TRD Augmentation: Antipsychotic vs. Lithium vs. T3 vs. Switch
Aripiprazole, lithium, T3, and switching to a new agent are all reasonable next steps after two failed antidepressant trials — none has ever shown clear superiority over the others in a real head-to-head comparison.
J.O., a 51-year-old union electrician with well-controlled hyperlipidemia on atorvastatin and no other chronic illness, is nine months into a depressive episode that has responded only partially to treatment. Sertraline, tried first for ten weeks at 150mg, produced no meaningful change. Escitalopram, the current agent, brought his PHQ-9 down from 24 to 15 over ten weeks at 20mg and has now plateaued there for the past month — a real but incomplete response, not a failure, which is its own distinct clinical situation from where he started.
He still works full shifts, still coaches his son's youth hockey team on weekends, and describes the current state honestly as 'functional but flat' rather than in crisis — his depression has stopped worsening without genuinely lifting. Nothing in his history — no renal disease, no thyroid dysfunction, no bipolar features on careful screening — rules out any of the standard next steps.
That's the actual shape of the decision in front of the team: augment with an atypical antipsychotic, augment with lithium, augment with T3, or switch to a different antidepressant entirely. Each has real trial evidence behind it, and none has ever demonstrated clear superiority over the others in a genuine head-to-head comparison — STAR*D, the largest real-world effectiveness trial available, produced broadly comparable remission rates for switching and for augmentation at this stage — though its equipoise-stratified design let patients decline whole strategy arms, so it never randomized switch against augment head-to-head and cannot settle the comparison on its own. The disagreement in the room isn't about which option is unsupported; it's about which tradeoffs matter most for a patient who is stable enough to weigh them carefully rather than needing the fastest option regardless of its downsides.
At the four-week medication-management visit
Aripiprazole augmentation has the most directly applicable evidence here — it's FDA-approved specifically as an adjunct for major depression with an inadequate response to antidepressant therapy, which is exactly his situation, and the effect size in the pivotal trials is real even at low augmenting doses like 2-5mg.
Metabolic monitoring is a genuine cost, but it's a manageable one, and he's already engaged enough in his own care — coaching, working full shifts — that I'd expect reasonable follow-through on the labs.
I'd put real weight on lithium instead. It has the longest track record as an antidepressant augmenting agent of anything on this list, decades of evidence in mood disorders, including an association with reduced suicide risk — though that signal comes largely from maintenance treatment across mood disorders rather than from unipolar augmentation trials, and a large randomized trial in veterans did not reproduce it, and his labs — renal function, thyroid — are clean going in, which is the main thing that would complicate starting it.
The monitoring burden is real and ongoing, not a one-time check, and that's a legitimate reason to prefer aripiprazole if adherence to routine lab draws is a concern — it isn't obviously a reason to prefer it on efficacy grounds.
I'd actually argue for switching before augmenting anything. He's only had one truly adequate SSRI trial plus one partial response — nothing in STAR*D clearly favors augmentation at this step, and its design can't adjudicate that comparison directly anyway, and a switch to an SNRI would add a norepinephrine mechanism he hasn't tried yet, without asking him to add a second daily medication and its own monitoring on top of an incomplete response to the first.
Aripiprazole 2mg was added to his existing escitalopram 20mg, with lithium and T3 both explicitly named as the next steps in that order if augmentation proves inadequate, and a metabolic panel scheduled at four weeks.
The team's documentation was explicit that this was a preference among reasonable options rather than a resolved clinical question — his engagement with routine care tipped the choice toward aripiprazole's monitoring profile, not new evidence that it outperforms lithium, T3, or a switch for a patient in his position.