Once bipolar disorder has been diagnosed, another practical question immediately follows:

Which mood stabiliser should we choose—lithium or valproate?

There is no universal answer.

Both are established treatments for bipolar disorder, particularly for mania and relapse prevention. WHO includes both lithium and valproate among effective treatments for acute bipolar mania and among options for maintenance treatment.

But they are not interchangeable.

A better way to decide is to ask:

What kind of bipolar illness does this patient have, what phase are we treating, what happened in previous episodes, and what are the patient’s medical and reproductive risks?

In many patients with classical episodic bipolar I disorder, lithium remains particularly attractive for long-term treatment.

In others—especially when lithium is unsuitable, monitoring is difficult, or the illness has a more complicated manic presentation—valproate may become a reasonable alternative.

The decision should therefore be phenotype-driven rather than simply prescription-driven.

First Question: Are We Treating Acute Mania or Preventing the Next Episode?

This distinction is fundamental.

A severely manic patient may need rapid control of:

agitation,

psychosis,

aggression,

severe insomnia,

disinhibition,

and dangerous behaviour.

In this situation, neither “lithium versus valproate” captures the entire treatment decision.

Antipsychotics frequently play an important role in acute mania.

WHO currently regards several antipsychotics as well as lithium and valproate as effective antimanic options.

NICE takes an even more sequential approach: an antipsychotic is generally used first for acute mania; lithium can be added when response is inadequate, and valproate considered when lithium is ineffective or unsuitable.

Once the acute episode settles, however, a different question emerges:

What treatment gives this particular patient the best chance of remaining well over the next several years?

That is where lithium often becomes especially important.

Lithium Has a Particularly Strong Position in Long-Term Bipolar Treatment

NICE specifically recommends lithium as first-line long-term pharmacological treatment for bipolar disorder.

Why?

Lithium has an unusually broad evidence base across:

mania prevention,

depression prevention,

overall relapse prevention,

and possibly suicide prevention.

A review comparing lithium and valproate concluded that lithium’s clinical advantage is most apparent when used for prevention of both manic and depressive recurrence rather than merely for acute symptom control.

Therefore, if I am looking at a patient with:

clear bipolar I disorder,

distinct episodes,

good recovery between episodes,

classical euphoric mania,

relatively few comorbidities,

and a need for long-term prophylaxis,

lithium moves high on the list.

The “Classic Lithium Responder”

Psychiatry has known for decades that certain bipolar presentations appear particularly lithium-responsive.

Modern meta-analysis supports at least some of these clinical observations.

A systematic review involving more than 12,000 patients found better lithium response associated with factors including:

a mania → depression → well interval pattern,

absence of rapid cycling,

absence of psychotic symptoms,

family history of bipolar disorder,

shorter duration of illness before lithium treatment,

and somewhat later illness onset.

None of these features guarantees response.

But together they describe something close to the classical episodic manic-depressive illness for which lithium has historically performed particularly well.

Family Response to Lithium May Actually Be Useful

Family history may tell us more than simply whether bipolar disorder runs in the family.

A 2026 multicentre study examined lithium response within families and found a striking pattern.

Among relatives of known lithium responders, 69% were themselves good lithium responders, compared with only 22% among relatives of lithium non-responders.

The odds of responding were approximately eight times higher in families of lithium responders.

So one question worth asking is:

“Does anyone in the family have bipolar disorder, and if so, did lithium work particularly well for them?”

That information may genuinely contribute to treatment selection.

What About Suicide Risk?

Lithium has long attracted particular interest because of a possible anti-suicidal effect.

Observational studies and older meta-analyses have suggested reductions in suicide and suicidal behaviour during lithium treatment.

However, more recent analyses of randomized trials have produced less definitive results because suicide is relatively rare and trials are usually underpowered for this outcome. A 2022 systematic review therefore concluded that the magnitude of lithium’s specific suicide-prevention effect remains statistically uncertain in randomized evidence.

Clinically, lithium still has the strongest historical evidence base among mood stabilisers in patients with bipolar disorder and substantial suicide risk.

But it is more scientifically accurate to say:

lithium has an important and suggestive anti-suicidal evidence base

rather than:

lithium has been conclusively proven to prevent suicide in every bipolar patient.

When Does Valproate Become More Attractive?

Valproate is also an effective antimanic and maintenance treatment.

WHO includes it alongside lithium among established treatments for mania and as a maintenance option.

Historically, clinicians have often considered valproate particularly when bipolar illness appears less “classical.”

Examples include patients with:

mixed or dysphoric manic states,

high episode burden,

rapid cycling,

significant psychiatric comorbidity,

or poor lithium suitability.

A review comparing lithium and valproate found that lithium tended to fit patients with classical bipolar features, fewer previous episodes and fewer comorbidities, whereas valproate was often favoured in patients with larger numbers of previous episodes or psychiatric comorbidity.

But this distinction should not be exaggerated.

Is Valproate Definitely Better for Mixed Mania?

Not necessarily.

Older clinical literature often taught:

classic euphoric mania → lithium

mixed/dysphoric mania → valproate

There is some evidence behind that tradition.

However, contemporary data are less decisive.

CANMAT/ISBD recommendations for mania with DSM-5 mixed features place divalproex as a second-line option, alongside medications such as asenapine, cariprazine and aripiprazole; the evidence was not strong enough to identify any mood stabiliser as unequivocal first-line treatment for DSM-5 mixed mania.

So I would interpret mixed features as:

a reason to be somewhat less confident that lithium will behave like the ideal “classic lithium responder” treatment—not proof that valproate must be chosen.

What About Rapid Cycling?

The same caution applies.

Historically, valproate has often been preferred over lithium in rapid-cycling bipolar disorder.

And meta-analytic studies of lithium response consistently find that rapid cycling predicts poorer lithium response. In one systematic review, rapid cycling was associated with substantially lower odds of good lithium response.

But that does not automatically prove that valproate is dramatically superior.

Rapid-cycling bipolar disorder is intrinsically difficult to treat, and comparative evidence remains imperfect.

So the clinical reasoning should be:

rapid cycling makes me somewhat less enthusiastic about lithium monotherapy

rather than:

rapid cycling means valproate will definitely work.

Lithium vs Valproate: A Practical Clinical Comparison

Clinical situation	Lithium	Valproate
Classical euphoric bipolar I	Often strongly favoured	Effective alternative
Long-term relapse prevention	Particularly strong evidence	Effective
Previous excellent lithium response	Strongly favoured	Usually no reason to switch
Family member strongly lithium-responsive	Supports lithium choice	—
High suicide risk	Often favoured	Less specific evidence
Mixed/dysphoric presentation	May respond less predictably	Often considered, though evidence is not absolute
Rapid cycling	Response may be poorer	Often considered
High number of previous episodes	Lithium may be less predictably effective	May be reasonable
Renal disease	Often problematic	Usually easier than lithium
Hypothyroidism	Can aggravate thyroid dysfunction	Usually less relevant
Obesity/metabolic concerns	Can cause weight gain but often modest	Weight gain can be substantial
Liver disease	Usually preferable from hepatic perspective	Potential problem
Thrombocytopenia	Usually not central issue	Potential problem
Reliable access to blood monitoring	Suitable	Suitable
Difficulty maintaining hydration / recurrent dehydration	Lithium becomes problematic	Often easier
Woman or girl of childbearing potential	Requires careful pregnancy planning	Generally avoid valproate

The table is a framework—not an algorithm.

Reproductive Safety Can Completely Change the Decision

This is perhaps the single clearest situation in which lithium and valproate diverge.

WHO advises that valproate should not be prescribed to women and girls of childbearing potential because of the high risk of major congenital malformations and neurodevelopmental problems following fetal exposure.

Therefore, in a young woman with bipolar disorder, the question should not casually be:

“Lithium or valproate—which is easier?”

Valproate should generally be avoided where pregnancy is possible, and reproductive plans must be explicitly discussed.

Lithium also has pregnancy-related risks and requires specialist management, but valproate’s teratogenic and neurodevelopmental risks make its reproductive restrictions particularly important. WHO similarly advises avoiding lithium during pregnancy and breastfeeding whenever possible, but the clinical risk-benefit situation is not identical to valproate.

This is an area where international regulatory requirements continue to evolve, so clinicians should also follow current local prescribing guidance.

Lithium’s Main Disadvantage: It Demands Respect

Lithium is not a “prescribe and forget” medicine.

It has a relatively narrow therapeutic range.

The same medicine that can provide excellent mood stabilization can become toxic when levels rise significantly.

Situations that can increase lithium levels include:

dehydration,

vomiting or diarrhoea,

kidney impairment,

certain antihypertensives,

diuretics,

and NSAIDs.

Patients therefore need both laboratory monitoring and education.

NICE recommends baseline assessment including weight/BMI, renal function, electrolytes including calcium, thyroid function and blood count, with ECG where cardiovascular risk warrants it. Lithium levels are then checked after initiation and dose changes and periodically thereafter.

For long-term treatment, renal function, thyroid function and calcium also require regular monitoring.

The “Dehydration Question” Matters More Than It Seems

Before prescribing lithium, I would specifically want to know:

Does the patient frequently become dehydrated?

Do they work outdoors in extreme heat?

Do they undertake prolonged endurance exercise?

Do they fast extensively?

Do they regularly develop diarrhoea or vomiting?

Do they take NSAIDs frequently?

Can they reliably attend blood monitoring?

Lithium can still sometimes be used in challenging circumstances.

But these factors alter the risk-benefit calculation.

In hot climates such as Chennai, educating patients about hydration and illness-related lithium toxicity is particularly relevant.

Lithium and Kidney Function

Lithium is cleared predominantly through the kidneys.

That makes renal function central to long-term decision-making.

NICE recommends more frequent monitoring when creatinine rises or eGFR falls and specifically advises weighing lithium’s clinical benefit against the extent and trajectory of renal impairment.

Therefore:

significant renal impairment pushes the decision away from lithium.

Mild abnormalities do not automatically require permanent discontinuation, but they may justify nephrology input and closer monitoring.

Lithium and Thyroid Function

Hypothyroidism is another well-known lithium-associated problem.

Therefore thyroid function should be measured before treatment and periodically afterwards.

Development of hypothyroidism does not automatically mean lithium must be stopped.

If lithium has produced exceptional mood stability, clinicians may sometimes treat the thyroid dysfunction and continue lithium.

This is another example of why mood stabiliser selection cannot be reduced to side-effect lists.

The question is always:

How much psychiatric benefit are we receiving in exchange for this medical burden?

Valproate Has a Different Medical Burden

Valproate does not require lithium-level style therapeutic monitoring in routine bipolar treatment.

But it has its own important adverse-effect profile.

Relevant concerns include:

weight gain,

sedation,

tremor,

gastrointestinal effects,

liver toxicity,

thrombocytopenia,

and major reproductive toxicity.

NICE recommends monitoring weight, full blood count and liver function during valproate treatment.

Therefore a patient with:

significant liver disease

or

important thrombocytopenia

is not an obvious valproate candidate.

Weight Gain Can Be a Deciding Factor

This matters enormously for long-term adherence.

Many patients with bipolar disorder are simultaneously exposed to antipsychotics that can themselves produce:

weight gain,

insulin resistance,

dyslipidaemia.

Adding valproate can increase the metabolic burden in some patients.

If a patient already has:

obesity,

prediabetes,

fatty liver,

or major concern about weight,

valproate becomes less attractive.

Lithium can also cause weight gain in some patients, but the decision should consider the entire medication regimen, not one drug in isolation.

What About Bipolar Depression?

This is where the simplistic “lithium versus valproate” comparison becomes less useful.

Neither drug should automatically be thought of as the perfect acute bipolar-depression treatment.

Modern bipolar depression treatment frequently involves other agents such as:

quetiapine,

lamotrigine,

lurasidone,

or selected combinations,

depending on the guideline, illness history and patient characteristics.

WHO includes lithium and valproate within broader treatment strategies for bipolar depression, but evidence strength varies considerably across specific interventions.

Lithium’s major advantage is often its longitudinal stabilizing value, rather than rapid antidepressant action.

Previous Response Usually Beats Theoretical Prediction

Suppose someone previously took lithium for five years and had:

no manic episodes,

no depressive episodes,

good functioning,

minimal adverse effects.

Then stopped it and relapsed.

There is very little reason to ignore that information because a textbook says another medication fits some theoretical phenotype.

Past individual response is one of the strongest predictors we possess.

The same applies to valproate.

If valproate has repeatedly produced excellent stabilization with acceptable tolerability, that real-world response carries substantial weight.

Sometimes the Answer Is Lithium AND Valproate

Bipolar disorder does not always respect our desire for elegant monotherapy.

Some difficult cases require combination treatment.

WHO and other guidelines recognize combination strategies, and NICE includes lithium–valproate combinations among later options for difficult maintenance treatment.

Combination treatment may increase efficacy in selected patients but also increases:

side effects,

monitoring,

polypharmacy,

and treatment complexity.

Therefore combination treatment should generally have a clear rationale rather than being accumulated indefinitely.

A Simple Way to Think About the Decision

If I see classical bipolar I, with distinct euphoric mania, clear remission between episodes, low comorbidity, good renal/thyroid status, ability to undergo monitoring and especially a strong family or personal lithium-response history:

Lithium becomes very attractive.

If I see predominantly acute manic illness with mixed/dysphoric features, high episode burden, rapid cycling tendencies or substantial comorbidity, and lithium appears poorly suitable or has already failed:

Valproate becomes more attractive—provided reproductive, hepatic, haematological and metabolic factors allow it.

If I see a woman of childbearing potential:

Valproate moves very far down—or effectively off—the list under current WHO safety recommendations.

If I see substantial renal impairment, repeated dehydration or inability to perform lithium monitoring:

Lithium becomes less attractive.

If I see substantial liver disease, thrombocytopenia, obesity/metabolic vulnerability or reproductive risk:

Valproate becomes less attractive.

And if the patient has previously demonstrated a spectacular response to one of them:

That previous response may outweigh many theoretical predictors.

We Still Cannot Predict Mood-Stabiliser Response Perfectly

This is important.

Despite decades of research, psychiatry does not yet have a blood test, QEEG marker, genetic test or MRI scan that reliably says:

“This patient will respond to lithium.”

or:

“This patient should receive valproate.”

A 2025 review of treatment prediction in bipolar disorder concluded that the strongest signals remain clinical characteristics, particularly for lithium—episodic course, absence of rapid cycling, lower comorbidity and family patterns of bipolar disorder or lithium response. Biomarkers remain promising but are not yet sufficiently validated for routine treatment selection.

So for the moment, the best “precision psychiatry” tool remains:

a detailed longitudinal clinical history.

Lithium or Valproate? The Better Question

Instead of asking:

“Which is the better mood stabiliser?”

ask:

“Which drug best fits this patient’s bipolar phenotype, previous response, medical status, reproductive circumstances and long-term treatment goals?”

That produces a much better decision.

Lithium often fits the classical, episodic bipolar patient requiring strong long-term prophylaxis.

Valproate remains a valuable antimanic and maintenance option, particularly when lithium is unsuitable or the illness phenotype is more complicated.

Neither should be selected simply because:

“This is what I usually prescribe for bipolar disorder.”

Clinical Takeaway

Choose the patient before choosing the mood stabiliser.

The most useful variables are often not complicated biomarkers.

They are:

episode polarity and sequence, episodicity, mixed features, rapid cycling, family history, previous treatment response, suicide risk, renal and thyroid health, liver and metabolic health, reproductive considerations, and the patient’s ability to undergo long-term monitoring.

That is where rational mood-stabiliser selection begins.

Dr. Srinivas Rajkumar T

Senior Consultant Psychiatrist
MD Psychiatry — AIIMS New Delhi

Clinical interests include Bipolar Disorder, Depression, Treatment-Resistant Mood Disorders, Diagnostic Psychiatry and Interventional Psychiatry

Chennai Consultations

Apollo Clinic, opposite Phoenix Market City, Velachery, Chennai

Appointments: +91 85951 55808
Email: srinivasaiims@gmail.com

The best mood stabiliser is not necessarily the newest medication or the easiest one to prescribe. It is the treatment that best matches the patient’s longitudinal illness pattern while remaining medically safe and sustainable for years.