Behavioral disorders in dementia are common and are the most important symp
toms with regard to socio-economic burden. Up to now there is no common int
ernational agreement of how to define and measure these disorders. Antideme
ntia trials focus mainly on cognition. Investigations of neurobiological co
rrolaries of disturbed behavior in the dementias are rare. The same holds t
rue for studies on the longitudinal course of behavioral disorders and thei
r interrelation. Many symptoms may be the expression of variable conditions
, e.g., agitation may be related to anxiety or akathisia. In primary care,
hospitals and nursing homes, antipsychotics are most often chosen for their
treatment. The available data demonstrate at least a modest efficacy. New
neuroleptics (risperidone, clozapine, olanzapine) offer some advantages wit
h regard to the risk benefit ratio. Benzodiazepines are frequently prescrib
ed, but seem to be superior to neuroleptics only for the treatment of sleep
disorders. Antidepressants, carbamazepine or valproic acid offer some bene
fits, but do not provide immediate effects, which may the reason why they a
re used much less. For long-term treatment of many behavioral symptoms, the
y may however be superior. Drugs should also be chosen with regard to demen
tia etiology. For example, physicians should consider the high neuroleptic
sensitivity in dementia of Lewy body type and the anticholinergic sensitivi
ty in dementia of Alzheimer type. Empirical evidence indicates overtreatmen
t of the demented population with sedating psychotropic drugs. With regard
to the instability of behavioral disorders in the time course the necessity
of drug treatment should always be (re)evaluated.