OUP user menu

Evidence-based pharmacotherapy for personality disorders

Luis H. Ripoll, Joseph Triebwasser, Larry J. Siever
DOI: http://dx.doi.org/10.1017/S1461145711000071 1257-1288 First published online: 1 October 2011


Patients with personality disorders are prescribed psychotropic medications with greater frequency than almost any other diagnostic group. Prescribing practices in these populations are often based on anecdotal evidence rather than rigorous data. Although evidence-based psychotherapy remains an integral part of treatment, Axis II psychopathology is increasingly conceptualized according to neurobiological substrates that correspond to specific psychopharmacological strategies. We summarize the best available evidence regarding medication treatment of personality disordered patients and provide optimal strategies for evidence-based practice. Most available evidence is concentrated around borderline and schizotypal personality disorders, with some additional evidence concerning the treatment of avoidant and antisocial personality disorders. Although maladaptive personality symptoms respond to antidepressants, antipsychotics, mood stabilizers, and other medications, evidence-based pharmacotherapy is most useful in treating circumscribed symptom domains and induces only partial improvement. Most available evidence supports use of medication in reducing impulsivity and aggression, characteristic of borderline and antisocial psychopathology. Efforts have also begun to reduce psychotic-like symptoms and improve cognitive deficits characteristic of schizotypy. Indirect evidence is also provided for psychopharmacological reduction of social anxiety central to avoidant personality disorder. Evidence-based practice requires attention to domains of expected clinical improvement associated with a medication, relative to the potential risks. The development of future rational pharmacotherapy will require increased understanding of the neurobiological underpinnings of personality disorders and their component dimensions. Increasing efforts to translate personality theory and social cognitive neuroscience into increasingly specific neurobiological substrates may provide more effective targets for pharmacotherapy.

Key words
  • Avoidant personality disorder
  • borderline personality disorder
  • medication
  • psychopharmacology
  • schizotypal personality disorder


Personality disorders are defined by an ‘enduring pattern of inner experience and behavior that … is inflexible and pervasive across a broad range of personal and social situations’, with symptomatic disturbances in cognition, affect, impulsivity, and interpersonal functioning leading to distress (APA, 1994). Until recently, guidelines recommended sparing use of pharmacotherapy, and expectations remained guarded regarding expected benefits from medications. Since then, distinctions between Axis I disorders, considered ‘genetic … biological … brain disorders’ treated with medications; and Axis II disorders, alternatively considered ‘psychological’ and therefore treated with psychotherapy, has undergone a paradigm shift (Siever & Davis, 1991). In this atmosphere, clinicians must rely on the most up-to-date, evidence-based practices for pharmacotherapy to be effective.

Component dimensions of personality, such as impulsivity or aggressiveness, have demonstrable neurobiological correlates, as shown via a variety of endocrine, electrophysiological, and neuroimaging measures (Brambilla et al. 2004; Goodman et al. 2004; Houston et al. 2004; Juengling et al. 2003; Levitt et al. 2004; Minzenberg et al. 2006; New et al. 1997, 2004; Ogiso et al. 1993; Oquendo et al. 2005; Prossin et al. 2010; Rusch et al. 2003; Russ et al. 1991; Simeon et al. 1992; Soderstrom & Foresman, 2004). Identifying neurobiological substrates of personality has allowed for increasingly specific pharmacotherapy. Nevertheless, improvement from effective pharmacotherapeutic interventions is often transient and/or restricted to several symptom domains. In the USA, there are no FDA-approved medications for treating personality disorders. Thus, pharmacotherapy for personality disorders remains off-label, and psychopharmacological strategies for evidence-based practices remain lacking.

The majority of psychopharmacological research on personality disorders has focused on borderline personality disorder (BPD). In the most recent treatment guidelines for BPD, the American Psychiatric Association (APA, 2001), acknowledges that ‘pharmacotherapy has an important adjunctive role’, along with ‘extended psychotherapy to attain and maintain lasting improvement in … personality, interpersonal problems, and overall functioning’. Similarly, others have described psychopharmacological treatment of BPD as resulting only in ‘a mild degree of symptom relief’ (Paris, 2005). Moreover, there remains a dearth of evidence-based medication treatments for other personality disorders.

Often, pharmacotherapy for severe personality disorders is used to stabilize patients' symptoms sufficiently in order to facilitate psychosocial interventions and foster reflective functioning. Close communication between psychotherapists and psychopharmacologists remains crucial. Although functional gains can be expected from medications, the magnitude and time-course vary. There is little evidence regarding distinctions between acute and maintenance pharmacotherapy, or how long to continue patients on a medication. Empirical data on recurrence or relapse is similarly scarce. Therefore, evidence-based practices must be judged case-by-case, weighing clinical risks and benefits.

Clinicians can refer to accompanying tables for the best available evidence regarding pharmacotherapy for personality disorders (see Tables 14). This data was compiled by searching the Medline database with the main combinations pharmacotherapy and each of the various DSM-IV personality disorder diagnoses. In addition, we paid particularly close attention to randomized, placebo-controlled trials (along with some lower-level evidence if this type of evidence was severely limited). We focused on studies published in the past 3 years, since the publication of the last World Federation of Societies of Biological Psychiatry Guidelines for the Biological Treatment of Personality Disorders (Herpertz et al. 2007). Additional research regarding medications for treating impulsive aggression was found via a similar Medline search on impulsivity, aggression, and pharmacotherapy.

View this table:
Table 1

View this table:
Table 2

View this table:
Table 3

View this table:
Table 4

Unfortunately, only a few novel trials of pharmacotherapy for personality disorders have been published during this recent period. Several recent meta-analyses have been published in this time, which we utilized to establish areas of consensus for evidence-based practice, and identify gaps that need to be addressed with future research. Many prior reviews cover only BPD, but we expanded our scope to include all personality disorders. Thus, we include a comprehensive summary of the best, current evidence, with commentary on recent consensus and recommendations for evidence-based practices, and future directions regarding pharmacotherapeutic strategies that have been insufficiently tested, but appear promising for further research. This situates this review as a nexus, compiling evidence-based practices for treating personality disorders for interested clinicians, as well as providing avenues for future psychopharmacological research.

Schizotypal personality disorder (SPD)

SPD is characterized by interpersonal deficits and psychotic-like symptoms. Like schizophrenia patients, SPD patients often demonstrate cognitive deficits in working memory, particularly sustained attention and executive functioning (Bergida & Lenzenweger, 2006; McClure et al. 2007a; Parc & McTigue, 1997), as well as significant abnormalities in empathic understanding (Langdon & Coltheart, 2004; Pickup, 2006; Ripoll et al. unpublished data). Unlike schizophrenic patients, there is greater preservation of frontal volume in SPD (Siever & Davis, 2004).

Overall, clinical trials for SPD have been complicated by comorbidity, particularly with other personality disorders. Most early RCTs on BPD also included SPD patients (Goldberg et al. 1986; Serban & Siegel, 1984; Soloff et al. 1986c), because both SPD and BPD were considered rooted in ‘borderline’ schizophrenia; but psychotic symptoms in SPD and BPD are clinically distinguishable.

The conceptualization of SPD within the schizophrenia spectrum supports treatment with antipsychotic medications. Antipsychotics appear to be useful in the treatment of SPD, particularly in terms of psychotic-like symptoms (Goldberg et al. 1986; Koenigsberg et al. 2003). Open-label studies have suggested a role for antidepressants in treating self-injury, psychotic-like, and depressive symptomatology (Jensen & Andersen, 1989; Markovitz et al. 1991), but the evidence is weaker. Recent RCTs targeting cognitive deficits in SPD compared performance on neuropsychological tasks before and after treatment with medication or placebo. Both pergolide, a dopaminergic agonist active at both the D1 and D2 receptor (McClure et al. 2010), and the noradrenergic α2A agonist guanfacine (McClure et al. 2007b) improved SPD patients' cognitive performance, on distinct neuropsychological measures. Whether this improvement extends to overall clinical functioning in SPD remains subject to future investigation.

In sum, SPD patients respond to low-dose, atypical antipsychotics, targeting psychotic-like symptoms and general functioning. First-generation antipsychotic medication and antidepressants may also play a role, although the evidence is not as reliable. Evidence-based practice requires weighing risk of extrapyramidal side-effects or tardive dyskinesia against potential benefits. Cognitive enhancement via noradrenergic α2A or dopaminergic agonism may be future avenues of research, given that, by analogy with schizophrenia, the cognitive impairment in SPD may be responsible for the overall dysfunction observed in the disorder. Research efforts to understand neurobiological substrates of social cognitive dysfunction have heretofore mainly focused on BPD and schizophrenia. Because SPD involves social isolation, relational paranoia, and empathic deficits, research on pharmacotherapeutic effects on social cognition may also be fruitful.

Antisocial personality disorder (AsPD)

Peer-reviewed trials of AsPD include studies on groups of individuals likely to have been antisocial based on histories of repeated violence and criminality and an absence of other stated causes for these behaviours. Lithium has been associated with decreases in serious rule infractions in incarcerated males (Sheard, 1971; Sheard et al. 1976). Prisoners treated with phenytoin committed fewer aggressive acts and evidenced decreased tension-anxiety and depression-dejection (although not anger-hostility), and improvements in aggression appeared to be limited to impulsive (not pre-meditated) aggression (Barratt et al. 1991, 1997). At present, evidenced-based pharmacotherapy for AsPD is restricted to treatment of impulsive aggression. Future neurobiological research in AsPD and psychopathy will probably increase our understanding of the dysfunctional emotional empathy often seen in this disorder (Blair, 2005) and whether this may be susceptible to psychopharmacological intervention.


Most RCTs on personality disorders focused on BPD, which consists of several domains of dysfunction: affective instability, impulsivity and anger, transient psychotic or dissociative symptoms, and intense, unstable relationships (Lieb et al. 2004; Zanarini et al. 1990). BPD patients often demonstrate high comorbidity (Zanarini et al. 2004a, c) and make numerous suicide attempts and parasuicidal gestures, conferring significantly higher risk for completed suicide (Welch & Linehan, 2000).

Early studies employed a distinct nosology in characterizing subjects, some of whom actually had what might be called BPD today (Rifkin et al. 1972). In studies on suicidal or parasuicidal subjects, the majority often have BPD (Battaglia et al. 1999; Montgomery & Montgomery, 1982; Montgomery et al. 1983; Verkes et al. 1998). Early studies often included combinations of BPD and SPD subjects (Goldberg et al. 1986; Serban & Siegel, 1984; Soloff et al. 1986c), and studies recruiting a range of all personality disorders ultimately include BPD as the most frequent diagnosis (Coccaro & Kavoussi, 1997; Hollander et al. 2003).

Clinicians should exercise caution in attempting to apply research findings to severely ill BPD patients, as many RCTs recruited only outpatients, who further were excluded if they expressed acute suicidality (Frankenburg & Zanarini, 2002; Tritt et al. 2005; Zanarini & Frankenburg, 2003; Zanarini et al. 2004b) or had made a recent suicide attempt (Bogenschutz & Nurnberg, 2004). In addition, small sample sizes predominated, and most studies lasted ≤3 months. The few trials lasting ⩾6 months suffered from high drop-out rates (Frankenburg & Zanarini, 2002; Zanarini & Frankenburg, 2001) or concomitant recruitment of subjects without BPD (Battaglia et al. 1999; Montgomery et al. 1983; Verkes et al. 1998). Moreover, RCTs with BPD subjects appear to be prone to high placebo response rates (Lieb et al. 2004; Salzman et al. 1995), meaning that open-label trial data should be interpreted with caution.

APA practice guidelines (APA, 2001) recommended a symptom-targeted approach in pharmacotherapy of BPD. This leaves open the possibility for patients to improve in some but not all symptom dimensions. Some clinicians have based their decision to implement polypharmacy on this, but there is actually little evidence as to the effectiveness of this strategy. The only study on combined pharmacotherapy in BPD (Zanarini et al. 2004b) found no superior efficacy for combination treatment compared to one medication alone. Using as few medications as possible to target central areas of clinical dysfunction, together with evidence-based psychotherapy, is usually the optimal treatment strategy. In light of this, although the 2001 guidelines suggest a prominent role for serotonergic pharmacotherapy, recent reviews have questioned this and instead emphasized anticonvulsants and antipsychotics (Abraham & Calabrese, 2008; Mercer et al. 2009).

Tricyclic antidepressants (TCAs)

Disturbances of serotonin have been associated with BPD, impulsive aggression, self-harm, and suicidality (Coccaro et al. 1995; Evenden, 1999; Malone et al. 1996; Pitchot et al. 2005). Low CSF levels of serotonin metabolites have been associated with suicide attempts and completion (Samuelsson et al. 2006; Traskman et al. 1981), impulsivity, aggression (Mehlman et al. 1994; Virkkunen et al. 1994), lifetime aggressiveness, and suicidal lethality (Placidi et al. 2001). Impulsive aggression with suicidality has been linked to blunted prolactin responses to the serotonergic probe fenfluramine (Coccaro et al. 1989). PET scans of personality-disordered subjects high in impulsive aggression have demonstrated reduced response to fenfluramine in orbitofrontal, ventromedial, and cingulate regions (Siever et al. 1999).

Nevertheless, early research on TCAs for BPD proved disappointing (Montgomery et al. 1983; Soloff et al. 1989). Amitryptiline has been associated with paradoxical increases in suicidality, paranoia, and behavioural dysregulation, attributed to ‘generalized disinhibition of cognitive and affective controls’ (Soloff et al. 1986a, 1987). Indeed, borderline patients have difficulty cognitively resolving conflict among stimulus dimensions (Posner et al. 2002), and prefrontal hypofunction can be seen after a serotonergic stimulus in subjects with prominent impulsive aggression (New et al. 2002). Thus, medications with adverse cognitive sequelae, including anticholinergic side-effects, may contribute to worsening impulsivity. As mentioned in prior reviews, the use of TCAs in treating BPD is discouraged (Abraham & Calabrese, 2008; Mercer et al. 2009). Their use is also associated with potentially significant risk of overdose.

Monoamine oxidase inhibitors (MAOIs)

Despite hesitancy in prescribing MAOIs to patients with prominent impulsivity or self-injurious behaviour, some recommend these medications for BPD patients who can take them safely and reliably. Interest in MAOIs for BPD is rooted in their differential efficacy for conditions such as hysteroid dysphoria or atypical depression, viewed as being related to one other and BPD (Kayser et al. 1985; Liebowitz & Klein, 1981). In a crossover trial with multiple medication phases, only tranylcypromine was associated with higher patient-rated improvement scores and completion rates (Cowdry & Gardner, 1988).

Similarly, relative prominence of BPD symptoms predicted superiority of phenelzine (Parsons et al. 1989). Phenelzine is beneficial in the treatment of hostility, anxiety, and borderline symptoms (Soloff et al. 1993). In some patients, it could cause uncomfortable excitement and emotional reactivity (Cornelius et al. 1993). Thus, although there is evidence for their efficacy, many patients may not tolerate these medications. Other associated risks of MAOIs include toxicity in overdose and potentially fatal hypertensive crises or serotonin syndrome.

Selective serotonin reuptake inhibitors (SSRIs)

SSRIs are thought to potentiate serotonergic neuromodulation but demonstrate more favourable side-effect profiles. Fluoxetine reduced anger in BPD independent of any antidepressant effect (Salzman et al. 1995). It also improved verbal and impulsive aggression, irritability, and overall functioning (Coccaro & Kavoussi, 1997). Similarly, a RCT with paroxetine demonstrated efficacy in preventing recurrent suicidal behaviour but no significant effect on depression, hopelessness, or anger (Verkes et al. 1998). By contrast, there was little added benefit from fluoxetine when added to dialectical behavioural therapy (DBT) (Simpson et al. 2004).

On the other hand, fluvoxamine decreased affective lability, but not scores of impulsivity or aggression (Rinne et al. 2002). Although SSRIs decrease impulsivity and aggression in BPD patients with comorbid intermittent explosive disorder (IED; Coccaro & Kavoussi, 1997; New et al. 2004), data from BPD subjects without comorbid IED are inconsistent (Rinne et al. 2002). Previous reviews have emphasized that effect sizes for antidepressant pharmacotherapy vary widely between classes and trials (Ingenhoven et al. 2010; Lieb et al. 2010; Mercer et al. 2009). Nevertheless, current evidence-based practice recommends use of SSRIs, due to potential benefits on impulsive aggression that may outweigh associated risks. There has been no evidence that antidepressants alleviate the chronic emptiness, shameful self-concept, and intrapsychic pain in BPD.

First-generation antipsychotics

An early interest in antipsychotic medications for treating BPD probably arose from a conception of BPD as a variant of schizophrenia (e.g. Deutsch, 1942). Antipsychotics have demonstrated partial efficacy, reflecting underlying abnormalities in dopaminergic signalling. Borderline subjects demonstrate high levels of the dopamine metabolite, homovallinic acid, in both plasma and cerebrospinal fluid (Siever et al. unpublished data). Prior to more widespread use of SSRIs, antipsychotics demonstrated efficacy in decreasing psychotic-like symptoms (Goldberg et al. 1986; Soloff et al. 1986b), depression (Soloff et al. 1986b), irritability (Cornelius et al. 1993), and general symptom severity (Cowdry & Gardner, 1988; Soloff et al. 1986b).

A recent Cochrane review suggests haloperidol is efficacious in reducing anger in BPD, and treatment with flupenthixol decanoate reduced suicidal behaviour (Lieb et al. 2010). By contrast, evidence for efficacy of neuroleptics on affective symptoms, psychosis, and anxiety remains inconsistent. The dosage of antipsychotic medication for evidence-based treatment of BPD is usually lower than schizophrenia. High drop-out rates are noted, and risk of extrapyramidal symptoms may further limit the utility of neuroleptics.

Second-generation antipsychotics

Classical neuroleptics have largely been superseded by atypical antipsychotics, whose broader therapeutic benefits may be explained by activity beyond the D2 receptor. As mentioned in prior reviews, olanzapine has proven beneficial in treatment of BPD patients' anxiety, anger, interpersonal sensitivity, and paranoia, but not depression (Zanarini & Frankenburg, 2001), as well as improving general clinical functioning and BPD symptomatology (Bogenschutz & Nurnberg, 2004). A recent, large RCT demonstrated no effect of olanzapine on BPD symptoms (Schulz et al. 2008), although the authors suggested that patients may have been underdosed. A study comparing olanzapine to haloperidol showed no between-group differences except with respect to side-effects, with more weight gain associated with olanzapine and more extrapyramidal side-effects with haloperidol (Shafti & Shahveisi, 2010).

Adding fluoxetine to olanzapine did not elicit further benefit, except that subjects receiving both medications gained less weight than those receiving only olanzapine (Zanarini et al. 2004b). The addition of olanzapine to DBT reduced depression, anxiety, and impulsive aggression, but the magnitude and timing of these benefits relative to DBT was difficult to interpret (Soler et al. 2005).

Aripiprazole has a novel mechanism of action (partial agonist at the dopamine D2 receptor and serotonin 5-HT1A receptor, antagonist at the 5-HT2A receptor). It may be more favourable than other atypicals with respect to metabolic side-effects. A longer half-life may be more effective for patients susceptible to non-adherence. In non-suicidal BPD patients, aripiprazole was effective in reducing aggression, anxiety, depression, psychosis, interpersonal symptoms, self-injurious behaviour, and subjective distress. There were no significant differences between groups in weight gain (Nickel et al. 2006). An 18-month, open-label follow-up showed sustained improvements and continued tolerability (Nickel et al. 2007).

Previously, open-label trials suggested possible efficacy of ziprasidone in BPD patients during acute exacerbations (Pascual et al. 2004, 2006). The side-effect and psychopharmacological profiles of ziprasidone indicated lesser metabolic risks and antidepressant and anxiolytic effects thought to be independent of antidopaminergic activity (Keck et al. 1998; Tandon, 2000; Wilner et al. 2002). Despite such promise, a recent RCT with ziprasidone was negative (Pascual et al. 2008). Thus, evidence-based practice supports use of aripiprazole but not ziprasidone in treating BPD.

In meta-analyses, the class of antipsychotics had moderate effect in treating aggression, but no significant effect on depression, although aripiprazole and olanzapine may be exceptions (Lieb et al. 2010; Mercer et al. 2009). For aypicals, metabolic side-effects may limit clinical utility. Because 29–53% of borderline patients fulfil criteria for an eating disorder at some point in their lives (Lieb et al. 2004), and a significant number suffer from obesity (Frankenburg & Zanarini, 2006), iatrogenic metabolic risks must be regarded as serious. Although evidence-based practices have advanced in treating aggression associated with BPD; chronic emptiness, affective lability, and interpersonal dysfunction lack effective, evidence-based medication treatments.

Mood stabilizers and anticonvulsants

Due to BPD patients' affective dysregulation and comorbidity with bipolar disorder, some have classified BPD within the bipolar spectrum (Akiskal, 2004; Smith et al. 2004), although most continue to distinguish between the two, particularly with regard to interpersonal dysfunction (Bolton & Gunderson, 1996; Henry et al. 2001; Paris, 2004). Mood stabilizers are indeed becoming a more integral component of evidence-based treatment practices for BPD.

Lithium is beneficial in treating BPD, particularly in terms of quieting affective instability (Links et al. 1990; Rifkin et al. 1972). Lithium toxicity and/or non-compliance may be problematic, due to BPD patients' characteristic impulsive, self-destructive behaviour.

However, anticonvulsants are more often recommended for treatment of rapid-cycling bipolar disorder, the variant most closely resembling BPD. Carbamazepine demonstrated ‘dramatic’ reductions in behavioural dyscontrol and improvements in global functioning, anxiety, anger, euphoria, impulsivity, and suicidality, but it was associated with worsening melancholic depression (Gardner & Cowdry, 1986a, b), and therapeutic benefits could not be replicated in in-patients (de la Fuente & Lotstra, 1994).

Although high drop-out rates were reported with divalproex (Hollander et al. 2001), it subsequently demonstrated benefits on interpersonal sensitivity, anger, and aggression in euthymic borderline women with bipolar II (Frankenburg & Zanarini, 2002). Divalproex reduced aggression, irritability, and overall disease severity in patients with Cluster B personality disorders and prominent impulsive aggression (Hollander et al. 2003). Differential treatment response in Cluster B subjects was enhanced by baseline trait impulsivity and state aggression, although not affective instability (Hollander et al. 2005).

Lamotrigine extends periods of euthymia in bipolar patients (e.g. Goodwin et al. 2004). Potential benefits also include pro-cognitive activity, as previously demonstrated in normal volunteers (Aldenkamp et al. 2002). Lamotrigine effectively reduced BPD patients' anger (Tritt et al. 2005), and an 18-month follow-up demonstrated maintenance of this anti-aggressive effect (Leiberich et al. 2008). More recently, BPD patients without comorbid bipolar disorder but with prominent affective instability demonstrated reduced affective lability and impulsivity, but no change in other BPD symptoms, when treated with lamotrigine (Reich et al. 2009). Documented effects on impulsivity, anger, and affective lability in BPD thus make lamotrigine an attractive pharmacotherapeutic option. Nevertheless, the latter study reported higher rates of skin rash than reported elsewhere. Due to this life-threatening risk, clinicians should monitor patients closely and titrate the dose slowly.

Although topiramate's utility in bipolar disorder is controversial, it is efficacious for BPD. Topiramate reduced anger in female BPD subjects (Nickel et al. 2004), and a similar RCT reported this effect in males with BPD (Nickel et al. 2005). A separate RCT conducted with female BPD patients taking topiramate also demonstrated improvements in somatization, anxiety, health-related quality of life, overall stress, interpersonal sensitivity, hostility, and other facets of interpersonal functioning (Loew et al. 2006). Although no drop-outs were due to side-effects; cognitive impairment, reduced appetite, and weight loss were commonly reported.

Open-label follow-up studies to these initial RCTs demonstrated maintenance of therapeutic gains and additional weight loss associated with topiramate, and the authors therefore encouraged longer-term use (Loew & Nickel, 2008; Nickel & Loew, 2008). The authors admit that the patients studied were not the most severe. Because cognitive side-effects of topiramate may more adversely affect severely impulsive or suicidal BPD patients, a careful risk/benefit analysis should be undertaken before prescribing.

Overall, mood stabilizers and anticonvulsants are effective in treating BPD, particularly symptoms of impulsivity and aggression. As a class, they also demonstrate a moderate effect in treating depression in BPD (Ingenhoven et al. 2010; Mercer et al. 2009). Although they are an important component of evidence-based practice, patients should be closely monitored, because some may not tolerate these medications. The relatively slow titration schedules and the necessity of drawing plasma levels to reach an optimal dose may limit clinical effectiveness, particularly in a population often characterized by impulsive non-compliance. Although impulsivity and aggression appear to respond to treatment, there is little evidence of any effect from mood stabilizers in improving interpersonal dysfunction or disturbances of identity. Future research should focus more closely on these domains.

Other medications

Although there have been case reports of improvement in BPD patients treated with alprazolam (Faltus, 1984), the class of benzodiazepines has been associated with disinhibition, worsening impulsivity, suicidal ideation, and behavioural dyscontrol in BPD (Cowdry & Gardner, 1988). Benzodiazepines are vehemently discouraged, due to these risks, as well as elevated risks of dependence. Patients may abuse benzodiazepines to self-medicate intrapsychic pain, interfering with progress in psychotherapy and adversely affecting cognition.

The omega-3 fatty acid, ethyl-eicosapentaenoic acid (E-EPA) decreased aggression and depression in women with moderate to severe BPD (Zanarini & Frankenburg, 2003). A similar anti-aggressive effect was observed in two other RCTs with healthy subjects (Hamazaki et al. 1996, 2002). Omega-3 fatty acids may act by inhibiting protein kinase C, a mechanism thought to be involved in lithium and valproic acid pharmacotherapy (Peet & Stokes, 2005).

Clonidine, a presynaptic α2 noradrenergic agonist, has been studied in a trial comparing two doses given to BPD patients amidst states of ‘acute aversive inner tension’. Although tension, dissociative symptoms, self-injurious urges, and suicidal ideation decreased for both doses, there was no difference between the two doses (Philipsen et al. 2004a). Ziegenhorn et al. (2009) conducted an RCT of clonidine with BPD subjects with prominent symptoms of hyperarousal. Most of them therefore also met criteria for comorbid PTSD, which limited generalizability of findings. In the total sample, clonidine treatment improved hyperarousal, subjective quality of sleep, and anxiety, but not borderline-specific symptoms, and these benefits were not seen in the minuscule non-PTSD subsample. Although clonidine and similar agents have been efficacious in the treatment of PTSD (e.g. Southwick et al. 1999; Strawn & Geracioti, 2008), their role in treating BPD remains unclear.

A subset of borderline patients engage in self-injurious behaviour or more indirect forms of self-destructiveness (e.g. bulimia, substance abuse), which may reflect disturbances in endogenous opioids. Some BPD patients become disinhibited and aggressive after receiving opiate medications (Saper, 2000), and morphine administration increased self-injurious behaviour in one patient with BPD (Thurauf & Washeim, 2000). Naloxone used during acute states of aversive tension and dissociation in BPD demonstrated no significant benefit (Philipsen et al. 2004b). Naltrexone has been used successfully in open-label trials to treat self-harm (Griengl et al. 2001; McGee, 1997; Roth et al. 1996) and dissociation (Bohus et al. 1999). Therefore, evidence for treatment of BPD with medications acting upon opioid receptors remains inconsistent. Treatment with full agonists or antagonists may be complicated by differences between chronic, effects on post-synaptic receptor density on the one hand, and distinct, acute effects of receptor agonism or antagonism on the other (Prossin et al. 2010; Stanley & Siever, 2010). The potential for abuse of full opioid agonists may pose too great a risk for an effective treatment. For both these reasons, future trials with partial opioid agonists may be more effective in reducing self-injury, interpersonal dysfunction, and intrapsychic pain.

Avoidant personality disorder (AvPD)

AvPD is a common personality disorder (Loranger et al. 1994), existing as a comorbid condition in up to one-third of all patients with anxiety disorders (Alden et al. 2002). Up to 56% of AvPD patients continue to meet criteria after 2 years (Skodol et al. 2005). Nevertheless, distinguishing between this and generalized social phobia has been difficult, due to similarities in diagnostic criteria as well as frequently reported comorbidity. No neurobiological evidence indicates how the aetiology and psychopathology of AvPD differs from social phobia.

At present, clinicians should ‘extrapolate from data which are primarily related to anxiety disorders … to apply treatment strategies … that have primarily been developed for social phobia’ (Herpertz et al. 2007). Evidence-based treatment for AvPD would thereby include venlafaxine and SSRIs as first-line agents. A potential caveat is mentioned for sertraline if symptoms began in childhood or adolescence, in which case lesser efficacy was reported (van Ameringen et al. 2004). Gabapentin (Pande et al. 1999) and pregabalin (Pande et al. 2004) have also demonstrated efficacy in social phobia. Second-line agents would include reversible MAOIs brofaromine and moclobemide, for which there is presently less robust evidence, and the irreversible MAOI phenelzine, which entails risk of serious side-effects.

Other personality disorders, maladaptive traits

Pharmacological research is strikingly absent from other personality disorders. In these cases, medication is particularly indicated in the treatment of comorbid Axis I disorders, particularly mood and anxiety disorders that frequently co-occur with narcissistic, histrionic, and dependent personality disorders. SSRIs may be of particular clinical benefit relative to TCAs, given their more favourable side-effect profile and the possibility of an independent effect on personality factors (Ekselius & von Knorring, 1998; Reich et al. 2002). With the advent of the next DSM, a greater emphasis on a dimensional diagnostic approach to personality disorders will probably cast greater importance upon pharmacotherapeutic interventions targeting dimensions common to a variety of current Axis II diagnoses.

One such dimension of personality dysfunction is impulsive aggression. Recent trials with levetiracetam and oxcarbazepine for impulsive aggression recruited individuals with IED without significant comorbidity (e.g. Mattes, 2005, 2008), while other trials recruited subjects with personality disorders and a history of impulsive aggression (e.g. Coccaro et al. 2009; Hollander et al. 2003). Coccaro et al. (2009) found an anti-aggressive effect of fluoxetine in patients with IED and personality disorders. Several anticonvulsants (most notably divalproex, oxcarbazepine, and phenytoin) have also demonstrated evidence in treating impulsive aggression across diagnoses (Huband et al. 2010).

Future directions

Although the past two decades of research have ushered a paradigm shift in personality disorders, most research has been limited to BPD and SPD. Future research should be directed towards the treatment of other Axis II diagnoses and dimensions of dysfunction across diagnoses. Although research has made great strides towards understanding impulsivity and aggression, similar neurobiological substrates should be sought for other dimensions of personality. Only by clarifying these gaps in the evidence base can clinicians anticipate more effective evidence-based psychopharmacological practices for the treatment of personality disorders.

Further efforts to understand to what extent AvPD differs from generalized social phobia are warranted. This may require understanding distinctions between these diagnoses in neurobiology of fear and social inhibition, and in the developmental trajectory of each disorder. For SPD, the effects of pro-cognitive interventions should be evaluated with respect to social isolation and overall functioning. More comprehensive efforts are needed to understand the underlying neurobiology of SPD to improve evidence-based practices. Further characterization of the interpersonal dysfunction and cognitive, sensory-gating abnormalities seen in SPD will probably improve the effect of treatment on general functioning. Efforts at understanding the neurobiology of schizophrenia and its prodrome will also assist in defining targets for pharmacotherapy. Clarifying the respective roles of genes and environment in shaping the course of the schizophrenia spectrum will also uncover future pharmacotherapeutic targets.

Within BPD, research has detailed more extensive evidence-based practices for treating impulsive aggression. Anticonvulsants and atypical antipsychotics are acquiring more prominent roles in the treatment of BPD, relative to SSRIs. Nevertheless, greater serotonergic specificity will probably improve the efficacy of treatments. For example, selective 5-HT2A antagonism, but not 5-HT2C antagonism, has been shown to decrease impulsivity (Higgins et al. 2003; Winstanley et al. 2004).

Future research will also focus on treating affective instability, intrapsychic pain, dissociation, and interpersonal dysfunction associated with BPD. BPD has been conceptualized as related to disturbed attachment (Fonagy & Luyten, 2009) and dysfunctional representations of self and other (Bender & Skodol, 2007), with other symptoms seen as sequelae to this core feature. Oxytocin, vasopressin, and opioids may therefore be of particular relevance for treating BPD (Stanley & Siever, 2010), given the developmental role of these neuropeptides in attachment and the relationship between attachment security and stable social cognitive representations of self and other (Fonagy & Luyten, 2009). Although these domains have been exclusively treated with psychotherapy, research in the neurobiology of affiliative behaviour (e.g. Depue & Morrone-Strupinsky, 2005) as well as self-injury and dissociation (Mauchnik & Schmahl, 2010) may eventually provide novel pharmacotherapeutic targets.

For all personality disorders, integrating psychopharmacology with neurobiological effects of psychotherapy may produce synergistic and long-lasting benefits. Evidence-based practice continues to recommend an approach that includes both psychotherapy and pharmacotherapy. Although experienced therapists' contributions to personality theory and empirical research often continue to be at odds with one another, future research should attempt to connect theory with empirically-validated psychopharmacological targets. By understanding the neurobiology underlying increasingly complex behavior, pharmacotherapy can be optimized and targeted to personality dimensions previously considered susceptible only to psychotherapy.



Statement of Interest



View Abstract