To the Editor: Kirk D. Denicoff, M.D., and colleagues presented interesting findings from their clinical trials of mood stabilizers in bipolar disorder patients (1). Clarification of important aspects of their protocol would greatly increase the usefulness of their data.

First, the title "Valproate Prophylaxis in a Prospective Clinical Trial of Refractory Bipolar Disorder" is imprecise. This title suggests valproate monotherapy, but 17 of the 18 patients were receiving combinations of mood stabilizers. The possibility that some of these 17 patients may have responded to valproate monotherapy was not considered in the article.

Furthermore, the authors reported that one patient who responded to valproate plus lithium experienced a manic relapse when the lithium was tapered off. They did not include data regarding how quickly this was done. It has been suggested that tapering off lithium too quickly may precipitate relapse (2).

The authors also did not include data regarding serum blood levels of these medications. Consideration of blood levels is essential for several reasons. One cannot conclude that the preceding trials of lithium and carbamazepine were ineffective unless serum levels were adequate. The same is true for the trials involving mood stabilizers in combination. Also, when carbamazepine is used with valproate, it is imperative to carefully monitor levels, since carbamazepine induces both its own metabolism and that of valproate (3). Furthermore, valproate can elevate carbamazepine levels (4).

Additionally, data were lacking regarding whether mood stabilizers in combination resulted in greater adverse reactions. There have been many reports of side effects when lithium, carbamazepine, and valproate in various combinations are used. It would also be important to know if bipolar patients on combinations of mood stabilizers were able to achieve good responses with lower blood levels of the drugs than are needed in monotherapy.

In conclusion, Dr. Denicoff and colleagues should be commended for investing much time and effort in this study. Their study is potentially quite significant for the clinical treatment of bipolar disorder, since more prospective studies of mood stabilizers in combination are needed. I look forward to subsequent and more comprehensive reports of their work.