The optimal strategy for the treatment of behavioral complications in patients with probable Alzheimer's disease (AD) remains unclear.

The objective of this study is to evaluate the risk of relapse following discontinuation of haloperidol in patients with Alzheimer's disease (AD) with psychosis or agitation who respond to it.

In Phase A of this study, AD outpatients with behavioral complications receive 20 weeks of open haloperidol treatment with an oral dose of 1-5 mg daily, titrated individually to achieve the optimal trade-off between efficacy and side effects. Responders to Phase A participate in Phase B, a 24-week continuation trial in which patients are randomized to continuation haloperidol or placebo.

The primary outcome is the time to relapse of psychosis or behavioral disturbance.

The study involves two phases. Outpatients with AD who meet inclusion/ exclusion criteria enter Phase A, the 20 week open acute treatment phase that uses a flexible dose regimen of haloperidol 1-5 mg daily. Haloperidol is started at an oral dose of 1 mg daily, with subsequent dose titration in 1 mg increments until the optimal dose is reached, i.e., optimal trade-off between efficacy and side effects. At the end of Phase A, patients who do not meet criteria for clinical response exit the protocol and is treated openly with alternative medications.

Phase A responders enter Phase B, a 24-week random assignment, placebo-controlled, continuation trial. Randomization is stratified by the severity of dementia and by the presence of psychosis. Half the patients are randomized to haloperidol (continuing at the same dose as at the end of Phase A), and the other half are randomized to placebo. Patients who relapse during Phase B exit the protocol and receive open treatment.

In Phase A, patients are followed at 0, 2, 4 weeks and every 4 weeks thereafter until 20 weeks. In the discontinuation trial, Phase B, patients are followed at 0, 1, 2, 4, week time points and every 4 weeks thereafter until 24 weeks. If a patient shows signs of relapse, the patient is brought in for more frequent visits, regardless of the stage of the protocol.

Inclusion Criteria:

• Meets DSM-IV criteria for dementia either sex, age 50-95 years
• Meets NINCDS-ADRDA criteria for probable Alzheimer's disease
• Meets Folstein Mini-Mental State Exam score of 5-26, inclusive
• Intellectual impairment reported for at least six months
• Availability of family member who has had direct contact with the patient for an average of at least once every week during the three months prior to study entry
• Has current symptoms of psychosis or agitation. Criteria for "psychosis" requires the presence of delusions and/or hallucinations identified by the Columbia University Scale for Psychopathology in Alzheimer's Disease (CUSPAD) and a minimum Brief Psychiatric Rating Scale (BPRS) psychosis factor score of at least 4 (moderate severity) on one of the following two items: These two items comprise the psychosis factor, excluding the item for conceptual disorganization. Agitation is defined as a score of greater than 3 (present at least 10 days per month) on one or more of the CERAD Behavioral Rating Scale for Dementia items for agitation, purposeless wandering, verbal aggression or physical aggression.
• Free of psychotropic medication for at least two weeks prior to study entry, or able to tolerate medication washout for this period.
• Informed consent by patient and family member, as per IRB procedures at New York State Psychiatric Institute.

Exclusion Criteria:

• Acute unstable medical condition, delirium, alcohol or substance abuse or dependence within the past 1 year
• Clinical evidence of stroke, other dementias including vascular or Lewy body or frontotemporal dementia, multiple sclerosis, Parkinson's disease, Huntington's disease, tardive dyskinesia
• Diagnosis of a psychotic disorder antedating the onset of dementia
• Antipsychotic medication usage during 4 weeks prior to study entry
• Contraindication to the use of haloperidol