Article Text
PDF
Abstract
Objective This study aims to conduct an overview on the comparative efficacy of valproate in acute mania, bipolar depression and maintenance treatment of bipolar disorder (BD).
Method We performed an overview of systematic reviews with meta-analyses of randomised controlled trials (RCTs), registered in PROSPERO (CRD42024497749). We searched Medline and Cochrane Database of Systematic Reviews. Summary measures comparing valproate with placebo or other active drugs were described.
Results We included 26 systematic reviews. For acute mania (31 RCTs, n=4376), valproate showed a significantly better response than placebo in two high-quality systematic reviews (RR=1.42; 95% CI: 1.19 to 1.71) (OR=2.05; 95% CI: 1.32 to 3.20). No significant differences with lithium were found in most outcomes. Valproate had similar efficacy to quetiapine and lower efficacy compared with risperidone, with conflicting results when compared with olanzapine. In bipolar depression (7 RCTs, n=399), valproate was more effective than placebo in reducing depressive symptoms (OR=2.80; 95% CI: 1.26 to 6.18) and achieving remission (OR=2.4; 95% CI: 1.09 to 5.29) (OR=2.15; 95% CI: 0.82 to 5.6), considering the results of three high-quality systematic reviews. No significant difference was observed with lithium, lurasidone, quetiapine or olanzapine plus fluoxetine, but valproate showed superior efficacy to aripiprazole, ziprasidone and agomelatine. In maintenance treatment (11 RCTs, n=1063), valproate was superior to placebo in preventing relapse of any mood episode in two high-quality systematic reviews (RR=0.63; 95% CI: 0.48 to 0.83) (RR=0.63; 95% CI: 0.47 to 0.83). No significant difference was found with lithium, olanzapine or lamotrigine.
Conclusion This overview highlights favourable results for valproate compared with placebo in all phases of BD, as well as presenting specific results in comparison with other active drugs. However, these results must be interpreted considering the methodological limitations of our study.
- Bipolar and Related Disorders
- Systematic Review
- Drug Therapy
- Adult psychiatry
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
This study provides a comprehensive overview of systematic reviews with meta-analysis on the efficacy of valproate in treating acute mania, bipolar depression and maintenance therapy for bipolar disorder.
This study is the first overview of systematic reviews with meta-analyses specifically designed to evaluate the effect of valproate in the treatment of bipolar disorder, including open-label, single-blind and double-blind randomised controlled trials.
This overview assessed the quality of evidence from published systematic reviews and meta-analyses on the efficacy of valproate in bipolar disorder, categorising each review into levels of high, moderate, low and critically low confidence.
Due to the limited number of randomised controlled trials assessing the comparative efficacy of valproate in bipolar disorder, the results of this overview are dependent on a small number of primary studies.
This study aims to extract and analyse the evidence regarding the efficacy of valproate in bipolar disorder, without summarising or aggregating the results into new meta-analytic measures.
Introduction
Bipolar disorder (BD) is a psychiatric condition in which individuals experience severe mood changes, swings or fluctuations, and sometimes, psychotic features.1 2 The lifetime prevalence of bipolar spectrum disorders was estimated at 0.6% for bipolar type 1, 0.4% for bipolar type 2 and 2.4% for the entire spectrum, including subthreshold BD.3 Individuals with BD may exhibit persistent functional and cognitive impairments, leading to substantial decline in their overall quality of life and contributing to societal harms.4 5 The impact of BD is long-lasting and wide-ranging, affecting not only mental and physical health but also the social and occupational aspects of an individual’s life.6 7 The WHO highlights that BD ranks among the major mental health contributors to global disability, primarily due to its impact on a young and economically active population.8 9 Besides, the frequent delay in diagnosis and initiation of appropriate treatment contributes to the worsening course of the disease.10
The treatment of BD is primarily pharmacological and should be tailored to each phase of the disorder: mania/hypomania, depression or maintenance. Mood stabilisers such as lithium, valproate and lamotrigine, along with second-generation antipsychotics, are considered the drugs of choice by major guidelines.11 12 The treatment goal is to alleviate current symptoms and prevent new episodes. A common challenge is the low patient adherence to treatment, especially, during manic episodes when insight into the illness is compromised. Some studies attempt to compare the effects of various drugs in BD management, and their findings indicate that mood stabilisers, particularly, valproate and lithium, play a crucial role in treatment of BD.13
In the 1960s, the use of valproate emerged as a promising treatment for BD. Researchers in Europe and the USA studied the mood-stabilising effects of valproate for three decades, initially using valproic acid and its derivatives, such as valpromide, and later, in the 1980s, transitioning to sodium divalproate.14 15 The Canadian Network for Mood and Anxiety Treatments and the International Society for Bipolar Disorders have established in their guidelines that valproate is a first-line treatment for acute mania and maintenance treatment, and a second-line treatment for bipolar depression.16 Despite a vast literature demonstrating the efficacy of valproate in treating BD,17–20 numerous studies show conflicting results. Previous evidence suggested that valproate might be less effective than lithium in the treatment of acute mania.21 22 However, recent network meta-analyses (NMAs) have not confirmed this finding23 nor have they corroborated the inferiority of valproate in bipolar depression19 and the maintenance treatment.24 Furthermore, comparisons regarding the efficacy of valproate vs antipsychotics also show conflicting results.21 23 25
To address the conflicting evidence regarding the efficacy of valproate in treating BD, both in comparison to placebo and active drugs, this study aims to conduct an overview that synthesises and discusses the primary findings of systematic reviews with meta-analyses on the comparative efficacy of valproate in treating acute mania, bipolar depression and maintenance therapy for BD.
Methods
This study was conducted following the Cochrane Method for Overviews of Reviews26 and we adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines.27 We defined a systematic review, based on Cochrane’s definition, as a comprehensive analysis of literature involving identifying, assessing and synthesising all relevant empirical evidence that meets predetermined eligibility criteria to address a specific research question. By employing explicit and systematic methods aimed at reducing bias, the systematic review aims to produce reliable findings that can inform decision-making processes.28
The search was systematically conducted in Medline (PubMed) and the Cochrane Database of Systematic Reviews (CDSR) from their inception through 15 January 2024, considering the following general terms: ‘Bipolar Disorder,’ ‘Valproic Acid’, ‘Valproate’, ‘Divalproex’ and ‘Systematic Review’. We used a search term specific to the systematic review study design, aligning with Cochrane’s recommendations for overviews.26 Furthermore, the general terms were expanded through synonym searches. There were no restrictions on publication date or language. We also conducted a manual examination of the reference lists of identified articles to ensure the inclusion of other reviews that might be eligible. The complete search strategy and the PRISMA 2020 Checklist are available in the online supplemental materials I and II.
Supplemental material
We included systematic reviews of open-label, single-blinded or double-blinded randomised controlled trials (RCTs) evaluating the efficacy of valproate in treating BD and comparing it to placebo or other active drugs. We only considered systematic reviews with meta-analysis, whether pairwise or NMAs. We excluded non-systematic literature reviews, reviews of materials that do not constitute original research, such as systematic reviews of guidelines, and reviews solely focused on the economic evaluation of medications. We also excluded (1) reviews lacking diagnostic confirmation of BD through the Diagnostic and Statistical Manual of Mental Disorders or International Classification of Diseases criteria; (2) reviews that exclusively focused on specific populations (children, adolescents, pregnant women or postpartum women) or outcomes; (3) reviews that encompassed populations with BD and psychiatric comorbidities, such as Substance Use Disorders, Anxiety Disorders, Obsessive-Compulsive Disorder, Neurocognitive Disorders and Suicide; (4) reviews that exclusively focused on adverse effects of valproate; (5) reviews that compared valproate with electroconvulsive therapy or psychotherapeutic interventions; (6) reviews on treatment guidelines for BD; (7) reviews that did not specifically evaluate the efficacy of valproate in BD; (8) systematic reviews that included preclinical trials; and (9) no direct comparisons between valproate and other drugs at the metanalytical estimates.
The studies’ selection was carried out in two stages: screening phase (title and abstracts reading) and eligibility phase (full article reading). Two investigators (JJM and DPB) executed the stages independently; in case of inconsistencies, a discussion involving a third analyst (NRS) was conducted to reach a consensus. After completing the study selection, two authors (JJM and DPB) independently extracted data from the systematic reviews and any discrepancies were carefully reviewed by a third author (NRS). We extracted the first author’s name, year of publication, identification of the RCTs included in the review, number of individuals included in each head-to-head comparison, main findings, and, if available, type of meta-analysis (pairwise or network meta-analysis) and meta-analytical estimates (effect sizes and measures of dispersion intervals). The OR, relative risk (RR), mean difference (MD) and standardised mean difference (SMD) were considered as effect size measures. As for measures of dispersion, we considered 95% CIs or 95% credibility intervals (95% CrI). Additionally, the efficacy definitions from the included systematic reviews were analysed. The term ‘valproate’ was employed in a generic sense, encompassing any form of valproate including valproic acid or divalproex.
We used the AMSTAR-2 (a measurement tool to assess systematic reviews-version 2) to assess the methodological quality of the systematic reviews included in this overview. AMSTAR-2 is a practical critical appraisal tool specifically designed for rapid assessments of the quality of conduct of systematic reviews of RCTs involving interventions.29 It categorises the certainty of the evidence as high, moderate, low or critically low. One investigator (DPB) applied this tool, and the results were discussed with two additional investigators (JJM and NRS). We emphasise that AMSTAR-2 is a fundamental tool in conducting an overview of reviews; however, it does not assess the risk of bias of the included RCTs nor the confidence in the evidence of the included reviews. We emphasised that in our overview, all identified studies were presented and discussed in the article, regardless of each study’s level of evidence. However, our final conclusions were primarily based on the findings from high-quality studies.
Additionally, we evaluated the overlap of primary studies included in each drug comparison by calculating the corrected covered areas (CCA).30 CCA values exceeding 11% indicate a high overlap of RCTs among the systematic reviews included in the respective comparison. This study was registered with PROSPERO (International prospective register of systematic reviews),31 number CRD42024497749.
Patient and public involvement
This study is based exclusively on existing literature and does not involve direct patient engagement. None of the authors or contributors to the manuscript are individuals with BD. The dissemination of findings will focus on engaging the clinical community, including presentations at internationally recognised scientific conferences to connect with clinicians involved in BD treatment. This will allow for discussion of the findings and their potential clinical implications. Additionally, the results will be shared with advocacy groups both in Brazil and internationally. Any author of this paper may participate in events aimed at educating patients and caregivers. Social media will be used to communicate the findings using visually appealing and concise language. The Brazilian Practical Psychopharmacology platform, which includes over 5000 psychiatrists, will be leveraged to share these findings and contribute to improving bipolar disorder treatment.
Results
The study selection process is shown in figure 1. Initially, we screened 188 records in total including 182 from the search strategy on the databases and six additional systematic reviews that were identified via citation searching: one through backward search and five through forward search strategies.17 19 25 32–34 From this total screening, it was selected 39 systematic reviews for full-text reading. Finally, we retained 26 articles for inclusion, encompassing 50 RCTs, with publication years ranging from 1996 to 2023. The list of included and excluded articles during full-text screening, along with the corresponding reasons for exclusion can be found in online supplemental material III. We did not find any systematic review with specific meta-analysis on the effect of valproate on mixed state or rapid cycling that met our eligibility criteria.
PRISMA 2020 flow diagram for articles included in the overview.
The complete results of our overview are described at the online supplemental table S1 to S6, and the list of included RCTs is presented in online supplemental material IV. To summarise the key findings and present the results concisely, table 1 (acute mania treatment), table 2 (bipolar depression treatment) and table 3 (maintenance treatment in BD) exclusively present comparisons between valproate and other FDA-approved medications for BD treatment, as well as with placebo. Additionally, since combinations of valproate with other active drug do not allow us to accurately isolate the effect of valproate on treatment efficacy in BD, we have decided not to discuss such results in the manuscript; however, we have retained them in the online supplemental table S1 to S6. In the summarised tables, we also excluded indirect evidence from NMAs to ensure a more concise presentation of the data. The assessment of methodological quality of the included reviews using the AMSTAR-2 tool, and the estimation of overlap of RCTs in each drug comparison, are both provided in online supplemental material V and VI, respectively.
Characteristics of included meta-analyses on the efficacy of valproate in the treatment of acute mania in bipolar disorder
Characteristics of included meta-analyses on the efficacy of valproate in the treatment of bipolar depression
Characteristics of included meta-analyses on the efficacy of valproate in maintenance treatment of bipolar disorder
Among the included systematic reviews, 11 assessed the efficacy of valproate in managing acute mania, while seven evaluated its effectiveness in bipolar depression and nine in maintenance treatment of BD. The average duration of the included RCTs in acute mania, bipolar depression and maintenance treatment was 7.51, 9.25, and 66.22 weeks, respectively.
Valproate in acute mania
11 reviews included in our analysis evaluated the comparative efficacy of valproate in treating acute mania, involving a total of 31 RCTs and 4376 participants. These reviews encompassed direct comparisons of valproate monotherapy with placebo, lithium, carbamazepine, olanzapine, quetiapine and risperidone (table 1). Additionally, they featured direct comparisons of valproate vs oxcarbazepine, topiramate, haloperidol and endoxifen (online supplemental table S1). We also identified statistically significant indirect evidence from NMAs comparing valproate with other active drugs, such as paliperidone and cariprazine (online supplemental table S2). Considering FDA-approved drugs for the treatment of acute mania, apart from some comparisons mentioned above, direct evidence comparing valproate to aripiprazole, asenapine and ziprasidone was not found. However, indirect evidence from NMAs did not show a significant difference between valproate and aripiprazole (RR=0.93, 95% CrI: 0.74, 1.16),35 (MD=1.50, 95% CrI: −0.43, 3.34),23 asenapine (RR=1.12, 95% CrI: 0.86, 1.44),35 (MD=0.82, 95% CrI: −1.38, 2.96)23 and ziprasidone (RR=1.05, 95% CrI: 0.78, 1.42),35 (MD=2.55, 95% CrI: −0.33, 5.37).23
Eight studies compared valproate vs placebo,17 21 23 25 34–37 with only two of them considered to be of high-quality systematic review according to the AMSTAR-2.21 35 These studies comprised the findings of seven trials,38–44 exhibiting a very high overlap of RCTs (CCA=42.8%). For treating acute mania, valproate has been shown to be more effective than placebo across all efficacy outcomes, as detailed in table 1. The exception occurred with data from Kishi and colleagues (2021), where valproate did not demonstrate statistical superiority compared with placebo in clinical remission (RR=1.25, 95% CrI: 0.98–1.59).35
Seven meta-analyses compared valproate to lithium,21–23 25 35–37 with three regarded as high AMSTAR-2 standard.25 35 37 These studies comprised the findings of seven RCTs,39 45–50 showing a very high overlap of RCTs (CCA=26.2%). No significant differences were found between valproate and lithium in most efficacy outcomes, as detailed in table 1, except for two reviews that showed valproate to be less effective than lithium in achieving clinical remission (OR 1.75, 95% CI: 1.04 to 2.94)22; (OR 0.69, 95% CI: 0.14 to 1.25).21
Four systematic reviews with meta-analyses compared valproate with carbamazepine in acute mania21 25 36 37 considering data from two clinical trials,46 51 thus, exhibiting a very high overlap of primary studies (CCA=50%). Two recent NMAs showed that valproate was less effective than carbamazepine in reducing mania symptoms (MD=3.23, 95% CrI: 0.51–5.93),23 (SMD=0.36, 95% CrI: 0.06–0.67),35 while a third NMA and two pairwise meta-analyses found no significant difference between these drugs (SMD=−0.12, 95% CrI: −0.42, +0.22),25 (OR=2.41, 95% CI: 0.52 to 11.1),21 (RR=0.66, 95% CI: 0.38 to 1.16)37 (table 1). Not surprisingly, when oxcarbazepine, a structural derivative of carbamazepine, was compared with valproate in the treatment of acute mania, no significant difference in effectiveness was found (MD=−0.12, 95% CrI: −4.40, 4.12),23 (MD = +0.02, 95% CrI: −0.60, 0.65),25 (OR=2.27, 95% CI: 0.51, 10.0),51 (OR=2.25, 95% CI: 0.51 to 9.99)21 (online supplemental table S1). When compared with topiramate, two NMAs demonstrated that valproate was more effective in improving mania symptoms (MD=−3.78, 95% CrI: −6.28 to –1.36),23 (SMD=−0.31, 95% CrI: −0.55 to –0.08),35 and reducing the treatment discontinuation due to inefficacy (RR=0.51, 95% CrI: 0.30, 0.79)35 (online supplemental table S1).
Six reviews compared valproate to olanzapine in the management of acute mania,21 23 25 35–37 using data from four RCTs,41 52–54 revealing a very high overlap of primary studies (CCA=70%). Three reviews found no significant difference between these drugs in reducing mania symptoms (RR=0.90, 95% CrI: 0.75, 1.07),35 (SMD=−0.16, 95% CrI: −0.34, 0.04),25 (OR=0.77, 95% CI: 0.48, 1.25)21; however, three other reviews showed the inferiority of valproate compared with olanzapine (MD=2.21, 95% CrI: 0.54, 3.84),23 (MD=2.81, 95% CI: 0.83, 4.79),37 (RR=0.76, 95% CI: 0.58, 0.99).36 Furthermore, three meta-analyses found no difference in the efficacy of these drugs regarding the risk of treatment discontinuation due to inefficacy (RR=1.24, 95% CrI: 0.91, 1.69),35 (RR=1.12, 95% CI: 0.66, 1.89),36 (RR=1.08, 95% CrI: 0.50, 2.36),37 and one study showed that valproate was less effective in changing symptom severity after 3 weeks of treatment (SMD=0.25, 95% CI: 0.11, 0.39).21
Four pairwise systematic reviews21 36 37 55 and three NMAs23 25 35 compared the effect of valproate vs haloperidol in acute mania; however, all these studies were based on a single RCT with 36 participants56 (table 1). All pairwise systematic reviews yielded no differences between valproate and haloperidol in terms of efficacy in acute mania (OR=1.82, 95% CI: 0.46, 7.18),21 (RR=1.43, 95% CI: 0.61, 3.32),36 (RR=0.79, 95% CI: 0.46, 1.35).37 55 However, the results of the NMAs were conflicting, which could be explained by the inclusion of different indirect evidence in their outcomes (Table S2). Kishi et al24 and Hong et al23 demonstrated that valproate was less effective than haloperidol in reducing manic symptoms (SMD=0.39, 95% CrI: 0.19, 0.59)35 (MD=2.52, 95% CrI: 0.47, 4.48),23 and led to higher rates of treatment discontinuation due to inefficacy (RR=1.59, 95% CrI: 1.12, 2.28).35 However, Yildiz et al (2015) found no significant difference between these drugs (SMD=−0.22, 95% CrI: −0.43, 0.02)25 (table 1).
Three NMAs showed that valproate was less effective than risperidone in managing manic symptoms with direct comparisons (MD=3.79, 95% CrI: 1.66, 5.93)23 (table 1), and indirect comparisons (SMD=0.38, 95% CrI: 0.16, 0.60)35 (online supplemental table S1), (SMD=−0.33, 95% CrI: −0.58 to –0.66)25 (online supplemental table S1). One NMA showed non-significant difference between valproate and quetiapine (MD=1.45, 95% CrI: −0.55, 3.40)23 (table 1). Online supplemental table S1 provides a comparison of the efficacy of valproate against other less used medications in addressing BD.
Valproate in bipolar depression
Seven reviews included in our analysis assessed the comparative efficacy of valproate in bipolar depression, involving 7 RCTs and 399 participants. These reviews encompassed direct comparisons only of valproate monotherapy with placebo (table 2). We also identified statistically significant indirect evidence from NMAs comparing valproate and aripiprazole, ziprasidone, agomelatine, levetiracetam, valproate plus memantine, valproate plus cytidine, valproate plus olanzapine and sertraline plus lithium (online supplemental table S4). No FDA-approved drug for treating bipolar depression was directly compared with valproate, and no indirect evidence from NMAs showed significant differences between valproate and lurasidone (OR=1.58, 95% CrI: 0.68, 3.47),19 quetiapine (OR=1.54, 95% CrI: 0.68, 3.47)19 or olanzapine plus fluoxetine (OR=1.13, 95% CrI: 0.47, 2.70)19 (online supplemental table S4).
Six studies compared valproate vs placebo,17–19 57–59 of which two were considered high-quality systematic reviews according to AMSTAR-2.18 19 The results of all these reviews were limited to six RCTs, exhibiting a very high overlap of RCTs demonstrating high overlap in their primary studies (CCA=56.67%). Both high-quality systematic reviews demonstrate the superiority of valproate over placebo in reducing depressive symptoms18 19; however, their results were conflicting regarding the reduction in the number of participants who achieved clinical remission (OR=2.15, 95% Crl: 0.82, 5.61),18 (RR=2.15,% 0.60, 7.74)59 (table 2). No direct comparisons between valproate and lithium were found, and only one NMA provided indirect evidence comparing these drugs, showing a non-significant difference between them (OR=2.41, 95% CrI: 0.99, 5.82)19 (online supplemental table S4).
A pairwise meta-analysis demonstrated that the combination of valproate and olanzapine was more effective than valproate monotherapy (SMD=−0.31, 95% CrI: −0.58 to –0.03),60 although the quality of the review scored low on the AMSTAR-2 (online supplemental table S3). Conversely, an NMA showed, through indirect evidence, that valproate was superior to aripiprazole (OR=2.73, 95% CrI: 1.15, 6.46), ziprasidone (OR=2.88, 95% CrI: 1.25, 6.65) and agomelatine (OR=2.80, 95% CrI: 1.10, 7.10) in reducing depressive symptoms19 (online supplemental table S3). The same study also revealed that the comparison between valproate monotherapy and valproate plus memantine indicated the superiority of add-on memantine over valproate monotherapy in reducing acute depressive symptoms (OR=3.46, 95% CrI: 1.29, 9.28),19 although these findings were based on a single clinical trial.61 Additionally, this review suggested that the combination of valproate plus cytidine did not demonstrate superiority over valproate alone in the management of bipolar depression (OR=3.19, 95% CrI: 0.66, 15.42)19 (online supplemental table S3). Online supplemental tables S3 and S4 also provide additional comparisons of the efficacy of valproate against other drugs in addressing bipolar depression.
Valproate in maintenance treatment of BD
Nine included reviews assessed the comparative efficacy of valproate in BD maintenance treatment, comprising a total of 11 RCTs and 1063 participants. These reviews encompassed direct comparisons of valproate monotherapy with placebo, lithium, olanzapine, valproate plus lithium and valproate plus aripiprazole (table 3). We also found statistically significant indirect evidence from NMAs comparing valproate vs asenapine; valproate vs paliperidone, risperidone, placebo and lamotrigine (online supplemental table S6). Considering FDA-approved drugs for the maintenance treatment of BD, apart from some comparisons mentioned above, we did not find direct comparisons between valproate and lamotrigine in monotherapy. However, indirect evidence from NMAs did not demonstrate a significant difference between these drugs in reducing relapse of any mood episode (RR=0.83, 95% CrI: 0.60, 1.14).24
Eight studies reporting meta-analysis compared valproate vs placebo in maintenance treatment.20 24 62–67 These reviews comprised the findings of five RCTs,68–72 with high level of primary studies overlap (CCA=22.5%). Three studies were considered high-quality systematic reviews according to AMSTAR-2.20 24 62 As detailed in table 3, two high-quality systematic reviews demonstrated the superiority of valproate over placebo in preventing relapse for any mood episode24 62; however, no differences were found in the rates of relapse of depressive symptoms, and the results regarding the relapse of manic episodes were conflicting. Moreover, another high-quality systematic review73 also demonstrated that valproate was more effective than placebo in reducing the risk of treatment discontinuation due to any mood episode or depressive episodes (table 3). Other five pairwise meta-analyses also found superior efficacy of valproate compared with placebo in reducing the relapse of any mood episode;63–67 whereas, only two demonstrated better results to valproate in reducing the relapse rate of depressive episodes.65 66 No meta-analyses found a significant difference between valproate and placebo in preventing manic episodes.
Six meta-analyses compared valproate with lithium,20 24 62 65–67 pooling data from four RCTs,68 74–76 with a very high level of overlap of primary studies (CCA=60%). Three studies were considered high-quality systematic reviews according to the AMSTAR-2,20 24 62 as detailed in table 3. No included reviews found a significant difference between valproate and lithium in the maintenance treatment of BD, considering all efficacy outcomes (rates of relapse of any mood episode, depressive episodes, or manic, hypomanic, or mixed episodes).
Three pairwise meta-analyses,65 66 73 all based on data from a single RCT,77 compared the efficacy of valproate vs olanzapine, revealing similar efficacy for both treatments in maintenance treatment for BD (OR=1.02, 95% CI: 0.32, 3.23),65 (RR=1.33, 95% CI: 0.78, 2.27),73 (OR=1.02, 95% CI: 0.32, 3.23)66 (table 3). A recent NMA demonstrated, through indirect evidence, that valproate was less effective than asenapine in reducing the relapse rates of any mood episodes (RR=2.42, 95% CrI: 1.16, 5.02)24 (online supplemental table S6). Three high-quality systematic reviews reported no significant differences between valproate vs valproate plus lithium (RR=1.21, 95% CrI: 0.86, 1.69),24 (RR=1.20, 95% CrI: 0.84, 1.72),62 (RR=1.28, 95% CI: 1.04, 1.59),20 and three meta-analyses reported no significant differences between valproate vs valproate plus aripiprazole (RR=2.17, 95% CrI: 0.88, 5.35),24 (RR=2.17, 95% CrI: 0.86, 5.55),62 (OR=0.37, 95% CI: 0.12, 1.07)64 (online supplemental table S5). Online supplemental table S5 and S6 provides direct and indirect comparison of the efficacy of valproate against other drugs in addressing maintenance treatment to BD.
Discussion
The present overview of systematic reviews with meta-analyses further corroborates that valproate is more effective than a placebo in all phases of BD. Our results align the literature supporting the use of valproate in treating acute mania. About bipolar depression, the present overview shows that valproate is superior to placebo in most efficacy outcomes. Furthermore, valproate demonstrates comparable efficacy to lithium in both the treatment of acute mania and maintenance treatment. Of note, the results for bipolar depression are limited to indirect evidence derived from NMAs and not from direct comparisons.
Valproate in acute mania
Multiple studies consistently position valproate as more effective than placebo across a range of efficacy measures.17 21 23–25 34 36 37 However, on meticulous examination, we found a notable exception where valproate did not exhibit statistical differences compared with placebo, particularly regarding one measure of clinical remission in the study conducted by Kishi and colleagues, 2021.24 This exception underscores the importance of considering individual study nuances and the potential impact of high-quality methodologies.
A comparative analysis between valproate with lithium, a widely acknowledged mood stabiliser, suggests no statistical differences between the two drugs across multiple outcomes.21–25 36 37 The consistency in these findings indicates a comparable efficacy profile between valproate and lithium, highlighting the importance of individualised treatment plans tailored to patient-specific factors.
In NMAs, valproate demonstrated poorer results compared with carbamazepine in improving manic symptoms23 25; however, pairwise meta-analyses indicated similar efficacy between these two drugs.21 36 37 Additionally, valproate exhibited comparability to oxcarbazepine in improving manic symptoms,23 25 clinical response,51 remission rate and change in symptom severity.21 The absence of significant differences between oxcarbazepine and valproate in various outcomes raises questions about potential class effects within antiepileptic medications and their applicability in the context of acute mania. That is not applicable to topiramate that did not show antimanic effects and was inferior to valproate in all efficacy outcomes in acute mania.
The evaluation of olanzapine, a second-generation antipsychotic, in comparison with valproate yields intriguing results. No significant differences were observed in terms of remission rate21 35 and withdrawal due to a failure to respond.36 37 However, olanzapine outperformed valproate in terms of change in symptom severity,21 clinical response37 and improvement of manic symptoms.35 Risperidone showed superiority to valproate in reducing manic symptoms in a single NMA.23 Valproate demonstrated similarities to haloperidol21 36 37 55 and quetiapine,23 indicating comparable efficacy of valproate with the use of these two antipsychotics in acute mania. In addition to these comparisons, recent literature reviews have recommended valproate, along with other antipsychotics, for treating acute mania.12 These include asenapine, cariprazine, risperidone, aripiprazole, ziprasidone11 78 79 and paliperidone.11 78
Valproate in bipolar depression
The assessment of systematic reviews and meta-analyses examining the utilisation of valproate in bipolar depression suggests its superiority over placebo in reducing depressive symptoms, as evidenced by two high-quality systematic reviews.18 19 However, conflicting findings emerge regarding the reduction in the number of participants achieving clinical remission, with discrepant results reported by two studies.59 Furthermore, the use of an adjunctive antipsychotic, such as olanzapine, was found to be more effective than valproate monotherapy in symptom reduction. Additionally, the combination of memantine and valproate demonstrated superiority over valproate monotherapy in the acute treatment of bipolar depression in a small non-replicated trial.61 Importantly, despite various meta-analyses, all estimates for valproate in bipolar depression were derived from six small clinical trials.61 80–84 The meta-analysis with the highest sample size, conducted by Yildz et al included 206 subjects. This reliance on relatively small-sample clinical trials undermines the generalisability of systematic reviews in demonstrating the superiority of valproate in bipolar depression. For robust evidence based on systematic reviews, the combined number of expected samples should be around one thousand patients, as suggested by Trikalinos et al.85 85 It is widely acknowledged that large randomised controlled trials have become increasingly prevalent in some medical fields, though this trend is less pronounced in mental health.85
It is noteworthy that the first-line efficacious treatments for acute episodes of bipolar depression primarily involve antipsychotics such as cariprazine and quetiapine that were shown to be first-line treatments in recent clinical reviews.12 19 78 79 Furthermore, adjunctive treatments involving valproate and antipsychotics can also be considered, particularly quetiapine,12 59 olanzapine60 and lurasidone.86
It is widely recognised that individuals with BD frequently experience depressive symptoms.11 However, there has been a lack of emphasis on clinical trials specifically addressing bipolar depression. The use of antidepressants remains a subject of controversy and it is advised against as a standalone treatment.11 12 For bipolar disorder type 2, the use of antidepressants should always be in conjunction with mood stabilisers or antipsychotics due to the heightened risk of a manic switch.11 Moreover, we emphasise that considering indirect comparisons from NMAs our study reveals that valproate was more effective than agomelatine and the combination of sertraline with lithium.19
Valproate in maintenance treatment of BD
The investigation into valproate’s efficacy in preventing relapses among individuals diagnosed with BD has greatly improved our understanding of its therapeutic potential. The included systematic reviews have consistently demonstrated valproate’s superiority over placebo.24 62–67 However, in certain comparisons, valproate was found to be comparable to placebo, particularly concerning the relapse of depressive episodes.24 62
The comparative analysis of valproate with other mood stabilisers, particularly lithium, introduces a layer of complexity to our understanding. Lithium and valproate exhibited comparability across a range of clinical outcomes. Importantly, no significant differences were identified in the recurrence or relapse of depressive episodes, or withdrawals for any acute episode.20 24 62 65–67 In the realm of antipsychotic comparisons, no significant differences were observed between valproate and olanzapine across all acute episodes—be they depressive, manic or mixed episodes.65 66 73 The addition of valproate to aripiprazole24 62 64 and lamotrigine did not demonstrate superiority over valproate monotherapy.24 64 Antipsychotics like olanzapine, risperidone, asenapine, aripiprazole and quetiapine play a crucial role in managing the maintenance treatment of BD.11 78 79
Limitations
This overview presents potential limitations. First, this overview limited its search to Medline and CDSR, even though other databases are available and could potentially yield additional results. However, PubMed and CDSR are widely recognised for their comprehensive coverage and relevance to the topics we examined, and they typically cover the most high-quality systematic reviews. Additionally, there is evidence that using data sources beyond PubMed has only a modest impact on the results of systematic reviews of therapeutic interventions.87 Also, the search strategy specifically targeted systematic reviews, which could potentially overlook articles. However, we followed the Cochrane recommendations for the overviews, which guide the use of specific terms related to the systematic review study design.26
Second, most of the included systematic reviews demonstrated an overlap of primary studies, potentially introducing bias by overemphasising the significance of certain RCTs. To mitigate this bias, the present overview assessed the magnitude of primary studies overlapping. Furthermore, we adhered to a standardised procedure, elucidating each step, and adopted clear inclusion and exclusion criteria for the analysed studies. To address potential omissions, articles that were not identified in the initial literature search were then cross-referenced with the reference lists of included studies. Contacting authors became necessary for specific studies to clarify uncertainties about the trials incorporated into their results. Notably, the systematic reviews that focused on maintenance and bipolar depression were conducted with a sample size of fewer than 1000 subjects, which falls short of the ideal number for a robust meta-analysis.85
Conclusion
The present overview of systematic reviews with meta-analyses showed that valproate stands out as a valuable agent in the treatment of mania, bipolar depression and maintenance treatment of BD, demonstrating efficacy in preventing relapses and stabilising mood. Future research endeavours should concentrate on refining our understanding of the comparative efficacy of valproate, particularly combinations with antipsychotics within the evolving landscape of treating BD.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Ethics approval
Not applicable.
Acknowledgments
DPD has a postdoctoral grant from the Instituto Nacional Saúde Cerebral (INSC, No 406020/2022-1), National Research Council. JJM is a senior researcher of the National Research Council (CNPq).
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors JJM is the guarantor author. JJM, FK and DPB designed the study and determined the methodology. DPB and NRS wrote the search strategy and participated in data collection. JJM, DPB and NRS applied statistical methods and extracted data from the meta-analysis. JJM, LHJD, DPB, MH, NRS and FK all contributed to writing the manuscript. JJM, LHJD, MH and FK had a major role to select existing literature and incorporate it into the manuscript. All authors reviewed and approved the final version of the manuscript before submission.
Funding The authors received a grant from Abbott Laboratories to conduct this study (grant number: RRTI-8293).
Competing interests This article was funded by a grant from the Abbott Laboratories in the provision of study material and on review process, and article processing charges, there was no involvement of medical writing. The paper was prepared by the investigators with independency and autonomy. Grant number: RRTI-8293. JMM has recieved honoraria for lectures for Apsen, Janssen, Abbot, EMS, and Biolab. LHJD has received fees as a consultant from Apsen, Abbot, AbbVie, Biolab, Cristália, Daichii-Sankyo, EMS, Gntech, Inspirali, LIBBS, Lundbeck, Mantecorp, Moksha 8, Pfizer, Sanofi, Servier, Viatris, FQM, Takeda-Shire, Torrent, Artmed, Biopas, Genon, Besins, Greencare, TEVA, Aché. MH has recieved fees as consultant from Apsen, Abbot, Biolab, Cristália, Daichii-Sankyo, EMS, Gntech, Johnson&Johnson, LIBBS, Lundbeck, Mantecorp, Moksha 8, Pfizer, Sanofi, Servier, Viatris, Takeda-Shire, Torrent, Biopas. FK has received personal fees as a speaker/consultant from Janssen (Johnson & Johnson), Daiichi-Sankyo, Libbs, Abbott and Teva Pharmaceutical Industries. DPB and NRS declare no conflict of interest.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.