The Use of Antiepileptic Drugs in Psychiatry

Valproic Acid

The acute anti-manic efficacy of valproic acid has been established by several controlled studies, as these randomized and placebo-controlled studies have found divalproex sodium to be an effective and safe treatment for a manic episode, even in cases where lithium treatment previously failed. In comparison with lithium, a randomized, open-label trial of 300 patients with acute mania showed no difference between the two medications in mania remission or tolerability at 12 weeks. A meta-analysis of multiple different anti-manic treatments showed that valproate was slightly (but not significantly) less efficacious than lithium, and slightly better tolerated (although as a class, antipsychotics were overall more effective than anticonvulsants and lithium in treating acute mania). The Food and Drug Administration (FDA) has approved divalproex sodium as a treatment for acute mania. There also appears to be a linear relationship between valproate serum concentration and symptom response, and the target dosage with optimal response in acute mania is above 94 μg/ml.

Valproate is not approved by the FDA for maintenance treatment of BPD; in fact, there is growing evidence in disfavor of such a role. The largest trial of valproate for BPD maintenance to date (n = 372) failed to find a difference between valproate, lithium, and placebo for the primary outcome measure (time to any mood episode) for up to 1 year. More recently, a large randomized, open-label trial (the BALANCE study; n = 330) compared valproate and lithium monotherapies in BPD maintenance with combination lithium and valproate, for up to 2 years. This innovative study examined physician intervention for a mood episode as the primary outcome, and utilized an initial run-in period where patients were stabilized on both medications before randomization to different maintenance arms, which ensured patient tolerance to either medication. The study found that patients on combination lithium and valproate were significantly less likely to relapse compared to patients on valproate alone, and were as equally likely to relapse as patients on lithium alone. The study was not designed to compare valproate and lithium monotherapies directly, although it appeared that lithium monotherapy was more effective than valproate alone. Thus valproate monotherapy is not recommended for maintenance therapy of BPD. For patients on valproate alone, the study indicates there is a potential benefit of adding lithium for maintenance, although the question remains whether patients should simply be on lithium alone in the first place. It is intriguing that the same investigators who demonstrated a role for divalproex sodium in acute mania treatment conducted a second, randomized double-blind study of the same drug in the same setting, but subsequently found no difference between divalproex sodium and placebo in acute mania. The investigators attributed this discrepancy to methodological differences in study design (lower drug dose, allowance of early study termination, more liberal use of adjunctive medications, and 2 : 1 randomization in favor of study drug); however, in the context of other well-established options for acute mania treatment and BPD maintenance therapy, one wonders whether overall use of valproate for any phase of BPD should be limited.

Anticonvulsants, including valproate, are often viewed as more effective in rapid cycling than is lithium. However, it appears that this may not be the case. In a rigorous and ambitious double-blind study, rapid-cycling patients were stabilized on open-label lithium and divalproex sodium and then randomized in a double-blind fashion to either lithium or divalproex and followed prospectively. There were no significant differences between the lithium-treated or the valproate-treated groups in time to drop-out or time to additional psychopharmacology.

Lastly, there is inadequate data to support a role for valproate in acute bipolar depression. Two independent meta-analyses of four randomized placebo-controlled trials suggested possible efficacy in acute depressive symptoms of BPD; however, the study sizes were very small (n = 9–28). These findings will need to be replicated before valproate can be recommended for this indication.

Lamotrigine

Lamotrigine represents a significant advance in the long-term management of bipolar depression, especially given the prominent burden of depression and depressive relapses in BPD. Lamotrigine is approved for the maintenance treatment of BPD, and is efficacious when compared to placebo in maintenance studies. Long-term studies found an overall reduction in bipolar depressive relapse compared with placebo. In a key study in which patients who were most recently depressed were first stabilized on lamotrigine and randomly assigned to maintenance treatment with lamotrigine, lithium, or placebo, the overall sustained response rate was 57% with lamotrigine, compared with 46% for lithium, and 45% for placebo; this indicates that lamotrigine is effective in the prevention of relapse to depression when compared with placebo. It is important to note, however, that a substantial proportion (43%) of bipolar patients remained unprotected against a depression relapse by continuation lamotrigine.

The evidence for efficacy of lamotrigine in the treatment of acute bipolar depression, on the other hand, remains limited. No single trial of the drug for acute bipolar depression found the drug better than placebo on the primary outcome measure of the study. One widely referenced study found that while lamotrigine was not statistically different from placebo for total Hamilton Depression Rating Scale (HDRS) scores, it was superior on several other measures, including the Montgomery-Asberg Depression Rating Scale (MADRS) and the Clinical Global Impression Improvement (CGI-I) scale.

This effect has never been replicated in individual trials, but a meta-analysis and meta-regression of pooled individual data from all five lamotrigine trials in acute bipolar depression (both bipolar I and II disorders) found that response (defined as a ≥ 50% decrease in scores) on both the HDRS and the MADRS was significantly greater than for placebo. The effect size was small, however, and the number-needed-to-treat was about 11; a finding the authors note is at the “margins” of clinical utility. Remission was greater than placebo on the MADRS, but not the HDRS. The antidepressant effect of lamotrigine was greatest in the most severely depressed subjects in a subgroup analysis. These data suggest that any potential benefit of lamotrigine in acute depression is likely to be small, except perhaps in severely depressed patients. It is notable that the study of lamotrigine in maintenance of bipolar depression was designed to follow patients after they were first stabilized on lamotrigine for an acute depressive episode ; taken together, these studies suggest that lamotrigine should be continued for maintenance therapy in patients who responded to lamotrigine during acute bipolar depression, and that the patients most likely to respond are those who are most severely depressed.

A small randomized, double-blind study was performed comparing lamotrigine versus placebo as add-on medications to lithium in acutely depressed bipolar patients, and showed that lamotrigine plus lithium significantly improved MADRS scores at the end of week 8 versus lithium alone. In order to be eligible for the study, however, patients had to be depressed despite taking a therapeutic dose of lithium for at least 2 weeks (most had been taking for at least 3 months), suggesting that this study population was likely enriched for non-responders to lithium monotherapy. The study also consisted of only 124 patients, and needs to be replicated. Subsequent follow-up of these patients showed that lamotrigine plus lithium was approximately as effective as lithium alone for preventing mood relapse or recurrence up to 68 weeks; however, the small sample size and the study design precluded formal statistical analyses. At this point, evidence for any additional benefit of lamotrigine–lithium combination therapy in treating BPD, versus either of these medications alone, remains limited.

Lamotrigine has not demonstrated efficacy for the acute treatment of mania. Multiple treatments meta-analysis of anti-manic drugs also indicated that lamotrigine is not more effective than placebo in treating acute mania. No single trial found benefit for lamotrigine in the prevention of manic episodes; however, pooled analysis revealed a small but significant effect size for the prevention of mania. While lamotrigine is more effective than lithium in the prevention of depressive episodes, lithium appears to be more effective than lamotrigine in preventing manic episodes.

No validated treatments for treatment-refractory bipolar depression exist, but a small, randomized, open-label trial of adjunctive lamotrigine, risperidone, or inositol in subjects who had depression in spite of trials of two consecutive standard antidepressants of adequate dose and duration was conducted as part of the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD). An equipoise randomization process allowed subjects to choose to be randomized to any pair of the study treatments. While no differences were found in primary pair-wise comparisons in randomized patients (n = 66), a secondary, post-hoc analysis found that 8 weeks of sustained recovery were seen in 23.8% of the lamotrigine-treated patients, 17.4% of inositol-treated patients, and 4.6% of risperidone-treated patients. These data must be viewed cautiously, as they are from a secondary analysis, but they represent one of few studies for the treatment of refractory bipolar depression.

Carbamazepine

Carbamazepine was the first anticonvulsant studied as a treatment of mania. More than 19 studies (most of which were small case series or open trials) evaluated carbamazepine for the treatment of mania, and until recently it was used in BPD despite little scientific support for its use. More recently, carbamazepine (in extended-release form) was found to be effective for acute mania in two large placebo-controlled trials. Multiple treatments meta-analysis of anti-manic drugs showed that carbamazepine was similar to valproate in terms of efficacy and acceptability in treating acute mania; however, all anticonvulsants as a class were outperformed by atypical antipsychotics. There are no data directly establishing carbamazepine as an effective maintenance treatment in BPD. A meta-analysis of four small studies comparing efficacy of carbamazepine versus lithium in BPD maintenance suggested a possible similarity in relapse rates; however, this was tempered by the finding that carbamazepine use was associated with significantly more study withdrawals due to adverse effects. At present, there are insufficient data to suggest that carbamazepine is more effective in these patients than any other treatment.

Valproic Acid

Valproic acid (di-n-propylacetic acid) is an anticonvulsant drug chemically unrelated to other psychiatric medications. One of the more commonly used preparations is divalproex sodium (Depakote), a compound of sodium valproate and valproic acid in a 1 : 1 molar ratio. Valproate is available as tablets (both delayed- and extended-release), capsules, enteric-coated capsules, sprinkles, and syrup. Absorption is different across the different preparations and it is delayed by ingestion of food. This may be of some importance when one switches from one preparation to another. Peak plasma levels are achieved between 2 and 4 hours after ingestion of the direct-release preparation, and the half-life ranges from 6 to 16 hours. More than 90% of plasma valproic acid is protein-bound. The time of dosing is determined by possible side effects, and, if tolerated, a once-a-day dosing could be employed. The therapeutic plasma levels generally used for the treatment of mania are the same as those used for anticonvulsant therapy (50–100 μg/ml), and the total daily dosage required to achieve these levels ranges from 500 mg to greater than 1,500 mg, although one study suggests a direct relationship between plasma valproate levels and response in acute mania, suggesting optimal levels in acute treatment of greater than 90 μg/ml.

Valproic acid is metabolized by the hepatic CYP 2D6 system but, unlike carbamazepine, does not auto-induce its own metabolism. Concomitant administration of carbamazepine will decrease plasma levels of valproic acid, and drugs that inhibit the CYP system (e.g., selective serotonin reuptake inhibitors [SSRIs]) can cause an increase in valproic acid levels. Valproate is known to increase the plasma levels of lamotrigine, so it is recommended that the lamotrigine dose be lowered in patients taking valproate. Dose-related and common initial side effects include nausea, tremor, and lethargy. Gastric irritation and nausea can be reduced by dividing the dose or by using enteric-coated preparations. Valproic acid has been associated with potentially fatal hepatic failure, usually occurring within the first 6 months of treatment and most frequently occurring in children under age 2 years and in persons with pre-existing liver disease. Transient, dose-related elevations in liver enzymes can occur in up to 44% of patients. Any change in hepatic function should be followed closely, and patients should be warned to report symptoms of hepatic failure (such as malaise, weakness, lethargy, edema, anorexia, or vomiting). Multiple cases of valproate-associated pancreatitis have also been reported, as has multi-organ failure. These can occur at any point during treatment.

Valproic acid may produce teratogenic effects, including spina bifida (1%) and other neural tube defects. Other potential side effects include weight gain, inhibition of platelet aggregation, hair loss, and severe dermatological reactions (such as Stevens-Johnson syndrome).

There is additional worry that valproate may cause endocrine abnormalities in women. Two hundred and thirty women were evaluated for polycystic ovarian syndrome (PCOS) as part of an ancillary study during the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD). Criteria for PCOS are met when oligomenorrhea (defined as ≤ 9 cycles in the past year) coincides with at least one feature of hyperandrogenism (including hirsutism, acne, male-pattern alopecia, or elevated serum androgen levels); it can ultimately result in an increased risk of type 2 diabetes mellitus, cardiovascular disease, and some types of cancer. Joffe and associates compared the rate of new-onset PCOS in women with BPD taking valproate compared to the rate in those taking other anticonvulsants and lithium. Nine (10.5%) of the 86 valproate users developed treatment-emergent oligomenorrhea with hyperandrogenism, compared to 2 (1.4%) of the 144 valproate non-users. The relative risk for developing PCOS on valproate versus on non-valproate mood stabilizers was 7.5 (95% confidence interval, 1.7 to 34.1, p = 0.02). The onset of oligomenorrhea usually began within 12 months of the beginning of valproate treatment. A later analysis found that discontinuation of valproate in women with valproate-associated PCOS may result in an improvement of the PCOS reproductive features, as these symptoms resolved in three of the four women who discontinued valproate, but persisted in all three women who continued on valproate.

Because of the risk of PCOS in women of child-bearing age who are exposed to valproate-containing products, the use of valproate as a first-line treatment for BPD is not recommended in these patients.