Patient P is a divorced 40-year-old woman with a history of depression. She was referred by a primary care colleague for psychiatric consultation secondary to ongoing depressive symptoms after she failed to respond to her first 2 antidepressants.
She had been in treatment for the past year, but her symptoms had not fully abated despite antidepressant treatment. Her symptoms of depression developed after she lost her job because of the current economy. She has had difficulty finding a new job and all of her bills are overdue.
This patient reports that she suffers from depressed mood, sad and tearful affect, anhedonia, low energy, poor motivation, and problems with concentration. During the most severe point in her depression, she was barely able to eat and her energy was so low that she wanted to sleep most of the day and night. She denied suicidal thoughts.
What is a rational next step in treating this patient?Complete the interview with emphasis on differential diagnosis and establishing antidepressant treatment regimen and resistance.So far, we have established that the patient is depressed based on Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria, but we know very little about the type of depression, comorbidity, and previous antidepressant trials. She may or may not have mild, moderate, or severe 'treatment-resistant' depression, depending on this information. No future treatment should be planned without first completing a thorough assessment.
This patient is typical of the outpatients seen at our outpatient mental health clinic. Patients are often referred from outside sources as being treatment resistant in that they continue with symptoms of depression after being treated with antidepressants. In these scenarios, patients have typically been on a few antidepressants, often from the same class, ie, selective serotonin-reuptake inhibitors (SSRIs). Furthermore, they are often on these agents for a limited number of weeks at the minimum therapeutic dose.
Sometimes patients who present with depression and dysphoria are found, on further evaluation, to suffer from anxiety disorders such as generalized anxiety disorder (GAD), post-traumatic stress disorder, substance use disorder, or borderline personality disorder. If major depressive disorder (MDD) is believed to be the primary psychiatric disorder, then these other disorders are often comorbid in nature and considered co-occurring.[1,2] A key aspect of determining if a patient suffers from treatment-resistant depression (TRD) is to make an accurate initial diagnosis and to assess whether the medications tried were utilized for a clinically appropriate duration at a clinically appropriate dose. Upon interviewing the patient, you confirm the diagnosis of MDD with no apparent psychiatric comorbidity. A family history suggests that no one else has suffered from MDD or other mental disorders. Medically, the patient suffers from migraine headaches, but is otherwise healthy. She has no history of psychiatric hospitalizations and has never been suicidal.
Reviewing her medications, you determine that she had initially taken fluoxetine 40 mg/day for at least 3 consecutive months and then was switched to the newest SSRI, escitalopram, which she continues to take at 10 mg/day. She denies major side effects aside from mild daytime fatigue on this agent.
What intervention would you initiate at this point?
Increase the escitalopram to the US Food and Drug Administration (FDA) limit of 20 mg/day .
Because this patient is not responding to the second SSRI, any of these items would fall within the standard of care for treating MDD. There are no definitive data suggesting that any one of these treatments would be superior, but a majority of clinicians would likely maximize the current antidepressant, escitalopram, to the full FDA-allowed 20 mg/day as the first and simplest step. This ensures an adequate dose and duration of treatment in the current clinician's practice. Clearly, not doing something will allow her MDD symptoms to continue.
We elected to increase this patient's escitalopram to the maximum FDA-allowed amount, 20 mg/day, and treated her for an additional 8 weeks. Based on the STAR*D trial[3] and other studies, we have no definitive data to dictate that one approach is superior to another, but increasing the patient's current medication to the maximum dose is a logical next step.
During this period, the patient had noted a decrease in sadness and an improvement in her ability to enjoy things, but also developed increased daytime fatigue. We have now established that this patient has MDD that has failed to respond to 2 full SSRI trials, and that she has been suffering from her symptoms for at least a year. Is she "treatment resistant"? In some ways, the answer is "yes" in that she is still depressed despite initial treatment. (We could even argue that after her initial treatment with a moderate amount of fluoxetine she was treatment resistant as she received no benefit.) At a minimum, patients can only be resistant after a first bona fide treatment occurs and fails to resolve MDD symptoms. This would be the minimum definition of TRD.
More formally, a clinical definition would consider a patient to have TRD if he/she has been on 2 antidepressant treatments (ADTs) at moderate-to-full dosing for several weeks, and each antidepressant would need to be from different chemical classes.[4] The most formal, research definition used would be that proposed by Thase and colleagues,[5] which resembles the staging often used in cancer to designate different levels of illness (Table 1).
Table 1. Staging of Treatment Resistance
Stage Treatment Response
0 No single adequate trial of medication
1 Failure to respond to an adequate trial of 1 medication
2 Failure to respond to 2 different monotherapy trials of medications with different pharmacologic profiles
3 Stage 2 plus failure to respond to augmentation of 1 of the monotherapies
4 Stage 3 plus failure of a second augmentation strategy
5 Stage 4 plus failure to respond to ECT
From Thase ME and Rush AJ[5]
Our current patient would be considered to be in stage 1, in that despite being on 2 reasonable ADTs (antidepressant treatments), these medications were both from the SSRI class and used only 1 antidepressant mechanism of action.
We must also discuss chronicity, because MDD is often chronic or recurrent. Chronic MDD occurs when there are 2 or more years of symptoms without a euthymic break in symptoms. Recurrent MDD occurs when there is a euthymic episode followed by a clear MDD recurrence.[6] Between 25% and 40% of patients with MDD will experience a recurrence within 2 years of the index episode, and 60% will experience recurrence after 5 years.[7] Anywhere from 20% to 35% of patients who experience an initial, index episode of MDD will continue to have a chronic course.[8-10] Therefore, chronicity does not equal treatment resistance, but statistically, they often travel together. Although patients with chronic and recurrent MDD may be more treatment resistant over time, we usually assess treatment resistance in the current major depressive episode only.
After 8 weeks of treatment at a full dose of escitalopram, we re-evaluated the patient's symptoms and discussed her concerns about side effects. She said the improvement was welcome but the adverse effects likely would impair her ability to work successfully. After informed consent, the escitalopram was continued and bupropion XL was added, titrated to 300 mg/day, and dosed for 6 weeks in this manner. At her return visit, she noticed an improvement in her daytime fatigue, but denied other positive symptom responses. We elected to increase the bupropion XL to the FDA maximum of 450 mg/day and asked her to return in 8 weeks. At this most recent return visit she stated that there was not a robust change. During this period, she began a new job. Although her energy greatly improved, she noted only partial improvement in regard to her mood, enjoyment level, affective range, and ability to concentrate fully. She was thought to be a partial responder to this often combined pair of antidepressants.[4]
Questions answered incorrectly will be highlighted.
What stage of treatment resistance is this patient at now?
Stage 3.This patient has now fulfilled 2 stages of treatment resistance in failing at least 1 full clinical trial of an SSRI and full trial of an NDRI, which is from a separate antidepressant class. In clinical discussions, this scenario is often referred to being treatment resistant.
What are the negative connotations of prolonged treatment resistance? The Sequenced Treatment Alternatives to Relieve Depression (STAR*D) initiative was a large scale National Institute of Mental Health study which featured prospective use of fully dosed ADTs for 12 weeks of treatment. In STAR*D, among patients who remained symptomatic after 2 full trials, 25% achieved remission after a third complete trial; among those who were non- or partially responsive after 3 trials, 12% achieved remission on the fourth trial.[11] These numbers suggest that as clinicians we must continue to treat MDD aggressively, use good diagnostic practices, use moderate-to-full doses of ADTs, and wait the full duration for effect to reach better patient outcomes.
Finally, as TRD occurs and patients move up in staging toward refractory illness, researchers have suggested that these patients will experience more chronicity and recurrence in their MDD, a higher suicide rate, and overall increases in medical and psychiatric healthcare use.[12-14] The MDD illness therefore becomes more risky to the patient and harder to treat to full, sustained remission.
Is TRD a bona fide clinical entity, a distinct type of MDD, or just a descriptive term? With the advent of the staging system above, and certainly in clinical conversation, treatment resistance clearly exists. With the FDA approval of vagus nerve stimulation (VNS), the term TRD began to be used in association with this FDA approval. Insurance companies and some clinicians, however, suggested that TRD is a term invented for marketing purposes and that the term TRD had no clinical relevance and that devices used for TRD should be deemed experimental.[15,16] The atypical antipsychotic aripiprazole was next approved as augmentation for MDD that fails to respond to an initial ADT, suggesting that a population of patients with TRD exists. Following that, the prescription medical food L-methylfolate was approved as a strategy to boost the effects of an initial ADT as well. Finally, this year, the first antidepressant clearly for use in TRD was approved -- the olanzapine-fluoxetine combination. At this point, we should agree that treatment resistance in MDD exists, is common, has supporting FDA language, and prospective studies already conducted in this special patient population are available for review.
Questions answered incorrectly will be highlighted.
What of the following is not a clear predictor in regard to the development of TRD?Being married vs single.All of the above statements are correct and, when present, increase the odds of the development of TRD. However, there are little data regarding marriage being preventative vs deleterious in TRD. Clinicians must be thorough and astute in all of these areas if they would like to achieve better outcomes in TRD.
Our patient has returned and we are rightfully concerned that she has not achieved remission after 3 different full clinical ADT trials (fluoxetine, escitalopram, bupropion XL). The patient states she is fully compliant with her medications, but she has continued in her TRD and is in stage 2 resistance. As treatment resistance increases, it often makes sense to re-evaluate the patient. The fact that she is presently not in remission is the greatest risk factor for TRD.[17] This forces us to increase our effort and be more diligent in providing aggressive care.
She is re-screened for medical conditions because these may increase TRD.[18] Basic laboratory values such as thyroid function, liver function, electrolytes, vitamin B12 and folate levels, and cell counts are all normal. A release was obtained from her primary care physician, who furnishes a report suggesting that the patient has diffuse muscle pain and tender points and likely meets fibromyalgia criteria in addition to her migraines. Under-treated comorbid medical problems may increase TRD,[19] so these issues will need to be better addressed.
In case our previous descriptive interviewing failed to detect additional psychiatric comorbidity, we next issued the patient an Axis I screening instrument, the Psychiatric Diagnostic Screening Questionnaire; and an Axis II screening instrument, the Personality Diagnostic Questionnaire. These are scored and show that the patient likely has generalized anxiety disorder (GAD) and symptoms of borderline personality disorder that were not readily apparent and were overshadowed by her marked depression on initial evaluation. These comorbidities also lend to TRD.[19] There is no evidence of bipolar depression, and no history of mania, psychosis, or substance use disorder.
Questions answered incorrectly will be highlighted.
What next pharmacologic intervention would be most in line with FDA approvals and indications for this patient in particular?
Remove the current 2 agents, switch to duloxetine because it is approved for MDD, GAD, and fibromyalgia.All of the options above are reasonable for a patient with stage 2 TRD. Clinicians must weigh safety and tolerability, the current evidence base, FDA approval status, drug-drug interactions, cost, and patient adherence when choosing the next strategy. When possible, choose FDA-approved agents that may cover all or most of the patient's disease states. In the absence of this, choose agents that cover certain disease states, but may be off-label for others. Use the evidence base to aid your choice and provide medico-legal protection. Be aware that some agents with greater evidence base and FDA approvals; ie, aripiprazole, lithium, olanzapine-fluoxetine, may have a greater side effect burden (ie, tardive dyskinesia, extrapyramidal symptoms, metabolic effects, renal disease) than some off-label agents with less robust evidence. This clinical decision-making should be interactive and documented in the medical record.
After full informed consent, the patient elects to switch off her medications to duloxetine because it has FDA approvals for her 3 main illnesses. It is titrated to the maximum dose of 120 mg/day over 3 months without added clinical effect (stage 3). The patient is given the above options again, with additional choices of adding mirtazapine or a stimulant, or changing to a tricyclic antidepressant or a monoamine oxidase inhibitor. She chooses mirtazapine because it has the ability to manipulate serotonin and norepinephrine through 4 distinct pharmacodynamic mechanisms[20] and was used effectively in STAR*D.[11]
The patient returns after several weeks on duloxetine 120 mg/day and mirtazapine 30 mg/day. She states that her fatigue is back, but her crying spells have completely stopped and she clearly notices a lack of GAD. Her fibromyalgia pain is dampened by 50%. However, she states that she still has poor concentration, guilty ruminations, anhedonia, and is not functioning well in the work setting. She is afraid she could lose her job. She appreciates these improvements but feels she is only mildly better.
Questions answered incorrectly will be highlighted.
The patient is currently taking duloxetine and mirtazapine, and previously failed on bupropion, escitalopram, and fluoxetine -- what intervention would you choose next?Add buspirone to the above 2 medications because it is FDA approved for her GAD and has some supporting evidence that it may augment antidepressant effects in an off-label manner.All of the options in the polling question above are reasonable for a patient in stage 4 TRD. Clinicians must weigh safety and tolerability, the current evidence base, FDA approval status, drug-drug interactions, cost, and patient adherence when choosing an appropriate intervention.
After discussion and informed consent with the patient, she refuses monoamine oxidase inhibitor treatments due to her perceived risk for stroke, and fear of losing the partial benefit from her current medications. She does not want to add any other medications that induce weight gain because she has already gained 15 pounds. As the threat of losing her job may be imminent, she elects to take a medical leave for 4 weeks and receive outpatient ECT treatment.
ECT has clear data and treatment effectiveness for TRD.[21-23] It is particularly fast, with many patients responding within 2-3 weeks. VNS would be a fitting option for her over the long term because 1 of 6 patients may experience a remission based on open label data and have a 2 to 1 chance of responding over non-VNS treatment as usual.[24] It is FDA approved for stage 4 TRD. Thus, while VNS is a good option, it may take months to become effective and our patient could lose her job waiting for improvement. The other device recently FDA approved is transcranial magnetic stimulation.[25] This device is approved for stage 1 TRD and may be applied when a first ADT has failed, with an end-goal of remission. However, some data have shown good effects when used in more resistant patients.[26]
Unlike this patient, many patients are afraid of ECT because of its effects and stigma. Other reasonable strategies might include adding aripiprazole within its FDA-approval as an augmentation strategy for patients failing to respond to SSRI/SNRIs. Thinking in rational polypharmacy,[27,28] with the goal of using a drug's known pharmacodynamic profile, we could assume that this patient's current regimen is elevating serotonin and norepinephrine activity by its dual reuptake blockade from duloxetine, and elevating the same 2 transmitters by way of mirtazapine's alpha-2 receptor blockade and serotonin-2c receptor blockade. Thinking through this serotonin-2c blockade further there is likely some norepinephrine and dopamine released into the dorsolateral prefrontal cortex downstream as well. This makes 6 antidepressant mechanisms from the initial 2 agents. Aripiprazole, in theory, would contribute serotonin-2a blockade and serotonin-1a stimulation, both of which have known antidepressant and anti-anxiety effects. This would allow for stimulation of dopamine in brain areas low on this transmitter (antidepressant effect) and blockade of high dopamine areas (anti-agitation effect), allowing for 4 more antidepressant moieties to converge to better treat this patient. This option carries an FDA-approval for MDD augmentation and utilizes theoretical application of pharmacodynamic properties to support this complicated polypharmacy regimen.[20] No other atypical has partial dopamine agonism, but others may share similar 5-hydroxytriptamine properties that may/or may not help treat depression. The combination of olanzapine and fluoxetine also is approved for TRD; other atypical antipsychotics also have been investigated. Whenever considering adding medication in this class, it is important to carefully consider the potential side effects, such as weight gain, metabolic changes, or extrapyramidal effects; and how these may affect patient health, coexisting physical conditions, and the patient's ability to adhere to treatment.
Lithium likely has the most controlled trials of any augmentation strategy,[4,29] but carries the stigma of lithium use, end-organ damage issues (kidney, thyroid), and requires fairly routine blood monitoring because it has a narrow therapeutic window. Pharmacodynamically speaking we do not know how it works as an antidepressant. It may boost monoamine synthesis, stabilize neuronal membranes, or promote neurotrophic factors.[30] This actually makes lithium unique in that all of our antidepressants primarily manipulate the monoamines (serotonin, norepinephrine, dopamine), but lithium does not. Its truly unique mechanism makes it a nonredundant, nonoverlapping strategy when adding it to other monoamine facilitating agents.
Our patient completed a course of 8 ECT treatments and, according to symptom reports and rating scales used by the ECT team, she is now in complete remission of her symptoms and she remains on her pre-ECT medication regimen.
Questions answered incorrectly will be highlighted.
What intervention has the most data regarding post-ECT maintenance of antidepressant response?Change the patient to nortriptyline, lithium, or both because they have strong evidence for maintaining the antidepressant response to ECT over the long term.All of the options above are reasonable for a patient in remission after 2-3 years of suffering from TRD; however, the use of nortriptyline and/or lithium may have the most stringent data where one might wish to enhance and maintain euthymia after acute ECT.
Maintaining Remission
Once a patient who has TRD has attained remission from his/her symptoms, it is imperative to maintain remission for as long as possible. Data suggest that even in this best case example, our patient has an approximately 70% chance of experiencing a relapse over 525 weeks of treatment per Judd and colleagues.[17] It is now the clinician's job to keep the patient on her medications and recommend maintenance ECT if that is required to maintain remission. Unfortunately, maintenance ECT is not robustly effective and may perpetuate side effects such as short-term memory problems.[31] This would, in all likelihood, preclude our patient from working. It is true that lithium and nortriptyline have reasonable post-ECT maintenance data,[32] but the side effect burden issues might preclude these changes as well. Rating scales are often underutilized in psychiatric care, but mounting evidence suggests that clinicians who routinely use scales detect mental illness to a greater extent, detect the absence of remission vs remaining residual symptoms more often, treat patients more aggressively as such, and obtain better outcomes eventually.[33]
Rating scales may be used when a patient is in remission to detect early relapse as well, thus allowing the clinician to change treatments to thwart a complete recurrence of MDD. Finally, psychotherapy, even in asymptomatic remission, may be helpful for patients to gain insight into their maladaptive, stress-causing patterns. This may decrease life stressors and triggers for future MDD recurrences as well.
