Borderline personality disorder

The concept of borderline personality disorder can be as difficult and frustrating to understand as the people it attempts to define. Hence, there is an abundance of literature on the topic attempting to explain its meaning, its aetiology, its parameters, its place in psychiatry and its management.

Problems of definition

The term “borderline” has interesting historical antecendents, as outlined below, as it was initially considered to be very close to schizophrenia. Currently, it can be used both to describe a spectrum, as described by Meissner, or a personality organization, as used by Kernberg. These ideas will be explored later in this paper. However, it has, since DSM-III, increasingly acquired a diagnostic entity which is reasonably specific. This has not stopped it being an increasingly overused diagnosis, suggesting that patients who are diagnostically confusing (or unpleasant in their interactions with the psychiatric professions, thereby causing strong countertransference reactions) are receiving the label “borderline” by default. Unfortunately, problems of definition will persist as DSM-IV and ICD-10 classifications are attempting to incorporate psychodynamic concepts into an empirical classification system. This issue is borne out nicely in the study done by Zimmerman and Mattia (1999), which looked at the differences between clinical and research practice in diagnosing borderline personality disorder. Individuals who were assessed in a structured interview using the Structured Interview for DSM-IV Personality were compared with those who were assessed in an unstructured clinical interview. They found that clinicians were much less likely to diagnose the disorder than those who had administered the structured, directed interview.

Historical background

The concept of abnormal personality predates modern theories of normal personality. In 1801, the French psychiatrist Pinel described “manie sans delire” as a condition in which patients were prone to outbursts of rage and violence but were not deluded.

In 1835, J.C. Prichard (in Britain) suggested an alternative term – “moral insanity” – which he defined as a “morbid perversion of the natural feelings, affections, inclinations, temper, habits, moral dispositions and natural impulses without any remarkable disorder or defect of the intellect or knowing or reasoning faculties and in particular without any insane delusion or hallucination.”

In 1891, Koch coined the term “psychopathic inferiority”, which was later changed, by Kraepelin, to “psychopathic personality” to remove the judgemental overtones of the word “inferiority”. Kraepelin devised seven categories of psychopathic personality but they really only applied to people causing inconvenience or suffering to others. Schneider extended the concept to include those who caused suffering to themselves.

The concept of “psychopathic personality” has been carried well into the 20th century, with varying interpretations. The term “borderline” had its antecedents in the concept of “pseudoneurotic schizophrenia”, which was coined in the late 40’s to describe the symptomatic pattern of “panneurosis”, “pananxiety” and “pansexuality” which affected those “in between” people who were not schizophrenic but were too disturbed for psychoanalytic treatment. In 1954, Robert Knight further categorised this ill-defined group by focusing on several impairments in ego functioning, including the inability to plan realistically, the incapacity to defend against primitive impulses, and the predominance of primary process thinking over secondary process thinking.

Grinker, in 1968, outlined 4 diagnostic subgroups: the psychotic border, core borderline syndrome, the as-if group, and the neurotic border. Regardless of the sub-type, however, Grinker identified 4 features common to all with the borderline syndrome: 1) anger as the main or only affect; 2) defects in interpersonal relationships; 3) absence of consistent self-identity; and 4) pervasive depression. Grinker’s studies found, importantly, that this syndrome was distinct from schizophrenia (something which had previously not been clear) and that these patients did not deteriorate into schizophrenia over time.

Current definitions

1. DSM-IV: borderline personality disorder is part of cluster B in the 3 current cluster groups. Clusters started with DSM-I and were based on purely phenomenological similarities. DSM-IV was the first DSM to take into account 4 quite distinct theoretical frameworks: dynamic, trait, biological and sociological.

The grouping of spectrum disorders, self disorders and trait disorders has also evolved, with the spectrum disorders having likely biogenetic sources in common with certain Axis I diagnoses; self disorders being characterised by severe developmental pathology; and trait disorders being extreme and maladaptive variants of normal personality types. The self disorders are schizoid, antisocial and borderline, in this system. According to Gunderson, the rationale for this grouping is that these disorders are associated with traumatic early childhoods and markedly disturbed family environments. Such sources led to diffuse or fragile identities that are sensitive to social context. Their courses are more unstable and variable than are the courses of other personality disorders.

At the same time, DSM-IV attempted to take into account a data base of 5000 structured interviews with personality-disordered patients, leading to some – minor – changes from DSM-IIIR criteria.

2. ICD-10: does not use “clusters”, instead using 9 separate disorders with the 4th being “emotionally unstable – impulsive type and borderline type”, the latter according fairly closely with DSM-IV’s borderline personality disorder.

3. DSM-5 (Categorical Model): Despite a number of alternative models being proposed, the decision was ultimately made to retain the DSM-IV categorical approach to diagnosis of personality disorders. The main change in DSM-5 is a shift away from the multiaxial system that previously separated personality disorders (Axis II) from other mental disorders (Axis I). This shift recognises that the division between personality disorders and other mental disorders is somewhat arbitrary. In the DSM-5, Axes I, II and III (from past editions of DSM) are combined to give a single way of documenting all mental and other medical diagnoses. Separate notations are then added for significant psychosocial and contextual factors (formerly Axis IV) and disability (formerly Axis V).

DSM-5 (Alternative Dimensional Model): Similar to the official, current DSM-5 categorical model of diagnosis, the proposed alternative dimensional model retains some of the same essential features of personality disorder diagnosis. First, in order to meet the proposed dimensional model’s criteria there must be evidence of impaired functioning. In the dimensional model, the impairments are described in terms of impairment with respect to self, and impairment with respect to others. The second key feature is inflexibility. In other words, the impairments and personality traits are stable across a broad range of situations and across time. Therefore, both the official categorical diagnostic model, and the proposed dimensional model, both require significant impairment and inflexibility.

The DSM-5 proposed dimensional model includes two dimensions: Criterion A: level of personality functioning; and Criterion B: pathological personality traits.

Aetiological theories

1. Genetic

Family-pedigree studies have suggested that patients with borderline personality disorder (BPD) may be non-specifically predisposed toward poor regulation of both impulses and moods. So far, small twin studies have failed to support a genetic transmission. Indeed, in Torgensen’s 1994 study, there was no concordance between monozygotic twins compared with 29% in dizygotic twins. At present, it is clear only that there is some familial nongenetic transmission of BPD. Family studies show a high frequency of relatives with substance abuse, violence and anti-social PD, as well as a 5-fold increase in BPD itself.

2. Neurobiological

Some studies have shown that there are deficiencies in central serotonergic function in BPD, which may be linked to the impulsive and hostile behaviours often displayed. This could explain the favourable results of studies of SSRI’s in BPD (see Mx section). Hormones have been considered and impulsivity has been associated with increased levels of testosterone and some patients with BPD have abnormal cortisol levels. There are also “neurological soft signs”, subtle neuropsychological deficits and EEG changes (specifically slow-wave activity) reported in the literature. However, these are nonspecific.

3. Developmental

Kernberg (1975) targetted a specific developmental phase in infancy as being pathogenic in relation to BPD. After the infant, in Mahler’s developmental scheme, successfully traverses the symbiotic phase so that he appreciates the distinction between self and other (by 16 months), he becomes fixated during the separation-individuation phase (between 16 and 30 months). At this stage, the child becomes alarmed about the potential for his mother to disappear which is normally settled by consistent reappearance over time. Kernberg postulates that, for those who develop BPD, the fixation is due to a disturbance of the mother’s emotional availability during this period, due either to a constitutional excess of aggression in the child or to maternal problems with parenting, or even a combination of both. Thus, borderline patients can be viewed as repeatedly reliving this early infantile crisis.

Masterson and Rinsley (1975) also focused on the rapprochement phase of separation-individuation. However, they stressed the behaviour of the mother rather than the innate aggression of the child. They saw them as typically borderline themselves and highly conflicted about their children growing up. In the interaction the child receives the message that growing up and moving away will result in the loss of maternal love; remaining dependent (regressed and clinging) is the only means of maintaining the maternal bond. The fixation at this fragmented level leaves the borderline patient feeling that there are only two choices – you can feel abandoned and bad or you can feel good by denying reality and never growing up.

Adler (1985) based his model on the developmental framework of Fraiberg (cf. Mahler, above), in which at around 18 months of age a child normally can summon up an internal image of a maternal figure even in her physical absence. This was termed “evocative memory”. Adler believed that this capacity is only tenuously established in the borderline patient and, in situations of stress, he will regress to a developmental stage prior to the establishment of evocative memory. Adler saw this as the lack of a holding-soothing introject, which creates feelings of emptiness and depressive tendencies. In the absence of selfobject responses from significant others, borderline individuals have inadequate internal resources to sustain them and are prone to fragmentations of the self.

These three major psychodynamic models have been criticised for several reasons, including the fact that they fixate on only one developmental stage and that they heavily blame the mother. Others see later childhood trauma and abuse as more relevant to the development of BPD. Empirical research has largely substantiated the estimation that many borderline patients have experienced real victimisation in their childhood relationships with parents and other caretakers.

An important study in this respect, carried out by Herman et al. in 1989, found that 68% of borderline patients had been sexually abused as children, 71% had been physically abused, and 62% had witnessed serious domestic violence. It has been hypothesised by Gunderson that traumatic experiences in childhood may contribute to image-distorting defenses such as splitting, denial, and projective identification. In 1997, Zanarini et al. (in a larger study of 358 subjects) found of patients with BPD that 91% had experienced abuse (61% sexual) and 92% had experienced neglect. These authors speculated that sexual abuse is an important aetiological factor in BPD but that neglect may predispose both to BPD and make the child more at risk of sexual abuse. Importantly, the general family atmosphere of chaos and neglect by both parents is now seen as more important than the mother’s role alone.

Schachnow et al. (1997) emphasises the importance of pathology of both parents in the development of BPD. They believed that this is transmitted by both genetic and environmental pathways. Major studies by figures such as Parker, Frank and Paris, Soloff and Millward, and Zweig-Frank (in the 80’s and 90’s) have shown that both parents are implicated in the child’s psychopathology. A pattern of “affectionless control” was seen, with both parents failing to provide support and preventing the child from separating.

Gunderson et al. (1980) noted that borderline families were “…characterised by the rigid tightness of the marital bond to the exclusion of the attention, support or protection of the children”. This meant that a poor or non-existent relationship with one parent was not offset by a positive one with the other. Thus, the parental unit failed to provide nurturing or caring. Gunderson saw that inadequate nurturing in the first 2 years contributed to the development of BPD.

Mandelbaum (1980) suggests that parents of children who later develop BPD use them as targets for projecting their own marital and personal problems. Both parents are enmeshed with their own families of origin, although this is obscured by a distant relationship with them, which is not to be mistaken for differentiation and autonomy. Feldman and Guttman (1984) believed that there was a pattern in which one parent is severely personality disordered, while the other fails to protect the child against the adverse effects of this. Other studies support the epidemiological findings of familial psychopathology with depression, alcoholism, substance abuse, and so on, in the parents.

4. Multifactorial

From the above findings, it is clear that BPD is multifactorial in aetiology. A recent Australian paper has attempted to present a coherent multifactorial picture. Russell Meares et al. (1999) have developed what they term a Jacksonian and biopsychosocial model for BPD. They note that “Recent observations suggest that BPD is associated with “soft” neurophysiological deficits, particularly of a frontal kind”, though they feel it unlikely that any neurophysical deficit in BPD will be of a simple localised kind. They also point out that these findings may not be aetiological but rather the consequences of severe trauma. For example, the finding of reduced hippocampal volume may accord with the diminished expression of “emotional memory” (which is thought to centred in the amygdala) in BPD.

Overall though, as they state, “The vulnerability factor which seems most likely to be operative in the development of BPD is a dysfunctional family environment.” They cite several examples of this in other studies and state that “…the developmental background of those with BPD disrupts the emergence of higher order psychological functions which include the system of self.” The concept of disturbance of “self” was first espoused by Hughlings Jackson in the late 19th century: “He suggested that those functions which have evolved last and which emerge late in human development are more fragile, more easily disrupted by insults to the brain-mind system than those functions which, as it were, are more hard-wired and appear earlier in evolutionary history. This process, the reverse of evolution, he called ‘dissolution’.”

Jackson conceived of mind, or self, as double, made up of two poles, subject and object, one pole of awareness, the other of the objects of awareness. He thought of “self” as an awareness of the stream of consciousness. He saw “self” as being a complete system of unification of the whole organism, dependent upon the evolution of anatomically new structures, especially the prefrontal cortex. Thus, as a consequence of insults to the brain-mind system, the development of “self” will be deficient: “The stream of consciousness will be stunted, or perhaps absent, leaving a painful emptiness which is a cardinal feature of BPD. Much of the impulsive and maladaptive behaviour of BPD can be understood as a desperate attempt to fill this emptiness.”

Studies suggest that affect regulation is dependent upon prefrontal activity. Jackson’s hypothesis would fit with this, as affect dysregulation is another cardinal feature of BPD. Again, prefrontal activity is vital to attentional focus. Meares hypothesises that the somatisation found in BPD may be a consequence, at least in part, of a disturbance of attentional focus. Meares also addresses the symptoms of dissociation and depersonalisation in terms of Jackson’s schema.

The concept of maturation as involving the coordination of ‘biogenesis’ (the genetically given biological programme) with ‘sociogenesis’ (appropriate environmental provision) was originally developed by Vygotsky and Luria. Meares believes that the borderline and related phenomena are a consequence of failure of ‘sociogenesis’, thus leading to deficient maturation. However, he suggests that :”since the suggested maturation failure is conceived as ‘experience-dependent’, it is considered to be reversible, at least in part”.


According to DSM-IV, BPD occurs is 2-3% of the general population and is by far the most common personality disorder in clinical settings. It is estimated to occur in 11% of outpatient populations, 19% of inpatient populations and in 27-63% of PD clinical populations. It would seem to be 3 times more common in women than men, but several sources have suggested that this is because men are more often given the diagnosis of antisocial or narcissistic PD, when in some cases they may have BPD.

It is 5 times more common among 1st-degree relatives of those with BPD than in the general population. There is also an increased familial risk for substance abuse, antisocial PD and depressive disorders.


There are high rates of co-morbidity with depression, substance abuse, and other personality disorders, as has been seen in many studies. Some authors have suggested that these high rates raise questions about the validity of the diagnostic category. DSM-IV has improved sensitivity and specificity over its predecessors, but makes no claims about its validity. This will remain an ongoing issue.

When the term “borderline” was first used, it related to the belief at the time that it was on the border with schizophrenia. Time and many studies have shown this not to be the case – rather, that schizotypal personality disorder has a greater claim to this status. Some believe that there is still an association as those with BPD quite commonly have psychotic episodes. However, these are transient and stress-related in BPD.

Depression has been seen to have high rates of co-morbidity. Long-term follow-up studies have shown a 70% to 80% lifetime prevalence of major depression in patients with BPD. Torgensen cites a higher frequency of depression among relatives of BPD probands compared with schizophrenic probands, but only in those probands who had co-morbid depression. Despite the clear evidence of coexistence, there is no evidence that the 2 disorders are related. The partial efficacy of antidepressants to treat BPD is cited as evidence of a relationship, though it is not proof. Gabbard speaks of the need to differentiate the “characterological depression” typical of BPD from major depressive disorder. “Depression” may be used to describe chronic feelings of boredom, emptiness, loneliness and abandonment by those with BPD.

Atre-Vaidya and Hussain (1999) found in their study that less than 50% of the patients receiving the clinical diagnosis of BPD actually satisfy the DSM criteria. They had set out to find if BPD was a variant of bipolar disorder and if it was distinct from other cluster B personality disorders. They used an instrument that incorporates both biological and psychosocial aspects of personality and found that borderline patients could be distinguished from bipolar mood disorder based on temperament and character deviation. They also found that BPD is not distinguishable from other cluster B disorders.

PTSD has often been compared with BPD, as there is often the common finding of a history of significant childhood trauma. Gabbard believes that one third of borderline patients satisfy the criteria for PTSD. However, this is a controversial topic as there are strong arguments against it. Zanarini et al. (1988) found that PTSD was a common but not universal comorbid disorder and concluded that their findings were inconsistent with the view that BPD is actually a form of chronic PTSD. They also found that anxiety disorders were almost as common as mood disorders in BPD, that substance use disorders were more common in males with BPD and eating disorders in females with BPD. They argue that the lifetime pattern of axis I comorbidity characteristic of borderline patients and distinguishing for the disorder is a particularly good marker for BPD.


An Australian study by Clarke et al.(1995) found that “the treatment of BPD in the state mental health system was generally haphazard and ineffective.” As BPD does not fit the criteria for “serious mental illness”, its sufferers are often taken into care only when a crisis emerges and are then often only managed cursorily. Limited mental health funds and health practitioners’ attitudes to BPD make effective management even less likely, combined with the fact that people with BPD are difficult to engage and to keep engaged in any meaningful therapy.

Hospital admission is usually only provided for brief respite in times of crisis. Ongoing follow-up therapy is not always provided with enthusiasm. The private mental health sector bears the brunt of most people with BPD, though the most severe cases are, of necessity, in the public domain.

Gabbard describes extended, psychodynamically informed hospital treatment with the goals of ‘ego-building and modification of the patient’s internal world”. Quaytman and Sharfstein also argue the benefits of long-term hospitalisation. However, they are realistic about the costs of this and believe that a short hospital stay followed by closely managed, consistent care may have the same results over time. They also address the negative aspects of long hospitalisation, in terms of dependency. Gabbard sees a degree of dependent regression as an important part of the hospitalisation process and a stage to be passed through. Other negatives include the occurrence of splitting by the patient and the presence of negative countertransference by staff. Gabbard believes that these can be addressed and minimised, for the benefit of all. Dolan et al.(1997) believe that inpatient care is vital for long-term benefit, in terms of changes in PD psychopathology, and that there were was a positive relationship in their study to the length of stay in treatment.

Pharmacological treatment in BPD has been unimpressive overall. Some studies cite benefit from antipsychotic medications – on symptoms of psychosis, impulsive behaviour and hostility – and others refute this. A study carried out by Chengappa et al.(1999), looks at the role of clozapine in severe self-mutilation and aggression. Though theirs was a small, and therefore preliminary, study, they found that there was a statistically significant reduction in incidents as well as a two-fold increase in the patients GAF scores.

Antidepressants are the other frequently used medications for BPD. Past studies have not been impressive, with tricyclics. Open trials of fluoxetine in patients with severe chronic illness showed dramatic reductions in self-injury, as have open trials of sertraline and venlafaxine. However, a randomised controlled trial of fluoxetine gave equivocal results. Hirschfeld (1997) believes that pharmacotherapy should be directed towards the different symptom groups. He classifies 4 groups: affective, impulsive, ego-interpersonal, and psychotic. In relation to affective and impulsive symptoms, he states that overall the MAOI’s, the SSRI’s and venlafaxine provide the widest spectrum of effective treatment.

Benzodiazepines can be problematic in BPD: firstly, because of the potential for abuse and, secondly, because they can have a disinhibiting effect which can worsen self-mutilation and cause ‘rage reactions’. Mood stabilisers, particularly carbamazepine and, to a lesser extent, sodium valproate have shown modest benefits on impulsive behaviour.

Psychotherapy on an individual basis is the mainstay of treatment of BPD. The most effective form of psychotherapy, however, depends on many factors: the severity of the patient’s illness, the presence of comorbid conditions, the response to pharmacotherapy, the therapists’s perspective (i.e., psychodynamic, CBT, etc.) and the ability of the patient to be engaged in the therapy process over time. Josephine Beatson argues that patients with borderline pathology who have sustained severe early developmental trauma will often require long-term psychotherapeutic treatment to achieve lasting psychological change.

Psychodynamic psychotherapy, well described in Gabbard, may not be appropriate to many patients with BPD who find it difficult to cope with silences and lack of direction and limits. For these patients, a more supportive, directive type of therapy is needed. Gunderson describes therapy as passing through several phases: in the first year of treatment there is a reduction in self-harm accompanied by dependence on the therapist; in the second year some stable role functioning is seen; in the third year meaningful transference work becomes possible. A major difficulty with measuring the success of any treatment of BPD is the high rate of drop-out, given the nature of the patient population.

1. Atre-Vaidya, N. & Hussain, S.M., “Borderline Personality Disorder and Bipolar Mood Disorder: Two Distinct Disorders or a Continuum?” in Journal of Nervous & Mental Disease, 187(5):313-5, May 1999.
2. Beatson, Josephine A., “Long-term psychotherapy in borderline and narcissistic disorders: when is it necessary?” in ANZJP 29:591-7, 1995.
3. Bloch, Sidney, et al., The Family in Clinical Psychiatry. Oxford, 1995.
4. Brodsky, B.S., et al., “Characteristics of borderline personality disorder associated with suicidal behaviour” in Amer. J. Psych. 154(12):1715-9, Dec., 1997.
5. Chengappa, K.N., et al., “Clozapine reduces severe self-mutilation and aggression in psychotic patients with borderline personality disorder” in J. Clin. Psych. 60(7):477-84, July, 1999.
6. Dawson, D. & Macmillan, H.L., Relationship Management of the Borderline Patient: From Understanding to Treatment. N.Y., 1993.
7. Gabbard, Glen O., “Finding the ‘Person’ in Personality Disorders” in Am. J. Psychiatry 154:7, July 1997.
8. Gabbard, Glen O., Psychodynamic Psychiatry in Clinical Practice. Washington, 1994.
9. Gelder, Michael, et al., Oxford Textbook of Psychiatry (3rd edition). Oxford, 1996.
10. Hirschfeld, R.M., “Pharmacotherapy of borderline personality disorder” in J. Clin. Psych. 58 Suppl.14:48-53, 1997.
11. Hooley, J.M. & Hoffman, P.D., “Expressed emotion and clinical outcome in borderline personality disorder” in Amer. J. Psych.156(10):1557-62, Oct., 1999.
12. Kaplan, H.I. & Sadock, B.J.(eds.) Comprehensive Textbook of Psychiatry. 6th edition, Baltimore, 1995.
13. Leichsenring, F., “Primitive defense mechanisms in schizophrenics and borderline patients” in J. Nervous & Mental Disease 187(4):229036, April, 1999.
14. Levine, D., et al., “Emotion processing in borderline personality disorders” in J. Nervous & Mental Disease 185(4):240-6, April, 1997.
15. Meares, Russell, et al, “A Jacksonian and biopsychosocial hypothesis concerning borderline and related phenomena” in ANZJP 33(6):831-40, December, 1999.
16. Rutter, Michael, et al., Child and Adolescent Psychiatry: Modern Approaches. Oxford, 1994.
17. Schimmel, P., “Swimming against the tide? A review of the therapeutic community” in ANZJP 31(1):1220-7, Feb., 1997.
18. Shachnow, J., et al., “Biparental psychopathology and borderline personality disorder” in Psychiatry 60(2):171-81, Summer, 1997.
19. Share, Isaiah A. (ed.), The Psychiatric Clinics of North America: Borderline Personality Disorder. Philadelphia, December, 1994.
20. Zanarini, M.C., et al., “Reported pathological experiences associated with the development of borderline personality disorder” in Amer. J. Psych. 154(8):1101-6, August, 1997.
21. Zanarini, M.C., et al., “Axis I comorbidity of borderline personality disorder” in Amer. J. Psych. 155(12):1773-9, Dec., 1998.
22. Zimmerman, M. & Mattia, J.L., “Differences between clinical and research practices in diagnosing borderline personality disorder” in Amer. J. Psych. 156(10):1570-4, Oct., 1999.

Henry CaudleAUTHOR | Dr Henry Caudle
Dr Caudle is a psychiatrist with special interests in the management of eating disorders, psychosomatic medicine, pain psychiatry, addiction psychiatry and mental health problems associated with general medical and surgical illness.


Consulting suite for outpatient appointments
PH | 03 9805 4309  FAX | 03 9805 4388

Intake clinician for inpatient referrals and enquiries
PH | 03 9805 4338  FAX | 03 9805 4233 FAX |


Keep up to date with relevant GP information including upcoming professional development events by subscribing to our email newsletter.
Your email address will not be sold or shared with anyone else. Refer to our privacy policy for more details.