Delirium occurs in up to 30% of hospitalised patients and is associated with prolonged hospital stay and increased morbidity and mortality. Recently published reports have suggested that the standard drug for delirium, haloperidol, a typical antipsychotic that may cause adverse extrapyramidal symptoms among patients, may be replaced by atypical antipsychotics such as risperidone, olanzapine or quetiapine, that are as effective as haloperidol in controlling delirium, but that have a lower incidence of extrapyramidal adverse effects.To compare the efficacy and incidence of adverse effects of haloperidol with risperidone, olanzapine, and quetiapine in the treatment of delirium.The trials were identified from a search of the Specialized Register of the Cochrane Dementia and Cognitive Improvement Group on 7 August 2006 using the search terms:haloperidol or haldol or risperidone or risperdal* or quetiapine or seroquel* or olanzapine or zyprexa* or aminotriazole or sertindole or leponex* or zeldox* or ziprasidone.Types of studies included unconfounded, randomised trials with concealed allocation of subjects. For inclusion trials had to have assessed patients pre- and post-treatment. Where cross-over studies are included, only data from the first part of the study were examined. Interrupted time series were excluded. Length of trial and number of measurements did not influence the selection of trials for study. Where indicated, individual patient data were requested for further examination.Two reviewers extracted data from included trials. Data were pooled where possible, and analysed using appropriate statistical methods. Odds ratios of average differences were calculated. Only 'intention to treat' data were included. Analysis included haloperidol treated patients, compared with placebo.Three studies were found that satisfied selection criteria. These studies compared haloperidol with risperidone, olanzapine, and placebo in the management of delirium and in the incidence of adverse drug reactions. Decrease in delirium scores were not significantly different comparing the effect of low dose haloperidol ( 4.5 mg per day) in one study was associated with an increased incidence of extrapyramidal adverse effects, compared with olanzapine. Low dose haloperidol decreased the severity and duration of delirium in post-operative patients, although not the incidence of delirium, compared to placebo controls in one study. There were no controlled trials comparing quetiapine with haloperidol.There is no evidence that haloperidol in low dosage has different efficacy in comparison with the atypical antipsychotics olanzapine and risperidone in the management of delirium or has a greater frequency of adverse drug effects than these drugs. High dose haloperidol was associated with a greater incidence of side effects, mainly parkinsonism, than the atypical antipsychotics. Low dose haloperidol may be effective in decreasing the degree and duration of delirium in post-operative patients, compared with placebo. These conclusions must be tempered by the observation that they are based on small studies of limited scope, and therefore will require further corroborating evidence before they can be translated into specific recommendation for the treatment of delirium.