, we point out that these studies were undertaken among psychiatric inpatients whose BPD symptomatology likely manifested in a traditional manner rather than the somatic preoccupation encountered in the subgroup of BPD patients seen in primary care settings.

To provide a broader context, it is important to clarify that BPD may co-exist with genuine medical illness. Given the tendency of Axis I disorders to be exacerbated by Axis II disorders, we suspect that bona fide medical illness may be exacerbated by BPD as well. In summary, BPD may affect immunity to medical illness, manifest as somatic preoccupation, or co-exist with genuine medical illness.

Classic BPD Symptoms with Primary-Care Nuances

In the following section, we overview several symptoms or syndromes that tend to manifest predominantly in the primary care setting, but have the traditional dynamic underpinnings of BPD.

Medically self-sabotaging behavior. According to the DSM-IV, chronic self-destructive behavior is one of nine diagnostic criteria for BPD. The importance of these behaviors is supported by other diagnostic measures including the Diagnostic Interview for Borderlines,³⁸ which is considered by many to be the gold standard for BPD diagnosis in the research setting. It is not surprising that self-destructive behavior would have a slightly different manifestation in the primary setting—one with a more medicalized focus.

In one of our initial studies in this area,³⁹ we examined whether primary care patients would openly acknowledge medically self-sabotaging behaviors. In a sample of 411 outpatients being seen in a university-based family practice clinic, we found that 6.6 percent of the sample acknowledged the intentional sabotage of their medical care. Examples of endorsed behaviors included the following: exposed him- or herself to an infected person on purpose; created symptoms to attract the attention of a physician or nurse; exaggerated symptoms to attract the attention of a physician or nurse; purposefully misused prescription medications to worsen an illness; did not follow directions given by a physician or nurse in order to prolong illness; tampered with medical equipment to create false readings; lied about treatment recommendations to family to prolong illness; and prevented a wound from healing. At the outset of analyses, we excluded two high-response items (e.g., intentionally not taking a prescribed medication [25%] and intentionally not seeking medical care when needed [37%] ) because of concerns about participants' interpretation.

From this previous research, we confirmed that medicalized self-harm behavior existed, but we could not assert its relationship with BPD. In a second study,⁴⁰ we examined 118 internal medicine outpatients to determine the relationship between medically self-harming behaviors and two measures of BPD. Among medicalized self-harmers, 80 percent and 87 percent exceeded the cutoff scores for the two BPD measures, compared with 15 percent and 43 percent of non-self-harmers. These between-group differences were statistically significant (p<0.00001, p<0.002). These data confirmed a correlation between medicalized self-harm behavior and BPD.

We conducted another study to determine if medically self-sabotaging behavior was evident among psychiatric patients.⁴¹ We did so in an inpatient psychiatric sample of 77 patients and examined three specific behaviors. For patients without BPD, the mean number of endorsed medically self-sabotaging behaviors was 0.41, compared with 1.19 for the 32 patients who scored positively on all three measures of BPD, which was a significant difference. So, as in primary care settings, we found evidence of medically self-harming behavior in a psychiatric setting, as well. This suggests that both populations are capable of medically self-harming behavior, but the predominance or severity of such behavior probably determines the treatment setting.

As a caveat, medically self-sabotaging behavior might be interpreted as co-existing factitious illness; however, factitious illness itself may actually be a variant of BPD (please see Factitious Illness). In factitious illness, there is the intentional production of symptoms, as in the intentional generation of self-harm behavior encountered in BPD. In factitious illness, the motivation is to assume the sick role, as we are proposing for the role of somatic symptoms in BPD. External incentives are absent in factitious illness, which leaves emotional incentives, such as eliciting caring responses from others, as found in BPD.

Perceptions of illness (diabetes). The negative misperception of one's illness is relevant to BPD in terms of fostering a sense of medical victimhood. We examined perceptions of illness in a small sample of diabetic patients with and without BPD.⁴² While there were no between-group differences with regard to body mass index, the subsample with BPD symptoms was more likely to be on insulin. In addition, in the BPD subsample, 50 percent acknowledged one diabetes-related complication, yet no such complications were noted in the medical record. In the non-BPD subsample, nearly 90 percent correctly acknowledged a diabetes complication that was confirmed in the medical record. These findings suggest that the BPD group perceived themselves as more disabled. Whether these negative misperceptions of illness extend to other disease states is unknown, but likely. Again, in keeping with traditional BPD dynamics, negative distortions or a sense of disability in the perception of illness perpetuate the victim role. These negative distortions are likely to affect a patient's willingness to engage in and comply with a treatment.

Pain syndromes. Given that BPD is characterized by self-regulation difficulties, it is perhaps not surprising that disturbances in the regulation of pain sensation and pain states would emerge in primary care settings. Using the Diagnostic Interview for Borderlines,³⁸ we found in a primary-care sample of chronic pain patients that 50 percent met the diagnostic criteria for BPD.⁴³ Using projective testing, Merceron and colleagues⁴⁴ encountered borderline personality features among chronic pain patients. Finally, Burton and colleagues⁴⁵ found that chronic pain in combination with BPD significantly predicted less likelihood of return-to-work. Again, these findings indicate that a subset of chronic pain patients suffers from BPD, with the relationship most likely being mediated by self-regulation difficulties.

Not all studies are consistent with the preceding findings. For example, Kaylor⁴⁶ examined 166 chronic pain patients with the Personality Assessment Inventory and concluded that chronicity was not related to borderline personality features. However, this study did not address chronic pain and its association with BPD features.

Prescription misuse/abuse. In keeping with the theme of self-regulation difficulties, prescription substance abuse would seem to be a practical concern. Indeed, those with BPD clearly have self-regulation difficulties that result in the abuse of various substances. In this regard, in a study among 379 inpatients with BPD, Zanarini and colleagues⁴⁷ found that 64 percent had comorbid substance abuse problems. Dulit and colleagues⁴⁸ found similar results (67%). In the Dulit study, 15 percent of the BPD sample abused opioids, which is comparable to the findings of Skinstad and Swain.⁴⁹ These data suggest that in primary care settings, BPD patients may abuse potentially addicting medications, such as analgesics and high-potentcy benzodiazepines.

Human immunodeficiency virus (HIV). Because self-regulation difficulties may relate to both substance abuse and promiscuity, HIV patients may be at particular risk for comorbid BPD. Such relationships have been confirmed in two studies.^50,51 In the study by Ellis and colleagues,⁵⁰ HIV patients were diagnosed more often with BPD than control patients who were referred for psychiatric consultation.

Skin picking/excoriation. Because BPD is associated with self-harm behavior, any form of self-mutilation would be suspicious for this disorder, such as skin damage. Arnold and colleagues⁵² confirmed a diagnosis of BPD in patients with psychogenic excoriation. Wilhelm and colleagues⁵³ examined 31 subjects with repetitive skin picking and found that 26 percent suffered from BPD. Finally, Reis and colleagues⁵⁴ described a patient with self-induced skin ulcerations and BPD.

Factitious illness. Gunderson⁵⁵ describes one function of self-harm behavior as the elicitation of caring responses from others. A variation of this dynamic—the need to have a medical symptom to elicit emotional involvement from others—may explain the relationship between BPD and several factitious disorders.

Several studies have examined the relationship between factitious disorders and BPD. In a review of the literature, Sutherland and Rodin⁵⁶ found that factitious disorders were associated with BPD, among other psychiatric disorders. Several authors have associated pseudoseizures or psychogenic seizures with BPD.^57–59 Finally, BPD has been reported in factitious vomiting,⁶⁰ iron deficiency anemia actually caused by intentional blood-letting,⁶¹ psychogenic purpura,⁶² and Munchausen's syndrome alone or by proxy.⁶³ Indeed, factitious illness may be a variant of BPD in some individuals.

Other Medical Phenomena Associated with BPD

Plastic surgery. In a study of plastic surgery patients, Napoleon⁶⁴ found that compared with other patients, those with BPD requested a much higher number of areas for surgery, perceived plastic surgery as more serious (except for those with paranoid personality features), and experienced the lowest levels of post-operative satisfaction. These findings may relate to the very negative self-image harbored by BPD patients and the strong need to externally alter that image, but without much subsequent satisfaction.

Rheumatoid arthritis. Among 15 patients with rheumatoid arthritis, Marcenaro and colleagues⁶⁵ found that 40 percent met the criteria for BPD. Again, rather than a direct relationship, the association between BPD and rheumatoid arthritis may be mediated by early developmental trauma and its subsequent effects on immunity. If so, it would seem reasonable to determine the prevalence of BPD among other types of autoimmune diseases.

Obesity. Given that obesity is a multidetermined disorder, one contributory variable may be BPD because of the associated difficulties with self-regulation. In a review of the literature,⁶⁶ we found that the prevalence rates of BPD among the obese varied anywhere between 2.2 and 94.1 percent; however, two thirds of the assessment measures detected BPD rates of 25 percent or higher in study populations. In our own study, using semi-structured interviews, we found the prevalence of BPD among obese female primary care patients to be seven percent.⁶⁷ In examining the prevalence of BPD among those with binge-eating disorder, a disorder often characterized by obesity, we found that 12 percent of 479 subjects in eight studies met the criteria.⁶⁸

Disability. Given that BPD unfortunately seems to harbor some relationship with childhood victimization, it is not surprising that this theme perpetuates itself in adulthood. Among medical patients, the theme of victimization may manifest as medical disability. In support of this, several studies have found associations between various facets of disability and personality disorder. For example, Ekselius and colleagues⁶⁹ found that Cluster B personality disorders (e.g., erratic/dramatic personality disorders such as histrionic, borderline, antisocial, and narcissistic personalities) predicted an earlier age of longstanding work disability.

Jackson and Burgess⁷⁰ found that personality disorder was a significant predictor of disability among back pain patients, and personality disorder was found to be associated with a lack of return to work at six months by other investigators.⁷¹ Among chronic pain patients with personality disorder, Wijeratne and colleagues⁷² found a relatively higher level of physical disability. In a pain clinic setting, Allaz and colleagues⁷³ found a higher frequency of personality disorder among those with litigation neurosis versus those without.

As for studies explicitly examining for BPD, we found among a sample of 45 internal medicine patients that 72 percent of the disabled versus 26 percent of nondisabled participants met criteria on at least one of two measures for BPD.⁷⁴ The relationship between disability and BPD has been reported for psychiatric disability as well.⁷⁵

Conclusion

In this article, we have attempted to summarize the data between BPD and the primary care setting in terms of prevalence, healthcare utilization, and a variety of medical syndromes including somatic preoccupation and disability. While BPD may have more medicalized manifestations in the primary care setting (e.g., somatic preoccupation), the routine dynamic themes seen in psychiatric settings (e.g., self-regulation difficulties, self-destructive behavior, role of victimhood in adulthood) are unmistakably present.

From the perspectives of diagnosis and treatment, and teaching and research, we believe that the BPD concept needs to be broadened to include those characteristics of the disorder that may be encountered in primary care settings. Without an awareness of the manifestations of this disorder in primary care settings, there is less opportunity for diagnosis. Without diagnosis, effective intervention strategies are less likely.

We have previously recommended explicit treatment strategies for this unique group of patients.⁷⁶ These include, for example, very conservative medical management (i.e., avoidance of addicting medications, careful referral to conservative practitioners), clearly defined treatment plans with the patient, firm boundaries, structured office environments (e.g., consistent staff working with the patient), reasonable accommodation of dependency needs (e.g., frequent but contained appointments), neutral limit setting, acceptance of limited symptom resolution, etc. Only through awareness and effective diagnosis can we begin to really fully understand the impact of BPD and in turn, provide better medical management for these difficult patients.