Levetiracetam and valproic acid in glioma: antiseizure and potential antineoplastic effects

ABSTRACT

Seizures are a frequent complication in glioma. Incidence of brain tumor-related epilepsy (BTRE) in high-grade glioma (HGG) is an estimated > 25% and in low-grade glioma (LGG) is approximately 72%. Two first-line antiseizure medications (ASMs) for BTRE include levetiracetam (LEV) and valproic acid (VPA). Use of VPA has decreased because of a broader side effect profile, potential interaction with chemotherapeutic drugs, and availability of newer generation agents. In refractory BTRE, LEV and VPA may be prescribed together to enhance seizure control. VPA and LEV have gained attention for their purported antineoplastic effects and synergistic role with temozolomide. VPA is suggested to modulate anticancer activity in vitro through multiple mechanisms. In addition, retrospective studies indicate increased overall survival in patients with epileptogenic HGGs who are managed with LEV or VPA rather than other ASMs. However, these studies have numerous limitations. It is also reported that patients with glioma and a seizure history have a longer survival. This extended survival, if one exists, may be only observed in certain gliomas with corresponding patient characteristics. We provide a brief overview of the management of BTRE, VPA and LEV as anticonvulsants and antineoplastics, and the factors that may be associated with survival in epileptogenic glioma.

1. Introduction

The clinical course of glioma is often complicated by seizures. Incidence of seizures is inversely related to the World Health Organization (WHO) grade of glioma [1]. While exact numbers are difficult to obtain, patients with high-grade glioma (HGG) have an incidence in the range of 25–60% [2], while low-grade glioma (LGG), mainly with an IDH1 mutation, have an incidence around 72% [3]. Numerous factors are involved in brain tumor-related epilepsy (BTRE), some of which include excitation-inhibition imbalance, structural and metabolic aberrations, genetic mutations, and immunological signals [4]. In addition to negatively [5] impacting quality of life, cortical hyperexcitability associated with seizures has been implicated [6,7] in glioma progression. Given that there may be a link between a hyperexcitable peritumoral environment in glioma and its disease burden, antiseizure medications (ASMs) have been suggested to confer a survival benefit in glioma [8].

Optimal management of brain-tumor related epilepsy (BTRE) is not fully codified. ASM prophylaxis is unwarranted for glioma patients who do not have a history of a seizure [9,10]. Patients with brain tumors are diagnosed with epilepsy in the event of their first seizure as there is a > 60% likelihood of additional seizures, meeting ILAE (International League Against Epilepsy) criteria for the diagnosis [11]. Although there has been limited data to provide high-level guidelines, BTRE is often managed with the same ASMs utilized for newly diagnosed nonlesional focal epilepsies. Current first-line monotherapies for adults with nonlesional focal epilepsy include but are not limited to: levetiracetam (LEV), lamotrigine, lacosamide, zonisamide, carbamazepine, and oxcarbazepine, with lamotrigine as a preferred option for first line when feasible [12–14]. ASMs favored as first-line for BTRE by both the Society for Neuro-Oncology (SNO) [15] and provider members of the European Association of Neuro-Oncology (EANO) [16] are lacosamide, lamotrigine, and LEV. Lamotrigine has the limitation of a slow titration schedule, taking up to 8–12 weeks for therapeutic dosing, limiting its use as a first-line agent in BTRE. Enzyme-inducing ASMs (EIASMs) are not recommended in the management of BTRE due to their modulation of the cytochrome P450 system and their potential impact on drug-drug interactions. Currently, ASM selection is irrespective of histology and localization of tumors [17]. Dosage titrations to maximum tolerated dose levels followed by alternate monotherapies or multidrug regimens are sought to achieve seizure control if patients are unresponsive to initial ASM treatment. Besides avoiding EIASMs, ASM preference depends on patient-specific factors. These include concurrent medications, drug-metabolizing enzyme polymorphisms, comorbidities, effective clearance (i.e., hepatic and kidney function), consideration of potential pregnancy, ease (or absence) of drug level monitoring, insurance coverage, and copays. Surgical resection of gliomas often has a benefit on seizures [18] as does minimally invasive stereotactic ablation [19] of epileptogenic tumors. In a subset of patients, management of BTRE remains challenging because it can be medically intractable [20]. In addition, there are unfavorable drug-drug interactions, and there is still incomplete insight into the pathophysiology of BTRE. Ultimately, the management of BTRE is driven by seizure control, quality of life, and survival [21].

LEV and VPA are two first-line ASMs utilized for BTRE [15–17,22]. At times, LEV-VPA dual therapy is used to manage refractory glioma-related epilepsy (GRE) [23]. There is support from a large multicenter retrospective study that these two agents in combination have the greatest overall efficacy in GRE, and their dual therapy may not be associated with more adverse effects compared to other ASM co-therapies [23]. However, this has not been validated in prospective trials. In addition to managing seizures, there are suggestions that the off-target activities of LEV and VPA may increase overall survival (OS) in glioma [8,24,25]. The majority of this data is based on retrospective studies [24]. The initially promising preliminary observations noted in these investigations have not been conclusively recapitulated in larger and more powerful investigations.

2. Mechanism of action

LEV is a pyrrolidine derivative utilized in the management of focal and generalized seizures [14]. Its main mechanism of action to decrease excitatory drive is blockade of synaptic vesicle protein 2A, which results in a reduction in the release of neurotransmitter vesicles (Figure 1) [26]. VPA is a branched short-chain fatty acid also used in the treatment of focal or generalized seizures in children and adults [14]. There are a few mechanisms by which VPA decreases neuronal excitation within the cortex (Figure 1). VPA increases the concentration of γ-aminobutyric acid (GABA) neurotransmission via inhibition of GABA transaminase [27]. Furthermore, VPA reduces epileptiform discharges through blockade of Na⁺ and Ca⁺ channels [28]. Given that LEV and VPA blunt neuronal activation through different mechanisms, they have demonstrated to be more effective together rather than independent agents in the management of intractable GRE [23].

Figure 1.

Mechanism of action levetiracetam blocks the release of neurotransmitter vesicles by inhibition of SV2A. Valproic acid increases the concentration of GABA via inhibition of GABAt. Valproic acid blocks Na+ and Ca2+ channels.

Abbreviations: AMPAr, AMPA receptor; GABA, -aminobutyrate; GABAt, GABA transaminase; NMDAr, N-methyl-Daspartate receptor; SV2A, synaptic vesicle protein 2A.

3. Monotherapy – levetiracetam

LEV is a first-line agent and likely the most common ASM used in the management of BTRE. It carries the following pharmacokinetic advantages: absence of cytochrome P450 modulation, low potential of drug-drug interactions, good bioavailability in IV formulations, rapid titration to an effective dose, and relative safety at pregnancy [29,30]. Additionally, LEV is suitable for a wide patient population in the United States because it is available in liquid formulations and is offered in a generic prescription. LEV may be associated with some of the highest rates of seizure freedom compared to other ASMs in GRE, with a 74% six-month seizure-free weighted average [31]. Furthermore, LEV has been associated with the highest likelihood of achieving a ≥ 50% reduction in seizure frequency and demonstrated the lowest treatment failure among other ASMs [31]. In a comparative study, LEV demonstrated similar efficacy as phenytoin in managing GRE, but LEV was preferred because of less seizure-unrelated dose adjustments and a more favorable side-effect profile [32]. After 2 years of treatment, the rates of seizure freedom (i.e., no seizures for 12 months) in WHO grade 2 GRE appears best achieved with LEV compared to other ASMs [33]. At this time, the ASM provided with the greatest support in the management of BTRE is LEV.

The most concerning side effect recognized with LEV is neuropsychiatric disturbance (e.g., emotional lability, anxiety, depression) [34,35]. These mood disturbances associated with LEV are most common when brain tumors are located in the frontal lobe [36]. Therefore, as a treatment option for BTRE, LEV can be impacted by the dose-limiting toxicity of irritability observed in a subset of patients. This is of importance, as many glioma patients are treated concurrently with steroids, which may further heighten the neuropsychiatric side effects of LEV. It is essential for clinicians to monitor for mood changes, depression, and irritability. The side effect profile of LEV is summarized in Table 1.

Table 1.

Side Effect Profile.

Side Effect Profile
Levetiracetam [29,37,40]	Somnolence, fatigue, infection, dizziness, anxiety, depression, aggression, agitation
Valproic Acid [41,45]	Weight gain, hair loss, polycystic ovary, pancreatitis, gastrointestinal disturbances, hepatoxicity, encephalopathy, coagulopathy, teratogenicity, tremor

4. Monotherapy - valproic acid

Focal seizures with or without secondary generalization are the most common manifestation of epileptogenic glioma [38,39]. VPA is a reasonable option for BTRE, given that it is has level B evidence for nonlesional focal epilepsies [14]. In addition to epilepsy management, VPA is also approved for the management of bipolar disorder and migraine prophylaxis. In an EANO survey of mostly European physicians, 21% chose VPA as a first-line alternative to LEV in managing BTRE [16]. Alternative options for the management of BTRE were not queried in another survey of American surgeons, but 85% selected LEV and 0% selected VPA as a first-line agent for postoperative prophylaxis [40]. Compared to LEV, VPA has been shown to result in more treatment failures in WHO grade 2–4 GRE 12 months after ASM initiation (50% vs 33%) [41]. Nonetheless, VPA has demonstrated efficacy in GRE in other studies. In a cohort of patients with glioblastoma (GBM), VPA and LEV had comparable seizure freedom at a follow-up ≥6 months (41% and 43%) [42]. At 6 months after surgery, VPA achieved postoperative seizure control in 61% of WHO grade 2 GRE [43]. Even though VPA attains efficacy for the management of GRE, it has fallen out of favor over LEV and newer generation ASMs, which might be directly related to its side effect profile.

VPA has a broad side effect profile, with the ability to cause injury across multiple organs. VPA may lead to neurological (e.g., tremor), hematological (e.g., thrombocytopenia), and gastrointestinal (e.g., nausea, constipation) symptoms [44]. Albeit uncommon, life-threatening toxicities are associated with VPA (e.g., cutaneous hypersensitivity reactions, hepatitis, and pancreatitis) [45]. Due to the incidence of transaminitis (15%) [46], hematologic abnormalities (33%) [47], and hyperammonemia (28%) [48], routine lab monitoring should be considered. A baseline complete blood count with differential and liver function tests should be performed with subsequent routine monitoring, especially within the first 6 months of treatment. The majority of adverse effects of VPA are often controlled with appropriate dosage adjustments. Importantly, VPA is a cytochrome P450 enzyme inhibitor; therefore, VPA may lead to drug-drug interactions and exclude patients from clinical drug trials [49]. Valproic acid should be avoided in women of childbearing age as it carries the highest risk of teratogenic effects including risk of neural tube defects [50], lower intelligence quotient [13], neurodevelopmental disorders (i.e., autism) [51] in offspring. The side effects of VPA are summarized in Table 1.

5. Dual therapy - levetiracetam & valproic acid

GRE may be refractory to ASMs [20,52,53]. In these cases, a trend toward an add-on ASM is the choice of management rather than sequential monotherapy trials. LEV-VPA combinatorial therapy has been shown to enhance seizure control for GRE compared to each of their stand-alone managements [23]. This is most likely because VPA and LEV mitigate cortical hyperexcitability through different mechanisms. A SNO consensus recommends LEV-VPA as a potential treatment option for BTRE if a first-line agent is ineffective [15]. Among dual ASM combinations, the value of LEV-VPA for intractable BTRE is documented in a few reports [23,42,54]. In a cohort of various brain neoplasms, LEV-VPA dual therapy, unsurprisingly, provided better seizure control compared to VPA or LEV monotherapies after ≥6 months of treatment [54]. Among patients with WHO grade 2–4 GRE, a large multicenter analysis further distinguished LEV-VPA dual therapy to provide superior seizure control than other alternative dual therapy combinations that included either VPA or LEV in their regimen (e.g., clobazam, phenytoin, lacosamide, etc.) [23]. Furthermore, in a study of GBM-related epilepsy, a subset of patients continued to have seizures with VPA or LEV. Among them, 54% achieved seizure freedom after LEV-VPA co-therapy [42]. Given their different mechanisms of action and combined tolerability, LEV-VPA dual therapy is a suitable, but certainly not the only, next-step treatment option for intractable BTRE.

6. Thrombocytopenia - temozolomide & valproic acid

Therapy with either temozolomide (TMZ) or VPA can lead to low blood cell counts, bringing into question if there is an increased risk of cytopenia for patients treated simultaneously with both of these agents [55,56]. In particular, platelet levels are vulnerable to decline with TMZ or VPA treatment [47,57]. Although the exact mechanism is unknown, VPA is thought to peripherally destroy platelets [47]. Myelosuppression is often seen with TMZ because of the low expression of MGMT in the bone marrow [58]. In matched cohorts of GRE (±TMZ) managed with either VPA (n = 429) or LEV (n = 429), ASM failure rates from adverse events were comparable among the two agents at 16% for LEV and 17% for VPA. However, low platelet counts led to 14% of all discontinuations of VPA. Ten patients (2.3%) developed grade 1/2 and 6 patients (1.3%) developed grade 3/4 thrombocytopenia [40]. In a clinical trial subanalysis of newly diagnosed GBM patients receiving chemoradiotherapy, VPA led to grade 3 or 4 thrombocytopenia. Patients had delay in their TMZ treatment, but ultimately, they did not have more dose reductions [59]. In a retrospective study of GBM patients treated with TMZ-VPA, TMZ was the only variable that reached the platelet threshold (<100,000/mm³) for TMZ dosage modification [60]. During and 4–8 weeks after chemoradiotherapy, relative platelet stability has been observed in VPA compared to LEV or no prescribed ASM. Levels of neutrophils, lymphocytes and thrombocytes decreased unanimously across these three treatment groups [61]. It is worthy to mention, in addition to TMZ and VPA, radiation that is a part of the standard treatment of HGG may contribute to hematological toxicity [62]. In the aggregate, in contrast to other ASMs, VPA has a slightly increased risk for thrombocytopenia in patients with HGG who are prescribed concurrent TMZ treatment. Therefore, routine hematological surveillance should be performed to monitor patients if VPA is chosen as agent to manage GRE.

7. Seizures & overall survival

Although perhaps counterintuitive, a preoperative seizure has been associated with prolonged OS in glioma, and this association has been most described in WHO grade 3–4 glioma [63–66]. About one-fourth of patients are diagnosed with GBM after they present with a seizure as their initial symptom [38,67,68]. In a retrospective nationwide study, mortality risk was lower in patients with a seizure <1 year prior to WHO grade 2–4 glioma diagnosis [69]. A notable limitation of many studies in this space is the incorporation of both isocitrate dehydrogenase (IDH)-wild-type (older age, unfavorable prognosis, lower seizure incidence) and IDH-mutant (younger [70], favorable [71], higher [3]) glioma patient populations into single cohorts. Moving forward, studies exclusively analyzing patients based on the 2021 WHO Classification of CNS Tumors [72], which codified adult infiltrating glioma classification on the IDH mutational status, will hopefully overcome this substantial limitation.

Patients with HGG are susceptible to a delay in diagnosis when presenting with nonspecific signs (e.g., headache, weakness, fatigue) [73]. In the event of an abrupt, unprovoked seizure, HGG patients may be more inclined to seek medical attention (e.g., present to an emergency room) and receive brain imaging. Thus, glioma patients with a seizure may present in an earlier stage of disease [74] and may receive earlier treatment [75]. On the other hand, an initial seizure may simply cause a diagnostic lead-time bias and skew longer OS. Contrast-enhanced MRI of newly diagnosed GBM have also shown a distinctive microstructure for epileptogenic tumors which might indicate a less aggressive growth pattern [76]. Volumetric analyses measure epileptogenic HGGs smaller than their non-epileptogenic analog [68,76–79]. In turn, a smaller tumor increases the probability to achieve a greater extent of resection, a robust prognostic marker [80]. Moreover, epileptogenic glioma may be surgically advantageous, given their more accessible location at the superficial cortex [78]. Patients presenting with seizures are also younger at the time of primary glioma resection [1,37,78,79,81,82]. An increased duration of functional independence (a correlate to extended OS) is furthermore observed with a preoperative seizure in patients with GBM after surgery [83].

Other reports have not found any association between longer OS and seizures in GBM [79,84,85]. The prognosis of GBM and seizures seems to be impacted differently across age groups, with older patients demonstrating decreased survival with epilepsy [86]. The general comorbidities associated with seizures (e.g., stroke) can also lead to worsened survival [87]. Lastly, preclinical data have expounded on the association of neuronal excitability and malignancy in glioma [6,7]. The direct synaptic connection between neurons and glioma has been implicated to drive glioma progression and invasiveness [6,7]. There are a multitude of variables to take into consideration while evaluating the effect of seizures on glioma survival. Currently, more data is required to determine whether there is a connection between epileptogenic glioma and OS, but this outcome appears to be different between specific subgroups of patients.

8. Prolonged survival

TMZ, a DNA methylating agent, is a part of the standard management of malignant glioma [88]. LEV is thought to hold value in the management of HGG, in part due to its purported bioaction to augment the apoptotic effects of TMZ via O⁶-methylguanine-DNA methyltransferase (MGMT) transcriptional inhibition in vitro [89–91]. Single-institutional retrospective studies have shown prolonged median OS with TMZ-LEV dual therapy versus TMZ monotherapy [92–95]. The impact of these studies is limited by their intrinsic design. A single institution retrospectively case-matched GBM patients (n = 460) who either received continuous, intermittent or no LEV during standard chemotherapy. The investigators found that continuous LEV treatment was associated with longer OS compared to the two other cohorts [95]. Perioperative administration of LEV has been correlated with increased progression-free survival (PFS) and OS [96]. In an open-label single-arm phase 2 study, LEV was associated with prolonged GBM survival in younger patients and those with unmethylated MGMT promoter status [97]. These studies indicating increased survival with LEV have several methodological limitations. As an example, in one study, patients who had extended survival with LEV were diagnosed with HGG after a seizure and a higher percentage of patients underwent complete resection [92]. The survival benefit of LEV treatment has not been observed in other studies [98–100]. A double-blind, randomized control trial is ongoing to determine the survival outcome of TMZ-LEV treatment in Chinese GBM patients [101]. Before impacting clinical practice, more data is needed to validate LEV favorably impacts OS.

The finding that VPA behaves as a histone deacetylase (HDAC) inhibitor has also implicated its potential use as an antitumor agent [102]. In addition to this mechanism, there are various other modalities by which VPA might facilitate an antitumor effect. VPA is reported to mitigate tumoral proliferation [103], invasion [104], and angiogenesis [105], in addition to attaining a synergistic [106–108] effect with the antineoplastic effects of TMZ in several glioma cell cultures (Figure 2). It also demonstrates a radio-sensitizing [109,110] and radio-protecting [109] effect upon GBM cells and healthy cells, respectively. It is possible that the benefit of HDAC inhibition may be most pronounced in the IDH-mutant glioma population, where this approach has not been as extensively studied [111]. Retrospective studies initially suggested that VPA co-medication during chemoradiation extends OS [42,59,112,113]. This observation was supported by systematic reviews and meta-analyses [24,114,115]. However, an increase in OS has not been recapitulated in more recent studies [24]. It is possible enhanced survival with VPA can be confounded by age. In a nationwide analysis, TMZ-VPA treatment prolonged OS to the greatest extent in younger patients [116]. Lastly, after pooling data from four randomized control trials, there was no improvement of OS and PFS in newly diagnosed GBM patients when they were treated with VPA (or LEV), indicating that ASMs should only be used to manage seizures and not malignancy [117].

Figure 2.

Levetiracetam and valproic acid attain a synergistic effect with temozolomide in cell culture studies. Valproic acid has shown anti-glioma activity in vitro through mitigation of cellular proliferation, tumor invasion and angiogenesis. Valproic acid induces radiosensitization of tumor cells and inhibits HDAC.

Abbreviations: HDAC, histone deacetylase.

9. Future perspective

Randomized clinical trials are one means to evaluate both the anticonvulsant efficacies and antineoplastic properties of LEV and VPA. However, as resources are limited, and the expectation is modest for a marked clinical impact with the addition of an ASM to standard of care treatments, the likelihood of such trials moving forward and completing enrollment is somewhat low [118]. These trials, if conducted, could provide data regarding seizure control, toxicities, and survival outcomes of LEV and VPA. Analyses of such trials should separate tumors based on molecular markers. IDH is one example, as an increased risk of seizures and increased OS is observed with this mutational status [3]. D-2-Hydroxyglutarate, an oncometabolite produced in IDH-mutant glioma, has shown to contribute to this seizure activity [119], and IDH inhibitors appear to decrease seizure burden [120]. Stratification of glioma based on tumor grade is also insightful, given that the tumor grade is inversely related to seizure incidence [1]. Furthermore, there is uncertainty to whether LEV or VPA penetrate malignant glia to produce a significant tumoricidal environment in vivo. The appropriate dosages, if a difference exists, for seizure control compared to anti-glioma activity needs to be determined for these agents. Phase 0 clinical trials can help evaluate these queries by measuring intratumoral concentrations of LEV and VPA that is administered prior to surgical resection. While such trials have been proposed, to our knowledge, none are actively accruing patients at this time.

10. Conclusion

LEV and VPA are two commonly used agents for seizure control in glioma. They are both categorized as first-line agents, but VPA is more often utilized as an add-on or alternative agent when a monotherapy is ineffective. LEV and VPA have demonstrated antineoplastic effects in preclinical studies. In addition, some retrospective studies have also suggested increased survival with these agents. At this time, the proposal that LEV and VPA increases glioma survival and incorporation of these agents into clinical practice for this reason has not been validated in randomized trials and the current evidence is not strong enough to recommend clinical use beyond antiseizure activity. The majority of studies supporting prolonged survival are utilizing data extrapolated from retrospective studies where compliance and dosage of LEV and VPA were not quantified. In addition, the extended survival observed with these agents has been most pronounced in the younger population [97,116]. VPA is suggested to slow malignant progression in preclinical studies via inhibition of HDAC. Previous clinical studies investigating HDAC inhibition (e.g., vorinostat, romidepsin) as a therapeutic option for GBM have been ineffective [121,122]. In comparison to these trialed agents, VPA is a less potent HDAC inhibitor, and it appears therefore less probable VPA is able to slow glioma progression through this mechanism. Nonetheless, numerous modalities have been described by which VPA exerts its purported antineoplastic effects [123]. The association of seizures and glioma survival is a complex topic with heterogeneous findings. Survival is most likely influenced by a multitude of factors, including but not limited to tumor grade, molecular subtype, transcriptional modification (e.g., MGMT methylation) and baseline cohort characteristics (e.g., age, functional status). The idea of repurposing ASMs as an adjuvant agent for glioma progression will require further investigation through clinical trials prior to impacting the clinical practice.

Funding Statement

R.V.L has received funding from P50CA221747 and BrainUp.

Disclosure statement

R.V.L has received research support (drug only) from BMS, honoraria for serving on Advisory Boards for AstraZeneca, Cardinal Health, Curio, Merck, Novartis, Novocure, Servier, and Telix, honoraria for consulting for Novartis and Servier, honoraria for serving on the Speakers’ Bureau for Merck, Novocure, and Servier, and honoraria for editing for EBSCO, Elsevier, Medlink Neurology, and Oxford University Press. T.W. is on the Advisory Board of Alexion, Springworks, Servier and the Data Monitoring Board of Novocure, IQVIA. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Article highlights

• Two first-line agents utilized in the management of glioma-related epilepsy include levetiracetam and valproic acid.

• Beyond seizure control, an extended survival via anticancer mechanisms have been attributed to both levetiracetam and valproic acid.

• Levetiracetam and valproic acid in combination may increase seizure freedom rates in patients with refractory glioma-related epilepsy.

• Seizures are believed to influence survival in glioma; although, these findings are heterogeneous across tumor and patient characteristics.

• Prospective studies are needed to validate whether antiseizure medications have a role in the oncologic management of glioma.

Author contributions

BFK, RVL, and JWT conceptualized the content and structure of the review. BFK wrote the initial draft of the manuscript. BFK, TW, KD, CH, KG, HT, MCT, RS, RVL, and JWT extensively revised the manuscript for intellectual content. All authors approved the final draft.