OBJECTIVE:

To investigate the contribution of neuroleptic drugs to cognitive decline in dementia.

DESIGN:

Two year prospective, longitudinal study consisting of interviews every four months, with necropsy follow up.

SETTING:

Community settings in Oxfordshire.

SUBJECTS:

71 subjects with dementia, initially living at home with informant.

MAIN OUTCOME MEASURES:

Cognitive function (score from expanded minimental state examination); behavioural problems (physical aggression, hallucinations, persecutory ideas, and disturbance of diurnal rhythm); and postmortem neuropathological assessment (cortical Lewy body pathology).

RESULTS:

The mean (SE) decline in cognitive score in the 16 patients who took neuroleptics was twice that in the patients who did not (20.7 (2.9) v 9.3 (1.3), P = 0.002). An increased rate of decline was also associated with aggression, disturbed diurnal rhythm, and persecutory ideas. However, only use of neuroleptics and severity of persecutory ideas were independently associated with more rapid cognitive decline when all other variables were adjusted for. The start of neuroleptic treatment coincided with more rapid cognitive decline: median rate of decline was 5 (interquartile range 8.5) points per year before treatment and 11 (12) points per year after treatment (P = 0.02). Cortical Lewy body pathology did not account for association between neuroleptic use and more rapid decline.

CONCLUSIONS:

Neuroleptic drugs that are sometimes used to treat behavioural complications of dementia may worsen already poor cognitive function. Randomised controlled trials are needed to confirm a causal relation.