Antidepressant-Associated Manic Switch: Missed Versus Anticipated Risk
Two patients, same underlying risk. One received antidepressant monotherapy with no hypomania screen and was unmasked as bipolar by the switch itself; the other had known bipolarity and established mood-stabilizer coverage before any antidepressant was added.
K.T. is a 29-year-old woman, a night-shift labor and delivery nurse who has been living alone in a studio apartment since a difficult breakup four months ago. She went to her primary care doctor six weeks ago with low energy, poor sleep, tearfulness, and difficulty concentrating at work, scored in the moderately-severe range on a PHQ-9, and was started on sertraline for major depressive disorder — no formal screen for a history of hypomania was documented in that visit, which is typical of a busy primary care encounter built around a nine-item depression questionnaire rather than a structured mood-disorder interview. She was titrated to 100mg three weeks ago, and her depressive symptoms genuinely improved for the first two of those weeks, which is part of what makes tonight confusing to everyone around her. Her sister brought her to the emergency department last night after finding her awake at 4 a.m. reorganizing her entire kitchen, talking rapidly about a business plan to start a home-birth doula service, having already emailed three former coworkers about investing, and irritable to the point of shouting when questioned about any of it.
On direct questioning tonight, K.T. also describes a two-week stretch in nursing school at age 22 with almost identical features — barely sleeping, finishing a semester's worth of coursework in days, feeling unstoppable, spending well beyond her student budget on new clothes — that she never mentioned to a doctor because it felt more like a productive stretch than an illness at the time, and because it resolved on its own within a few weeks. Read against that history, the sertraline did not create a new problem so much as unmask an old one: an antidepressant given without a bipolarity screen, in a patient whose true baseline was likely bipolar II depression rather than unipolar depression, producing exactly the switch that screening exists to prevent. The early improvement she felt in week one is consistent with this reading too — an antidepressant response curve that keeps climbing past ordinary euthymia into a hypomanic register is a recognized pattern, not a coincidence layered on top of an unrelated new problem.
At the bedside, overnight
My working differential when she arrived was activation or akathisia from the sertraline rather than a true mood switch — pressured speech and irritability three weeks into an SSRI can look similar on the surface, and akathisia doesn't require a bipolar diagnosis to explain or treat.
The history she gave once we asked directly changed that read. Akathisia doesn't come with grandiose business plans and a decreased need for sleep that started before the agitation did.
This is an antidepressant-induced hypomanic switch in a patient whose real underlying diagnosis is bipolar II, not unipolar depression — the college episode she just described meets criteria for hypomania on its own, it was simply never brought to anyone's attention. The sertraline needs to stop, and she needs a mood stabilizer, not just a taper.
I want to be specific about what failed here: not the prescribing itself, but the absence of a hypomania screen before an antidepressant was started for a first depressive episode. That is a primary-care-level gap, not a rare edge case.
Mechanistically this fits — SSRIs increase net monoaminergic tone, and in a substrate already prone to mania that tone shift can precipitate a switch that would not occur in a true unipolar patient. The switch itself is the diagnostic test result here: it is strong retrospective evidence for bipolarity that the PHQ-9 alone could never have provided.
Going forward, she should not be treated with an antidepressant again without mood-stabilizer coverage in place first — the same logic Case B in this pair was built around from the start.
Sertraline discontinued, divalproex sodium ER started at 500mg twice daily, and lorazepam used briefly for agitation and sleep. Her diagnosis was revised from major depressive disorder to bipolar II disorder, and she was referred to outpatient psychiatry with an explicit note that any future antidepressant trial requires mood-stabilizer coverage first.
R.O. is a 35-year-old man, an avid distance runner training for his fourth marathon, who lives with his longtime partner and has a well-documented bipolar II disorder diagnosis, confirmed four years ago after two clearly characterized hypomanic episodes — one during a period of intense work travel, one following a period of poor sleep during his partner's medical residency. He has been maintained on lamotrigine 200mg daily for the past three years with good mood stability and no further hypomanic episodes, and has been reliable about his twice-yearly psychiatry follow-ups the entire time. He presents today with three weeks of low energy, anhedonia, difficulty finishing long runs he used to complete easily, and a flat, joyless quality to time with his partner — a depressive episode, his first since starting lamotrigine, without any features suggesting an emerging mixed or hypomanic state layered underneath it.
The question is whether to add an antidepressant, and it carries the same underlying risk this pair's Case A illustrates directly: antidepressants can precipitate a hypomanic switch in bipolar spectrum illness, and that risk does not disappear just because a patient is already diagnosed. What is different here is that the risk is anticipated rather than discovered after the fact — his bipolarity is already known, his mood-stabilizer coverage is already established and has already proven adequate for three years, and any antidepressant added now would be layered onto that existing protection and watched for deliberately, rather than given as unmonitored monotherapy to a patient nobody had screened for hypomania in the first place.
At routine follow-up
Bupropion, added onto his existing lamotrigine, is a reasonable next step. It carries one of the lower switch risks among antidepressants because it acts primarily on dopamine and norepinephrine reuptake rather than serotonin, and he already has three years of proven mood-stabilizer coverage in place — this is close to the textbook version of doing it correctly.
I agree with adding it, but "lower switch risk" is not "no switch risk," and I don't want the coverage from lamotrigine to be treated as a reason to relax monitoring rather than a reason the risk is acceptable to take. Case A is a reminder of what an unmonitored switch looks like — the difference here should be vigilance, not indifference.
Concretely: a two-week follow-up call specifically asking about sleep need and racing thoughts, not just a routine refill visit in a month.
Bupropion XL 150mg daily started, added to his unchanged lamotrigine 200mg. A two-week phone check specifically screening for decreased sleep need, racing thoughts, and irritability was scheduled, rather than deferring to his routine one-month follow-up.