Retrospective analysis of serum valproate levels and need for an antidepressant drug.

We sought to determine whether patients receiving valproate plus an antidepressant had significantly lower serum valproate levels before initiation of the antidepressant than those patients receiving valproate without an antidepressant. We further sought to identify the prevalence of antidepressant-induced mania and to determine if valproate provided a protective effect against antidepressant-induced mania. A computer database search from January 1, 1990-June 30, 1998, identified patients with bipolar or schizoaffective disorder treated with valproate. Patients receiving an antidepressant during valproate therapy were identified as the treatment group (9 patients), and the remaining patients served as the control group (17 patients). Serum valproate levels were recorded just before starting the antidepressant for the treatment group and monthly during a comparable period for the control group. The mean time to antidepressant initiation was 15 +/- 8 weeks. The mean serum valproate level just before antidepressant initiation was significantly lower for the treatment group compared with the mean serum valproate level averaged over 16 +/- 6 weeks for the control group (54 +/- 24 vs 73 +/- 13 microg/ml, p<0.05). Four patients (44%) developed antidepressant-induced mania. Three required discontinuation of the antidepressant; their serum valproate levels were 54, 60, and 71 microg/ml. Patients requiring the addition of an antidepressant had significantly lower valproate serum levels than those who did not require an antidepressant. Further study is necessary to determine whether higher serum valproate levels are needed for prevention of depressive symptoms in bipolar and schizoaffective disorders.