Efficacy and safety of haloperidol versus atypical antipsychotic medications in the treatment of delirium

Delirium is a common, complex neuropsychiatric disorder with a high prevalence among elderly hospitalized patients [1‐3], postsurgical patients [4, 5], and cancer patients [6‐8] in advanced stages of illness. Typically, delirium shows an abrupt, rapid onset and a fluctuating course [9, 10]. The core features of delirium consist of disturbances in cognitive function such as attention, memory, thought, and language [9, 10]. However, its clinical presentation can be highly variable with a broad range of associated non-cognitive, behavioral symptoms that reflect the influence of distinct etiologies, medical comorbidities, or pharmacological treatments [10, 11].

In hospitalized elderly patients, the prevalence of delirium ranges from 10 to 40% [3, 12, 13]. Delirium is also associated with major adverse outcomes such as increased mortality, functional impairment, prolonged hospitalization, and increased cost of care [14‐16]. Regardless of the evident clinical significance, delirium tends to be under-diagnosed and under-treated [17]. Therefore, early identification and effective psychiatric treatment of delirium is important in the comprehensive care of elderly hospitalized patients [18].

The management of delirium includes ensuring safety with environmental or supportive interventions, identifying and treating the cause of delirium, and enhancing the patient’s functioning [19, 20]. Regarding pharmacological intervention, antipsychotic medication has been considered as first-line pharmacotherapy of delirium except in the case by sedative or alcohol withdrawal [19, 20]. Haloperidol, a typical antipsychotic, has continued to be the most frequently used antipsychotic drug [19‐21] due to its effectiveness, relatively lesser sedative and hypotensive effects and fewer anticholinergic properties [19, 21]. However, haloperidol may induce adverse side effects such as extrapyramidal symptoms (EPSs) [19, 21] or prolongation of the QTc interval and fatal arrhythmia such as torsade de pointes among patients with delirium [19, 21‐23]. EPSs are more likely to occur in elderly and seriously medically ill patients, who are also the most susceptible to delirium [24]. In addition, it may be difficult to distinguish agitation, a common behavioral symptom of delirium, from akathisia, a frequent EPSs induced by haloperidol [25, 26].

Recently, atypical antipsychotics such as risperidone, olanzapine, and quetiapine have been increasingly used to treat delirious patients due to the lower incidences of EPSs associated with these drugs [20, 27‐29]. Although a number of studies have been conducted to evaluate the efficacy or safety of atypical antipsychotics in the treatment of delirium, most of these reports have been in the form of case reports or open-label trials [30‐32]. Only two placebo-controlled, randomized trials of atypical antipsychotics for the treatment of delirium have been reported [33, 34]. Some randomized comparative trials assessing the efficacy of various antipsychotics in the treatment of delirium have been conducted in critical care units or consultation-liaison psychiatric services [26, 28, 32, 35‐39]. Five trials compared the efficacy between one atypical antipsychotic agent and haloperidol [28, 35, 36, 38, 39] and two trials compared the effectiveness of two different atypical antipsychotics [26, 37]. One trial assessed the comparative efficacy among two atypical antipsychotics and haloperidol [32]. Previous trials comparing the treatment response of atypical antipsychotics based on age (<70 years old, ≥70 years old) suggested that older age might predict a poor response to the treatment of delirium [26, 40]. Recently, researchers in the field of psychiatry often divide older subjects into two groups (young-old and old-old), with an age of 74 being the cutoff point [41‐43]. However, no previous research has compared the response rate to various atypical antipsychotics in the treatment of delirium based on this age grouping.

Although previous trials have reported that there was no significant difference in the efficacy between haloperidol and atypical antipsychotics in the treatment of delirium [20, 21, 32, 44], the reported data are not sufficient to form conclusions regarding the efficacy of various atypical antipsychotics compared to haloperidol. Only a few trials have considered age as a factor when comparing the response rates of atypical antipsychotics in the treatment of delirium [26, 40]. To our knowledge, no previous trial has compared the efficacy and safety of haloperidol with more than two atypical antipsychotics in the treatment of delirium. For these reasons, we investigated the comparative efficacy and safety of haloperidol versus atypical antipsychotic medications in the treatment of delirium. The primary objective of this study was to compare the efficacy and safety of haloperidol versus three atypical antipsychotic medications (risperidone, olanzapine, and quetiapine) for patients with delirium. The secondary objective was to investigate whether response rate of haloperidol and three atypical antipsychotic medications differ depending on age, dividing the study cohort into two age groups, in the treatment of delirium.

Recently, a number of researchers have reported that atypical antipsychotics may be as effective as haloperidol in treating delirium [20, 21, 28, 32, 35, 36, 38, 39]. Risperidone, olanzapine, and quetiapine have been increasingly used for pharmacologic intervention of delirium [20, 26‐29]. Previous researches have shown that the efficacy of risperidone and olanzapine is not different from that of haloperidol in the treatment of delirium [28, 32, 35, 36, 44, 57‐59]. Quetiapine has been reported to be as efficacious as haloperidol [58] and to reduce the severity of the symptoms of delirium more rapidly than placebo [33]. To date, most randomized comparative trials of atypical antipsychotics in the treatment of delirium have compared the efficacy of one atypical antipsychotic agent and haloperidol [28, 35, 36, 38, 39] or two different atypical antipsychotics [26, 37]. Only one randomized comparative study has compared the efficacy of two different atypical antipsychotics and haloperidol [32]. To our knowledge, this study is the first trial to compare the efficacy and safety of haloperidol versus three atypical antipsychotics. In the present study, haloperidol, risperidone, olanzapine and quetiapine were equally effective in improving the symptoms of delirium. There was no significant difference in the rate of reduction of DRS-K severity score and improvement of K-MMSE score with time among the four groups. Recently, one comparative efficacy study of haloperidol, risperidone, and olanzapine showed that risperidone and olanzapine were as efficacious as haloperidol in treating delirium [32]. Our result supports the findings of previous researches with regard to the comparative efficacy of haloperidol versus three atypical antipsychotics in managing symptoms of delirium [28, 32, 35, 36, 44, 57‐59]. The mean daily doses of risperidone, olanzapine, and quetiapine were not largely different from those of previous studies [26, 32, 35, 36, 57, 60‐62]. This finding also suggests that a relatively low dose of atypical antipsychotics may be effective in managing the symptoms of delirium [26, 28, 32, 36].

There have been no previous studies to have assessed the difference in the efficacy of atypical antipsychotics in terms of the cognitive and non-cognitive subscale scores of the DRS-R-98, with the exception of one placebo, controlled trial [33]. The group difference was not significant over this study period in terms of the rate of reduction of the cognitive and non-cognitive subscale scores of the DRS-K. In fact, improvement of cognitive and non-cognitive symptoms of delirium could occur naturally [33], due to nonspecific environmental care [63] or to treating the underlying etiologies of delirium [21]. Nevertheless, our findings might be meaningful in that the effectiveness of haloperidol versus three atypical antipsychotics was compared in the two different symptom domains of delirium at the same time in a real clinical setting.

The demographic characteristics of the subjects enrolled in this study were not significantly different from those of subjects in previous studies [32, 36, 37]. In regards to the causes of delirium, most subjects had more than one contributing cause of delirium. The most common definite cause of delirium was metabolic or endocrine disturbance, a finding which was similar to that of previous studies of patients with delirium referred to a consultation-liaison psychiatric service [32, 47].

In the present study, the response rate to olanzapine was poor in subjects over 75 years old compared to those under 75 years old. However, the response rate to the other three antipsychotics was not significantly different between age groups. A previous study reported that old age was associated with poorer response to olanzapine in hospitalized cancer patients with delirium [40], while another study reported that the response rate of olanzapine was similar depending on age, but that the response rate to risperidone was much lower in an older age group [26]. The major neurotransmitter hypothesized to be involved in delirium is acetylcholine, and several studies have reported that a variety of delirium-inducing factors are associated with decreased acetylcholine activity in the brain [64]. Actually, many medications with anticholinergic side effects can induce or aggravate delirium [21]. In regards to the pharmacological profile, olanzapine is known to have a significant affinity for muscarinic receptors and to induce relatively more anticholinergic adverse effects than the other three antipsychotics [65, 66]. Age-related differences in susceptibility to anticholinergic adverse effects might have affected the response rate in the olanzapine group. Another possibility for the reduced effectiveness of olanzapine in the older age group is that some of the patients may have had undiagnosed dementia. In addition, this might be related to differences in the general underlying medical condition, as the frequency of organ insufficiency was relatively higher in the olanzapine group than in the other three groups. These results suggest that advanced age is not only a risk factor [21, 64, 67] for delirium, but also may be a predictor of poor response to delirium treatment with atypical antipsychotics. Therefore, further investigation of the impact of age on treatment response is required.

There were no significant group differences in the number of subjects experiencing adverse events or in the type of adverse events. Although previous review articles have suggested that atypical antipsychotics are safe, with a lower rate of adverse events compared to haloperidol in the treatment of delirium [20, 21], a Cochrane review reported that haloperidol at a low dosage (<3.5 mg/day) was safe, with a similar frequency of adverse events compared to atypical antipsychotics [44]. In this study, the mean daily dose of haloperidol was relatively low (1.2 ± 0.4 mg/day) and the total duration of medication was relatively short. Thus, the results of the present study suggest that a low dose of haloperidol is safe and does not show a greater frequency of EPSs [44] compared to atypical antipsychotics over a relatively short period of treatment.

This study has several limitations: ⓐ Our study did not include a placebo control group. The absence of comparative placebo control group with active treatment groups limited the interpretation of our findings; ⓑ The use of empirical judgment of clinician to assign antipsychotic medications without randomization might be important source of bias and limit our findings. However, we found no significant difference in demographic and clinical characteristics among four groups; ⓒ Heterogeneity of study population including medical and surgical patients could be a major limiting factor in the interpretation of our main findings even though there was no significant difference in clinical characteristics among four groups; ⓓ Since the consultation-liaison psychiatric service might recruit a skewed patient cohort most likely to have active psychiatric symptoms, this should be acknowledged as a source of bias; ⓔ Since this was a pilot study, we did not consider sample size as a requirement for carrying out the proposed objectives and the small subject numbers in each group might limit the strength of the conclusions from our work. Thus, further studies with a larger number of subjects are needed to test our findings; ⓕ Although we made efforts to exclude patients with dementia by reviewing detailed clinical history and by obtaining information from reliable informants, dementia was not evaluated by a standardized screening instrument. Given that the number of subjects with dementia was relative low (N = 8) in screening process, subjects with dementia might be misclassified; ⓖ Although the rater was blind to which study drug was being administered, as the rater knew that all subjects were receiving active treatment, the ratings could have been affected; ⓗ Confounding factors associated with rescue medication could not be rigidly controlled due to ethical considerations. The permission to use rescue medication might not have seriously affected the results of this study because the total mean dose of intramuscular rescue injection of haloperidol was not significantly different among the four groups. If the dosing titration of the study drugs had been escalated more rapidly, the need for rescue medication could have been decreased; ⓘ The dropout rate was relatively high, and the missing data caused by dropouts could affect the result of this study. In order to overcome this limitation, we used a linear mixed model, in which all available data can be included and missing data can be appropriately addressed [54, 55]. By using the average area under each subject’s rating scale trajectory, we could compare treatment groups across whole study period; ⓙ The safety analysis was focused only on extrapyramidal side effects and could not evaluate other potential adverse events, such as QTc prolongation or arrhythmia associated with use of antipsychotic drugs [19, 21‐23]. Overall, six patients were transferred to the intensive care unit due to aggravation of an underlying medical or postsurgical condition during the study. Given that potential adverse effects of antipsychotic drugs, which had not been captured by data, might contribute to aggravation of underlying medical conditions, transfer to the intensive care unit during the study needs to be considered as a potential adverse effect of antipsychotic medications; ⓚ Changes in the various medical or surgical conditions of the study subjects might have affected the symptom severity of delirium, regardless of the use of antipsychotics; ⓛ Finally, as the symptoms of delirium can fluctuate and improve irrespective of the treatment given [33], the findings of our study must be understood in the background of these limitations. Regardless of these limitations, the results of the present study could provide important clinical information regarding the usefulness of commonly prescribed antipsychotics in the treatment of delirium with various underlying etiologies in a tertiary hospital setting.

In conclusion, the atypical antipsychotics risperidone, olanzapine, and quetiapine and low dose haloperidol were equally effective and safe in the treatment of delirium. The treatment response rate for the only olanzapine group was significantly lower in subjects over 75 years old than in subjects under 75 years old. The factor of age needs to be considered in the choice of antipsychotic medication for the treatment of delirium. Further prospective randomized placebo controlled trials of a larger patient group with delirium should be carried out to test the generalizability of our findings.