Bipolar Treatment

Bipolar treatment goes wrong in the first appointment more often than in any later one. If the diagnosis is wrong, every drug after it is wrong. And bipolar gets missed both ways. Plenty of people who actually have bipolar II spend a decade on antidepressants for what looks like recurrent depression. Plenty of people who are just moody, or hung over, or trauma-reactive, get told they have bipolar and started on lithium when they don’t need it.

The Hollywood version is someone awake for four days, spending the rent on a sports car, flying to Vegas to start a band. That’s bipolar I in a full manic episode. It’s real, and it’s the loud edge of a much wider category. Most of what walks into clinic is quieter. Someone who’s spent three weeks barely getting out of bed, then a week or two of feeling weirdly productive and not needing much sleep, and doesn’t quite connect the two states because the productive part felt like finally being normal.

Bipolar I, II, and cyclothymia are different diseases that share a name

Bipolar I requires at least one full manic episode. Manic, not just energetic. A week or more (or any duration if it landed you in a hospital) of elevated or irritable mood plus enough of the other stuff that life comes off the rails. Racing thoughts, no sleep and not missing it, pressured speech, grandiosity, impulsive spending or sex, sometimes psychosis. The depressions in bipolar I exist and are bad, but the manias define the diagnosis.

Bipolar II is the one that gets missed. The high side is hypomania, same cluster of symptoms but milder and shorter, four days or more, without psychosis, without hospitalization. The depressions are usually long and severe. So patients show up describing their depressions, the hypomanias get remembered as “the good weeks,” and the antidepressant gets prescribed.

Cyclothymia is the low-grade version. Two years of mood swinging between sub-threshold hypomanic and depressive symptoms, never quite hitting full episodes, but never really stable. People with cyclothymia often feel like their personality is the problem because the swings are mild enough that nothing ever quite “happens.” It still responds to mood stabilizers, and treating it matters because cyclothymia frequently graduates into bipolar I or II if left alone.

I had a woman come in last spring, late 30s, on her fourth SSRI in six years. Each one worked for about three weeks and then either pooped out or made her “wired and weird.” Nobody had ever asked about the wired-and-weird weeks. She wasn’t treatment-resistant. She was bipolar II with antidepressant-induced hypomania every time someone reached for the prescription pad.

Mood stabilizers, ranked roughly by what actually works

Lithium is still the gold standard. It’s been around since the 1950s and seventy years of data later it’s still the only mood stabilizer with a clear signal for reducing suicide independent of its mood effects. Typical dosing lands in the 600 to 1200 mg/day range, titrated to a blood level usually between 0.6 and 1.0 mEq/L. The price of admission is monitoring. Levels, kidney function, thyroid, every few months at first and then a couple times a year forever. Lithium is famously narrow-therapeutic-index, so the line between “working” and “toxic” is closer than most drugs. People get scared off by the monitoring and miss that the drug itself is remarkably well-tolerated once you find the dose.

Valproate (Depakote) is the workhorse for acute mania and rapid-cycling presentations. Faster on than lithium, less monitoring fuss, worse long-term partner. Weight gain, hair thinning, liver enzyme bumps, and absolutely off the table in anyone who could become pregnant unless you’ve had a long conversation about teratogenicity. The neural tube defect data on valproate is one of the clearest “don’t do this” signals in psychiatry.

Lamotrigine is the one for the depressive side of bipolar II. Quiet on mania, so not a first pick for someone with prominent highs, but for the bipolar II patient whose problem is mostly the depressions, it’s often the cleanest option. The catch is the titration. You start at 25 mg and creep up over six weeks because of the Stevens-Johnson rash risk. Rush it and you can land someone in a burn unit.

Carbamazepine still has a niche for mixed states and people who haven’t responded to the others. The drug-drug interactions are a nightmare because it induces half the cytochromes you care about, so I reach for it third or fourth.

The treatment goal is a mood steady enough that it stops costing you jobs, relationships, and years.

The atypicals filled a real gap, especially for bipolar depression

For decades the depressed side of bipolar was a mess to treat. Mood stabilizers alone often weren’t enough, and antidepressants alone were risky. Then a handful of atypical antipsychotics started accumulating real evidence for bipolar depression specifically, and the playbook changed.

Lurasidone (Latuda) has a clean indication for bipolar depression and a reasonable metabolic profile. It has to be taken with at least 350 calories of food or it doesn’t absorb properly, which trips people up constantly. Quetiapine (Seroquel) is the older option in the same lane, effective but sedating and metabolically expensive over years. Caplyta (lumateperone) is the newer entry, similar efficacy with a friendlier side effect profile on weight and prolactin. It’s earned a spot in the rotation, particularly for patients who couldn’t tolerate Latuda’s GI issues or quetiapine’s sedation.

None of these are substitutes for a mood stabilizer in bipolar I. They’re partners. Lithium plus an atypical is one of the most common stable long-term regimens I write.

Antidepressants in bipolar are a controlled-handling problem

You can use them. Sometimes you have to. But putting a bipolar patient on an SSRI or SNRI without a mood stabilizer underneath is one of the most reliable ways to make things worse. The risk is induced mania, mixed states, and rapid cycling, and it doesn’t always show up immediately. I’ve seen people do fine for two months on bupropion and arrive in clinic in week ten not having slept in four days.

If an antidepressant is going on board, it goes on top of a mood stabilizer, at a lower dose than you’d use in unipolar depression, and with somebody in the patient’s life who’s been told what to watch for. SNRIs and tricyclics are higher-risk for the switch than SSRIs. The order matters: stabilizer first, antidepressant second, never the other way around.

The rest of the picture: sleep, schedule, and the therapy that actually moves the needle

Bipolar is the diagnosis where lifestyle stuff stops being a wellness suggestion and becomes part of the prescription. Sleep deprivation is a known trigger for manic episodes. Shift work, jet lag, a new baby, finals week, any of it can flip a stable patient. The schedule itself, what time you wake up, eat, work, and sleep, matters more than for almost any other psychiatric condition.

This is what IPSRT, interpersonal and social rhythm therapy, was built around. It treats your daily rhythm as a clinical variable. Patients log wake times, meal times, social contact, and sleep onset, and over weeks the rhythm gets steadier. The data on IPSRT plus medication is solid enough that I bring it up with every newly diagnosed bipolar patient, even the ones who roll their eyes at charting their wake-up time. Family-focused therapy matters more than people expect too, because the people who live with a bipolar patient are usually the first to notice the early signs of an episode.

Bipolar is a long game. Most people will be on something for the rest of their lives. The patients who do best stop fighting that fact somewhere around year two or three, settle into a regimen they can actually live with, get their sleep boring, and accept that “stable” is the goal. The ones who chase being cured tend to stop their meds during a good stretch, and the good stretch ends about six weeks later in a way that costs them a job, a relationship, or worse.