Quantity of Research Available
A total of 177 citations were identified in the literature search. Following screening of titles and abstracts, 148 citations were excluded, and retrieved 29 potentially relevant reports for full-text review. Of these potentially relevant articles, 23 publications were excluded for various reasons. Further, two relevant publications from the grey literature and hand searches were retrieved. Ultimately, eight reports were included. Appendix 1 presents the flowchart of the study selection.
Appendix 2 provides additional references of potential interest that were found from the grey literature and hand searches. These are mostly studies that do not appear to focus on an intervention of interest, but may provide information about the efficacy or safety of other drugs for rapid tranquilization in the population of interest. Most studies are protocols of ongoing systematic reviews.
Summary of Study Characteristics
Appendix 3 summarizes the study characteristics.
Study Design
Eight relevant reports were identified – three RCTs, four SRs, and one CPG.15–22
All SRs exclusively included RCTs (nine to 32 trials per review) and pooled data for at least one outcome of interest.18–20,22 The publication year of the included studies within the SRs ranged from 1969 to 2015. The total number of patients enrolled in the eligible studies ranged from 552 to 3877.
All three RCTs directly compared two or more active therapies.15–17 Two of the three studies were superiority trials,15,17 while one was a non-inferiority trial.16 Two RCTs involved a single study centre.15,17
One evidence-based CPG, published by the National Institute for Health and Care Excellence (NICE) in 2015, provided recommendations on the short-term management of violence and aggression in mental health, health and community settings.21 Part of the recommendations included use of rapid tranquilization. It included evidence from SRs and RCTs, and, if needed, observational studies.
Country of Origin
Lead authors of three SRs originated from the United Kingdom (UK),18,20,22 while one was from Japan.19 The individual trials that the reviews included originated from numerous jurisdictions worldwide.
Among the RCTs, two studies originated from Australia,15,17 while the remaining was from China.16
The NICE guidelines were from the UK.21
Patient Population
All SRs focused on managing patients who exhibited agitated or aggressive behavior due to an underlying mental illness.18–20,22 Most of the included trials enrolled patients with schizophrenia.
The three trials enrolled different patient populations. Zhang et al., for instance, only enrolled Chinese patients who were in the acute phase of schizophrenia.16 The other two trials included a mixed patient population. Calver et al. mostly (56%) enrolled aggressive patients with mental illness, although about one in three patients (27%) they included presented with drug-induced psychosis.15 Conversely, more than half of the patients that Isbister et al. included exhibited agitation that was secondary to alcohol intoxication.17
The NICE guidelines targeted mental health service users who exhibit violent and aggressive behavior.21
Interventions and Comparators
The SRs evaluated a range of interventions and comparators. One review tested the effectiveness and safety of benzodiazepines alone versus eight different comparators, of which two – placebo and antipsychotic drugs alone – were relevant to this report.22 Although the authors sought trials of any benzodiazepine administered at any dose or via any route, they reported the majority of trials (against the two relevant comparators) tested IM routes of benzodiazepines, mostly lorazepam. The other reviews focused on antipsychotic drugs, including IM zuclopenthixol acetate,18 IM olanzapine,19 and any administration of haloperidol.20 Relevant comparators of interest in these reviews varied, and included oral and IM antipsychotic drugs and IM benzodiazepines alone, as well as a combination of haloperidol and promethazine. Powney et al., for instance, compared haloperidol with aripiprazole, chlorpromazine, droperidol, loxapine, olanzapine, perphenazine, risperidone, thiotixene, ziprasidone, zuclopenthixol acetate, and a combination of haloperidol and promethazine.20 The authors reported that for all but one study that they included, trial investigators administered haloperidol and other drugs via the IM route. Jayakody et al. compared IM zuclopenthixol acetate 50–100 mg (high-dose, as defined by the trial authors) with IM zuclopenthixol acetate 25–50 mg (low-dose), and compared IM zuclopenthixol acetate (dosage not specified) with standard therapy. Standard therapy was comprised of any of the following treatments: oral haroperidol, IM haloperidol, oral zuclopenthixol, oral chlorpromazine, IM chlorpromazine, oral clothiapine, or IM clothiapine.18 Lastly, Kishi et al. tested the efficacy and safety of IM olanzapine versus placebo, IM haloperidol, IM ziprasidone, a combination of IM haloperidol and promethazine, and IM lorazepam.19 The doses of the therapies in the above reviews varied.
Interventions and comparators among the three eligible RCTs varied. Two studies evaluated IM haloperidol at varying doses. Calver and colleagues compared IM haloperidol 10 mg with IM droperidol 10 mg,15 while Zhang et al. tested IM haloperidol up to 20 mg versus IM ziprasidone up to 20 mg.16 The third trial tested IM droperidol 10 mg versus two therapies – IM midazolam 10 mg and a combination of IM droperidol 5 mg and midazolam 5 mg – although the combination therapy was not a comparator of interest in this report.17
The NICE guidelines focused on antipsychotic drugs, including aripiprazole, chlorpromazine, haloperidol, loxapine, olanzapine, quetiapine, risperidone, benzodiazepines, and antihistamines.21 For the purposes of rapid tranquilization, it focused on use of these medications by the parenteral route, typically IM, and, in rare cases, intravenously. Comparators of interest included placebo or another intervention (not specified).
Outcomes
Efficacy outcomes of interest included sedation, and management of agitation and aggression. The four SRs and three RCTs evaluated treatment effects across a number of outcomes. Common instruments to measure efficacy included the Agitated Behavior Scale (ABS), the Agitation-Calmness Evaluation Scale (ACES), the Positive and Negative Syndrome Scale (PANSS), as well as the PANSS-Excited Component (PANSS-EC). Investigators reported scores on these measures in a number of ways, including change from baseline to a certain time-point, end of study scores, or both.
Safety outcomes of interest included rates of adverse events (AEs) (overall and serious) and extrapyramidal symptoms (EPS). Some studies used the Simpson-Angus Scale (SAS) to evaluate EPS.16,19,20
Relevant efficacy and safety outcomes that the NICE guidelines considered in formulating recommendations were rates of violence and aggression. tranquillization. sedation/somnolence, and AEs.21 The guidelines only considered short-term outcomes, i.e. 72 hours after treatment. The guideline used the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach to evaluate the quality of the evidence. Further, the strength of each recommendation was qualified by the choice of wording, although this is not described in detail.
Summary of Critical Appraisal
Appendix 4 summarizes the results of the critical appraisal.
Three of the four studies were Cochrane reviews, which meant they adhered to rigorous methodology and high reporting standards.18,20,22 This included assessment of risk of bias among individual studies, exploration of heterogeneity (clinical, methodological, and statistical) between studies, robust statistical analyses, and evaluation of quality of the body of evidence using Grading of Recommendations Assessment, Development and Evaluation (GRADE). Strengths of the one non-Cochrane review, by Kishi et al., include adherence to the PRISMA guideline for reporting systematic reviews, duplicate study selection and data extraction processes, and exploration of heterogeneity between studies.19 A key limitation, however, was lack of risk of bias assessment of included studies, hence limiting our ability to evaluate the internal validity of the studies, and ultimately judge the underlying quality of the evidence. Further, the authors did not use comprehensive search strategies (few search terms used, few databases searched) to find relevant studies, nor did they evaluate the quality of the body of evidence.
Authors of all three RCTs reported registering their studies on an international or national registry.15–17 Although described as randomized trials, the studies by Calver et al. and Zhang et al. did not mention the manner in which the investigators generated the randomization scheme.15,16 Further, Zhang et al. did not describe the method they used to conceal the randomization sequence, which raises the possibility of selection bias.16 The extent to which investigators across all trials blinded participants and study personnel also varied. Isbister et al., for instance, reported blinding patients, health care providers, and study investigators,17 but Zhang and colleagues reported not blinding participants and most study personnel (except outcome raters) in their study.16 In addition, although Calver and colleagues describe their trial as masked, the group(s) to which they are referring is unclear.15 All trials generally used robust techniques to conduct analyses. In particular, Zhang and colleagues, who conducted a non-inferiority trial, provided appropriate rationale for setting their non-inferiority margin, and conducted per-protocol analyses to support their conclusions.16
Overall, the NICE guidelines were very well-conducted.21 This included clear description of the objectives, the health questions, as well as the populations of interest. The guideline development group comprised of representatives from several clinical roles, and the development process involved several patients/patient groups, and public organizations. The guideline used systematic methods to search for evidence. A key limitation, however, is that the manner in which the guideline development group used the evidence to formulate the final recommendations was unclear.
Limitations
Several important limitations exist among studies of antipsychotics or benzodiazepines as rapid tranquilization in in-patients of mental facilities and emergency departments, which leaves uncertain the validity of the results. In particular, shortcomings of the primary studies limited the utility of the systematic reviews, primarily due to sparse data, and poor quality of reporting and design of the individual trials. Further, a limitation of most reviews (except for Kishi et al.) is that the literature searches were outdated. A key limitation among the three relevant RCTs that we found is the variability in methodology and rigor, including the extent to which trialists blinded participants and study personnel. As well, investigators often used a range instruments to measure efficacy outcomes, in particular, which limits the utility of results and precludes comparisons across trials. In the NICE guidelines, the quality of the underlying evidence that led to recommendations was very poor. Further, a key limitation of the guidelines was the unclear link between the evidence and final recommendations.
