Augmentation Strategies in Psychiatry: When First-Line Treatment Fails — Lithium, Atypical Antipsychotics, and Combination Approaches

In psychiatric practice, first-line pharmacotherapy fails to produce adequate response in a substantial proportion of patients. The landmark STAR*D (Sequenced Treatment Alternatives to Relieve Depression) trial — the largest prospective study of depression treatment ever conducted — demonstrated that only approximately 37% of patients with major depressive disorder (MDD) achieved remission with an initial adequate trial of citalopram. Cumulatively, even after four sequential treatment steps, roughly one-third of patients did not achieve remission. These findings underscore a clinical reality: treatment resistance is not exceptional — it is common.

Augmentation refers to the addition of a second agent to an existing antidepressant (or other primary medication) that has produced partial but insufficient response, with the goal of enhancing therapeutic efficacy without switching the base treatment. This is distinct from combination therapy, which involves two agents from the same pharmacological class (e.g., two antidepressants), and from switching, which involves replacing one medication with another. The distinction matters clinically because augmentation preserves whatever partial benefit the first agent has achieved and adds a mechanistically distinct pharmacological action.

This article provides a detailed, evidence-based review of the major augmentation strategies used in psychiatry — with particular focus on lithium augmentation, atypical antipsychotic augmentation, and key combination approaches. We examine the neurobiological mechanisms, comparative efficacy data from randomized controlled trials and meta-analyses, specific protocols, predictors of response, side effect profiles, and considerations for special populations. The primary focus is on treatment-resistant depression (TRD), defined here as failure to achieve remission following at least two adequate antidepressant trials, though we address augmentation in other conditions where evidence supports it.

Before initiating augmentation, clinicians must confirm that treatment resistance is genuine rather than apparent. Pseudo-resistance — inadequate response due to subtherapeutic dosing, insufficient trial duration, poor adherence, unrecognized comorbidity (especially substance use, personality disorders, or medical conditions like hypothyroidism), or pharmacokinetic factors (e.g., ultrarapid CYP2D6 metabolism) — must be systematically excluded. The Thase and Rush staging model and the Massachusetts General Hospital Staging Method both provide frameworks for quantifying degree of treatment resistance, with higher stages indicating more failed trials and worse prognosis.

The decision to augment versus switch depends on several clinical factors:

• Augmentation is preferred when the patient has achieved partial response (≥25% but <50% reduction in symptom severity scores) on the current agent, suggesting some mechanistic engagement that can be built upon.
• Switching is preferred when the patient has shown no response or has intolerable side effects from the current agent.
• Combination therapy (adding a second antidepressant) occupies a middle ground and is sometimes preferred when the clinician wants to broaden monoaminergic coverage without introducing a non-antidepressant agent.

Per the DSM-5-TR, treatment-resistant depression is not a formal diagnostic category but is widely operationalized in clinical research as failure to respond to at least two antidepressant trials of adequate dose and duration (typically ≥6–8 weeks at therapeutic dose). The ICD-11 does not include a specific code for treatment resistance but allows specifiers for severity and course that capture chronicity.

Major clinical guidelines — including those from the American Psychiatric Association (APA), the Canadian Network for Mood and Anxiety Treatments (CANMAT), the National Institute for Health and Care Excellence (NICE), and the British Association for Psychopharmacology (BAP) — all include augmentation as a core strategy at the point of inadequate first-line response, though they differ in which agents they prioritize.

Mechanism of Action

Lithium's augmentation effects are mediated through multiple neurobiological mechanisms that complement and extend serotonergic and noradrenergic antidepressant actions. At the molecular level, lithium inhibits glycogen synthase kinase-3β (GSK-3β), a serine/threonine kinase implicated in neuroplasticity, circadian regulation, and apoptotic signaling. GSK-3β inhibition upregulates brain-derived neurotrophic factor (BDNF) and promotes neurogenesis in the hippocampal dentate gyrus — processes increasingly recognized as central to antidepressant response. Lithium also inhibits inositol monophosphatase, modulating phosphoinositide signaling cascades involved in neuronal excitability. Additionally, lithium enhances serotonergic neurotransmission — increasing presynaptic tryptophan uptake, serotonin synthesis, and serotonin release — which provides a direct pharmacodynamic rationale for augmenting SSRI and SNRI antidepressants.

At the systems level, lithium modulates glutamatergic transmission, enhances GABAergic inhibition, and influences hypothalamic-pituitary-adrenal (HPA) axis activity, contributing to neuroprotective and mood-stabilizing effects that extend beyond monoamine modulation alone.

Efficacy Data

Lithium augmentation of antidepressants is the most extensively studied augmentation strategy in psychiatry, with evidence spanning four decades. A pivotal meta-analysis by Nelson and Papakostas (2009), published in the American Journal of Psychiatry, analyzed 10 randomized controlled trials (total N = 269) and found that lithium augmentation was significantly more effective than placebo augmentation, with a pooled response rate of approximately 41.2% versus 14.4% for placebo (odds ratio ~3.11, 95% CI 1.80–5.37). The number needed to treat (NNT) was approximately 5. An earlier meta-analysis by Bauer and Döpfmer (1999) reported broadly concordant findings across 9 RCTs, with response rates for lithium augmentation roughly three times those of placebo.

Most of the pivotal RCTs studied lithium augmentation of tricyclic antidepressants (TCAs). Evidence supporting lithium augmentation of SSRIs and SNRIs specifically is more limited but generally positive, with open-label studies and smaller RCTs showing comparable effect sizes. The LiTMUS (Lithium Treatment – Moderate Dose Use Study) trial, a practical randomized trial in treatment-resistant depression, found that adjunctive lithium added to optimized treatment had modest but non-significant advantages over optimized treatment alone, though this study was not powered to detect small effects and used a pragmatic design that diluted group differences.

Protocol

Clinical guidelines generally recommend targeting a serum lithium level of 0.5–0.8 mEq/L for augmentation, which is lower than the 0.8–1.2 mEq/L range typically targeted for bipolar mania prophylaxis. Starting doses are usually 300 mg once or twice daily, titrated based on serum levels drawn 12 hours post-dose after 5 days of stable dosing (approximately 5 half-lives). Response to lithium augmentation is often observed within 2–4 weeks, which is notably faster than the typical antidepressant onset. Some older studies reported rapid responses within 48 hours to 1 week, though a minimum trial of 4 weeks at therapeutic levels is recommended before concluding non-response.

Monitoring Requirements

Lithium requires baseline and ongoing monitoring of:

• Renal function (BUN, creatinine, eGFR) — lithium can cause chronic tubulointerstitial nephropathy
• Thyroid function (TSH, free T4) — lithium-induced hypothyroidism occurs in approximately 5–15% of patients
• Serum lithium levels — initially every 1–2 weeks during titration, then every 3–6 months at steady state
• Calcium and parathyroid hormone — hyperparathyroidism occurs in approximately 10% of chronic users
• ECG in patients over 40 or with cardiac risk factors

The narrow therapeutic index (therapeutic range 0.5–1.2 mEq/L, toxicity risk above 1.5 mEq/L) demands clinical vigilance, particularly regarding drug interactions (NSAIDs, ACE inhibitors, and thiazide diuretics reduce lithium clearance) and dehydration states.

Mechanism of Action

Atypical (second-generation) antipsychotics augment antidepressant action through a combination of receptor interactions that extend beyond dopamine D2 blockade. The augmentation rationale centers on several key pharmacological properties:

• 5-HT2A antagonism: Blockade of serotonin 5-HT2A receptors disinhibits dopaminergic and noradrenergic projections in the prefrontal cortex, enhancing executive function and motivation circuitry. This is mechanistically synergistic with SSRIs, which increase synaptic serotonin at all receptor subtypes, including 5-HT2A (which mediates some of the emotional blunting and sexual side effects of SSRIs).
• 5-HT1A partial agonism: Aripiprazole and brexpiprazole act as partial agonists at 5-HT1A receptors, facilitating serotonergic neurotransmission in a manner analogous to buspirone augmentation but with greater potency.
• D2 partial agonism (aripiprazole, brexpiprazole, cariprazine): These agents modulate dopaminergic tone rather than producing frank blockade, theoretically restoring dopaminergic function in hypodopaminergic prefrontal circuits implicated in anhedonia and cognitive impairment.
• D3 preferential binding (cariprazine): D3 receptors are concentrated in the mesolimbic system and may be particularly relevant to motivational and reward-processing deficits in depression.
• Norepinephrine reuptake inhibition and α2 antagonism (quetiapine): Quetiapine's active metabolite norquetiapine is a potent norepinephrine transporter (NET) inhibitor, giving it an intrinsic antidepressant mechanism independent of its antipsychotic properties.

FDA-Approved Agents and Efficacy Data

Three atypical antipsychotics have received FDA approval specifically for adjunctive treatment of MDD:

Aripiprazole (Abilify) — Approved 2007. The registration trials (Berman et al., 2007; Marcus et al., 2008; Berman et al., 2009) demonstrated response rates of approximately 32–33% versus 16–17% for placebo when added to ongoing SSRI/SNRI therapy. The pooled NNT from the registration program was approximately 8. Effect sizes (Cohen's d) for MADRS change versus placebo ranged from approximately 0.35 to 0.40 across trials. Typical augmentation doses are 2–15 mg/day, with most patients responding at 5–10 mg/day — substantially lower than the 15–30 mg range used for schizophrenia.

Quetiapine XR (Seroquel XR) — Approved 2009. The pivotal trials demonstrated efficacy at 150 mg/day and 300 mg/day as adjuncts to SSRI/SNRI. Response rates were approximately 52–58% versus 40–47% for placebo, with an NNT of approximately 9–12. Effect sizes (Cohen's d) ranged from approximately 0.30 to 0.44. The 300 mg/day dose had a higher response rate but significantly more sedation and metabolic effects.

Brexpiprazole (Rexulti) — Approved 2015. Registration trials (Thase et al., 2015) showed MADRS improvement of approximately 1.3–1.8 points greater than placebo at 2–3 mg/day, with NNT values of approximately 11–13. Effect sizes were in the modest range (Cohen's d ≈ 0.20–0.29). Brexpiprazole has a lower incidence of akathisia than aripiprazole (~6% vs. ~23% in clinical trials).

Cariprazine (Vraylar) — Approved for adjunctive MDD treatment in 2022. Phase 3 trials demonstrated efficacy at 1.5 mg/day with MADRS score reductions approximately 1.5–2.1 points greater than placebo. Its D3-preferring receptor profile may offer advantages for anhedonia-predominant depression, though head-to-head data comparing it with other atypicals in this subtype are not yet available.

Head-to-Head Comparisons

Direct comparisons between augmentation agents are limited. The most informative network meta-analysis on this topic was conducted by Zhou et al. (2015), published in the Journal of Clinical Psychiatry, which compared augmentation and combination strategies for TRD across 48 RCTs. This analysis found that aripiprazole, lithium, quetiapine, and thyroid hormone (T3) all significantly outperformed placebo, with aripiprazole and quetiapine showing the largest effect sizes among atypical antipsychotics. Lithium and aripiprazole performed comparably in indirect comparisons, though direct head-to-head RCTs between these agents are lacking — a significant gap in the evidence base.

A more recent network meta-analysis by Strawbridge et al. (2019) broadly confirmed these findings, identifying aripiprazole and quetiapine as having the strongest evidence bases among atypical antipsychotic augmenters.

Triiodothyronine (T3) Augmentation

Triiodothyronine (liothyronine, Cytomel) augmentation has been studied since the 1960s. The rationale involves the bidirectional relationship between thyroid function and mood: subclinical hypothyroidism is common in depression, and T3 may enhance noradrenergic and serotonergic neurotransmission by upregulating post-synaptic β-adrenergic receptor sensitivity. T3 also crosses the blood-brain barrier more readily than T4 and modulates gene expression in limbic regions.

The STAR*D Step 3 trial directly compared T3 augmentation (25–50 μg/day) with lithium augmentation in patients who had failed two prior antidepressant trials. Remission rates were 24.7% for T3 versus 15.9% for lithium, though this difference was not statistically significant (p = 0.11). However, T3 was significantly better tolerated, with fewer dropouts. A meta-analysis by Aronson et al. (1996) pooled 8 RCTs (N = 292) and found T3 augmentation approximately twice as likely to produce response as placebo (relative response rate ≈ 2.09). The NNT was approximately 4–5 in these earlier trials, though most studied T3 augmentation of tricyclics rather than SSRIs.

T3 is generally dosed at 25–50 μg/day, is well-tolerated, requires no serum monitoring in most patients, and is inexpensive. It remains underutilized relative to its evidence base, potentially because it falls outside the marketing apparatus that supports atypical antipsychotic augmentation.

Buspirone Augmentation

Buspirone, a 5-HT1A partial agonist, augments SSRI action by desensitizing presynaptic 5-HT1A autoreceptors, thereby enhancing serotonin release. In the STAR*D Step 2 trial, buspirone augmentation (up to 60 mg/day) was compared with bupropion augmentation. Remission rates were 30.1% for bupropion SR versus 29.7% for buspirone — no significant difference. However, bupropion was better tolerated, and buspirone's evidence base from RCTs beyond STAR*D is modest. CANMAT guidelines rate buspirone as a second-line augmentation agent.

Bupropion Combination

Adding bupropion (a norepinephrine-dopamine reuptake inhibitor) to an SSRI or SNRI is one of the most widely used combination strategies in clinical practice, leveraging complementary monoaminergic mechanisms. In STAR*D Step 2, bupropion SR augmentation of citalopram yielded a remission rate of approximately 30%. The CO-MED (Combining Medications to Enhance Depression Outcomes) trial compared escitalopram + placebo, bupropion SR + escitalopram, and venlafaxine XR + mirtazapine. At 12 weeks, remission rates were 38.8%, 38.9%, and 37.7% respectively — no significant differences. At 7-month follow-up, the escitalopram monotherapy group showed slightly better outcomes, challenging the common clinical assumption that combination therapy from the outset is superior.

Mirtazapine Augmentation

Mirtazapine (a noradrenergic and specific serotonergic antidepressant) augments SSRIs via α2-adrenergic antagonism and 5-HT2A/5-HT2C/5-HT3 blockade. The MIR trial (Kessler et al., 2018) randomized patients with TRD to SSRI/SNRI + mirtazapine versus SSRI/SNRI + placebo. At 12 weeks, there was no statistically significant difference in BDI-II scores between groups (mean difference 1.83 points, 95% CI −0.59 to 4.25). This pragmatic UK trial was a significant negative result that dampened enthusiasm for this widely used combination, though effect sizes were in the small-to-modest range and some subgroups may still benefit.

Several novel augmentation approaches have generated substantial recent interest:

Ketamine and Esketamine

Esketamine (Spravato), the S-enantiomer of ketamine administered intranasally, received FDA approval in 2019 for treatment-resistant depression and in 2020 for MDD with acute suicidal ideation. Its mechanism involves NMDA receptor antagonism leading to a rapid surge in glutamate signaling, activation of AMPA receptors, and downstream upregulation of BDNF and synaptogenesis in prefrontal and hippocampal circuits — a fundamentally different mechanism from monoaminergic antidepressants.

In the TRANSFORM trials, esketamine 56 mg or 84 mg intranasal + oral antidepressant produced MADRS score reductions approximately 3.5–4.0 points greater than placebo spray + oral antidepressant at 4 weeks (Cohen's d ≈ 0.30–0.45). Response rates in the SUSTAIN trial were approximately 52–69% in the initial phase. The NNT to prevent relapse (in the SUSTAIN-1 maintenance withdrawal trial) was approximately 5–6. Esketamine requires administration in a certified healthcare setting with 2-hour post-dose monitoring due to risks of dissociation, sedation, and blood pressure elevation.

Psilocybin-Assisted Therapy

Psilocybin, a 5-HT2A agonist that produces profound alterations in consciousness and promotes neural plasticity, has shown promising results in open-label and randomized trials for TRD. A randomized trial by Carhart-Harris et al. (2021) compared psilocybin with escitalopram for MDD and found comparable efficacy at 6 weeks, with psilocybin showing greater improvements on secondary measures including emotional connectedness and wellbeing. However, this was a small trial (N = 59) and the psilocybin group received psychological support that the escitalopram group did not, limiting interpretation. Psilocybin remains a Schedule I substance in the United States and is not FDA-approved, though breakthrough therapy designation has been granted.

Anti-Inflammatory Augmentation

Elevated inflammatory biomarkers (CRP, IL-6, TNF-α) are consistently associated with depression, particularly anhedonic and somatic subtypes. A meta-analysis by Köhler-Forsberg et al. (2019) found that anti-inflammatory agents (NSAIDs, cytokine inhibitors, statins, minocycline) had a pooled effect size of Cohen's d ≈ 0.55 for depressive symptoms when used as augmentation. However, evidence is heterogeneous, and the most promising results are in patients with demonstrably elevated baseline inflammation (CRP >3 mg/L). The anti-inflammatory agent celecoxib (200–400 mg/day) has the most consistent positive augmentation data among NSAIDs, with NNT estimates of approximately 6–8 in selected populations, though cardiovascular risks limit long-term use.

Where Augmentation Works Best

Major Depressive Disorder with partial response: This is the condition with the deepest evidence base. Patients who have achieved some improvement (typically 25–49% symptom reduction) on an SSRI or SNRI are the most appropriate candidates for augmentation. The evidence is strongest for lithium, aripiprazole, quetiapine XR, and T3 augmentation in this context.

Depression with psychotic features: Augmentation of antidepressants with antipsychotics is not just an augmentation strategy but a first-line approach. Combination of an antidepressant with an antipsychotic (the combination of olanzapine + fluoxetine, marketed as Symbyax, is FDA-approved for this indication) produces response rates of approximately 50–60% versus 30–35% for either drug class alone.

Bipolar depression: Quetiapine monotherapy (300 mg/day) and the olanzapine-fluoxetine combination are FDA-approved for bipolar depression. Lithium and lamotrigine serve as mood-stabilizer augmentation. Atypical antipsychotic augmentation of mood stabilizers has a strong rationale here, though augmenting antidepressants in bipolar depression is controversial due to switch risk.

Obsessive-Compulsive Disorder (OCD) with inadequate SSRI response: Augmentation with low-dose atypical antipsychotics (particularly risperidone 0.5–2 mg, aripiprazole 5–15 mg) has shown efficacy in approximately one-third of SSRI-refractory OCD patients. A meta-analysis by Dold et al. (2015) found a pooled response rate of approximately 30% with antipsychotic augmentation versus 12% for placebo (NNT ≈ 5–6). Risperidone had the strongest evidence.

Where Augmentation Is Less Effective or Uncertain

Generalized Anxiety Disorder (GAD): Atypical antipsychotic augmentation of SSRIs/SNRIs in GAD has limited RCT support. Quetiapine has shown some anxiolytic benefit, but metabolic risks generally outweigh benefits for a condition that is neither life-threatening nor typically associated with psychotic features.

Persistent Depressive Disorder (Dysthymia): The chronic, characterological quality of persistent depressive disorder often responds less robustly to pharmacological augmentation. Psychotherapy augmentation (particularly CBASP — Cognitive Behavioral Analysis System of Psychotherapy) may be more appropriate than pharmacological augmentation in this population.

Depression with active substance use disorders: Augmentation efficacy is substantially diminished when active substance use is perpetuating depressive symptoms. Treating the substance use disorder may be more effective than pharmacological augmentation in these patients.

Identifying patients most likely to benefit from specific augmentation strategies remains a critical research priority. Several moderators and predictors have emerged from clinical trials and post-hoc analyses:

Clinical Predictors

• Degree of partial response: Patients with partial response (≥25% improvement) on the base antidepressant are more likely to benefit from augmentation than complete non-responders, for whom switching may be more appropriate.
• Symptom profile: Anhedonia, psychomotor retardation, and cognitive symptoms may preferentially respond to dopaminergically-active augmentation agents (aripiprazole, brexpiprazole, cariprazine). Anxiety and insomnia symptoms may respond better to quetiapine or mirtazapine. Melancholic features have historically been associated with better lithium augmentation response.
• Duration of current episode: Longer episodes prior to augmentation are associated with lower response rates across augmentation strategies, consistent with the broader finding that chronicity predicts poorer treatment outcomes.
• Number of prior failed trials: STAR*D demonstrated a step-wise decline in remission rates with each successive treatment level: 37% at Step 1, 31% at Step 2, 14% at Step 3, and 13% at Step 4. This gradient applies equally to augmentation and switching strategies.
• Comorbid personality disorders: Particularly borderline personality disorder, which reduces the probability of antidepressant and augmentation response by approximately 50% according to several studies, likely reflecting the contribution of interpersonal and characterological factors to the depressive presentation.

Biological Predictors

• Thyroid function: Subclinical hypothyroidism (TSH 4.5–10 mIU/L) predicts better response to T3 augmentation, though patients with normal thyroid function can also respond.
• Inflammatory biomarkers: Elevated CRP (>3 mg/L) has been associated with preferential response to anti-inflammatory augmentation and, in some analyses, to augmentation with agents possessing anti-inflammatory properties (e.g., lithium, quetiapine).
• Family history of bipolar disorder or lithium response: Positive family history predicts better response to lithium augmentation, consistent with the hypothesis that some treatment-resistant unipolar depression represents a bipolar-spectrum illness.
• Pharmacogenomics: CYP2D6 and CYP2C19 genotype influence metabolism of many SSRIs and atypical antipsychotics, potentially affecting both base antidepressant and augmenting agent levels. However, prospective evidence that pharmacogenomic testing improves augmentation outcomes remains limited. The GUIDED trial showed modest improvements in response when pharmacogenomic-guided prescribing was used, but the clinical significance of findings has been debated.

Lithium

Lithium's side effect profile is its primary clinical limitation as an augmentation agent. Common adverse effects include:

• Renal: Nephrogenic diabetes insipidus (polyuria/polydipsia) in up to 40% of long-term users; chronic kidney disease with GFR decline of approximately 0.7–1.0 mL/min/year over decades
• Thyroid: Clinical or subclinical hypothyroidism in 5–15% of patients; goiter in approximately 5%
• Neurological: Fine tremor (27%), cognitive dulling, ataxia at supratherapeutic levels
• Gastrointestinal: Nausea, diarrhea, especially during initiation
• Metabolic: Weight gain (average 4–6 kg over 2 years), hyperparathyroidism
• Cardiac: T-wave flattening, sinus node dysfunction (rare)
• Teratogenicity: Ebstein's anomaly risk approximately 1 in 1000 (elevated ~20-fold over baseline, but absolute risk remains low); historically overestimated

Contraindications: Severe renal impairment (eGFR <30), severe cardiovascular disease, untreated Addison's disease, significant sodium depletion. Relative contraindications: Brugada syndrome, myasthenia gravis, concurrent use of nephrotoxic agents.

Atypical Antipsychotics

Metabolic effects are the primary concern, particularly with quetiapine and olanzapine:

• Weight gain: Quetiapine augmentation associated with mean weight gain of 1.5–3.0 kg over 6–8 weeks in registration trials; olanzapine causes significantly more (5–8 kg over similar periods)
• Metabolic syndrome: Increased fasting glucose, triglycerides, and LDL cholesterol, with olanzapine and quetiapine carrying the highest risk. Aripiprazole and brexpiprazole are metabolically relatively neutral.
• Akathisia: Particularly with aripiprazole (incidence approximately 23% in registration trials vs. 4% placebo), often dose-dependent and a major cause of discontinuation. Starting at 2 mg and titrating slowly mitigates risk.
• Sedation: Significant with quetiapine (up to 30% in trials) and olanzapine; less with aripiprazole and brexpiprazole.
• Tardive dyskinesia: Risk is lower with atypicals than typicals but not zero; estimated at approximately 0.5–1% per year of exposure with atypical antipsychotics. This risk is particularly relevant because augmentation may continue indefinitely.
• Prolactin elevation: Minimal with aripiprazole (which actually lowers prolactin as a D2 partial agonist) and quetiapine; significant with risperidone.

Limitations Common to All Augmentation Strategies

• Polypharmacy burden: Augmentation inherently increases pill burden, side effect risk, drug interaction potential, and cost.
• Insufficient long-term data: Most augmentation RCTs are 6–8 weeks in duration. Long-term maintenance data are sparse, and the optimal duration of augmentation treatment is poorly defined. NICE guidelines suggest continuing augmentation for at least 6 months after remission, but this is based primarily on expert consensus.
• Publication bias: Positive augmentation trials are more likely to be published, particularly for industry-sponsored atypical antipsychotic augmentation studies. The true effect sizes may be smaller than published estimates.

Youth (Children and Adolescents)

Augmentation evidence in pediatric populations is markedly more limited than in adults. The TORDIA (Treatment of Resistant Depression in Adolescents) trial found that switching antidepressants plus adding CBT produced better outcomes than switching alone, but pharmacological augmentation with atypical antipsychotics was not studied in this trial. Off-label use of aripiprazole augmentation in adolescents with TRD occurs in clinical practice, but RCT data are extremely limited. Lithium augmentation has been studied in adolescents in small open-label trials with mixed results. The FDA black box warning regarding suicidality risk with antidepressants in youth adds complexity to any augmentation decision, as does the heightened sensitivity of developing brains to metabolic and endocrine effects of atypical antipsychotics.

Clinical recommendation: In youth with treatment-resistant depression, augmentation with evidence-based psychotherapy (particularly CBT or interpersonal therapy for adolescents) should be prioritized before pharmacological augmentation. If pharmacological augmentation is pursued, lithium and aripiprazole at the lowest effective doses are the most studied options, with close metabolic and weight monitoring.

Elderly (≥65 years)

Older adults face distinct augmentation challenges: age-related declines in renal and hepatic function alter pharmacokinetics; polypharmacy risks increase with each added medication; and vulnerability to specific side effects — orthostatic hypotension (quetiapine), falls (all sedating agents), cognitive impairment (lithium, anticholinergic agents), and cardiac conduction abnormalities — is heightened.

Lithium augmentation in the elderly requires lower target serum levels (typically 0.4–0.6 mEq/L) and more frequent monitoring due to age-related GFR decline. Aripiprazole augmentation at low doses (2–5 mg) has shown benefit in older adult samples in post-hoc analyses of registration trials. The IRL-GRey (Incomplete Response in Late-Life Depression: Getting to Remission) study found that aripiprazole augmentation of venlafaxine in adults ≥60 years produced a remission rate of approximately 44% versus 29% for placebo (NNT ≈ 7).

Importantly, all atypical antipsychotics carry an FDA black box warning for increased risk of death in elderly patients with dementia-related psychosis, which must be communicated even when the indication is depression augmentation in a patient without dementia.

Pregnancy and Lactation

Augmentation during pregnancy involves complex risk-benefit analysis. Lithium, historically classified as Category D, carries a risk of Ebstein's anomaly (approximately 1 in 1000 exposed pregnancies, compared to a baseline risk of approximately 1 in 20,000) and neonatal toxicity (floppy infant syndrome, nephrogenic diabetes insipidus). However, these risks must be weighed against the well-documented risks of severe untreated maternal depression on fetal development, maternal functioning, and suicide risk.

Atypical antipsychotics have variable reproductive safety data. Quetiapine and olanzapine have the most pregnancy exposure data and are generally not associated with major teratogenic risk, though neonatal extrapyramidal symptoms and withdrawal effects have been reported. Aripiprazole has less exposure data but no clear signal of teratogenicity.

The Massachusetts General Hospital National Pregnancy Registry for Psychiatric Medications and the ENTIS (European Network of Teratology Information Services) databases are the primary sources for ongoing reproductive safety surveillance. Decisions about augmentation during pregnancy should involve shared decision-making between the patient, psychiatrist, and obstetrician, ideally through a reproductive psychiatry consultation.

Augmentation strategy selection is influenced not only by efficacy data but by practical considerations of access, cost, and clinical infrastructure:

Cost Considerations

• Lithium: Among the least expensive psychotropic medications, available generically at approximately $10–30/month in the US. However, the ongoing monitoring requirements (serum levels, renal function, thyroid function) add $200–500/year in laboratory costs.
• T3 (liothyronine): Generic formulations are available at approximately $15–40/month — highly cost-effective given the evidence base.
• Aripiprazole: Available generically since 2015 at approximately $15–50/month, a dramatic reduction from the $800+/month brand-name cost. This has substantially improved access.
• Quetiapine XR: Generic extended-release formulations are available at approximately $20–60/month.
• Brexpiprazole (Rexulti): Remains under patent; brand-name cost approximately $1,200–1,500/month without insurance, significantly limiting access.
• Cariprazine (Vraylar): Brand-name cost approximately $1,400–1,700/month; no generic available as of 2024.
• Esketamine (Spravato): Approximately $600–900 per administration session, administered twice weekly initially then weekly to biweekly. Annual costs can exceed $20,000–30,000, plus facility fees for the required supervised administration. Insurance coverage varies substantially by plan and region.

Provider Training and Practice Requirements

Most pharmacological augmentation strategies fall within the standard scope of psychiatry residency training. However, some newer approaches have specific requirements:

• Esketamine: Requires enrollment in the Spravato REMS (Risk Evaluation and Mitigation Strategy) program. Administration must occur in a certified healthcare setting with observation capabilities. This creates significant infrastructure barriers for community practices.
• Lithium management: While technically straightforward, surveys suggest that younger psychiatrists receive less training in lithium use than previous generations, leading to underuse despite strong evidence. Comfort with interpreting serum levels, managing interactions, and monitoring for toxicity is essential.
• Psychotherapy augmentation: Adding evidence-based psychotherapy (CBT, behavioral activation, CBASP) as a non-pharmacological augmentation strategy requires therapists trained in specific manualized protocols. Access to trained therapists varies enormously by geography, with rural and underserved areas facing particular shortages.

Equity and Access Disparities

Augmentation access is unequally distributed across populations. Patients with Medicaid or no insurance face formulary restrictions, prior authorization requirements, and inability to afford brand-name agents. Racial and ethnic minorities are less likely to receive guideline-concordant augmentation strategies and more likely to receive higher doses of individual agents rather than evidence-based augmentation. Addressing these disparities requires systemic changes including improved insurance parity, expanded formularies, and culturally informed shared decision-making about treatment options.

Synthesizing the available evidence, a pragmatic clinical algorithm for augmentation in treatment-resistant depression can be proposed:

Step 1: Confirm True Treatment Resistance

Verify adequate dose, duration (≥6–8 weeks), and adherence of current antidepressant. Rule out contributing comorbidities (substance use, medical illness, personality disorder, unrecognized bipolar disorder). Confirm diagnosis.

Step 2: First Augmentation Trial

Based on current guideline rankings and evidence synthesis:

• First-line options: Aripiprazole (2–15 mg/day), lithium (target 0.5–0.8 mEq/L), or quetiapine XR (150–300 mg/day). Choice should be individualized based on patient profile — aripiprazole for patients concerned about sedation and weight gain; quetiapine for patients with prominent insomnia and anxiety; lithium for patients with a family history of bipolar disorder, suicidality, or melancholic features.
• High-value alternative: T3 (25–50 μg/day) — underutilized, well-tolerated, inexpensive, with comparable evidence to lithium in the STAR*D comparison.
• Allow 4–6 weeks at adequate dose before determining response.

Step 3: If First Augmentation Fails

Switch to an alternative first-line augmentation agent. If lithium was tried, switch to aripiprazole or quetiapine (or vice versa). Consider combination antidepressant therapy (e.g., adding bupropion or mirtazapine to the existing SSRI/SNRI) as an alternative strategy.

Step 4: For Refractory Cases

Consider esketamine (if accessible and insurance covers it), MAOIs (tranylcypromine or phenelzine — highly effective but underutilized due to dietary restrictions and drug interaction concerns), electroconvulsive therapy (ECT — the most effective treatment for severe TRD, with response rates of 50–70% even after multiple failed pharmacological trials), or referral for emerging interventions such as repetitive transcranial magnetic stimulation (rTMS), vagus nerve stimulation (VNS), or clinical trials.

Throughout all steps: Evidence-based psychotherapy should be offered concurrently. The combination of pharmacotherapy and psychotherapy consistently outperforms either alone, with particular evidence for CBT, behavioral activation, and CBASP as augmentation of pharmacotherapy in TRD.