If remission has not been achieved when reevaluated up to six weeks later, consider:

1. Whether adequate engagement of patient/family exists and whether recommendations are being followed (adherence).
2. Optimize antidepressant dose. A systematic review and meta-analysis found higher doses of SSRIs were associated with an increased likelihood of response. Higher doses were also associated with an increased risk of side effects. The overall rate of treatment dropouts was reduced, possibly due to improved efficacy. Balancing these results, the authors recommended titrating SSRI doses in patients who had not responded to lower doses (Jakubovski, 2016). Additionally, it has been well established that raising the dose of tricyclics or MAOIs may improve response. Similarly, a controlled study showed that raising the dose of fluoxetine (from 20 mg to 40 or 60 mg) in partially responsive patients was more effective than adding desipramine (25-50 mg per day) or lithium (300-600 mg daily). In non-responders, raising the fluoxetine dose was as effective as adding lithium, and both were more effective than adding desipramine (Fava, 1994; Perry, 1994).
3. Switching to a different antidepressant medication. After a failed trial of citalopram, remission rates in the STAR*D study were 21.3% for bupropion SR, 17.6% for sertraline and 24.8% for venlafaxine XR (Rush, 2006), although the differences were not statistically significant. Failure of a drug in one family does not rule out possible benefit from other drugs in that family. This is particularly true for SSRIs (Bull, 2002; Thase, 1997; Brown, 1995).
4. Reconsider treatment modality:
   • Adding, switching or substituting treatment modality. A switch from an antidepressant to psychotherapy or vice versa appears useful for non-responders to initial treatment (Schatzberg, 2005). If there is less than 25% reduction of symptoms after six weeks at therapeutic dose (i.e., partial positive response to medication), add, switch or substitute another treatment modality. If there is a partial medication response and side effects are not prohibitive, increase the dose. As part of the evaluation, use a standardized assessment tool to gauge progress.
   • Adding cognitive psychotherapy or adding another medication such as buspirone or bupropion. Both augmentation strategies showed similar improvement rates in the STAR*D study; however, the addition of medication resulted in a significantly more rapid response (Thase, 2007).
5. Reevaluating the diagnosis and the possibility of a bipolar diagnosis. Bipolar patients require a different treatment approach and may not consistently come forward with their hypomanic, mixed or manic histories (Sharma, 2005). Consult with a behavioral health clinician if personality disorders are present.
6. Looking for comorbidities, such as substance abuse issues, and involving addiction specialists as needed.
7. Augmentation strategies (such as lithium or low-dose thyroid, making a referral to psychiatry for possible MAOI treatment). Many patients unresponsive to tricyclics are responsive to monoamine oxidase inhibitors (MAOIs). Rarely, the combination of tricyclics and MAOIs is used. This combination should be undertaken with extreme caution. Studies measuring response to MAOIs in SSRI non-responders have not been done (McGrath, 1994; McGrath, 1993). See the "Augmentation Therapy" section in this annotation for more information.
8. Other strategies: light therapy, ECT and hospitalization. See the "Consider Other Strategies" section in this annotation for more information.

Augmentation Therapy

Augmentation therapy is used for those situations in which the patient's depression is either treatment resistant or partially responsive to treatment. This is a good time to consult and/or refer to a mental health specialist. Augmentation strategies may be considered for partial responders, and combinations of antidepressants (when each has a different mechanism) have been shown to be options in those who fail to achieve remission.

Augmentation methods include:

• Bupropion or buspirone-SSRI combination.
  - Augmentation of citalopram with bupropion or buspirone after non-remission with a trial of citalopram alone yielded a remission rate of 29.7 and 30.1%, respectively, in the STAR*D study. These differences were statistically insignificant, but bupropion SR was better tolerated (Trivedi, 2006a).
  - Three open series of cases and two other case reports have described beneficial results. The basis of this combination is the addition of a noradrenergic agent to a serotonergic agent to enhance effects; bupropion may also have dopaminergic actions (Spier, 1998; Bodkin, 1997; Marshall, 1996).
  - Five open studies supported potential utility of this treatment, and a response rate of approximately 60% was observed (Dimitriou, 1998; Bouwer, 1997).
  
  
• Mirtazapine-SSRI combination.
  - The addition of the alpha-2 antagonist mirtazapine is used to augment SSRI. Three controlled studies have found evidence of more rapid effects (Maes, 1999; Dam, 1998; Cappiello, 1995).
  
  
• T3 augmentation of antidepressants.
  - Antidepressant augmentation with T3 had a remission rate of 24.7% in the STAR*D study (Nierenberg, 2006). There was no significant difference between T3 augmentation or lithium augmentation (13.2%), but T3 was better tolerated, despite being more vigorously dosed (Rush, 2009).
  - Placebo-controlled studies found mixed results. Usual dose of T3 varied between 25 and 50 micrograms/day (Nelson, 2000).
  
  
• Stimulant augmentation of TCA-SSRI ("jump-start response").
  - Some open label studies of modafinil augmentation of SSRI have reported benefit in sleepiness and fatigue, either disease-state-induced or secondary to the SSRI. The sample size and length of treatment are both small, and thus conclusions need to be taken with caution (Schwartz, 2004; Ninan, 2004).
  - Further research with larger higher-quality trials is needed to establish the benefit of stimulant augmentation and the clinical situations where this might be indicated (Candy, 2008; Dunlop, 2007; Fava, 2005).
  - Cases of sudden death, stroke and myocardial infarction have been reported in adults taking stimulant medications at usual doses for ADHD. Adults with serious structural cardiac abnormalities, cardiomyopathy, serious heart rhythm abnormalities, coronary artery disease or other serious cardiac problems should not be treated with stimulant medications.
  
  
• TCA-SSRI combination (caution – elevated TCA level – to be monitored).
  - A 1991 study by Nelson reported this combination to be more rapidly effective and indicated that remission was more likely. The dose of TCA should be adjusted to achieve effective TCA levels because SSRIs may increase TCA levels. Fluoxetine and paroxetine raise TCA (desipramine) levels three- to fourfold, and citalopram and sertraline have modest effects (Nelson, 1991; Preskorn, 1990).
  - If a combination is used, monitor side effects and consider checking blood levels.
  
  
• Lithium augmentation with TCAs. Lithium augmentation with SSRI (caution – case reports of serotonin syndrome).
  - Augmentation with lithium at stage 3 of STAR*D yielded remission rate of 15.9% (Nierenberg, 2006).
  - Seven placebo control studies have found positive evidence of efficacy of lithium augmentation. Combination of lithium and SSRIs have been relatively well studied. In early studies, the usual dose of lithium was 300 mg three times a day. At this dose, serum lithium levels were usually above 0.4 mEq/L (Delgado, 1998; Baumann, 1996; Katona, 1995; Joffe, 1993).
  
  
• Atypical antipsychotic-antidepressant combination.
  - Several studies have been published supporting the use of atypical antipsychotics as augmentation agents with antidepressants for treatment-resistant depression. A meta-analysis study of 1,500 treatment-resistant patients indicated pooled remission and response rates for atypical antipsychotics and placebo were 47.4% vs. 22.3% and 57.2% vs. 35.4%, respectively. The atypical antipsychotics used were risperidone, olanzapine and quetiapine (Papakostas, 2007).
  - A meta-analysis of 16 trials that included a total of 3,480 patients with treatment-resistant, non-psychotic, unipolar major depressive disorder found augmentation with atypical antipsychotics was significantly more effective than placebo in measures of both response and remission. The agents reviewed included risperidone, olanzapine, quetiapine and aripiprazole. No significant differences in efficacy were noted among the reviewed medications. The rate of patient discontinuation due to adverse events was higher in patients receiving augmentation with atypical antipsychotics, compared with placebo (Nelson, 2009).
  - Aripiprazole, quetiapine and the olanzapine-fluoxetine combination are FDA-approved adjunctive agents for the acute treatment of major depressive disorder in adults. In two studies, patients diagnosed with major depressive disorder who had at least two documented trials of incomplete response to antidepressant medications were randomized to aripiprazole (2 mg to 20 mg a day) or placebo. Patients receiving aripiprazole experienced significant improvements in depression symptoms within one to two weeks of initiated aripiprazole. Average doses were approximately 10 mg a day by mouth. Patients receiving aripiprazole experienced higher rates of akathisia and fatigue, compared to those randomized to placebo (Marcus, 2008; Berman, 2007).

Consider Other Strategies

• If patients do not respond to intensive outpatient treatment, partial or full hospitalization may be considered in patients who have not responded to outpatient management, particularly if safety issues are a concern.
• Electroconvulsive treatment is effective and can sometimes be administered safely in an outpatient setting.
• Use of bright light therapy for treatment of major depression with a seasonal specifier is well established.

Treatment-resistant depression has several definitions in the literature. It is important to distinguish treatment resistance from a lack of completion of a full course of treatment. The literature tends to focus on pharmacological treatments in the definition of treatment resistance without consistently incorporating psychotherapeutic modalities. True treatment resistance is seen as occurring on a continuum, from failure to reach remission after an adequate trial of a single antidepressant to failure to achieve remission despite several trials of antidepressants, augmentation strategies, ECT and psychotherapy. For our purposes of making recommendations for primary care clinicians, we define true treatment resistance as failure to achieve remission with an adequate trial of therapy and three different classes of antidepressants at adequate duration and dosage (Nierenberg, 2006; Keller, 2005; Geddes, 2003).

Hospitalization

Partial or full hospitalization may be indicated in patients with unrelenting depressive symptoms, particularly if safety issues are a concern (American Psychiatric Association, 2010).

The most important consideration from a primary care standpoint is the hospital communicating the details of partial and full hospitalization back to primary care, and patients having follow-up visits for chronic or acute physical and psychiatric problems arranged with their clinician prior to hospital discharge. Patients without a primary care provider should be connected with shortly after hospital discharge for a physical assessment and preventive interventions to help decrease the rate of readmission. If no medication reconciliation was done in the hospital, there should be follow-up on it during the primary care visit.

The following are most commonly referred from a primary care setting. For other specialized therapies, see Appendix C, "Specialized Therapies."

Electroconvulsive Therapy (ECT)

Response and remission rates are higher with ECT than with any other form of antidepressant treatment, with 70-90% of patients showing improvement (Kellner, 2006; UK ECT Review Group, The, 2003). Patients may express a choice for ECT; shared decision-making should be engaged to determine if it is appropriate. Electroconvulsive treatment is usually performed on an inpatient basis, but for some individuals, it can be administered safely in an outpatient setting. A patient considering ECT would need to be able to tolerate anesthesia and should consult with a psychiatrist about the risks and benefits (UK ECT Review Group, The, 2003; Sackeim, 2001a).

One study showed that 64.2% of patients referred for ECT achieved major depression remittance (Kellner, 2006). In addition to its use as a treatment in the acute phase, ECT is an effective maintenance therapy for major depression. The same study compared continuous ECT with nortriptyline and lithium treatment and found no difference in relapse rates (Kellner, 2006).

ECT is also effective for treating major mental illness during pregnancy, and the risks of adverse events are low. It should be strongly considered in pregnant women with severe symptoms of mental illness, such as psychotic symptoms, catatonia or strong suicidal urges (Anderson, 2009). For more information regarding the treatment of depression in pregnant women, refer to Appendix D, "Special Populations."

Factors that may suggest a given patient may be an ECT candidate include:

• Geriatric depression (Mitchell, 2005)
• If antidepressant medications have not been tolerated or pose a significant medical risk
• If antidepressant medication trials have not been successful
• If ECT has been successful in previous episodes
• If catatonia is present
• When a rapid response is needed because of severe suicide risk or because the patient's health has been significantly compromised by the depression (e.g., severe cachexia, inability to attend to the activities of everyday living). ECT has been shown to be effective in resolving expressed suicidal intent (Kellner, 2006).
• If depression has psychotic features
• If melancholic symptoms are predominant
• Depression and Parkinsonism

(National Institute for Clinical Excellence, 2003)

Common side effects associated with ECT include headaches, myalgias, nausea, drowsiness, confusion and amnesia. More serious and rare side effects include hypertension, tachycardia, myocardial infarction and cerebrovascular accident. ECT is a relatively safe and effective treatment for patients with treatment-refractory and severe depressive illness, and death is very rare.

Light Therapy

Use of bright light therapy for treatment of major depression with a seasonal specifier is well established (Leppämäki, 2002; Golden, 2005). Additionally, there is evidence to support the use of bright light therapy for other types of depressive symptom patterns, including non-seasonal depression and milder variations of seasonal depressive patterns (Jorm, 2002; Prasko, 2002). For non-seasonal depression, light therapy's benefit as an adjunctive treatment is more robust than its benefit as monotherapy (Freeman, 2010). Bright light therapy may also quicken and enhance the effects of antidepressant medication (Benedetti, 2003). In two small pilot studies, promising results were seen in pregnant and postpartum women with non-seasonal depression (Epperson, 2004; Oren, 2002).

Dosage. The standard starting dose for depression with a seasonal specifier is 10,000 lux for 30 minutes each morning (Freeman, 2010). Research on bright light therapy for other types of depression has not necessarily utilized standard dosages and exposure times.

Side effects. The most common side effects are nausea, jitteriness and headache (Freeman, 2010).

Equipment. It is important for light therapy treatment to utilize equipment that eliminates ultraviolet frequencies and produces bright light of known spectrum and intensity. For these reasons, use of client-constructed light therapy units is contraindicated.

Overall recommendation. The APA Task Force concluded that "light therapy is an evidence-based, effective, well-tolerated treatment for seasonal affective disorder, as well as an augmentation strategy for antidepressant treatment of nonseasonal depression" (Freeman, 2010).

Additional Specialized Therapeutic Options

There are other more specialized therapies available, as well. Refer to psychiatry for consideration of vagus nerve stimulation (VNS), repetitive transcranial magnetic stimulation (rTMS), magnetic seizure therapy (MST) and deep brain stimulation (DBS). See Appendix C, "Specialized Therapies."