FormalPara Key Points

Available information about the efficacy of ketamine is unavoidably biased as it is only based on observational studies, most retrospective.

The methodological limitations of the studies prevent any meaningful conclusions on comparative efficacy between ketamine and conventional anaesthetics.

There is a need for clinical trials designed to assess the efficacy of ketamine as early third-line therapy, thus avoiding endotracheal intubation.

The commonly held opinion that ketamine has a better safety profile than conventional anaesthetics needs additional evidence.

The neuroprotective properties of ketamine in humans need to be assessed through analysis of plasma and cerebrospinal fluid biomarkers for neuroinjury and prospective neuropsychological and neuroimaging studies.

1 Introduction

Status epilepticus (SE) is a life-threatening medical emergency and is traditionally defined as “an acute epileptic condition characterised by continuous seizures for at least 30 min, or by 30 min of intermittent seizures without full recovery of consciousness between seizures” [1]. Based on improved understanding of pathophysiology, there is now consensus that any seizure lasting longer than 5 min should be treated as SE [2]. Status epilepticus lasting longer than 120 min and not responding to first- and second-line treatments is defined as “refractory” (RSE) and requires intensive care unit admission [3]. Super-refractory SE is defined as SE that has continued or recurred despite therapy with general anaesthesia for 24 h or longer [3]. Based on the clinical features and severity, SE is distinguished as being either “convulsive” or “non-convulsive”, the former being the most common and harmful.

There is general consensus regarding the first and second lines of treatment for SE [4]. Although the same types of drugs are used in different countries, the algorithm/protocols may differ, even among institutions in the same country. At present, there is no definitive evidence or agreement to guide an optimal treatment choice for RSE [3, 5,6,7]. Refractory convulsive SE is generally treated with coma induction using high-dose midazolam (MDZ) or conventional anaesthetics such as thiopental, pentobarbital or propofol [5, 7]. Conversely, some concerns exist about the opportunity to use conventional anaesthetics in the less severe forms of non-convulsive SE (NCSE) with no impairment of consciousness [8, 9].

Status epilepticus results from the failure of inhibitory GABA-mediated mechanisms responsible for seizure termination and from the activation of excitatory glutamate-mediated mechanisms, which lead to abnormally prolonged seizures with consequent neuronal injury and death [10, 11]. In this scenario, N-methyl-d aspartate (NMDA)-receptor antagonist modulating molecules offer an attractive alternative in SE [12, 13]. Experimental models have demonstrated that ketamine (KE), a potent NMDA-receptor antagonist, controls prolonged SE while it is ineffective in its early phase [14]. It has also been demonstrated that the efficacy of KE is increased by the concomitant use of benzodiazepines [15, 16]. Ketamine is a phencyclidine derivative, with a chiral structure consisting of two optical isomers. Racemic KE is the most commonly used form and is a mixture of equal amounts of the two enantiomers (R)-KE and (S)-KE, the latter displaying analgesic and anaesthetic potency about three-fold superior to (R)-KE. Ketamine has a half-life of 2–3 h and is metabolised by cytochrome P450 3A and cytochrome P450 2B6 enzymes mainly owing to its active metabolite, norketamine. It is water and lipid soluble, reaching extensive distribution in the body. However, because of extensive first-pass metabolism, oral bioavailability is poor and vulnerable to pharmacokinetic drug interactions [17, 18].

The pharmacological profile is characterised by the so-called “dissociative anaesthetic state” described as a form of anaesthesia characterised by catalepsy, catatonia, analgesia and amnesia that does not necessarily cause loss of consciousness. These analgesic and anaesthetic properties of KE combined with its typical sympathomimetic effects are mediated by different sites of action. N-methyl-d aspartate-receptor antagonism is the most important neuropharmacological mechanism for the analgesic and anaesthetic effects and contributes to its neuroprotective action [19]. Analgesic state and dysphoric reactions are mediated by opiate receptors, whereas the enhancement of central peripheral monoaminergic transmission contributes to its sympathomimetic properties. Moreover, the inhibition of central and peripheral cholinergic transmission may contribute to the induction of the anaesthetic state and hallucinations, whereas the hyperpolarisation-activated cyclic nucleotide channels (HCN1 channels) contribute to the sedating actions of KE.

Compared with other drugs used for the treatment of SE, KE-induced respiratory depression is rare and this effect, such as the increase of bronchial secretions, can be prevented and reduced by the administration of a muscarinic antagonist such as atropine. The sympathomimetic properties of KE, in particular, subtend its vasopressor-sparing effect that reduces the need for vasoactive compounds to counteract hypotension, which is frequently seen with conventional intravenous anaesthetics commonly used in SE [18, 20]. Along with hallucinations and hypersalivation, nausea and vomiting are the most relevant adverse events reported with the use of KE [17, 18].

Based on its promising efficacy and good safety profile, KE may be considered as the anaesthetic agent of choice in specific situations and as an out-of-hospital treatment option of SE [18, 20, 21].

Ketamine is currently administered in patients with RSE only when conventional anaesthetics have failed [3]; however, based on its potential efficacy and good safety profile, more recent studies suggest [22] and recommend [23] an earlier administration. Here, we performed a systematic review of the literature on the efficacy and safety of KE in treating RSE in paediatric and adult populations.

2 Methods

We conducted a systematic review and reported it according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist [24]. We performed a MEDLINE literature search using PubMed to identify all articles published as of February 2018 with the following research details: “ketamine”[MeSH Terms] OR “ketamine” [All Fields]) AND (“status epilepticus” [MeSH Terms] OR (“status” [All Fields] AND “epilepticus” [All Fields]) OR “status epilepticus” [All Fields]. We also searched the Cochrane Database of Systematic Reviews (keyword: “ketamine” OR “status epilepticus”) for related systematic reviews and the ClinicalTrials.gov website for ongoing clinical studies on RSE.

Titles and abstracts obtained through the literature search were screened for inclusion in the review using the following criteria: KE efficacy and safety as primary outcome in SE, both in the paediatric and adult populations; only studies published in English. Exclusion criteria were: preclinical studies, editorials, letters and non-English publications. After full-text reading of the accordingly selected papers, articles included in the review were further reviewed in their reference list for other publications relevant to the topic (secondary search).

The following data were extracted: study type and design, patient demographics, number of patients, type of SE (convulsive, non-convulsive and subtle SE, focal and generalised SE), aetiology of SE, dose, timing, duration and route of KE administration, prior and concomitant therapies, outcome defined as electrographic SE control and adverse events. Taking into account the different aetiologies of SE across ages, data on paediatric and adult populations were analysed separately. A systematic assessment of the available evidence was conducted in accordance with the GRADE methodology [25, 26]. A meta-analysis was not possible because of the lack of prospective randomised trials.

3 Results

The search strategy yielded 135 MEDLINE abstracts, no Cochrane systematic review and two ongoing clinical trials on KE use in refractory convulsive SE in children (ClinicalTrials.gov identification number: NCT02431663) and adults (ClinicalTrials.gov identification number: NCT03115489). Eighty-six full texts were analysed and 63 included in the review. A further nine articles emerged from the secondary search and a total of 72 articles contributed to this review. Results of the search strategy across all databases and other sources are summarised in Fig. 1.

Fig. 1
figure 1

Flow diagram

Full size image

No randomised controlled trials were available and the current evidence consisted of 27 case reports accounting for 30 individuals and 14 case series, 6 of which included children. Most of the studies were retrospective and three were companion publications expanding on the original data set of Meyer Children’s Hospital [22, 27, 28]. For this reason, only one case series [22] was considered in the data analysis of the current review. Table 1 shows the characteristics of the selected case series by study design and population (adults vs. children). A total of 248 individuals (29 children) with a median age of 43.5 years (range 2 months to 67 years) were treated in 12 case series with a sample size ranging from 5 to 67 individuals (median 11). Table 2 reports the GRADE assessment for case series only.

Table 1 Selected case series
Full size table
Table 2 GRADE assessment of the selected case series 
Full size table

3.1 Adults

A total of 219 adults (median age 54.5 years, range 24–67 years) were treated in eight case series [29,30,31,32,33,34,35,36] with a sample size ranging from 7 to 67 individuals (median 11). In 16 case reports [37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52], 19 individuals and 20 RSE episodes were treated with KE (Tables 3, 4). Infections and anoxia were the most frequent aetiologies [30, 33, 35, 36]. In more than half of 60 RSE episodes described by Gaspard et al. [31] the aetiology remained unknown. The type of RSE was not specified in four out of eight case series and NCSE was the type of SE most commonly treated with KE, both in case series and in case reports (Tables 3, 4). The mean duration of SE prior to KE administration was highly heterogeneous, regardless of SE type, and ranged from 24 h to 26.5 days in case series (Table 3) and from 12 h to 5 months in case reports (Table 4).

Table 3 Demographics and outcome of refractory status epilepticus (RSE) in adults treated with ketamine (KE): case series
Full size table
Table 4 Demographics and outcome of refractory status epilepticus (RSE) in adults treated with ketamine (KE): case reports
Full size table

Considering both case reports and case series, KE was always administered after conventional anaesthetics, with the exception of the patient described by Pizzi et al. [50] while propofol was the most common third-line treatment administered. Benzodiazepines, especially MDZ, were the most commonly used drugs in add-on. Ketamine dosage ranged from 0.07 to 15 mg/kg/h. The duration of KE infusion ranged from 6 h to 29 days. The proportion of individuals where KE was effective (resolution of RSE) ranged from 11% in the study by Gosselin-Lefebvre et al. [32] which enrolled nine patients, to 100% in the case series described by Synowiec et al. [33] which included 11 individuals. Considering all RSE episodes, 156/222 (70.3%) were controlled by KE administration. Electroencephalography (EEG) features were not specified in the majority of the case series and a burst-suppression pattern was only observed in three out of seven patients as reported by Bleck et al. [29] as well as in three case reports [44, 47, 52]. Diffuse slowing and diffuse beta activity were EEG patterns observed in RSE episodes in which KE was effective. Adverse events, including shock, sepsis, renal failure, pneumonia and acidosis were only reported only in the series of Gaspard et al. [31] Cerebellar atrophy and cardiac arrest were documented by Ubogu et al. [38] and Koffman et al. [51] Endotracheal intubation was avoided in two patients in whom KE was effective [42, 50].

3.2 Children

A total of 29 children (age range 2 months to 18 years) accounting for 35 RSE episodes were treated in four case series [22, 53,54,55] with a sample size ranging from five to 13 individuals (median 5.5). Eleven patients treated with KE were also documented in corresponding case reports [56,57,58,59,60,61,62,63,64,65,66]. In case series, epileptic encephalopathy was the most frequent underlying condition (Table 5). Heterogeneous aetiologies were documented in case reports (Table 6). The type of RSE was not specified in two of the four case series, although convulsive SE was the most common form treated with KE, both in case series and in case reports (Tables 5, 6). The mean duration of SE prior to KE administration was highly heterogeneous, regardless of the RSE type, and ranged from 5 h to 26 days in case series and from 10 to 73 days in case reports.

Table 5 Demographics and outcome of refractory status epilepticus (RSE) in children treated with ketamine (KE): case series
Full size table
Table 6 Demographics and outcome of refractory status epilepticus (RSE) in children treated with ketamine (KE): case reports
Full size table

Both in case reports and case series, KE was always administered after conventional anaesthetics, with MDZ the third-line treatment most commonly employed. Ketamine dosage ranged from 0.04 to 10 mg/kg/h. The duration of KE infusion ranged from 1 to 21 days. The proportion of individuals in whom KE was effective (resolution of RSE) ranged from 20% in the study by Al-Otaibi et al. [55] which enrolled five patients, to 100% in the case series described by Mewasingh et al. [53] which included five NCSE children treated with oral KE. Of the 19 refractory convulsive SE episodes treated by Ilvento et al. [22] KE was effective in 14 (74%). Overall, 28/46 (61%) of RSE episodes were controlled by KE administration.

A burst-suppression pattern was observed during the initial bolus of 3 mg/kg in most responders in the case series of Ilvento et al. [22] and was followed by diffuse theta and beta activity in four children; endotracheal intubation was avoided in two of them. Mewasingh et al. [53] documented diffuse theta activity in five children who were treated orally. Adverse events were only reported in the series of Ilvento et al. [22] and consisted of a slight increase of saliva secretion in all patients and a transient mild increase of liver enzymes in four out of 13 children. Endotracheal intubation was unnecessary in the five individuals with NCSE treated with oral KE [53] and in 5 out of 13 children with refractory convulsive SE [22].

4 Discussion

Ketamine has proven effective in treating both convulsive and non-convulsive RSE in the adult and paediatric populations. In the case series with the largest sample, resolution of RSE was obtained in 74% of 19 patients with SE prospectively followed [22] and in 91% of 67 adults in a retrospective review of medical records [35]. The available information on the efficacy of KE is biased by the design of the available studies that is always observational, mostly retrospective. Indeed, only single-arm studies without a control group were available, and only two paediatric case series out of four and none in the adult population had a prospective design [22, 53].

In adults, the efficacy of KE was higher in convulsive SE compared with NCSE in the two largest series [31, 36]. Evidence of the efficacy of KE in paediatric NCSE was reported only in five children successfully treated with oral administration after a mean RSE duration of 4 days [53]. Regardless of SE type, KE was twice as effective if given early, with efficacy dropping from 64% in the 42 patients with RSE lasting 3 days [36] to 32% when the mean duration of 60 RSE was 26.5 days [31]. A similarly good efficacy was observed in children, with a response rate of 74% in the 19 refractory convulsive SE episodes treated after a mean RSE duration of 7 days [22].

Data from experimental models suggest the efficacy of KE in the treatment of SE when the drug is administered not too early (after 15 min) but at least 1 h after the onset of symptoms [14]. Likewise, the increased efficacy of KE when administered in the early stages of SE is confirmed even in the clinical setting, although the definition “early” is widely heterogeneous in different studies, ranging from a few hours to some days after onset.

Timing in KE administration and convulsive SE both appear to be the most relevant predictive determinants of the efficacy of KE in adults and children, while KE dosage, which was extremely heterogeneous throughout the studies, does not seem to be an independent prognostic factor, both in case series and in individual case reports. Moreover, timing and modalities of KE dosage titration were not reported in most articles, making it impossible to estimate the timing of SE control after the start of KE treatment.

The missing information, mainly owing to the retrospective design of the studies and the absence of centre-specific and standardised protocols for the treatment of SE, represents the most important limitation to assess the possible advantages from the use of KE in this clinical scenario.

Benzodiazepines, and especially MDZ, were the drugs most commonly used in add-on but no conclusions or speculations about a possible synergic effect, as documented in animal models [15, 16], were possible in this review because of a lack of detailed information.

As expected, in most instances, the delay using KE was related to prior administration of conventional anaesthetics, such as MDZ, thiopental, pentobarbital and/or propofol. Experimental models suggest that, with continuing seizures, inhibitory GABAA receptors are internalised in clathrin-coated vesicles, and excitatory NMDA receptors are mobilised to the membrane [10, 11]. However, the mechanisms underlying the refractoriness of SE are likely to be multifactorial and more complex as suggested by preclinical evidence indicating that polytherapy is more effective than monotherapy [67]. Conventional anaesthetics, all acting on GABAA receptors will be, therefore, less active, prompting administration of higher doses, which will in turn enhance their untoward effects, especially hypotension [20]. Ketamine represents an attractive alternative for SE, also in relation to its sympatico-mimetic action [12]. Owing to its pharmacological properties, KE use does not necessarily require amine administration or mechanical ventilation. Nevertheless, in the largest adult case series, a higher use of vasopressors was reported in association with KE compared with earlier treatments [31, 35]. Additional evidence is needed to confirm the safety profile of KE and the real feasibility in avoiding amine administration.

Ketamine has neither cardiac nor respiratory depressant properties, therefore, its use does not routinely require endotracheal intubation. This is considerably advantageous as intubation represents per se a negative prognostic factor of increased morbidity and mortality in critically ill adults and children [20, 68, 69]. The risk-benefit profile for conventional anaesthetics is questionable and their use seems to be unjustified especially in NCSE where endotracheal intubation is unnecessary, or when consciousness impairment does not occur [9]. The good safety profile of KE is furthermore documented by the paucity of adverse events emerging from both case series and case reports in adults and in children.

An EEG burst-suppression pattern is typically observed in patients treated with conventional anaesthetics and represents, along with seizure control, the goal of treatment. The clinical efficacy of KE is associated with a more heterogeneous EEG pattern in which diffuse slowing and diffuse beta activity should be considered the targets to achieve and retain on a par with the burst-suppression pattern.

Neuroprotection from glutamate-induced neurotoxicity is another potential advantage associated with KE use [70,71,72]. This hypothesis was substantiated by the ongoing KIND trial, examining the ability of subanaesthetic doses of KE to improve outcome and mitigate neuronal injury [as assessed using 3 Tesla magnetic resonance imaging and analysis of plasma and cerebrospinal fluid neuroinjury biomarkers in patients with grade I–IV subarachnoid haemorrhage (ClinicalTrials.gov identification number: NCT02636218)].

Finally, preliminary evidence is available about immunomodulatory and anti-inflammatory properties of KE, which seem to contribute to its anti-epileptogenic activity [73] and to a more favourable outcome [74]. Appropriately designed studies are needed to confirm this possible additional beneficial effect of KE in SE.

A systematic prospective collection of data accounts for more solid clinical, EEG and therapeutic information. Therefore, a targeted clinical database should be established to attempt to answer key questions such as: (1) is KE efficacy increased by add-on benzodiazepines use? (2) does KE require amine use and if, when? (3) is there any specific or more frequently observed EEG pattern in responders? (4) how long is KE infusion to be continued in responders? and (5) does KE have a neuroprotective action and can it be assessed by neuro-injury biomarker sampling in plasma and cerebrospinal fluid and prospective neuropsychological and neuroimaging evaluations?

5 Conclusion

As reported in a recent editorial by Dorandeu [75], despite the poor quality of the available evidence on the efficacy and safety of KE, data are encouraging and support the interest in developing future specifically designed clinical trials to investigate its role in the early stages of SE. The less pronounced hypotensive and respiratory depressive effects of KE and the potentially favourable risk/benefit profile compared with conventional anaesthetics, as well as the plausible neuroprotective effect, draw a concrete possibility of a future widespread application. We are confident that the two specifically designed ongoing trials will provide unbiased evidence on the efficacy and safety of KE in this particular scenario.