Mood Stabilizers

A high-level examination of lithium, the anticonvulsants, and the antipsychotics used in bipolar disorder — their receptors, likely mechanisms, real-world value, and problems

The Definitional Problem

"Mood stabilizer" is the most conceptually unstable label in psychiatric pharmacology. Unlike "SSRI," which names a mechanism, the term names a hoped-for clinical profile with no agreed definition. The most demanding version requires that a true mood stabilizer treat acute mania, treat acute bipolar depression, prevent recurrence of both poles, and do all this without destabilizing the patient toward the opposite pole — the "above, below, and from above and below" criterion. By that standard, only lithium fully qualifies, and even lithium is weaker for acute bipolar depression than for the other tasks. Most drugs called mood stabilizers are strong at one or two of these jobs and weak or absent at others: valproate is anti-manic but a poor antidepressant; lamotrigine prevents depressive recurrence but is nearly useless for acute mania; the antipsychotics span the range unevenly.

This matters clinically because bipolar disorder is not one treatment target but several distinct ones — acute mania, acute bipolar depression (the hardest), and long-term maintenance — and a drug excellent for one may be useless or harmful for another. The single most important fact in the entire field is the asymmetry of bipolar disorder: patients spend far more of their ill time depressed than manic, yet most "mood stabilizers" are far better at treating and preventing mania than depression, and the one intervention that reliably treats unipolar depression — antidepressants — is at best marginally helpful and at worst destabilizing in bipolar illness. Bipolar depression is the great unsolved problem of mood-disorder pharmacology, and the inadequacy of the "mood stabilizer" toolkit against it is this document's recurring theme.

This overview covers the three families that do the work: lithium (the prototype and still, by important measures, the best); the anticonvulsant mood stabilizers (valproate, lamotrigine, carbamazepine); and the atypical antipsychotics, which have quietly become the workhorses of modern bipolar treatment despite the metabolic price. The honest framing: these are genuinely effective, often life-saving drugs for a severe and recurrent illness — and a toolkit that is unevenly effective across the illness's phases, burdened with serious long-term toxicities, increasingly stretched (with weak evidence) to treat emotional dysregulation that is not bipolar at all, and anchored by a prototype drug (lithium) that is underused precisely because it is old, unprofitable, and demanding to monitor.

Lithium: The Prototype and the Enigma

What it is and why it still matters

Lithium — a simple monovalent cation, an element, not a designed molecule — was the first effective psychiatric drug of the modern era (Cade, 1949) and remains, three-quarters of a century later, the gold standard against which every other mood stabilizer is measured, and on several dimensions still the best. This is remarkable: a cheap, unpatentable salt with a mechanism we still do not fully understand outperforms or matches every rationally designed successor on the things that matter most. Its persistence despite commercial neglect is one of psychiatry's most telling facts.

What it does

Lithium is the closest thing to a complete mood stabilizer:

• Acute mania: effective, though slower than antipsychotics (which is why mania is often treated with both initially).
• Maintenance / relapse prevention: the strongest long-term prophylactic in bipolar disorder, reducing both manic and depressive recurrences — genuinely "from above and below."
• Bipolar depression: modestly effective (weaker than for mania, but real).
• The antisuicide effect: lithium uniquely and robustly reduces suicide — not just suicidal ideation but completed suicide and all-cause mortality — across multiple studies and meta-analyses, an effect that appears at least partly independent of its mood-stabilizing action and is unmatched by any other psychiatric drug. This alone justifies lithium's place and makes its underuse a genuine public-health failure. (The effect is strong enough that ecological studies have explored whether trace lithium in drinking water correlates with lower population suicide rates — suggestive, not conclusive, but striking.)
• Neuroprotection: lithium is associated with preserved or increased gray matter volume, and observational signals suggest possible protective effects against dementia — consistent with its molecular actions below.

Mechanism: still genuinely mysterious

Lithium's mechanism remains incompletely understood — a humbling fact for a drug this old and important. The leading candidates, none individually sufficient:

• Inositol depletion / IMPase inhibition. Lithium inhibits inositol monophosphatase, depleting myo-inositol and dampening the phosphatidylinositol second-messenger signaling cascade — the basis of the influential "inositol depletion hypothesis." This connects to overactive intracellular signaling in mania.
• GSK-3β inhibition. Lithium inhibits glycogen synthase kinase-3 beta, a master regulator of numerous intracellular pathways governing circadian rhythm, neuroplasticity, apoptosis, and gene expression. GSK-3β inhibition is currently the most theoretically central candidate, linking lithium to neuroprotection, plasticity, and circadian regulation in one mechanism.
• Neurotrophic / neuroprotective effects. Lithium upregulates BDNF and the cytoprotective protein Bcl-2, promotes neurogenesis, and protects against various neuronal insults — the likely basis of the gray-matter and possible anti-dementia signals, and a convergence (again) on the plasticity/neuroprotection theme that runs through this entire series.
• Circadian and other effects. Lithium lengthens circadian period (relevant given the centrality of circadian disruption to bipolar illness).

The honest synthesis: lithium acts on fundamental intracellular signaling and neuroprotective machinery (GSK-3β and inositol signaling foremost) rather than on a single neurotransmitter receptor — which is probably why it does so many things, why its effects build slowly, and why it has been so hard to replicate with a cleaner designed drug. Its mechanism is a window into the deep biology of mood regulation that we have not fully opened.

The problems

Lithium's burdens are the reason it is underused, and they are real:

• Narrow therapeutic index. The therapeutic range (roughly 0.6–1.0 mEq/L for maintenance) sits uncomfortably close to the toxic range; lithium toxicity (tremor, ataxia, confusion, seizures, renal failure, death) is a medical emergency, and the margin for error is small. This mandates regular blood-level monitoring — a burden patients and busy clinicians dislike, and a major driver of underuse.
• Renal effects. Long-term lithium causes nephrogenic diabetes insipidus (polyuria) and, in a subset, chronic kidney disease over decades — the most serious long-term concern, requiring ongoing renal monitoring and sometimes forcing discontinuation.
• Thyroid and parathyroid. Hypothyroidism (common, manageable) and hyperparathyroidism/hypercalcemia (less appreciated, requires calcium monitoring).
• Teratogenicity. Associated with cardiac malformations (Ebstein's anomaly) — a real but smaller risk than once believed, requiring individualized risk–benefit in pregnancy rather than absolute avoidance.
• Interactions and dehydration risk. NSAIDs, ACE inhibitors, diuretics, and any dehydrating illness raise levels dangerously — patients must understand this.
• Tolerability. Tremor, weight gain, cognitive dulling ("lithium fog"), GI effects, acne/psoriasis — the day-to-day complaints that erode adherence.

The tragic irony: the drug with the best maintenance and the unique antisuicide effect is the one most neglected, precisely because it is old, cheap, unmarketed, and demanding to monitor — a case study in how non-pharmacological factors (patent status, marketing, monitoring burden) distort prescribing away from what the evidence supports.

The Anticonvulsant Mood Stabilizers

The observation that some antiepileptic drugs stabilize mood (the "kindling" analogy — the idea that recurrent mood episodes, like recurrent seizures, may sensitize the brain to further episodes) drove the adoption of anticonvulsants in bipolar disorder. Three matter, and they could hardly be more different from one another.

Valproate (valproic acid / divalproex)

What it does: A first-line anti-manic agent, effective and faster-acting than lithium for acute mania, widely used for maintenance, and favored for mixed states and rapid cycling. Like lithium, it is a poor antidepressant — strong on the manic pole, weak on the depressive.

Mechanism: Multiple and incompletely resolved — enhanced GABAergic transmission, sodium-channel modulation (reducing neuronal excitability), and, of growing interest, histone deacetylase (HDAC) inhibition, an epigenetic mechanism altering gene expression that connects valproate to the neuroplasticity/neuroprotection themes (and, ominously, to its developmental toxicity).

The problems — and a major safety scandal: Valproate's teratogenicity is among the most serious in all of medicine and the subject of a genuine regulatory and public-health reckoning. It causes neural tube defects, other major malformations, and — critically — a dose-dependent reduction in childhood IQ and a markedly increased risk of autism and neurodevelopmental disorders in children exposed in utero. The neurodevelopmental harm, under-recognized for years, has led many regulators to contraindicate valproate in women of childbearing potential except under strict pregnancy-prevention programs — a dramatic restriction reflecting how badly the risk was previously managed. Beyond teratogenicity: weight gain, polycystic ovary syndrome and menstrual disturbance, hepatotoxicity (rare but serious), pancreatitis, hyperammonemia, thrombocytopenia, and tremor/sedation. The valproate story is a cautionary tale about a widely-used drug whose gravest harm took decades to be properly weighed.

Lamotrigine

What it does: The mirror image of valproate — lamotrigine is the bipolar toolkit's best agent for the depressive pole, with a genuine role in maintenance and prevention of depressive recurrence (and bipolar depression), but it is nearly useless for acute mania and too slow to titrate for acute use of any kind. It is the closest thing to an antidepressant-leaning mood stabilizer, and a mainstay for bipolar II and predominantly-depressive bipolar illness.

Mechanism: Sodium-channel blockade with consequent inhibition of glutamate release — reducing excitatory transmission, which connects (via the glutamate theme of the rapid-acting document) to its mood effects, though the link to its specifically antidepressant-leaning profile is not fully explained.

The problems: The defining issue is the rash: lamotrigine can cause Stevens-Johnson syndrome and toxic epidermal necrolysis — rare but potentially fatal. This mandates a slow, weeks-long titration (rushing the dose sharply increases SJS risk), which makes lamotrigine useless for any acute situation and demands patient education about stopping at the first sign of rash. The slow titration and rash anxiety limit its use despite an otherwise favorable tolerability profile (it is weight-neutral and cognitively cleaner than most alternatives — significant advantages for long-term use).

Carbamazepine and oxcarbazepine

What they do: Carbamazepine is an effective anti-manic agent, second-line to lithium and valproate, useful in some treatment-resistant and mixed presentations; oxcarbazepine is a better-tolerated relative with thinner mood evidence.

Mechanism: Sodium-channel blockade reducing neuronal excitability.

The problems: Carbamazepine is a pharmacokinetic menace — a potent inducer of hepatic enzymes (including its own metabolism, "autoinduction") that lowers levels of numerous co-administered drugs (a major problem in the polypharmacy typical of bipolar treatment, and a notorious cause of contraceptive failure). It causes hyponatremia (more with oxcarbazepine), blood dyscrasias (rare agranulocytosis, aplastic anemia), rash including SJS/TEN — with a critical pharmacogenetic dimension: the HLA-B*1502 allele (common in some Asian populations) sharply elevates SJS/TEN risk, mandating screening before use in at-risk groups. Teratogenic (neural tube defects). The interaction burden alone relegates it to later-line use.

The Atypical Antipsychotics as Mood Stabilizers

The quiet revolution in bipolar treatment over the past two decades is that second-generation antipsychotics have become the de facto workhorses — used across all three phases, often where lithium and the anticonvulsants fall short, and increasingly as first-line agents in their own right. They are not traditionally what "mood stabilizer" evoked, but functionally they now do much of the stabilizing.

What they do, by phase:

• Acute mania: the antipsychotics (olanzapine, risperidone, quetiapine, aripiprazole, ziprasidone, asenapine, cariprazine) are first-line and fast — faster than lithium or valproate — and are the backbone of acute manic treatment, often combined with a traditional stabilizer.
• Bipolar depression — the hard problem — is where several antipsychotics have carved out their most valuable niche, because they are among the few agents with genuine efficacy here: quetiapine (robust bipolar-depression evidence), lurasidone (effective, metabolically cleaner — a significant advantage), cariprazine (a newer partial agonist with bipolar-depression efficacy), and the olanzapine–fluoxetine combination. In an illness whose depressive burden dominates and whose depressive treatments are weakest, this is a major contribution.
• Maintenance: several (aripiprazole, quetiapine, others, including long-acting injectables) have maintenance indications, alone or adjunctive.

Mechanism: The mood effects are tied to the broader antipsychotic pharmacology — dopamine D2 receptor antagonism (or, for aripiprazole/cariprazine/brexpiprazole, D2 partial agonism — "dopamine system stabilization") plus, for most, 5-HT2A antagonism and varied actions across serotonergic, adrenergic, histaminergic, and muscarinic receptors. The anti-manic effect plausibly relates to dampening hyperdopaminergic states; the bipolar-depression and quetiapine-specific effects are less cleanly explained (quetiapine's active metabolite norquetiapine is a norepinephrine reuptake inhibitor and 5-HT2C/partial-5-HT1A agent — a quasi-antidepressant pharmacology hidden inside an antipsychotic, which may explain its depressive-phase efficacy).

The problems — chiefly metabolic: The price of the antipsychotics' versatility is a serious burden: weight gain, dyslipidemia, insulin resistance, and type 2 diabetes (the metabolic syndrome), worst with olanzapine and quetiapine, lesser with aripiprazole, lurasidone, and ziprasidone — a major driver of the reduced life expectancy associated with serious mental illness, and a particular concern given that bipolar patients take these for years to decades. Add sedation, extrapyramidal symptoms and akathisia (less than first-generation agents but real), tardive dyskinesia (a potentially irreversible movement disorder from long-term D2 blockade), hyperprolactinemia (risperidone especially), and QT effects. The metabolic toll is the central trade-off: antipsychotics work across the phases where the traditional stabilizers fail, but they exact a long-term physical-health cost that lithium and lamotrigine do not.

The Treatment Logic of Bipolar Disorder

Pulling the agents together into the clinical reasoning they serve:

Acute mania: rapid control with an antipsychotic and/or lithium or valproate, often in combination; benzodiazepines adjunctively for agitation/sleep; antidepressants stopped.

Acute bipolar depression — the hard problem: the evidence-based options are narrow — quetiapine, lurasidone, cariprazine, the olanzapine–fluoxetine combination, lamotrigine (slow), and lithium — not, as first instinct, an antidepressant, because antidepressant monotherapy in bipolar disorder risks manic/hypomanic switching and cycle acceleration, and its efficacy even as add-on is marginal and contested. The asymmetry bites hardest here: the phase patients suffer most has the fewest good treatments.

Maintenance: lithium first by the weight of evidence (and uniquely for its antisuicide effect), with lamotrigine for depression-predominant illness, and antipsychotics and valproate by profile and history — usually long-term, often combination, with the perennial tension between relapse prevention and cumulative side-effect burden.

The cross-cutting mechanistic note: there is no unified mechanism of mood stabilization. Lithium (GSK-3β/inositol/neuroprotection), valproate (GABA/sodium channels/HDAC), lamotrigine (glutamate release), carbamazepine (sodium channels), and the antipsychotics (dopamine/serotonin) act through entirely different molecular routes to overlapping clinical ends. The suggestive convergence — as everywhere in this series — is on intracellular signaling, neuroprotection, and plasticity (lithium and valproate both affect these directly), hinting that mood stabilization, like antidepressant action, may ultimately be about regulating the brain's capacity for and constraint on change rather than any single transmitter. But this remains a hint, not a theory.

Problems With Their Use

Beyond the agent-specific toxicities above, several cross-cutting problems define this class:

Narrow therapeutic windows and monitoring burden. Lithium and carbamazepine require blood-level monitoring; lithium, valproate, and carbamazepine require organ monitoring (renal, thyroid, hepatic, hematologic). This burden is real, drives underuse of the best agents (lithium especially), and demands a level of ongoing engagement that fragmented care systems deliver poorly.

Serious long-term toxicities. Lithium's renal risk, valproate's teratogenic and neurodevelopmental harm, the antipsychotics' metabolic syndrome and tardive dyskinesia — these are not nuisance side effects but conditions that shorten or damage lives, accruing over the decades of treatment a chronic relapsing illness requires. The benefit is real, but so is the cumulative cost, and honest long-term care means continuously re-weighing them.

Teratogenicity in a young, often female population. Bipolar disorder commonly presents in women of reproductive age, and the two oldest stabilizers (valproate, carbamazepine, and to a lesser graded degree lithium) carry significant teratogenic risk — making reproductive planning, contraception, and pregnancy management central rather than peripheral, and making valproate, in particular, increasingly contraindicated in this group.

The bipolar-depression gap. The toolkit's structural weakness: the phase that dominates the illness burden is the phase the drugs treat worst, with a narrow evidence base and the ever-present hazard that the intuitive treatment (antidepressants) may worsen the course.

Adherence. Bipolar patients face a brutal adherence challenge — the side effects are immediate and the benefit (a relapse that doesn't happen) is invisible, and the manic phase itself can feel desirable, so patients stop, relapse, and re-present, in a cycle that the drugs' tolerability problems actively worsen. Adherence is arguably the field's biggest practical determinant of outcome, and the side-effect burden is its biggest enemy.

Diagnostic overreach and indication creep. A subtler problem: "mood stabilizer" has become a loose label applied well beyond bipolar disorder — for emotional dysregulation, irritability, "mood swings," and borderline personality features, where the evidence is far weaker than for bipolar illness, where bipolar disorder is frequently overdiagnosed (affective instability is not bipolarity), and where patients are committed to long-term drugs with serious toxicities for indications those drugs were never well shown to treat. The valproate-for-borderline and antipsychotic-for-irritability patterns deserve the same skepticism this series applied to antidepressants-for-mild-distress: real drugs, real toxicities, weak evidence outside their proven niche.

A Theoretical Synthesis

Mood stabilization is the least mechanistically unified concept in this series — a clinical target ("keep the patient out of both poles without inducing either") served by drugs with nothing molecular in common. And yet a faint convergence appears: lithium and valproate, the two prototypes, both act on intracellular signaling cascades, neuroprotection, and gene expression (GSK-3β, inositol, HDAC, BDNF, Bcl-2) rather than on neurotransmitter receptors, and both are associated with neuroprotective and gray-matter-preserving effects. This hints — no more than hints — that mood stabilization, like antidepressant and rapid-acting action, may ultimately be a matter of regulating the brain's plastic and homeostatic machinery: where antidepressants and psychedelics increase plasticity to permit escape from depressive patterns, mood stabilizers may constrain the excess plasticity/excitability of the manic pole and protect against the neurotoxicity of recurrent episodes (the kindling idea). The antipsychotics reach overlapping clinical ends through a wholly different (dopaminergic) route, reminding us that the clinical category is defined by effect, not mechanism.

What the field most lacks is a real treatment for bipolar depression and a stabilizer with lithium's efficacy and antisuicide effect but without its toxicity and monitoring burden. Sixty years of rational drug design have not bettered an element discovered by accident — which is either an indictment of our mechanistic understanding or a clue that lithium's promiscuous action on fundamental cellular machinery is exactly the point, and hard to improve upon with a cleaner molecule.

The Clinical Bottom Line

Lithium remains the reference standard and is underused: the best maintenance agent, uniquely antisuicidal, neuroprotective — warranting first consideration for bipolar maintenance despite its monitoring demands, with honest attention to renal, thyroid, and parathyroid surveillance and the narrow therapeutic window. Its neglect relative to its evidence is a genuine failure of modern prescribing.

Valproate is an effective anti-manic agent that is contraindicated in women of childbearing potential outside strict pregnancy-prevention programs, owing to its grave teratogenic and neurodevelopmental harms — a restriction that should be taken seriously rather than worked around.

Lamotrigine is the agent for the depressive pole and depression-predominant maintenance, weight- and cognition-friendly, but slow to titrate and shadowed by the SJS/TEN risk that mandates that slow titration.

Carbamazepine is a later-line anti-manic agent whose enzyme-induction interactions and (HLA-B*1502-linked) rash risk limit it.

The atypical antipsychotics are now central across all phases — first-line and fast for mania, among the few effective options for bipolar depression (quetiapine, lurasidone, cariprazine), and useful in maintenance — at the cost of a serious, cumulative metabolic burden and tardive-dyskinesia risk that must be monitored and weighed over the long haul.

Across all of them: bipolar disorder is a serious, recurrent, life-shortening illness for which these drugs are genuinely effective and often life-saving — and a toolkit that is unevenly effective (weakest where the illness hurts most), seriously toxic over the decades of use it requires, and too often stretched to treat affective instability that is not bipolar disorder at all. The disciplined position: treat genuine bipolar illness vigorously and for the long term, favor lithium more than current practice does, weigh the cumulative toxicities continuously, manage reproductive risk centrally, and resist the indication creep that commits non-bipolar patients to these powerful drugs on thin evidence.

Selected References and Further Reading

1. Cade, J.F.J. (1949). Lithium salts in the treatment of psychotic excitement. Medical Journal of Australia, 2(10), 349–352.
2. Geddes, J.R., & Miklowitz, D.J. (2013). Treatment of bipolar disorder. The Lancet, 381(9878), 1672–1682.
3. Cipriani, A., et al. (2011). Comparative efficacy and acceptability of antimanic drugs in acute mania: A multiple-treatments meta-analysis. The Lancet, 378(9799), 1306–1315.
4. Cipriani, A., Hawton, K., Stockton, S., & Geddes, J.R. (2013). Lithium in the prevention of suicide in mood disorders: Updated systematic review and meta-analysis. BMJ, 346, f3646.
5. Geddes, J.R., Goodwin, G.M., et al. (2010). Lithium plus valproate combination therapy versus monotherapy for relapse prevention in bipolar I disorder (BALANCE). The Lancet, 375(9712), 385–395.
6. Malhi, G.S., et al. (2013). Potential mechanisms of action of lithium in bipolar disorder. CNS Drugs, 27(2), 135–153.
7. Quiroz, J.A., Machado-Vieira, R., Zarate, C.A., & Manji, H.K. (2010). Novel insights into lithium's mechanism of action. Neuropsychobiology, 62(1), 50–60.
8. Goodwin, G.M., et al. (2016). Evidence-based guidelines for treating bipolar disorder: Revised third edition (BAP guidelines). Journal of Psychopharmacology, 30(6), 495–553.
9. Yatham, L.N., et al. (2018). CANMAT and ISBD guidelines for the management of patients with bipolar disorder. Bipolar Disorders, 20(2), 97–170.
10. Geddes, J.R., Calabrese, J.R., & Goodwin, G.M. (2009). Lamotrigine for treatment of bipolar depression: Independent meta-analysis. British Journal of Psychiatry, 194(1), 4–9.
11. Meador, K.J., et al. (2013). Fetal antiepileptic drug exposure and cognitive outcomes at age 6 years (NEAD study). Lancet Neurology, 12(3), 244–252.
12. Christensen, J., et al. (2013). Prenatal valproate exposure and risk of autism spectrum disorders. JAMA, 309(16), 1696–1703.
13. Calabrese, J.R., et al. (2005). A randomized, double-blind, placebo-controlled trial of quetiapine in bipolar depression (BOLDER). American Journal of Psychiatry, 162(7), 1351–1360.
14. Loebel, A., et al. (2014). Lurasidone monotherapy in the treatment of bipolar I depression. American Journal of Psychiatry, 171(2), 160–168.
15. Severus, E., et al. (2014). Lithium for prevention of mood episodes in bipolar disorders: Systematic review and meta-analysis. International Journal of Bipolar Disorders, 2, 15.
16. McKnight, R.F., et al. (2012). Lithium toxicity profile: A systematic review and meta-analysis. The Lancet, 379(9817), 721–728.
17. Phiel, C.J., et al. (2001). Histone deacetylase inhibition by valproic acid. Journal of Biological Chemistry / molecular pharmacology literature.
18. Pisanu, C., Heilbronner, U., & Squassina, A. (2018). The role of pharmacogenomics in bipolar disorder. CNS Drugs, 32(6), 533–546.
19. Vieta, E., et al. (2018). Bipolar disorders (Primer). Nature Reviews Disease Primers, 4, 18008.
20. Post, R.M. (2007). Kindling and sensitization as models for affective episode recurrence. Neuroscience & Biobehavioral Reviews, 31(6), 858–873.