Section 3: Valproate/Divalproex: Clinical Niche, Evidence, and Safety Considerations

3.1 Where Valproate Fits

Valproate (encompassing valproic acid, sodium valproate, and the stable coordination compound divalproex sodium) is a first-line agent for acute mania with comparable efficacy to lithium in the acute setting (CANMAT/ISBD 2018; APA Practice Guideline; NICE CG185). Many patients have been well-stabilized on valproate for years, and for specific bipolar phenotypes (mixed features, rapid cycling), valproate remains the preferred mood stabilizer based on evidence of superior efficacy over lithium in these presentations. However, valproate has a weaker maintenance profile than lithium and a teratogenic risk profile that has reshaped its regulatory status, particularly for patients under 55.

Valproate's clinical value lies in two areas: rapid acute mania treatment (especially via oral loading strategies) and its superior efficacy in mixed features and rapid cycling, the two bipolar phenotypes where lithium performs least well. Understanding when valproate is the right choice, and when the risks make it the wrong one, is a critical prescribing skill.

3.2 How Valproate Works

Valproate operates through a pleiotropic set of mechanisms rather than a single receptor target (StatPearls, Valproic Acid; multiple pharmacology reviews). These mechanisms work on different timescales, which helps explain why valproate can achieve rapid antimanic effects (especially with loading doses) while also providing longer-term neuroprotective benefits.

Ion channel modulation. Valproate prolongs the inactive state of voltage-gated sodium channels in a use-dependent manner, preferentially targeting neurons firing at pathologically high frequencies while sparing normal activity. It also blocks T-type calcium channels, further stabilizing neuronal resting membrane potential.

GABA enhancement. Valproate inhibits GABA transaminase (GABA-T), the enzyme that degrades synaptic GABA, and also inhibits succinate semialdehyde dehydrogenase further downstream in the GABA degradation pathway. This dual blockade substantially increases GABA availability, enhancing inhibitory tone across cortical and limbic circuits and providing a pharmacological brake on the escalating excitation of mania.

HDAC inhibition: the epigenetic dimension. Valproate is a direct inhibitor of class I histone deacetylases (HDACs), particularly HDAC1 (multiple HDAC review sources). By inhibiting HDACs, valproate promotes histone acetylation, opening chromatin structure and activating genes involved in neural plasticity, neurogenesis, and BDNF expression. Because altering gene expression requires time, this mechanism likely explains the delayed onset of valproate's full prophylactic efficacy and its neuroprotective properties.

PKC pathway attenuation. Valproate inhibits brain fatty-acyl-CoA synthetase, reducing arachidonic acid turnover and downstream PKC signaling, which is frequently hyperactive during mania (Stahl and other psychopharmacology references).

3.3 Efficacy in Acute Mania

The foundational evidence comes from Bowden et al. (1994, JAMA), a multicenter, double-blind, placebo-controlled trial of 179 hospitalized patients randomized to divalproex, lithium, or placebo for 21 days (Bowden et al., 1994, JAMA). Response rates (50% or greater reduction in mania rating scores) were virtually identical between divalproex and lithium (48% and 49%, respectively, vs. 25% for placebo), establishing clear non-inferiority. The response rate difference yields a calculated effect size of approximately 1.0 (Cohen’s d equivalent), exceeding the 0.8 threshold for a “large” effect. The NNT for one additional responder over placebo was approximately 5 (calculated from the 48% vs. 25% response rate difference).

Note on funding: The Bowden 1994 trial was conducted by the Depakote Mania Study Group with manufacturer (Abbott Laboratories) support. This is important context, though the findings have been independently replicated in subsequent trials and meta-analyses.

Subsequent meta-analyses have consistently validated these findings. Yildiz et al. (2015, Psychological Medicine) confirmed that valproate is significantly superior to placebo for acute mania (RR of improvement: 2.05; 95% CI: 1.48 to 2.87) (Yildiz et al., 2015). A 12-week head-to-head trial by Bowden et al. (2010, International Clinical Psychopharmacology) demonstrated non-inferiority to lithium, with remission rates of 72.3% (valproate) versus 65.5% (lithium) and comparable tolerability (Bowden et al., 2010).

3.4 Maintenance Efficacy: Where Valproate Falls Short

Valproate's acute mania evidence is strong. Its maintenance evidence is notably weaker, and this distinction is clinically critical.

The BALANCE trial (Geddes et al., 2010, Lancet), a pragmatic, open-label, randomized trial of 330 patients followed for up to 24 months, found a clear hierarchy: both lithium monotherapy and lithium-plus-valproate combination were significantly more effective at preventing relapse than valproate monotherapy (Geddes et al., 2010, Lancet). The hazard ratio significantly favored lithium over valproate across the cohort. An important design consideration: BALANCE was open-label (not blinded) and did not include a separate placebo arm, so the valproate-versus-placebo comparison is indirect rather than directly tested within the trial. Nevertheless, the finding that valproate monotherapy did not clearly outperform the absence of a mood stabilizer in preventing relapse surprised many clinicians who had assumed comparable maintenance efficacy based on the strong acute data.

Broader meta-analytic consensus supports this position. Valproate monotherapy does show superiority over placebo for preventing mood episodes (RR 0.63; 95% CI: 0.47 to 0.83 in a network meta-analysis by Miura et al., 2014, Lancet Psychiatry; the Cipriani 2013 Cochrane review similarly found RR 0.68, 95% CI: 0.49 to 0.93), but in head-to-head comparisons, it consistently fails to match lithium's bidirectional preventive power (Cipriani et al., 2013, Cochrane Database of Systematic Reviews; Miura et al., 2014, Lancet Psychiatry). For maintenance, lithium remains the clear first choice, with valproate serving best as an adjunct to lithium or as the primary agent only when lithium is contraindicated or not tolerated.

3.5 The Clinical Niche: Mixed Features and Rapid Cycling

If lithium is the better maintenance agent, why does valproate remain so widely prescribed? Beyond tolerability preferences, valproate has genuinely distinct efficacy in two of the most challenging bipolar phenotypes.

Mixed features (dysphoric mania). Mixed states, where manic and depressive symptoms co-occur, represent the most dangerous and treatment-resistant form of bipolar illness. Decades of evidence show that mixed features strongly predict poor response to lithium. Swann et al. (2002, Neuropsychopharmacology) demonstrated that elevated depression scores during a manic episode were positively associated with favorable valproate response and negatively associated with lithium response (Swann et al., 2002, Neuropsychopharmacology). A systematic review of 32 studies (approximately 8,900 patients) confirmed valproate's superior efficacy over lithium specifically in mixed mania (multiple reviews). Valproate is universally recognized as the preferred mood stabilizer for dysphoric or mixed bipolar presentations.

Rapid cycling. Rapid cycling (four or more mood episodes per year) historically responds poorly to lithium prophylaxis. The original Bowden 1994 trial specifically found that divalproex was equally effective in rapid-cycling patients as in non-rapid cyclers, including patients who had previously failed lithium (Bowden et al., 1994, JAMA). While the long-term rapid cycling evidence is mixed, multiple analyses suggest valproate carries a lower relapse risk in rapid-cycling cohorts compared to lithium.

PsychHQ

Valproate/Divalproex: Niche, Efficacy & Safety