Most authorities in the world of borderline personality disorder (BPD) say that psychotherapy is the mainstay of treatment and that medications should only be used adjunctively to treat symptoms as they arise. Nevertheless, medications are used relatively frequently for this disorder. The Journal of Affective Disorders recently published a useful review of studies looking at the role of medications in treating borderline personality disorder (Abraham PF and Calabrese JR, J Aff Disord 2008, early online access). We’ve organized the information by medication class, and, in some cases, provide references to the original studies. For unlisted references, please see the review paper.

Antidepressants

The surprise here is that SSRIs are not very effective in BPD. The authors identified six trials of SSRIs, five of which were of fluoxetine. The benefits were modest, and were limited primarily to the easing of impulsive aggression. Oddly, SSRIs didn’t help much with depressive symptoms overall, with the exception of one study which included many patients with SSRI - responsive conditions. One study tested amitriptyline (mean dose 149 mg/day) with Haldol (4.8 mg/day). While both medications were superior to placebo for depressive symptoms and for measures of anxiety/hostility, Haldol actually was superior to amitriptyline on many other measures, including that of global functioning. The winner among antidepressants for BPD appears to be the MAOIs. Both tranylcypromine (Parnate) and phenelzine (Nardil) were effective for a range of borderline symptoms.

Anticonvulsants

Two identically designed placebo controlled trials found that topiramate (Topamax), ramped up to a target dose of 250 mg/day, was more effective than placebo specifically for anger symptoms. Carbamazepine received an early endorsement from one of the original studies of meds for BPD (the famous Cowdry and Gardner study, Arch Gen Psychiatry 1988; 45:111-119), but a 1994 placebo-controlled double-blind study of 20 hospitalized patients revealed no positive effects at all. Another study enrolled 17 patients in an open-label trial of oxcarbazepine (Trileptal), and found it to be broadly effective at doses ranging from 1200 to 1500 mg/day (Bellino S et al., J Clin Psychiatry 2005 Sep;66(9):1111-5). Three studies have compared divalproex sodium (Depakote) with placebo. It has been most impressive in decreasing measures of anger, and in one study, interpersonal sensitivity. It doesn’t appear to be helpful for depressive symptoms. Finally, Lamictal, titrated up to 200 mg/day, was more effective than placebo for anger symptoms in a study of 24 women.

Antipsychotics

Olanzapine (Zyprexa) has the most positive data of any medication tested for BPD. The authors review five placebo-controlled trials, all of which showed benefits of olanzapine for aggression, impulsivity, and, though not as consistently, depression. Patients on olanzapine had a mean weight gain of 3 kg. Average doses were low, ranging from 3.3 mg/day to 8.8 mg/day. One concerning finding, however, was published after this review. In a study adding either olanzapine or placebo to dialectical behavior therapy (DBT) in women with BPD with high irritability, olanzapine was superior to placebo in decreasing irritability, but there was a tendency for the placebo group to have less self-inflicted injury (Linehan M et al., J Clin Psychiatry 2008;69:999-1005). The sample size was quite small (24 women), so it’s not clear how believable the findings are. More recently, an open-label study has shown broad symptom benefit of quetiapine, mean dose 412.5 mg/day (Van den Eynde, F et al., J Clin Psychopharm 2008;28:147-155), and a placebo-controlled study endorsed aripiprazole 15 mg/day for a variety of borderline symptoms (Nickel M et al., Am J Psychiatry 2006;163:833-838). However, ziprasidone (Geodon) appears not to share the magic of its atypical cousins. A just-published study randomly assigned 60 patients with BPD to either Geodon (mean dose 84.1 mg/day) or placebo, and there were no differences between the two treatments after 12 weeks (Pascal J et al., J Clin Psychiatry 2008;69 (4):603-8). It might be that the dose of Geodon was too low, or maybe it just doesn’t work in this condition.

Miscellaneous

In a surprising result, 30 women with BPD were randomized to either omega-3 fatty acids (EPA, 1 g/day) or placebo. EPA was superior to placebo at 8 weeks for both aggressive and depressive symptoms. The only placebo-controlled trial of alprazolam used a mean dose of 4.7 mg/day, and patients assigned to this did more poorly on ratings of behavioral dyscontrol.

Conclusions

Although it appears from this brief review that anticonvulsants and antipsychotics have some fairly good efficacy data, these studies enrolled small numbers of subjects and the drop-out rates were high. For example, only 112 patients in total were randomized to Zyprexa in the six placebo-controlled studies combined, and the N’s for other medication trials decrease from there. Furthermore, since there are many different symptoms to track in BPD, the statistics risk becoming fishing expeditions, and the more outcome variables you look at, the higher the chance that apparently "significant" findings arise by chance. In fact, the Cochrane Collaborative, famous for its very stringent methodology standards, reviewed the literature on medications for BPD in 2006 and concluded that the data are too weak to make any solid recommendations (Binks CA et al., Cochrane Database Syst Rev. 2006 Jan 25;(1):CD005653).

Nonetheless, patients come to us in distress and we have to make decisions. Low dose Zyprexa was helpful in several studies. Given Zyprexa’s side effect burden, we hope that other antipsychotics will end up as effective, but there is little data on these, and Geodon’s poor performance was disappointing. Depakote can target symptoms of impulsive aggression. Unfortunately, SSRIs did not perform as impressively as MAOIs, and MAOIs are generally not the first choice for patients with impulsive suicidality. Tiny studies have found both Topamax and Lamictal helpful for anger symptoms, and since they are both relatively benign medications (aside from rare Stevens Johnson syndrome on Lamictal) it seems inevitable that they will become more popular for these challenging patients.