TCA Selection in TRD with Suicide Risk
Tricyclics remain among the most effective drugs available for treatment-resistant depression — and among the most lethal in overdose. Two patients with the same failed history put that tradeoff on opposite ends of the same decision.
D.M., a 45-year-old man who has taught high school chemistry for eighteen years, is six months into a depressive episode that has resisted three adequate antidepressant trials — fluoxetine for ten weeks, then venlafaxine titrated to 225mg over three months, then sertraline augmented with aripiprazole 5mg, each carried long enough to judge fairly before being abandoned for lack of response. His health is otherwise unremarkable: well-controlled GERD on omeprazole, no cardiac history, no smoking, and no family history of early cardiac disease that would change how the team weighs a tricyclic. This is his third lifetime depressive episode; the first, in his late twenties, resolved on sertraline alone, and the second, in his late thirties, responded to venlafaxine after a slower start — neither ever involved suicidal ideation of any kind.
His most recent PHQ-9 scored 19, essentially unchanged over the past two monthly visits, but item 9 — thoughts of self-harm — has read zero at every documented visit across all three episodes, and his wife, a nurse who manages the household's medications and has navigated both prior episodes alongside him, describes this one as exhausting rather than frightening; he still teaches full-time, still coaches the robotics club on weekends, and has never withdrawn from either. Nortriptyline is being considered as the next step: tricyclics carry a real efficacy edge over SSRIs and SNRIs in melancholic, treatment-resistant presentations, evidence that predates most of what has already been tried on him. Every prescriber in the room remains aware that a several-week supply of a tricyclic is potentially lethal in overdose — that fact shapes how the drug gets used even when nothing in his history suggests it will need to.
At the medication-management visit
Nortriptyline has the best-characterized therapeutic window of the tricyclics — plasma levels correlate with response in a way most antidepressants never established, and the 50-150 ng/mL range gives us something concrete to titrate against rather than guessing at dose.
A baseline EKG before starting is standard given the sodium-channel effects at supratherapeutic levels, not because anything here suggests cardiac risk — it's the same check I'd want for any patient starting a TCA regardless of history.
Nothing in his history changes how we dispense this. A one-month supply, plasma-level-guided titration, and the EKG the pharmacologist wants is the same regimen I'd use for any patient with his risk profile — his wife already manages his medications, which is real supervision most patients starting a TCA don't have.
Started nortriptyline 25mg at bedtime with a baseline EKG and plasma-level-guided titration toward the standard therapeutic window, dispensed as a routine one-month supply.
The team's reasoning here is the direct counterpoint to Patient B below: the same drug, the same disease, and a dispensing plan built on an actual risk assessment rather than a reflexive restriction applied to every TCA prescription regardless of who is receiving it.
S.T., a 39-year-old single mother of two who was laid off from her office-manager job four months ago, has also failed three adequate antidepressant trials for this recurrent episode — the same fluoxetine, venlafaxine, and sertraline-plus-aripiprazole sequence tried in Patient A, at comparable doses and durations, with the same lack of response. She has a history of intermittent asthma, well-controlled with an albuterol inhaler she rarely needs, and quit smoking two years ago after her second child was born — nothing in her general medical history complicates a tricyclic on its own. Three years ago, during a prior depressive episode compounded by a difficult divorce, she survived a suicide attempt by intentional overdose on a different medication, requiring a two-day medical admission; for the past several weeks, coinciding with the layoff and the resurfacing of the same financial pressures that surrounded her prior attempt, she has described persistent, passive thoughts of not wanting to be alive, without a specific plan.
She lives alone with her two children, manages her own medications, and has no other adult in the household to supervise dosing or secure a new prescription the way D.M.'s wife does for him. Nortriptyline would be pharmacologically reasonable for the same reasons it was reasonable for Patient A — it may be the most effective option left on her list — but a full month's supply of any tricyclic in the home of a patient with a prior overdose attempt and current passive ideation is a genuinely different risk than the identical prescription written for someone with no such history. The disagreement in the room is not about whether nortriptyline could help her; it is about whether that benefit can be delivered without handing her the means for another attempt during the highest-risk weeks of starting a new medication.
At the medication-management visit
I don't want to take an effective option off the table just because of what happened three years ago — that risks under-treating her depression, which is its own real danger, and TCAs have genuine efficacy advantages we'd be discarding for a risk we can manage through dispensing rather than avoidance.
The dispensing math is what changes my answer here, not the pharmacology. A ten-to-twenty-day supply of amitriptyline is a recognized fatal dose in a patient without tolerance; nortriptyline has a somewhat better margin but is still firmly in that category. The efficacy case for a TCA in her doesn't disappear — it just can't be delivered as a routine take-home month's supply.
Weekly pharmacist-dispensed quantities, or asking a family member outside her household to hold and administer doses, keeps the option genuinely open without keeping a lethal quantity in a home with no supervision.
I'd go further and ask whether we should be reaching for a TCA at all as our next step for her specifically, rather than solving the access problem after choosing the drug. Esketamine or an ECT referral gets her an option without ever putting a cardiotoxic-in-overdose medication in her hands, which is a cleaner answer than managing risk around a prescription she may still stockpile a dose or two at a time. Worth naming precisely: esketamine carries a second, separate FDA indication for depressive symptoms in MDD with acute suicidal ideation or behavior, established in the ASPIRE trials — that is closer to her actual presentation than the treatment-resistance indication we would otherwise be citing.
The team held nortriptyline as pharmacologically appropriate but clinically undeliverable in a standard prescription, and instead began an esketamine referral while safety planning was updated with her outpatient team.
Nobody in the room disputed that a TCA might still be tried later, and specifically only through restricted, pharmacist-controlled dispensing if esketamine access or response falls short — the disagreement was entirely about sequencing and delivery, never about whether her overdose history and current ideation were disqualifying on their own.