Good clinical guidelines must define the setting,patients and evidence: Benzodiazepines versus droperidol for acute behavioural disturbance in the emergency department

Jensen et al., (2016) challenge a recent guideline for ‘The sedation of patients in the emergency department with severe acute behavioural disturbance’. They suggest that these guidelines applied to young people with their first diagnosis of psychosis increase the risks of adverse drug reactions from antipsychotic drugs. They propose that treatment in these patients should follow the UK National Institute for Health and Care Excellence (NICE) guidelines for ‘Violence and aggression: short-term management in mental health, health and community settings’ (NICE, 2005), which suggest first-line use of lorazepam in young patients with psychosis, rather than an antipsychotic. To them, the titles suggest the guidelines cover the same scenario and therefore are in conflict. In fact, the guidelines are aimed at completely different settings, with minimal overlap of patients.

The NICE guidelines were developed for the patients in mental health institutions and only provide some guidance for emergency departments for mental health patients, as acknowledged: ‘This guideline includes adults (aged 18 and over), children (aged 12 and under) and young people (aged 13 to 17) with a mental health problem who are currently service users within mental health, health and community settings’. The guidelines are written for patients in mental health units where treatment is aimed at the underlying disorder (i.e. psychosis). These patients usually have a working diagnosis, have been admitted for hours to days and agitation and violence escalate while they are an inpatient. Treatment with benzodiazepines or antipsychotics is likely to occur on multiple occasions, and the use of antipsychotics in large doses or without anticholinergic drug cover is likely to result in extrapyramidal side-effects (Gillies et al., 2013). This is of particular concern in young patients with their first diagnosis of psychosis. The NICE guideline authors suggest it is sensible to initially use benzodiazepines in this setting and slowly introduce antipsychotic drugs.

In contrast, the recently produced guideline on the ‘Management of patients with Acute Severe Behavioural Disturbance in Emergency Departments’ was developed for a different patient group and setting. These guidelines are aimed at patients in an emergency department, an open environment, full of dangerous items and vulnerable critically ill people. The patients who are violent and agitated mostly have no established diagnosis and pose a serious risk to themselves, other patients and staff. This patient group mostly have substance -use disorders, mainly acute alcohol intoxication or, less commonly, stimulants (Calver et al., 2015; Isbister et al., 2010). A primary mental health disorder (including psychosis) accounts for only about 15%, and first-episode psychosis is much rarer again. The three or four emergency departments that might see patients referred to Jensen’s clinic would have treated several hundred aggressive violent patients in the year of their audit; that 10/39 young people with an acute psychosis presenting to an emergency department received a dose of an antipsychotic is not a good reason to avoid evidence-based treatment for the vast majority. Importantly, these patients generally require just one dose of parenteral antipsychotic. Most are discharged within 6 h, with resolution of their behavioural disturbance as the effects of alcohol or drug intoxication wear off (Isbister et al., 2010).

Research on sedating patients with acute behavioural disturbance in emergency departments has shown droperidol is as effective as a benzodiazepine (midazolam), but results in fewer adverse effects and less requirement for additional medication (Isbister et al., 2010; Spain et al., 2008). A large study of 1403 administrations of droperidol in these patients found few adverse effects occurred and extrapyramidal side-effects occurred in only 7/1403 patients (0.5%) (Calver et al., 2015).

Young people diagnosed with acute psychosis in the emergency department will be transferred to a mental health setting as soon as possible. There are other guidelines designed for acute psychosis, such as those developed by NICE, which will be more appropriate.

There is surprisingly little evidence on the management of behavioural disturbance. The NICE guidelines on treatment of aggression in the mental health setting certainly support benzodiazepine use for aggression. However, Jensen et al., (2016) err when they state that ‘there was good evidence that benzodiazepines are at least as effective as antipsychotics for psychosis-induced aggression or agitation’ referencing Gillies et al. (2013). Gillies et al. conclude, ‘The evidence from trials for the use of benzodiazepines alone is not good’. Remarkably across all comparisons in this review, the studies are small, mostly company-sponsored and generally use poorly tolerated haloperidol as the comparison treatment against the sponsors’ proposed treatment. All comparisons in this review have grade ratings that they are based on very low or low quality evidence, indicating that much better evidence is required.

Although we encourage clinicians to refer to clinical guidelines, it is important to understand when they are not based on good evidence. Both evidence and guidelines must be applied in the correct setting. In the setting of aggressive persons in the emergency department, we have better evidence, substantially arising from recent Australian research (Calver et al., 2015; Isbister et al., 2010). We also now have Australian guidelines supporting an evidence-based approach which promotes first-line use of a single intramuscular dose of droperidol. It is both safer and more effective than benzodiazepines in the emergency department. These emergency department guidelines should not be unthinkingly applied in other settings, just as guidelines from other settings should not be extrapolated to the emergency department.