Failure to respond to antiepileptic drugs in uncontrolled seizure activity such as refractory status epilepticus (RSE) has led to the use of anaesthetic drugs. Coma is induced with anaesthetic drugs to achieve complete control of seizure activity. Thiopental sodium and propofol are popularly used for this purpose. Both agents have been found to be effective. However, there is substantial lack of evidence as to which of the two drugs is better in terms of clinical outcome.
To compare the efficacy, adverse effects, and short- and long-term outcomes of RSE treated with one of the two anaesthetic agents, thiopental sodium or propofol.
We searched the Cochrane Epilepsy Group Specialized Register (10 May 2012), the Cochrane Central Register of Controlled Trials (CENTRAL Issue 4 of 12, The Cochrane Library 2012), and MEDLINE (1946 to May week 1, 2012). We also searched (10 May 2012) ClinicalTrials.gov, The South Asian Database of Controlled Clinical Trials, and IndMED (a bibliographic database of Indian Medical Journals).
All randomised or quasi-randomised controlled studies (regardless of blinding) of control of RSE using either thiopental sodium or propofol.
Data collection and analysis
Two review authors screened the search results and reviewed abstracts of relevant and eligible trials before retrieving the full text publications.
One study was available for review. This study was a small, single-blind, multicentre trial studying adults with RSE and receiving either propofol or thiopental sodium for the control of seizure activity (Rossetti 2011). This study showed a wide confidence interval suggesting that the drugs may differ in efficacy up to more than two-fold. There was no evidence of a difference between the drugs with respect to the outcome measures such as control of seizure activity and functional outcome at three months.
There is lack of robust and randomised controlled evidence that can clarify the efficacy of propofol and thiopental sodium over each other in the treatment of RSE. There is a need for large, randomised controlled trials for this serious condition.
Plain Language Summary
Propofol versus thiopental sodium for the treatment of refractory status epilepticus
Persistent convulsions are a major medical emergency associated with significant morbidity and mortality. At times, these convulsions fail to respond to first- and second-line drug therapy and may be observed in up to 31% of patients suffering from persistent seizure or convulsive activities. Persistent seizure activity may become refractory to antiepileptic drugs. Anaesthetics such as thiopental sodium and propofol are frequently given for control of seizures in such situations. Both agents have inherent side effects and complications. This review appraises the evidence for the these anaesthetic drugs in controlling seizure activity in patients with RSE.
There is a lack of good-quality evidence to support the choice of anaesthetic drug for the treatment of patients with RSE. Only one trial could be identified, which was terminated early due to recruitment problems. The two drugs did not show any difference among each other. The only difference noted was the requirement of prolonged mechanical ventilation for patients in the thiopental group. This could be due to the long half-life of the drug. There is a clear need for a large randomised controlled trial to study the efficacy of anaesthetic agents in the treatment of RSE.
Status epilepticus (SE) is defined as a condition in which there is either continuous seizure activity for more than 30 minutes, or two or more seizure activities in a sequence without return of full consciousness between the episodes (Prasad 2005; Working Group 1993). There is no universally accepted definition of SE. There is no consensus on the duration of seizure activity that may be required to define SE. SE may be broadly classified in two types, convulsive and non-convulsive. The common aetiologies for SE are stroke, traumatic brain injury, brain tumours, central nervous system infection, metabolic or toxic encephalopathies and electrolyte disorders. SE is a major medical emergency associated with significant morbidity and mortality (16% to 23%). Refractory status epilepticus (RSE) is defined as SE that fails to respond to first- and second-line therapy and it is observed in 9% to 31% of patients with SE (Mayer 2002; Treiman 1998). To be categorised as RSE, some authors have suggested a time frame (Mayer 2002), whereas others have not (Holtkamp 2007; Rossetti 2005). The first line of treatment for SE includes benzodiazepines; the second line includes antiepileptic drugs such as phenytoin, phenobarbital or valproic acid. Assistance of an anaesthetist is required for managing RSE, where coma may be induced with anaesthetic agents in order to achieve complete control of seizures. The use of anaesthetic agents such as thiopental sodium and propofol for managing RSE is common in many centres (Parviainen 2002; Van Gestel 2005). Thiopental sodium belongs to barbiturates group of drugs unrelated to propofol, which is a phenolic compound. Both agents have been found to be effective in controlling seizures in RSE (Parviainen 2002; Van Gestel 2005). There is substantial lack of evidence as to which of the two drugs, thiopental sodium and propofol, is better in terms of clinical outcome of patients with RSE. In a systematic review published in 2002 by Claassen 2002, the authors compared the efficacy of midazolam, propofol and pentobarbital for terminating seizures in RSE patients. Considering all possible limitations, the authors concluded that pentobarbital was more effective than any other strategy suggested for treatment of RSE.
Description of the condition
RSE develops when patients become resistant to antiepileptic drugs with passage of time. In hospital-based series, it develops in 31% to 44% of patients with SE (Mayer 2002). Significant morbidity and mortality are associated with RSE. Failure to respond to antiepileptic drugs has led to the use of anaesthetic agents in controlling seizures. The popular anaesthetic agents are the barbiturates, propofol and isoflurane.
Description of the intervention
Anaesthetic agents have been used for the treatment of RSE. Barbiturates and propofol have been commonly used in this regard. However, most of the published literature is anecdotal. There is no consensus as to which of the two agents is better in terms of clinical outcome.
How the intervention might work
Thiopental sodium, a barbiturate, is a γ-aminobutyric acid-A (GABAA) agonist with possible actions on calcium channels (Rogowski 2004). Barbiturates have a prolonged duration of action, mainly due to their accumulation in the body. They are also known to produce hypotension during use. In contrast, propofol is gaining popularity because of its shorter duration of action and little tendency to accumulate in the body. Similar to barbiturates, propofol also produces hypotension, and reduces intracranial pressure and brain metabolic requirements (Marik 2004). Prolonged use of propofol as an infusion has been shown to result in potentially fatal cardiovascular collapse associated with lactic acidosis, hypertriglyceridaemia and rhabdomyolysis, the so-called "propofol infusion syndrome" (Zarovnaya 2007). Both agents are also N-methyl-D-aspartate (NMDA) antagonists in vitro (Zhan 2001).
Why it is important to do this review
The current literature provides enough evidence to suggest that both thiopental and propofol are effective in the treatment of RSE (Parviainen 2002; Van Gestel 2005). As both agents are associated with inherent side effects and complications, the choice of the agent is usually left at the discretion of the attending anaesthetist. There is a lack of evidence to suggest the superiority of one drug over the other. A systematic review of published literature would help us reach to some conclusion. To date, no Cochrane systematic review comparing thiopental and propofol for treatment of RSE has been published. The aim of this review is to establish which of the two commonly used anaesthetic agents, thiopental or propofol, is better suited for the treatment of RSE.
To compare the efficacy, adverse effects, and short- and long-term outcomes of SE treated with one of the two anaesthetic agents, thiopental sodium or propofol.
Criteria for considering studies for this review
Types of studies
We included all relevant randomised or quasi-randomised controlled studies, regardless of blinding. Diagnosis of RSE was based on any given standard definition specified in the articles and treatment consisting either of propofol or thiopental sodium. We excluded studies that did not define RSE, prior treatment with any other intravenous anaesthetic before treatment with thiopental sodium or propofol and use of intermittent boluses of thiopental sodium or propofol for treating RSE.
Types of participants
We included individuals of any age group and gender diagnosed with RSE of any aetiology.
Types of interventions
Patients receiving either thiopental sodium or propofol for the treatment of RSE, in addition to standard antiepileptic drugs used in SE.
Types of outcome measures
1.Total control of seizures.
3.Length of ICU stay.
1.Adverse events, such as infection, hypotension and propofol infusion syndrome.
2.Duration of mechanical ventilatory support.
3.Duration of hospital stay.
5.Long-term outcomes, such as dependence for daily activities (walking, eating, bathing, dressing and toileting).
Search methods for identification of studies
We searched the following databases:
1.Cochrane Epilepsy Group Specialized Register (10 May 2012);
2.The Cochrane Central Register of Controlled Trials (CENTRAL Issue 4 of 12, The Cochrane Library 2012) using the search strategy outlined in Appendix 1;
3.MEDLINE (Ovid, 1946 to May week 1, 2012) using the search strategy outlined in Appendix 2.
We did not apply any language restrictions.
Searching other resources
We searched the following web sites for information about relevant ongoing trials:
1.IndMED (www.indmed.nic.in) (10 May 2012) using the search terms: 'Propofol AND (thiopental OR thiopentone) AND (epilepsy OR epileptic)';
2.South Asian Database of Controlled Clinical Trials (www.cochrane-sadcct.org, accessed 10 May 2012) using the search term 'propofol';
Using the results of the above searches, we screened all titles and abstracts for eligibility. Two review authors (AB and GPS) independently performed this screening. We obtained and assessed the full articles of all eligible RCTs for relevance based on the pre-planned check-list. Each author documented the reason for each trial that was excluded. We resolved any disagreement between the two review authors by discussion with a third review author (HP), who decided on the inclusion or exclusion of the study. We compiled a list of all eligible trials.
Data extraction and management
Two review authors (AB and GPS) planned to independently extract the data and assess the trial quality. We resolved any disagreement through consultation with the third review author (HP). In case of additional information being required, AB was chosen to contact the first author of the relevant trial.
Assessment of risk of bias in included studies
Two review authors assessed the methodological quality of the eligible trials independently (AB and GPS). We resolved any disagreement by discussion with the third review author (HP). We performed the assessment as suggested in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008). We judged the quality of the study on the basis of the following:
1.method of randomisation,
2.concealment of allocation,
3.blinding and outcome,
4.incomplete outcome data,
6.any other bias.
We included a 'Risk of bias' table as part of the 'Characteristics of included studies' and a 'Risk of bias summary' figure (Figure 1), which details all of the judgements made for all included studies in the review.
Measures of treatment effect
We planned to record the number (percentage) of participants experiencing each categorical outcome and mean (standard deviation (SD))/median (range) for continuous outcome per randomised group. All primary analyses were dealt by 'intention to treat'.
Unit of analysis issues
We planned to include only RCTs in our review with parallel design. The nature of the intervention here suggested that unit of analysis issues, such as those associated with cluster randomisation, were unlikely to arise. If we included any cluster randomised studies we planned to assess the risk of bias following the suggestions in Section 16.3.2 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008), and the approach to analysis suggested in the subsequent sections. We would have taken a similar approach to any cross-over trials included.
Dealing with missing data
We contacted the first author of the relevant trial and collect the required information.
Assessment of heterogeneity
We planned to assess the clinical heterogeneity of included studies as assessed as methodological diversity, such as distribution of patient characteristics (age, seizure type and number of drugs taken at the time of randomisation) and trial factors (randomisation concealment, blinding and loss to follow-up). We planned to use the Q statistics to test the statistical heterogeneity between trials and the I2 statistic to assess the magnitude of heterogeneity (Higgins 2002).
Assessment of reporting biases
We planned to assess publication bias/small-study effect in a qualitative manner, using a funnel plot. Due to limited data, we did not assess this bias.
We quantitatively reviewed the included data and combined it by intervention, outcome and population using the Cochrane Collaboration's statistical software, Review Manager (RevMan 2011). We planned to assess statistical heterogeneity using the Chi2 statistic and consider P values of 0.05 or less as statistically significant. We planned to assess the level of inconsistency across the studies using the I2 statistic where I2 greater than 50% indicates substantial heterogeneity. Had we found statistically significant heterogeneity that we could not readily explain, we would have assessed it using a random-effects model.
Due to only one study meeting our inclusion criteria, we did not perform a meta-analysis.
Subgroup analysis and investigation of heterogeneity
We planned to perform subgroup analyses for the following: age groups (children (< 14 years of age) and adults), gender, aetiology and type of seizure (convulsive or non-convulsive).
Due to limited data, we did not perform a subgroup analysis.
We planned to perform a sensitivity analysis to assess the influence of including studies judged low in methodological quality and characteristic of intervention; that is, the doses of propofol and thiopental used.
Due to limited data, we did not perform a sensitivity analysis.
The study conducted by Rossetti et al (Rossetti 2011) was a randomised, single-blind, multicentre trial studying adults with RSE. Patients received either propofol or barbiturates for control of seizures. There primary end point was the proportion of patients with RSE controlled after first course of study drug and secondary end point included drug tolerability. The trial was terminated after three years with only 24 patients recruited of the 150 required. The trial was under-sampled and found prolonged mechanical ventilation requirement in the barbiturate arm. The treatment-related complications were comparable for both propofol and barbiturates.
Results of the search
Our searches yielded 34 references (eight from MEDLINE, three from CENTRAL, two from the Cochrane Epilepsy Group Specialized Register, 20 from the South Asian Database of Controlled Clinical Trials, and one from ClinicalTrials.gov. The search of IndMED yielded no references). After de-duplication, 24 references remained. After further scrutiny, only one study was identified for inclusion in this review.
Only one study was included in the review (Rossetti 2011). The overview of the study is given in Table 1.
Risk of bias in included studies
The study conducted by Rossetti et al (Rossetti 2011) was a single-blind study. The small sample size and lack of double blinding could have influenced results.
It is difficult to understand how allocation was done, as the authors have not commented on this in their paper. Personal communication with the author failed to clarify this issue.
This was a single-blind study and so there is definitely a chance of performance bias and detection bias.
Incomplete outcome data
Attrition bias is unlikely, as there are no incomplete outcome data.
Reporting bias is unlikely as the authors have reported all the outcomes they have described in the methodology.
Other potential sources of bias
No other potential sources of bias could be identified.
Effects of interventions
The initial search in May 2011 identified no results. Searching other databases in July 2011, we identified only one study (Rossetti 2011) that met our inclusion criteria. This was a randomised, single-blind study that included patients with RSE treated with either propofol or barbiturates. The details of the study have been outlined above.
Total control of seizures
There was no statistical difference in total control of seizures after the first course of study drug between propofol and thiopental sodium; 6/14 patients versus 2/7 patients in propofol and thiopental sodium group, respectively (RR 1.50; 95% CI 0.40 to 5.61) (Analysis 1.1).
There was no statistical difference in mortality between propofol and thiopental sodium; 3/14 patients versus 1/7 patients in propofol and thiopental sodium group, respectively (RR 1.50; 95% CI 0.19 to 11.93). In the propofol group, deaths presumed to be in-hospital were secondary to Creuzfeldt-Jakob disease (day five), cardiac asystole (day six) and progressive brain tumour (day 11). In the thiopental group, one patient died on day five, secondary to colic ischaemia. Deaths on days 21, 29 and 42 were due to paraneoplastic encephalitis, pneumonia and sepsis, respectively, which are presumed to be out-of-hospital (Analysis 1.2).
There was no statistical difference in adverse events between propofol and thiopental sodium (Table 2). Infection was seen in 7/14 patients versus 5/7 patients in propofol and thiopental sodium group, respectively (RR 0.70; 95% CI 0.35 to 1.41).
Hypotension during administration of study drugs and requiring use of vasopressors was seen in 7/14 patients versus 4/7 patients in propofol and thiopental sodium group, respectively (RR 0.87; 95% CI 0.38 to 2.00) (Analysis 1.3).
The other severe complication noted was non-fatal propofol infusion syndrome in one patient.
Duration of mechanical ventilation
The number of days of mechanical ventilation were more in the thiopental group when compared with propofol group (median (range): 17 days (5 to 70 days) with thiopental sodium versus 4 days (2 to 28 days) with propofol).
There was no statistical difference in the functional outcome between propofol and thiopental sodium at three months; 5/14 patients versus 3/7 patients in propofol and thiopental sodium group, respectively (RR 0.83; 95% CI 0.28 to 2.52) (Analysis 1.4).
Our search identified only one study (Rossetti 2011) that addressed the issue of treatment of RSE using thiopental sodium and propofol in a randomised, single-blind multicentric trial. This trial was under-sampled and had to be terminated prior to completion. Twenty-four patients were recruited in five centres; 14 received propofol and seven received thiopental sodium. The primary end point, that is, control of seizure with first course of drug treatment, was achieved in 43% in the propofol group and 22% in the barbiturate group (seven patients received thiopental sodium and two patients received pentobarbital). The overall mortality was 43% and 34% in the propofol and barbiturate group, respectively. However, the authors fail to report the in-hospital mortality. Patients returning to baseline condition at the three months' follow-up were similar in the two groups. No information was provided on the length of intensive care unit (ICU) and hospital days of the patients. However, days of mechanical ventilation were significantly more in the thiopental group. This could be due to the long elimination half-life of the drug when compared with propofol. The fact that the trial was prematurely stopped could have introduced bias. At the same time, the under-sampling resulted in loss of power to detect difference between the two treatment arms.
This study confirms that RSE is a serious clinical condition carrying high morbidity and mortality. The authors of the study agree that a larger multicentric study is needed with larger sample size and adequate funding is needed to reach conclusive results. The authors also suggest that a third treatment arm using midazolam as the treatment drug may be included in the study that may address the issue related to tolerability of the drugs, propofol and thiopental.
Summary of main results
Both propofol and thiopental sodium are broadly comparable in terms of seizure control, mortality, rate of complications and long-term outcome of patients with RSE. The 95% confidence interval was wide and allowed for a up to a more than two-fold difference between the two drugs. Days of mechanical ventilation were more in patients receiving thiopental sodium when compared with propofol.
Overall completeness and applicability of evidence
There is lack of evidence that can clarify the efficacy of propofol and thiopental sodium over each other in the treatment of RSE. We are unable to detect a difference between the two drugs due to methodological issues. There is a need for a large, randomised controlled trial for this serious condition.
Quality of the evidence
Potential biases in the review process
Agreements and disagreements with other studies or reviews
To date this is the only eligible study available.
Implications for practice
There are no data from randomised controlled trials that can clarify the superiority of thiopental sodium and propofol over each other in the treatment of RSE. Clinicians managing such cases of RSE should be aware of the adverse effects of the two anaesthetics drugs.
Implications for research
RSE is a serious clinical condition where conducting clinical trials may be difficult. Various ethical and methodological issues may arise. However, the problem itself is important and needs to be resolved. Use of thiopental sodium and propofol in RSE patients should be assessed with good-quality multicentric, randomised controlled trials, for their effect and efficacy in terms of total control of seizures, mortality, length of ICU and hospital stay, adverse effects and long-term outcome, such as dependence for daily activities. Many centres may have to be involved to enrol a suitable number of patients so that adequate power of the study can be achieved. A standard method may have to be followed with uniform outcome measures.
We wish to thank The Cochrane Epilepsy Group for their continuous support in helping us prepare the protocol. We would like to thank Professor A Rossetti for providing the additional data and information on the included study. We wish to thank the South Asian Cochrane Network and Centre, CMC, Vellore, India, who conducted the workshop at the Prof. BV Moses and ICMR Center for Advanced Research and Training in Evidence-Informed Healthcare, where this review was completed.
Characteristics of studies
Characteristics of included studies [ordered by study ID]
EEG: electroencephalography; iv: intravenous; RSE: refractory status epilepticus.
Participants: adults (> 16 years) with RSE not due to cerebral anoxia, who clinically required coma
Sex (female/male): propofol group: 50%/50%; barbiturate group: 66%/34%
Age (median [range]): propofol group: 57 years (26 to 87 years); barbiturate group: 64 years (16 to 78 years)
Ethnic groups: not reported
Duration of epilepsy: not reported
Inclusion criteria: patients > 16 years of age suffering from RSE receiving at least 1 first-line and 1 second-line drug in adequate doses
Exclusion criteria: patients with known pregnancy, known intolerance to the study drugs, mitochondrial disorders, egg allergy, hypertriglyceridaemia (> 5 mmol/L) or significant rhabdomyolysis (creatinine kinase > 1500 U/L) on admission
Diagnostic criteria: RSE not due to cerebral anoxia, defined as ongoing clinical or electrographic seizures, or repetitive seizures without return to baseline for at least 30 minutes despite administration of 1 first-line (benzodiazepine) and 1 second-line antiepileptic drug (phenytoin, valproate, phenobarbital and levetiracetam) in adequate doses
Total randomised: 24 patients; 14 allocated to propofol group and 10 allocated to barbiturate group; (1 patient in thiopental sodium group did not require treatment and so excluded, remaining 9 analysed)
Number of control centre: 2
Country/location: Switzerland and US
Setting: CHUV et Universite de Lausanne, Lausanne and Brigham and Women's Hospital, Harvard School of Medicaine, Boston
Intervention I: 2 mg/kg titrated to burst suppression or 4 mg/kg until EEG was available
Intervention II: 2 mg/kg iv titrated to burst suppression or 5 mg/kg if no EEG available
Treatment before study: first- and second-line antiepileptic drugs
Time to treatment since onset of status: not reported
Duration of follow-up: 3 months
2 treatment arms: propofol and barbiturates (thiopental sodium or pentobarbital)
Primary outcomes (as stated in the publication): to assess the effectiveness (RSE control, adverse events) of a first course of propofol versus barbiturates
Secondary outcomes (as stated in the publication): none
Outcome used in our review:
total control of seizures
duration of mechanical ventilation
Stated aim of study: "This prospective study was undertaken to assess the effectiveness (SE control, adverse events) of a first course of PRO versus barbiturates, the two most commonly used agents according to the aforementioned surveys"
Language of publication: English
Commercial funding: yes
Non-commercial funding: no
Publication status(peer review journal): yes
Publication status (journal supplement): no
Publication status (abstract): no
Funded by AstraZeneca (Switzerland) and UCB (Switzerland)
No conflict of interest
Clinical Trial.gov ID: NCT00265616
Risk of bias
Support for judgement
Random sequence generation (selection bias)
"After written consent was obtained by proxy, randomisation was stratified by institution using sealed envelopes"
Comment: probably done
Allocation concealment (selection bias)
Comment: authors contacted. No information provided
Blinding (performance bias and detection bias)
Being a single-blind study, only the patient was blinded. Assessors were not blinded
Incomplete outcome data (attrition bias)
No loss to follow-up. All participants randomised completed the study and were included in final analysis
Selective reporting (reporting bias)
All outcomes that were mentioned in the methodology have been reported
Trial was terminated before completion due to inadequate recruitment