Gamma aminobutyric acid (GABA) receptor agonists for acute stroke

  • Review
  • Intervention

Authors


Abstract

Background

Gamma aminobutyric acid (GABA) receptor agonists have been shown to have a neuroprotectant effect in reducing infarct size and improving functional outcome in animal models of cerebral ischemia. However, the sedation effects of GABA receptor agonists have limited their wider application in acute stroke patients due to the potential risk of stupor.

Objectives

To determine the efficacy and safety of GABA receptor agonists in the treatment of acute stroke.

Search methods

We searched the Cochrane Stroke Group Trials Register (January 2012), the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2012, Issue 1), MEDLINE (1949 to March 2012), EMBASE (1980 to March 2012), CINAHL (1982 to March 2012), AMED (1985 to March 2012) and 11 Chinese databases (March 2012). In an effort to identify further published, unpublished and ongoing trials we searched ongoing trials registers, reference lists and relevant conference proceedings, and contacted authors and pharmaceutical companies.

Selection criteria

We included randomized controlled trials (RCTs) investigating GABA receptor agonists versus placebo for acute stroke patients (within 12 hours after stroke onset), with the outcomes of death or dependency, functional independence and adverse events.

Data collection and analysis

Two review authors independently screened the titles and abstracts of identified records, selected studies for inclusion, extracted eligible data, cross-checked the data for accuracy and assessed the methodological quality.

Main results

We included five trials with 3838 patients. The methodological quality of the included trials was generally good, with low risk of bias. Four trials measured death and dependency at three months in chlormethiazole versus placebo without significant difference (risk ratio (RR) 1.03, 95% confidence interval (CI) 0.95 to 1.11). One trial measured this outcome between diazepam and placebo (RR 0.94, 95% CI 0.82 to 1.07). In the subgroup analysis of total anterior circulation syndrome (TACS), a higher percentage of functional independence was found in the chlormethiazole group (RR 1.33, 95% CI 1.09 to 1.64). The frequent adverse events related to chlormethiazole were somnolence (RR 4.56, 95% CI 3.50 to 5.95) and rhinitis (RR 4.75, 95% CI 2.67 to 8.46).

Authors' conclusions

This review does not provide the evidence to support the use of GABA receptor agonists (chlormethiazole or diazepam) for the treatment of patients with acute ischemic or hemorrhagic stroke. Chlormethiazole appeared to be beneficial in improving functional independence in patients with TACS according to the subgroup analysis, but this result must be interpreted with great caution. More well-designed RCTs with large samples of TACS would be required for further confirmation. However, somnolence and rhinitis are frequent adverse events related to chlormethiazole.

Résumé

Agonistes des récepteurs de l'acide gamma-aminobutyrique (GABA) pour le traitement de l'AVC aigu

Contexte

Il a été montré que les agonistes des récepteurs de l'acide gamma-aminobutyrique (GABA) ont un effet neuroprotecteur puisqu'ils réduisent la taille des infarctus et qu'ils améliorent le résultat fonctionnel dans les modèles animaux d'ischémie cérébrale. Cependant, les effets sédatifs des agonistes des récepteurs du GABA ont limité leur plus large application chez les patients victimes d'un AVC aigu en raison du risque potentiel de stupeur.

Objectifs

Déterminer l'efficacité et l'innocuité des agonistes des récepteurs du GABA dans le traitement de l'AVC aigu.

Stratégie de recherche documentaire

Nous avons effectué une recherche dans le registre des essais du groupe Cochrane sur les accidents vasculaires cérébraux (janvier 2012), le registre Cochrane des essais contrôlés (CENTRAL) (The Cochrane Library 2012, numéro 1), MEDLINE (de 1949 à mars 2012), EMBASE (de 1980 à mars 2012), CINAHL (de 1982 à mars 2012), AMED (de 1985 à mars 2012) et 11 bases de données chinoises (mars 2012). Dans un effort visant à identifier les autres essais publiés, non publiés et en cours, nous avons effectué des recherches dans les registres des essais en cours, les listes bibliographiques et les résumés de conférences pertinents, et nous avons contacté les auteurs et les laboratoires pharmaceutiques.

Critères de sélection

Nous avons inclus les essais contrôlés randomisés (ECR) étudiant les agonistes des récepteurs du GABA comparés à un placebo chez les patients victimes d'un AVC aigu (dans les 12 heures suivant l'apparition de l'AVC), présentant les résultats de décès ou de dépendance, d'indépendance fonctionnelle et d'événements indésirables.

Recueil et analyse des données

Deux auteurs de la revue ont indépendamment examiné les titres et les résumés des dossiers identifiés, sélectionné les études à inclure, extrait les données éligibles, recoupé les données pour en vérifier l'exactitude et évalué la qualité méthodologique.

Résultats Principaux

Nous avons inclus cinq essais portant sur 3 838 patients. La qualité méthodologique des essais inclus était généralement bonne, le risque de biais étant faible. Quatre essais ont mesuré les décès et la dépendance après trois mois de traitement par chlorméthiazole versus placebo sans différence significative (risque relatif (RR) 1,03, intervalle de confiance (IC) à 95 % 0,95 à 1,11). Un essai a mesuré ce résultat entre le diazépam et un placebo (RR 0,94, IC à 95 % 0,82 à 1,07). Dans l'analyse en sous-groupe du syndrome de la circulation antérieure totale (SCAT), un pourcentage plus élevé d'indépendance fonctionnelle a été observé dans le groupe recevant du chlorméthiazole (RR 1,33, IC à 95 % 1,09 à 1,64). Les événements indésirables fréquents associés au chlorméthiazole étaient la somnolence (RR 4,56, IC à 95 % 3,50 à 5,95) et la rhinite (RR 4,75, IC à 95 % 2,67 à 8,46).

Conclusions des auteurs

Cette revue ne fournit pas de preuves permettant de corroborer l'utilisation des agonistes des récepteurs du GABA (chlorméthiazole ou diazépam) dans le traitement des patients ayant subi un AVC ischémique ou hémorragique aigu. Le chlorméthiazole est apparu comme étant bénéfique pour améliorer l'indépendance fonctionnelle chez les patients atteints du SCAT conformément à l'analyse en sous-groupe, mais ce résultat doit être interprété avec une grande prudence. Des ECR bien conçus supplémentaires avec de grands échantillons de SCAT sont nécessaires pour confirmer les conclusions. Cependant, la somnolence et la rhinite sont des événements indésirables fréquents associés au chlorméthiazole.

Plain language summary

Gamma aminobutyric acid (GABA) receptor agonists for acute stroke

GABA receptor agonists are a type of neuroprotective agent that may help protect the brain in the treatment of acute stroke. This class of drugs, which includes diazepam and chlormethiazole, are traditional sedatives that have been used for several decades and have been found to be beneficial in animal models of stroke. However, the sedation effect of GABA receptor agonists can be harmful for acute stroke patients with the potential risk of causing stupor. We included five trials with 3838 patients to assess the benefit and safety of GABA receptor agonists for patients with acute stroke. In conclusion, there was no convincing evidence to support the use of GABA receptor agonists for the treatment of patients with acute ischemic or hemorrhagic stroke. The most frequent adverse events of chlormethiazole were drowsiness and rhinitis.

Résumé simplifié

Agonistes des récepteurs de l'acide gamma-aminobutyrique (GABA) pour le traitement de l'AVC aigu

Agonistes des récepteurs de l'acide gamma-aminobutyrique (GABA) pour le traitement de l'AVC aigu

Les agonistes des récepteurs du GABA sont un type d'agent neuroprotecteur qui peut contribuer à la protection du cerveau lors du traitement de l'AVC aigu. Les médicaments de cette classe, qui comprennent le diazépam et le chlorméthiazole, sont des sédatifs traditionnels qui sont utilisés depuis plusieurs décennies et qui se sont avérés bénéfiques dans les modèles animaux d'AVC. Cependant, l'effet sédatif des agonistes des récepteurs du GABA peut être nocif chez les patients victimes d'un AVC aigu et présente un risque potentiel de provoquer une stupeur. Nous avons inclus cinq essais portant sur 3 838 patients afin d'évaluer les bénéfices et l'innocuité des agonistes des récepteurs du GABA chez les patients victimes d'un AVC aigu. En conclusion, il n'existait aucune preuve convaincante permettant de corroborer l'utilisation des agonistes des récepteurs du GABA dans le traitement des patients ayant subi un AVC ischémique ou hémorragique aigu. Les événements indésirables les plus fréquents du chlorméthiazole étaient la somnolence et la rhinite.

Notes de traduction

Traduit par: French Cochrane Centre 1st March, 2013
Traduction financée par: Pour la France : Ministère de la Santé. Pour le Canada : Instituts de recherche en santé du Canada, ministère de la Santé du Québec, Fonds de recherche de Québec-Santé et Institut national d'excellence en santé et en services sociaux.

Background

Description of the condition

Acute stroke is defined as a clinical syndrome of sudden onset of focal or global disturbance of central nervous system function due to an interruption of the cerebral circulation (Warlow 2001). Ischemic stroke (80% of all strokes) is the most frequent type, followed by intracerebral hemorrhage (15%) and subarachnoid hemorrhage (5%). The estimated annual incidence of stroke is 0.25% and increases with ageing (Simon 2009; WHO 2011). Two-thirds of all strokes occur in those older than 65 years (AHA 2002). The common risk factors include smoking, hypertension, diabetes, carotid stenosis, hypercholesterolemia, hyperhomocysteinemia, alcohol abuse and a high-fat diet (Cruz-Flores 2011). The prognosis is poor, with one-third of patients dying and one-third left with permanent disability (WHO 2011). In addition, stroke is a costly condition incurring treatment, care and indirect costs (Saka 2009).

Description of the intervention

Neuroprotective agents have attracted a lot of attention for the treatment of acute stroke, and are expected to be helpful in protecting vulnerable neurons and salvaging the ischemic penumbra (ischemic but still viable tissue). The most common neuroprotectants include excitatory amino acid antagonists, gangliosides, calcium channel antagonists, lubeluzole, methylxanthine derivatives and tirilazad - each with different modes of action. It is disappointing that none of these treatments has been confirmed to be effective in the acute phase of stroke (Bath 2001; Bath 2004; Candelise 2001; Gandolfo 2002; Horn 2000; Muir 2003). Therefore, it is necessary to examine other potential neuroprotectants for stroke. Gamma aminobutyric acid (GABA) receptor agonists (e.g. diazepam and chlormethiazole) are traditional sedatives that have been used for several decades. They have also been found to be effective in reducing infarct size in histology and improving functional outcome in animal models of cerebral ischemia (Gasior 2004; Marshall 2003; Sydserff 2002).

How the intervention might work

GABA is the main inhibitory neurotransmitter in the central nervous system and acts by reducing the depolarization-induced and ischemia-induced glutamate release (Nelson 2000; Vaishnav 2002). Firstly, GABA can trigger hyperpolarization of neurons through anion channels (GABAA) and presynaptic G-protein coupled receptors (GABAB) (Wilby 2004). This hyperpolarization counteracts the depolarization which is the initiating event in the biochemical ischemic cascade (Tuttolomondo 2009). Secondly, there is no shortage of GABA in ischemic conditions, but the affinity of GABA receptors is decreased (Alicke 1995). Activation of GABA reduces respiratory rate, preserving glucose and reducing acidosis, which facilitates local cerebral blood flow (Chi 2011; Zubcevic 2010). Finally, GABA receptor agonists can induce hypothermia, which is also regarded as a neuroprotective condition for acute stroke (Klassman 2011; Visser 2005).

Why it is important to do this review

There are existing clinical trials on the effects of GABA receptor agonists in acute stroke based on the results of preclinical in vivo studies. However, conflicting results limit their wider application in sedation caused by GABA receptor agonists in acute stroke with edema (Hanna 1996). To date, there is no systematic review with meta-analysis of this subject and thus it is necessary to evaluate the efficacy and safety of GABA receptor agonists through high-quality randomized controlled trials.

Objectives

To determine the efficacy and safety of GABA receptor agonists in the treatment of acute stroke.

Methods

Criteria for considering studies for this review

Types of studies

All randomized controlled trials (RCTs) investigating GABA receptor agonists for acute stroke patients were eligible. We excluded quasi-randomized or confounded studies.

Types of participants

We included patients who suffered from acute stroke within 12 hours after stroke onset. There were no limitations in gender, age or subtype of stroke.

Types of interventions

Intervention: GABA receptor agonists administered orally or intravenously, regardless of length of treatment period and dosage of treatment, alone or combined with other treatment.

Comparator: placebo.

We included other concomitant therapies providing that they were administered in both the intervention and control groups.

Types of outcome measures

We assessed the following outcomes measured at three-month follow-up.

Primary outcomes
Efficacy

Death or dependency at the end of long-term follow-up (at least three months). We defined dependency as a Barthel Index (BI) score of 60 or less, or the modified Rankin Scale (mRS) graded 3 to 5 (Sulter 1999), or used the definition provided by the researchers.

Safety

The number of patients with adverse events

Secondary outcomes
  1. Functional independence, defined as BI score > 60, or mRS < 3

  2. Neurological function measured by other stroke scales, e.g. National Institutes of Health Stroke Scale (NIHSS) or Scandinavian Stroke Scale (SSS)

Search methods for identification of studies

See the 'Specialized register' section in the Cochrane Stroke Group module. We searched for relevant trials in all languages and arranged translation of relevant papers published in languages other than English.

Electronic searches

We searched the Cochrane Stroke Group Trials Register (January 2012) and the following electronic bibliographic databases:

  • Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2012, Issue 1);

  • MEDLINE (1949 to March 2012) (Appendix 1);

  • EMBASE (1980 to March 2012);

  • CINAHL (1982 to March 2012);

  • AMED (Allied and Complementary Medicine Database) (1985 to March 2012);

  • Chinese Stroke Trials Register (March 2012);

  • CBM-disc (China Biological Medicine Databases) (1979 to March 2012);

  • CNKI (China National Knowledge Infrastructure) (1979 to March 2012);

  • Chinese MD and DD Dissertations in CNKI (March 2012);

  • CACP (Chinese Academic Conference Papers Database) (1998 to March 2012);

  • CDDB (Chinese Dissertations Database) (1977 to March 2012);

  • Chinese Evidence-Based Medicine Database (March 2012);

  • CMAC (China Medical Academic Conferences) (1994 to March 2012);

  • CMCC (Chinese Medical Current Contents) (1994 to March 2012);

  • Chinese Science and Technique Journals Database (VIP) (1989 to March 2012);

  • Wanfang Data (www.wanfangdata.com/) (1984 to March 2012).

We developed the MEDLINE search strategy with the help of the Cochrane Stroke Group Trials Search Co-ordinator and adapted it for the other databases.

To identify further published, unpublished and ongoing trials we searched the following trials registers in March 2012:

Searching other resources

We also:

  1. used Science Citation Index Cited Reference Search for forward tracking of important articles;

  2. searched reference lists of reviews and retrieved articles;

  3. searched conference proceedings, including the 6th to 20th European Stroke Conference (from 1997 to 2011) and the 4th, 5th, 6th and 7th World Stroke Congress (2000, 2004, 2008 and 2010);

  4. contacted authors where necessary for missing information;

  5. contacted the manufacturer (AstraZeneca pharmaceutical company) for updated information.

Data collection and analysis

Selection of studies

Two review authors (LJ, WL) independently screened titles and abstracts of the citations produced by the database searches and excluded obviously irrelevant studies. We obtained the full text of all remaining studies and the same two authors independently selected studies that met the inclusion criteria. Both authors evaluated the eligibility and assessed the methodological quality of these studies. We resolved any disagreements by discussion, or referred them to an independent party if necessary.

Data extraction and management

Two review authors (LJ, WL) independently extracted eligible data from the published reports onto pre-standardized forms, and cross-checked them for accuracy. We used checklists to independently record relevant details including methods of generating randomization schedule, method of concealment of allocation, blinding of assessors, intention-to-treat analysis, adverse events and drop outs for all reasons, important imbalance in prognostic factors, participants (socio-demographic and related clinical information), interventions (medications and non-pharmacological interventions) and outcomes. We resolved disagreements by consensus .

Assessment of risk of bias in included studies

Two review authors (LJ, WL) independently assessed methodological quality in terms of method of randomization, blinding of participant and outcome assessor, completeness of follow-up and whether intention-to-treat analysis was carried out or was possible from the published data according to the Cochrane Collaboration's tool for assessing risk of bias (Higgins 2011). We assessed whether the treatment groups were comparable with regards to demographics, clinical characteristics and the number of patients excluded or lost to follow-up within each trial and whether the definition of outcomes and entry and exclusion criteria were comparable across trials. We considered sources of bias on a study-by-study basis. We specified the reasons for exclusion where appropriate.

Measures of treatment effect

We expressed results for dichotomous outcomes as risk ratios (RR) with 95% confidence intervals (CI). We entered and analyzed data in RevMan 5.1 (Review Manager 2011).

Unit of analysis issues

We dealt with any unit of analysis issues according to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

Dealing with missing data

We attempted to contact the authors of the studies for further details if any data were missing and to establish the characteristics of unpublished trials through correspondence with trial co-ordinators or principal investigators.

Assessment of heterogeneity

We tested heterogeneity using the I2 statistic and made a judgement as to whether significant heterogeneity was present (Higgins 2011). We took I2 values over 50% as suggestive of substantial heterogeneity. However, the direction and magnitude of effects was taken into account.

Assessment of reporting biases

We planned to use the funnel plot method (Egger 1997), if there were sufficient numbers of trials to allow for a meaningful presentation.

Data synthesis

If a sufficient number of comparable studies with a low risk of bias were available, we carried out meta-analyses routinely by a fixed-effect model. If we found statistically significant heterogeneity, we calculated the overall effects by a random-effects model instead of a fixed-effect model. If substantial heterogeneity between the studies prevented us from combining outcome data, we gave a descriptive summary of the results.

Subgroup analysis and investigation of heterogeneity

We analyzed subgroups of studies categorized by the type of stroke (ischemic or hemorrhage) and time from stroke onset to treatment administration.

Sensitivity analysis

We undertook sensitivity analyses to assess the robustness of results in fixed-effect versus random-effects models, and studies at high risk versus low risk of bias. We also used these sensitivity analyses to examine potential sources of methodological heterogeneity.

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies.

Results of the search

From our searches of the electronic databases and after removing duplicates we identified a total of 782 unique references. After screening the titles and abstracts, we obtained the full papers of 16 articles and assessed these for eligibility (Figure 1). Agreement between the review authors on exclusion was 100%. We found no relevant studies in the Chinese databases and no relevant ongoing trials from the searches of the trials registers.

Figure 1.

Study flow diagram.

Included studies

We included five studies, with 3838 patients, according to the criteria. One study (EGASIS) evaluated the efficacy and safety of diazepam for acute stroke in 880 patients within 12 hours of stroke onset. The Chlormethiazole Acute Stroke Study (CLASS) mainly focused on the efficacy and safety of chlormethiazole in 1360 acute stroke patients within 12 hours of stroke onset: 95 patients had hemorrhagic stroke and were analyzed separately. A subgroup analysis of total anterior circulation syndrome (TACS) was undertaken in 545 patients. After the completion of CLASS, another chlormethiazole acute stroke study in ischemic, hemorrhagic and t-PA treated stroke (CLASS-IHT) was designed. All participants were included within 12 hours after stroke onset. There were 1198 patients randomized in CLASS-I, 201 in CLASS-H and 200 in CLASS-T. We have provided relevant information about the included trials in the Characteristics of included studies table.

Excluded studies

We excluded eight studies after full-text evaluation. We have provided the reasons for exclusion in the Characteristics of excluded studies table.

Risk of bias in included studies

Allocation

All the included trials stated that participants were randomized into intervention and placebo groups, but only one trial described the actual method of randomization and allocation concealment. Therefore, we judged random sequence generation and allocation concealment of the trials as unclear risk of bias.

Blinding

All the participants and investigators were blinded to trial medication in the included trials. We assessed blinding as low risk of bias.

Incomplete outcome data

All the five studies reported the reason and number of participants who had discontinued treatment at the endpoint. Only one trial stated an intention-to-treat analysis (EGASIS). In general, 80 of 3838 (2%) randomized participants were not included in the efficacy analysis. Therefore, we assessed this as low risk of bias.

Selective reporting

In the included trials, all the pre-specified outcomes were reported. Therefore, we assessed this as low risk of bias.

Other potential sources of bias

We did not find any potential publication bias. Insufficient numbers of trials were available for a funnel plot analysis.

Effects of interventions

Primary outcome measures

Death or dependency at the end of long-term follow-up (at least three months)

For acute stroke, four trials (CLASS; CLASS-H; CLASS-I; CLASS-T) with 2909 patients reported death and dependency at three months after chlormethiazole or placebo administration: 710 of 1457 (49%) and 689 of 1452 (47%) deaths or dependencies occurred in the chlormethiazole and placebo groups respectively (RR 1.03, 95% CI 0.95 to 1.11). One trial (EGASIS), with 849 patients, reported this outcome between diazepam and placebo, in which 206 of 428 (48%) and 216 of 421 (51%) deaths or dependencies occurred in the diazepam and placebo groups respectively (RR 0.94, 95% CI 0.82 to 1.07). In total, 916 of 1885 (49%) patients in the gamma aminobutyric acid (GABA) receptor agonists group and 905 of 1873 (48%) patients in the placebo group experienced deaths or dependencies (RR 1.01, 95% CI 0.94 to 1.07) (Analysis 1.1).

Adverse events

Five trials reported serious adverse events (SAEs). Four trials (CLASS; CLASS-H; CLASS-I; CLASS-T) compared chlormethiazole with placebo, and one trial (EGASIS) compared diazepam versus placebo. No significant differences were found in either the number of patients with SAEs or the number of SAEs in all trials.

The more frequent adverse events in the chlormethiazole group included somnolence and rhinitis. Two trials reported the information (CLASS; CLASS-I): in the chlormethiazole and placebo groups, 654 of 1266 (52%) patients and 142 of 1261 (11%) patients respectively experienced somnolence (RR 4.56, 95% CI 3.50 to 5.95), while 165 of 1266 (13%) patients and 32 of 1261 (3%) patients experienced rhinitis (RR 4.75, 95% CI 2.67 to 8.46). In these two outcomes, there was significant heterogeneity (I2 = 62% and 53% respectively). Therefore, we used a random-effects model for the analyses (Analysis 1.2; Analysis 1.3).

Secondary outcomes measures

Functional independence

Four trials (CLASS; CLASS-H; CLASS-I; CLASS-T) reported functional independence, which was defined as a Barthel Index (BI) score more than 60. All four trials compared chlormethiazole with placebo, with 747 of 1457 (51%) patients and 763 of 1452 (53%) patients respectively scoring more than 60 on the BI score (RR 0.98, 95% CI 0.91 to 1.05). The EGASIS trial, of diazepam versus placebo, reported functional independence based on the mRS, with 222 of 428 (52%) and 205 of 421 (49%) respectively having a mRS score of less than 3 (RR 1.07, 95% CI 0.93 to 1.22) (Analysis 1.4).

National Institutes of Health Stroke Scale (NIHSS) score

We intended to calculate the mean change of NIHSS score. However, only two trials (CLASS-H; CLASS-I) reported this and neither of them gave it as 'mean (SD)'. In CLASS-H, the mean change of NIHSS score was -4.5 in the chlormethiazole group (N = 96) and -4.0 in the placebo group (N = 102) (P = 0.36). In CLASS-I, the change of NIHSS score (median (quartiles)) was -5.5 (-11, 17) in the chlormethiazole group (N = 586) and -6.0 (-10, 16) in the placebo group (N = 583) (P = 0.68).

Scandinavian Stroke Scale (SSS) score

The results of the SSS score were given by three trials (CLASS; CLASS-H; CLASS-I). In CLASS, no significant difference was found between the placebo and chlormethiazole groups for the score change in the 48-point Scandinavian Stroke Scale (SSS-48) (P = 0.56) and the Scandinavian Stroke Scale motor power score (SSS-MP) (P = 0.96). In CLASS-H and CLASS-I, the change of score in the SSS was not significant in the two groups (P = 0.06 and P = 0.23, respectively).

Subgroup analysis

Efficacy for acute ischemic stroke

For acute ischemic stroke, three trials (CLASS; CLASS-I; CLASS-T) compared death and dependency at three months in the chlormethiazole and placebo groups; 652 of 1327 (49%) and 628 of 1319 (48%) deaths or dependencies were found respectively (RR 1.03, 95% CI 0.96 to 1.12) (Analysis 2.1). The same three trials reported functional independence, with 662 of 1327 (50%) patients in the chlormethiazole group and 675 of 1319 (51%) patients in the placebo group (RR 0.97, 95% CI 0.90 to 1.05). One other trial (EGASIS) tested functional independence, with 203 of 380 (53%) patients in the diazepam group, and 178 of 368 (48%) patients in the placebo group (RR 1.10, 95% CI 0.96 to 1.27) (Analysis 2.2).

Efficacy for acute hemorrhagic stroke

For acute hemorrhagic stroke, two trials (CLASS; CLASS-H) compared death and dependency at three months in the chlormethiazole and placebo groups; 58 of 143 (41%) and 61 of 149 (41%) deaths or dependencies were estimated respectively (RR 0.99, 95% CI 0.75 to 1.30) (Analysis 3.1). Functional independence was found: 85 of 143 (59%) in the chlormethiazole group and 88 of 149 (59%) in the placebo group (RR 1.01, 95% CI 0.83 to 1.22). In addition, EGASIS, which tested diazepam versus placebo, found that 18 of 46 (39%) and 24 of 49 (49%) respectively were functionally independent (RR 0.80, 95% CI 0.50 to 1.27) (Analysis 3.2).

Efficacy for total anterior circulation syndrome (TACS)

For TACS stroke patients, two trials (CLASS; CLASS-T) investigated functional independence at three months in the chlormethiazole and placebo groups. In total, 144 of 338 (43%) patients and 94 of 297 (32%) patients respectively were found to be functionally independent (RR 1.33, 95% CI 1.09 to 1.64) (Analysis 4.1).

Efficacy for early-treated acute stroke

We extracted data for early-treated acute stroke. Two trials (CLASS; CLASS-I) focused on functional independence at three months in those patients treated within six hours of stroke onset, with 267 of 590 (45%) patients in the chlormethiazole group and 282 of 592 (48%) patients in the placebo group (RR 0.93, 95% CI 0.73 to 1.19). EGASIS defined early treatment as within three hours from onset and compared functional independence at three months in the diazepam and placebo groups: 37 of 70 (53%) patients and 27 of 62 (44%) patients respectively were estimated to be functionally independent (RR 1.21, 95% CI 0.85 to 1.74) (Analysis 5.1).

Sensitivity analysis

Fixed-effect versus random-effects models

For those outcomes with significant heterogeneity (I2 > 50%), we used a random-effects model to analyze the data. We also assessed the robustness of results in fixed-effect versus random-effects models. No changes of outcomes were found.

Excluding studies with potential selection bias

Only one trial (EGASIS) clearly described random sequence generation and allocation concealment, and it was the only trial of diazepam for acute stroke, while the other trials tested chlormethiazole. Thus, the data were inadequate to conduct a sensitivity analysis.

Discussion

Summary of main results

According to the inclusion criteria, we analyzed five eligible trials with 3838 patients in our review. The methodological quality of the included trials was generally good, with low risk of bias. However, there was no convincing evidence to support the use of gamma aminobutyric acid (GABA) receptor agonists (chlormethiazole or diazepam) for the treatment of patients with acute ischemic or hemorrhagic stroke.

Regarding the primary outcomes, four trials reported death and dependency at three months in the chlormethiazole versus placebo groups with no significant difference (RR 1.03, 95% CI 0.95 to 1.11). One trial reported this outcome between diazepam and placebo (RR 0.94, 95% CI 0.82 to 1.07).

The most frequent adverse events caused by chlormethiazole were somnolence and rhinitis. Two trials reported these outcomes in the chlormethiazole and placebo groups, with somnolence (RR 4.56, 95% CI 3.50 to 5.95) and rhinitis (RR 4.75, 95% CI 2.67 to 8.46).

Overall completeness and applicability of evidence

All the included studies demonstrated no benefits from GABA receptor agonists (chlormethiazole or diazepam) for acute stroke patients compared with placebo; that is, neither a decrease in death or dependency, nor an increase in functional independence were found. The subgroup analysis in total anterior circulation syndrome (TACS) illustrated a positive result based on only two studies. Meanwhile, the subgroup analysis in acute hemorrhagic stroke did not find any increase in death or dependency, which meant GABA receptor agonists appeared to cause no harm in hemorrhagic stroke when compared with placebo. Readers should note these conclusions need further confirmation by more randomized controlled trials (RCTs) with large samples. At present, no clear evidence supports the clinical application of GABA receptor agonists in any type of acute stroke.

Quality of the evidence

The evidence from this review is credible because it was derived from different multicenter RCTs with large samples. The search methods were rigorous and well performed. The methodological quality of the included trials was generally good, with low risk of bias. We also specified the reasons for excluding trials. However, the conclusions from subgroup analyses should be interpreted with caution, due to the limitations of the available data.

Potential biases in the review process

For subgroup analysis, some data were not provided. For instance, the CLASS study consisted of 1360 acute stroke patients: 7% of them (95 patients) had a hemorrhagic stroke. However, there were no independent data for ischemic stroke available. We therefore used subtraction to indirectly get the ischemic stroke data for the dichotomous outcomes. This may lead to potential bias.

Agreements and disagreements with other studies or reviews

We found only one review (Zingmark 2003) on this topic, which mainly focused on the pharmacokinetics of chlormethiazole in acute stroke patients. Therefore, it is not comparable, considering the different objectives and outcomes.

Authors' conclusions

Implications for practice

This review does not provide the evidence to support the use of gamma aminobutyric acid (GABA) receptor agonists (chlormethiazole or diazepam) for the treatment of patients with acute ischemic or hemorrhagic stroke. Based on a subgroup analysis, chlormethiazole may be beneficial for improving functional independence in patients with total anterior circulation syndrome (TACS). However, any subgroup analysis must be interpreted with great caution, and more trials would be required to confirm this result. Somnolence and rhinitis seem to be the most frequent adverse events related to chlormethiazole, and further investigations are required.

Implications for research

Well-designed, double-blind RCTs would be required to test the efficacy of chlormethiazole in a large group of patients with TACS.

Acknowledgements

The authors would like to acknowledge the help provided by the Cochrane Stroke Group.

Data and analyses

Download statistical data

Comparison 1. Efficacy and safety for acute stroke
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Death or dependency53758Risk Ratio (M-H, Fixed, 95% CI)1.01 [0.94, 1.07]
1.1 Chlormethiazole versus placebo42909Risk Ratio (M-H, Fixed, 95% CI)1.03 [0.95, 1.11]
1.2 Diazepam versus placebo1849Risk Ratio (M-H, Fixed, 95% CI)0.94 [0.82, 1.07]
2 Somnolence22527Risk Ratio (M-H, Random, 95% CI)4.56 [3.50, 5.95]
2.1 Chlormethiazole versus placebo22527Risk Ratio (M-H, Random, 95% CI)4.56 [3.50, 5.95]
3 Rhinitis22527Risk Ratio (M-H, Random, 95% CI)4.75 [2.67, 8.46]
3.1 Chlormethiazole versus placebo22527Risk Ratio (M-H, Random, 95% CI)4.75 [2.67, 8.46]
4 Functional independence53758Risk Ratio (M-H, Fixed, 95% CI)0.99 [0.94, 1.06]
4.1 Chlormethiazole versus placebo42909Risk Ratio (M-H, Fixed, 95% CI)0.98 [0.91, 1.05]
4.2 Diazepam versus placebo1849Risk Ratio (M-H, Fixed, 95% CI)1.07 [0.93, 1.22]
Analysis 1.1.

Comparison 1 Efficacy and safety for acute stroke, Outcome 1 Death or dependency.

Analysis 1.2.

Comparison 1 Efficacy and safety for acute stroke, Outcome 2 Somnolence.

Analysis 1.3.

Comparison 1 Efficacy and safety for acute stroke, Outcome 3 Rhinitis.

Analysis 1.4.

Comparison 1 Efficacy and safety for acute stroke, Outcome 4 Functional independence.

Comparison 2. Efficacy for acute ischemic stroke
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Death or dependency32646Risk Ratio (M-H, Fixed, 95% CI)1.03 [0.96, 1.12]
1.1 Chlormethiazole versus placebo32646Risk Ratio (M-H, Fixed, 95% CI)1.03 [0.96, 1.12]
2 Functional independence43394Risk Ratio (M-H, Fixed, 95% CI)1.00 [0.94, 1.07]
2.1 Chlormethiazole versus placebo32646Risk Ratio (M-H, Fixed, 95% CI)0.97 [0.90, 1.05]
2.2 Diazepam versus placebo1748Risk Ratio (M-H, Fixed, 95% CI)1.10 [0.96, 1.27]
Analysis 2.1.

Comparison 2 Efficacy for acute ischemic stroke, Outcome 1 Death or dependency.

Analysis 2.2.

Comparison 2 Efficacy for acute ischemic stroke, Outcome 2 Functional independence.

Comparison 3. Efficacy for acute hemorrhagic stroke
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Death or dependency2292Risk Ratio (M-H, Fixed, 95% CI)0.99 [0.75, 1.30]
1.1 Chlormethiazole versus placebo2292Risk Ratio (M-H, Fixed, 95% CI)0.99 [0.75, 1.30]
2 Functional independence3387Risk Ratio (M-H, Fixed, 95% CI)0.96 [0.81, 1.15]
2.1 Chlormethiazole versus placebo2292Risk Ratio (M-H, Fixed, 95% CI)1.01 [0.83, 1.22]
2.2 Diazepam versus placebo195Risk Ratio (M-H, Fixed, 95% CI)0.80 [0.50, 1.27]
Analysis 3.1.

Comparison 3 Efficacy for acute hemorrhagic stroke, Outcome 1 Death or dependency.

Analysis 3.2.

Comparison 3 Efficacy for acute hemorrhagic stroke, Outcome 2 Functional independence.

Comparison 4. Efficacy for TACS
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Functional independence2635Risk Ratio (M-H, Fixed, 95% CI)1.33 [1.09, 1.64]
1.1 Chlormethiazole versus placebo2635Risk Ratio (M-H, Fixed, 95% CI)1.33 [1.09, 1.64]
Analysis 4.1.

Comparison 4 Efficacy for TACS, Outcome 1 Functional independence.

Comparison 5. Efficacy for early-treated acute stroke
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Functional independence31314Risk Ratio (M-H, Random, 95% CI)0.99 [0.80, 1.21]
1.1 Chlormethiazole versus placebo (< 6 hours)21182Risk Ratio (M-H, Random, 95% CI)0.93 [0.73, 1.19]
1.2 Diazepam versus placebo (< 3 hours)1132Risk Ratio (M-H, Random, 95% CI)1.21 [0.85, 1.74]
Analysis 5.1.

Comparison 5 Efficacy for early-treated acute stroke, Outcome 1 Functional independence.

Appendices

Appendix 1. MEDLINE (Ovid) search strategy

1. cerebrovascular disorders/ or exp basal ganglia cerebrovascular disease/ or exp brain ischemia/ or exp carotid artery diseases/ or exp cerebrovascular trauma/ or exp intracranial arterial diseases/ or exp intracranial arteriovenous malformations/ or exp "intracranial embolism and thrombosis"/ or exp intracranial hemorrhages/ or stroke/ or exp brain infarction/ or vasospasm, intracranial/
2. (stroke or poststroke or post-stroke or cerebrovasc$ or brain vasc$ or cerebral vasc$ or cva$ or apoplex$ or SAH).tw.
3. ((brain$ or cerebr$ or cerebell$ or intracran$ or intracerebral) adj5 (isch?emi$ or infarct$ or thrombo$ or emboli$ or occlus$)).tw.
4. ((brain$ or cerebr$ or cerebell$ or intracerebral or intracranial or subarachnoid) adj5 (haemorrhage$ or hemorrhage$ or haematoma$ or hematoma$ or bleed$)).tw.
5. ((transi$ adj3 isch?em$ adj3 attack$) or TIA$1).tw.
6. 1 or 2 or 3 or 4 or 5
7. gaba agonists/ or exp gaba-a receptor agonists/ or exp gaba-b receptor agonists/ or exp gaba modulators/
8. exp gamma-Aminobutyric Acid/tu [Therapeutic Use]
9. exp Receptors, GABA/de [Drug Effects]
10. ((gamma aminobutyric acid or gamma-aminobutyric acid or gaba or gaba-A or gaba-B) adj5 (agonist$ or modulator$ or stimulat$ or stimulant$)).tw.
11. ((gabaergic or gaba-ergic or gabamimetic) adj5 (agent$ or drug$ or stimul$)).tw.
12. (Adipiplon or Alprazolam or Amobarbital or Arbaclofen or Atagabalin or AZD3355 or AZD9343 or Baclofen or Barbital or Benzodiazepine$ or Bromazepam or Chlordiazepoxide or Chlormethiazole or Clomethiazole or Clonazepam or Clorazepate or Dipotassium or Diazepam or Dihydromuscimol or Estazolam or Fengabine or Flumazenil or Flunitrazepam or Flurazepam or Gabapentin or Gaboxadol or Hexobarbital or Hopantenate calcium or Lesogaberan or Lorazepam or Medazepam or Mephobarbital or Midazolam or muscimol or Nitrazepam or Nordazepam or Oxazepam or Oxybate sodium or Pagoclone or Pentobarbital or Phenobarbital or Picamilon or Prazepam or Pregabalin or Progabide or Secobarbital or Temazepam or Thiamylal or Thiopental or THIP or TPA023 or Triazolam or Valproic acid or Vigabatrin or Zolazepam or XP19986).tw,nm.
13. 7 or 8 or 9 or 10 or 11 or 12
14. 6 and 13
15. Randomized Controlled Trials as Topic/
16. random allocation/
17. Controlled Clinical Trials as Topic/
18. control groups/
19. clinical trials as topic/ or clinical trials, phase i as topic/ or clinical trials, phase ii as topic/ or clinical trials, phase iii as topic/ or clinical trials, phase iv as topic/
20. double-blind method/
21. single-blind method/
22. Placebos/
23. placebo effect/
24. Drug Evaluation/
25. Research Design/
26. randomized controlled trial.pt.
27. controlled clinical trial.pt.
28. (clinical trial or clinical trial phase i or clinical trial phase ii or clinical trial phase iii or clinical trial phase iv).pt.
29. random$.tw.
30. (controlled adj5 (trial$ or stud$)).tw.
31. (clinical$ adj5 trial$).tw.
32. ((control or treatment or experiment$ or intervention) adj5 (group$ or subject$ or patient$)).tw.
33. (quasi-random$ or quasi random$ or pseudo-random$ or pseudo random$).tw.
34. ((singl$ or doubl$ or tripl$ or trebl$) adj5 (blind$ or mask$)).tw.
35. placebo$.tw.
36. controls.tw.
37. or/15-36
38. 14 and 37
39. exp animals/ not humans.sh.
40. 38 not 39

Contributions of authors

Liu J and Wang L formulated the idea for the review and developed the basis for the review. Liu J took the primary role in searching, identifying and assessing studies, in data extraction and analyses, and in writing up the full review. Wang L provided general advice on this review, as well as helping to identify trials, assess studies and extract data. Liu J supervised the quality of the methodology and statistics used. Both authors wrote or revised the full review, and Liu J will be responsible for updating the review.

Declarations of interest

None known.

Sources of support

Internal sources

  • None, Not specified.

External sources

  • None, Not specified.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

CLASS

MethodsRandomized, double-blind, multicenter, placebo-controlled study to test the efficacy and safety of the neuroprotective drug chlormethiazole for acute stroke
ParticipantsPatients aged 40 to 90 years were included with full consciousness before treatment. The symptoms should last more than 1 hour and less than 12 hours. SSS-48 of ≦ 40 with a sum of scores on arm, hand and leg motor items of ≦ 14. 1360 eligible patients from 85 clinical centers in 7 European countries and Canada had been randomized into the trial, in which 546 patients had TACS and 95 patients had hemorrhagic stroke
InterventionsChlormethiazole (75 mg/kg) or placebo were given as an intravenous infusion over a 24-hour period
OutcomesIndependence (BI ≧ 60); SSS-48; SSS-MP; adverse events; mortality
NotesFollow-up: 3 months
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskRandomization was stratified by center, but the method of random sequence generation was not described
Allocation concealment (selection bias)Low riskAll validations were made with the treatment allocation blinded
Blinding of participants and personnel (performance bias)
All outcomes
Low riskOnly the independent data monitoring committee had access to unblinded data during the course of the study
Blinding of outcome assessment (detection bias)
All outcomes
Low riskOnly the independent data monitoring committee had access to unblinded data during the course of the study
Incomplete outcome data (attrition bias)
All outcomes
Low risk16 of 1360 (1%) randomized patients discontinued at the endpoint. 4 patients did not receive treatment (1 randomized to chlormethiazole, 3 to placebo). 4 of 1360 (0.3%) randomized patients were not available in the safety analysis. In subgroup analyses, 1 of 95 (1%) randomized hemorrhagic stroke patients and 6 of 546 (1%) randomized TACS patients discontinued at the endpoint
Selective reporting (reporting bias)Low riskAll pre-specified outcomes were reported
Other biasLow riskNo other bias was found

CLASS-H

MethodsThe safety of chlormethiazole versus placebo in hemorrhagic stroke patients was evaluated in a randomized, double-blind trial
ParticipantsPatients aged 18 to 90 years with consciousness were included within 12 hours after stroke onset. 201 eligible patients were recruited and randomized into the trial
InterventionsChlormethiazole (68 mg/kg) or placebo was given as an intravenous infusion over a 24-hour period
OutcomesAdverse events; mortality; independence (BI ≧ 60 or mRS < 3); NIHSS; SSS-48
NotesFollow-up: 3 months
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskThe method of random sequence generation was not described
Allocation concealment (selection bias)Unclear riskThe allocation concealment was not reported
Blinding of participants and personnel (performance bias)
All outcomes
Low riskChlormethiazole and placebo were supplied in identical bottles to keep the treatment assignment blinded
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll the measurements were to be made by an assessor who was not involved during the administration of the study drug to maintain blinding of treatment assignment
Incomplete outcome data (attrition bias)
All outcomes
Low riskThe study drug was not administered to 1 patient in the chlormethiazole group and to 2 patients in the placebo group. Therefore, 3 of 201 (1%) patients were not included in the analysis of safety or efficacy
Selective reporting (reporting bias)Low riskAll pre-specified outcomes were reported
Other biasLow riskNo other bias was found

CLASS-I

MethodsA randomized, double-blind, multinational, placebo-controlled investigation of the efficacy and safety of chlormethiazole for acute ischemic stroke
ParticipantsPatients aged 18 to 90 years with consciousness were included within 12 hours after stroke onset. NIHSS score ≧ 3. 1198 eligible patients were recruited from 139 US and 14 Canadian centers and randomized into the trial
InterventionsChlormethiazole (68 mg/kg) or placebo was given as an intravenous infusion over a 24-hour period
OutcomesIndependence (BI ≧ 60 or mRS < 3); NIHSS; SSS-48; adverse events; mortality
NotesFollow-up: 3 months
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskThe method of random sequence generation was not described
Allocation concealment (selection bias)Low riskThe allocation was conducted by a central randomization scheme via telephone
Blinding of participants and personnel (performance bias)
All outcomes
Low riskChlormethiazole and placebo were supplied in identical bottles to keep the treatment assignment blinded
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll the measurements were to be made by an assessor who was not involved during the administration of the study drug to maintain blinding of treatment assignment
Incomplete outcome data (attrition bias)
All outcomes
Low risk29 of 1198 (2%) patients were not available in efficacy analysis. Treatment was never started in 27 patients: 12 in the chlormethiazole group and 15 in the placebo group. In addition, 2 patients (1 per group) provided no efficacy data but were included in the safety analysis
Selective reporting (reporting bias)Low riskAll pre-specified outcomes were reported
Other biasLow riskNo other bias was found

CLASS-T

MethodsA randomized, double-blind, multicenter, placebo-controlled study to explore the safety of t-PA combined with chlormethiazole
ParticipantsThere were 101 patients randomized to the chlormethiazole group and 99 to the placebo group by 76 of the 142 hospitals involved in the study
InterventionsAll patients received 0.9 mg/kg t-PA beginning within 3 hours of stroke onset and then either 68 mg/kg chlormethiazole (N = 97) IV over 24 hours or placebo (N = 93) beginning within 12 hours of stroke onset
OutcomesAdverse events; mortality; independence (BI ≧ 60)
NotesFollow-up: 3 months
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskThe stratified randomization was implemented and the method of random sequence generation was not described
Allocation concealment (selection bias)Unclear riskThe allocation concealment was not reported
Blinding of participants and personnel (performance bias)
All outcomes
Low riskChlormethiazole and placebo were supplied in identical bottles to keep the treatment assignment blinded
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll the measurements were to be made by an assessor who was not involved during the administration of the study drug to maintain blinding of treatment assignment
Incomplete outcome data (attrition bias)
All outcomes
Low riskAfter randomization, 10 of 200 (5%) patients (4 in the chlormethiazole group and 6 in the placebo group) did not receive the study drug and thus were not included in the safety analysis. All 10 patients showed signs of clinical deterioration after randomization before the study drug could be initiated
Selective reporting (reporting bias)Low riskAll pre-specified outcomes were reported
Other biasLow riskNo other bias was found

EGASIS

  1. a

    BI: Barthel Index score
    CT: computerized tomography
    MRI: magnetic resonance imaging
    mRS: modified Rankin Scale
    NIHSS: National Institutes of Health Stroke Scale
    SSS-48: 48-point Scandinavian Stroke Scale
    SSS-MP: Scandinavian Stroke Scale motor power score
    t-PA: tissue-type plasminogen activator
    TACS: total anterior circulation syndrome

MethodsA multicenter, randomized, stratified, double-blind, placebo-controlled clinical trial to examine the efficacy and safety of diazepam in acute stroke
ParticipantsAdult males and females were included within 12 hours after stroke onset. CT or MRI within 7 days was mandatory. Patients with a clear indication for or contraindication to benzodiazepines (at the discretion of the attending physician) were excluded, as were patients with unresponsive coma. 880 eligible patients from 35 hospitals in 5 European countries had been randomized into the trial
InterventionsDiazepam 10 mg or placebo by rectiole, as soon as possible, followed by 10 mg tablets twice daily for 3 days versus placebo
OutcomesIndependence (mRS < 3); complete recovery (BI ≧ 95 or mRS ≦ 1); adverse events; mortality
NotesFollow-up: 3 months
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskPatients were randomized using a computer-generated random listing of the 2 treatment assignments blocked in groups of 4 and stratified for center
Allocation concealment (selection bias)Low riskTrial medication was packed and labeled by the hospital's pharmacist according to a medication code schedule generated before the trial, and sent to the participating centers in boxes of 20 treatment packs
Blinding of participants and personnel (performance bias)
All outcomes
Low riskAll the patients were blinded to trial medication
Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll the investigators, treating physicians and nurses were blinded to trial medication
Incomplete outcome data (attrition bias)
All outcomes
Low risk31 participants (3.5%) discontinued after randomization with explicit reasons
Selective reporting (reporting bias)Low riskAll pre-specified outcomes were reported
Other biasLow riskAll efficacy and safety outcomes were analyzed by intention-to-treat

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
  1. a

    RCT: randomized controlled trial

Cucchiara 2003Not a RCT
Cucchiara 2004Not a RCT
Lodder 2000Not a RCT
Lyden 1998Neurological outcome of patients was not addressed
Lyden 2004Not a RCT
Millis 2007Not a RCT
van Raak 2002Neurological outcome of patients was not addressed
Wester 1998Not a RCT

Ancillary