The seasoned approach
The layered approach to refractory depression. The discipline is to escalate systematically rather than cycle through serial monotherapy. Each layer addresses a specific clinical scenario and adds defined benefit; the patient on adequate layered treatment is biologically different from the patient on serial SSRI substitutions.
Layer 1 — First — confirm adequacy of trials
An "adequate trial" requires sufficient dose for sufficient duration. For SSRIs and SNRIs: maximum tolerated dose (or label maximum) for at least 6–8 weeks before declaring failure. Most "TRD" patients have had multiple inadequate trials — undertreated, undertreated duration, or mistakenly switched at the partial-response point. Confirm adequacy before escalating; if not adequate, optimize first.
Layer 2 — Augmentation — lithium remains the gold standard
Lithium 600–900 mg targeting levels of 0.4–0.8 mEq/L (lower than the bipolar range) has the strongest evidence base for SSRI augmentation, supports a small but real reduction in suicide risk, and adds neuroprotective benefit (Stage 16). Limitations: monitoring burden, thyroid and renal effects with long-term use, narrow therapeutic index. The augmentation that has more evidence than any other and the lowest cost.
Layer 3 — T3 augmentation — frequently effective, underused
Liothyronine 25–50 mcg daily added to antidepressant. The STAR*D data and earlier work support meaningful response rates in TRD. Particularly useful when subclinical hypothyroidism or low T3 is present, but effective in many euthyroid TRD patients as well. Monitor TSH; usually well-tolerated. The augmentation that needs the workup to identify the strongest candidates.
Layer 4 — Atypical antipsychotic augmentation
Aripiprazole 2–10 mg, brexpiprazole 1–3 mg, cariprazine 1.5–3 mg, quetiapine XR 150–300 mg. Strong efficacy data; significant side effects vary by agent — aripiprazole/brexpiprazole/cariprazine more metabolically favorable, quetiapine more sedating and metabolically problematic. The augmentation with the most consistent efficacy data but the highest side-effect cost for chronic use.
Layer 5 — Antidepressant combination — bupropion or mirtazapine
Bupropion 150–300 mg added to SSRI: useful when sexual dysfunction, fatigue, or weight concerns are present; modest evidence for additive benefit. Mirtazapine 15–45 mg added to SSRI: classic "California rocket fuel" combination; useful when sleep, appetite, and anxiety are residual concerns. Both are reasonable augmentations with lower side-effect burden than antipsychotic augmentation.
Layer 6 — Move to interventional treatment
When augmentation has been adequately tried and remission is not achieved, move to interventional treatment — TMS, ECT, ketamine — rather than cycling through more medication combinations. The delay between TRD declaration and interventional treatment is one of the most costly delays in psychiatric care. See Stage 15 for the interventional landscape.
Special situations
- TRD with prominent anxiety: Augment with low-dose pregabalin (50–150 mg bid) or buspirone (15–30 mg bid) for anxiety-component; reduces benzodiazepine pressure. Address the anxiety as a residual symptom worth specific intervention.
- TRD with chronic pain: SNRIs (duloxetine, venlafaxine) plus low-dose TCA (nortriptyline 25–75 mg) often produce response where either alone fails. Address the pain as part of the depression treatment.
- TRD with severe insomnia: Mirtazapine combination or trazodone augmentation; address CBT-I for behavioral foundation. Sleep is often the residual symptom that maintains the depression.
- TRD in perimenopausal women: Hormonal evaluation and consideration of hormonal intervention (Stage 19) alongside antidepressant. The window of perimenopausal vulnerability often resolves with hormonal support that pure antidepressants cannot match.
Generally avoid
- Cycling through SSRIs as the augmentation strategy — substituting sertraline for fluoxetine for citalopram without escalation produces serial inadequate treatment and the kindling cost of unremitted depression.
- Adding more medications without removing what is not working — the patient on five psychiatric medications who is still depressed is on a polypharmacy regimen that may be making things worse; audit and simplify when augmentation is not working.
- Delaying ECT or interventional treatment past the point where it is the right answer — particularly in patients with melancholic features, psychotic features, suicidality, or catatonia, where ECT remains the most effective treatment and delay produces real cost.
- Treating "TRD" without confirming adequate trial history — the patient with three "failed" SSRI trials at 25 mg has not actually had three failed trials. Confirm adequacy before escalation.
The chief-resident note: Most patients labeled TRD have had inadequate treatment, not refractory treatment. Most augmentation strategies are underused because they require ongoing monitoring and dose optimization that the standard 15-minute visit does not support well. Build the practice to support layered treatment — longer initial visits, structured PHQ-9 tracking, willingness to move through layers rather than parking the patient on inadequate treatment. The clinical work of getting the recurrent-depression patient to durable remission is one of the highest-leverage interventions in psychiatry; treat it accordingly.
The layered approach to refractory depression is the discipline that distinguishes effective TRD care from serial monotherapy cycling. Most patients labeled "treatment resistant" have actually had a series of inadequate trials — undertreated doses, premature switches at the partial-response point, missed adequacy criteria — and the first move in the TRD workup is to confirm whether prior trials actually qualified as adequate. Adequacy requires sufficient dose for sufficient duration: for SSRIs and SNRIs, typically maximum tolerated dose (or label maximum) for at least 6–8 weeks. The patient with three "failed" SSRIs at starting doses for 4 weeks each has not been treated; they have been sampled.
Augmentation logic prioritizes adding to a partially-effective treatment over substituting it. When a current antidepressant has produced partial response, adding a second agent often produces durable remission more reliably than switching to a third antidepressant in the same class. Lithium augmentation has the strongest evidence base — meta-analyses consistently show response rates of 40–50% in lithium-augmented TRD, with the added benefit of suicide risk reduction and neuroprotection (Stage 16). T3 augmentation is similarly evidence-based and frequently underused. Atypical antipsychotic augmentation (aripiprazole, brexpiprazole, cariprazine, quetiapine) has the most consistent efficacy data among recently studied augmentation strategies but carries significant chronic side-effect cost.
Antidepressant combination strategies are useful where mechanism complementarity makes sense. Bupropion added to SSRI addresses sexual dysfunction, fatigue, and concentration deficits while adding noradrenergic and dopaminergic activation to serotonergic effect. Mirtazapine added to SSRI ("California rocket fuel") addresses sleep, appetite, and anxiety while adding noradrenergic and serotonergic effects through alpha-2 and 5HT2 antagonism. Both combinations have favorable side-effect profiles compared to antipsychotic augmentation for chronic use. The combination approach is particularly useful when residual symptoms are specific (sexual dysfunction, sleep, anxiety) and the augmentation can target them directly.
The case for moving to interventional treatment — TMS, ECT, ketamine — exists when augmentation has been adequately tried and remission has not been achieved. One of the most costly errors in TRD care is the prolonged delay between TRD declaration and interventional referral; patients spend years on ineffective polypharmacy when they could have achieved remission with TMS or ECT. The clinical judgment includes severity, safety, prior response history, patient preference, and access. Severe TRD with melancholic features, psychotic features, suicidality, or catatonia is where ECT remains the most effective treatment and where delay produces meaningful cost. TMS is reasonable for moderate-to-severe TRD without those features; ketamine has emerged as an option with specific clinical considerations (Stage 10.4).
The special situations within TRD modify the strategy. TRD with prominent anxiety often benefits from low-dose pregabalin or buspirone augmentation. TRD with chronic pain often responds to SNRI plus low-dose TCA combination. TRD with severe insomnia benefits from mirtazapine or trazodone augmentation plus CBT-I. TRD in perimenopausal women may respond preferentially to hormonal intervention (Stage 19) alongside or instead of further antidepressant manipulation. The clinical move is to identify which special situation applies and to target the augmentation accordingly, rather than running through standard augmentation in order.
The discipline of layered treatment is what separates effective TRD care from cycling. Each layer addresses a specific clinical scenario and adds defined benefit; the patient on adequate layered treatment is biologically different from the patient on serial substitutions. Most patients can be moved from TRD to remission through systematic application of the layers — adequacy first, augmentation second, interventional third — with the workup at each step refining the choice. The clinician's task is to build the practice that supports this work — longer initial visits, structured PHQ-9 tracking, willingness to escalate rather than park, and follow-through on the layered approach. Most TRD is treatable; the trick is to actually treat it.
The anchor
The layered approach to TRD: confirm adequate trials, then augment systematically (lithium, T3, atypical antipsychotic, antidepressant combination), then move to interventional treatment if needed. Substitution alone produces inadequate treatment; combination and augmentation produce remission.