Seizure: European Journal of Epilepsy 102 (2022) 74–82

Contents lists available at ScienceDirect

Seizure: European Journal of Epilepsy

journal homepage: www.elsevier.com/locate/seizure

Review

Treatment of benzodiazepine-resistant status epilepticus: Systematic review

and network meta-analyses
Puneet Jain a, Satinder Aneja b, Jessie Cunningham c, Ravindra Arya d, e, Suvasini Sharma a, f, *
a
Epilepsy Program, Division of Neurology, Department of Pediatrics, Hospital for Sick Children (HSC), University of Toronto, Toronto, ON, Canada
b
Department of Pediatrics, Sharda Hospital, Sharda University, Greater Noida, Uttar Pradesh, India
c
Hospital Library and Archives, Learning Institute, Hospital for Sick Children, Toronto, ON, Canada
d
Comprehensive Epilepsy Center, Division of Neurology, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio, USA
e
Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, Ohio, USA
f
Neurology Division, Department of Pediatrics, Lady Harding Medical College (LHMC) and Kalawati Saran Children Hospital, Delhi University, New Delhi, Delhi, India

A R T I C L E I N F O A B S T R A C T

Keywords: Purpose: Multiple interventions have been studied for benzodiazepine-resistant status epilepticus (SE) in children
Status epilepticus and adults. This review aimed to summarize the available evidence and provide estimates of comparative
Benzodiazepine-resistant SE effectiveness and ranking of treatment effects.
Prolonged seizures
Methods: All randomized controlled trials studying patients (>1 month of age) with benzodiazepine-resistant SE
Levetiracetam
Valproate
were included. Outcomes including seizure cessation within 60 min, seizure freedom for 24 h, death, respiratory
Phenobarbital depression warranting intubation and cardiovascular instability were studied. Conventional and network meta-
Phenytoin analyses (NMA) were done.
Fosphenytoin Results: Seventeen studies were included (16 in NMA). Phenobarbital and high-dose levetiracetam were signif­
icantly superior to phenytoin with respect to seizure cessation within 60 min. Network ranking demonstrated
that phenobarbital had the highest probability of being the best among the studied interventions followed by
high-dose levetiracetam and high-dose valproate. Network meta-analysis was limited by predominant indirect
evidence and high heterogeneity.On pairwise comparisons, phenobarbital was found to be associated with a
higher risk of need for intubation and cardiovascular instability. Levetiracetam had a better safety profile than
fosphenytoin.
Conclusions: Based on low quality evidence, phenobarbital appears to be the most effective agent for seizure
cessation within 60 min of administration in patients with benzodiazepine resistant status epilepticus. High-dose
levetiracetam, high-dose valproate and fosphenytoin are probably equally effective. Choice of medication may be
guided by effectiveness, safety concerns, availability, cost and systemic co-morbidities.

1. Introduction Benzodiazepine-resistant SE or established SE. This may be partially

explained by the time-dependent changes in GABAA receptor function
Status epilepticus (SE) is a common neurological emergency, which and transmembrane chloride gradients [6].
may be associated with significant short and long term neurological and Traditionally intravenous phenytoin or phenobarbital have been the
systemic morbidity. A staged treatment protocol for management of agents of choice in benzodiazepine resistant status epilepticus. However
status epilepticus is generally followed. The first line of treatment is phenytoin may cause cardiac arrhythmias, hypotension, extravasation
administration of benzodiazepines; intravenous lorazepam, intravenous and purple glove syndrome [7]. The use of phenytoin is also limited by
diazepam or intramuscular midazolam [1,2]. However, approximately the speed of infusion, it needs to be infused at the rate of ≤ 1
30–40% of all patients fail to respond to initial treatment with benzo­ mg/kg/min, so the infusion takes at least 20 min [8,9]. In status epi­
diazepines [3]. The estimates may be higher based on recent lepticus management, time is of the essence, as ongoing seizures lead to
population-based studies [4,5]. This condition is referred to as incremental brain injury. Fosphenytoin, a phenytoin pro-drug, is a

* Corresponding author at: Neurology Division, Department of Pediatrics, Lady Harding Medical College (LHMC) and Kalawati Saran Children Hospital, Delhi
University, New Delhi, Delhi, India.
E-mail address: [email protected] (S. Sharma).

https://doi.org/10.1016/j.seizure.2022.09.017
Received 12 July 2022; Received in revised form 6 September 2022; Accepted 25 September 2022
Available online 26 September 2022
1059-1311/© 2022 British Epilepsy Association. Published by Elsevier Ltd. This article is made available under the Elsevier license (http://www.elsevier.com/open-
access/userlicense/1.0/).
P. Jain et al. Seizure: European Journal of Epilepsy 102 (2022) 74–82

useful alternative to phenytoin, as it can be administered at three times in their administration, adverse effect profile, cost and availability
faster than phenytoin. However, higher cost and limited availability in across countries. Levetiracetam was further categorized as low dose (≤
low resource settings is an impediment. Further, there are concerns of 30 mg/kg/dose) and high dose (> 30 mg/kg/dose). Similarly, valproate
respiratory depression with phenobarbital, especially if it is adminis­ was also further categorized as low dose (≤ 30 mg/kg/dose) and high
tered after benzodiazepines. dose (> 30 mg/kg/dose).
In the last decade, intravenous formulations of other anti-seizure
medications such as levetiracetam and valproate have been used in 2.3. Data extraction
benzodiazepine resistant status epilepticus. These drugs may offer ad­
vantages in terms of safety and improved tolerability, but availability The data was extracted as per a pre-designed data abstraction sheet
and cost are still limitations. There are now several randomized and included author, year of publication, study period, country of origin,
controlled trials comparing levetiracetam with either phenytoin or fos­ study design, presence of blinding, number of study participants, age
phenytoin and a few RCTs comparing valproate with phenytoin/fos­ and sex of study participants, details of intervention in each treatment
phenytoin and levetiracetam [10–15]. There is also some data on the use arm, history of previous epilepsy, etiology, seizure cessation within 60
of lacosamide in this setting [16,17]. min, seizure freedom at 24 h, time to seizure cessation, death during
We aimed to perform a systematic review and a network meta- hospital admission, respiratory depression requiring intubation/ me­
analysis (NMA) of randomized controlled trials comparing the efficacy chanical ventilation, cardiovascular adverse effects including hypoten­
and safety of different agents used in benzodiazepine resistant status sion and/or arrhythmias, and other adverse effects (extravasation, rash,
epilepticus in adults and children. NMA provides estimates of compar­ hepatic dysfunction). Two authors independently extracted the data and
ative effectiveness and ranking of treatment effects incorporating both any disagreement in the extracted data was resolved by reaching a
direct and indirect comparisons. The results may help to guide clinicians consensus through discussion.
to decide the most appropriate agent for benzodiazepine resistant status
epilepticus in their setting. 2.4. Outcome measures

2. Methods The primary outcome measure was the proportion of study partici­
pants who achieved seizure cessation within 60 min of initiation/
This systematic review was conducted in accordance with the termination of treatment infusion. The secondary outcome measures
Preferred Reporting Items for Systematic Reviews and Meta-Analysis included proportion of study participants who remain seizure free for 24
(PRISMA) extension statement for network meta-analyses [18] and h after initiation/termination of treatment infusion, who died during the
updated 2020 guidelines [19]. It was registered with the PROSPERO hospital admission, who develop respiratory depression requiring intu­
International prospective register of systematic reviews bation/ mechanical ventilation, and who developed cardiovascular
(CRD42022313496). adverse effects including hypotension and/or arrhythmias.

2.1. Search strategy 2.5. Data analysis

Literature search was carried out by searching the following biblio­ The analysis was carried out by using ``intention-to-treat’’ popula­
graphic databases: MEDLINE (1946–March 31, 2022) with in-process tion for each outcome.
records and daily updates via Ovid; Embase+Embase Classic
(1974–2022 week 12) via Ovid; and Cochrane Central Register of 2.5.1. Conventional meta-analysis
Controlled Trials – March 2022. The search strategy consisted of both The meta-analysis was performed for the available primary and
controlled vocabulary, such as the National Library of Medicine’s MeSH secondary outcomes using Review Manager 5.4. The results were pre­
(Medical Subject Headings) and Emtree Subject Headings (Embase), and sented as percentages, and risk ratios with 95% confidence interval for
keywords. Some terms that were used included “Status epilepticus”, categorical outcomes. The statistical heterogeneity among studies was
``Seizure cessation’’ and ``Drug-Related Side Effects and Adverse Re­ assessed by Chi-square and I2 statistics. A fixed-effects model was
actions.’’ Please see supplemental Table 1 for the full list of search terms initially conducted, and if significant heterogeneity found between tri­
and controlled vocabulary subject headings. Methodological filters were als, potential sources of heterogeneity were considered, and where
applied to limit the study types to health technology assessments, sys­ appropriate a random-effects model was used.
tematic reviews, meta-analyses, randomized controlled trials (RCTs) and
controlled clinical trials. Conference abstracts were excluded from the 2.5.2. Network meta-analyses
search results. Date limits of 1990-current date and English language Network meta-analyses were performed for the primary outcome.
limits were also applied. See supplemental Table 1 for the detailed Difference in the proportion of patients achieving the outcome was used
search strategy. as the effect estimate. A frequentist, fixed effects, NMA was first per­
formed with phenytoin as the reference treatment. Heterogeneity in the
2.2. Types of studies, participants, and intervention overall model was quantified with I2, whereas significance of the be­
tween and within design heterogeneities were tested with respective Q
All randomized controlled trials (RCTs) reporting on children (> 1 statistics. Full design-by-treatment interaction model was used for
month of age) and/or adults with established convulsive status epi­ network decomposition. Effects of source data quality on network esti­
lepticus (SE), resistant to first line benzodiazepine treatment, were mates were assessed with relative proportions of direct and indirect
included. Only RCTs were included to minimize bias. Trials enrolling evidence, mean path length, and graph parallelism. Pooled network
patients with SE irrespective of their response to first-line benzodiaze­ effect estimates and their 95% confidence intervals for all possible
pines were excluded (as these patients may/may not be resistant to comparisons were computed, along with probabilistic rankings for each
benzodiazepines). Trials with duplicate study participants from the treatment. For comparisons having both direct and indirect data, node
same center/study group were also be excluded; most recently published splitting was performed to evaluate the relative magnitude and direction
study or study with larger sample size was then included in this scenario. of both effects.
Two authors (PJ, SS) independently agreed on selection of eligible Next, a Bayesian random effects NMA was performed using Markov
studies and achieved consensus. Chain Monte-Carlo simulation. Model parameters were optimized by
Phenytoin and fosphenytoin were analysed separately as they differ minimizing the potential scale reduction factor. Network effects

75
P. Jain et al. Seizure: European Journal of Epilepsy 102 (2022) 74–82

estimates along with their 95% credible intervals were computed for all 3.1. Pairwise comparisons
possible comparisons. Treatments were probabilistically ranked for their
effect with Surface Under Cumulative Ranking (SUCRA) score. All an­ 3.1.1. Efficacy measures
alyses were formed with R version 4.2 using `netmeta’ and `gemtc’ li­ For the primary outcome of seizure cessation within 60 min of
braries. Bayesian hierarchical models fitted with `gemtc’ relied on `Just treatment initiation, phenytoin (1 trial; risk ratio 0.53, 95% confidence
Another Gibbs Sampler’ using `rjags’ dependency. interval [CI] 0.36, 0.78; p = 0.002) in children and valproate in adults (2
The bias in the included studies was done using Risk of Bias (ROB) trials; risk ratio 0.65, 95% CI 0.51, 0.82; I2=42, p = 0.0004) were found
2.0 tool; [20] overall bias, based the five domains, was classified as to be significantly inferior to phenobarbital. Levetiracetam was similar
``low’’, ``some concerns’’ or ``high’’. To assess the quality of evidence, to phenytoin (5 trials; risk ratio 1.05, 95% CI 0.99, 1.11, I2=54; p =
we used GRADE Profiler software (V.3.6). [21] The rating was classified 0.12) and fosphenytoin (4 trials; risk ratio 1.05, 95% CI 0.94, 1.17,
as – no, serious and very serious limitation. I2=36; p = 0.42) (Fig. 1). Results for other comparisons are tabulated in
Table 1.
3. Results Phenytoin (1 trial; risk ratio 0.47, 95% CI 0.27, 0.84; p = 0.01) in
children was found to be inferior to Phenobarbital for the outcome of
After screening 3454 records, seventeen studies were included in this seizure cessation within 24 h (Supplemental Table 3).
systematic review; three studies [16,22,23] were excluded from con­
ventional meta-analyses (as single study available for respective com­ 3.1.2. Safety outcomes (Supplemental Table 4–6)
parison) and one study was excluded from NMA [12,15,22-36]. Eighteen The rates of deaths were comparable across different interventions.
studies were excluded from the review: SE not benzodiazepine resistant Phenytoin had significantly higher rates of intubation in children than
[37–49], third-line treatment reported [50,51], duplicate population phenobarbital in one study (risk ratio 3.05, 95% CI 1.24, 7.55, p = 0.02).
[52], outcomes reported at 7 days [53], and only nonconvulsive seizures Levetiracetam was safer than fosphenytoin (3 trials; risk ratio 0.62, 95%
included [54]. PRISMA flow diagram is shown in Supplemental Figure 1. CI 0.42, 0.91; I2=0, p = 0.01) and fosphenytoin was worse than val­
The study characteristics have been summarized in supplemental proate (1 trial, risk ratio 1.79, 95% CI 1.15, 2.79; p = 0.01). Valproate
Table 2A/2B Studies were done across many countries: India (7 studies), had lower rates of intubation than phenobarbital (3 trials; risk ratio
China (3 studies), Iran (2 studies) and 1 each from Australia/New Zea­ 0.18, 95% CI 0.04, 0.79; I2=0, p = 0.02). With respect to cardiovascular
land, Pakistan, South Africa, United Kingdom and United states of instability, valproate was safer than phenytoin (2 trials; risk ratio 0.13,
America. The study participants included children (10 studies), adults (6 95% CI 0.02, 0.98; I2=0, p = 0.05).
studies) and mixed age group (1 study). Different benzodiazepines
through different routes were used for initial treatment across studies 3.2. Network meta-analysis for primary outcome
(Supplemental Table 2B). Benzodiazepine resistance was defined as
refractoriness to either 1 (5 studies) or two doses (12 studies). Outcome 3.2.1. Network geometry
assessment also varied: clinical (13 studies), both clinical and EEG (4 The network graph (Fig. 2A) showed that 8 interventions have been
studies). reported for this outcome with multiple closed loops. The comparisons
All studies had low risk of bias [20] except four studies where ``some of Valproate-LD/phenobarbital (4 trials) and phenytoin/levetiracetam-
concerns’’ were noted due to the randomization process (two studies HD (3 trials) had most head-to-head trials. The most frequently re­
were quasi-randomized [25,27] and two had inadequate information on ported interventions included phenytoin (7 trials), valproate-LD (7 tri­
randomization process [24,33]). Seven studies were funded; conflict of als), phenobarbital (5 trials), levetiracetam-HD (5 trials) and
interest was unknown in one [36]. Four studies did not receive any levetiracetam-LD (4 trials). Most network estimates resulted from indi­
funding and the rest six studies did not declare any sources of funding. rect evidence, with most comparisons having a mean path length >2 and
unitary parallelism (Fig. 2B).

Fig. 1. Forest plot for the primary outcome ``seizure cessation within 60 min’’ for levetiracetam vs phenytoin/fosphenytoin comparison.

76
P. Jain et al. Seizure: European Journal of Epilepsy 102 (2022) 74–82

Table 1
Primary outcome (Seizure cessation within 60 min of treatment initiation) – Pairwise comparisons.
Comparison Population Number of studies/Total Risk ratio with 95% P value I [2] Interpretation
number of study participants confidence intervals (%)

Phenytoin/valproate-LD Mixed 2/170 1.00 (0.88,1.13) 1.00 0 Similar

[15,24]
Phenytoin/Phenobarbital Pediatric 1/69 0.53 0.002 NA PHB better
[22] (0.36, 0.78)
Levetiracetam/Phenytoin Mixed 5/1233 1.05 0.12 54 Similar
[15,25,26,29,32] (0.99, 1.11)
Levetiracetam/Phenytoin Pediatric 4/1189 1.03 0.64 60 Similar
[15,26,29,32] (0.92,
1.15)
Levetiracetam-LD/ Mixed 2/114 0.93 (0.77, 0.48 0 Similar
Phenytoin [15,25] 1.13)
Levetiracetam-HD/ Pediatric 3/975 1.04 (0.90, 0.61 67 Similar
Phenytoin [26,29,32] 1.20)*
Levetiracetam/ Mixed 4/544 1.05 (0.94,1.17) 0.42 36 Similar
Fosphenytoin [12,27,31,
33]
Levetiracetam/ Pediatric 4/383 1.07 0.23 40 Similar
Fosphenytoin [12,27,31, (0.96,
33] 1.19)
Levetiracetam-LD/ Pediatric 2/166 1.01 (0.92, 0.79 0 Similar
Fosphenytoin [27,33] 1.12)
Levetiracetam-HD/ Mixed 2/378 1.14 (0.87, 0.34 61 Similar
Fosphenytoin [12,31] 1.50)*
Levetiracetam-HD/ Mixed 1/320 0.96 (0.76,1.20) 0.71 NA Similar
Valproate-HD [12]
Valproate-LD/ Pediatric 1/67 0.93 (0.64–1.35) 0.74 NA Similar
Levetiracetam-LD [15]
Fosphenytoin/Valproate- Mixed 1/287 0.95 0.67 NA Similar
HD [12] (0.75, 1.21)
Valproate-LD/ Mixed 4/282 0.74 (0.49, 1.10)* 0.14 85 Similar
Phenobarbital [28,
34-36]
Valproate-LD/ Pediatric 2/140 0.82 0.63 92 Similar
Phenobarbital [28,36] (0.37,
1.83)*
Valproate-LD/ Adult 2/142 0.65 0.0004 42 PHB
Phenobarbital [34,35] (0.51, better
0.82)
Valproate-LD/Lacosamide Adult 1/66 1.09 0.60 NA Similar
[16] (0.78, 1.54)
Valproate infusion/ Adult 1/66 0.90 0.66 NA Similar
diazepam infusion [23] (0.57, 1.43)

*Random effect model.

Fig. 2. A. Network graph showing 8 reported interventions for the primary outcome. The size of the bubble is proportional to the risk difference between that
intervention and phenytoin. B. Proportion of direct and indirect evidence, mean network path length, and parallelism for each treatment comparison for pri­
mary outcome.

3.2.2. Fixed-effect frequentist meta-analysis possible pairwise comparisons along with their 95% CI are provided in
Phenobarbital and levetiracetam-HD were significantly superior to Table 2A.
phenytoin (Fig. 3A). The network risk difference estimates for all Network ranking (Table 3) showed that phenobarbital (p

77
P. Jain et al. Seizure: European Journal of Epilepsy 102 (2022) 74–82

Fig. 3. A- Forest plot for comparison of various interventions vs phenytoin for primary outcome. B- Network forest plot for the for primary outcome (n = 10) with
both direct and indirect evidence.

78
P. Jain et al. Seizure: European Journal of Epilepsy 102 (2022) 74–82

Table 2
Network risk difference estimates (with 95% confidence intervals) from fixed effect frequentist meta-analysis (A) and network risk difference estimates (with 95%
credible intervals) from Bayesian Hierarchical Random Effect Network Meta-analysis (B) for outcome ``Seizure cessation within 60 min’’. Risk difference <0 favors the
drug in the row and risk difference >0 favors the drug in the column. Bold figures indicate statistical significance. Figures in blue indicate presence of direct evidence
within the network estimate.
A - Fixed effect frequentist meta-analysis
FOS

0.02 (− 0.23; 0.28) LAC

− 0.07 (− 0.15; 0.02) − 0.09 (− 0.34; 0.16) LEV HD
− 0.00 (− 0.08; 0.08) − 0.02 (− 0.28; 0.23) 0.06 (− 0.03; 0.16) LEV LD
¡0.26 (¡0.40; ¡0.28 (¡0.52; ¡0.19 (¡0.31; ¡0.26 (¡0.40; PHB
¡0.12) ¡0.04) ¡0.07) ¡0.12)
− 0.00 (− 0.09; 0.09) − 0.02 (− 0.27; 0.22) 0.06 (0.02; 0.11) − 0.00 (− 0.09; 0.09) 0.26 (0.14; PHT
0.37)
− 0.06 (− 0.17; 0.05) − 0.08 (− 0.35; 0.19) 0.01 (− 0.10; 0.11) − 0.06 (− 0.18; 0.07) 0.20 (0.04; − 0.06 (− 0.17; VPA HD
0.36) 0.05)
− 0.04 (− 0.16; 0.08) − 0.06 (− 0.29; 0.17) 0.03 (− 0.07; 0.13) − 0.04 (− 0.16; 0.08) 0.22 (0.13; − 0.04 (− 0.13; 0.02 (− 0.12; 0.16) VPA LD
0.31) 0.06)
B - Bayesian Hierarchical Random Effect Network Meta-analysis
FOS
− 0.08 (− 0.52, 0.38) LAC
− 0.09 (− 0.25, 0.10) − 0.003 (− 0.44, 0.41) LEV_HD
− 0.02 (− 0.22, 0.15) 0.06 (− 0.43, 0.48) 0.06 (− 0.16, 0.27) LEV_LD
¡0.36 (¡0.65, − 0.27 (− 0.66, 0.10) ¡0.27 (¡0.53, ¡0.33 (¡0.61, PHB
¡0.07) ¡0.03) ¡0.04)
− 0.06 (− 0.26, 0.14) 0.03 (− 0.41, 0.42) 0.03 (− 0.13, 0.16) − 0.03 (− 0.24, 0.17) 0.30 (0.09, PHT
0.52)
− 0.003 (− 0.27, 0.27) 0.07 (− 0.45, 0.57) 0.08 (− 0.21, 0.36) 0.02 (− 0.29, 0.32) 0.36 (− 0.05, 0.053 (− 0.27, VPA_HD
0.74) 0.35)
− 0.14 (− 0.40, 0.12) − 0.05 (− 0.43, 0.30) − 0.05 (− 0.28, 0.17) − 0.12 (− 0.37, 0.14) 0.22 (0.06, − 0.09 (− 0.28, − 0.14 (− 0.49, VPA_LD
0.39) 0.11) 0.24)

Abbreviations: FOS-Fosphenytoin; LAC-lacosamide; LEV HD-Levetiracetam High dose; LEV LD-Levetiracetam low dose; PHB-Phenobarbital; PHT-Phenytoin; VPA HD-
Valproate high dose; VPA LD-Valproate low dose.

3.2.3. Bayesian hierarchical random effect meta-analysis

Table 3
The network estimates for risk differences for all possible pairwise
Network Rankings for various interventions for the primary outcome ``seizure
comparisons along with 95% credible intervals are shown in Table 2B.
cessation within 60 min’’, by both frequentist fixed effect and Bayesian random
effect network meta-analysis.
For eight pairwise comparisons having both direct and indirect evidence
to estimate network effects, no significant differences were found. The
Frequentist Fixed Effect Meta-analysis Bayesian Random Effect Meta-analysis
hierarchical ranking (Table 3) showed phenobarbital had the highest
Rank Intervention P score Rank Intervention SUCRA
probability of being the best among the analyzed interventions
1 Phenobarbital 0.99 1 Phenobarbital 0.74 (SUCRA=0.74) followed by valproate-LD (SUCRA=0.68) and
2 Levetiracetam HD 0.69 2 Valproate LD 0.68
3 Valproate HD 0.62 3 Levetiracetam HD 0.67
levetiracetam-HD (SUCRA=0.67).
4 Valproate LD 0.52 4 Lacosamide 0.59
5 Lacosamide 0.30 5 Phenytoin 0.52
6 Levetiracetam LD 0.30 6 Levetiracetam LD 0.46 3.3. Grade of evidence
7 Fosphenytoin 0.29 7 Valproate HD 0.34
8 Phenytoin 0.28 8 Fosphenytoin 0.02 Regarding the primary outcome measure of seizure control within
60 min, the quality of evidence was high for these comparisons; Leve­
score=0.99) had the highest probability of being the best among the tiracetam HD versus Valproate HD (no difference) and Fosphenytoin
studied interventions followed by levetiracetam-HD (p score=0.69) and versus Valproate HD (no difference); and low to very low quality for all
valproate-HD (p score=0.62). the other comparisons (Supplemental Table 7). For the secondary
Both direct and indirect risk difference estimates were available for outcome of need for intubation, the quality of evidence was moderate
ten pairwise comparisons (Fig. 3B). This showed that phenobarbital was for these comparisons; Levetiracetam HD versus Valproate HD (favoring
superior to both phenytoin (network risk difference estimate 0.26, 95% Valproate HD), Valproate LD versus Phenobarbital (favoring Valproate)
CI 0.14, 0.37) and valproate-LD (network risk difference estimate 0.22, and Fosphenytoin versus Valproate HD (favoring Valproate HD). The
95% CI 0.13, 0.31). Levetiracetam-HD was also significantly better than quality of evidence for need for intubation was low to very low for all
phenytoin (network risk difference estimate 0.06, 95% CI 0.02, 0.11). other comparisons. The quality of evidence for the other secondary
Rest of the comparisons did not show significant difference. Node- outcomes such as death, cardiovascular instability and seizure freedom
splitting did not show significant inconsistency between direct and in­ at 24 h was low to very low across all the comparisons.
direct evidence for any comparison.
A high degree of heterogeneity (I2=66.6%; 95% CI: 38.7%, 81.8%) 4. Discussion
was observed, contributed mainly by significant ``within design’’
inconsistency (75.8%, p = 0.0003). Phenobarbital/valproate-LD This systematic review and meta-analysis were aimed at generating
(68.2%, p = 0.0007) and phenytoin/ levetiracetam-LD (28.3%, p = best available effectiveness estimates for anti-seizure medications in
0.03) comparisons contributed most to ``within design’’ inconsistency. children and adults with benzodiazepine-resistant status epilepticus,
Under the assumption of a full design-by-treatment interaction random based on the current available evidence. Comparisons between the
effects model, the between-designs inconsistency remained following agents were studied- fosphenytoin, phenytoin, levetiracetam
insignificant. (HD and LD), valproate (LD and HD), phenobarbital, lacosamide and
diazepam infusion; in a total of 17 studies.

79
P. Jain et al. Seizure: European Journal of Epilepsy 102 (2022) 74–82

4.1. Efficacy side effects, in terms of raised transaminases and elevated ammonia
levels. Valproate has the risk of hepatotoxicity, and is contra-indicated
Based on low quality of evidence, for primary outcome, phenobar­ in patients with liver disease, which may not be evident when the pa­
bital was significantly better than phenytoin in children and valproate in tient is brought convulsing and needs emergency treatment. Valproate is
adults. Levetiracetam was similar to phenytoin or fosphenytoin. The contra-indicated in inherited metabolic including mitochondrial disor­
rates of deaths were comparable across different interventions. With ders, which may manifest in children with seizures and status
respect to intubation, phenobarbital was safer than phenytoin in chil­ epilepticus.
dren; levetiracetam was safer than fosphenytoin; and valproate was Another factor of concern is the speed of administration. Levetir­
superior to phenobarbital and fosphenytoin. Valproate was also noted to acetam, fosphenytoin and valproate can be administered within 10 min,
produce less cardiovascular instability when compared to phenytoin. whereas phenytoin and phenobarbital need to be infused over 20 min.
Network meta-analysis was limited by predominantly indirect evi­ Time is of the essence while treating status epilepticus. Other advan­
dence and high heterogeneity. Phenobarbital and levetiracetam-HD tages of levetiracetam and valproate include lesser risk of adverse effects
were significantly superior to phenytoin with respect to seizure cessa­ as compared to fosphenytoin. Both levetiracetam and valproate are
tion within 60 min. Network ranking demonstrated that phenobarbital broad spectrum drugs active against all types of seizures, and hence may
had the highest probability of being the best among the studied in­ be a good agent for maintenance therapy after the acute control of sei­
terventions followed by levetiracetam-HD and valproate-HD. Similar zures in children and adults with genetic generalized epilepsy or when
results were obtained in NMA done in children [55] and adults [56]. the type of seizure/ epilepsy syndrome is not clear.
Phenobarbital has the advantage of easy availability and low cost, a
huge benefit in low- and middle-income countries. 4.2. Strengths and limitations
This result however has to be viewed with caution. There were only
five trials with small number of patients which evaluated phenobarbital This systematic review and meta-analysis were carried out with
in BZD resistant SE. Four of these compared low-dose valproate with rigorous methodology. However, the source data comprised of multiple
Phenobarbital (total 282 patients), and one compared phenytoin with studies with small sample sizes and poor quality of evidence. Further,
phenobarbital (69 patients). No direct comparisons of phenobarbital NMA was limited by predominant indirect evidence and high
versus levetiracetam were available. Differences in dosages should be heterogeneity.
also taken into account in interpreting the results. The high relative The studied population was heterogenous with respect to age (chil­
efficacy of phenobarbital in the network comparisons for seizure dren, adults, elderly). The subgroup analyses (conventional meta-
cessation within 60 min could possibly be resulting from the high dose of anlyses; see Table 1) for pediatric age-group for comparisons levetir­
phenobarbital used in the adult studies [34,35]. These studies used an acetam/phenytoin, levetiracetam/fosphenytoin, levetiracetam-LD/
initial 20 mg/kg loading, which is higher than the 15 mg/kg recom­ fosphenytoin and valproate LD-phenobarbital, showed similar results
mendation [57]. Further, all five studies emanated from Asia where to their respective mixed age-group analyses. There were not enough
etiological profile for SE may be different, thus limiting its generaliz­ studies available to do similar meaningful subgroup analyses for NMA.
ability [58–61]. Further, the pathophysiology of status epilepticus and effect of anti-
When direct pairwise comparisons were used, based on high quality seizure medications on neuronal receptors are similar from children
of evidence, there was no difference between levetiracetam HD versus (beyond neonatal period) through adults [57]. This may allow to take
valproate HD, and fosphenytoin versus Valproate HD with respect to the more unified approach to SE management in children and adults.
primary outcome measure (seizure control within 60 min). There is a The heterogeneity may also arise from variability in reported etiol­
low quality evidence that phenobarbital may be better than low-dose ogies of SE, resource-setting and initial benzodiazepines used (agent,
valproate in the treatment of BZD resistant SE in adults (2 studies, 142 dose, route). Important variables like etiology and time to treatment
patients). were not uniformly reported/classified across studies. Frequent reliance
on clinical seizure cessation alone (and not EEG cessation of seizures)
4.1.1. Safety may also affect the robustness of the primary outcome measures.
In the direct comparisons for safety issues, there was a higher risk of
need for intubation and cardiovascular instability in the patients treated 5. Conclusion
with phenobarbital, as compared to those treated with valproate, though
this evidence was of low to moderate quality. Respiratory depression is a Our network meta-analysis of randomized trials, limited by pre­
known side effect of phenobarbital. Administration of phenobarbital dominant indirect evidence and high heterogeneity, revealed pheno­
after benzodiazepines, which also cause respiratory depression, is likely barbital to be the most effective agent for seizure cessation within 60
to lead to a higher risk of respiratory depression. Also, even though the min of administration in patients with benzodiazepine resistant status
direct comparisons of valproate versus phenobarbital, and phenytoin epilepticus. However, in direct comparisons, phenobarbital was found to
versus phenobarbital favored phenobarbital, this evidence was graded be associated with a higher risk of need for intubation and cardiovas­
as low quality. Another practical consideration is that high dose cular instability. Further, based on probabilistic rankings, levetiracetam
phenobarbital infusions are often used in refractory and super-refractory HD was also very effective and ranked higher than levetiracetam LD.
status epilepticus [62]. Hence, purely from a safety standpoint, it may be Direct comparisons revealed that levetiracetam HD, valproate HD and
desirable to try other medications (phenytoin, levetiracetam or val­ fosphenytoin are probably equally effective for seizure cessation within
proate) as second line, and perhaps reserve phenobarbital for refractory 60 min. Levetiracetam had a better safety profile than fosphenytoin.
SE. However, phenobarbital has the advantages of effectiveness, wide­ When deciding the drug to be used in patients with benzodiazepine
spread availability, and low cost. Thus, the choice of second line resistant status epilepticus, the choice needs to be decided by patient
anti-seizure medication must be modified in a given setting by other characteristics (age, underlying co-morbidities such as cardiac or liver
considerations including effectiveness, availability, and cost. disease, seizure type and need for long term anti-seizure medications),
There is low to moderate quality evidence that the safety profile of drug availability and cost, and evidence of efficacy and safety.
levetiracetam was better than fosphenytoin, in terms of lesser cardio­
vascular adverse events. There is low to moderate quality evidence that Funding
the safety profile of valproate was better than phenytoin or fospheny­
toin, in terms of less requirement for intubation and lesser cardiovas­ Research presented in this manuscript did not receive any funding
cular side effects. However, valproate has been associated with hepatic from federal, private, or not-for-profit agencies.

80
P. Jain et al. Seizure: European Journal of Epilepsy 102 (2022) 74–82

Declaration of Competing Interest [22] Burman RJ, Ackermann S, Shapson-Coe A, Ndondo A, Buys H, Wilmshurst JM.
A comparison of parenteral phenobarbital vs. parenteral phenytoin as second-line
management for pediatric convulsive status epilepticus in a resource-limited
The authors declare that they have no known competing financial setting. Front Neurol 2019;10:506.
interests or personal relationships that could have appeared to influence [23] Chen WB, Gao R, Su YY, et al. Valproate versus diazepam for generalized
the work reported in this paper. convulsive status epilepticus: a pilot study. Eur. J. Neurol. 2011;18(12):1391–6.
[24] Agarwal P, Kumar N, Chandra R, Gupta G, Antony AR, Garg N. Randomized study
of intravenous valproate and phenytoin in status epilepticus. Seizure 2007;16(6):
Acknowledgment 527–32.
[25] Chakravarthi S, Goyal MK, Modi M, Bhalla A, Singh P. Levetiracetam versus
phenytoin in management of status epilepticus. J Clin Neurosci 2015;22(6):
None. 959–63.
[26] Dalziel SR, Borland ML, Furyk J, et al. Levetiracetam versus phenytoin for second-
Supplementary materials line treatment of convulsive status epilepticus in children (ConSEPT): an open-
label, multicentre, randomised controlled trial. Lancet (London, England) 2019;
393(10186):2135–45.
Supplementary material associated with this article can be found, in [27] Handral A, Veerappa BG, Gowda VK, Shivappa SK, Benakappa N, Benakappa A.
the online version, at doi:10.1016/j.seizure.2022.09.017. Levetiracetam versus fosphenytoin in pediatric convulsive status epilepticus: a
randomized controlled trial. J Pediatr Neurosci 2020;15(3):252–6.
[28] Khajeh A, Yaghoubinia F, Yaghoubi S, Fayyazi A, Miri Aliabad G. Comparison of
References the effect of phenobarbital versus sodium valproate in management of children
with status epilepticus. Iran J Child Neurol 2018;12(4):85–93.
[1] Leppik IE, Derivan AT, Homan RW, Walker J, Ramsay RE, Patrick B. Double-blind [29] Lyttle MD, Rainford NEA, Gamble C, et al. Levetiracetam versus phenytoin for
study of lorazepam and diazepam in status epilepticus. JAMA 1983;249(11): second-line treatment of paediatric convulsive status epilepticus (EcLiPSE): a
1452–4. multicentre, open-label, randomised trial. Lancet (London, England) 2019;393
[2] Silbergleit R, Lowenstein D, Durkalski V, Conwit R, Neurological Emergency (10186):2125–34.
Treatment Trials (NETT) Investigators. RAMPART (Rapid Anticonvulsant [30] Misra UK, Dubey D, Kalita J. A randomized controlled trial of lacosamide versus
Medication Prior to Arrival Trial): a double-blind randomized clinical trial of the sodium valproate in status epilepticus. Epilepsia 2017.
efficacy of intramuscular midazolam versus intravenous lorazepam in the [31] Nalisetty S, Kandasamy S, Sridharan B, Vijayakumar V, Sangaralingam T,
prehospital treatment of status epilepticus by paramedics. Epilepsia 2011;52 Suppl Krishnamoorthi N. Clinical effectiveness of levetiracetam compared to
8(Suppl 8):45–7. fosphenytoin in the treatment of benzodiazepine refractory convulsive status
[3] Trinka E, Höfler J, Zerbs A, Brigo F. Efficacy and safety of intravenous valproate for epilepticus. Indian J Pediatr 2020;87(7):512–9.
status epilepticus: a systematic review. CNS Drugs 2014;28(7):623–39. [32] Noureen N, Khan S, Khursheed A, et al. Clinical efficacy and safety of injectable
[4] Guterman EL, Betjemann JP, Aimetti A, et al. Association between treatment levetiracetam versus phenytoin as second-line therapy in the management of
progression, disease refractoriness, and burden of illness among hospitalized generalized convulsive status epilepticus in children: an open-label randomized
patients with status epilepticus. JAMA Neurol 2021;78(5):588–95. controlled trial. J Clin Neurol 2019;15(4):468–72.
[5] Strzelczyk A, Ansorge S, Hapfelmeier J, Bonthapally V, Erder MH, Rosenow F. [33] Senthil Kumar CS, Selvakumar P, Kowsik M. Randomized controlled trial of
Costs, length of stay, and mortality of super-refractory status epilepticus: a levetiracetam versus fosphenytoin for convulsive status epilepticus in children. Int
population-based study from Germany. Epilepsia 2017;58(9):1533–41. J Pediatr Res 2018;5(4):237–42.
[6] Burman RJ, Rosch RE, Wilmshurst JM, et al. Why won’t it stop? The dynamics of [34] Su Y, Huang H, Jiang M, et al. Phenobarbital versus valproate for generalized
benzodiazepine resistance in status epilepticus. Nat Rev Neurol 2022;18(7): convulsive status epilepticus in adults (2): a multicenter prospective randomized
428–41. controlled trial in China (China 2-P vs. V). Epilepsy Res. 2021;177:106755.
[7] Trinka E. What is the relative value of the standard anticonvulsants: phenytoin and [35] Su Y, Liu G, Tian F, et al. Phenobarbital versus valproate for generalized convulsive
fosphenytoin, phenobarbital, valproate, and levetiracetam? Epilepsia 2009;50 status epilepticus in adults: a prospective randomized controlled trial in China.
(Suppl 12):40–3. CNS Drugs 2016;30(12):1201–7.
[8] DeToledo JC, Ramsay RE. Fosphenytoin and phenytoin in patients with status [36] Malamiri RA, Ghaempanah M, Khosroshahi N, Nikkhah A, Bavarian B, Ashrafi MR.
epilepticus: improved tolerability versus increased costs. Drug Saf 2000;22(6): Efficacy and safety of intravenous sodium valproate versus phenobarbital in
459–66. controlling convulsive status epilepticus and acute prolonged convulsive seizures
[9] Popławska M, Borowicz KK, Czuczwar SJ. The safety and efficacy of fosphenytoin in children: a randomised trial. Eur. J. Paediatr. Neurol. 2012;16(5):536–41.
for the treatment of status epilepticus. Expert Rev Neurother 2015;15(9):983–92. [37] Chitsaz A, Mehvari J, Salari M, Gholami F, Najafi M-R. A comparative assessment
[10] Abdelgadir I, Hamud A, Kadri A, et al. Levetiracetam for convulsive status the efficacy of intravenous infusion of sodium valproate and phenytion in the
epilepticus in childhood: systematic review and meta-analysis. Arch Dis Child treatment of status epilepticus. Int J Prev Med 2013;4(Suppl 2):S216–21.
2020. [38] Gilad R, Izkovitz N, Dabby R, et al. Treatment of status epilepticus and acute
[11] Angurana SK, Suthar R. Efficacy and safety of levetiracetam vs. phenytoin as repetitive seizures with i.v. valproic acid vs phenytoin. Acta Neurol Scand 2008;
second line antiseizure medication for pediatric convulsive status epilepticus: a 118(5):296–300.
systematic review and meta-analysis of randomized controlled trials. J Trop [39] Gujjar AR, Nandhagopal R, Jacob PC, et al. Intravenous levetiracetam vs phenytoin
Pediatr 2021;67(2). for status epilepticus and cluster seizures: a prospective, randomized study. Seizure
[12] Chamberlain JM, Kapur J, Shinnar S, et al. Efficacy of levetiracetam, fosphenytoin, 2017;49:8–12.
and valproate for established status epilepticus by age group (ESETT): a double- [40] Misra UK, Kalita J, Patel R. Sodium valproate vs phenytoin in status epilepticus: a
blind, responsive-adaptive, randomised controlled trial. Lancet (London, England) pilot study. Neurology 2006;67(2):340–2.
2020;395(10231):1217–24. [41] Mundlamuri RC, Sinha S, Subbakrishna DK, et al. Management of generalised
[13] DeMott JM, Slocum GW, Gottlieb M, Peksa GD. Levetiracetam vs. phenytoin as convulsive status epilepticus (SE): a prospective randomised controlled study of
2nd-line treatment for status epilepticus: a systematic review and meta-analysis. combined treatment with intravenous lorazepam with either phenytoin, sodium
Epilepsy Behav 2020;111:107286. valproate or levetiracetam–Pilot study. Epilepsy Res. 2015;114:52–8.
[14] Feng Y, Chen Y, Jia Y, et al. Efficacy and safety of levetiracetam versus (fos) [42] Nazir M, Tarray RA, Asimi R, Syed WA. Comparative efficacy of IV phenytoin, IV
phenytoin for second-line treatment of epilepticus: a meta-analysis of latest valproate, and IV levetiracetam in childhood status epilepticus. J Epilepsy Res
randomized controlled trials. Seizure 2021;91:339–45. 2020;10(2):69–73.
[15] Vignesh V, Rameshkumar R, Mahadevan S. Comparison of phenytoin, valproate [43] Nene D, Mundlamuri RC, Satishchandra P, et al. Comparing the efficacy of sodium
and levetiracetam in pediatric convulsive status epilepticus: a randomized double- valproate and levetiracetam following initial lorazepam in elderly patients with
blind controlled clinical trial. Indian Pediatr 2020;57(3):222–7. generalized convulsive status epilepticus (GCSE): a prospective randomized
[16] Misra UK, Dubey D, Kalita J. Comparison of lacosamide versus sodium valproate in controlled pilot study. Seizure 2019;65:111–7.
status epilepticus: a pilot study. Epilepsy Behav 2017;76:110–3. [44] Shaner DM, McCurdy SA, Herring MO, Gabor AJ. Treatment of status epilepticus: a
[17] Strzelczyk A, Zöllner JP, Willems LM, et al. Lacosamide in status epilepticus: prospective comparison of diazepam and phenytoin versus phenobarbital and
systematic review of current evidence. Epilepsia 2017;58(6):933–50. optional phenytoin. Neurology 1988;38(2):202–7.
[18] Hutton B, Salanti G, Caldwell DM, et al. The PRISMA extension statement for [45] Sreenath TG, Gupta P, Sharma KK, Krishnamurthy S. Lorazepam versus diazepam-
reporting of systematic reviews incorporating network meta-analyses of health care phenytoin combination in the treatment of convulsive status epilepticus in
interventions: checklist and explanations. Ann Intern Med 2015;162(11):777–84. children: a randomized controlled trial. Eur. J. Paediatr. Neurol. 2010;14(2):
[19] Page MJ, Moher D, Bossuyt PM, et al. PRISMA 2020 explanation and elaboration: 162–8.
updated guidance and exemplars for reporting systematic reviews. Bmj 2021;372: [46] Treiman DM, Meyers PD, Walton NY, et al. A comparison of four treatments for
n160. generalized convulsive status epilepticus. Veterans Affairs Status Epilepticus
[20] Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias Cooperative Study Group. N Engl J Med 1998;339(12):792–8.
in randomised trials. Bmj 2019;366:l4898. [47] Wani G, Imran A, Dhawan N, Gupta A, Giri JI. Levetiracetam versus phenytoin in
[21] Guyatt GH, Oxman AD, Schünemann HJ, Tugwell P, Knottnerus A. GRADE children with status epilepticus. J Family Med Prim Care 2019;8(10):3367–71.
guidelines: a new series of articles in the Journal of Clinical Epidemiology. J Clin [48] Pooja Sharma SM, Bamnawat Subhash. Levetiracetam versus phenytoin for
Epidemiol 2011;64(4):380–2. treatment of convulsive status epilepticus in pediatric population: a randomized
controlled trial. Int J Contemp Pediatr 2019;6(2):741–5.

81
P. Jain et al. Seizure: European Journal of Epilepsy 102 (2022) 74–82

[49] Kanika Singh AA, Faridi MMA, Sharma S. IV Levetiracetam versus IV Phenytoin in systematic review and network meta-analysis. Epilepsy Behav 2019;101(Pt B):
childhood seizures: a randomized controlled trial. J Pediaric Neurosci 2018;12(2): 106466.
158–64. [57] Glauser T, Shinnar S, Gloss D, et al. Evidence-based guideline: treatment of
[50] Abbaskhanian A, Sheidaee K, Charati JY. Comparison of the effect of continuous convulsive status epilepticus in children and adults: report of the guideline
intravenous infusion of sodium valproate and midazolam on management of status committee of the American Epilepsy Society. Epilepsy Curr 2016;16(1):48–61.
epilepticus in children. Arch Pediatr 2021;28(8):696–701. [58] Chetan C, Sharma S, Mathur SB, Jain P, Aneja S. Clinical profile and short-term
[51] Mehta V, Singhi P, Singhi S. Intravenous sodium valproate versus diazepam outcome of pediatric status epilepticus at a tertiary-care center in Northern India.
infusion for the control of refractory status epilepticus in children: a randomized Indian Pediatr 2020;57(3):213–7.
controlled trial. J. Child Neurol. 2007;22(10):1191–7. [59] Gaillard W, Goodkin H, Abend N, et al. Etiology of pediatric refractory convulsive
[52] Kapur J, Elm J, Chamberlain JM, et al. Randomized trial of three anticonvulsant status epilepticus. Results from the pediatric status epilepticus research group
medications for status epilepticus. N Engl J Med 2019;381(22):2103–13. (pSERG) (S29.005). Neurology 2014;82(10 Supplement).
[53] Amiri-Nikpour MR, Nazarbaghi S, Eftekhari P, et al. Sodium valproate compared to [60] Holtkamp M, Othman J, Buchheim K, Meierkord H. Predictors and prognosis of
phenytoin in treatment of status epilepticus. Brain Behav 2018;8(5):e00951. refractory status epilepticus treated in a neurological intensive care unit. J Neurol
[54] Husain AM, Lee JW, Kolls BJ, et al. Randomized trial of lacosamide versus Neurosurg Psychiatry 2005;76(4):534–9.
fosphenytoin for nonconvulsive seizures. Ann Neurol 2018;83(6):1174–85. [61] Lingappa L, Konanki R, Patel R, Vooturi S, Jayalakshmi S. Clinical profile and
[55] Zhang Y, Liu Y, Liao Q, Liu Z. Preferential antiseizure medications in pediatric outcome of refractory convulsive status epilepticus in older children from a
patients with convulsive status epilepticus: a systematic review and network meta- developing country. Seizure 2016;36:31–5.
analysis. Clin Drug Investig 2021;41(1):1–17. [62] Tiamkao S, Mayurasakorn N, Suko P, et al. Very high dose phenobarbital for
[56] Brigo F, Del Giovane C, Nardone R, Trinka E, Lattanzi S. Intravenous antiepileptic refractory status epilepticus. J Med Assoc Thai 2007;90(12):2597–600.
drugs in adults with benzodiazepine-resistant convulsive status epilepticus: a

82