Lurasidone (Latuda)

Lurasidone is the drug I keep reaching for when somebody walks in with bipolar depression and a body that’s already been wrecked by Seroquel. It’s an atypical antipsychotic, brand name Latuda, FDA-approved for schizophrenia and bipolar I depression. On paper it does what every other atypical does. In practice it’s one of the few in this class that won’t add thirty pounds and a metabolic panel that looks like a cry for help.

The catch, and there’s always one, is that you have to take it with food. Not a cracker. Not a piece of toast. At least 350 calories of actual food. Miss that and you’ve absorbed maybe half the dose, sometimes less. I’ve watched patients tell me lurasidone “stopped working” when what actually happened is they switched from dinner-with-pill to bedtime-with-pill three weeks ago and haven’t eaten anything substantial after 6 PM since.

That food rule is the single most underrated counseling point in the modern antipsychotic toolkit. It’s printed on the bottle. Patients still don’t do it. Pharmacists mention it in passing if you’re lucky. By the time somebody lands in my office saying the drug failed, the first question I ask is what they ate when they took it last night, and about a third of the time the answer is “nothing, I take it before bed.”

Why I pick it over Seroquel or Abilify

For bipolar depression specifically, the FDA-approved options are narrow. Quetiapine (Seroquel). Olanzapine-fluoxetine (Symbyax, which nobody prescribes anymore because olanzapine is a metabolic disaster). Lurasidone. Cariprazine (Vraylar) got the indication more recently. Lumateperone (Caplyta) joined the list. That’s basically it for monotherapy with an actual depression indication.

Seroquel works. It also sedates people into oblivion at the doses that treat bipolar depression (300mg, sometimes 600mg), and it puts weight on almost everyone. I had a woman in her early thirties last year, schoolteacher, bipolar II, who’d been on 300mg of Seroquel for two years and gained 42 pounds. Her A1c had crept into prediabetic range. She was sleeping fourteen hours on weekends. The Seroquel was doing its job for the depression, but the cost was eating her quality of life from the other side.

We cross-tapered to lurasidone, started at 20mg with dinner, worked up to 60mg over three weeks. The first ten days were rough with akathisia, which is the catch with this drug. After that she stabilized, dropped eleven pounds in the next three months without trying, and her A1c came back down. That’s the trade. You give up the sedation (which some people actually want) and you take on a real risk of akathisia in the first month, in exchange for a metabolic profile that looks almost like a placebo.

Abilify is the other comparator people ask about. Abilify is weight-neutralish, but it isn’t FDA-approved for bipolar depression as monotherapy, only as adjunctive treatment for major depression. The data for Abilify in bipolar depression is genuinely underwhelming. Lurasidone has the indication and the placebo-controlled trials behind it. If bipolar depression is the target, Latuda beats Abilify on actual evidence.

Akathisia is the thing that derails it

Akathisia is internal restlessness. Not anxiety exactly. Patients describe it as needing to move, skin crawling, can’t sit through a meeting, legs won’t quit. It looks like agitation from the outside and feels like torture from the inside. With lurasidone it shows up in roughly 10-15% of patients, usually in the first two weeks, usually dose-dependent.

The clinical mistake I see other prescribers make is interpreting akathisia as worsening bipolar agitation and bumping the dose. That makes it worse. The fix is either to drop the dose, slow the titration, or add propranolol 10-20mg twice a day, which works well for most people and is cheap. Benztropine doesn’t help akathisia much despite being lumped in as an EPS treatment. Propranolol is the move. Sometimes a low-dose benzo short-term while you sort out the dose.

Half the lurasidone failures I see come down to somebody taking it on an empty stomach, or akathisia that nobody named correctly in week two.

If a patient calls me at week one saying they feel jittery and can’t sit still, I don’t tell them to push through. That’s how you lose them. We back off to 20mg, add propranolol, and revisit in a week. Most of the time the akathisia fades by week three or four and we can climb back up.

Dosing, food, and the math nobody does

Standard dose range is 20-120mg once daily. For bipolar depression, the trial data shows 20-60mg works about as well as 80-120mg with fewer side effects, so I usually park people at 40-60mg unless they need more. For schizophrenia, 40-160mg, with 80mg being the most common landing spot.

The food thing again, because I can’t say it enough. Below 350 calories of food, absorption tanks. Studies have shown bioavailability drops by half or more on an empty stomach. So if your patient is taking it before bed with a glass of water, they’re effectively on half their dose, which usually means they’re getting the side effects of a real dose and the efficacy of a fake one.

I tell patients to pick one meal and anchor the pill to it. Dinner usually, because they’re more likely to eat enough at dinner than breakfast. Same chair, same time, same routine. The drug doesn’t care if you take it morning or night, only that there’s real food in your stomach when you do.

How it stacks against Vraylar and Caplyta

Vraylar (cariprazine) is the closest comparator. Also bipolar depression-approved, also reasonably weight-neutral, also has akathisia as the main early side effect. Vraylar has a much longer half-life, which means side effects take longer to clear when you stop. It doesn’t require food. That’s a real advantage for patients who can’t reliably eat at the same time every day, which honestly is a lot of people with bipolar disorder.

Caplyta (lumateperone) is the newer one, approved for bipolar I and II depression. Similar weight profile, low akathisia rate (lower than lurasidone or Vraylar in head-to-head-ish comparisons), no food requirement. The downside is it’s still on patent, runs about $1,500 a month without insurance, and a lot of plans don’t cover it without a prior auth fight.

Lurasidone went generic in 2023. Pre-generic, it was running $1,300 a month and basically nobody could afford it without good insurance. Now it’s $30-80 a month cash at most pharmacies. That cost shift changed who I prescribe it to. It used to be a drug for people with great insurance. Now it’s a drug for anybody who can commit to eating dinner.

The honest decision tree: if your patient eats consistently and can tolerate akathisia management, lurasidone first because it’s cheap and the data is solid. If they can’t eat reliably, Vraylar. If money isn’t an issue and they want the lowest side-effect burden, Caplyta. If they need the sedation, Seroquel still has a role despite everything.

What I tell patients on day one

Three things, in this order. Eat real food when you take it, every single day, no exceptions. If you feel restless or jittery in the first two weeks, call me before you stop the drug. Give it six weeks before you decide whether it’s working, because at week two you’re going to feel the side effects without the benefit and that’s a common quitting point.

The patients who do well on lurasidone are the ones who treat the food rule like part of the prescription instead of a footnote. The ones who fail are almost always the ones who weren’t told clearly enough, or who heard it once at the pharmacy and forgot. A drug that requires behavioral consistency is only as good as the consistency. That’s the whole story with this one.