As reported in this issue of Critical Care, Reade and colleagues [1] have completed a pilot open-label trial comparing the efficacy of haloperidol with dexmedetomidine in the treatment of ventilated patients with agitation. This is an important contribution to the literature on delirium management despite the limitations of what is essentially a pragmatic feasibility study. Agitation leads to adverse events such as accidental extubation, removal of arterial and central lines and injury. The current delirium treatment used by the majority of intensivists worldwide is haloperidol. With agitation, clinicians often also use benzodiazepines for rapid control. This may result in continuing or starting respiratory support for over-sedation. Also, benzodiazepines and physical restraints are deliriogenic. A recent Cochrane review concluded that benzodiazepines cannot be recommended for the control of this condition [2].

Dexmedetomidine is a highly selective α2-receptor agonist that has been available in the US since 1999 as a short-term sedative in the intensive care unit (ICU). It has an affinity for the α2-receptor that is eight times higher than that of clonidine and a half-life of 2 hours. It is clinically effective, has opiate-sparing properties and is well tolerated. It is not yet available for general use in Europe. Reade and colleagues studied 20 intubated agitated patients and demonstrated impressive decreases in time to extubation and ICU length of stay in patients on dexmedetomidine as opposed to haloperidol. There are case reports regarding the successful use of dexmedetomidine in agitated patients with withdrawal syndromes and facilitating extubation in 13 of 20 agitated patients difficult to wean [3, 4]. Conversely, a retrospective assessment of 40 patients found transitioning to dexmedeto-midine to be associated with increased agitation and pain. Five patients self-extubated and four of these needed emergency reintubation [5].

Delirium is common, life-threatening and largely ignored in most ICUs. In fact, the majority of patients with delirium are hypoactive, are quiet and lethargic, or present with a mixed hypo/hyperactive variant [6]. If clinicians do not use a simple tool to screen for diagnostic features of delirium, they are likely to miss it 70% of the time [7]. The incidence in a UK critical care unit is documented as two out of three mechanically ventilated patients [8]. The issue regarding the pharmacological management of delirium, agitated or otherwise, is the lack of evidence. Although haloperidol has been used on hundreds of thousands of patients since 1950 and is the drug of choice in all delirium guidelines, this is solely on the basis of case series or case reports rather than any robust clinical effectiveness trials. Regardless, most patients with delirium are untreated because it is usually undiagnosed. This is frustrating for those of us who are clear about the impact of delirium. The undeniable association between delirium and long-term cognitive impairment alone should prompt us to take notice [9]. Discharging patients so they can spend the rest of their lives in a nursing home or struggling to remember names goes against all we are trying to achieve. While this study is small and unblinded and lacks objective endpoints, it is welcome not because it shows that dexmedetomidine is effective in the treatment of agitated delirium (although the findings suggest it may be), but because it shows that haloperidol may not be.

Delirium has bad outcomes. Hyperactive delirious patients are often violent and remain days longer on sedation and ventilated due to agitation. Dexmedetomidine is a good choice for investigation. But pure hyperactive delirium is just the tip of the iceberg; many more patients are quietly delirious. We do not know whether treating delirium will improve outcomes; the study has not been done. We do not know whether haloperidol decreases mortality as hypothesised by Milbrandt and colleagues [10] or increases mortality as retrospective cohort studies in demented patients have suggested [11]. Two other Cochrane reviews into delirium conclude that more trials are urgently needed [12, 13]. Now clinicians need to decide whether to implement delirium screening as recommended and how to manage hypoactive delirium in their patients. Delirium screening resources are available on ICU delirium information websites [14].