Authors: Mohammed Alsabri, Sarah Makram Elsayed, Shree Rath, Hamza A. Abdul-Hafez, Dina Essam Abo-elnour, Elizabeth Kurian
Categories: Systematic Review, Levetiracetam, Phenytoin, Valproate, Epilepsy, Pediatric, Emergency department
Source: BMC Neurology
Authors: Mohammed Alsabri, Sarah Makram Elsayed, Shree Rath, Hamza A. Abdul-Hafez, Dina Essam Abo-elnour, Elizabeth Kurian
Status epilepticus is one of the most commonly occuring emergencies among children across the world. Time is crucial in the treatment of SE, with increasing risk of long term adverse events and sequelae with delay in treatment or action of drugs. This systematic review and meta-analysis aims to compare levetiracetam, a drug known for its comparatively safer profile, with other routinely used drugs in pediatric SE like phenytoin, fosphenytoin and valproate.
A comprehensive literature search was conducted across four databases from 1996 till November 2024. All original studies evaluating the efficacy of levetiracetam vis-a-vis other anti-seizure medications in pediatric children in an emergency setting were included in the study. Data analysis was conducted using RevMan software, using a random-effects model.
A total of fourteen studies, comprising a patient population of 2,473, were included for further quantitative analysis. No differences were noted between the drugs when comparing seizure termination and recurrence at 24 h. Levetiracetam notably reduced the time to cessation of seizures when compared to phenytoin or fosphenytoin (MD=-3.97, 5% CI [-6.18, -1.76], p = 0.0004) and length of ICU stay over phenytoin (MD = 0.77, 95% CI [0.54, 1.00], p < 0.00001). A lower risk of adverse events was noted on use of levetiracetam over fosphenytoin (RR = 0.62, 95% CI [0.40, 0.96], p = 0.03); however risk of agitation was the least on use of phenytoin (RR = 3.90, 95% CI [1.42, 10.73], p = 0.008). Non significant differences in mortality rates were observed.
The study concludes better immediate effects of levetiracetam such as faster cessation of seizures. Levetiracetam was also suggested to stabilize patients faster, as implied by the lesser ICU stay and lower risk of adverse events. Further studies are needed to evaluate the efficacy of levetiracetam over other anti-seizure medications.
The online version contains supplementary material available at 10.1186/s12883-025-04323-0.
Convulsive Status epilepticus (CSE) is the most common neurological emergency, especially among children with an incidence of 20 per 100,000 individuals annually. CSE is characterized by continuous seizure activity lasting more than 5 min or by more than one sequential seizure without return to baseline consciousness [1]. If left untreated, CSE can lead to significant morbidity and mortality, reaching up to 20% of cases [2–4].
Initial management of CSE focuses on stabilizing the patient’s airway, breathing, and circulation, followed by seizure control using anti-seizure medications, such as Benzodiazepines (BZDs), phenytoin, levetiracetam, and fosphenytoin. Once seizure is controlled, the underlying cause of CSE must be identified and treated. Timely and effective management of CSE is essential to lower the risks of seizure progression, mortality, neurological damage, and other complications associated with prolonged seizures [1, 5].
Benzodiazepines (BZDs), such as lorazepam, diazepam, and midazolam, are the best choice as a first-line treatment of CSE, supported by evidence from randomized controlled trials (RCTs) and meta-analyses [6, 7]. In managing status epilepticus, treatment typically begins with weight-based intravenous benzodiazepines. For example, lorazepam is usually administered at 0.1 mg/kg with a dose of up to around 4 mg, while midazolam is given at roughly 0.2 mg/kg, not exceeding 5–10 mg per dose. If seizures persist, a second dose is repeated after 5–15 min. In situations where IV access is delayed or patients do not respond to the initial dosing, alternative routes, such as intramuscular or intranasal administration, and alternative agents like clonazepam, owing to its unique pharmacokinetic properties, can be considered [8]. However, one-third of patients fail to respond to the maximum dose of BZDs. This condition is called benzodiazepine-refractory CSE. In such cases, second-line treatments, including levetiracetam, phenytoin, fosphenytoin, and valproic acid, are commonly needed. There are no sufficient studies about the second-line intervention in pediatric CSE who fail to improve after BZD [9].
While intravenous phenytoin has historically been the preferred second-line agent for pediatric CSE, its use is associated with significant adverse effects, including hypotension, cardiac arrhythmias, pancytopenia, and Stevens-Johnson syndrome (SJS). In contrast, levetiracetam is emerging as a promising alternative due to its less adverse effect, easy use, and rapid onset, levetiracetam is emerging agent as an alternative second-line of CSE in pediatrics with better safety and efficacy [10].
While status epilepticus occurs across all pediatric age groups, treatment approaches may vary significantly between neonates and older children. Neonates (particularly those under 1–2 months of age) present unique pathophysiological and pharmacokinetic challenges, for which phenobarbital remains the standard first-line therapy. However, high-quality comparative evidence in this population remains limited. Given the heterogeneity in treatment response and safety profiles across age groups, this analysis focuses specifically on children older than one month, for whom levetiracetam, phenytoin, and fosphenytoin are commonly considered second-line options. This systematic review aims to evaluate the efficacy and safety of levetiracetam as a second-line treatment for pediatric CSE in the emergency department, providing a comprehensive analysis of its role in this setting.
This systematic review strictly adhered to the PRISMA framework [11] and the Cochrane Collaboration Guidelines [12] to ensure comprehensive and methodologically sound research. Every step of the process was designed to increase reproducibility, avoid bias, and ensure the reliability and clinical relevance of the findings for pediatric status epilepticus management in emergency settings.
An exhaustive and meticulous literature search was conducted across the premier medical databases—PubMed, Embase, Cochrane Library, ClinicalTrials, Scopus, and Web of Science—up to November 2024. The search strategy was crafted with a precise combination of terms, including “Pediatric,” “Children,” “Emergency department,” “convulsive status epilepticus,” “levetiracetam,” “fosphenytoin,” “valproate,” “placebo,” “phenytoin,”and “seizure,” ensuring a far-reaching scope to capture all relevant studies.
To reinforce the robustness of the search, we incorporated a backward reference search of the included studies and explored systematic reviews to uncover any potentially omitted articles. Additionally, we proactively contacted the corresponding authors of key studies when further clarification or additional data was required, thus ensuring data completeness and minimizing the risk of incomplete reporting.
The four-step screening process applied was designed to systematically refine the initial search
All citations were imported into EndNote 20 for de-duplication.Two independent reviewers rigorously screened titles and abstracts to exclude studies outside the scope.The full texts of the remaining studies were scrutinized to confirm eligibility.The reference lists of included studies were reviewed for any additional relevant articles.
This independent, multi-layered review process, with conflicts resolved through consensus or a third-party adjudicator, ensured the methodological integrity of study selection, leaving no room for subjective bias.
The eligibility criteria were deliberately stringent to ensure relevance to the pediatric SE
Studies included children patients aged 1 month to 18 years, with convulsive status epilepticus presenting to emergency departments or pediatric intensive care units.We included randomized controlled trials and observational studies that addressed diagnostic strategies, management approaches, or outcomes related to pediatric SE.
Excluded
Studies involving adults (> 18 years) only.Non-acute setting (e.g. Outpatient or rehabilitation setting).Did not meet the focus on pediatric SE in emergency settings,Were not published in English, or.Were abstracts, conference proceedings, editorials, studies without a comparator group, or case reports.
This strict filtering ensured that only the most robust and clinically significant data were included in the review.
A structured and systematic approach to data extraction was implemented. Two reviewers independently extracted data using a pre-defined, comprehensive extraction form, capturing all necessary study variables. These included patient demographics, clinical presentation, diagnostic methods, treatment modalities, and key outcomes such as mortality and ICU admissions. Any disagreements during extraction were resolved through thorough discussion or referral to a third expert reviewer, guaranteeing both accuracy and completeness.
Quality assessment was conducted separately for included trials and cohort studies by two independent authors (H.I, S.R). RCTs were assessed using RoB-2 tool, grading studies as having “low”, “some concerns”, or high risk of bias. Further, observational studies were graded using Newcastle-Ottawa Scale, with studies having > 6 points judged as “high” quality. All conflicts were resolved by a third reviewer (M.A).
The analysis was conducted on RevMan software. Dichotomous data were presented as risk ratio (RR), and 95% confidence interval (CI), while the continuous data were presented as mean difference (MD), and 95% CI. The data were considered significant when p-value was less than 0.5, and homogenous when p was equal or more than 0.1. For resolving the heterogeneity, the random-effect model was used. Also, subgroup analyses were conducted according to different parameters when applicable (Comparators: Phenytoin, Fosphenytoin, and Valproate/Study Designs: RCTs, and observational/Doses or quality assessment of studies), and leave-one-out was also done when applicable.
The initial search from different databases yielded 1,811 records. Following duplicate removal, the number of records eligible for title and abstract screening was 1,375 studies. Following title and abstract screening, 55 eligible records were further assessed via full-text screening. Finally, 14 studies [10, 13–25] were included in our study. (Fig. 1)
Fig. 1PRISMA flow diagram
The included studies were randomized controlled trials (RCTs), except for three studies which were cohort studies, and Quasi-experimental [10, 17, 25]. The countries involved were India, Pakistan, Turkey, the UK, and the USA. The included sample size was 2,473 participants. The follow-up duration mostly ranged between 15 min and 24 h, but included up to three weeks. Regarding Levetiracetam dose, it ranged between 20 and 60 mg/kg, all were intravenous (IV), except for Dalziel et al. [14], which involved both IV or intraosseous, with no add-on medications. The control involved Phenytoin, Fosphenytoin, and Valproate. (Supplementary Table 1) The included trial by Chamberlain et al. is a substudy of the ESETT trial [26].
The mean age of the included population in the interventional group mainly ranged between 2.5 and 8.31 years, and it was mostly similar among the control groups. Among the studies that reported the seizure type, the generalized one was more common. Several etiological factors were reported in different studies, which included fever, trauma, meningitis, cryptogenic, and others. (Supplementary Table 2)
Included RCTs and quasi-randomized trials were assessed using RoB-2 too. Regarding the included RCTs, three of the studies had a low risk of bias, three had a high risk of bias due to high risk in measurement of outcome, missing outcome data, or selection of reported results, and the others had some concerns (Supplementary Fig. 1). As for the observational study, Besli et al. scored eight points, while Kole et al. scored six points, with Besli et al. being judged as “high quality”, and Kole et al. judged as “poor quality” due to inadequacy of follow-up (Supplementary Table 3).
The overall analysis, based on different comparators from 11 studies (N = 2466 participants), did not show significant differences amongst the groups (RR = 1.03, 95% CI [0.96, 1.10], p = 0.45), and the results were heterogeneous (p = 0.004; I² =57%). The subgroup analyses according to different comparators (Phenytoin, Fosphenytoin, and Valproate) also had non-significant variations between the (RR = 1.04, 95% CI [0.96, 1.13], p = 0.31), (RR = 0.99, 95% CI [0.89, 1.09], p = 0.82), and (RR = 0.98, 95% CI [0.81, 1.20], p = 0.85), respectively, as shown in Fig. 2A (Forest Plot)/Supplementary Fig. 2 (Funnel plot). In Levetiracetam VS Valproate, the heterogeneity was solved after removing Vignesh et al. The analysis did not show significant differences between the groups (RR = 0.88, 95% CI [0.76, 1.03], p = 0.11), and the results were homogenous (p = 0.30; I² =5%).
Fig. 2A Forest plot of different comparators for seizure termination rate, 24 h. B Forest plot of different study designs for seizure termination rate, 24 h
Meanwhile, the subgroup analyses based on study designs also did not show significant differences between Levetiracetam and Phenytoin or Fosphenytoin (RR = 1.05, 95% CI [0.98, 1.12]p = 0.20). However, the results were heterogeneous. No significant differences were found among both the RCTs or observational studies (RR = 1.03, 95% CI [0.96, 1.11], p = 0.36), and (RR = 1.18, 95% CI [0.86, 1.62], p = 0.31), respectively, but the results were homogenous in the observational studies (p = 0.11, I² =60%) (Fig. 2B).
The overall analysis, based on different comparators representing 1535 patients from eight studies, had non-significant differences between the groups (RR = 0.92, 95% CI [0.76, 1.13], p = 0.44), and the results were heterogeneous (p = 0.06; I² =45%). The subgroup analyses according to different comparators (Phenytoin/Fosphenytoin, and Valproate) also showed no significant differences between the groups, p = 0.44, and p = 0.58, respectively, as shown in Fig. 3A. The results were homogenous in Levetiracetam VS Valproate analysis (p = 0.20; I² =40%). Also, in Levetiracetam VS Phenytoin/Fosphenytoin, the heterogeneity was solved after removing Vignesh et al. The analysis still did not show significant differences between the groups (RR = 0.99, 95% CI [0.79, 1.24], p = 0.93), and the results were homogenous (p = 0.13; I² =39%).
Fig. 3A Forest plot of different comparators for ICU admission. B Forest plot of different study designs for ICU admission
In the subgroup analyses according to the study designs, in the observational studies, there was significant favoring of Levetiracetam compared to Phenytoin/Fosphenytoin (RR = 0.43, 95% CI [0.22, 0.83] p = 0.01), and non-significant differences between them amongst the RCTs (RR = 1.05, 95% CI [0.91, 1.21], p = 0.49, respectively) and both results were homogenous. (Fig. 3B)
The analysis of seven included studies, accounting for 727 participants, did not show significant differences between the groups (RR = 0.89, 95% CI [0.63, 1.26], p = 0.51), and the results were homogenous (p = 0.28; I² =20%). Figure 4A.
Fig. 4A Forest plot of 24 h seizure recurrence. B Forest plot of time of cessation, doses
The overall analysis, which involved 912 participants from seven studies, showed a significant favoring of Levetiracetam compared to Phenytoin/Fosphenytoin (MD=−3.97, 5% CI [−6.18, −1.76], p = 0.0004) however, the results were heterogeneous (p < 0.00001); I² = 92%). In terms of subgroup analyses based on the doses, such significant favoring was found in the 40 mg/kg group dose (p = 0.004), however, the results were also heterogeneous which could not be resolved. Figure 4B.
The overall analysis showed a significant favoring of the control compared to Levetiracetam; however, the results were heterogeneous. Regarding the subgroup analysis based on the comparators, the significant favoring was only found with phenytoin over Levetiracetam (MD = 0.77, 95% CI [0.54, 1.00], p < 0.00001), and the results were homogenous (p = 0.11; I² =62%), with non-significant differences amongst other comparators. (Supplementary Fig. 3)
The overall analysis did not show significant differences between the groups (MD = 0.07, 95% CI [−0.19, 0.32], p = 0.61), and the results were homogenous (p = 0.59; I² =0%). The subgroup analysis also showed non-significant variations and homogenous results. (Supplementary Fig. 4)
The overall analysis showed a significant favoring of Levetiracetam over the control (RR = 0.62, 95% CI [0.40, 0.96], p = 0.03), and the results were heterogeneous (p = 0.0008; I² =64%). In the subgroup analysis, such significance was only found in Levetiracetam VS Fosphenytoin (RR = 0.37, 95% CI [0.25, 0.55], p < 0.00001), and the results were homogenous (p = 0.89; I² =0%). Meanwhile, in Levetiracetam VS Valproate analysis, there were non-significant differences amongst groups (RR = 0.81, 95% CI [0.49, 1.33], p = 0.41) and the results were homogenous (p = 0.42; I² =0%). (Fig. 5A).
Fig. 5A Forest plot of Number of Adverse Events. B Forest plot of Mortality
In Levetiracetam VS Phenytoin, the initial heterogeneity, shown in Fig. 5A (p = 0.003; I² =65%), was solved after removing Besli et al., which is an observational study while the others are RCTs. The analysis did not show significant differences between the groups (RR = 0.86, 95% CI [0.59, 1.25], p = 0.42), and the results were homogenous (p = 0.24; I² =24%).
The overall analysis of 1028 patients from six studies did not show significant differences between the groups (RR = 0.89, 95% CI [0.37, 2.12], p = 0.79), and the results were homogenous (p = 0.87; I² =0%). Similarly, the subgroup analysis, based on different comparators (Phenytoin/Fosphenytoin or Valproate) showed non-significant variations (RR = 1.02, 95% CI [0.39, 2.66], p = 0.97, and (RR = 0.56, 95% CI [0.07, 4.31], p = 0.58), respectively), and homogenous results (Fig. 5B).
The analysis of 426 participants from three studies showed a significant favor with phenytoin over Levetiracetam (RR = 3.90, 95% CI [1.42, 10.73], p = 0.008), and the results were homogenous (p = 0.28; I² =14%). Supplementary Fig. 5.
The overall analysis did not show significant differences between the groups (RR = 0.15, 95% CI [0.02, 1.13], p = 0.006), however, the results were heterogeneous (p = 0.07; I² =62%). In the overall analysis, the heterogeneity was solved after removing Besli et al., which is an observational study while the others are RCTs. The overall analysis did not show significant differences between the groups (RR = 0.41, 95% CI [0.09, 1.91], p = 0.26), and the results were homogenous (p = 0.31; I² =1%). Supplementary Fig. 6.
From this study, it was noted that levetiracetam and other anticonvulsant medications like valproate and phenytoin equaled in efficacy when comparing seizure termination rates and recurrence at 24 h, incidence of ICU stay, and length of hospital stay. On the other hand, on quantifying immediate outcomes like time to cessation of seizure, fastest recovery was noted in the levetiracetam arm, more so at 40 mg/kg when compared to 20 mg/kg dosage. Levetiracetam also reduced the length of ICU admission, thus implying faster stabilization of patients. While considering the drugs’ safety profile, levetiracetam significantly reduced risk of incident adverse events. However, no differences were noted in mortality rates and incidence of hypotension between the drugs. Lesser risk of agitation was noted on use of phenytoin over levetiracetam.
Levetiracetam was found to be non-inferior, and equal in efficacy to drugs like phenytoin, fosphenytoin and valproate when comparing seizure termination and recurrence rates at 24 h. A previous meta-analysis by Feng et al. noted similar outcomes, when analyzing efficacy of levetiracetam versus fosphenytoin (RR] 0.94; 95% CI 0.87 to 1.01) [27]. While these results were noted in an adult population, most of the included studies in this meta-analysis noted similar findings [14, 24]. However, meta-analysis by Angurana et al. concluded a higher risk of seizure recurrence at 24 h on use of phenytoin over levetiracetam in children [28]. This would imply that recent included studies have found no significant effects of levetiracetam over other anti-seizure. However, of note is the study by Naeem et al. [19], where a lower risk of cessation of seizure was noted over Levetiracetam. However, these findings may have been biased due to the clinical judgment of cessation of seizures over usage of EEG. Additionally, non-blinding of the physicians administering the drugs may have contributed to some degree of observation bias. In all, no advantages of usage of levetiracetam is noted over other anti-seizure when quantifying seizure cessation or recurrences at 24 h. Similarly, no benefits were noted in the length of hospital stay, implying similar end-outcomes of all drugs. However, certain causative factors and etiology may influence these outcomes; a study by Berg et al. describes the higher risk of recurrence among individuals with a history of febrile seizure [29]. Family history of seizure also increases the risk of recurrence manifold times [30]. Thus, an objective evaluation of seizure termination and recurrence rates needs to be made, while considering etiological factors.
Levetiracetam was significantly effective over phenytoin or fosphenytoin resulting in faster termination of seizure, with better efficacy noted using 40 mg/kg dose of levetiracetam. These findings were uniformly reflected across all included studies except Nalisetty et al. [20] and Vignesh et al. [24], who noted a faster, albeit insignificant termination of seizure on use of levetiracetam. Lower sample size and open-label trials may have influenced the strength of the findings in the two studies. Lyttle et al. [18] hypothesizes that the faster rate of infusion of the drug over phenytoin may have led to effective termination of seizures. Additionally, the higher dose ensures immediate actions at the neuro-synaptic junction, thus resulting in faster onset of action. This would also result in reduced length of ICU stay due to the faster onset of action, as evidenced by the findings in our study.
ICU admission rates were similar across included studies, with no benefits noted on use of levetiracetam. A significant effect of levetiracetam was noted only across included studies, while non-significant differences were noted overall and among RCTs, thus strengthening the findings. Convulsive SE among pediatrics require intensive observation and treatment, resulting in a uniform admission to the ICU, even when there is improvement in vitals of the patient [31]. Our patient population focuses on children specifically presenting to the emergency department, and these patients are refractory to benzodiazepines, thus necessitating the use of levetiracetam. These patients would have established refractory SE, and guidelines recommend the admission of these patients to the pediatric ICU [32]. Thus, ICU admission rates are similar across all cases of refractory SE, as noted in this study.
Finally, levetiracetam was associated with lower incidence of adverse events over other anti-seizure. This is mainly attributed to the lesser receptor-level interactions when compared to other drugs, which have a multi-systemic effect [33]. The safety profile of levetiracetam over other drugs is the major reason for its recommendation for use in pregnant women [34]. Nakamura et al. noted higher side effects on use of fosphenytoin over levetiracetam in adults [35]. The previous meta-analysis by Klowak et al. noted lesser rates of respiratory side effects on use of levetiracetam, similar to our findings on lesser risk of hypotension and total adverse events [36]. However, levetiracetam is known to cause certain behavioral side effects in children; Halma et al. noted a relative risk of nearly two times for hostility and other behavioral adverse events on use of levetiracetam over placebo [37]. In this study, a higher rate of agitation was noted on use of levetiracetam over phenytoin. Thus the choice of drugs must be made carefully while considering previous adverse events in children, especially if they have a history of neurobehavioral adverse events on use of levetiracetam. However, clinicians may prefer levetiracetam in practice due to its favorable safety profile, easier administration, and minimal need for cardiac monitoring. These practical considerations, though beyond efficacy alone, are particularly relevant in acute and resource-limited settings. On the other hand, phenytoin and fosphenytoin have varied safety profiles. Local reactions like venous thromboembolism are significantly lower on the use of fosphenytoin [38]. This holds greater significance in EDs, wherein drugs are administered through intravenous routes and patient discomfort and pain may significantly lead to poorer outcomes [39]. However, fosphenytoin may not be available or commonly used across all regions. Low-middle income countries have limited availability of anti-seizure medications, and fosphenytoin is not commonly used [40]. Thus, a comprehensive evaluation is needed on the cost vs. benefits offered by fosphenytoin over phenytoin.
This study offered a comprehensive review of levetiracetam over phenytoin, fosphenytoin and valproate, focusing primarily on children admitted to emergency departments. The study aimed to quantify outcomes in serious children needing immediate attention following SE. A rigorous methodology was followed, and subgroup analyses based on study designs and comparator drugs in order to generate stronger evidence. However, the study is not without its limitations; all observational studies and trials were included, which may have resulted in biasing of the pooled results. Age of the children was not considered during the analysis, resulting in lack of age-specific recommendations of use of drugs. This was due to the wide variability across studies in inclusion of their patient population, with many studies considering all age groups from infants till 18 years of age. Heterogeneity across outcomes such as seizure termination rates may have been influenced by the wide age-group of inclusion, different management protocols across studies, and different causes of epilepsy in these patients.
Newer drugs like lacosamide are being increasingly used in the settings of emergency departments in children, and may offer a better safety profile [41]. Efficacy of lacosamide from preliminary case studies have shown rapid seizure termination rates, and are well tolerated in children [42]. There is a need to comparatively evaluate the efficacy of lacosamide vis-à-vis other routinely prescribed medications like levetiracetam or phenytoin [43].
Future trials need to be conducted, adopting a double-blinded randomization, ensuring results of highest quality. A further look is needed into age-specific drug effects among pediatrics, and the nullification of etiological factors in influencing seizure recurrence and cessation. Additionally, guidelines of management and admission of children vary significantly across regions. The difference in first-approach, admission and care administered may widely influence outcomes. A thorough region-based analysis is needed to evaluate each of these guidelines in order to maximize patient outcomes.
This study suggests that levetiracetam is comparable in efficacy to other second-line antiepileptic agents for pediatric convulsive status epilepticus, with our analysis noting faster seizure cessation and stabilization on use of levetiracetam. However, heterogeneity in study designs and populations limits the certainty of these findings. There is a need for more modern clinical trials following a rigorous methodology to generate more conclusive evidence. Future research should prioritize well-designed, age-stratified randomized controlled trials, including newer agents such as brivaracetam and lacosamide. Expanding the evidence base to encompass neonates and other underrepresented pediatric subgroups will be critical for refining individualized, evidence-based treatment pathways.
Supplementary Material 1.