Background: Studies of pharmacotherapy for agitation in dementia have primarily been limited to single-drug trials and have not determined if some forms of agitated behaviors are more responsive to treatment than others.
Objective: The goal of this study was to determine if manifestations of agitation (ie, physical aggression, physically nonaggressive behavior, and verbally agitated behaviors) show different degrees of response to divalproex sodium (extended release or sprinkles) alone or in combination with second-generation antipsychotic agents.
Methods: This was a 6-week, open-label, naturalistic pilot study of patients aged >60 years recruited from a geriatric psychiatry inpatient unit, 2 nursing homes, and 4 assisted living residences. Patients were referred for study if they displayed Behavioral and Psychological Symptoms of Dementia. The primary outcome measure was the Cohen-Mansfield Agitation Inventory (CMAI), an instrument whose subscales allow measurement of physically aggressive behaviors, physically nonaggressive behaviors (eg, wandering), and verbally agitated behaviors. The Neuropsychiatric Inventory-Nursing Home version (NPI-NH) was also used to assess patients' behavior.
Results: Fifteen patients were included in the study (10 men, 5 women; mean [SD] age, 81.9 [7.7] years). Total CMAI improved by 17.1% at week 1, showed an additional improvement of 3.4% at week 3, and no further improvement by week 6 (total score dropped by 1.1%). Aggressive behavior improved at weeks 3 and 6 while physically nonaggressive and verbally agitated behaviors improved at weeks 1 and 3, and lost much of the gains by week 6. The NPI-NH agitation/aggression score decreased a mean (SE) of 1.3 (0.5) points (P = 0.03), the irritability/lability subscale decreased a mean (SE) of 2.3 (0.6) points (P = 0.005), and the disinhibition subscale decreased a mean (SE) of 1.4 (0.4) points (P < 0.01). Seven patients were on divalproex monotherapy and 8 patients were on combination therapy. Daily doses of divalproex (mean dose, 656 mg/d) in combination with a second-generation antipsychotic were 28% lower than divalproex monotherapy (mean dose, 914 mg/d). The most common adverse events were somnolence (7 of 15) and gait disturbance (5 of 15). One death was observed in the study; the death was due to a prior medical condition and was judged unlikely to be study related.
Conclusions: Patients with higher levels of agitation receiving divalproex had reduced agitation on the physical aggression subscale of the CMAI. Divalproex was less effective on physically nonaggressive behavior and verbal agitation. Irritability, as measured on the NPI-NH, was also reduced. Patients who received both divalproex and an antipsychotic agent were responsive at lower doses of divalproex. In either case, the effective dosage of divalproex was lower than that commonly used for epilepsy or mania in elderly patients. The most common adverse events included somnolence and gait disturbance.