Lacosamide Boluses Decreased Seizure Burden and were Well Tolerated in Neonates with Acute Seizures: A Single-Center Retrospective Case Series

Abstract

Introduction:

Neonatal seizures are associated with worsened neurodevelopmental outcomes. Phenobarbital, the only FDA approved treatment for neonatal seizures, can cause neuronal apoptosis and may worsen neurodevelopmental outcomes. Lacosamide (LCM) may be an efficacious treatment for neonatal seizures.

Methods:

We assessed the impact of LCM boluses on seizure burden in a retrospective cohort of 15 neonates monitored with video electroencephalography (EEG). Medication bolus times and seizure start/end times on EEG tracings determined change in seizure burden.

Results:

Seven patients received LCM as first- or second-line treatment and eight as third-line or later. Average 4-hour seizure burden decreased from 13% to 3% following LCM boluses (P=0.002). Reduction in seizure burden greater than 30% followed 79% of boluses. LCM was well tolerated; one patient experienced mild asymptomatic episodic bradycardia that medication taper resolved.

Conclusions:

LCM significantly decreased seizure burden in this cohort. Prospective studies of LCM treatment for neonatal seizures are warranted.

INTRODUCTION

Neonatal seizures are the most common neurologic emergency encountered in neonates. A recent review reports that the estimated incidence of neonatal seizures is between 1 to 5 per 1000 live births in population-based studies, with higher incidence among premature infants.¹ The majority of neonatal seizures are provoked by an acute insult, with hypoxic-ischemic encephalopathy (HIE, 38%); ischemic stroke (18%); and intracranial hemorrhage (11%) being the most common causes.² High seizure burden and status epilepticus are associated with negative neurodevelopmental outcomes; therefore, it is essential to promptly recognize and treat neonatal seizures.^2–7

Currently, phenobarbital is the only US Food and Drug Administration (FDA)-approved treatment for neonatal seizures.⁸ Although phenobarbital has demonstrated efficacy in decreasing seizure burden,^8–11 it has also been shown to cause neurodegeneration and neuronal apoptosis in neonatal rodent and primate animal models,^12–14 as well as lower full scale IQ in children treated during the first 3 years of life.¹⁵ These findings raise concern that phenobarbital itself could contribute to worsened neurodevelopmental outcomes with prolonged use, and highlight the need for alternative, efficacious first-line antiseizure medications (ASMs) that are not associated with potential comorbidities in this population.

Lacosamide (LCM) may be such a medication. LCM is a third-generation ASM that acts by enhancing slow inactivation of voltage-gated sodium channels;^16,17 recent studies suggest that it is well tolerated in neonates and may be efficacious for decreasing seizure burden.^18,19 A study in neonatal mice found that LCM reduced neocortical seizure-like activity without an acute increase in apoptosis, both in vitro and in vivo, raising the possibility that it may have a superior risk/benefit ratio to phenobarbital.¹⁹ Another recent study also found that lacosamide was not toxic to the developing mouse brain.²⁰ A clinical trial in progress is studying the efficacy, safety, and pharmacokinetics of LCM for neonatal seizures.²¹

While a growing body of literature suggests LCM may be a safe and efficacious treatment for neonatal seizures, reports to this point have not described use of LCM as a first or second-line treatment.²² Additionally, previous studies have not evaluated response to LCM with EEG, a major limitation since most neonatal seizures have no clinical change and are only detected by EEG.²³ In this study, we assess the tolerability and efficacy of LCM for reducing seizure burden on video EEG, and report the efficacy of LCM used as a first- or second-line treatment.

MATERIALS AND METHODS

Study population

We conducted a retrospective chart review of all neonates admitted to a level IV neonatal intensive care unit at a single institution from January 1, 2021 to August 4, 2022 who had neonatal seizures that were treated with LCM. This study was approved by the University of Wisconsin-Madison Institutional Review Board. We queried the medical records for all patients with a diagnosis of neonatal seizures treated with LCM. We included all neonates who were treated with LCM prior to 44 weeks postconceptual age while being recorded on video EEG. We excluded neonates if raw EEG data were not available for review in its entirety, or if all lacosamide doses were less than 2mg/kg. Neonatal birth history, hospital course, ASMs treatment course (all ASM doses, dates, and times), and adverse events were abstracted from the medical record. Maintenance doses of ASMs were not included in the analysis.

EEG review

All video EEG tracings were reviewed by a board-certified pediatric epileptologist who marked start and end times of all seizures (AK, DH, MC). Original video EEG reports (also read by board-certified pediatric epileptologists) were then reviewed; if there was disagreement between the study reviewer and the original report, the study was reviewed by a second board-certified pediatric epileptologist, and majority consensus was used to determine seizure start and end times.

A Python²⁴ script was used to calculate the seizure burden (percentage of time seizing over the next 1-hour window) at all time points for the duration of each EEG study. The primary outcome was absolute change in seizure burden in four-hour windows before and after LCM boluses. The pre-bolus window is defined to end at initiation of LCM bolus; the post-bolus window begins one hour after initiation of LCM bolus, as shown in Figure 1. Analysis was carried out for (1) all boluses of LCM, (2) all initial boluses of LCM, (3) boluses in which LCM was the first or second ASM given, (4) boluses in which LCM was the third or greater ASM, and (5) boluses in which the patient did not receive phenobarbital or midazolam in the four hours preceding the LCM bolus.

Figure 1:

Timeline demonstrating the pre- and post-bolus windows used in seizure burden calculations in relation to LCM bolus initiation.

Statistical analysis

Statistical analysis was carried out in Python using the SciPy statistics module.²⁵ Two-sided Wilcoxon signed-rank sum test was used to identify significant change in pre- and post-LCM bolus. Fisher’s Exact Test or two-sided Mann-Whitney U test was used to compare categorical and continuous demographic data respectively. Significance was defined as a p-value less than 0.05. Multivariate analysis was not attempted due to small sample size.

RESULTS

Patients

Fifteen neonates met inclusion criteria for this study (Figure 2). Seven (47%) received LCM as a first- or second-line treatment and eight (53%) received it as a third-line or greater treatment. Demographics of the sample population are summarized in Table 1. Ten neonates (67%) were born after 37 weeks gestation, four (27%) were born between 33 and 37 weeks gestation, and one (7%) was born at 24 weeks and two days gestation. Eleven neonates (73%) had acute symptomatic seizures, and at the time of publication two (13%) had been diagnosed with epilepsy, one with ongoing seizures in the neonatal period and one with emergence of hemi-hypsarrhythmia at 6 months of age. Seven patients (47%) had seizures due to HIE (five severe, one moderate, and one of undetermined severity). Three (20%) neonates had seizures due to central nervous system infection and two (13%) developed seizures following an endoscopic third ventriculostomy with choroid plexus cauterization. Etiologies in single neonates included myelomeningocele repair, brain malformation, self-limited neonatal epilepsy, incontinentia pigmenti, and unknown. Five neonates (33%) received an ASM prior to admission to the study institution; none received LCM prior to transfer.

Figure 2:

Flow diagram showing the breakdown of patients included in the study and the number who received LCM as a first- or second-line versus third-line or greater treatment for seizures. Two patients were excluded due to lack of complete EEG tracings. One patient was mistakenly treated with a single negligible dose of lacosamide (1mg/kg rather than 10mg/kg), and so was excluded.

Table 1:

Patient demographic information

Patient characteristics	Total patients (N=15)	Patients who received LCM as a first- or second-line ASM (N=7)	Patients who received LCM as a third-line or greater ASM (N=8)	P-value
Median gestational age (IQR)	38+0 (34+0 – 39+1)	38+0 (36+1 – 39+0)	36+4 (31+4 – 39+3)	0.43
Median birth weight in kg (IQR)	2.97 (2.77 – 3.41)	2.92 (2.78 – 3.29)	3.04 (2.77 – 3.44)	0.43
Sex	47% (N=7) female 53% (N=8) male	29% (N=2) female 71% (N=5) male	63% (N=5) female 38% (N=3) male	0.31
Median Apgar score at 1 minute (IQR)	4.0 (1.0 – 8.0)‡	4.0 (2.5 – 5.5)‡	4.5 (1.0 – 9.0)	0.47
Median Apgar score at 5 minutes (IQR)	8.0 (5.0 – 9.0)‡	8.0 (4.75 – 9.0)‡	7.5 (5.0 – 9.0)	0.47
Percent with HIE	47% (N=7)	29% (N=2)	63% (N=5)	0.31
Percent with diffusion changes on brain imaging	60% (N=9)	43% (N=3)	75% (N=6)	0.58
Median age at seizure onset in days (IQR)	3.0 (1.0 – 14.0)	4.0 (1.0 – 12.0)	3.0 (0.75 – 12.5)	0.41
Mean seizure burden in the four hours prior to treatment with LCM (IQR)	15% (4 – 22)	20% (8 – 25)	11% (5 – 20)	0.32
Percent deceased prior to discharge	13% (N=2)	14% (N=1)	13% (N=1)	1.0

Abbreviations: EEG, electroencephalography; HIE, hypoxic-ischemic encephalopathy; IQR, interquartile range; LCM, lacosamide

^‡Excludes one patient for whom this information was not available

Lacosamide efficacy

Seizure times, seizure burden, medication dose, and medication administration times for each neonate are shown in Figure 3. Fifteen neonates received 24 LCM boluses during the study period, with bolus doses ranging from 5 mg/kg – 10 mg/kg. When considering all LCM boluses, there was a significant decrease in average seizure burden from 13% to 3% (P=0.002). No seizures were seen in the four-hour window following 17 LCM boluses (71%), and 19 boluses (79%) reduced seizure burden by more than 30% (Table 2). In order to test for potential confounders, we repeated analysis limited to: (1) first bolus of LCM given to each patient, (2) boluses of LCM for which phenobarbital and midazolam drips were not administered during the pre or post bolus windows, (3) boluses of LCM which were given as a first or second line ASM, and (4) boluses of LCM given as a third-line or greater ASM. The post-bolus decrease in seizure burden remained statistically significant in all subgroup analyses except for that in which LCM was given as a first- or second-line ASM; in this group, mean seizure burden decreased from 12% to 5% but was not statistically significant (P=0.17, Table 2). However, a medication effect still appeared to be present, as five of seven (71%) neonates who received LCM first- or second-line had no seizures in the following four hours.

Figure 3:

Raster plot of seizure times and seizure burden for each neonate. All lacosamide boluses (L) are included. Seizures are denoted as dark vertical lines, seizure burden is color coded as shown in the legend, midazolam drip is denoted by the gray horizontal line at the base of the row for the associated patient, and absence of the thicker horizontal light green bar indicates the end of the video electroencephalographic (EEG) study. Color at a given point indicates seizure burden over the next hour. The medications patients received in a bolus are denoted with the following abbreviations: Lacosamide (L), levetiracetam (V), phenobarbital (P), lorazepam (Z), and fosphenytoin (F). The number following the letter abbreviation represents the bolus dose the patient received at that time in mg/kg. Time 0 is the start of each EEG study.

Table 2:

Changes in seizure burden following LCM boluses (bolus dose range 5 mg/kg - 10 mg/kg)

	Average 4-hour seizure burden		P-value for seizure burden change	Boluses with no seizures in the 4 hours post-bolus (%)	Boluses with 30% or greater reduction in seizure burden (%)
	Pre-bolus	Post-bolus
All LCM boluses (N=24)	13%	3%	0.002	71%	79%
First LCM bolus (N=15)	17%	4%	0.02	73%	80%
LCM bolus with no PHB or MDZ ggt in pre/post windows (N=17)	14%	4%	0.01	77%	82%
LCM boluses given as a first- or second-line ASM (N=11)	12%	5%	0.17	73%	73%
LCM boluses given as a third-line or greater ASM (N=13)	13%	2%	0.004	70%	85%

Abbreviations: LCM, lacosamide; ASM, antiseizure medication; PHB, phenobarbital; MDZ, midazolam; ggt, continuous medication drip

Adverse events

One full term neonate with moderate HIE and neonatal seizures had mild intermittent sinus bradycardia (80–100 BPM) starting 24–28 hours following initiation of therapy with LCM. An electrocardiogram (ECG) at that time showed sinus bradycardia (84 BPM) without other abnormalities; the PR interval was within the normal neonatal range at 120ms. LCM was slowly weaned over the next seven days; episodes of mild bradycardia resolved on the last day of LCM therapy.

DISCUSSION

This retrospective study of neonates with video EEG–confirmed seizures provides the best characterized evidence of efficacy of LCM for treating neonatal seizures to date. We found that LCM boluses significantly decreased four-hour seizure burden from 13% to 3% (p=0.002) with 71% of patient’s seizure free during the four-hour post-bolus window. Despite the small sample size, this effect was still observable when only considering subgroups of (1) the first LCM bolus given to each patient, (2) LCM boluses without concomitant phenobarbital bolus or midazolam continuous infusion in pre and post bolus windows, (3) LCM boluses that followed treatment with at least two other ASMs.

LCM boluses given first- or second-line reduced seizure burden from a mean of 12% to 5% but did not reach statistical significance (P=0.17), primarily due to a smaller number of boluses (11 LCM boluses in 7 neonates) and one outlier (ID 9) whose seizure burden doubled in the post-treatment window and ultimately required a high-dose midazolam infusion to control seizures. Although this metric was not statistically significant, five of seven (71%) neonates who received LCM first- or second-line had no seizures in the following four hours. Thus, LCM shows promise as an efficacious first or second line ASM in neonates.

LCM was well tolerated in our cohort. One patient experienced mild intermittent bradycardia without PR prolongation or other arrhythmias on ECG; LCM was tapered over the course of a week, with resolution of bradycardia and no other complications. LCM does have known risks of cardiac side effects including PR prolongation and atrioventricular block,^26,27 and cardiac side effects have been described in neonates. A case series of 38 neonates found that LCM was well tolerated in 95% of patients but two patients had atrial bigeminy on ECG and telemetry.¹⁸ A case report describes a neonate who developed second-degree AV heart block and suffered cardiac arrest.²⁸ A more recent study of 62 patients across 10 institutions found no adverse events related to LCM.²²

While it is important to consider potential cardiac risks associated with LCM, they must be balanced against the risks of untreated seizures and other ASMs. Phenobarbital is the only US FDA-approved medication to manage neonatal seizures, but has been found to cause neurodegeneration and neuronal apoptosis in neonatal rodent and primate animal models,^12–14 as well as lower cognitive measures among children in the first 3 years of life.¹⁵ While phenobarbital has been shown to be effective in achieving immediate seizure freedom,¹⁰ the long-term net benefit (or harm) of treatment with phenobarbital compared to other medications remains unknown. A small retrospective cohort study comparing the use of phenobarbital and levetiracetam as the initial treatment for neonatal seizures found both medications to have comparable rates of seizure freedom and sustained reduction in seizure burden.²⁹ Neonatal animal models with levetiracetam have not shown this medication to cause neuronal apoptosis,^12,30 and a study of both LCM and levetiracetam found neither to be toxic to the developing mouse brain.²⁰ In contrast to phenobarbital, a neonatal HIE mouse model showed LCM to be efficacious for treatment of neonatal seizures without inducing neuronal apoptosis,¹⁹ and adult studies in humans have found no negative and potentially positive impact on cognition.^31,32 Given these findings and other neonatal seizures showing tolerability and potential efficacy as a third or greater line treatment,^18,22 studies of LCM as an early treatment for neonatal seizures are clearly warranted.

LIMITATIONS

The primary limitation of this study is the small cohort size and retrospective nature of the study. This likely caused some statistical analyses (use of LCM as a first or second-line ASM) to be under-powered. Our observations of ASM efficacy has several potential confounders: (1) doses of maintenance ASMs were not included in analysis, (2) variable timing of LCM relative to other ASMs (LCM is not a part of an institutional protocol for treatment of neonatal seizures), and (3) clinical information not captured in our retrospective chart review that affected provider decision to treat with LCM. Due to the reliance on provider documentation in this retrospective review, we were not able to determine the reasoning behind LCM bolus dosing for individuals in the cohort. Additionally, the decision process for when LCM was used rather than other ASM was not consistently documented. Finally, due to the design of this study, we were not able to compare efficacy of LCM to other ASMs.

CONCLUSIONS

In this small retrospective cohort monitored with video EEG, we found LCM to be an effective treatment for neonatal seizures, both as a late treatment and possibly also as a first- or second-line treatment. LCM boluses led to at least 4 hours of seizure freedom in approximately 70% of patients who were treated, both in the subgroup who received LCM as a first or second-line ASM as well as in the entire cohort. Given this data supporting efficacy, absence of LCM-induced-apoptosis in animal models, and an otherwise good side effect profile, further studies of LCM as an early treatment for neonatal seizures are warranted. This pilot study may serve as good template for design of larger prospective trials with active ASM comparators.