Bipolar Depression: The 2026 PMHNP Treatment Algorithm

Roughly 40 percent of patients diagnosed with major depression actually sit somewhere on the bipolar spectrum, and the first script a new PMHNP writes can push them into mixed features, rapid cycling, or frank mania within weeks.

That is not a theoretical risk. Missed bipolarity is the single most common cause of “treatment-resistant depression” referrals that land on a mentor’s desk.

Carlat Publishing’s 8th edition Medication Fact Book, released for 2026, finally adds a standalone bipolar depression treatment algorithm — the single most requested addition in the prior edition’s reader survey. The reason is simple: early-career psychiatric nurse practitioners are asked to make this call before most training programs have taught the nuance.

The algorithm is short, evidence-grounded, and worth committing to memory. Four drugs own the first-line space. One meal changes whether your prescription works. Two labs determine which option is even safe.

This is the 2026 algorithm, the clinical pearls behind each choice, and the monitoring protocol Lindsay Hill, DNP, PMHNP-BC uses inside the Psych NP Fellowship to keep new grads out of the “why is she still depressed” trap.

Why bipolar depression breaks the MDD algorithm

Unipolar depression responds to serotonergic pressure. Bipolar depression punishes it.

Starting an SSRI in an unrecognized bipolar patient can trigger a switch into hypomania in roughly 10 to 20 percent of cases and a full manic episode in a smaller but meaningful minority. The risk is highest in bipolar I, younger patients, and those with a family history of bipolar illness.

Therefore, every depression evaluation needs three questions before a prescription is written. First — any past episode of decreased need for sleep with increased goal-directed activity? Second — any family history of bipolar disorder, psychiatric hospitalization, or completed suicide? Third — any past antidepressant trial that produced unusual energy, irritability, or insomnia instead of recovery?

However, the Mood Disorder Questionnaire (MDQ) catches most bipolar I and maybe half of bipolar II. It is not sufficient. In short, a longitudinal history — ideally with a collateral informant — remains the standard.

Specifically, once bipolarity is established, the algorithm shifts. First, antidepressant monotherapy drops off the list. Then, lithium, lamotrigine, and second-generation antipsychotics with proven bipolar depression efficacy move to the front.

“The question I ask every new Fellow is whether they can name four first-line options for bipolar depression without pausing. If they hesitate, we rebuild the algorithm before we touch a single chart.” — Lindsay Hill, DNP, PMHNP-BC

The distinction matters because bipolar depression accounts for the majority of time patients with bipolar I and II spend symptomatic across their lifetime. Indeed, mania gets the attention in the literature. However, depression takes the years.

Therefore, getting the first script right is the most important clinical decision a PMHNP makes in the first six months of practice.

First-line bipolar depression monotherapy options for 2026

Four drugs carry a first-line indication for bipolar depression in the 2026 algorithms. Specifically, each has a specific niche.

Quetiapine (Seroquel, Seroquel XR)

First-line for both bipolar I and bipolar II depression, and the only monotherapy listed as first-line for bipolar II in the most recent CANMAT/ISBD guidelines. Typical dosing runs 300 mg at bedtime, titrated over four days — 50 mg day 1, 100 mg day 2, 200 mg day 3, 300 mg day 4. Importantly, the sedation is the feature, not the bug, for patients with depression-driven insomnia. However, metabolic monitoring is required.

Lurasidone (Latuda)

First-line for bipolar I depression as monotherapy or added to lithium or valproate. Starting dose is 20 mg once daily, raised to effect in the 40 to 60 mg range, with a 120 mg maximum. The critical pearl — lurasidone must be taken with a meal of at least 350 calories or absorption drops roughly 50 percent. In other words, a patient taking it on an empty stomach is effectively taking half a dose. Therefore, food consistency drives response.

Cariprazine (Vraylar)

Moved from second-line to first-line for bipolar I depression in the 2023 CANMAT update and kept in 2026 algorithms. Starting dose is 1.5 mg once daily, raised to 3 mg on day 15, with a 3 mg maximum. The 14-day interval matters — faster titration produced a 10 percent rate of akathisia in trials versus 6 percent at the slower schedule. Side effects may first appear several weeks after starting the drug because the parent drug and active metabolites build up.

Lamotrigine (Lamictal)

First-line for bipolar II maintenance and bipolar depression prevention, with a weaker acute antidepressant effect than the antipsychotics above. The 6-week titration — 25 mg for 2 weeks, 50 mg for 2 weeks, 100 mg for 1 week, then 200 mg target — is required because Stevens-Johnson syndrome risk is dose- and titration-speed-dependent. Notably, valproate doubles lamotrigine levels and requires a slower schedule.

Olanzapine-fluoxetine combination (Symbyax) still carries an FDA indication but has fallen out of first-line recommendations in most 2026 algorithms because of its metabolic burden relative to lurasidone and cariprazine.

Bipolar depression adjunctive combinations with lithium or valproate

When a patient is already stabilized on a mood stabilizer and the depressive episode persists, the algorithm moves to adjunctive therapy rather than a new monotherapy.

Lithium + lurasidone. Lurasidone is FDA-approved as adjunctive therapy with lithium in bipolar I depression. The evidence base is strong and the combination tolerates a full lurasidone dose range of 20 to 120 mg. Lithium levels need to sit in the 0.6 to 0.8 mEq/L range for efficacy — lower trough levels predict relapse.

Valproate + lurasidone. Likewise, the same FDA adjunctive indication applies as for the lithium combination. Valproate trough levels run 50 to 125 mcg/mL. Monitor LFTs and CBC given valproate’s hepatic and hematologic signals.

Lithium + lamotrigine. Favored for maintenance, particularly when the predominant pole is depression. Lamotrigine’s anti-manic effect is weaker than lithium’s, so lithium covers the mania-prevention side and lamotrigine handles the depression-prevention side.

Lithium + quetiapine. Used when rapid effect is needed during an acute depressive episode and the patient is already on lithium. The combination adds sedative and metabolic burden, so weight, waist circumference, and A1c need a baseline and quarterly follow-up.

Avoid lithium plus an SSRI as a routine first-line move in bipolar depression. The SSRI still carries switch risk. Adding lithium does not eliminate it.

Avoid valproate monotherapy for acute bipolar depression. The acute antidepressant evidence is weaker than lithium’s, and valproate is teratogenic with a Category X designation in pregnancy for neural tube defects, requiring pregnancy prevention counseling for patients of reproductive potential.

Adjunctive decisions also factor in what the patient failed. A patient who failed lithium monotherapy for depression rarely gets adequate symptom control by adding a partial dopamine antagonist without revisiting the lithium level first.

Bipolar depression prescribing: what most people get wrong

Three errors show up in almost every new PMHNP’s first six months of prescribing.

Starting an SSRI “while we figure out bipolarity.” This is the highest-cost mistake. A two-week quick start of sertraline before a full mood history can convert a bipolar II patient’s hypomania into a mixed episode and erase the diagnostic clarity needed for the actual algorithm. Get the mood history first. Hold the prescription until it is complete.

Dosing lurasidone without teaching the 350-calorie rule. The food requirement is not a preference. Without 350 calories, the AUC drops about 50 percent and Cmax drops about 67 percent. A patient who takes it at bedtime on an empty stomach and reports “it isn’t working” is correct — they are taking a functionally subtherapeutic dose. The fix is patient education at the first script, not a dose increase at the follow-up.

Rapid titration of lamotrigine or cariprazine. Lamotrigine’s Stevens-Johnson syndrome risk is dose- and titration-rate-dependent — doubling the schedule is never worth saving two weeks. Cariprazine’s 14-day minimum interval before escalation exists because the active metabolites accumulate slowly, and faster titration raises akathisia from 6 percent to 10 percent without improving efficacy.

A fourth, quieter error is skipping baseline labs. Every first-line bipolar depression agent has a monitoring requirement. Quetiapine and cariprazine demand metabolic panels. Lithium requires renal and thyroid function. Valproate requires LFTs and CBC. A prescription written before the labs is a prescription that will need to be re-evaluated at the first abnormality.

“Most of the ‘treatment-resistant’ bipolar depression consults I see are not treatment-resistant at all. They are dosing, adherence, or diagnosis errors that were never audited.” — Lindsay Hill, DNP, PMHNP-BC

Bipolar depression monitoring and safety by drug class

Baseline labs and interval follow-up change based on the agent. A monitoring table lives in every Fellowship workbook for exactly this reason.

Quetiapine, lurasidone, cariprazine (second-generation antipsychotics)

Baseline. Weight, waist circumference, fasting glucose or A1c, lipid panel, blood pressure, and personal and family history of diabetes and dyslipidemia.

Follow-up. Weight at each visit, waist and A1c at 12 weeks then yearly, lipids at 12 weeks then yearly. AIMS exam at baseline and every 6 months. Quetiapine has the heaviest metabolic signal; lurasidone and cariprazine are lighter but not weight-neutral.

Lithium

Baseline. eGFR, urinalysis, TSH, calcium, pregnancy test in patients of reproductive potential, and an ECG in patients over 40 or with cardiac risk.

Follow-up. Lithium level 5 to 7 days after any dose change and at steady state. TSH and creatinine every 6 months. A rising creatinine or a falling eGFR is the signal to slow down and involve nephrology — not a reason to panic-discontinue.

Valproate

Baseline. LFTs, CBC with platelets, pregnancy test. Reliable pregnancy prevention is required for patients of reproductive potential given teratogenicity.

Follow-up. Level, LFTs, and CBC at 1 to 3 months, then every 6 months. Watch for thrombocytopenia and hyperammonemia.

Lamotrigine

Baseline. None routine.

Follow-up. Patient education about rash is the highest-yield step. Any new rash in the first 8 weeks stops the drug and gets a same-day evaluation. Stevens-Johnson syndrome risk is highest in the first 8 weeks and with rapid titration.

Documentation of labs, dose, and patient education at each visit is what separates a prescription that defends itself from one that becomes a malpractice exhibit.

The clinical bottom line

Bipolar depression is not a detour within the MDD algorithm. It is a separate tree.

Four drugs own the first-line space. One meal changes whether one of them works. One 14-day interval changes whether another is tolerated. Get those three facts right and the algorithm runs.

The gap most new PMHNPs hit is not intelligence or effort. It is the absence of a trained clinician reviewing the first 30 scripts in real time, catching the SSRI that shouldn’t have been written or the lurasidone dose that was taken on an empty stomach.

This content is for educational purposes and does not replace individualized clinical judgment or supervision.