Levetiracetam for the prevention of busulfan-induced seizures in pediatric hematopoietic cell transplantation recipients

Introduction

High-dose busulfan is frequently used for hematopoietic cell transplantation (HCT) conditioning regimens. In this setting, the busulfan dose ranges from 2 to 4 mg/kg/day for 1–4 days for a total dose of 9 to 16 mg/kg. ¹ High-dose busulfan is associated with an increased risk of seizures, typically generalized tonic-clonic seizures.^2,3 Neurotoxicity is likely a result of the drug’s ability to freely cross the blood–brain barrier and achieve high concentrations in the central nervous system. ³ Without prophylaxis, the incidence of seizures is ∼10% (range 1.8–40%).^2,3 Due to this risk, patients standardly receive an antiepileptic drug (AED) during busulfan conditioning to prevent seizures. The prophylactic AED should ideally meet criteria to ensure it does not negatively impact clinical HCT outcomes. These specific criteria are that the prophylactic AED: (1) has a short onset of action allowing for a short course of therapy; (2) be well tolerated; (3) does not add to the toxicity of the conditioning regimen and thus has minimal hematologic, hepatic and dermatologic toxicity; and (4) does not alter the pharmacokinetics of conditioning agents. ³ When busulfan-based conditioning regimens were being developed in the 1980s, phenytoin best fit these criteria in the context of AEDs available at that time. Thus, phenytoin has historically been the AED used most often clinically. Phenytoin reduces the prevalence of busulfan-induced seizures to 0–5.5%. ² However, phenytoin has unfavorable pharmacokinetics, and it has high potential for drug–drug interactions (DDI). Given these limitations, phenytoin would ideally be replaced with a different AED to prevent busulfan-induced seizures.

Many novel AEDs have been approved by the Food and Drug Administration (FDA) over the past 20 years. These second generation AEDs, such as lamotrigine, topiramate, zonisamide and levetiracetam, are now more commonly prescribed than phenytoin for first-line treatment of generalized-tonic-clonic seizures.^4,5 Thus, in the field of neurology, the standard of care for preventing generalized tonic-clonic seizures has moved away from phenytoin and toward the second generation AEDs. Of these AEDs, levetiracetam best fits the criteria for prophylactic AED in HCT patients receiving busulfan.

Levetiracetam is FDA approved for the adjunctive treatment of myoclonic, partial-onset, and primary generalized tonic-clonic seizures. Levetiracetam is becoming the standard of care for both seizure prophylaxis and treatment in the settings of traumatic brain injury, brain tumors and neurosurgery.^6–8 It has favorable pharmacokinetic characteristics, such as 100% bioavailability, a predictable time to maximum plasma concentration (Tmax), short half-life (∼7 h), and low potential for DDI. ³ It is well tolerated, with a lower incidence of rash compared to phenytoin. ⁹ The most common toxicity of levetiracetam is somnolence. A comparison of important characteristics of phenytoin and levetiracetam is provided in Table 1. A few single-centers have reported successful use of levetiracetam for the prevention of busulfan-induced seizures.^10,11 We are reporting our institutional experience with pediatric HCT recipients.

OPEN IN VIEWER

Table 1.

Pharmacologic comparison of phenytoin and levetiracetam as prophylaxis for busulfan-induced seizures.

	Phenytoin	Levetiracetam
Pediatric dosing12,13	Loading dose: 10–15 mg/kg PO divided over three doses Maintenance dose: 4–8 mg/kg/day PO divided every 12 h (≤ 40 kg) OR 300 mg PO QPM (≥40 kg)	10 mg/kg/dose PO/IV BID
PK12,13	Oral bioavailability: 20–100% Tmax: 1.5–12 h Clearance: hepatic, Michaelis-Menten pharmacokinetics Highly protein bound Half-life: 7–42 h	Oral bioavailability: 100% Tmax: 1 h Clearance: renal Half-life: 6–8 h
Dose personalization	Pharmacokinetic-guided dosing (i.e. therapeutic drug monitoring) 14 Pharmacogenetic-guided dosing with CYP2C9 genotype 15	None
DDI potential 16	High DDI risk Induces: CYP1A2, CYP2B6, CYP2C9, CYP2C19, CYP2C8, CYP3A4, P-glycoprotein, UGTs Substrate: CYP2C9, CYP2C19	No – low DDI risk
Reported DDI in HCT	HCT conditioning: phenytoin administration lead to a marked reduction in AUC of cyclophosphamide and an increased AUC of two cyclophosphamide metabolites (4-hydroxycyclophosphamide and carboxyethylphosphoramide mustard) 17	None
HCT-related adverse effects	Sedation/incoordination (dose related) 12 Rash (5–10%) 18 Serious dermatologic reaction such as Stevens-Johnson syndrome or toxic epidermal necrolysis (case reports) 12 Hypersensitivity syndrome (0.01–0.1%) 19 Myelosuppression (0.004%) 20 Hepatotoxicity (case reports) 12	Sedation/incoordination (8–15%) 13 Rash (not reported) 13 Serious dermatologic reaction such as Stevens-Johnson syndrome or toxic epidermal necrolysis (case reports) 13 Hypersensitivity syndrome (case reports) 13 Myelosuppression (case reports) 13 Hepatotoxicity (case reports) 13

Methods

We conducted a study to determine the efficacy and safety of levetiracetam for the prevention of busulfan-induced seizures. This was an IRB-approved, single-center, retrospective analysis of pediatric HCT patients receiving intravenous (IV) busulfan for HCT conditioning from January 2009 to July 2014. In 2013, institutional pediatric HCT protocols changed the preferred AED for busulfan-induced seizure prophylaxis from phenytoin to levetiracetam, thus allowing for retrospective comparison of the two agents at our transplant center.

The inclusion criteria were that subjects must be less than 21 years of age and had received IV busulfan as a part of their HCT conditioning regimen. Exclusion criteria included history of prior seizures, AED therapy prior to HCT, and use of more than one AED for busulfan-induced seizure prophylaxis during conditioning. The primary study endpoint was the incidence of seizure during busulfan administration (day −8 to 0). Key transplant related-outcomes were also collected, including the incidence of graft rejection, sinusoidal obstruction syndrome (SOS), relapse, and death. The modified Seattle criteria were used to identify cases of SOS and grade severity. ²¹ Baseline liver function tests, including aspartate aminotransferase (AST), alanine aminotransferase (ALT) and bilirubin, were collected and compared to peak concentrations post-graft infusion (day 0 to +20). Cases were classified as liver disease of unknown etiology if liver toxicity was likely attributable to other etiologies (graft-versus-host disease (GVHD), drug-induced, etc.). Other safety endpoints included the incidence of rash and somnolence documented during AED administration (day −8 to 0).

Results

A total of 22 patients were screened and of these, 20 patients met the criteria for inclusion in the study. One patient was excluded due to lamotrigine use prior to HCT and another for administration of both phenytoin and levetiracetam for seizure prophylaxis. The median patient age was 4.3 years (range 0.5–20.0) and each received one of seven HCT conditioning regimens containing IV busulfan. Patient diagnoses included both malignant (N = 8) and non-malignant (N = 12) conditions, with the most common being acute myeloid leukemia (N = 4) and immunodeficiency disorders (N = 4), respectively. The population was heterogeneous with regard to the indication for HCT, donor type, stem cell source, and conditioning regimen (Table 2). Nine patients (45%) received levetiracetam and 11 (55%) received phenytoin for seizure prophylaxis. Phenytoin was administered orally as a loading dose divided over three doses followed by a maintenance dose given every 12 h (≤40 kg) or once daily at night (≥40 kg) (median loading dose 11.8 mg/kg, range 8.9–15.3 mg/kg; median maintenance dose 5.6 mg/kg/day, range 4.1–6.8 mg/kg/day). Levetiracetam had no loading dose; it was administered either intravenously or orally every 12 h (median dose 20.1 mg/kg/day, range 17.1–21.2 mg/kg/day).

OPEN IN VIEWER

Table 2.

Patient characteristics, as N (%) or median (range).

	N = 20
Male sex	10 (50%)
Age (years)	4.3 (0.5–20.0)
Weight (kg)	16.8 (8.0–55.9)
Diagnosis
Malignant disorders a	8 (40%)
Non-malignant disorders b	12 (60%)
Donor type
Matched unrelated	10 (50%)
Mismatched unrelated	6 (30%)
Matched related	3 (15%)
Autologous	1 (5%)
Stem cell source
Peripheral blood	9 (45%)
Marrow	7 (35%)
Cord blood	4 (20%)
Conditioning regimen
Targeted Bu/CY ± ATG (equine)	9 (45%)
Targeted Bu, clofarabine	4 (20%)
Targeted Bu/CY, melphalan	2 (10%)
Targeted Bu/CY, fludarabine, ATG (rabbit)	2 (10%)
Other c	3 (15%)
GVHD prophylaxis
Tacrolimus OR cyclosporine + Methotrexate	13 (65%)
Cyclosporine + mycophenolate mofetil	4 (20%)
Cyclosporine	2 (10%)
None	1 (5%)

No seizures or graft rejections were observed in the study population. The rates of relapse or death following HCT were low and similar between groups. One relapse and two deaths occurred in the levetiracetam group, while no relapses and one death occurred in the phenytoin group. The incidence of SOS was also similar between groups, with one case in the levetiracetam group and two cases in the phenytoin group (Table 3). For this study, a patient was categorized as having liver function test abnormalities that were grade 3 or worse in the NCI – Common Toxicity Criteria Version 4. ²² These definitions are: bilirubin >3× upper limit of normal (ULN), ALT > 5 × ULN, or AST > 5 × ULN. Post-graft infusion (day 0 to +20), liver function abnormalities occurred at similar rates in those receiving phenytoin (55%, 6 of 11 patients) compared to levetiracetam (56%, 5 of 9 patients). The incidence of rash was slightly higher in the phenytoin group (64%) compared to the levetiracetam group (33%) and the incidence of somnolence was low overall, with only one patient in the phenytoin group experiencing notable toxicity.

OPEN IN VIEWER

Table 3.

Incidence of liver toxicity, N (%).

	Levetiracetam	Phenytoin/ fosphenytoin
N	9	11
No liver disease	6 (67%)	5 (45%)
Liver disease of unknown etiology	2 (22%)	4 (36%)
SOS
Mild	0	1 (9%)
Moderate	1 (11%)	1 (9%)
Severe	0	0

Discussion

Our data suggest similar effectiveness and safety between levetiracetam and phenytoin for the prevention of busulfan-induced seizures in this pediatric HCT population. Many HCT centers have already transitioned to levetiracetam as their standard AED prophylaxis because of its favorable pharmacokinetics and lower potential for DDI compared to phenytoin. Our data provide further support for the use of levetiracetam, as there was no difference in HCT outcomes compared to phenytoin. Further case series are needed to confirm the safety and efficacy of levetiracetam for the prevention of busulfan-induced seizures.

Although phenytoin has been used most commonly for the prevention of busulfan-induced seizures, there are several important limitations to consider. Phenytoin has unfavorable pharmacokinetics, including variable bioavailability, variable time to maximum plasma concentration (Tmax), and dose-dependent capacity limited (Michaelis-Menten) hepatic metabolism. Most importantly, phenytoin has high potential for DDI. Conversely, levetiracetam has favorable pharmacokinetics and low potential for DDI, making it an ideal alternative to phenytoin. Levetiracetam is also preferred due to its ease of administration: (1) equivalent dosing between formulations and (2) simple twice-daily dosing (no loading dose required). Both AEDs are available generically, making their costs comparable. Given the short duration (five to six days) of busulfan seizure prophylaxis and the high cost of HCT, the potential financial impact of the AED is minimal. Another limitation of levetiracetam is less evidence for its use compared to phenytoin; case series such as these provide data to support its use for the prevention of busulfan-induced seizures.

The DDI potential of busulfan and the AED used to prevent busulfan-induced seizures cannot be separated since they are concomitantly administered. Busulfan itself is not known to cause DDI. Phenytoin is a well-known potent inducer of multiple cytochrome P450 (CYP) enzymes. Phenytoin induces oral busulfan clearance, which reflects the effects of phenytoin upon gastrointestinal and hepatic enzymes. However, the effect of phenytoin upon IV busulfan pharmacokinetics is not clear. The package insert states that phenytoin increases IV busulfan clearance by 15% or more possibly due to induction of glutathione-S-transferase. ²³ However, phenytoin administration has had either a slight effect ²⁴ to no effect^25–27 upon IV busulfan clearance.

Importantly, busulfan is not administered alone so the impact of busulfan and its seizure prophylaxis upon other HCT medications must be considered. Busulfan is often administered with cyclophosphamide (CY), fludarabine or melphalan for HCT conditioning. Busulfan without phenytoin has been shown to inhibit CY metabolism. ²⁸ Phenytoin affects CY pharmacokinetics in both the targeted busulfan (BU)/CY regimen and in the setting of post-transplant CY. In the setting of the targeted BU/CY regimen, phenytoin putatively led to a marked reduction in the area under the curve (AUC) of CY and an increased AUC of two of its metabolites 4-hydroxycyclophosphamide (HCY) and carboxyethylphosphoramide mustard (CEPM).^17,29,30 Unfortunately, there was minimal correlation between the AUC of CY and its metabolites HCY and CEPM, ²⁹ such that the AUC of CY cannot be used to predict the impact of phenytoin on levels of HCY (the word should be HCY) and CEPM. The mechanism for the DDI between phenytoin and CY is complex and comprehensively explained in our previous publication. ¹⁷ With post-transplant CY, phenytoin putatively led to a marked reduction in the AUC of CY and a decreased AUC of the metabolite 4-keto-cyclophosphamide. The AUCs of other metabolites, CEPM and deschloroethyl-CY, did not show statistically significant differences. ³¹ The impact of the DDI between phenytoin and cyclophosphamide upon clinical outcomes has not been examined.

Fludarabine and melphalan are also commonly administered with busulfan for HCT conditioning. Severe neurotoxicity has been reported in up to 36% of patients receiving high-dose fludarabine (96 mg/m²/day for 5 to 7 days). ³² This syndrome is characterized by delayed progressive encephalopathy with seizures, blindness, paralysis, and coma. Fludarabine-associated neurotoxicity is thought to be dose-dependent and is rarely seen (<0.2%) with lower doses, ³² such as those used in the HCT setting. Fludarabine is rapidly dephosphorylated to 2-flouro-ara-A and then phosphorylated intracellularly by deoxycytidine kinase to the active triphosphate, 2-flouro-ara-ATP. Melphalan is eliminated primarily via hydrolysis in the plasma. Neither agent relies on CYP enzyme hepatic metabolism, and thus there is minimal potential for DDI between phenytoin and these two conditioning agents.

In summary, AED prophylaxis is warranted for the prevention of busulfan-induced seizures. The ideal AED would be well tolerated and have minimal potential for DDI with other agents used in HCT conditioning regimens. Our data, along with published reports from other transplant centers, show similar effectiveness and safety between levetiracetam and phenytoin. Given its favorable pharmacokinetic profile and decreased risk for DDI, levetiracetam is preferred over phenytoin for the prevention of busulfan-induced seizures in children.