Results of Phase II Pharmacokinetic Study of Levetiracetam for Prevention of Posttraumatic Epilepsy

Abstract

Levetiracetam (LEV) has anti-epileptogenic effects in animals and is a candidate for prevention of epilepsy after traumatic brain injury. Pharmacokinetics of LEV in TBI patients are unknown. We report pharmacokinetics of TBI subjects ≥6 years with high PTE risk treated with LEV 55 mg/kg/day orally, nasogastrically or intravenously for 30 days starting ≤8 hours after injury in a phase 2 safety and pharmacokinetic study. 41 subjects (26 adults, 15 children) were randomized to PK studies on treatment days 3 and 30. 36/41 randomized subjects underwent PK study on treatment day 3, and 24/41 subjects on day 30. On day 3, mean T_max was 2.2 hours, C_max 60.2 μg/ml and AUC 403.7 h*/μg/ml. T_max was longer in the elderly than in children and non-elderly adults (5.96 h vs. 1.5 h and 1.8 h, p=0.0001). AUC was non-significantly lower in children compared with adults and the elderly (317.4 h*/μg/ml vs. 461.4 and 450.2, p=0.08). C_max trended higher in i.v.- versus tablet- or n.g.-treated subjects (78.4 μg/ml vs. 59 and 48.2, p=0.07). AUC for n.g. and i.v. administrations were 79% and 88% of oral administration. There were no significant PK differences between days 3 and 30. Treatment of TBI patients with high PTE risk with 55 mg/kg/day LEV, a dose with antiepileptogenic effect in animals, results in plasma LEV levels comparable to those in animal studies.

INTRODUCTION

Head injury is the cause of 5% of all epilepsy in both children and adults, with estimated 20,000 new cases annually [1–2]. The risk of post-traumatic epilepsy (PTE), defined as unprovoked seizures occurring > 7 days after traumatic brain injury (TBI) is 5–7% [1] overall, but is higher, 17–21% among patients with severe TBI, defined as the presence of either intracranial hemorrhage or loss of consciousness or amnesia lasting >24 hours after TBI [1].

Levetiracetam (LEV) has potent anti-epileptogenic effects in animal models in doses similar to those used for treatment of epilepsy [3] and may be useful in preventing the development of epilepsy after TBI. There has been no experience in administering LEV rapidly to individuals with acute TBI. TBI patients, often critically ill, may differ in their susceptibility to adverse events and drug metabolism from otherwise healthy patients with epilepsy. Side effects and pharmacokinetics of LEV in TBI patients are unknown.

We recently reported the safety, tolerability, and feasibility of acute and chronic administration of LEV to adults and children with TBI with a high risk for PTE, using a dose with demonstrated antiepileptogenic effects in animals (55 mg/kg/day [4]). The study was a first step of a project aimed at evaluating the efficacy of LEV in prevention of PTE. Here, we report the study’s pharmacokinetic findings. Pharmacokinetic (PK) studies were obtained on days 3 and 30 of LEV administration to TBI subjects with a high risk for developing PTE. Our goals were to determine whether acute (day 3) and chronic (day 30) treatment results in plasma levels comparable to LEV concentrations in positive animal antiepileptogenic studies; whether LEV pharmacokinetics vary with different routes of administration (tablet, nasogastric liquid, and intravenous) and across ages (children, adults, elderly); and whether salivary concentrations correlated with plasma concentrations.

METHODS

Study Design

This PK study was done as part of a fixed dose, open label, non-randomized phase 2 study with two arms comparing LEV treatment versus observation at two level 1 trauma centers, Children’s National Medical Center and Washington Hospital Center in Washington, D.C. Only the treatment arm is described in this paper; the non-treatment arm has been reported separately [4]. Both institutions’ IRBs approved the study. All subjects/parents or their legal surrogates signed the IRB-approved informed consent form before enrollment. The study was performed in accordance with the ethical standards of the 1964 Declaration of Helsinki.

Subjects

Children and adults aged >6 years were enrolled (age range 6–87). Inclusion criteria were TBI with any intracranial hemorrhage except for isolated subarachnoid hemorrhage, or with penetrating wound injury, depressed skull fracture or early post-traumatic seizure. These criteria identify adult patients with an average 20% risk of PTE [1]. Exclusion criteria were Glasgow Coma Score (GCS) < 6, based on best GCS within 4 hours of injury, to avoid high mortality; elevated serum creatinine (≥1.7 for adults, ≥ 1.0 for age 13–17 and ≥ 0.7 for age 6–12); history of psychosis, prior unprovoked seizure, CVA, TBI or encephalitis within 3 years, AED use within 3 months, unstable medical disease, moderate to severe mental retardation, and pregnancy. Computerized tomographs (CT) scans were used to evaluate injury type.

Treatment

LEV 55 mg/kg/day was administered initially within 8 hours of injury, followed by 2 divided doses at 8 am/8 pm for 30 days. This dose was chosen to approximate the maximally effective antiepileptogenic dose, 54 mg/kg/day, in animal kindling models [3]. LEV was administered as 250, 500 or 750 mg tablets for subjects who could swallow, as liquid solution of 100 mg/ml via nasogastric (n.g.) tube for subjects who could not swallow before the i.v. formulation became available, or intravenously (i.v.) as 5 cc vials of 100 mg LEV/ml purified water infused over 15 minutes (the i.v. formulation became available during the study). Those initially unable to swallow were changed to tablets once they could swallow. Children unable to swallow tablets received the oral liquid formulation. Study medication was provided by UCB Pharma, Inc. Compliance was assessed at each visit by pill count/liquid volume and by plasma trough (am) LEV levels.

In addition, all subjects received PHT for one week after TBI as standard (adult) care for prevention of “early” (<7 days after TBI) post-traumatic seizures [5] starting at 20 mg/kg. There is no standard of care for children, who received fosphenytoin unless PHT was contraindicated. PHT/fosphenytoin was administered initially i.v. at 20 mg/kg within 8 hours of injury.

Evaluations

Pharmacokinetic (PK) profiles were assessed on treatment days 3 and 30 in 26 adults and 15 children randomized to participate in the PK evaluation. For PK evaluation, blood was obtained through an indwelling catheter immediately prior to dosing (0hr) and 0.5, 1, 2, 3, 4, 6, 8, and 12 hours post-dose on days 3 and 30 for orally and n.g.-treated subjects, with additional samples at 20 and 40 minutes in i.v. treated subjects. In addition to the PK studies, blood was drawn from all treated subjects for trough (8–9 a.m.) LEV levels on days 2,3,4,7,14 and 30.

Visits were conducted on study days 3, 7, 14, 30 and 60 (one month post-treatment) for adverse events (AEs), concurrent illnesses and medications, and medication compliance.

Saliva was collected within 3 minutes of all LEV blood drawing. This was done by spitting into a cup or by using a syringe to withdraw saliva from the oropharynx.

Determination of LEV levels

LEV concentrations in plasma and saliva were measured at the CNMC Bioanalytical Core Laboratory, using a high performance liquid chromatography-electrospray tandem mass spectrometry (HPLC-ESI-MS/MS) assay developed by one of the investigators (SJS) [6]. An API-3000 or API-4000 triple quadruple mass spectrometer (Sciex, Concord, Canada) coupled with the IonSpray source and Shimadzu HPLC system (Shimadzu Scientific Instruments, Columbia, MD) was used employing ritonavir as an internal standard (IS) for LEV. One hundred microliters of plasma or saliva was deproteinized by adding 150 μl of acetonitrile containing internal standard. After centrifugation, 100 μl of supernatant was diluted with 300 μl of water and 10 μl aliquot was injected onto a C-18 column. After a 2.5 min wash the valve was activated to initiate the isocratic elution program which eluted the LEV and internal standard into the MS/MS system. Quantitation by multiple reaction monitoring (MRM) analysis was performed in the positive ion mode. Reliability and accuracy of this method were assessed by comparison with external quality control samples (ChromSystems), between laboratory comparisons and by recovery studies. Within-day coefficients of variation (CVs) were <6.1% and between-day CVs were <8.2%. The average spiked recoveries of LEV added to the drug-free human plasma samples were 108% at low concentration level and 103% at high concentration level. The method correlated well with the commonly used Quest/Chantilly tandem mass spectrometric method (r=0.983).

Statistical methods

Pharmacokinetic data were analyzed using WinNonLin. The parameters that were estimated for each participant from these studies included the area under the 12-hr concentration curve (AUC), the maximum concentration of drug (C_max), and the time to maximum concentration (T_max). Half-life was not assessed because the concentration curves did not return to their origin before the next dose of drug needed to be administered. Plasma and salivary LEV concentrations were summarized by age (6–17, 18–65, > 65), administration route (p.o, n.g., i.v.) and timing of administration (acute/day 3 and chronic/day 30) using descriptive statistics. AUC, C_max, and T_max estimates were compared by ANOVA across groups with treatment (acute/chronic), subject age and route of administration as factors. The 95% CI of the geometric mean ratio was computed. All statistical tests were 2-tailed, with significance set at p<0.05.

RESULTS

Demographics, enrollment and disposition of subjects are shown in Table 1 and Table e-1. Eight/66 subjects discontinued LEV early (15% adults, 5% children), two (3%, both adults) for LEV-related toxicity (chiefly somnolence), and 6 (9%) for medication-unrelated reasons, including 3 elderly (≥ 65 years).

Table 1.

Subject demographics and injury severity by GCS. All % figures are rounded to the whole number. GCS is best GCS ≤ 4 hours after TBI.

Characteristic	n	%
Total	66	100.0
Male	50	75.8
Race
White	29	43.9
Black	33	50.0
Other	4	6.1
Ethnicity
Hispanic	10	15.2
Other	56	84.8
Age
6 –12	15	22.7
13–17	5	7.6
18–39	24	36.4
40–59	10	15.2
60–69	7	10.6
70+	5	7.6
Glasgow Coma Score
6–8	18	27.3
9–11	6	9.1
12–15	42	63.6

Pharmacokinetic studies (Table 2) were obtained on 36/41 subjects randomized for PK evaluation on treatment day 3 (22/26 randomized adults, 14/15 randomized children), and on 24/41 subjects on treatment day 30 (15/26 adults, 9/15 children). Reasons for non-participation in PK study on day 3 are shown in Table e-2.

Table 2.

Levetiracetam plasma pharmacokinetic parameters on treatment days 3 and 30, all subjects combined and grouped by age and route of administration. Figures are rounded to one decimal point.

Treatment day		All1	Age groups1						Administration routes1
			6–17,all	6–11	12–17	18–64	>65	P2	p.o.	n.g.	i.v.	P2
		N=36	N=14	N=7	N=7	N=17	N=5		N=23	N=7	N=6
Day 3	Tmax (hr)	2.2	1.7	1.97	1.37	1.5	6	0.0001	2.7	1.5	1.1
	Cmax (μg/mL)	60.2	57.6	58.1	57.2	63.4	56.4		59	48.5	78.4	0.07
	AUC (hr*μg/mL)	403.7	317.4	341.2	293.6	461.1	450.2	0.08	429.7	341	377.3
		N=24	N=9	N=7	N=2	N=12	N=3		N=20	N=2	N=2
Day 30	Tmax (hr)	1.8	1.9	1.94	1.8	1.2	4.2	0.01	2	0.9	0.8
	Cmax (μg/mL)	57.2	49.2	49.8	47.2	59.6	71.5		58.4	41.6	60.6
	AUC (hr*μg/mL)	431.6	352.6	349.1	364.7	448.3	602.1		453.4	235.6	409.9

¹ANCOVA analysis for PK estimates

²p= >0.1 where no p value is shown

C _max: maximal plasma concentration

T_max: time to reach maximal plasma concentration

AUC: area under the plasma-concentration time curve

Overall PK

On day 3, mean T_max was 2.2 hours (1.7 h children, 1.5 h adults ≤ 65 years [=“adults”], 5.96 h adults > 65 years [=“elderly’]). Mean C _max was 60.2 μg/ml (57.6 children, 63.4 adults, 56.4 elderly) and AUC 403.7 h*/μg/ml (317.4 children, 461.1 adults, and 450.2 elderly). When normalized for a 1-mg/kg dose, C_{max norm} was 1.1 μ/ml overall (1.1 μ/ml children, 1.2 μ/ml adults and 1 μ/ml elderly), and AUC _norm was 7.3 h*/μg/ml overall (5.8 h*/μg/ml children, 8.4 h*/μg/ml adults and 8.2 h*/μg/ml elderly).

There was no statistically significant difference in any PK parameters between acute (day 3) and chronic (day 30) LEV administration. Age- and route of administration comparisons on day 30 were limited by the small number of elderly (n=3) and n.g.- and i.v.-treated PK subjects (n= 2 each).

Age

On day 3, the elderly had longer T_max than children and non-elderly adults (5.96 h vs. 1.5 h and 1.7 h, respectively, p=0.0001). AUC was non-significantly lower in children compared with adults and the elderly: 317.4 h*/μg/ml vs. 461.4 and 450.2, respectively (p=0.08). C_max did not differ significantly across ages.

Administration route

On day 3, C_max trended to be higher in i.v.- versus tablet- or n.g.-treated subjects: 78.4 μg/ml vs. 59 and 48.5, respectively (p=0.07). AUC for n.g. and i.v. administrations were 79% and 88% of oral administration AUC (341 h*/μg/ml n.g. vs. 429.7 p.o. and 377.3 i.v., NS).

Salivary PK

Collection of good quality saliva was technically difficult in critically ill, intubated subjects. Frequently there was no saliva, or it was thick, tenaceous and coalesced into mucous plugs which could not be analyzed. Thus, salivary analysis was stopped when 2/3 of subjects completed treatment. Of 44 subjects treated up to that point, 27 subjects were randomized to PK evaluation. 24/27 completed day 3 plasma PK analyses, but only 8/27 completed salivary analyses (Table e-3).

Comparison of saliva and plasma PK (Table 3)

Table 3.

Levetiracetam pharmacokinetic plasma and salivary parameters on treatment days 3 and 30, adults and children combined. Figures are rounded to one decimal point.

	Day 3			Day 30
	Saliva	Plasma	% sal/plasma	Saliva	Plasma	% sal/plasma
	N=8	N=36		N=14	N=24
Tmax (hr)	2	2.2		1.9	1.8
Cmax (μg/mL)	50.2	60.2	83.4	54.9	57.2	96
AUC (hr*μg/mL)	259.1	403.7	64.2	389.7	431.6	90.3

C _max and AUC values were non-significantly lower in saliva than in plasma: 83.4% and 64.2% of plasma values on day 3, and 96% and 90.3% on day 30. T_max was similar for saliva and plasma on both days 3 and 30. All but one salivary PK analyses were done on orally treated subjects. Differences between salivary and plasma PK parameters between children and adults aged 18–64 were statistically non-significant, possibly because of the small number of samples (Table e-4). There was no analyzable saliva from the elderly on day 3, and only one on day 30.

Levetiracetam plasma trough levels (Table 4)

Table 4.

Mean and 95% confidence intervals (C.I.) of LEV plasma trough level (μg/ml) by treatment day number, rounded to the nearest decimal point.

Treatment day #	Mean	C.I.
2	20.2	16.2–24.2
3	24.4	19.1–26.6
4	21	16.3–25.7
7	19.6	15–24.2
14	22	17.5–26.5
30	26.7	21.9–31.5

Mean plasma trough LEV levels on days 2, 3, and 30 were 20.2, 24.4, and 26.7 μg/ml, respectively, with no significant difference between the different days.

DISCUSSION

The main finding of this study is that acute and chronic treatment of TBI patients with a high risk for developing PTE with a LEV dose that is antiepileptogenic in animal studies (55 mg/kg/day) results in plasma LEV levels comparable to or higher than those seen in animal antiepileptogenic studies. These levels are achieved rapidly and maintained chronically with oral, n.g. and i.v. routes of administration.

There have been repeated attempts at prevention of PTE in the past, mainly with AEDs, all unsuccessful. There are reasons for these failures: all AEDs evaluated to-date have had either no antiepileptogenic effect in animals (PHT, CBZ) or had it at doses too high for human use (VPA, PB,CZP) [7–8].

Unlike previously tested AEDs, LEV has antiepileptogenic effect in animals in clinically applicable doses with clinically relevant blood levels [3]. It dose-dependently retards kindling with a true antiepileptogenic (vs. anticonvulsant) effect in electrical limbic, pentylenetetrazole, corneal and audiogenic rat and mouse models in a dose range of 13–54 mg/kg/day [3]. At 80 mg/kg/day, LEV prevents epilepsy in a genetic model of epilepsy, the “spontaneously epileptic rat”. In status epilepticus (SE) models of epileptogenesis, 50–300 mg/kg/day of LEV attenuates the development of spontaneous seizures after self-sustaining SE induced by 30 minute-long stimulation of the perforant pathway, although not after 4 hour-long stimulation of the amygdala. In the kindling and SE models, the same LEV doses block increases in neuronal excitability and synchronization, two key processes of epileptogenesis. Other AEDs, such as PB, VPA, lamotrigine, or clonazepam lack these effects.

The 55 mg/kg/day dose was selected because 54 mg/kg/day was the most effective dose in the kindling studies where LEV antiepileptogenic effect has been most studied [3 ]. LEV AUCs are similar between rats and humans [9]. In the kindling model plasma LEV levels were 10–25 mcg/ml with 54 mg/kg/day [3]. In the pilocarpine-SE model, increase in neuronal excitability and synchronization was blocked with LEV levels of 12.7–55.2 μg/ml [3]. Our mean trough levels exceeded the lower end, and mean C max levels exceeded the upper end of this range. These levels were attained quickly and were sustained: mean day 2 and 30 trough levels were 20.2 μg/ml, and 26.7 μg/ml. Thus, LEV dose similar to that associated with the best antiepileptogenic effect in animal kindling studies results in human plasma LEV levels similar to or higher than those seen in the animal studies. The 55 mg/kg/day dose is higher than the maximum adult recommended dose of 3000 mg/day, but lower than 60 mg/kg/day commonly used in children [10]. Started without titration at full dose it was well tolerated; this is reported separately [4].

LEV pharmacokinetics in this TBI study do not differ significantly, when dose differences are taken into account, from previously published oral and i.v. pharmacokinetics in healthy humans and in epilepsy patients [11–15], with the exception of prolonged T_max in the elderly. Day 3 mean T_max of 2.2 h for all subjects is longer than the 1.3 h seen in healthy humans and epilepsy patients [11], but this is due to the longer T_max, 5.96 h, in the elderly. Day 3 C_max and AUC, dose -normalized for a 1-mg/kg dose,were similar to previously published pediatric and adult C_{max norm} values of 1.2–1.3 μ/ml and 1.4 μ/ml [11–13]. AUC _norm was slightly lower than previously published pediatric (7.2–12.4 h*/μg/ml) and adult (11.5 h*/μg/ml) values with tablet formulation [11–13]. The i.v.-administered C _{max norm} and AUC _norm are comparable to those seen in healthy adult (18–55 yrs) subjects after a single and multiple 1500 mg 15 minute infusions [14].

There was no statistically significant difference in any PK parameters between acute (day 3) and chronic (day 30 PK) treatment, indicating stability of metabolism of LEV over time.

Comparisons of PK of different routes of administration and across age groups was limited to day 3 because by day 30 most subjects were treated orally and the number of elderly subjects evaluated was small (n=3). There were no statistically significant differences in the PK parameters between oral, i.v. and n.g. administrations. However, AUC and C _max both tended to be smaller with n.g. than with oral or i.v. administrations. AUC and C_max of n.g. administration were 79% of 82% of oral administration. The data suggest that i.v. or oral LEV administration may be preferable to n.g. administration.

The longer T_max in the elderly, compared to both non-elderly adults and children in this study and to T_max in healthy and epilepsy elderly subjects [11], could be due to the fact that our subjects were acutely/criticially ill. The longer T_max in the elderly could be important when treating acutely ill elderly subjects, e.g. for stroke-related seizures. Children showed a statistically non-significant trend to have smaller AUC, in line with previous studies [12–13]. Unlike previous studies, we did not find a significant difference in PK parameters between children < 12 years and > 12 years old.

We sought to determine whether there was correlation between LEV plasma and salivary concentrations to allow substitution of saliva for plasma LEV concentrations in the planned phase 3 PTE prevention study. Saliva collection in the critically ill patients was difficult, with a high failure rate. This makes use of salivary LEV level in critically ill patients impractical.

CONCLUSION

In summary, this study shows that acute and chronic treatment of TBI patients with a high risk for developing PTE with a LEV dose that is antiepileptogenic in animal studies (55 mg/kg/day) results in plasma LEV levels comparable to or higher than those seen in effective animal antiepileptogenic studies. These levels are achieved rapidly and maintained chronically with oral, n.g. and i.v. routes of administration in both adults and children.

HIGHLIGHTS.

• LEV PK was evaluated in a phase 2 study in TBI patients with high risk of PTE

• LEV 55 mg/kg/d was given ≤8 h after TBI for 30 days, with PK done on days 3/30

• PK was similar to PK in chronic epilepsy patients, across different administration routes and across ages

• Plasma levels were similar to those seen in positive animal antiepileptogenic studies

• The study lays the groundwork for LEV evaluation in a phase 3 study of PTE prevention