Review of: Missio G et al, Trials 2019;20(1):608

STUDY TYPE: Randomized, open-label controlled trial

It’s rare to see full recovery in bipolar I disorder with a single medicine, so we often depend on some combination of mood stabilizers and/or antipsychotics. Even then, weight gain and metabolic problems are deal-breakers. Some experts favor the combination of lithium (Li) and carbamazepine (CBZ), based in part on early reports that the two cancel out some of each other’s side effects (eg, neutropenia and hyponatremia from CBZ, diabetes insipidus from Li; see TCPR, Nov/Dec 2018). This study compared Li+CBZ with an alternative strategy, Li+valproic acid (Li+VPA), in bipolar I disorder.

The independently funded study recruited 64 young adults with bipolar I from a tertiary referral center in Brazil. Of the 60 who were treated, 27 were depressed, 17 were hypomanic or manic, and 16 had mixed states. They were randomized to receive Li+VPA (n = 33) or Li+CBZ (n = 27), with each drug titrated to its therapeutic serum level. Only sertraline (given to a third of the depressed patients) and lorazepam were allowed as add-ons. The main outcome measure was reduction in bipolar symptoms, but tolerability was also key, using standard rating scales for each.

All patients were followed for 2 months with weekly to biweekly assessments, and responders (≤ 50% reduction in symptoms) were followed for 2 years to track remission. Neither response nor remission rates differed significantly between the two combinations. There was a non-significant trend favoring Li+VPA against mania and Li+CBZ against depression. Where the therapies differed was in tolerability. Li+VPA had significantly more side effects than Li+CBZ, particularly fatigue, decreased libido, and weight gain. After 2 months, the Li+VPA group gained an average of 4.6 pounds, while the average Li+CBZ patient lost half a pound.

The results were weakened, however, by the small sample size and high dropout rate. Less than half of those who entered the study made it through the first 2 months. The dropout rates were similar for both groups. Most dropouts either didn’t get better or left because of side effects.

TCPR’s Take
A high dropout rate and small sample size make these results tentative, but they suggest that CBZ may be more patient friendly than VPA when adding a mood stabilizer to lithium. Although CBZ does not interact with lithium, its other drug interactions can complicate the picture.