Haloperidol for agitation in dementia

Agitation includes wandering, crying out, abusive vocalization, and assaultive behavior and occurs in up to 70% of patients with dementia. Although the neuroleptic haloperidol has been used for decades to control disruptive behavior in psychotic and demented patients, the effectiveness of this drug for agitated dementia remains in question. The first meta-analysis on the effectiveness of haloperidol for agitated dementia, published in 1990, was limited in scope and was unable to provide clear guidelines for the use of haloperidol for demented patients who are agitated. Meta-analyses in 1998 and 2000 examined haloperidol compared with other neuroleptics as well as with placebo and omitted a number of databases, including non-English language publications. To determine the effect of haloperidol, compared with placebo, in the control of agitated dementia and to make recommendations for future research in this area a more widely based, yet more highly focussed review was carried out.

The main objective was to determine whether evidence supports the use of haloperidol to treat agitation in demented patients.

The CDCIG Specialized Register was searched to identify all available reports on haloperidol treatment of agitated dementia.

We examined randomized, placebo-controlled trials, with concealed allocation, where subjects' dementia and agitation were assessed. Trials involving treatment of less than one week were not included.

1. Two reviewers extracted data from included trials. 2. Data were pooled, where possible, and analysed using appropriate statistical methods. 3. Odds ratios or average differences were calculated. 4. Only 'intention to treat' data were included. Where a cross-over design was employed (Devanand, 1998), only the initial phase of the study was used to compare haloperidol versus placebo. 5. Sensitivity analysis was applied to heterogeneity of results and to gauge the effect of the included studies of small sample size. 6. In addition to the overall meta-analysis, individual analyses of the reports were carried out to examine the effect of degree of dementia, dose of haloperidol, and duration of therapy on agitated dementia. Analysis included the following groups: All patients treated with haloperidol compared with placebo.

There were five included trials. All studies stated "intention to treat" analysis of their results. Three studies were from the United States, and two studies were from Europe. Two studies examined patients with various forms of dementia, and three studies included only patients with diagnosed Alzheimer's dementia. 1. Overall meta-analysis of the response of agitated patients to haloperidol, compared with controls, showed no improvement in agitation. There is some evidence that haloperidol helps to control aggression. Adverse reactions and dropouts were more frequent among haloperidol treated patients, compared with controls. This meta-analysis provided no information about the relationship between the degree of dementia, the kind of agitation manifested, or the dosage and duration of therapy with haloperidol and response to treatment of demented patients with agitation. 2. The results of this meta-analysis were too broad to permit specific recommendations for treatment of agitated dementia with haloperidol. 3. Higher dose haloperidol, or prolonged haloperidol (12 weeks compared with 3 - 6 weeks) was associated with increased side effects, largely related to Parkinsonian symptoms of rigidity and bradykinesia.

1. Haloperidol appeared to provide no improvement in agitation among demented patients compared with placebo, but side effects were frequent. 2. Dropout rates were higher for haloperidol compared with placebo treated patients, suggesting that side effects led to discontinuation of treatment in some patients. 3. Because of the wide focus of this meta-analysis, not enough information was provided to permit recommendations linking haloperidol treatment of agitated dementia to degree of dementia, manifestations of agitation, or dosage and duration of treatment of haloperidol. 4. Individual analysis of reports indicated that higher dose haloperidol (more than 2 mg per day) may have been more effective than lower dose haloperidol (less than 2 mg per day) in controlling aggression, but not other manifestations of agitation, among patients with mild to moderate dementia. 5. Similar analysis suggested that prolonged therapy with haloperidol (more than 3 - 6 wks) or higher dosage (more than 2 mg per day) was more likely to result in side effects than were short term therapy (3 weeks) or lower dose haloperidol (less than 2 mg per day). 6. The reports provided too little information to permit interpretation of the effect of degree or type of dementia on response to haloperidol. Except for a favorable response of aggression to haloperidol, no other manifestations of agitated dementia were found to have improved following therapy with haloperidol, compared with controls.