INTRODUCTION — Many patients presenting with unipolar major depression (major depressive disorder) do not recover after their initial treatment. As an example, one prospective observational study found that among 3671 outpatients who were treated with citalopram, remission occurred in only 37 percent [1]. In addition, patients who fail their initial treatment often do not respond to subsequent trials and frequently experience chronic depression, impaired psychosocial functioning, and poor overall general health [2].
This topic reviews choosing a specific treatment for resistant depression. The general principles of treating resistant depression, and the epidemiology, risk factors, assessment, and prognosis of treatment-resistant depression are discussed separately, as are the management of highly resistant (refractory) depression, the initial treatment of depression, and clinical features and diagnosis of depression.
●(See "Unipolar depression in adults: General principles of treating resistant depression".)
●(See "Unipolar depression in adults: Management of highly resistant (refractory) depression".)
●(See "Unipolar major depression in adults: Choosing initial treatment".)
●(See "Unipolar depression in adults: Clinical features".)
●(See "Unipolar depression in adults: Assessment and diagnosis".)
DEFINITIONS
●Unipolar major depression – Unipolar major depression (major depressive disorder) (table 1) is diagnosed in patients who have suffered at least one major depressive episode and have no history of mania or hypomania [3]. A major depressive episode is a period lasting at least two weeks, with five or more of the following symptoms, at least one of which is depressed mood or anhedonia: depressed mood, anhedonia, insomnia or hypersomnia, change in appetite or weight, psychomotor retardation or agitation, low energy, poor concentration, thoughts of worthlessness or guilt, and recurrent thoughts about death or suicide. Additional information about the clinical presentation and diagnosis of major depressive disorder is discussed separately. (See "Unipolar depression in adults: Assessment and diagnosis".)
●Treatment-resistant depression – The term “treatment-resistant depression” often refers to major depressive episodes that do not respond satisfactorily to at least two trials of optimally dosed antidepressant monotherapy; however, the definition has not been standardized. The definition of treatment-resistant depression is discussed separately. (See "Unipolar treatment resistant depression in adults: Epidemiology, risk factors, assessment, and prognosis", section on 'Treatment resistant depression'.)
●Treatment-refractory depression – The term “treatment-refractory depression” typically refers to unipolar major depressive episodes that do not respond satisfactorily to numerous (eg, 5 to 10) sequential treatment regimens. However, the definition has not been standardized, and there is no clear demarcation between treatment-resistant and treatment-refractory depression. The definition and management of refractory depression are discussed separately. (See "Unipolar depression in adults: Management of highly resistant (refractory) depression".)
GENERAL PRINCIPLES — The general principles and issues that are involved in treating resistant depression in adults include the following:
●Reassess the diagnosis
●Comorbidity
●Assess adherence
●Treatment strategies
●Nonspecific care management
●Duration of an adequate drug trial
●Referral
These general principles are discussed in detail separately. (See "Unipolar depression in adults: General principles of treating resistant depression".)
MILD TO MODERATE DEPRESSION — Mild to moderate major depression is characterized by the following clinical features:
●No suicidal or homicidal ideation or behavior, or ideation that does not pose an imminent risk. Ideation that does not pose an imminent risk includes the wish or hope that death will overtake oneself (eg, “Life is not worth living” or “I would be better off dead”); or fleeting thoughts of killing oneself, with nonexistent or vague plans to commit suicide and no intent.
●No psychotic features (eg, delusions or hallucinations).
●Little to no aggressiveness.
●Intact judgment such that the patient or others are not at imminent risk of being harmed.
●Impaired functioning is not obvious.
In addition, mild to moderate major depression is indicated by a score <20 points on the Patient Health Questionnaire – Nine Item (PHQ-9) (table 2). The PHQ-9 is a self-report assessment that is discussed separately. (See "Using scales to monitor symptoms and treat depression (measurement based care)", section on 'Patient Health Questionnaire - Nine Item'.)
Alternatively, one study classified episodes of major depression as mild to moderate in those individuals who had only five to seven of the nine symptoms that define major depression (table 1) [4]. Many studies have assessed severity of depression using clinician-administered instruments, such as the Hamilton Rating Scale for Depression (table 3) [5] or Montgomery-Asberg Depression Rating Scale (figure 1A-C) [6]. However, these rating scales are generally not used as part of standard care.
Mild to moderate depression can generally be treated in an outpatient or partial (day) hospital program setting.
Treatment algorithm — For patients with unipolar major depression who do not respond to initial treatment with an optimally dosed antidepressant medication, treatment strategies include augmentation (adding a treatment) and switching treatment (eg, switching antidepressants) [7]. Regardless of which strategy is used, we make one change at a time, which allows us to better understand whether a particular therapeutic is helpful.
For mild to moderate unipolar major depression that is treatment resistant, our general approach is as follows (algorithm 1):
●We suggest initially using augmentation interventions; some evidence suggests that the benefit of augmentation is modestly superior to switching antidepressants (see 'Efficacy of augmentation compared with switching' below). Augmenting an antidepressant with a second drug may provide faster, complimentary, or synergistic effects, compared with switching antidepressants [8]. In addition, augmentation avoids withdrawal symptoms that may arise when the initial antidepressant is discontinued. If patients do not respond to augmentation, we then switch antidepressants (ie, administer monotherapy with a new antidepressant) or switch to psychotherapy or repetitive transcranial magnetic stimulation (TMS). This approach, augmentation followed by switching, is consistent with multiple treatment guidelines and systematic reviews [9-11].
●However, it is reasonable to reverse the order of these treatment strategies and initially switch treatments. Some studies of treatment-resistant depression suggest that the benefits of switching and augmentation are comparable, and some patients may prefer antidepressant monotherapy. In addition, switching antidepressants may be preferable to augmentation because adherence is generally better with monotherapy than combination treatment [12,13]. Monotherapy may also cost less and may be less likely to cause adverse events and drug-drug interactions, compared with adding a second drug [8].
●Treatment-resistant patients who cannot tolerate an adequate dose of the initial antidepressant, or who encounter drug-drug interactions, should initially switch antidepressants [14].
Given that the efficacy of augmentation is not clearly superior to switching antidepressants, shared decision making with patients is important. Patients who partially benefit from the initial antidepressant and experience few adverse effects generally prefer adjunctive pharmacotherapy rather than switching [15]. Conversely, patients who experience less symptomatic improvement and more side effects with the antidepressant prescribed at first presentation typically prefer switching antidepressants. For treatment-resistant depression, multiple practice guidelines suggest either augmentation or switching [9,10,16-20].
Among patients with treatment-resistant depression who augment antidepressants with a second drug and do not respond, it is not clear how many trials of add-on therapy that clinicians should administer before switching the antidepressant. We generally provide one to three courses of augmentation before switching the antidepressant. When switching the antidepressant, we typically continue the current adjunctive drug, based upon the principle of making only one change at a time.
Although monotherapy often causes fewer adverse effects than multidrug regimens [16], this is not always the case. The prospective Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study compared side effects in 269 treatment-resistant patients who selected switching (from citalopram to bupropion, sertraline, or venlafaxine) as next-step treatment, and 269 patients who selected augmentation of citalopram (with bupropion or buspirone); propensity scoring was used to match the two groups for potential confounders observed at baseline [21]. The overall incidence of distressing side effects for each group was similar.
For patients with treatment-resistant depression who initially switch antidepressants and do not respond, it is not clear how many trials of antidepressant monotherapy that clinicians should administer before augmenting the antidepressant with a second treatment. We generally provide one to three courses of next-step antidepressant monotherapy before using augmentation.
An alternative to switching antidepressants or augmenting the antidepressant with a second medication is to switch from pharmacotherapy to psychotherapy (eg, cognitive-behavioral therapy [CBT]) or to TMS, or retain the initial antidepressant and add psychotherapy or TMS [9,22,23]. In addition, it is reasonable to augment the initial antidepressant with both pharmacotherapy and psychotherapy. However, psychotherapy is often not available, and many patients decline it [15]. The efficacy of switching to and augmenting with psychotherapy are each discussed elsewhere in this topic. (See 'Psychotherapy' below and 'Psychotherapy' below.)
Implementation of augmentation and switching are discussed elsewhere in this topic. (See 'Augmentation' below and 'Switching to a different treatment' below.)
Efficacy of augmentation compared with switching — For treatment-resistant depression, relatively few head-to-head randomized trials have compared augmentation with switching treatments. Although results across the trials are mixed, the most compelling evidence suggests that add-on therapy may be at least modestly superior to switching antidepressants:
●A 12-week, open-label, randomized trial enrolled patients (n = 1522; 85 percent male) with unipolar major depression who remained depressed despite treatment with at least one course of antidepressant therapy [24]. Nearly half of the patients had comorbid posttraumatic stress disorder, and most patients were currently receiving psychotherapy. Patients were assigned to one of three treatment strategies: augment the current antidepressant with aripiprazole (target dose 5 to 15 mg/day), augment with bupropion sustained release (target dose 300 to 400 mg/day), or switch to bupropion monotherapy. The primary findings included the following:
•Remission was statistically greater in the augment-aripiprazole group than the switch group, but the clinical difference was modest (29 versus 22 percent of patients). In the augment-bupropion group, remission occurred in 27 percent, which did not differ statistically from the other two groups. In patients 65 years and older, the benefit of augmentation with aripiprazole vis a vis switching to bupropion was more prominent (n = 226, 38 versus 21 percent) [25].
•Patients who remitted during acute treatment (n = 396) received continuation treatment for another 24 weeks; relapse at week 36 was comparable for the three treatment groups (approximately 25 percent of patients in each group).
•Adverse effects that occurred more often in the augment-aripiprazole group, compared with the other two groups, included akathisia, somnolence, and weight gain, as well as multiple abnormal laboratory tests. As an example, at week 36, weight gain from baseline ≥7 percent occurred in 25 percent of the augment-aripiprazole group, compared with 5 percent in each of the other two groups.
Anxiety occurred more often in the augment-bupropion group and the switch group, compared with the augment-aripiprazole group.
One limitation of the study was attrition; during the 12-week acute phase, 25 percent of the patients withdrew from the study. Other limitations included the lack of blinding for patients and treating clinicians, as well as the predominantly male sample; major depression occurs twice as often in females than in males. Nevertheless, other randomized trials indicate that aripiprazole augmentation may be more efficacious in females than in males [26]; if true, the present study may have underestimated the benefit of aripiprazole augmentation [27].
●Another study enrolled 96 patients (77 percent female) with unipolar major depression who did not respond to their initial antidepressant within six weeks and randomly assigned them to add-on treatment with aripiprazole (mean dose 4 mg/day) or to switch antidepressants [28]. Study treatments were administered for six weeks, and patients and treating clinicians were not blind to treatment. Remission occurred more often with aripiprazole augmentation than switching antidepressants (54 versus 20 percent of patients) and functioning also improved more with aripiprazole.
In addition, tolerability appeared to be comparable, such that discontinuation of treatment due to adverse effects for augmentation and switching occurred in 6 and 10 percent of patients. Weight gain during the six weeks with augmentation and switching was 0.6 and 1.0 kg, and akathisia and sexual functioning in the two groups were also comparable.
More limited evidence suggests that for treatment-resistant depression, the benefits of augmenting with pharmacotherapy and switching antidepressants are comparable:
●An eight-week randomized trial enrolled 375 patients and randomly assigned them to various therapies that included five augmentation options, two switch options, and continuing paroxetine monotherapy [29,30]. Pooled remission rates for the augmentation and the switch strategies appeared to be comparable (37 and 41 percent; difference was not statistically tested).
●Multiple reviews that examined randomized, placebo-controlled trials of augmentation and separate randomized, placebo-controlled trials of switching concluded that the two strategies achieved comparable results [31,32]. As an example, a pooled analysis in one review found that the mean remission rates for augmentation and switching were 27 and 22 percent; response (reduction of baseline symptoms ≥50 percent) rates for augmentation and switching were 38 and 40 percent [33].
●A prospective observational study compared outcomes in 269 patients who selected augmentation of citalopram (with bupropion or buspirone) and 269 patients who selected switching (from citalopram to bupropion, sertraline, or venlafaxine) as next-step treatment; propensity scoring was used to match the two groups for potential confounders observed at baseline [34]. The probability of remission for the two groups was comparable.
In some cases, switching antidepressants may be more efficacious than augmentation. One open-label, 10-week trial randomly assigned patients who did not remit with venlafaxine (n = 112) to switch to imipramine or to add-on treatment with mirtazapine [35]. Remission occurred in more patients who switched to imipramine than those who added mirtazapine (71 versus 39 percent).
Augmentation
Overview — Standard augmentation strategies for managing treatment-resistant depression include pharmacotherapy and psychotherapy [7]. The choice between the two is generally based upon availability and patient preference because there is no compelling evidence that one is superior to the other for acute outcomes [36]. Pharmacotherapy is typically used for augmentation because it is more available and often preferred. However, patients acutely ill with unipolar major depression who improve with pharmacotherapy and subsequently discontinue it appear to be at greater risk for relapse, compared with patients who improve with and discontinue psychotherapy (eg, CBT). (See "Unipolar depression in adults: Continuation and maintenance treatment", section on 'Relapse/recurrence in the absence of treatment'.)
Patients who do not remit with either pharmacotherapy or psychotherapy may respond to add-on treatment with the other modality. In a study of patients with unipolar major depression who were initially randomized to monotherapy for 12 weeks with an antidepressant or CBT, 112 did not remit and subsequently received adjunctive treatment with the other modality for another 12 weeks [37]. Remission with combination treatment occurred in 62 percent. Higher levels of anxiety during both treatment phases was associated with lack of remission during combination treatment.
Treatment-resistant depression that is managed with adjunctive pharmacotherapy or psychotherapy may also benefit from supplementary interventions such as exercise. (See "Unipolar major depression in adults: Choosing initial treatment", section on 'Supportive care'.)
Pharmacotherapy — For patients with treatment-resistant depression who receive augmentation, drug-drug interactions between antidepressants and add-on medications (eg, second-generation antipsychotics, lithium, or triiodothyronine) are generally not a problem. However, combining a monoamine oxidase inhibitor (MAOI) with another antidepressant, such as a selective serotonin reuptake inhibitor (SSRI), can cause the serotonin syndrome or a hypertensive crisis [8,38]. Specific interactions between an antidepressant and another medication may be determined using the Lexicomp drug interactions tool (Lexi-Interact Online) included in UpToDate.
Treatment-resistant patients who are treated with an add-on drug and do not respond within 6 to 12 weeks of reaching the target dose, or do not tolerate the combination, should be treated with a second medication combination. We suggest tapering and discontinuing the failed adjunctive medication over one to two weeks at the same time another adjunctive medication is started and titrated up. The failed medication is generally tapered by the same amount for each dose decrease. As an example, aripiprazole 15 mg per day is decreased by 5 mg per day every one to three days. At the same time, the antidepressant is continued at the same dose. However, it is reasonable to switch the antidepressant after the adjunctive drug has been switched. Switching antidepressants is discussed elsewhere in this topic. (See 'Switching to a different treatment' below.)
Choosing a drug — For patients with mild to moderate depression who are treatment resistant and receiving augmentation with pharmacotherapy, several options are available. The most widely studied drugs include [7,24,31,39-44]:
●Second-generation antipsychotics
●Second antidepressant from a different class
●Thyroid hormone
Our specific choice of add-on therapy depends upon the degree to which patients benefit from the antidepressant that is prescribed at first presentation, that is, whether patients achieve little symptom relief (minimal response) or definite symptom relief that is not satisfactory (partial response). This approach is described in the subsections below. (See 'Minimal response to initial treatment' below and 'Partial response to initial treatment' below.)
However, it is reasonable for clinicians to select any of these four add-on pharmacotherapy options, because in the few head-to-head trials that compared different drugs, efficacy was generally comparable [1]. As an example:
●An eight-week randomized trial compared paroxetine plus risperidone, paroxetine plus trazodone, and paroxetine plus thyroid hormone in 140 treatment-resistant patients and found that remission was statistically comparable (27, 43, and 38 percent of patients) [29].
●A six-week, open-label randomized trial compared adjunctive quetiapine (target dose 300 mg per day) with adjunctive lithium (target serum concentration 0.6 to 1.2 mEq/L [0.6 to 1.2 mmol/L]) in 450 treatment-resistant patients; remission was comparable (32 and 27 percent) [45].
In addition, a network meta-analysis of 48 randomized trials (n >6000 depressed patients) evaluated the efficacy of augmentation agents by using results from direct comparisons between the drugs (in head-to-head trials), as well as indirectly comparing drugs through their relative effect with a common comparator (typically placebo) [46]. Response (reduction of baseline symptoms ≥50 percent) or remission occurred more often with add-on aripiprazole, lithium, olanzapine, quetiapine, risperidone, or thyroid hormone (T3 or T4), compared with placebo; the benefits of each active treatment were comparable. However, discontinuation of treatment due to side effects was greater with aripiprazole, lithium, olanzapine, and quetiapine, compared with placebo. Information about meta-analyses, including network meta-analyses, is discussed separately. (See "Systematic review and meta-analysis".)
Other factors to consider in choosing add-on pharmacotherapy include past response, safety, comorbid general medical conditions, ease of use, patient preference, and cost. As an example, patients with a history of extrapyramidal side effects should avoid aripiprazole; overweight patients should avoid quetiapine, risperidone, and olanzapine; patients with renal or thyroid disease should avoid lithium; and patients with compromised cardiovascular function should avoid thyroid hormone. In addition, adding a second antidepressant may involve fewer baseline laboratory tests and monitoring than adding an antipsychotic, lithium, or triiodothyronine.
Minimal response to initial treatment — For patients who obtain little symptom relief (eg, reduction of baseline symptoms <25 percent) and can tolerate the initial antidepressant, we suggest augmenting antidepressants with a second-generation antipsychotic as first-step treatment and augmentation with lithium for patients who do not respond to second-generation antipsychotics. However, it is reasonable to first augment with lithium. The rationale for using second-generation antipsychotics or lithium is that their efficacy is better established, compared with other drugs that are commonly used for augmentation, such as a second antidepressant or thyroid hormone. However, adverse effects also appear to be greater with second-generation antipsychotics and lithium, and it is thus reasonable to augment with a second antidepressant or thyroid hormone instead.
For patients with treatment-resistant depression who do not respond to one second-generation antipsychotic within 6 to 12 weeks of reaching the target dose, or do not tolerate the drug, we suggest tapering and discontinuing the failed medication over one to two weeks at the same time that a different second-generation antipsychotic is started and titrated up. We generally attempt treatment with no more than two second-generation antipsychotics before augmenting with lithium.
Evidence supporting add-on therapy with a second-generation antipsychotic includes a study that enrolled patients (n = 103) who continued to meet diagnostic criteria for unipolar major depression despite treatment with an SSRI [47]. Patients were randomly assigned to open-label augmentation with aripiprazole (mean dose 3 mg/day) or bupropion (mean dose 199 mg/day) for six weeks. Although most outcomes were comparable for the two groups, remission occurred in more patients who received adjunctive aripiprazole than bupropion (55 versus 34 percent). All-cause discontinuation of treatment was comparable for aripiprazole and bupropion (25 versus 23 percent of patients), no patients discontinued treatment because of adverse events, and the incidence of extrapyramidal symptoms and of akathisia was comparable for the two groups.
●Specific second-generation antipsychotic – In choosing a second-generation antipsychotic as augmentation for treatment-resistant depression, our general order of preference is as follows, based upon the evidence of benefits and harms, as well as formulary availability: aripiprazole or brexpiprazole, quetiapine, risperidone, and less often, cariprazine, ziprasidone, or olanzapine [36,41,46,48-50]. If aripiprazole fails because of intolerance due to akathisia, we discontinue aripiprazole and initiate brexpiprazole, based upon studies that indicate akathisia occurs in approximately half as many patients taking brexpiprazole than aripiprazole [51]. Aripiprazole is often used first because it appears to cause fewer side effects than other second-generation antipsychotics (table 4). However, few head-to-head trials have compared the second-generation antipsychotics for augmentation and it is reasonable to use these drugs in a different sequence.
Multiple studies indicate that adjunctive treatment with several second-generation antipsychotics can be efficacious for unipolar, nonpsychotic major depression that has not responded to antidepressant monotherapy [11]. However, adverse effects and discontinuation of treatment are more likely with these drugs, compared with antidepressant monotherapy. In addition, many of the trials that studied the antipsychotics were funded by the manufacturer.
Evidence regarding the benefits and harms of specific second-generation antipsychotic drugs includes the following:
•Aripiprazole, olanzapine, quetiapine, or risperidone – A pooled analysis of 16 randomized trials compared adjunctive aripiprazole, olanzapine, quetiapine, or risperidone with placebo in 3480 patients with nonpsychotic, unipolar major depression who failed at least one course of antidepressant monotherapy [41]. Remission occurred in more patients who received an adjunctive antipsychotic compared with placebo (31 versus 17 percent). However, discontinuation because of adverse effects was greater in patients who received an antipsychotic (9 versus 2 percent). Specifically, the pooled rate of discontinuation of treatment due to adverse events for aripiprazole, olanzapine, quetiapine, and risperidone was 4, 11, 12, and 7 percent.
In addition, a meta-analysis of 11 randomized trials compared adjunctive aripiprazole, olanzapine, quetiapine, or risperidone with placebo in more than 3000 patients and found that the efficacy of atypical antipsychotic augmentation may increase with increasing treatment resistance [52]. As an example, augmentation may be more helpful for patients who do not respond to three or four antidepressant trials, compared with patients who do not respond to one antidepressant trial.
-Aripiprazole – A subsequent 12-week, open-label, randomized trial enrolled veterans (n = 1522) with unipolar major depression who remained depressed despite treatment with at least one course of antidepressant therapy [24]. Patients were assigned to one of three treatment strategies: augment the current antidepressant with aripiprazole (target dose 5 to 15 mg/day), augment with bupropion sustained release (target dose 300 to 400 mg/day), or switch to bupropion. Response (reduction of baseline symptoms ≥50 percent) was greater in the augment aripiprazole group than either the augment bupropion group or switch group (74 versus 66 and 62 percent). Adverse effects that occurred more often in the aripiprazole group included akathisia, somnolence, and weight gain, as well as multiple abnormal laboratory tests, whereas patients in the two groups that received bupropion were more likely to become anxious.
-Olanzapine – Pooled analyses have found that adjunctive olanzapine combined with fluoxetine is efficacious for treatment-resistant depression [46,53,54]. Nevertheless, we often avoid olanzapine, especially for longer-term treatment (eg, ≥12 weeks), because among the second-generation antipsychotics that are used for depression, it carries the highest risk of weight gain and metabolic disturbances such as diabetes [55-57].
•Brexpiprazole – In one randomized trial, patients who did not respond to an antidepressant were randomly assigned to add-on brexpiprazole (2 to 3 mg/day; n = 191), quetiapine extended release (150 to 300 mg/day; n = 99), or placebo (n = 205) [58]. Improvement of depression was greater with brexpiprazole than placebo, whereas quetiapine did not separate from placebo. However, the study did not report whether improvement with brexpiprazole was greater than that with quetiapine. In addition, the study was funded by the manufacturer of brexpiprazole.
•Cariprazine – An eight-week trial enrolled patients (n = 808) with an inadequate response to ongoing treatment with SSRIs, serotonin-norepinephrine reuptake inhibitors (SNRIs), or other antidepressants and randomly assigned them to add-on cariprazine 1 to 2 mg/day, cariprazine 2 to 4.5 mg/day, or placebo [59]. Response occurred in more patients who received lower or higher dose cariprazine, compared with placebo (48 and 49 versus 38 percent). However, discontinuation of treatment due to adverse events was two to four times greater with lower or higher dose cariprazine than placebo (7 and 13 percent versus 3 percent).
•Ziprasidone – One trial found that both response and discontinuation of treatment due to side effects were greater with add-on ziprasidone than placebo [60].
Additional information about the efficacy of second-generation antipsychotics for treatment-resistant depression, including patients with a minimal response to the initial antidepressant, as well as information about safety issues (eg, metabolic syndrome and tardive dyskinesia) and the administration and side effects of these drugs (table 4), is discussed separately. (See "Unipolar depression in adults: Treatment with second-generation antipsychotics" and "Second-generation antipsychotic medications: Pharmacology, administration, and side effects".)
First-generation antipsychotics are typically not used for nonpsychotic, treatment-resistant major depression due to concerns about tardive dyskinesia, and because there are less efficacy data for these drugs than second-generation antipsychotics [61].
●Lithium – Lithium augmentation has been used for treatment-resistant depression since the 1960s [62], and multiple studies have subsequently demonstrated its efficacy [36]. As an example, a meta-analysis of nine randomized trials (237 patients) compared adjunctive lithium with placebo and found that response was superior with lithium [63]. In addition, subgroup analyses found that lithium was efficacious for augmenting either first-generation antidepressants (eg, tricyclics) or second-generation antidepressants (eg, SSRIs). Another possible benefit of lithium is reduced risk of suicide. (See "Suicidal ideation and behavior in adults", section on 'Pharmacotherapy'.)
Nevertheless, augmentation with drugs other than lithium may be preferred because lithium can be difficult to use due to the risk of toxicity, the need to monitor serum concentrations, and adverse effects. The efficacy, dose, safety issues, and side effects of lithium in treatment-resistant depression are discussed separately. (See "Unipolar depression in adults: Treatment with lithium".)
Partial response to initial treatment — For patients who obtain definite symptom relief that is not satisfactory (partial response; eg, reduction of baseline symptoms 25 to 49 percent) and can tolerate the initial antidepressant, we suggest augmentation with a second antidepressant from a different class as first-step treatment and augmentation with thyroid hormone (typically triiodothyronine) for patients who do not respond to a second antidepressant. However, it is reasonable to first augment with triiodothyronine.
For depressed patients with partial response to initial treatment with an antidepressant, the rationale for using a second antidepressant or thyroid hormone is that their adverse effects appear to be less, compared with other drugs that are commonly used for augmentation, such as second-generation antipsychotics and lithium. However, the efficacy of a second antidepressant or thyroid hormone is less well established, and it is thus reasonable to augment with a second-generation antipsychotic or lithium instead.
A reasonable alternative to a second antidepressant or thyroid hormone is buspirone. In addition, for patients with a partial response and either cognitive impairment secondary to depression or comorbid attention-deficit hyperactivity disorder, we augment with stimulants. (See "Unipolar major depression in adults: Augmentation of antidepressants with stimulants and stimulant-like drugs".)
●A second antidepressant – Depressive syndromes that respond incompletely to antidepressant monotherapy are often treated by adding a second antidepressant from a different class (commonly referred to as combination therapy). However, the evidence suggests that the clinical benefit of combination therapy is generally limited:
•A meta-analysis of 20 randomized trials compared antidepressant combination therapy with antidepressant monotherapy in patients (n >4500) with depressive disorders that did not respond to initial treatment [64]. Improvement was greater with combination therapy, and withdrawal from treatment due to adverse effects appeared to be comparable for the two groups. However, the clinical advantage of combination therapy was small.
The results suggested that it is preferable to combine a monoamine reuptake inhibitor (SSRI, SNRI, or tricyclic antidepressant) with mianserin, mirtazapine, or trazodone. In addition, antidepressant combinations that included bupropion were superior to monotherapy.
•A 10-week, open-label randomized trial compared add-on citalopram with add-on lithium in patients (n = 104) who did not respond to initial treatment with imipramine [65]. Citalopram was titrated up to 30 mg/day and target lithium serum concentrations were 0.6 to 0.8 mEq/L (0.6 to 0.8 mmol/L). Remission occurred in more patients who received adjunctive citalopram than lithium (40 versus 21 percent). Tolerability was not systematically assessed.
Additional information about the use and efficacy of antidepressant combinations is discussed separately. (See "Unipolar depression in adults: Treatment with antidepressant combinations", section on 'Treatment resistant depression'.)
Certain antidepressant combinations should be avoided; as an example, an MAOI plus an SSRI, an SNRI, a serotonin modulator, an atypical antidepressant, or a tricyclic can cause the serotonin syndrome or a hypertensive crisis [38]. (See "Serotonin syndrome (serotonin toxicity)" and "Monoamine oxidase inhibitors (MAOIs): Pharmacology, administration, safety, and side effects".)
Specific interactions between antidepressants may be determined using the Lexicomp drug interactions tool (Lexi-Interact Online) included in UpToDate.
●Thyroid hormone – Thyroid hormone (eg, triiodothyronine) has been used as augmentation for treatment-resistant depression since the 1960s [66], and subsequent studies have provided low-quality evidence supporting its efficacy. As an example, a meta-analysis of four randomized trials (95 depressed patients unresponsive to a tricyclic) compared adjunctive triiodothyronine (T3) to a control condition (either adjunctive placebo or thyroxine [T4]) [43]. Although the clinical benefit of T3 was moderately large, and response (reduction of baseline depressive symptoms ≥50 percent) occurred in 53 percent more patients who received T3, the difference in the frequency of response between T3 and the control condition was not statistically significant (relative response 1.53, 95% CI 0.70-3.35), and heterogeneity across studies was significant. In addition, it is not clear whether T3 augmentation is efficacious with antidepressants other than tricyclics [67,68]. The efficacy, dose, safety issues, and side effects of thyroid hormone in treatment-resistant depression are discussed separately. (See "Unipolar depression in adults: Augmentation of antidepressants with thyroid hormone" and "Unipolar depression in adults: Treatment with lithium", section on 'Triiodothyronine (T3)'.)
Psychotherapy — For patients with unipolar major depression who receive an antidepressant as initial treatment and do not improve sufficiently, augmentation with psychotherapy is often beneficial [7,69,70]. As an example:
●A meta-analysis of six randomized trials compared antidepressants plus add-on psychotherapy with antidepressants alone in 635 patients with treatment-resistant depression [71]. Remission was nearly twice as likely with adjunctive psychotherapy (relative risk 1.9, 95% CI 1.5-2.5), and discontinuation of treatment was comparable for the two groups. Heterogeneity across studies was small to moderate.
●A subsequent meta-analysis of 20 randomized trials compared antidepressants plus add-on psychotherapy with antidepressants alone in nearly 3000 patients [72]. The primary findings included the following:
•Improvement was greater with adjunctive psychotherapy and the clinical benefit was small to moderate. However, heterogeneity across studies was moderate.
•The efficacy of specific treatments, including CBT, interpersonal psychotherapy, and mindfulness-based cognitive therapy (MBCT), appeared to be comparable.
•Greater baseline severity (intensity) of depressive symptoms was associated with greater improvement.
However, treatment-resistant patients may decline psychotherapy despite its demonstrated benefits [15]. In addition, psychotherapy is often not available.
Psychotherapies suitable for add-on treatment in patients with resistant depression include:
●CBT (see "Overview of psychotherapies", section on 'Cognitive and behavioral therapies')
●Behavioral activation (see "Behavioral activation therapy for treating unipolar major depression")
●Interpersonal psychotherapy (see "Interpersonal Psychotherapy (IPT) for depressed adults: Indications, theoretical foundation, general concepts, and efficacy" and "Interpersonal Psychotherapy (IPT) for depressed adults: Specific interventions and techniques")
●Family and couples therapy (see "Unipolar depression in adults: Family and couples therapy")
●Problem solving therapy (see "Overview of psychotherapies", section on 'Integrated primary and specialty care')
●Psychodynamic psychotherapy (see "Unipolar depression in adults: Psychodynamic psychotherapy")
●Supportive psychotherapy (see "Unipolar depression in adults: Supportive psychotherapy")
The specific choice of an adjunctive psychotherapy is usually based upon availability and patient preference because few head-to-head randomized trials have compared different psychotherapies in treatment-resistant depression. In addition, randomized trials in patients who present for initial treatment of depression indicate that there is no compelling evidence that one psychotherapy is superior to the rest. (See "Unipolar major depression in adults: Choosing initial treatment".)
Specific psychotherapies that have demonstrated efficacy for treatment-resistant depression include the following:
●Cognitive-behavioral therapy – The most widely studied and utilized add-on psychotherapy for treatment-resistant depression is CBT. Multiple randomized trials indicate that adding CBT to pharmacotherapy provides a small to moderate clinical benefit:
•A meta-analysis of three randomized trials compared antidepressants plus add-on CBT with antidepressants alone in 522 patients who did not respond to antidepressants [71]. Improvement was greater with adjunctive CBT, and the clinical effect was small to moderate. Results from the specific trials indicated that adjunctive CBT led to remission in 30 to 40 percent of patients [73,74].
•A second meta-analysis included seven randomized trials in patients who did not respond to pharmacotherapy (sample size not reported), and compared adjunctive CBT plus pharmacotherapy with pharmacotherapy alone [72]. Improvement was greater in patients who received CBT and the clinical benefit was small to moderate.
The principles of CBT are discussed separately. (See "Overview of psychotherapies", section on 'Cognitive and behavioral therapies'.)
●Interpersonal psychotherapy – Evidence supporting add-on interpersonal psychotherapy with pharmacotherapy for treatment-resistant depression includes a meta-analysis of three randomized trials in patients who did not respond to pharmacotherapy (n = 233) [72]. Improvement was greater in those treated with add-on interpersonal psychotherapy, and the clinical effect was small to moderate.
As an example, a four month trial enrolled patients (n = 64) who had not responded to an average of three antidepressants and were acutely depressed on average for approximately 26 months, and randomized them to an intervention consisting of interpersonal psychotherapy, occupational therapy, and pharmacotherapy or to treatment as usual (pharmacotherapy and/or psychotherapy) [75]. Remission occurred in more patients who received interpersonal psychotherapy (35 versus 13 percent).
General information about interpersonal psychotherapy is discussed separately. (See "Interpersonal Psychotherapy (IPT) for depressed adults: Indications, theoretical foundation, general concepts, and efficacy" and "Interpersonal Psychotherapy (IPT) for depressed adults: Specific interventions and techniques".)
●Mindfulness-based cognitive therapy – Multiple randomized trials indicate that adjunctive MBCT can help patients with treatment-resistant depression. A meta-analysis of four trials in 325 patients who had not responded to pharmacotherapy compared ongoing pharmacotherapy plus add-on MBCT with pharmacotherapy alone [72]. Improvement was greater with adjunctive MBCT and the clinical benefit was moderate.
Specific trials indicate that response is two times greater in treatment-resistant patients who receive MBCT than controls:
•In the largest trial of the meta-analysis described immediately above, patients (n = 173) treated with antidepressants for resistant depression were randomized to eight weeks of MBCT or an active control that included exercise, music therapy, and dietary education [76]. The mean length of the current depressive episode was seven years. Response (reduction of baseline symptoms ≥50 percent) occurred in more patients who received adjunctive MBCT than the adjunctive active control (30 versus 15 percent).
•Subsequent to the meta-analysis, a trial enrolled patients (n = 92) with unipolar major depression that did not respond to pharmacotherapy and did not respond to either CBT or interpersonal psychotherapy, and randomly assigned them to eight weeks of usual care plus MBCT or to usual care alone [77]. Usual care included pharmacotherapy and/or psychotherapy. Partial remission, defined as ongoing symptoms that no longer met criteria for major depression, occurred in more patients who received add-on MBCT than usual care alone (41 versus 22 percent). In addition, quality of life improved more with MBCT.
General information about mindfulness-based cognitive therapy is discussed separately. (See "Unipolar major depression: Treatment with mindfulness-based cognitive therapy".)
●Psychodynamic psychotherapy – Psychodynamic psychotherapy has also demonstrated efficacy for treatment-resistant unipolar major depression:
•A 20-week randomized trial (n = 60 patients) compared ongoing pharmacotherapy plus weekly add-on psychodynamic psychotherapy with usual care consisting of pharmacotherapy and/or psychotherapy (eg, CBT) [78]. Remission occurred in more patients who received add-on psychodynamic psychotherapy than usual care (36 versus 4 percent).
•An 18-month randomized trial compared usual care plus add-on long-term psychodynamic psychotherapy (60 sessions) with usual care alone in 129 patients [79]. Usual care was administered according to practice guidelines; however, short-term psychotherapies included in the guidelines were disallowed for patients assigned to adjunctive psychodynamic psychotherapy. Although partial remission at the end of treatment was comparable for the two groups, follow-up assessments revealed that partial remission occurred in more patients who received adjunctive psychodynamic psychotherapy than usual care alone, starting at month 24 and persisting at the final assessment at month 42 (30 versus 4 percent).
The principles and administration of psychodynamic psychotherapy are discussed separately. (See "Unipolar depression in adults: Psychodynamic psychotherapy".)
Switching to a different treatment
Overview — For patients with treatment-resistant depression who are switching to a different treatment, standard options include:
●A different antidepressant
●Psychotherapy
●TMS
The choice between these three options is generally based upon availability and patient preference because there is no compelling evidence that one is superior to the others for acute outcomes. Most patients switch antidepressants because this option is readily accessible and often preferred. In addition, patients with unipolar depression who do not respond to initial treatment with either pharmacotherapy or psychotherapy may respond to switching to the other modality. (See 'Psychotherapy' below.)
Other factors may determine the choice when switching treatments. As an example, patients acutely ill with unipolar major depression who improve with pharmacotherapy and subsequently discontinue it appear to be at greater risk for relapse than patients who improve with and discontinue psychotherapy (eg, CBT). In addition, maintenance treatment with antidepressants or psychotherapy can forestall relapse. Although maintenance treatment with TMS for unipolar major depression is common, its efficacy has not been demonstrated in randomized trials. (See "Unipolar depression in adults: Continuation and maintenance treatment" and "Unipolar depression in adults: Indications, efficacy, and safety of transcranial magnetic stimulation (TMS)".)
Patients with treatment-resistant depression who switch to a different treatment may also benefit from supplementary interventions such as exercise. (See "Unipolar major depression in adults: Choosing initial treatment", section on 'Supportive care'.)
For treatment-resistant patients who are switching antidepressants, we generally cross-taper, that is, taper and discontinue the failed medication over one to two weeks at the same time that another antidepressant is started and titrated up. The failed medication is generally tapered by the same amount for each dose decrease. As an example, venlafaxine extended release 225 mg per day is decreased by 37.5 to 75 mg per day every one to three days. Implementing switches is discussed separately. (See "Switching antidepressant medications in adults", section on 'Switching antidepressant medications'.)
If cross-tapering is used for antidepressant switches, clinicians should be aware of overlapping side effect profiles (table 5), as well as potential drug-drug interactions such as the serotonin syndrome (see "Serotonin syndrome (serotonin toxicity)") [80]. Specific interactions between antidepressants may be determined using the Lexicomp drug interactions tool (Lexi-Interact Online) included in UpToDate.
Although some studies suggest that switching antidepressants may not be effective for treatment-resistant depression, the methods used are problematic. As an example, a meta-analysis of eight randomized trials, which enrolled patients (n = 1627) who did not respond to an antidepressant, found that switching antidepressants was no better than continuing the initial antidepressant [81]. However, the duration of treatment with the initial antidepressant was typically inadequate, such that the initial treatment trial lasted only two weeks in 29 percent of the patients and only four weeks in another 38 percent. The duration of an adequate treatment trial with an antidepressant is discussed separately. (See "Unipolar major depression in adults: Choosing initial treatment", section on 'Duration of an adequate trial'.)
Antidepressants — For patients with major depression who are resistant to treatment with an SSRI at first presentation and are switching antidepressants, many options are available (table 6). The most commonly studied antidepressants are as follows, and are presented in our general order of preference based upon the number and quality of randomized trials that studied each option, as well as safety issues, side effect profiles (table 5), potential for drug-drug interactions, and ease of use (algorithm 2) [1,39,82]:
●SNRIs (eg, venlafaxine)
●Atypical antidepressants (eg, bupropion or mirtazapine)
●Tricyclic antidepressants (eg, imipramine or nortriptyline)
●MAOIs (eg, phenelzine or tranylcypromine)
However, it is reasonable to use these drugs in a different sequence, or to switch to a different SSRI at any point in the sequence [83]. In the few head-to-head studies that have compared different antidepressants for treatment-resistant depression, efficacy is often comparable [1]. As an example, an eight-week randomized trial (n = 105 patients) compared switching to either venlafaxine extended release (225 mg per day) or mirtazapine (45 mg per day) and found that remission was comparable (42 and 36 percent) [30]. Other factors to consider in changing antidepressants include treatment history, patient preference, cost, and comorbid general medical conditions. Patients with seizure disorders should avoid bupropion, patients with obesity should avoid mirtazapine, and patients with cardiovascular disease should avoid tricyclics and MAOIs.
We typically use a drug from a different class when switching antidepressants for treatment-resistant depression, rather than switching to an antidepressant within the same class, especially if there are problematic, class-wide adverse effects (eg, sexual dysfunction, which can occur with most SSRIs) [14]. Evidence supporting a switch to an antidepressant in a different class includes a pooled analysis of four randomized trials involving 1496 patients with unipolar major depression who were resistant to initial treatment with an SSRI, and were switched either to a non-SSRI antidepressant (bupropion, mirtazapine, or venlafaxine) or to a different SSRI [82]. Remission was statistically greater in patients who switched to a different drug class (28 versus 24 percent); in addition, discontinuation due to side effects was comparable for the two groups. However, given the modest difference in remission (4 percent), switching from one SSRI to another is reasonable [84].
Further support for switching from one SSRI to another SSRI, rather than to a different drug class, includes an eight week trial in depressed patients (n = 160) who primarily received SSRIs as initial treatment, did not respond, and were then randomized to escitalopram (10 to 20 mg/day) or duloxetine (40 to 60 mg/day) [85]. All-cause discontinuation occurred in fewer patients who received escitalopram than duloxetine (5 versus 19 percent). Patients unresponsive to a second SSRI should be switched to an antidepressant from a different class.
●Venlafaxine – For patients with mild to moderate unipolar major depression who do not respond to initial treatment with an SSRI and are switching antidepressants, we generally choose venlafaxine extended release because it has been most widely studied [86]. However, other SNRIs are reasonable alternatives.
Multiple trials support switching to venlafaxine in treatment-resistant depression [82]. As an example, a pooled analysis of three randomized trials, in 3375 patients with major depression who did not respond sufficiently to initial treatment with an SSRI, compared switching to venlafaxine with switching to a different SSRI [83]. Remission occurred in more patients who received venlafaxine (54 versus 45 percent), and the number of dropouts because of side effects was comparable for the two groups.
The pharmacology, administration, and side effects of venlafaxine and other SNRIs are discussed separately. (See "Serotonin-norepinephrine reuptake inhibitors: Pharmacology, administration, and side effects".)
●Atypical antidepressant – For depressed patients who are resistant to initial treatment with an SSRI, randomized trials indicate that remission is statistically comparable for patients who switch to either an atypical antidepressant (bupropion or mirtazapine) or a different SSRI. Clinicians may expect that an atypical antidepressant will lead to remission in roughly 20 to 40 percent of patients:
•The STAR*D study included a 14-week trial that compared bupropion sustained release (mean dose 283 mg per day) with sertraline (mean dose 136 mg per day) in 477 treatment-resistant patients; medications were administered on an open-label basis and assessment of outcome was blinded [87]. Remission was comparable for bupropion and sertraline (21 and 18 percent), as was tolerability.
•An eight-week trial compared mirtazapine (45 mg per day) with paroxetine (20 mg per day) in 100 treatment-resistant patients and found that remission was comparable (36 and 47 percent), as was tolerability [30].
•An eight-week trial compared mirtazapine (mean dose 30 mg per day) with sertraline (mean dose 120 mg per day) in 250 treatment-resistant patients; remission was comparable (38 and 28 percent) [40,88,89]. Adverse events that occurred more often with mirtazapine included sedation, fatigue, weight gain, and dry mouth.
The pharmacology, administration, and side effects of atypical antidepressants are discussed separately. (See "Atypical antidepressants: Pharmacology, administration, and side effects".)
●Tricyclic antidepressant – Tricyclic antidepressants are fourth- or fifth-line drugs for treatment of depression due to their greater safety hazards (eg, cardiotoxicity and potential lethality with overdose) and less favorable side effect profiles [90]. However, for patients with treatment-resistant depression, the efficacy and tolerability of tricyclics may be comparable to other antidepressants and SSRIs, and clinicians may expect that a tricyclic will lead to remission in 20 to 70 percent of patients [91]:
•In the STAR*D study, an open-label, 14-week randomized trial (n = 235) compared switching to nortriptyline (mean dose 97 mg per day) with switching to mirtazapine (mean dose 42 mg per day) and found that remission was statistically comparable (20 and 12 percent), as was tolerability [92].
•A 12-week study enrolled 168 patients who did not respond to randomly assigned treatment with either imipramine or sertraline and crossed the patients over (under double-blind conditions) to the alternate drug [93]. Remission in patients who switched from sertraline to imipramine, or from imipramine to sertraline, was comparable (23 and 32 percent). However, discontinuation of treatment due to side effects was greater among patients who switched to imipramine (9 versus 0 percent).
•In one study, patients were initially treated with venlafaxine; those who did not remit (n = 112) were randomly assigned to switch to imipramine or to add-on treatment with mirtazapine 30 mg/day [35]. Imipramine was dosed to achieve a combined serum imipramine plus desipramine concentration of 175 to 300 ng/mL. After 10 weeks, remission occurred in more patients who switched to imipramine than those who added mirtazapine (71 versus 39 percent). However, all-cause discontinuation was two-fold greater with imipramine (9 and 4 percent).
The pharmacology, administration, and side effects of tricyclics are discussed separately. (See "Tricyclic and tetracyclic drugs: Pharmacology, administration, and side effects".)
●Monoamine oxidase inhibitor – MAOIs are seldom prescribed because of potentially lethal drug-drug and drug-food interactions, as well as adverse effects and the danger that MAOIs pose in overdoses [90,94]. Nevertheless, switching to an MAOI may be beneficial for patients with major depression that is resistant to other drug classes (table 6) [83,95]. Evidence that supports switching to MAOIs includes the following:
•A randomized trial included 46 depressed patients who did not respond to imipramine and were crossed over to phenelzine (45 to 90 mg per day) for six weeks under double-blind conditions, and 22 patients who did not respond to phenelzine and were crossed over to imipramine (150 to 300 mg per day) [96]. Response occurred in more patients who switched to phenelzine than imipramine (67 versus 41 percent).
•A network meta-analysis included 52 randomized trials that compared an MAOI to another antidepressant and/or placebo in patients with depressive disorders (n >6400) [97]. The meta-analysis used results from direct comparisons between drugs, as well as indirect comparisons of the drugs through their relative effect with a common comparator (typically placebo). The efficacy (response rate) and acceptability (all-cause discontinuation) of moclobemide, phenelzine, selegiline, and tranylcypromine were comparable to the three SSRIs and six tricyclics that were included. Among the 13 antidepressants that were studied, the drug with the greatest efficacy was phenelzine.
Although some evidence suggests that major depression with atypical features (hypersomnia, hyperphagia, rejection sensitivity, heavy or leaden feelings in limbs, and/or mood brightens in response to positive events) may respond preferentially to MAOIs [98], other studies call into question the clinical utility of diagnosing major depression with atypical features. (See "Unipolar depression in adults: Clinical features", section on 'Atypical'.)
The pharmacology, administration, dietary restrictions (table 7), and side effects of MAOIs are discussed separately. (See "Monoamine oxidase inhibitors (MAOIs): Pharmacology, administration, safety, and side effects".)
●Serotonin modulators – In addition, it is reasonable to switch patients from an ineffective SSRI to a serotonin modulator such as vortioxetine and vilazodone [44]. (See "Serotonin modulators: Pharmacology, administration, and side effects".)
Preliminary studies have examined the clinical utility of selecting an antidepressant based upon tests that assess genes involved in a drug’s pharmacokinetics and pharmacodynamics [99,100]. However, using these pharmacogenetic tests is not standard practice because they do not consistently lead to clinically meaningful outcomes for treating depression [101]. As an example, a 24-week, open-label randomized trial in patients with treatment-resistant unipolar major depression (n = 1944) compared pharmacogenetic testing with usual care for choosing a new antidepressant [102]. Although remission at week 12 was greater with testing-guided care than usual care, the clinical effect was small and at week 24, remission in the genomic testing and control groups was nearly identical (17 and 16 percent).
Psychotherapy — For treatment-resistant major depression, switching to psychotherapy is a reasonable option for patients who prefer it [44,69]:
●A 12-week trial enrolled 122 patients who did not respond to or tolerate citalopram and randomly assigned them to switch to cognitive therapy or to a different antidepressant (bupropion, sertraline, or venlafaxine) [103]. Remission was statistically comparable for cognitive therapy and pharmacotherapy (25 and 28 percent of patients), as was discontinuation of treatment due to side effects (17 and 27 percent). Acceptability of cognitive therapy was greater among patients with more education (eg, college graduates) and a family history of mood disorders [15].
●In one study, patients with persistent depressive disorder (dysthymia) who were initially randomized to nefazodone or CBT and did not respond, were switched to the other modality for 12 weeks [104]. Remission in patients switched from nefazodone to CBT (n = 61, mean number of sessions = 17) or from CBT to nefazodone (n = 79, mean dose 466 mg/day) was comparable (36 and 27 percent). However, all-cause discontinuation of treatment occurred in fewer patients who received CBT than nefazodone (13 versus 28). Nefazodone is often not used due to case reports of life-threatening hepatic failure.
Although switching from an antidepressant to psychotherapy can be effective for treatment-resistant depression, many patients decline this option [15]. In addition, psychotherapy is often not available.
Transcranial magnetic stimulation — For patients with major depression that does not respond to initial treatment with an antidepressant and who decide to switch treatment, another reasonable option is repetitive TMS. Meta-analyses of randomized trials that compared TMS with sham treatment (placebo) indicate that TMS can be efficacious for acute treatment-resistant depression. (See "Unipolar depression in adults: Indications, efficacy, and safety of transcranial magnetic stimulation (TMS)", section on 'Acute TMS'.)
However, patients who respond to acute TMS generally require continuation and maintenance treatment with antidepressants, psychotherapy, or both. It is not known if maintenance treatment with TMS for unipolar major depression is beneficial. The efficacy of maintenance treatment with antidepressants, psychotherapy, or TMS are discussed separately, as is the technique for performing TMS. (See "Unipolar depression in adults: Continuation and maintenance treatment" and "Unipolar depression in adults: Indications, efficacy, and safety of transcranial magnetic stimulation (TMS)" and "Technique for performing transcranial magnetic stimulation (TMS)".)
SEVERE DEPRESSION — Severe major depression is indicated by a score ≥20 points on the self-report Patient Health Questionnaire – Nine Item (PHQ-9) (table 2). Additional information about the PHQ-9 is discussed separately. (See "Using scales to monitor symptoms and treat depression (measurement based care)", section on 'Patient Health Questionnaire - Nine Item'.)
Alternatively, one study classified episodes of major depression as severe in those individuals who had eight or nine of the nine symptoms that define major depression (table 1) [4]. Many studies have assessed severity of depression using clinician administered instruments, such as the Hamilton Rating Scale for Depression (table 3) [5] or Montgomery-Asberg Depression Rating Scale (figure 1A-C) [6]. However, these rating scales are generally not used as part of standard care.
Patients who are severely ill with major depression often report suicidal ideation and behavior, typically demonstrate obvious impairment of functioning, and are more likely to develop complications such as psychotic features and catatonic features. These patients should be referred to a psychiatrist for management and frequently require hospitalization [17,18]. Treatment of major depression with psychotic features or catatonia is discussed separately. (See "Unipolar major depression with psychotic features: Acute treatment" and "Catatonia: Treatment and prognosis".)
Choosing treatment — For patients with severe unipolar major depression that is treatment resistant, electroconvulsive therapy (ECT) is often the treatment of choice [16-19,105,106]. Indications for ECT as first-line treatment include [106-110]:
●Persistent suicidal ideation with intent (see "Suicidal ideation and behavior in adults")
●Severe weight loss, malnutrition, or dehydration secondary to refusal of food and fluids
●Malignant catatonia (see "Catatonia: Treatment and prognosis", section on 'Treatment algorithm')
ECT may also be indicated for psychotic features (eg, delusions or hallucinations), as well as prior favorable response and patient preference. (See "Unipolar major depression with psychotic features: Acute treatment", section on 'Electroconvulsive therapy'.)
Prior to using ECT, many patients receive a few (eg, one to four) courses of next-step pharmacotherapy using the same regimens that are employed for patients with mild to moderate depression that is treatment resistant; these medication options include antidepressant monotherapy (see 'Pharmacotherapy' below) as well as combination treatment with an antidepressant plus a second drug and/or psychotherapy (see 'Mild to moderate depression' above). Although response to ECT may be decreased in patients who have not responded to multiple medication trials, the data are inconsistent. (See "Unipolar major depression in adults: Indications for and efficacy of electroconvulsive therapy (ECT)", section on 'Predictors of response'.)
For severely depressed patients who do not respond to ECT or who decline ECT, we suggest pharmacotherapy, as well as psychotherapy if it is feasible. However, a reasonable alternative is transcranial magnetic stimulation (TMS). (See 'Pharmacotherapy' below and 'Psychotherapy' below and 'Transcranial magnetic stimulation' below.)
Electroconvulsive therapy — ECT is superior to pharmacotherapy for unipolar major depression, based upon meta-analyses of randomized trials [111-113]. As an example, a meta-analysis of 18 trials (1144 patients) compared ECT with pharmacotherapy and found that ECT was more efficacious [114]. In one open-label randomized trial that compared ECT with paroxetine in 39 patients with treatment-resistant depression (mean number of failed antidepressant trials was five), response (reduction of baseline symptoms ≥50 percent) occurred in more patients who received ECT than paroxetine (71 versus 28 percent) [115].
A review of studies that compared ECT with ketamine for treatment-resistant depression concluded that the literature is insufficient to determine their relative efficacy [116]. Nevertheless, perhaps the most rigorous study suggests that ECT is superior to ketamine. A small, two-week randomized trial compared six sessions of ECT (bifrontal or high-dose right unilateral) with intravenous ketamine (0.5 mg/kg) in patients hospitalized for unipolar or bipolar depression (n = 25) [117]. Remission occurred in more patients who received ECT than ketamine (92 versus 50 percent). Assessment of adverse cognitive effects was limited because many of the patients were too impaired at baseline to complete the battery of cognitive tests.
The efficacy of ECT is comparable or superior to other neuromodulation interventions for major depressive episodes. A network meta-analysis of randomized trials evaluated the efficacy of nonsurgical neuromodulation interventions by pooling results from direct comparisons between the therapies in head to head trials, as well as indirectly comparing therapies through their relative effect with a common comparator (sham stimulation) [118]. The probability of response was greatest with bitemporal ECT and high-dose right unilateral ECT, relative to repetitive TMS and transcranial direct current stimulation.
ECT is generally regarded as the most efficacious treatment for major depression [17-19,107,119,120], and use of ECT for severe, treatment-resistant depression is consistent with recommendations in multiple practice guidelines [10,16-19,106,121]. However, ECT is associated with safety risks, adverse effects, logistical constraints, and patient refusal, and relapse rates following remission are high, especially in patients with treatment-resistant depression [122,123]. An overview of ECT is discussed separately, as are indications for and efficacy of ECT in unipolar major depression, medical consultation for ECT, and the technique for performing ECT. (See "Overview of electroconvulsive therapy (ECT) for adults" and "Unipolar major depression in adults: Indications for and efficacy of electroconvulsive therapy (ECT)" and "Medical evaluation for electroconvulsive therapy" and "Technique for performing electroconvulsive therapy (ECT) in adults".)
Pharmacotherapy — Pharmacotherapy options for severe, treatment-resistant depression include antidepressants, intravenous ketamine, and intranasal esketamine. No head-to-head randomized trials have compared these three options; thus, the choice depends upon availability and patient preference. Factors that influence the choice include the following [124]:
●Efficacy and safety – Efficacy and safety are better established for antidepressants than ketamine or esketamine, because far more patients have been treated with antidepressants in high-quality studies. In addition, the longer-term efficacy of ketamine and esketamine are not as well established, compared with antidepressants.
●Ease of use – Treatment with antidepressants is typically less complicated compared with intravenous ketamine, which requires patients to visit a facility and work with multiple clinicians, such as psychiatrists and anesthesiologists, to make the diagnosis of severe depression, establish the treatment plan, administer the drug, and monitor outcomes. Treatment with antidepressants is also less complicated than intranasal esketamine, which is available only through a Risk Evaluation and Mitigation Strategy program, in which the drug is sold to certified medical offices for specific patients who are enrolled in a registry.
●Psychotic features – Ketamine and esketamine are not suitable for patients with a history of psychosis.
●Onset of action – The onset of action is typically faster with ketamine or esketamine (hours to days) than antidepressants (weeks to months).
●Duration of treatment – Clinicians should prescribe ketamine or esketamine for only a short period of time (eg, one month), and patients who respond to ketamine or esketamine need to transition to maintenance treatment with antidepressants and/or psychotherapy. By contrast, patients who respond to an antidepressant typically continue the drug for maintenance treatment.
Antidepressants — For severely depressed patients who are treatment resistant, it is not clear that one class of antidepressants (table 6) is superior to others. However, some evidence suggests that tricyclics may be preferred [19]:
●A meta-analysis of 25 randomized trials compared tricyclics with selective serotonin reuptake inhibitors (SSRIs) in 1377 hospitalized patients who were not selected for treatment resistance [125]. Although tricyclics were more efficacious than SSRIs, the difference was small, and heterogeneity across studies was significant. In addition, discontinuation of treatment occurred in more patients who received tricyclics than SSRIs (14 versus 9 percent).
●In one study, patients who were initially treated with venlafaxine and did not remit (n = 112) were randomly assigned on an open-label basis to switch to imipramine or to add-on treatment with mirtazapine 30 mg/day [35]. Imipramine was dosed to achieve a combined serum imipramine plus desipramine concentration of 175 to 300 ng/mL. After 10 weeks, remission occurred in more patients who switched to imipramine than those who added mirtazapine (71 versus 39 percent).
Additional information about choosing an antidepressant for treatment resistant, severe depression is discussed elsewhere in this topic in the context of mild to moderate depression. (See 'Antidepressants' above.)
Augmentation of the antidepressant with a second drug is often indicated for treatment-resistant, severe major depression. Examples include patients with:
●Unrelenting suicidal thoughts and intent – Lithium may reduce the risk of suicide. (See "Unipolar depression in adults: Treatment with lithium", section on 'Preventing suicide'.)
●Psychotic features – Meta-analyses of randomized trials have found that an antidepressant plus an antipsychotic is more efficacious than antidepressant monotherapy (or antipsychotic monotherapy). (See "Unipolar major depression with psychotic features: Acute treatment", section on 'Evidence of efficacy'.)
●Catatonia – Depressed patients with catatonia who refuse or do not require ECT are treated with antidepressants plus lorazepam. (See "Catatonia: Treatment and prognosis", section on 'Treatment'.)
Other severely ill, treatment-resistant patients may benefit from augmentation with another medication, such as a second antidepressant. In one study, patients with a baseline score on the 17-item Hamilton Rating Scale for Depression (table 3) ≥24 were initially treated with imipramine, which was dosed to achieve a combined serum imipramine plus desipramine concentration of 175 to 300 ng/mL. Patients who did not remit (n = 104) were randomly assigned, on an open-label basis, to add-on either citalopram titrated up to 30 mg/day or lithium dosed to achieve a serum concentration of 0.6 to 0.8 mEq/L (0.6 to 0.8 mmol/L) [65]. After 10 weeks, remission occurred in more patients who received adjunctive citalopram than adjunctive lithium (40 versus 21 percent).
Additional information about augmentation is discussed elsewhere in this topic, in the context of mild to moderate depression. (See 'Augmentation' above.)
Ketamine or esketamine — Ketamine is a racemic mixture of two enantiomers, S-ketamine (esketamine) and R-ketamine [126]. For patients with severe, treatment-resistant unipolar major depression without psychotic features, who decline or do not respond to ECT, a reasonable alternative to antidepressants is short-term use of esketamine (eg, one to four weeks) [127,128]. Alternatively, patients can benefit from short-term use of intravenous ketamine: one to two infusions per week for one to four weeks [44,124,129,130]. Esketamine or ketamine may be particularly helpful for patients with refractory, active suicidal ideation that includes a plan and intent (algorithm 3). The onset of action for esketamine/ketamine often occurs within hours of administration; thus, one potential indication may be short-term treatment of severe depressive symptoms while patients wait for the delayed effect of concomitant, standard antidepressants.
We suggest that clinicians use esketamine or ketamine cautiously [131], consistent with recommendations from the American Psychiatric Association, the Canadian Network for Mood and Anxiety Treatments, and other experts [124,132,133]. The acute benefit of esketamine/ketamine is only short lived, and the drug can cause adverse effects. As an example, esketamine/ketamine can produce psychotomimetic effects and should thus not be used in patients with previous psychotic symptoms [134]. In addition, esketamine/ketamine may be misused as an intoxicant/euphoriant and can lead to diversion and physiologic and psychological dependence [135,136]. Ketamine is also potentially neurotoxic, particularly with longer-term administration [124,135].
We do not prescribe ketamine or esketamine for self-administration at home due to concerns about safety, including misuse and diversion [124,132]. In addition, intranasal esketamine is available in the United States only through a program entitled Risk Evaluation and Mitigation Strategy, in which the drug is sold to certified medical offices for specific patients who are enrolled in a registry. Patients self-administer the drug in the office and are then monitored for at least two hours by clinicians in the office. Esketamine is kept under lock and key and is not allowed to leave the office. This program is intended to ensure safety and prevent misuse and diversion.
Evidence supporting short-term use of esketamine/ketamine for severe, treatment-resistant unipolar major depression without psychotic features includes multiple meta-analyses of randomized trials that consistently demonstrated ketamine/esketamine can rapidly improve depression, including symptoms of suicidal ideation, for a relatively short period of time (eg, 10 days) and that the drugs are relatively well tolerated [36,128,133,137]. However, evidence for the safety and efficacy nonintravenous formulations of ketamine is limited [133].
For severe, treatment-resistant unipolar major depression without psychotic features, we recommend that treatment with short-term esketamine/ketamine include the following aspects of care [124,132,133,136]:
●Assessment – Prior to the intervention, a psychiatrist should assess the patient’s current and past psychiatric history to make the diagnosis, establish the indication for esketamine/ketamine, and formulate the treatment plan. The evaluation should include the patient’s treatment history and history of substance-related and addictive disorders, as well as psychotic disorders. A baseline urine toxicology screen can corroborate the history regarding substance use disorders.
In addition, an internist or anesthesiologist should establish that patients are medically stable by taking a general medical history with a full review of systems, performing a physical examination, and obtaining weight and vital signs. Laboratory tests (eg, electrocardiogram) are pursued as indicated by the history and examination. Contraindications include poorly controlled hypertension (eg, blood pressure >140/90 mmHg) and unstable general medical illnesses such as cardiovascular or respiratory disease.
●Written informed consent – Prior to the intervention, clinicians should provide information about depression and the risks and benefits of esketamine/ketamine and alternative treatments. Informed consent requires that patients are capable of understanding and acting reasonably on this information and that patients are given the opportunity to consent in the absence of coercion.
●Setting – Administer ketamine in facilities staffed and equipped to provide advanced cardiac life support to manage potential adverse cardiovascular and respiratory effects. Clinicians should also be prepared for managing agitation and other dangerous behavioral symptoms. Esketamine is available in the United States only through certified medical offices for specific patients who are enrolled in a registry.
●Monitoring – Clinicians should monitor the patient’s general medical and psychiatric status during treatment with esketamine/ketamine. This includes assessing cardiovascular and respiratory function (eg, blood pressure, pulse, respiratory rate, and pulse oximetry) during ketamine infusions and subsequently until the patient is discharged from the facility. During the course of treatment, monitoring should include depressive symptoms (table 1); adverse effects such as urinary symptoms, cognitive impairment, dissociation, and psychotomimetic symptoms; and signs and symptoms of ketamine/esketamine use disorder.
●Stopping the ketamine infusion and intervening – Establish criteria for discontinuing the ketamine infusion and intervening. As an example, it is reasonable to stop the infusion and intervene to maintain systolic blood pressure <160 to 180 mmHg and diastolic blood pressure <100 to 110 mmHg.
We recommend that patients who respond to short-term esketamine or ketamine receive maintenance treatment with antidepressants, psychotherapy, or both. Although there are no compelling data to support transitioning patients to a previously untried antidepressant, this approach is nevertheless reasonable and consistent with practice guidelines [133]. Only one randomized trial indicates that maintenance esketamine treatment is efficacious, and the role of esketamine as a maintenance treatment continues to evolve and remains uncertain. Some clinicians offer continuation and maintenance ketamine treatment [138,139], but there are no high-quality studies that indicate this practice is efficacious or safe [124,131,133].
A small, short trial included 34 patients who initially received four infusions of intravenous ketamine for one week, and were randomly assigned to lithium (up to 1200 mg/day) or placebo as continuation treatment for approximately two weeks [140]. Lithium was no better than placebo.
The general topic of maintenance treatment for patients who remit with acute treatment is discussed separately. (See "Unipolar depression in adults: Continuation and maintenance treatment".)
The administration, efficacy, and adverse effects of ketamine and esketamine for treatment-resistant depression are discussed separately. (See "Ketamine and esketamine for treating unipolar depression in adults: Administration, efficacy, and adverse effects".)
Additional information about prescribing intravenous ketamine according to best practices is discussed in the Consensus Statement on the Use of Ketamine in the Treatment of Mood Disorders, which was issued by a task force from the American Psychiatric Association [124].
Psychotherapy — We typically include psychotherapy for severe episodes of major depression that are treatment resistant. Most hospitalized patients receive individual and/or group psychotherapy, provided that they are well enough to participate in therapy [141]. Indirect evidence supporting the use of psychotherapy for severe episodes of treatment-resistant depression includes randomized trials in patients with mild to moderate episodes of treatment-resistant depression (see 'Psychotherapy' above). In addition, randomized trials indicate that patients who present for initial treatment of severe major depression can benefit from pharmacotherapy plus psychotherapy. (See "Unipolar major depression in adults: Choosing initial treatment", section on 'Choosing a treatment regimen'.)
Only low-quality studies have evaluated psychotherapy in severe, treatment-resistant depression. As an example, two relatively small and old studies yielded conflicting results:
●A 12-week randomized trial compared cognitive-behavioral therapy (CBT) plus pharmacotherapy with pharmacotherapy alone in 20 patients hospitalized for chronic (duration ≥2 years) depression; improvement was comparable for the two groups [142].
●A prospective observational study included 174 patients with treatment-resistant depression who were hospitalized for four to seven months and treated with pharmacotherapy, CBT, milieu therapy, occupational therapy, and couples therapy, as well as ECT if indicated [141]. Response (reduction of baseline symptoms ≥50 percent) occurred in 47 percent.
Transcranial magnetic stimulation — Repetitive TMS is another option for severe episodes of treatment-resistant, unipolar major depression. (See 'Transcranial magnetic stimulation' above.)
MAINTENANCE TREATMENT — Patients who remit from treatment-resistant depression typically require maintenance treatment. (See "Unipolar depression in adults: Continuation and maintenance treatment".)
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Depressive disorders".)
INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, “The Basics” and “Beyond the Basics.” The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on “patient info” and the keyword(s) of interest.)
●Basics topics (see "Patient education: Depression (The Basics)" and "Patient education: When you have depression and another health problem (The Basics)")
●Beyond the Basics topics (see "Patient education: Depression in adults (Beyond the Basics)" and "Patient education: Depression treatment options for adults (Beyond the Basics)" and "Patient education: Electroconvulsive therapy (ECT) (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Definitions
•Unipolar major depression – Unipolar major depression (major depressive disorder) is diagnosed in patients who have suffered at least one major depressive episode (table 1) and have no history of mania or hypomania. (See 'Definitions' above and "Unipolar depression in adults: Assessment and diagnosis".)
•Treatment-resistant depression – Treatment-resistant depression typically refers to major depressive episodes that do not respond satisfactorily to at least two trials of antidepressant monotherapy; however, the definition has not been standardized. (See "Unipolar treatment resistant depression in adults: Epidemiology, risk factors, assessment, and prognosis", section on 'Treatment resistant depression'.)
●Patients with mild to moderate depression
•General principles – The general principles and issues that are involved in treating resistant depression in adults include reassessing the diagnosis, comorbidity, assessing adherence, treatment strategies, nonspecific care management, duration of an adequate drug trial, and referral. (See "Unipolar depression in adults: General principles of treating resistant depression".)
•Preferred approach – Our approach in patients with mild to moderate treatment-resistant major depression is to augment the initial antidepressant with a second drug and/or psychotherapy as first-line treatment, rather than switching to other treatments, such as a different antidepressant, psychotherapy, or repetitive transcranial magnetic stimulation (TMS) (algorithm 1). However, switching treatments is a reasonable alternative to augmentation. Patients who cannot tolerate an adequate dose of the initial antidepressant should switch to another antidepressant. (See 'Treatment algorithm' above.)
-For patients who obtain little or no symptom relief and can tolerate the initial antidepressant, we suggest add-on therapy with a second-generation antipsychotic as first-step treatment, rather than augmenting with lithium, a second antidepressant from a different class, or thyroid hormone (Grade 2C). However, lithium, a second antidepressant, or thyroid hormone are reasonable alternatives to a second-generation antipsychotic. Among second-generation antipsychotics, our general order of preference is aripiprazole or brexpiprazole, quetiapine, risperidone, and less often, cariprazine, ziprasidone, or olanzapine. (See 'Choosing a drug' above and 'Minimal response to initial treatment' above.)
-For patients who obtain definite symptom relief that is not satisfactory (partial response) and can tolerate the initial antidepressant, we suggest augmentation with a second antidepressant from a different class as first-step treatment, rather than augmenting with a second-generation antipsychotic, lithium, or thyroid hormone (Grade 2C). We avoid using monoamine oxidase inhibitors as the second antidepressant. However, a second-generation antipsychotic, lithium, or thyroid hormone are reasonable alternatives to a second antidepressant from a different class. (See 'Choosing a drug' above and 'Partial response to initial treatment' above.)
•Next steps – For patients with treatment-resistant depression who augment antidepressants with a second drug and do not respond, we generally provide one to three courses of augmentation before switching the antidepressant or switching to psychotherapy or repetitive TMS. When switching the antidepressant, we typically maintain the current adjunctive drug. (See 'Treatment algorithm' above.)
Our approach in patients who switch antidepressants is to select a drug from a different class (table 6) rather than the same class. In choosing a new antidepressant for patients who fail a selective serotonin reuptake inhibitor, our general order of preference is serotonin-norepinephrine reuptake inhibitors, atypical antidepressants, tricyclics, and monoamine oxidase inhibitors (algorithm 2). However, it is reasonable to use these drugs in a different sequence or to switch to a different selective serotonin reuptake inhibitor or a serotonin modulator at any point in the sequence. (See 'Antidepressants' above.)
•Other options – Other treatment options are available for patients who do not respond satisfactorily to several (eg, 5 to 10) courses of initial and next-step treatments. (See "Unipolar depression in adults: Management of highly resistant (refractory) depression".)
●Patients with severe depression
•Initial approach – For the general population of severely depressed patients who fail two to four next-step pharmacotherapy trials with antidepressant monotherapy or an antidepressant plus an adjunctive drug, we suggest electroconvulsive therapy (ECT) rather than ketamine or esketamine (Grade 2C). However, some patients may prefer ketamine or esketamine. (See 'Choosing treatment' above.)
Additionally, some patients with very concerning symptoms of depression (eg, suicidal behavior and intent, malnutrition/dehydration secondary to food/fluid refusal, or malignant catatonia), regardless of the number of prior treatment trials, are also appropriate candidates for ECT, as discussed elsewhere. (See 'Choosing treatment' above and "Unipolar major depression in adults: Indications for and efficacy of electroconvulsive therapy (ECT)", section on 'Indications' and "Catatonia: Treatment and prognosis", section on 'Malignant catatonia'.)
•Next steps – Treatment for severely depressed patients, who do not respond to or accept ECT or ketamine/esketamine, consists of antidepressants, as well as psychotherapy if it is feasible. We often use tricyclics, but other antidepressant classes (table 6) are reasonable alternatives. (See 'Pharmacotherapy' above and 'Psychotherapy' above.)
ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Wayne Katon, MD, and Paul Ciechanowski, MD, who contributed to earlier versions of this topic review.
