Safety and tolerability of divalproex sodium in the treatment of signs and symptoms of mania in elderly patients with dementia: results of a double-blind, placebo-controlled trial

Abstract

Objective: The aim of this study was to assess the efficacy, tolerability, and safety of divalproex sodium in the treatment of elderly patients with dementia who exhibited manic symptoms. Methods: This 6-week, multicenter, randomized, double-blind, placebo-controlled, parallel-group study enrolled elderly nursing home residents with dementia and secondary mania, as distinct from the actual manic episodes that might be associated with bipolar disorder. Patients were randomly assigned to receive placebo or divalproex sodium at a starting dose of 125 mg BID, followed by daily titration in increments of 125 mg/d up to a target dosage of 20 mg/kg/d (maximum dosage of 30 mg/kg/d). Efficacy was assessed using the Bech-Rafaelsen Mania Scale (BRMS), the Cohen-Mansfield Agitation Inventory (CMAI), the Brief Psychiatric Rating Scale (BPRS), and the severity of illness and improvement scales of the Clinical Global Impressions (CGI). Because of a disproportionate number of withdrawals due to somnolence in the active-treatment group, the study was terminated prior to full enrollment. Results: A total of 172 patients were enrolled; 87 received divalproex sodium and 85 received placebo. Forty (46%) divalproex sodium-treated and 60 (71%) placebo-treated patients completed the study. There was no appreciable difference between drug and placebo groups in scores on the BRMS, the primary outcome measure. Scores on the CMAI, however, showed significant improvement in the divalproex sodium-treated group compared with the placebo group ( P < 0.05) for the total score and the verbal agitation subscale. There was no drug-placebo difference in scores on the BPRS, whereas the CGI improvement score at the final assessment was significantly better in the placebo group ( P = 0.035). Nineteen divalproex sodium—treated patients (22%) and 3 placebo-treated patients (4%) withdrew from the study because of adverse events ( P < 0.001), primarily somnolence; this typically occurred at or above daily dosages of ∼15 mg/kg. Lower doses were well tolerated. Somnolence and thrombocytopenia occurred significantly more frequently ( P = 0.027 for somnolence; P = 0.029 for thrombocytopenia) with divalproex sodium use than with placebo. Conclusions: Divalproex sodium did not improve signs and symptoms of mania associated with dementia in this sample of nursing-home residents, but did improve symptoms of agitation. Further study of divalproex sodium for agitation, at lower doses and a slower titration rate, is needed to define the optimal regimen for this drug in elderly, medically compromised populations.