Background In order to limit neuroleptic use in the elderly, because of cardiovascular events, specialists in charge of behaviour disorders don’t have many therapeutic options in cognitive-behavioural units (CBU).

Purpose Valproic acid (VPA) is an anticonvulsant and/or a mood stabiliser that can be used in a behavioural way in CBU. One side effect of VPA is hyperammonaemia, which can lead to sedation and changes in behaviour or personality.

Materials and Methods Inclusion criteria were opposition, agitation, aggressiveness or impulsiveness. Ammoniemia levels were assessed before starting the VPA, between 2 and 4 days and after 5 days with VPA. For each person included, Cockroft’s creatinine clearance, medical background and neuroleptic co-prescriptions were identified. Results are presented with mean±SEM.

Results The population was defined by an average age of 79.3 y ±1.74, a sex ratio of 15 men for 6 women; a creatinine clearance of 65.4 mL/min ± 8.9, no patients had liver troubles or a history of epilepsy. 21 patients received VPA in the CBU, for at least one of the following indications: opposition (n = 9), agitation (n = 13), aggressiveness (n = 16) or impulsiveness (n = 6). 9/21 patients came out of the CBU with VPA (42.85%), 13/21 without VPA (61.9%), 5/21 with a neuroleptic (23.8%) and 8/21 without VPA or a neuroleptic (38.1%). Ammoniemia rates at D-1, between D2 and D4 and after D5 were respectively 47.47 µM ± 3.71, 51.4 µM ± 6.43 and 63.76 µM ± 4.95. Response rate to VPA was 55% (5/9 patients) for opposition, 37.5% (6/16) for aggressiveness, 38% (5/13) for agitation and 66.6% (4/6) for impulsiveness.

Conclusions Those results show that only one of every two patients with VPA were responders, and average ammoniemia increases during treatment. However, 100% of patients going out with VPA didn’t have any neuroleptics and for 33%, VPA contributed to stopping neuroleptics.

No conflict of interest.