The Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial is the largest and most clinically consequential antidepressant effectiveness study conducted to date. Its findings directly established that first-line antidepressant monotherapy produces remission in only a minority of patients, validated a sequential treatment framework that remains the standard structure for clinical decision-making in major depressive disorder (MDD), and quantified the diminishing returns of successive treatment steps in a real-world patient population.
STAR*D enrolled 4041 adult outpatients with nonpsychotic MDD across 41 clinical sites, using a measurement-based care framework with the Quick Inventory of Depressive Symptomatology (QIDS-SR) as the primary outcome instrument and remission defined as a QIDS-SR score of five or below.1 Of particular clinical significance, STAR*D used an effectiveness rather than efficacy design — it enrolled patients with comorbid general medical conditions and psychiatric comorbidities who would have been excluded from most randomized controlled trials, making its findings more representative of patients encountered in clinical practice than traditional registration trial populations. All patients began with citalopram at Step 1. Those who did not achieve remission or could not tolerate citalopram were offered options at Step 2: switch to an alternative antidepressant, augment citalopram with bupropion or buspirone, or switch to cognitive behavioral therapy (CBT). Subsequent non-remitters progressed through Steps 3 and 4 with further combinations and switches.
The overall pattern of remission rates across the four steps was: Step 1 (citalopram) 36.8%; Step 2 approximately 30.6%; Step 3 approximately 13.7%; Step 4 approximately 13.0%.1 Cumulative remission over all four steps reached approximately 67%, but the trial also documented that the probability of sustaining remission declined steeply with each successive step, with only 3.0% of those who achieved remission at Step 4 remaining in remission at follow-up, compared with 40% of those who achieved remission at Step 1.
The most important clinical takeaways from STAR*D are four: first, remission — not merely response — should be the treatment target, because patients who achieve response without remission have substantially higher relapse rates and continued functional impairment; second, approximately one-third of patients with MDD will not achieve remission on their first antidepressant, making a systematic plan for next-step treatment an essential component of initial prescribing rather than an afterthought; third, each successive treatment step produces lower remission rates and shorter durations of sustained remission, providing a strong pharmacological argument for optimizing the first treatment step through dose adequacy and full trial duration rather than premature switching; and fourth, the effectiveness design of STAR*D demonstrated that comorbidity — particularly anxiety disorders, which were present in over 50% of participants — substantially reduces the probability of remission at Step 1 and should inform initial agent selection and dose targets.2
An adequate antidepressant trial is defined as treatment at a therapeutic dose for a sufficient duration to assess efficacy — generally four to six weeks at a dose within the established therapeutic range, confirmed by tolerability assessment and, where available, therapeutic drug monitoring.2 Premature switching before four weeks at an adequate dose misses patients who would have achieved remission with additional time, while waiting beyond eight to twelve weeks without meaningful improvement and without dose optimization constitutes an unnecessarily prolonged course of ineffective treatment. The practical clinical algorithm derived from STAR*D and subsequent guidelines recommends: initiate at lowest effective dose; titrate to the target dose within two to four weeks if tolerated; reassess at four weeks for early trajectory; if partial response, continue and optimize dose; if no response at four to six weeks at adequate dose, consider switching, augmenting, or combining, with the decision informed by the patient's partial versus no response, tolerability of the current agent, and clinical urgency.2
Only one-third of patients achieve remission on the first antidepressant. This is not a failure of the drug class — it is the expected pharmacological reality, and the clinical response should be a systematic, prospectively planned sequential treatment strategy, not reactive switching driven by frustration. STAR*D validated that persistence through a structured sequence, not more rapid switching, maximizes cumulative remission rates.
Augmentation in antidepressant pharmacology refers to adding a second agent to a partially effective primary antidepressant, with the goal of converting partial response to full remission. The rationale for any augmentation strategy should be grounded in the pharmacology of both agents: what mechanism is the primary antidepressant not providing, and how does the augmenting agent supply it? Augmentation is pharmacologically distinct from switching, which abandons the primary agent entirely, and from combining, which employs two antidepressants of different classes.
Atypical antipsychotics are the most evidence-rich augmentation class in antidepressant pharmacology, with FDA-approved indications for adjunctive use in MDD for aripiprazole, quetiapine extended-release, brexpiprazole, and cariprazine when added to antidepressants in patients with inadequate response.3 The mechanistic rationale for each differs and is pharmacologically instructive. Aripiprazole is a dopamine D2/D3 receptor partial agonist and 5-HT1A partial agonist; its augmenting effect in depression is attributed to partial agonism at D2 receptors in the mesocortical dopamine pathway, partially restoring dopaminergic tone in the prefrontal cortex without producing hyperprolactinemia or extrapyramidal effects, and to 5-HT1A partial agonism contributing anxiolytic effects. Quetiapine's augmenting activity at low doses (25 mg to 200 mg) appears primarily mediated by potent 5-HT2A antagonism and H1 antagonism, improving sleep quality and reducing anxiety rather than through D2 blockade. Brexpiprazole, a D2 partial agonist with additional 5-HT1A partial agonism and 5-HT2A antagonism, has demonstrated statistically significant augmenting effects with more favorable akathisia rates than aripiprazole.
The shared adverse effect burden of atypical antipsychotic augmentation includes weight gain (most prominent with quetiapine and olanzapine; less with aripiprazole and brexpiprazole), metabolic dysregulation, and the general risks of tardive dyskinesia and extrapyramidal symptoms with long-term use, which require explicit informed consent and periodic neurological monitoring.3
Lithium augmentation of tricyclic antidepressants (TCAs) was the first pharmacologically characterized augmentation strategy in psychiatry, described in case series in the early 1980s and subsequently supported by randomized controlled trials.4 The mechanism by which lithium enhances antidepressant response remains not fully elucidated, but the leading hypotheses involve serotonergic sensitization: lithium increases presynaptic serotonin synthesis and release, enhances postsynaptic 5-HT1A receptor sensitivity, and may augment the SERT-blocking effect of antidepressants by increasing the serotonin available for reuptake blockade to act on. Lithium also modulates intracellular signaling pathways relevant to neuroplasticity, including glycogen synthase kinase-3 beta (GSK-3beta) inhibition, which promotes beta-catenin stability and neurotrophin signaling. The effective augmenting plasma concentration for lithium is generally 0.6 to 1.0 mEq/L — within the standard mood stabilizer therapeutic range — and requires the same monitoring infrastructure as lithium therapy for any indication: baseline renal and thyroid function, regular serum concentration monitoring, and hydration counseling. Lithium's narrow therapeutic index and the renal and thyroid adverse effects of long-term use have made it a less frequently used augmentation option in contemporary practice relative to the atypical antipsychotics, despite its more extensive historical evidence base.4
Bupropion augmentation of SSRIs is the most widely used augmentation strategy in primary care settings and was included as a STAR*D Step 2 option on the basis of its distinct mechanism and favorable tolerability profile.5 Bupropion inhibits the norepinephrine transporter (NET) and the dopamine transporter (DAT) without significant SERT activity, providing noradrenergic and dopaminergic enhancement in circuits where the primary SSRI provides only serotonergic input. The mechanistic rationale is that residual depressive symptoms after SSRI monotherapy — particularly fatigue, hypersomnia, reduced motivation, cognitive slowing, and anhedonia — are more closely linked to insufficient noradrenergic and dopaminergic tone in prefrontal circuits than to insufficient serotonergic tone, and that adding bupropion addresses the catecholaminergic deficit specifically. Bupropion augmentation also mitigates SSRI-induced sexual dysfunction through its dopaminergic activity, making it doubly attractive in patients with residual anhedonia and sexual adverse effects. The primary pharmacokinetic consideration is that bupropion is a potent CYP2D6 inhibitor, which increases plasma concentrations of SSRIs metabolized by CYP2D6 and requires awareness of this interaction, particularly in patients on paroxetine, fluoxetine, or TCAs.
Buspirone is a partial agonist at the 5-HT1A receptor and a weak D2 receptor partial agonist, with anxiolytic properties primarily mediated through 5-HT1A agonism at postsynaptic hippocampal receptors and autoreceptor desensitization analogous to the mechanism attributed to vilazodone. Its augmenting mechanism in combination with SSRIs is theorized to involve presynaptic 5-HT1A autoreceptor desensitization — the same mechanism responsible for the delayed onset of SSRI efficacy — but achieving that desensitization more rapidly or completely through direct 5-HT1A agonism rather than waiting for the autoreceptor to respond to chronically elevated synaptic serotonin.5 STAR*D included buspirone as one of two augmentation options at Step 2; neither augmentation arm (bupropion or buspirone) was superior to the other in terms of remission rates, though bupropion showed a marginally more favorable tolerability profile. Buspirone augmentation is most rationally applied in patients with residual anxiety as a prominent feature of their partial response, where its anxiolytic activity adds a mechanistically distinct benefit beyond antidepressant enhancement. The absence of dependence, withdrawal, or respiratory depression makes it a particularly suitable option in patients with anxiety disorders who would otherwise be candidates for benzodiazepines.
Triiodothyronine (T3) augmentation is one of the oldest pharmacological augmentation strategies in psychiatry, with case series data extending back to the 1960s and a consistent signal in small randomized trials that T3 added to TCAs or SSRIs in euthyroid patients with inadequate response accelerates response and may improve ultimate remission rates.5 The mechanism is incompletely understood but likely involves thyroid hormone's role in regulating central nervous system serotonergic and noradrenergic receptor sensitivity; hypothyroid states are associated with depression and with reduced responsiveness to antidepressants, and even within the normal reference range, patients at the lower end of normal thyroid function may have insufficient thyroid hormone signaling to support optimal antidepressant drug action. T3 is preferred over T4 (levothyroxine) for augmentation because it does not require peripheral conversion to the active form and has a shorter half-life that permits more rapid dose adjustment. The typical augmenting dose is 25 mcg to 50 mcg daily. STAR*D included T3 as a Step 3 augmentation option, where it showed similar remission rates to lithium augmentation but with a more favorable adverse effect profile in that trial's population.
Choose the augmenting agent based on the residual symptom pattern and the pharmacological gap: fatigue, anhedonia, and cognitive slowing favor bupropion (NET/DAT) or a low-dose atypical antipsychotic with dopaminergic activity; residual anxiety favors buspirone or low-dose quetiapine; sleep disruption favors quetiapine or mirtazapine; inadequate overall antidepressant response with no clear residual pattern favors lithium, aripiprazole, or brexpiprazole. The augmenting pharmacology should address a mechanism the primary antidepressant is not covering.
Treatment-resistant depression (TRD) lacks a universally agreed regulatory or clinical definition, but the most widely used operational criterion is failure to achieve an adequate response — typically defined as at least 50% reduction in validated symptom scale scores — after two or more adequate antidepressant trials of different classes at therapeutic doses for sufficient duration (generally four to eight weeks).6 This definition is the basis for the FDA approval of esketamine in TRD and for most research inclusion criteria. The distinction between TRD and pseudo-resistance is clinically critical: pseudo-resistance encompasses apparent treatment failure caused by inadequate dosing, insufficient trial duration, poor medication adherence, pharmacokinetic variability (including unrecognized CYP2D6 poor metabolizer status), or unaddressed comorbidities such as hypothyroidism, substance use disorders, sleep apnea, or untreated anxiety, any of which can prevent antidepressant response regardless of drug selection. A systematic clinical assessment before concluding TRD should explicitly rule out each of these factors.
The Maudsley Staging Method (MSM) is the most clinically validated TRD staging instrument, assigning scores based on treatment duration, number of failed trials, nature of trials (monotherapy versus augmentation or combination), and whether ECT has been tried.6 Total scores categorize patients into mild (2–6), moderate (7–10), and severe (11–15) TRD, with higher stages associated with lower probability of response to subsequent treatment steps and longer time to remission. The MSM is useful clinically not because staging itself alters individual treatment decisions, but because it provides a structured framework for documenting treatment history, identifying gaps in the treatment sequence, and predicting which patients are likely to require more intensive or specialized interventions. Patients scoring in the moderate or severe range on the MSM are those most likely to benefit from psychiatric consultation, consideration of esketamine, ECT evaluation, or enrollment in a clinical trial.
Several biological and pharmacological features are associated with genuine treatment resistance rather than pseudo-resistance. Elevated inflammatory markers, particularly C-reactive protein (CRP) above 3 mg/L and interleukin-6 (IL-6), predict poor response to SSRIs and may predict relatively better response to agents with anti-inflammatory properties or to bupropion.7 Hyperactivation of the HPA axis with persistent hypercortisolemia impairs neuroplasticity and antidepressant drug action by reducing hippocampal neurogenesis and downregulating BDNF expression; patients with sustained HPA hyperactivation require antidepressant strategies that directly address this biological substrate rather than simple transporter blockade. CYP2D6 ultra-rapid metabolizer genotype produces subtherapeutic antidepressant plasma concentrations at standard doses without obvious clinical signs of inadequate drug exposure; pharmacogenomic testing identifying ultra-rapid metabolizer status should prompt substantial dose increases or switching to agents not metabolized by CYP2D6. Structural neuroimaging findings of hippocampal volume loss, and functional imaging patterns of altered amygdala-prefrontal connectivity, are increasingly being studied as biomarkers that may eventually guide treatment selection, though they are not yet part of routine clinical decision-making.7
A proportion of patients presenting as TRD in unipolar depression framework are in fact suffering from bipolar disorder, most commonly bipolar II disorder, in which the hypomanic episodes may be mild, brief, or unrecognized by the patient and clinician. Antidepressant monotherapy in bipolar depression is associated with mood destabilization, cycle acceleration, and mixed state induction in a significant proportion of patients, and may produce an apparent antidepressant treatment failure that is actually a pharmacologically driven worsening of the bipolar course.8 Every patient presenting with apparent TRD warrants explicit assessment for bipolar features: a structured mood history extending back to adolescence or early adulthood, screening for hypomanic symptoms with instruments such as the Mood Disorder Questionnaire (MDQ), and review of family history for bipolar disorder or related conditions. The discovery of bipolar disorder in apparent TRD completely changes the pharmacological treatment strategy, with mood stabilization using lithium, lamotrigine, quetiapine, or lurasidone becoming the primary therapeutic goal, and with antidepressant monotherapy either discontinued or used only in combination with a mood stabilizer under close monitoring.
Confirm adequate dosing and duration of each prior trial. Confirm adherence (plasma concentration measurement if available). Assess for comorbid medical conditions reducing response: hypothyroidism, sleep apnea, chronic pain, inflammatory disease. Assess for comorbid psychiatric conditions: anxiety disorders, substance use, ADHD, personality disorders. Screen explicitly for bipolar spectrum illness using structured instruments. Only after ruling out pseudo-resistance is a patient appropriately classified as having true TRD.
Non-pharmacological interventions in TRD are not alternatives to pharmacology but operate through overlapping neurobiological mechanisms that complement drug-based approaches. Clinicians who understand the pharmacology of antidepressants must also understand the mechanistic basis of these interventions to integrate them rationally into a comprehensive treatment plan.
Electroconvulsive therapy (ECT) remains the most effective acute antidepressant intervention available, with remission rates in TRD populations of 50% to 70% in prospective series, substantially exceeding pharmacological alternatives at the same stage of treatment resistance.9 The mechanism of ECT's antidepressant action is not fully established but encompasses multiple biological pathways: it produces massive, synchronous neural discharge that increases serotonin and norepinephrine release from central monoaminergic pathways; it downregulates beta-adrenergic receptors and 5-HT2 receptors in a pattern analogous to chronic antidepressant drug treatment; it substantially increases BDNF expression and adult hippocampal neurogenesis, with the neurogenic response to ECT among the most robust of any antidepressant intervention; and it normalizes HPA axis hyperactivation and reduces hypercortisolemia more rapidly and reliably than pharmacological agents.9 ECT is particularly indicated in patients with severe or life-threatening depression (psychotic depression, refusal to eat or drink, imminent suicidal risk), in pregnant patients for whom pharmacological alternatives carry unacceptable risk, and in patients with known prior response to ECT. The primary limitation of ECT is the cognitive adverse effects — particularly anterograde and retrograde memory impairment — that occur during the acute treatment course and that may persist for weeks to months, though most cognitive effects resolve fully within six months.
Repetitive transcranial magnetic stimulation (rTMS) applies rapidly alternating magnetic fields to the scalp overlying the left dorsolateral prefrontal cortex (DLPFC), generating focal electrical currents that modulate neuronal excitability in the underlying cortex and its connected subcortical circuits.10 High-frequency rTMS (10 Hz) applied to the left DLPFC increases excitability of the target cortex, which is characteristically hypoactive in MDD, and produces downstream effects on limbic circuitry through corticolimbic connectivity. The FDA has cleared rTMS for MDD in patients who have failed one prior antidepressant trial, a lower treatment resistance threshold than ECT. Response rates in FDA-cleared protocols are approximately 50% to 60%, with remission rates of approximately 30% to 37% — substantially lower than ECT but produced without the cognitive adverse effects, anesthesia requirement, or hospitalization of ECT. Deep TMS using an H-coil capable of reaching deeper cortical and subcortical structures, and theta-burst stimulation (TBS) — a more compressed protocol that achieves comparable efficacy to standard rTMS in a fraction of the session time — have received FDA clearance and are increasingly used in clinical practice. The primary pharmacological integration point is that rTMS and antidepressant medications appear to work synergistically, with continuation of pharmacotherapy during and after rTMS associated with better durability of response than rTMS followed by drug discontinuation.10
Cognitive behavioral therapy (CBT) and other structured psychotherapies are not merely supportive adjuncts in TRD — they produce measurable neurobiological changes including normalization of amygdala hyperreactivity, restoration of prefrontal cognitive control circuitry, and changes in cortisol reactivity that overlap with the biological effects of antidepressants.11 CBT combined with antidepressant pharmacotherapy produces higher remission rates and more durable responses than either treatment alone in moderate-to-severe MDD. STAR*D included CBT as a Step 2 option; remission rates with CBT alone or CBT combined with medication were comparable to medication-based options, though CBT required longer to achieve remission. In TRD, the addition of CBT specifically designed for treatment resistance — addressing hopelessness, avoidance behaviors, negative cognitive schemas, and medication adherence beliefs that develop in the context of chronic treatment failure — is supported by randomized evidence showing meaningful improvements in remission rates over pharmacotherapy alone.
The optimal treatment of TRD combines pharmacological optimization with the most appropriate somatic intervention and structured psychotherapy. ECT for severe, urgent, or life-threatening presentations; rTMS for moderate TRD where ECT is declined or not yet indicated; CBT or its variants as a consistent companion to pharmacotherapy at every stage. Pharmacologists should understand that ECT, rTMS, and CBT all work through overlapping neurobiological pathways — BDNF, HPA normalization, monoamine regulation — that are also the targets of antidepressant drugs.
Achieving remission from a depressive episode is the acute treatment goal, but preventing relapse and recurrence is the long-term pharmacological imperative. The distinction between relapse (return of the index episode before full recovery) and recurrence (a new depressive episode following a period of full recovery) has pharmacological implications that directly govern treatment duration decisions, dose maintenance strategies, and the threshold for recommending indefinite treatment.
Antidepressants do not cure MDD in the way an antibiotic cures an infection; they suppress active depressive episodes and reduce the probability of relapse and recurrence while they are being taken. This distinction is fundamental to informed consent and to duration-of-treatment decisions. The neurobiological basis for continuing antidepressant treatment beyond acute remission involves the ongoing neuroplasticity-maintaining effects of antidepressants: continued BDNF expression, maintenance of newly formed dendritic spines and synaptic connections, and sustained normalization of HPA axis activity all appear to require continued drug exposure, because these adaptations partially reverse when the drug is withdrawn.12 Pooled data from randomized withdrawal trials consistently demonstrate that patients who continue antidepressants after remission have substantially lower relapse rates over 12 to 24 months than those randomized to placebo substitution, with number needed to treat estimates of approximately three to six patients to prevent one relapse per year of maintenance treatment.12
The continuation phase of antidepressant treatment encompasses the period from acute remission through the natural duration of the depressive episode, typically defined as six to twelve months following remission. The pharmacological principle governing this phase is that early discontinuation of effective antidepressant treatment within this window results in relapse rates of approximately 50% within six months, because the underlying depressive episode has not yet run its natural biological course.13 The drug does not shorten the episode; it suppresses its manifestation. Premature discontinuation therefore exposes the patient to the biological substrate of the episode without pharmacological protection. Current clinical guidelines universally recommend continuing the effective antidepressant at the same dose that produced remission, for at least six to twelve months following remission, before considering dose reduction or discontinuation. The same dose that produced remission should be maintained — dose reduction during continuation increases relapse risk without reducing adverse effect burden sufficiently to justify the tradeoff in most patients.
Following the continuation phase, the decision of whether to discontinue antidepressant treatment or to enter a long-term maintenance phase depends on the patient's individual recurrence risk, which is governed by the number of prior depressive episodes, episode severity, age at first episode, and the presence of residual symptoms after remission.13 The recurrence rate after a first depressive episode is approximately 50%; after two episodes, approximately 70%; after three or more episodes, approximately 90%. Clinical guidelines recommend considering indefinite maintenance therapy in patients with three or more lifetime depressive episodes, a history of severe or suicidal episodes, a chronic or recurrent course without full interepisode recovery, or a strong family history of recurrent MDD. The pharmacological rationale for long-term maintenance is supported by randomized withdrawal trials demonstrating progressively increasing recurrence rates with each successive episode in unmedicated patients, and by evidence that long-term antidepressant maintenance reduces the biological kindling-like sensitization that makes each successive episode more likely and more severe. In patients maintained on antidepressants for recurrence prevention, the same agent and dose that produced the most recent remission is recommended, as switching during the maintenance phase introduces unnecessary pharmacological risk without demonstrated benefit.
Measurement-based care (MBC) does not end when acute remission is achieved. Standardized symptom monitoring using validated instruments such as the PHQ-9 at regular intervals — three to six months during maintenance — provides early detection of subsyndromal symptom return, which is a strong predictor of full relapse if not addressed promptly. Residual symptoms after apparent remission are among the most powerful predictors of subsequent relapse: patients with residual symptoms have two to three times the relapse rate of those with complete symptomatic remission at the end of acute treatment.14 Residual symptoms in patients on adequate maintenance antidepressant doses should trigger a systematic reassessment: confirm dose adequacy, assess for new or ongoing psychosocial stressors, evaluate comorbid conditions that may be driving breakthrough symptoms, and consider augmentation, switching, or the addition of structured psychotherapy to address the gap between partial and full remission.
First episode: continue antidepressant for 6–12 months after remission, then consider gradual taper. Second episode: continue for 2 years after remission before considering discontinuation. Three or more episodes, or history of severe/suicidal episodes: indefinite maintenance therapy with the effective agent at the effective dose. All discontinuation should be gradual with a systematic tapering plan and scheduled follow-up within 4–8 weeks of dose reduction.
The eight modules of Chapter 17 have traced antidepressant pharmacology from its neurobiological foundations through the full clinical arc of treatment decision-making. The monoamine hypothesis provided the mechanistic framework for the first generation of antidepressants; the neuroplasticity, BDNF, and HPA axis models extended that framework to encompass the complexity of antidepressant action and the biological basis of treatment resistance. The receptor pharmacology of each drug class — SSRIs, SNRIs, TCAs, MAOIs, mirtazapine, bupropion, the novel receptor agents, and the glutamatergic rapid-acting agents — determines both the therapeutic effects and the predictable adverse effect profiles that govern rational prescribing. STAR*D transformed empirical sequential prescribing into evidence-based algorithmic practice. Augmentation pharmacology provided mechanistic rationale for rational combination strategies. And the principles of adequate trial duration, measurement-based monitoring, and individualized maintenance duration close the pharmacological loop from molecular mechanism to long-term clinical outcome. The clinician who holds all of these frameworks simultaneously is equipped not merely to prescribe antidepressants, but to deploy them with the precision that their complex pharmacology demands and their patients require.
Rush AJ, Trivedi MH, Wisniewski SR, et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am J Psychiatry. 2006;163(11):1905–1917.
doi:10.1176/ajp.2006.163.11.1905Trivedi MH, Rush AJ, Wisniewski SR, et al. Evaluation of outcomes with citalopram for depression using measurement-based care in STAR*D: implications for clinical practice. Am J Psychiatry. 2006;163(1):28–40.
doi:10.1176/appi.ajp.163.1.28Nelson JC, Papakostas GI. Atypical antipsychotic augmentation in major depressive disorder: a meta-analysis of placebo-controlled randomized trials. Am J Psychiatry. 2009;166(9):980–991.
doi:10.1176/appi.ajp.2009.09030312Bauer M, Bschor T, Pfennig A, et al. World Federation of Societies of Biological Psychiatry (WFSBP) guidelines for biological treatment of unipolar depressive disorders in primary care. World J Biol Psychiatry. 2007;8(2):67–104.
doi:10.1080/15622970701227829Trivedi MH, Fava M, Wisniewski SR, et al. Medication augmentation after the failure of SSRIs for depression. N Engl J Med. 2006;354(12):1243–1252.
doi:10.1056/NEJMoa052964Ruhé HG, van Rooijen G, Spijker J, Peeters FP, Schene AH. Staging methods for treatment resistant depression: a systematic review. J Affect Disord. 2012;137(1–3):35–45.
doi:10.1016/j.jad.2011.02.020Strawbridge R, Young AH, Cleare AJ. Biomarkers for depression: recent insights, current challenges and future prospects. Neuropsychiatr Dis Treat. 2017;13:1245–1262.
doi:10.2147/NDT.S115965Ghaemi SN, Boiman EE, Goodwin FK. Diagnosing bipolar disorder and the effect of antidepressants: a naturalistic study. J Clin Psychiatry. 2000;61(10):804–808.
doi:10.4088/JCP.v61n1014UK ECT Review Group. Efficacy and safety of electroconvulsive therapy in depressive disorders: a systematic review and meta-analysis. Lancet. 2003;361(9360):799–808.
doi:10.1016/S0140-6736(03)12705-5Lefaucheur JP, Aleman A, Baeken C, et al. Evidence-based guidelines on the therapeutic use of repetitive transcranial magnetic stimulation (rTMS): an update (2014–2018). Clin Neurophysiol. 2020;131(2):474–528.
doi:10.1016/j.clinph.2019.11.002DeRubeis RJ, Siegle GJ, Hollon SD. Cognitive therapy versus medication for depression: treatment outcomes and neural mechanisms. Nat Rev Neurosci. 2008;9(10):788–796.
doi:10.1038/nrn2345Geddes JR, Carney SM, Davies C, et al. Relapse prevention with antidepressant drug treatment in depressive disorders: a systematic review. Lancet. 2003;361(9358):653–661.
doi:10.1016/S0140-6736(03)12599-8Kupfer DJ. Long-term treatment of depression. J Clin Psychiatry. 1991;52(suppl):28–34.
Judd LL, Paulus MJ, Schettler PJ, et al. Does incomplete recovery from first lifetime major depressive episode herald a chronic course of illness? Am J Psychiatry. 2000;157(9):1501–1504.
doi:10.1176/appi.ajp.157.9.1501