If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Clinical and Experimental Pharmacology Unit, Department of Internal Medicine and Therapeutics, University of Pavia, Pavia, ItalyC. Mondino National Neurological Institute, Pavia, Italy
We analyzed the adverse events (AEs) profile of levetiracetam with a meta-analysis.
•
A combined analysis of children and adults with any disease was performed.
•
The risk of withdrawals was higher for patients taking levetiracetam than placebo.
•
A general good tolerability profile of levetiracetam has been evidenced.
•
We did not find a clear dose–response relationship for the significant AEs.
Abstract
Purpose
To analyze the adverse events (AEs) significantly associated with levetiracetam (LEV) therapy through a meta-analysis of all available double-blind, randomized placebo-controlled trials (RCTs), performed in any age, gender, ethnic background and disease. General tolerability and study withdrawals due to AEs associated with LEV treatment were also investigated. In addition, a dose–effect responses relationship for all variables was assessed.
Methods
RCTs were identified searching Medline (PubMed), Embase and Cochrane CENTRAL for the words “Levetiracetam” and “randomized controlled trial”, with different search strategies, setting the limits “humans” and “English”. Very common and common AEs according to the summary of product characteristics were investigated.
RevMan version 5.2 was used for the statistical analyses. Risk difference with 95% confidence intervals was used to investigate the association of any AEs and withdrawal with LEV.
Results
Twenty-six studies with 2832 patients were included in the RCTs analysis. Nasopharyngitis, somnolence, dizziness, nervousness/irritability and asthenia/fatigue were statistically significant associated with LEV. In addition, LEV was significantly associated with an increased risk of AEs-related withdrawals. No dose–response relationship was found for any of the assessed variables.
Conclusions
This first large meta-analysis suggests that participants were more likely to discontinue LEV than placebo.
The AE profile confirmed that LEV is associated with few unfavorable sedative, vestibulocerebellar and behavioral effects, such as nervousness and irritability. However, there does not seem to be a clear dose–response relationship.
Levetiracetam (LEV) is an antiepileptic drug with multiple mechanisms of action that correlates with his broad spectrum of activity. It inhibits neuronal hyper-synchronization, most likely by binding to synaptic vesicle protein 2a (SV2A) located in presynaptic membranes [
]. LEV also activates the γ-aminobutyric acid (GABA) and glycine system, interacts with the AMPA receptors, and partially depresses the N calcium current: all targets that have been associated with alcohol-seeking and drinking behavior [
The wide broad spectrum of activity of LEV is supported by various double-blind studies that demonstrated its effectiveness and safety as add-on therapy in generalized epilepsies, such as myoclonic seizures [
]. The FDA approved LEV as an adjunctive therapy for partial onset seizures in adults and children ≥4 years of age, adjunctive therapy in the treatment of myoclonic seizures in adults and adolescents ≥12 years of age with juvenile myoclonic epilepsy and as adjunctive therapy in the treatment of primary generalized tonic–clonic seizures in adults and children ≥6 years of age with idiopathic generalized epilepsy. In contrast, the EMEA approved LEV for the treatment of partial-onset seizures with or without generalization in patients from one month of age, juvenile myoclonic epilepsy in patients from 12 years of age, and of idiopathic generalized epilepsy in patients from 12 years of age.
LEV has a favorable pharmacokinetic profile that has been determined over a dose range of 500–5000 mg for adults and 20–30 mg/kg for children [
] have shown a good safety profile for this drug. However, up to now, there are no large meta-analyses performed on all available RCTs on LEV in any condition. In fact, the existing meta-analyses separately evaluated the RCTs performed only on a specific disorder.
The aim of our study was to carry out a comprehensive meta-analysis on the safety profile of LEV using a large sample of RCTs in adults and children with regard to dose-dependent effects.
2. Methods
2.1 Criteria for considering studies in this review
2.1.1 Types of studies
We included only double-blind, placebo-controlled RCTs with parallel or crossover design that investigated the therapeutic effect of oral LEV in children and adults with various diseases. Studies performed on healthy volunteers and/or lasting less than 4 weeks were excluded. Full journal publications or summary clinical trial reports were required. Brief abstracts were excluded. Other types of studies, including non-randomized trials, case reports, or clinical observations, were not included.
2.1.2 Types of participants
Subjects with any age and ethnic background taking LEV for any clinical disorder, including partial or generalized epilepsy, essential tremor, multiple sclerosis, alcohol dependence, neuropathic pain, polyneuropathy, bipolar depression, social anxiety disorder, dyskinesia in Parkinson's disease. Studies on participants who underwent a concomitant drug therapy were included.
2.1.3 Search methods for identification of studies
In April 2013, we conducted a computer search of Medline (PubMed version), EMBASE and Cochrane library for all double-blind placebo-controlled studies performed on LEV. Therefore, the limit “humans” was set and we restricted our search to articles written in English. Search strategies and a flow diagram of the selection process are available in Appendix A (Fig. A1).
Reference lists of retrieved articles were examined and summary clinical trial reports not available as full publication were searched on the Internet in order to find additional articles. Both studies randomizing patients to a fixed or a flexible dose of LEV were selected for the purpose of this meta-analysis.
Eligibility was determined after reading each study identified by our research; two blinded reviewers independently assessed trials for inclusion in the meta-analysis. Any disagreements were resolved by discussion. Further details are provided in Appendix B.
2.2 Analysis of adverse events
2.2.1 Selection of adverse events
Relevant adverse events (AEs) were selected following a three-step strategy: (1) identification of synonyms (Appendix B, Table B1), which were grouped under one main term; (2) exclusion of rare AEs and very rare AEs (respectively from 1/10,000 to 1/1000 and less than 1/10,000, as reported by information leaflet) among those randomized to LEV or placebo (PCB); (3) exclusion of not assessable AEs, such as those reported in a single study and those not ascribable to a clear number of patients. After selecting those AEs to be included in our analyses, we extracted the frequency of each AE reported from LEV and PCB groups of each study. This information was recorded into a separate data extraction sheet.
2.2.2 Statistical analysis
Statistical analyses were performed using RevMan version 5.2 and IBM SPSS Statistics 19.0 [
]. Statistical heterogeneity between trials was checked for each outcome using a Chi2 test for heterogeneity. The I2 test was also used to assess statistical heterogeneity, with an I2 > 70% indicating heterogeneity. Where significant heterogeneity was present (only for asthenia/fatigue and dizziness at 2000 mg/day), random-effect method was used. A fixed-effect model was used in all other outcomes showing no significant heterogeneity. The preferred estimator was the Mantel–Haenszel risk difference (RD). For the outcomes, 95% confidence intervals (CIs) were set.
We evaluated: (1) RDs to assess the association of general tolerability (any AE), included in this analysis with LEV treatment; (2) RDs to assess the association of patients withdrawing from studies, because of AEs due to LEV; (3) RDs of 12 AEs, those described in an adequate number (>3) of included studies, in order to find which AEs were significantly associated to LEV therapy; (4) RDs to investigate dose–response relationships of LEV in term of general tolerability, withdrawal and AEs. For this latter purpose, only those studies (n = 21) in which subjects were randomized to fixed-dose regimens (1000, 2000, 3000 mg/day) were selected, since in studies allowing flexible-dose regimens (n = 4) it was not possible to assign an AE to a specific drug dose. Therefore, all children taking a flexible dose were excluded from this analysis, while adolescents (>13 y) receiving the same dosage than adults were not excluded. Given that only one included study Betts et al. [
A multicentre, double-blind, randomized, parallel group study to evaluate the tolerability and efficacy of two oral doses of levetiracetam, 2000 mg daily and 4000 mg daily, without titration in patients with refractory epilepsy.
] adopted a drug regimen of 4000 mg/day, that could not be included in the dose–response analysis.
In order to easy interpret the dose–response relationships, each RD estimate was complemented by the computation of the number needed to harm (NNH). An inverse variance-weighted meta-regression model was also applied to test the relationship between LEV dose and RDs.
3. Results
3.1 Included studies
Twenty-five studies (2832 participants) satisfied our inclusion criteria and were entered into our analysis (Table 1). Twelve studies were performed with a total of 2114 epileptic patients, mainly affected by refractory partial seizures (1831, 86.6%), 283 (13.4%) with idiopathic generalized epilepsy. The remaining 13 studies were performed in 468 (22.1%) patients affected by other diseases (for more information see Appendix C). In all studies, AEs were calculated on the intention-to-treat (ITT) population. The total number of patients treated with the active drug was1698, whereas patients treated with PCB were 1329.
Table 1Characteristics of included studies divided into studies on epileptic patients (A) and non-epileptic patients (B).
A multicentre, double-blind, randomized, parallel group study to evaluate the tolerability and efficacy of two oral doses of levetiracetam, 2000 mg daily and 4000 mg daily, without titration in patients with refractory epilepsy.
Neurocognitive effects of adjunctive levetiracetam in children with partial-onset seizures: a randomized, double-blind, placebo-controlled, noninferiority trial.
Once-daily extended-release levetiracetam as adjunctive treatment of partial-onset seizures in patients with epilepsy: a double-blind, randomized, placebo-controlled trial.
Multicenter double-blind, randomized, placebo-controlled trial of levetiracetam as add-on therapy in patients with refractory partial seizures. European Levetiracetam Study Group.
Efficacy and safety of levetiracetam (up to 2000 mg/day) in Taiwanese patients with refractory partial seizures: a multicenter, randomized, double-blind, placebo-controlled study.
Multicenter double-blind, randomized, placebo-controlled trial of levetiracetam as add-on therapy in Chinese patients with refractory partial-onset seizures.
In two trials the participants were only children (N = 296) with an age range of 4–16 years. Thirteen studies analyzed only adults patients (N = 718, age range from 21 to 83 years). In the remaining 10 trials the study population includedboth adults and children (age range from 4 to 70 years).
The majority of trials had parallel study design, while only 7 had a crossover design.
Most trials recruited patients from USA and Europe, 2 trials recruited participants from China, 1 from Taiwan and the other studies recruited from various countries (centers in Europe, Mexico, Australia, New Zealand, USA, Canada, South Africa, India and Russia).
The treatment period included an up-titration followed by a maintenance phase in all but 2 trials, which did not include a dose-up titration period. Treatment duration ranged from 2 to 24 weeks (up-titration range 4 days to 7 weeks).
LEV dosage ranged from 30 mg/kg/day to 60 mg/kg/day for children and from 500 mg/day to 4000 mg/day for adults. The trial performed by Betts et al. [
A multicentre, double-blind, randomized, parallel group study to evaluate the tolerability and efficacy of two oral doses of levetiracetam, 2000 mg daily and 4000 mg daily, without titration in patients with refractory epilepsy.
] was the only one in which a 4000 mg dose was tested.
The RCTs included in our analysis investigated the use of LEV in very different disorders, including partial (10) and generalized epilepsies (2), essential tremor (2), multiple sclerosis (1), alcohol dependence (1), spinal cord injury pain (1), polyneuropathy (1), central neuropathic post-stroke pain (1), bipolar disorder (1), l-Dopa or tardive dyskinesia (1), social anxiety disorder (1) and post-mastectomy syndrome (1).
Many concomitant drugs were administered to patients during the trials (see Appendix C).
Data for AEs were available in all studies, while data for withdrawals were lacking in 3 trials.
The minimal incidence required for each AE was set between 5 and 10%, but not all authors set a minimal percentage.
3.2 General tolerability
General tolerability was assessed by evaluating the number of patients reporting at least one AE. The overall RD for ‘any AE’ did not reach statistical significance (RD 0.02, 95% CI −0.02 to 0.05). When compared by dosage, a staistical significance was found only for the 3000 mg dosage (RD 0.07; 95% CI 0.01–0.12) compared to 1000 mg and 2000 mg (RD −0.01; 95% CI −0.08 to 0.06 and RD 0.18; 95% CI −0.21 to 0.56 respectively), as shown also by NNH values (Fig. 2, Fig. 3, Appendix B, Table B2).
3.3 Treatment withdrawals
The overall RD discontinuation for AEs at any age was 0.03 (95% CI 0.01–0.05); thus there was significant evidence that participants were more likely to discontinue LEV than PCB. Dose-withdrawal analysis showed a lower RD in patients taking 1000 mg (−0.01, 95% CI −0.04 to 0.03) compared to patients taking 2000 mg and 3000 mg (respectively 0.05, 95% CI 0.01 to 0.09 and 0.04, 95% CI 0.01 to 0.06). (Fig. 1, Fig. 2, Fig. 3). In Fig. 2b data are reported as NNH. In the regression analysis, the risk of withdrawal due to AEs was not dose-dependent (p = 0.186).
Fig. 1Forest plot of RDs (95% CI) for all assessed AEs.
Fig. 2Number needed to harm (NNH) at LEV 1000, 2000, and 3000 mg/day for all* statistically significant AEs (A), risk of withdrawals and Any AE (B). (*) nasopharyngitis has been excluded because this AE did not show statistical significance when grouped by dosage.
Fig. 3(A) RDs (95% CIs) for all the assessed variables in patients randomized to LEV at 1000 mg/day, 2000 mg/day and 3000 mg/day. (B) Focus on RDs of withdrawals in patient randomized to LEV at 1000 mg/day, 2000 mg/day and 3000 mg/day.
From our analysis, the AEs associated with LEV were: nasopharyngitis (RD 0.03; 95% CIs 0.00–0.06), nervousness/irritability (RD 0.06; 95% CIs 0.02–0.09); somnolence (RD 0.07; 95% CIs 0.05–0.09); asthenia/fatigue (RD 0.07; 95% CIs 0.04–0.10) and dizziness (RD 0.04; 95% CIs 0.02–0.06) (Fig. 1, Fig. 2, Fig. 3).
3.5 Dose–effect correlations of statistically significant levetiracetam adverse effects
We excluded 4 studies because they use flexible-dose regimens or they provided unclear data. Therefore, the sample eligible for this analysis was represented by 21 studies, including adults and adolescents (>13 years). RDs (95% CIs) for the statistically significant LEV AEs at different doses are shown in Table 2 (see Table B2 in appendix for more information). When grouped by dosage, statistical significance was reached only by somnolence, nervousness/irritability, and dizziness at 1000 mg/day, somnolence at 2000 mg/day, whereas all four AEs were statistically significant at LEV 3000 mg/day. A statistically significant dose–response relationship was not found, as confirmed by the regression analyses (p = 0.134; p = 0.819; p = 0.135; p = 0.688 respectively for asthenia/fatigue, somnolence, dizziness and nervousness/irritability) and NNH values (Fig. 2a). Nasopharyngitis did not reach statistical significance when grouped by dosage.
Table 2Risk difference [95% CIs] for 5 statistically significant AEs, withdrawals and any AE at different doses.
Accurate analysis of AEs is essential for the optimal choice of AED, given that AEs may lead to treatment discontinuation in up to 25% of patients, and contribute to poor patient adherence [
Comparative efficacy and tolerability of anti-epileptic drugs for refractory focal epilepsy: systematic review and network meta-analysis reveals the need for long term comparator trials.
Efficacy and safety of pregabalin versus levetiracetam as adjunctive therapy in patients with partial seizures: a randomized, double-blind, noninferiority trial.
], but they failed to show a significant association between the drug and several AEs, as these studies were restricted to RCTs. Therefore, the number of recruited patients was hampered by insufficient large sample sizes to fully uncover the AE profile of a given AED.
Recent evidence, however, suggests that non-epilepsy studies can be merged with epilepsy studies in meta-analyses of RCTs of AEDs [
Adverse events of antiepileptic drugs across indications: can randomized controlled trials data from non-epilepsy indications be included in meta-analysis for AEDs used in epilepsy?.
in: Spec issue 29th int epilepsy congr Rome, 52, 28th August–1st September 20112001: 120
Drug interactions with the newer antiepileptic drugs (AEDs)—Part 2: pharmacokinetic and pharmacodynamic interactions between AEDs and drugs used to treat non-epilepsy disorders.
] we decided to extend the target group also to non-epileptic patients. Moreover, a combined analysis of children and adults was performed in order to extend our sample. In this meta-analysis of all available RCTs of LEV, we included 2832 patients and found 5 AEs significantly associated with LEV treatment, while Mbizvo et al. [
] meta-analysis, performed on 11 trials for a total of 1861 patients identified 2 AEs significantly associated with LEV: somnolence (RR 1.58; 99% CIs 1.14–2.18) and infection (RR 1.76; 99% CIs 1.03–3.02). Furthermore, in the Mbzivo study there was insufficient evidence that participants were more likely to discontinue LEV than PCB (RR 1.03; 95% CIs 0.80–1.33). Lo et al. [
] performed another meta-analysis of 10 RCTs: the likelihood of serious AEs requiring withdrawal from the study was not significantly different between LEV patients and controls (RR = 1.37; 95% CIs 0.88–2.13). The meta-analysis of Zaccara et al. [
] was performed on 5 RCTs and found only somnolence being associated with LEV (RD 0.07; 95% CIs 0.01–0.12).
In previous meta-analyses specifically focused on the assessment of tolerability of AEDs from all RCTs, 20 AEs were significantly associated with pregabalin (PGB) in a population of almost 12.000 subjects [
]. Compared to Levite spectrum of the AEs observed with the other AEDs is wider for PER and PGB and narrower for LCM. Vestibulocerebellar dysfunction resulted a common effect of all these AEDs [
Comparative efficacy and tolerability of anti-epileptic drugs for refractory focal epilepsy: systematic review and network meta-analysis reveals the need for long term comparator trials.
In the current study, the 5 AEs significantly associated with LEV were: nasopharyngitis, somnolence, asthenia/fatigue, dizziness and nervousness/irritability. The association between nasopharyngitis and LEV has already been described by Mbizvo et al. [
]. Although we restricted the synonyms to upper respiratory tract infections, a possible explanation for the high frequency of nasopharyngitis may be an overlap in reporting caused by the variety of terms. Other authors suggested that the improvement in seizure control may favor socialization and consequently increase the risk of respiratory infections [
]. However, the correlation between nasopharyngitis and LEV remains unclear, given that previous analyses did not show a concomitant decrease in white blood cell and neutrophil count suggesting a reduction of immunologic response to LEV [
Both somnolence and asthenia/fatigue may reflect the depressant CNS effect of LEV. Asthenia and fatigue might also result from decreased excitatory activity [
]. Somnolence and some cognitive disturbances (confusional state, abnormal thinking, disturbance of attention) were significantly associated with PGB [
A further AE associated with LEV was nervousness/irritability, which has already been reported in a previous systematic review of open label studies aimed at assessing behavioral effects of LEV in children [
]. Psychiatric effects, such as anxiety, depression and psychosis, were very rarely associated with LEV in RCTs. Only long-term observational studies provided data about an increased risk of psychiatric disorders in a subset of patients with pre-existing mood disorders [
]. Dizziness, which is a subjective AE, was the only AE related to vestibulocerebellar/brainstem dysfunction associated with LEV. Other objective and more severe AEs caused by involvement of the same brain structures (ataxia, diplopia, balance disorders, and so on) are commonly observed in individuals treated with AEDs acting on voltage-gated sodium channels, such as phenytoin, carbamazepine, lamotrigine, and oxcarbazepine [
Efficacy and safety of pregabalin versus levetiracetam as adjunctive therapy in patients with partial seizures: a randomized, double-blind, noninferiority trial.
Probably, the different spectrum of AEs linked to LEV is caused by its peculiar mechanism of action: the predominant effect of LEV on cortical function and alertness may be attributed to its action on the GABA and SV2A receptors. Moreover, the anti-epileptic actions of LEV seem to be mediated, at least partly, via the combination of inhibitory effects on depolarization-induced and Ca2+-induced Ca2+ release-associated neurotransmitter releases.
An interesting finding of our meta-analysis was the unclear relationship between LEV tolerability and dose.
Specifically, LEV tolerability, measured as percentage of patients withdrawing because of AEs and as percentage of patients with any AE, showed a particular dose-relationship (see Fig. 2, Fig. 3). Whereas tolerability clearly decreased by a change in drug dose from 1000 to 2000, a further dose increment (up to 3000 mg/d) was not associated with a worsening of tolerability. This is different from what has been observed with other AEDs, which show a clear worsening of tolerability at higher doses [
]. The characteristic LEV tolerability profile was even more evident when the dose–effect relationship was analyzed for specific AEs. Only somnolence seemed to be influenced by drug dose. Strikingly, dizziness and irritability have a similar frequency at all drug doses.
Even though the reasons for these atypical findings are not clear, the lack of a dose–effect relationship for several AEs can explain the clinical profile of LEV, which is considered a drug with a very good tolerability [
Some limitations of this meta-analysis need to be acknowledged. Firstly, RCTs have some limitations that tend to overstate the effect of a new treatment: the restrictive inclusion and exclusion criteria may lead to the enrollment of a sample not-representative of the target patient population. In addition, the trial relatively short observation period may underestimate long-term toxicity. Secondly, there may be demographic and clinical heterogeneity since adults and children with different disorders could receive many different concomitant therapies. Thirdly, our analysis focused on the NNH of LEV; however, it is important to take into account that clinical decisions need to weigh both harm and benefits, which were not assessed in our study. Fourthly, grouping of synonyms under one single term carries the risk of double-counting of patients, but this approach was necessary in our analysis due to the variety of terms used for AE description in RCTs. Collective efforts should be made to develop a single standardized terminology for RCTs on AEDs. Finally, our study is limited by the fact that we did not check the database for clinical studies and by a potential bias related to studies with high AEs which have beennot published.
Due to these limitations, the analysis of AEs profile of LEV requires further investigations and validation.
5. Conclusions
In conclusion, our results show a general good tolerability profile of LEV, which is characterized mainly by a sedative effect, reflected by somnolence and asthenia/fatigue, minimal neuropsychiatric impairment (nervousness/irritability), along with mild vestibulocerebellar manifestations, represented by dizziness, and a weak increase in upper respiratory tract infections.
Of interest, all these 5 AEs did not seem to show a clear dose–response relationship. Further longitudinal well-conducted investigations will provide additional data on potential toxic and/or idiosyncratic effects and help in fully characterize the tolerability profile of this drug.
Conflict of interest statement
The authors received no funding for this study.
GZ has received speaker's or consultancy fees from EISAI, Jansen-Cilag, Sanofi-Aventis, and UCB Pharma. AV has received a grant from Viropharm. VF is a former employee of Eisai s.r.l. Italy. GP, FDS, and FC report no disclosures.
Acknowledgment
The authors are grateful to Dr. Loredana Marcovecchio for language revision.
Appendix A. Supplementary data
The following are supplementary data to this article:
A multicentre, double-blind, randomized, parallel group study to evaluate the tolerability and efficacy of two oral doses of levetiracetam, 2000 mg daily and 4000 mg daily, without titration in patients with refractory epilepsy.
Comparative efficacy and tolerability of anti-epileptic drugs for refractory focal epilepsy: systematic review and network meta-analysis reveals the need for long term comparator trials.
Neurocognitive effects of adjunctive levetiracetam in children with partial-onset seizures: a randomized, double-blind, placebo-controlled, noninferiority trial.
Drug interactions with the newer antiepileptic drugs (AEDs)—Part 2: pharmacokinetic and pharmacodynamic interactions between AEDs and drugs used to treat non-epilepsy disorders.
Once-daily extended-release levetiracetam as adjunctive treatment of partial-onset seizures in patients with epilepsy: a double-blind, randomized, placebo-controlled trial.
Multicenter double-blind, randomized, placebo-controlled trial of levetiracetam as add-on therapy in patients with refractory partial seizures. European Levetiracetam Study Group.
Adverse events of antiepileptic drugs across indications: can randomized controlled trials data from non-epilepsy indications be included in meta-analysis for AEDs used in epilepsy?.
in: Spec issue 29th int epilepsy congr Rome, 52, 28th August–1st September 20112001: 120
Efficacy and safety of levetiracetam (up to 2000 mg/day) in Taiwanese patients with refractory partial seizures: a multicenter, randomized, double-blind, placebo-controlled study.
Multicenter double-blind, randomized, placebo-controlled trial of levetiracetam as add-on therapy in Chinese patients with refractory partial-onset seizures.
Efficacy and safety of pregabalin versus levetiracetam as adjunctive therapy in patients with partial seizures: a randomized, double-blind, noninferiority trial.
00167-3&id=gr1.jpg){kind=link}
00167-3&id=gr2.jpg){kind=link}
00167-3&id=gr3.jpg){kind=link}