The Haloperidol Tradition
There are currently no drugs officially approved by the U.S. Food and Drug Administration (FDA) for the treatment of hospital associated delirium. 1,2 Haloperidol, however is currently the medication recommended by both The American College of Critical Care Medicine (Grade of recommendation=C),3and the American Psychiatric Association 4 for the treatment of delirium in patients who are critically ill. The use of haloperidol as a first line agent for the treatment of delirium is based more on collective clinical experience than on rigorous, well designed studies.4-10 Most studies examining the use of pharmacological agents for the treatment of delirium are case studies or case series, 5,7,8,10 retrospective rather than prospective, 5,8,10 open label rather than blinded, 4-6,8,11 lacking placebo controls, 4-6,9-11 and—even when otherwise well designed—lacking in large enough numbers of patients to justify definitive conclusions favoring one agent over another. 3,5,6,8-11
Among haloperidol’s many benefits are: flexibility in mode of administration—it can be given orally, intramuscularly (IM) or intravenously (IV); 3,8,10 relatively few adverse effects when used in low doses for a short duration of time; and few active metabolites. Specifically, it has few anticholinergic, hypotensive, or sedating effects.8, 10 Among haloperidol’s potential risks are: the possibility of causing parkinsonism/ extrapyramidal symptoms (EPS), especially when given orally, in higher doses, and for longer periods of time; and prolongation of the QT interval when given intravenously.3,10
Haloperidol and the Atypical Antipsychotics
A Cochrane review, “Antipsychotics for delirium,” 6 was first published in 2007 and edited with no changes to conclusions in 2009. The specific objective of the Cochrane reviewers was “to compare the efficacy and incidence of adverse effects of haloperidol with risperidone, olanzapine, and quetiapine in the treatment of delirium.” 6(p.1) Only three studies met their selection criteria that trials be unconfounded, randomized, “with concealed allocation of subjects,” and include pre and post treatment assessment. 6 The authors were able to perform a meta-analysis comparing haloperidol with risperidone, and haloperidol with olanzapine, but there was not enough data to compare haloperidol with quetiapine. They found that, in terms of effect on delirium, olanzapine, risperidone and haloperidol were all comparable, and all showed an improvement when compared with placebo (when placebo controls were used). They reported one study by Hu in 2004 which showed earlier improvement in symptoms of delirium with olanzapine but no difference in end-of-study delirium scores.
The Cochrane reviewers also noted the study by Kalisvaart et al. in which haloperidol was used prophylactically. In that study 430 patients over the age of 70 undergoing emergency or elective hip repair were randomized to receive either placebo or haloperidol 0.5 mg three times a day from admission to 3 days after surgery. Although the incidence of post-operative delirium did not differ in the two groups, those patients in the haloperidol arm experienced delirium for shorter periods of time, had less severe delirium, and were discharged from the hospital sooner. 5,6,12 Comparable studies looking at the prophylactic use of atypical antipsychotics have not been reported.
Since the Cochrane review was published, Devlin et al., 11 published a study in 2010 comparing quetiapine to placebo in the treatment of delirium for ICU patients who were requiring as needed (prn) haloperidol. In their study of 36 patients, those who were in the quetiapine arm had a shorter interval before their first resolution of delirium, spent less time being agitated and/or delirious, and required both the study drug and haloperidol for a shorter overall duration. 100% of patients on quetiapine experienced at least one resolution of delirium compared with 78% of patients on placebo. No differences were noted between groups in terms of hospital mortality, how long patients had to remain on a ventilator, remain in the ICU, or remain in the hospital. Quetiapine patients, however, were more likely to go home or to a rehabilitation facility compared to patients in the placebo arm, who were more likely to be discharged to a long term care facility or to die.11
Compared with haloperidol, the atypical antipsychotic agents are less likely to be associated with the development of parkinsonism/EPS or QT prolongation. However, since 2004 risperidone and olanzapine, when used longer term (8-12 weeks) have, been associated with an increased stroke risk 5, 8 an increased risk of infection, 11 and increased mortality, 4,5,8 in older patients with dementia. No studies have specifically looked at the treatment of delirium in patients with underlying cognitive dysfunction and some, such as the quetiapine study discussed above 11 specifically excluded patients with dementia, a primary neurological condition or irreversible brain disease. In the limited studies that have been done, the atypical antipsychotics seem to have a comparable efficacy to haloperidol. 5,6,8-10,13
Additionally, no significant differences have been noted between the three atypical antipsychotics which are in widest use and are the most studied: risperidone, olanzapine and quetiapine.9 It has been suggested that atypicals be considered as first line agents for the treatment of delirium in patients who seem to require higher doses of haloperidol or who have been determined to be at risk for extra pyramidal symptoms or cardiac arrhythmias. 6,8 Otherwise, the incidence of side effects with low dose haloperidol seems not to be significantly different than the incidence of side effects with the atypical antipsychotics and, as the authors of the Cochrane review remind us, haloperidol remains considerably less expensive.
Why Aren’t There More Studies?
There are many difficulties inherent in the study of delirium which may contribute to the relative scarcity of clinical trials looking into pharmacological options for its treatment. First, by definition, delirium has a fluctuating clinical course with a waxing and waning of symptoms. It can therefore be unclear whether to credit clinical improvement to a particular intervention or to the natural course of the disorder. 4,9-11 Even more importantly perhaps, delirium is rarely treated with pharmacological agents alone. In virtually all cases, once delirium is recognized, efforts are directed at treating the underlying cause and hopefully at employing non-pharmacological measures as well. It would be very difficult and, in fact, not ethical, to have a delirium study in which these two confounding factors were not playing a major role. So, not only is improvement in symptoms potentially part of the natural history of delirium itself, but it may also attributable to improvement in the underlying disorder and/or to the use of non-pharmacological measures.
Another factor which complicates meta-analysis and precludes the drawing of definitive conclusions is the wide variation in the instruments used by different researchers to diagnose, measure, and monitor delirium. Among these are: the Confusion Assessment Method (CAM); Clinical Global Impressions (CGI); Delirium Index (DI); Delirium Rating Scale (DRS); Intensive Care Delirium Screening Checklist (ICDSC); Memorial Delirium Assessment Scale; Sedation Agitation Scale (SAS)—to name just a few.
Yet another limitation to most studies lies in the failure to distinguish between hyperactive or agitated, and hypoactive delirium-- which may actually have the worse prognosis.4,5 Not only might these two forms of delirium require different approaches, but it is possible that cases of hypoactive delirium are being missed completely and/or not considered for treatment or study. Alternatively, hypoactive phases of delirium may potentially be misinterpreted as resolution of symptoms.
Finally, no studies have looked at potential long term effects of pharmacological intervention on parameters such as cognitive function, functional status or quality of life post acute hospitalization. 5,6,11 Delirium itself, however, has long been associated with longer hospitalizations, irreversible cognitive decline and a higher likelihood of long-term care placement.14 Intuitively, it seems that any interventions for prevention or treatment of this complication would be welcome, but systematic, well-constructed and conducted studies looking at both short and long-term effects remain desperately needed.