Ziprasidone, haloperidol and clonazepam intramuscular administration in the treatment of agitation symptoms in Chinese patients with schizophrenia: A network meta-analysis

Data sources and search strategy

We searched PubMed, Web of Knowledge, Embase and Cochrane Library using the English keywords including schizophrenia, agitation, Chinese, China, injection, and intramuscular and searched Wanfang data, CNKI, SinoMed and VIP databases using the Chinese keywords including schizophrenia, agitation, injection, and intramuscular. Search strategy was based on the Cochrane Collaboration Handbook,6 and we used truncation to improve recall ratio. In addition, we hand-searched articles published in the past 6 months in the related psychiatric journals (Chinese Journal of Psychiatry, Chinese Journal of Nervous and Mental Diseases, Shanghai Archives of Psychiatry and so forth) and references of the retrieved articles (to identify studies not captured by our primary search strategy).7 8

Search results

The search was not restricted to initial published time, and the final search was run on 18 February 2018. A total of 227 retrieved results from four English databases and four Chinese databases were obtained, including 15 English articles and 212 Chinese articles (figure 1).

Request permission

Study selection

Inclusion and exclusion criteria were according to the Consolidated Standards of Reporting Trials statement.9 In general, the trials were considered eligible if they met the following criteria: (1) all the subjects were Chinese patients with schizophrenia; (2) randomised controlled trials (RCTs) with a clear randomised method; (3) correct statistical methods; and (4) complete data including efficacy and/or safety data for meta-analysis. The following criteria were used for data exclusion: (1) non-randomised controlled trials (non-RCTs) or self-control study; (2) statistical data for meta-analysis were unavailable; (3) reviews, case reports, animal experiments and so forth and (4) repeated reports, resulting in unreliable results.

We found 65 articles for screening and after excluding non-conforming studies (n=48), repeated reports (n=2), we included 15 trials for meta-analysis (figure 1).

Quality assessment and data extraction

Two independent reviewers (LS and ZL) assessed the quality of the included studies. Disagreements between the two reviewers were regularly resolved by consulting a third reviewer (SS) or the expert (YX). Major assessment items included: (1) description of research design, research sites and research environment; (2) randomisation, blinding methods, (3) statistical data, description of statistical methods and (4) whether the data of efficacy and side effects were complete. Data extraction and entry of the efficacy indicators and side effects of agitation were performed independently by two evaluators, and then they conducted a cross-check. The descriptions of studies included in the meta-analysis are shown in table 1.

Data analysis

The network meta-analysis was done with multivariate random-effects meta-analysis and meta-regression (mvMeta),10 and network graphs package11 12 in STATA statistical software. The primary outcome was the efficacy towards agitation symptoms, that is, the reduced scores of the Agitation-Calmness Evaluation Scale (ACES), and the secondary outcome was incidence of adverse side effects like excessive sedation, EPS, tachycardia and ECG Corrected QT Interval (QTc) interperiod extension. The specific statistical methods were listed in the published literature,5 13 14 and it should be made clear that the network diagram was a visualisation of network relationships between interventions, each node of which presented an intervention and the size of node indicated the sample size. Moreover, a solid line connecting two points indicated that two nodes had direct comparative evidence. The more comparisons were made, the thicker the connection was. The two points that were not connected can be used to discuss indirect comparative evidence according to the network relationship.11 The publication bias of the included literature were assessed by funnel plot.10 One of the most important functions of network meta-analysis was to sort the interventions by the effect value. In the end, we chose network rank15 to rank the therapeutic effects and adverse reactions of the interventions and used cluster rank14 for comprehensive ranking. Grades of Recommendation, Assessment, Development and Evaluation (GRADE) was used to evaluate the overall level of the evidence.

Included studies characteristics

A total of 15 studies were included in the meta-analysis, including 1 English paper and 14 Chinese papers. The risk bias of four studies were rated as ‘low risk’ (≥5 low-risk items), while 11 studies were rated as ‘high risk’ (≥2 high-risk items). The level of evidence for primary outcome was medium according to GRADE. There were 11 studies comparing ziprasidone and haloperidol and four comparisons between haloperidol and clonazepam in 15 articles. However, searches about chlorpromazine were excluded from the meta-analysis due to intravenous infusion. All the papers included were RCTs and 14 studies reported the efficacy data (ziprasidone vs haloperidol: n=11; haloperidol vs clonazepam: n=3), while the data of main adverse reactions caused by drugs were as follows: excessive sedation were reported by 12 articles (ziprasidone vs haloperidol: n=8; haloperidol vs clonazepam: n=4). Moreover, 12 articles reported EPS, 11 studies reported tachycardia and 8 studies reported abnormal EKG data; all the studies above were a comparison of ziprasidone and haloperidol. The mean age, gender composition and specific therapies are shown in table 1.

Primary outcome: efficacy for agitation symptoms

We were unable to perform an inconsistency test for circulated comparisons of the three drugs (i.e., direct comparison of ziprasidone and clonazepam, no triangular or quadratic loops) (figure 2). Test for heterogeneity were carried by direct comparisons of the drugs (haloperidol vs ziprasidone: Q=16.15, n=10, p=0.10, I²=38.07%; haloperidol vs clonazepam: Q=3.49, n=2, p=0.17, I²=42.65%) but heterogeneity was not statistically significant. Efficacy indicated the primary outcome, that is, the standardised mean difference (SMD) and SE of the differences in the scores of the ACES before and after treatment (haloperidol: SMD=2.278, 95% CI 1.836 to 2.719; ziprasidone: SMD=2.536, 95% CI 2.082 to 2.990; clonazepam: SMD=1.360, 95% CI 0.127 to 2.593). The results of the network meta-analysis suggested that three intramuscularly administered drugs were all effective for agitation symptoms in Chinese patients with schizophrenia. Furthermore, ziprasidone was better than haloperidol and clonazepam in accordance with effect value (see table 2).

Request permission

Secondary outcome indexes: comparison of excessive sedation

The target of interventions was alleviation of agitation symptoms and reducing the risk of impulsive or self-injurious behaviours. Also these interventions were given to enable patients to cooperate with the necessary assessments and related examinations and give them a smooth transition to subsequent standardised treatment. In this study, we used the acceptability indicators as secondary outcome measures, namely the proportion of excessive sedation in adverse events. The results of the network meta-analysis showed that clonazepam has the strongest sedative effect, while haloperidol was similar to ziprasidone (see table 3). The results of the overall efficacy and acceptability are shown in figure 3.

Request permission

Comparison of other adverse reactions such as EPS, tachycardia and abnormal ECG

Since the studies of clonazepam did not report adverse reactions data for EPS, tachycardia and abnormal ECG, we performed a traditional meta-analysis of the above adverse reactions by aligning the two injections of ziprasidone and haloperidol. The results suggested that the incidence of EPS was significantly higher in haloperidol than in ziprasidone (z=5.01, p<0.001) (see figure 4). There were no significant differences between the two groups in the adverse reactions of tachycardia and abnormal ECG (the combined statistics were: z=1.69, p=0.091; z=0.87, p=0.386), (see figures 5 and 6).

Request permission

Request permission

Request permission

Main findings

Evidence of intramuscular injections for treating agitation symptoms in patients with schizophrenia have not been seen in studies in China and abroad. Our study indicates that ziprasidone and haloperidol work better than clonazepam in the treatment of agitation symptoms in patients with schizophrenia. As for the acceptability, the incidences of excessive sedation in clonazepam are higher than haloperidol and ziprasidone. It is necessary for drug interventions that treat agitation symptoms in schizophrenia to work as quickly as possible and to make patients calm but not excessively sedated. Regarding the efficacy and acceptability of intramuscular administration, ziprasidone is superior to haloperidol and clonazepam, which is consistent with the research findings in China and internationally.16,17,18

Studies have shown that the incidence of aggression in mental disorders is 2 10 times higher than that in healthy individuals, with the highest incidence among those with schizophrenia.2 Agitation caused by schizophrenia has an acute onset, rapid progression and can present with violent aggression.3 Schizophrenia patients often have poor compliance, particularly in the acute phase of agitation symptoms and tend to refuse oral medication. Intramuscular injections work fast so they can alleviate agitation-related symptoms quickly and reduce the risk of violent behaviour or tendencies.19 Intramuscular injections are important in clinical practice to control agitation symptoms.

Ziprasidone and haloperidol stand as new and traditional antipsychotics and the oral dosage forms of which are widely used in the treatment of patients with schizophrenia. As for adverse reactions, haloperidol injections have a low risk of sedation and hypotension, but a high risk of EPS and QTc interval extension especially in patients with high-risk factors (such as heart diseases, electrolyte imbalance, combined with other prolonged QTc interval drugs and so forth).3 The chances for EPS occurrences with ziprasidone are small, but tachycardia and QTc interval extension have a possibility of happening (patients with high-risk factors).19 Benzodiazepines such as clonazepam have strong sedative effects and can also be used for the treatment of agitation symptoms, but it can induce excessive sedation, respiratory depression, hypotension, falls and so forth.2 This study tells us that the risks of EPS induced by haloperidol are higher than ziprasidone, and excessive sedations caused by clonazepam are obvious, which is consistent with the above findings. However, what surprised us is that after the treatment with ziprasidone or haloperidol injections, the occurrence probability of QTc interval extension is not high, and this may be related to the small sample size and the strict requirements of the RCT study for the enrolled samples.

Intravenous administration of benzodiazepines or antipsychotics can quickly work; however, it can also cause respiratory depression, hypotension, pain and thrombophlebitis easily.3 4 If we have to apply it to the patients in this way in clinical practice, we should have complete assurance of cardiopulmonary resuscitation equipment and pay close attention to changes in the patient's vital signs. Due to the high safety of intramuscular injection (I.M)., it is rare to report on intravenous injections, and previous studies did not meet the inclusion criteria of quality, so no study of intravenous injection drugs was included in this study.

Limitations

Limitations are as follows: first, studies included are few, and some of them have small sample sizes, which may lead to statistical bias. Second, due to lack of direct comparison of ziprasidone and clonazepam, the assessment of consistency tests is inadequate; the risk of bias in some studies is ‘high risk’; and the level of GRADE evidence for primary outcome measures is ‘medium’. Therefore, the accuracy and interpretation of the research results need to be handled with caution. Last but not least, we only performed a traditional meta-analysis of some adverse reactions because not all the studies reported the data of efficacy and adverse reactions. The conclusions we got need further study to obtain more comprehensive conclusions, which are more conducive to the clinical practice of agitation symptom therapy in patients with schizophrenia.

Research significance

Compared with traditional meta-analysis, network meta-analysis can simultaneously evaluate multiple interventions. In the absence of direct comparisons, indirect comparisons can also provide valuable information for health policy. When direct comparisons exist, combining results of direct and indirect comparisons can increase the accuracy of findings.20 At the same time, network meta-analysis can sort the interventions by their efficacies and calculate the probability of optimal interventions.14 In the present study, although there were not direct comparisons between clonazepam and ziprasidone, the network meta-analysis still showed that ziprasidone has the advantage in the efficacy and the probability of excessive sedation related to ziprasidone is smaller than that of clonazepam.