Introduction: The management of behavioural disorders in the elderly is a reality that is difficult to implement. Few studies have focused on the therapeutic arsenal for managing these numerous conditions which will only increase in the years to come, in view of our ageing population. In cases where medical and non-medical strategies fail, valproic acid could provide an alternative based on studies that have highlighted its effects in the management of behavioural disorders. However, VA can cause sedation due to excessive ammonia levels in the blood, which may prove toxic for the central nervous system. The purpose of this study was to monitor any beneficial effects of VA in the treatment of behavioural disorders in a cognitive-behavioural unit (CBU), to measure ammonia levels in the blood after initiating treatment and, finally, to study combinations of psychotropic treatments with and without VA on patients’ discharge prescriptions. Patients and Methods: This retrospective study was carried out in a 13-bed cognitive behavioural unit over an 18-month period from November 2011 to May 2013. All of the patients who had received VA on at least one occasion were enrolled in the study. Ammonia levels in the blood were assayed by nursing staff. The clinical course was documented in the medical records and NPI scores (on admission and discharge) were collected. Finally, the discharge prescriptions issued by the CBU were analysed. Statistical tests were carried out using Graph Pad Prism 6. Results: Out of 35 patients with an average age of 80 +/-1 year, who received VA, 20 (57%) discharged from the CBU with this treatment benefited from it. The levels of ammonia recorded in the blood of these patients did not differ from those of patients who either did not respond or responded unfavourably to VA (treatment discontinued), namely 53.29 +/-4.7 mg/L and 63.50+/-6.7 mg/L (p=0.21), respectively. Blood ammonia levels do not appear to be implicated in the favourable clinical course of patients receiving VA. Finally, the inclusion of VA in discharge prescriptions allows a significant saving to be made in terms of neuroleptic agents (p=0.0017) and benzodiazepines (p= 0.0006), which are known to be iatrogenic in the elderly. Conclusion: valproic acid proved to be beneficial for 1 in 2 patients admitted to a cognitive behavioural unit since the discharge prescriptions contain this medicinal product. In fact, these results are consistent with doctors’ medical observations. However, the statistics quoted by the NPI do not highlight the effect of valproic acid. When included in discharge prescriptions, VA promotes savings in terms of psychotropic medication.