Efficacy of vigabatrin in the treatment of infantile epileptic spasms syndrome: A systematic review and meta‐analysis

Abstract This systematic review and meta‐analysis aimed to evaluate the efficacy of vigabatrin (VGB) in treating infantile epileptic spasms syndrome (IESS). Databases of PubMed, Embase, Web of Science, MEDLINE, and Cochrane Library were systematically searched. All the relevant randomized controlled trials (RCTs) and observational studies (OSs) of VGB for IESS were included and analyzed separately. The primary outcome was the cessation of epileptic spasms (ES). Five RCTs and nine OSs compared the efficacy of VGB vs hormonal monotherapy for IESS. Meta‐analysis of the five RCTs showed that hormonal monotherapy was significantly better than VGB monotherapy (OR = 0.37, 95% CI = 0.20‐0.67) for patients with new‐onset IESS. Meta‐analysis of the nine OSs agrees with the result from RCTs (OR = 0.61, 95% CI = 0.43‐0.85). VGB was more effective in patients with TSC than in those with other etiologies (five OSs, OR = 5.59, 95% CI = 2.17‐14.41). There was no significant difference in the efficiency of VGB combined with hormonal therapy vs hormonal monotherapy for IESS (two RCTs, OR = 0.75, 95% CI = 0.09‐6.45). Hormonal monotherapy is better than VGB monotherapy for non‐TSC‐associated IESS. But for patients with IESS due to TSC, VGB is the first choice. VGB combined with hormone therapy does not definitely increase ES control rates compared with that of hormonal monotherapy.


| INTRODUCTION
Infantile epileptic spasms syndrome (IESS) is a common developmental and epileptic encephalopathy characterized by the onset of epileptic spasms (ES) between 1 and 24 (peak 3 and 12) months of age. 1 Patients with IESS-a devastating epilepsy syndrome-typically suggest refractory to drugs and poor prognosis. Exploring effective treatments and optimizing treatment regimens is essential. At present, adrenocorticotropic hormone (ACTH), steroids, and vigabatrin (VGB) are considered as the most effective treatments in the short-term control of ES. 2,3 The antiseizure effects of VGB are attributed to the selective and irreversible inhibition of γ-aminobutyric acid (GABA) transaminase and GABA-degrading enzyme activity, thereby increasing GABA levels-the major inhibitory neurotransmitter in the central nervous system. 4 Besides, one study also noted that elevated endogenous GABA concentrations could attenuate glutamate-glutamine cycling between astrocytes and neurons. 5 VGB was initially found to be effective and well-tolerated in adult patients with refractory epilepsy in 1983. 6 As early as 1989, VGB was found to be effective for refractory IESS as an add-on therapy and especially effective in patients with tuberous sclerosis complex (TSC). [7][8][9] Since then, there has been an increasing number of studies analyzing the effectiveness and safety of VGB for patients with IESS. Meanwhile, the researchers also found that VGB could cause some severe adverse events. A meta-analysis showed that the risks of VGB-attributed retinal toxicity, visual field defects, and VGB-related magnetic resonance imaging abnormalities in 29% (95% CI: 7%-69%), 28% (95% CI: 4%-78%), and 21% (95% CI: 15%-29%), respectively. 10 Despite these adverse events, the efficacy of VGB in treating IESS was well recognized, and VGB was approved by the United States Food and Drug Administration in 2009 for monotherapy of IESS and as adjunctive therapy for refractory complex partial seizures. 11 In recent years, many meta-analyses compared ACTH and steroids for patients with IESS. [12][13][14] Guidelines and consensus suggested that VGB is effective for IESS, especially for patients due to TSC. 2,3,15 But there were no reviews, including all trials of VGB treatments for IESS, performed a meta-analysis. To provide strong certainty and high-quality evidence for the clinical management of IESS, we conducted a systematic review and meta-analysis for the efficacy of VGB in the treatment of IESS. All the relevant randomized controlled trials (RCTs) and OSs of VGB for IESS were included and analyzed separately. The purposes of the present study were (a) to compare the efficacy of VGB vs hormonal monotherapy for IESS, (b) to compare the efficacy of VGB in IESS patients with TSC vs patients with other etiologies, and (c) to compare the efficacy of VGB combined with hormonal therapy vs hormonal monotherapy for IESS.

| METHODS
This systematic review and meta-analysis were reported in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, 16 following the instructions of the Cochrane Manual for systematic reviews, 17 and it was registered prospectively with the PROSPERO database (CRD42022344775).

| Eligibility criteria
Inclusion criteria: (a) all randomized controlled trials (RCTs) and observational studies (OSs) of VGB for IESS were included; (b) all the patients were diagnosed with IESS based on the hypsarrhythmia on electroencephalogram (EEG) at the early presentation of ES 1 ; (c) mean age at onset of ES ≤24 months; (d) information on patients with the cessation of ES was detailed, including the number of patients, etiology, medication regimen, method of outcome measure, and duration of follow-up; (e) for the trials of compared the efficacy of different treatments, the experimental group patients were treated with VGB or VGB combined with hormonal therapy (ACTH or steroids) as initial therapy, and patients in the control group were treated with hormonal therapy; and (f) the patients were newly diagnosed with IESS and had not previously received treatment.
Exclusion criteria: (a) case reports, literature reviews, animal experiments, conference abstracts, letters, and non-English literature; (b) lack of original data or unable to get a full text; and (c) lack of information on ES frequency change were excluded.

| Information sources and search strategy
Databases of PubMed, Embase, Web of Science, MEDLINE, and Cochrane Library were systematically searched for relevant original research from their inception to December 9, 2022. The combination of Medical Subject Heading (MeSH) terms and free text was used for the search strategy. The search strategy was detailed in Table S1, including an exhaustive list of search terms used. The database was searched independently by two reviewers (Z.X. and Y.N.) according to the search strategy, and the number of papers obtained from each database search was verified.

Key Point
• For non-TSC-associated IESS, hormonal therapy is superior to vigabatrin (VGB) and can be used as initial monotherapy.
• For patients with IESS due to TSC, VGB is the most effective.
• VGB combined with hormone therapy does not definitely increase ES control rates compared with hormonal monotherapy.

| Study selection and data collection
Based on the aforementioned eligibility criteria, each article's title/abstract or full text was independently screened within Endnote software (Version X9) by two reviewers (Z.X. and X.J.) to determine all eligible studies. If there are any conflicts or discrepancies, they will be discussed and adjudicated by the third reviewer (Z.Y.). Two reviewers (Z.X. and X.J.) independently extracted data for the eligible studies, and conflicts were adjudicated by the third reviewer (Z.Y). From studies that met the criteria, the following information was extracted: first author, year of publication, country, study design, number of patients, age of ES onset, sex, medication intervention in each group (including dose, sample size), seizure outcomes, EEG outcomes, and duration of the follow-up. Our primary outcome was the number of patients with cessation of ES with or without EEG remission (resolution of hypsarrhythmia) in the short-term.

| Risk of bias assessment
The risk of bias in RCTs was assessed using the revised Cochrane risk of bias tool for randomized trials (RoB2). 18 The risk of bias in case-controls and cohort studies was assessed using the Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I). 19 Two reviewers (Z.X. and Y.N.) independently evaluated all included studies for risk of bias. We resolved any disagreements by discussion.

| Data synthesis
All statistical analyses were performed with R (version 4.1.0), RStudio (version February 2, 2022 + 485) software, and Meta packages (version 5.2-0). In all procedures, a P ≤ 0.05 was considered to be statistically significant.
Data from RCTs and OSs were pooled separately. A meta-analysis was performed only if there were at least two studies. The odds ratio (OR) with 95% confidence intervals (95% CI) was used to analyze all dichotomous outcomes. The heterogeneity of the data was examined using I 2 statistics, in which I 2 > 50% was considered substantial heterogeneity. 20 A random-effects model was employed if there was statistical heterogeneity (I 2 > 50%). Otherwise, a common-effect model was applied. In the sensitivity analysis, a meta-analysis was performed after removing each study individually to evaluate the quality and consistency of the results. The Egger test was used to assess publication bias if there were more than 10 studies for a meta-analysis.

| Study selection
The search strategy acquired a total of 3935 articles from the PubMed (n = 556), Embase (n = 1984), Web of Science (n = 757), MEDLINE (n = 544), and Cochrane Library (n = 94) databases. A flowchart of the study selection process was displayed in Figure 1. After removing duplicates (2243 studies remained), reviews, animal experiments, conference abstracts, case reports, letters, and non-English literature (a total of 1206) were excluded with the assistance of Endnote software. Then, by reading the titles and abstracts, 1006 studies were excluded for not meeting the eligibility criteria, the research topic did not meet the requirements, or being unable to get the full text. A total of 31 studies were assessed for eligibility by reading the full texts, of which eight studies could not extract the data that we needed. And there may be duplication of patient data in four studies: from the same consortium (The National Infantile Spasms Consortium, USA), hospitals (Nationwide Children's Hospital, The Ohio State University, Columbus), and research team. [21][22][23][24] In addition, all four studies 21-24 enrolled patients with new-onset IESS between 2 months and 2 years of age, and data were collected between 2012 and 2018. The same data set in more than one publication likely will bias the aggregated effects in a meta-analysis. 25 So we excluded three of the trials and included the latest study by Yuskaitis et al. 21 with the largest number of patients. Finally, 20 studies. 21,[26][27][28][29][30][31][32][33][34][35][36][37][38][39][40][41][42][43][44] appropriately covered the topics, met the eligibility criteria, and were included in this review.

| Study characteristics
The details of the 20 included articles are listed in Tables 1  and 2 Aicardi-1996) were rated as having a moderate risk of bias due to classification of interventions, missing data, or measurement of outcomes. The four studies 21,26,30,33 with a serious risk of bias assigned almost all the patients with TSC to the VGB treatment group. Five OSs 34,38,39,41,42 were rated as having a low risk of bias ( Figure S2).

| VGB for TSC vs other etiologies
Five OSs compared the efficacy of VGB in patients with TSC and patients with other etiologies. 38,39,41,42,44 The results revealed that for patients with IESS, VGB was more effective in patients with TSC than in those with other etiologies (I 2 = 0%, OR = 5.59, 95% CI = 2.17-14.41, P < 0.0001; Figure 2C).

| VGB combined with hormonal therapy vs hormonal monotherapy
Two RCTs 27,31 excluded patients with TSC and compared the efficacy of VGB combined with hormonal therapy vs hormonal monotherapy for IESS. The I 2 statistic reflected a high heterogeneity across studies (I 2 = 83%), and the random effect model was used. There was no significant difference in the efficiency of VGB combined with hormonal therapy vs hormonal monotherapy for IESS (OR = 0.75, 95% CI = 0.09-6.45, P = 0.79; Figure 2D).

| DISCUSSION
This review provides a comprehensive synthesis of existing articles, comparing (a) the efficacy of VGB vs hormonal monotherapy for IESS, (b) VGB in IESS patients with TSC vs patients with other etiologies, and (c) VGB combined with hormonal therapy vs hormonal monotherapy for IESS. All the relevant RCTs and OSs were included in our review and were analyzed separately. Gershon et al. 45 proposed that both RCTs and OSs exist with consistent findings representing the best evidence. In our study, comparing the efficacy of VGB vs hormonal monotherapy for IESS, as shown in Figure 2A,B, the meta-analysis results in RCTs and OSs showed consistent findings: VGB was less effective than hormonal therapy. ACTH, steroids, and VGB, as standard therapies, have been strongly recommended as the first-line treatment for IESS. Many previous meta-analyses indicated no significant difference in  [12][13][14] But, clinical studies comparing the efficacy of VGB with hormonal monotherapy for IESS have yielded inconsistent conclusions. And no high-level evidence was provided to prove the superiority or inferiority of VGB compared with hormonal therapy (ACTH and steroids) for patients with IESS. We performed a meta-analysis of these relevant clinical trials and concluded that hormonal therapy (ACTH and steroids) was more effective than VGB in patients with new-onset IESS. Because only one RCT by Chiron et al. 46 compared the efficacy of VGB vs hormonal therapy in patients with IESS due to TSC, we did not include this study for meta-analysis. Chiron et al. 46 concluded that VGB was more effective than hormonal therapy for patients with IESS due to TSC.
Our study was the first to synthesize trials to compare the efficacy of VGB in patients with TSC vs patients with other etiologies and showed that the efficacy of VGB in treating patients with TSC was significantly better than in patients with other etiologies. Almost all previous clinical studies suggested that VGB was more effective for patients with TSC. Our result further strengthened and corroborated it.
Three studies (one OS and two RCTs) compared the efficacy of VGB combined with hormonal therapy vs hormonal monotherapy for IESS. 27,29,31 One OS 29 was excluded from this meta-analysis. The study by O'Callaghan et al. 31 suggested that combination therapy was superior to hormonal monotherapy, while the results of Knupp et al. 27 suggested the opposite. Our meta-analysis showed no difference in efficacy, but the heterogeneity between the studies was high. Further studies need to corroborate whether there is a difference in efficacy.
In general, our review has yielded robust findings that support the following recommendations: hormonal monotherapy was better than VGB monotherapy for patients with new-onset non-TSC-associated IESS; for patients with IESS due to TSC, VGB was the first choice; the combination of VGB and hormone therapy may not always be more effective than hormone monotherapy. These results can improve the management of IESS.
Our review also has some limitations. The meta-analysis of combination therapy vs hormonal monotherapy showed heterogeneity (I 2 > 50%) and still needed to be validated by further clinical studies. Besides, there were differences in the dosage and duration of hormonal therapy between studies, which may lead to bias. Last, in the meta-analysis comparing the efficacy of VGB in patients with TSC, the trials had a small number of patients with TSC. F I G U R E 3 A, Forest plot of sensitivity analysis (RCTs). B, Forest plot of sensitivity analysis (OSs) AUTHOR CONTRIBUTIONS All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.