Chronic and Treatment Resistant Depression (07.11.2014)

Depression is a common condition that often takes a chronic course and fails to respond to treatment (Box 1). The 12-month prevalence of depression in Europe is 6.9%, according to the largest and best study of the subject to date (1). Depression is thus one of the most common diseases of any kind. It has not, however, become more common in recent years (1). Although it classically takes an episodic course, depression becomes chronic in 15–25% of sufferers (2). 50% of the depressed patients in a longitudinal study were well again six months later (e1), but 7% were still depressed ten years later (e2). Resistance to treatment is also common: in the approval and efficacy trials of various antidepressants, one-third to one-half of the patients did not respond to several weeks of treatment (2–4). The largest-ever study of the treatment of depression was the STAR*D trial (Sequenced Treatment Alternatives to Relieve Depression), carried out in the USA and involving over 4000 participants: only 37% were in remission after an initial 12-week treatment trial with citalopram, and 67% after four treatment trials (5).

Learning aims

In this article, we answer the following questions:

• What distinguishes true treatment resistance from pseudoresistance?
• What treatment options are available—drugs, psychotherapy, other—that can prevent or overcome treatment resistance and chronification?
• What scientific evidence underlies these treatment strategies, and how are they applied in practice?

Methods

This article is based on a selective literature search making use of the authors’ personal experience, as well as on the most up-to-date German-language guideline for the treatment of depression: the S3 Guideline (German National Disease Management Guideline) Unipolar Depression (6). The S3 Guideline is, in turn, based on a systematic analysis of the literature.

Diagnostic evaluation

Often, when depression does not respond to treatment, the resistance to treatment is only apparent, because either the diagnostic evaluation or the trial of treatment was not properly conducted. This is called “pseudoresistance”; possible causes are listed in Box 2.

As there are many somatic and psychiatric diseases and comorbidities in the differential diagnosis of depression (9), only a systematic diagnostic approach can ensure that the diagnosis is correct. To be both scientifically well-founded and cost-efficient, the diagnostic evaluation should be performed in stepwise fashion, i.e., according to an algorithm (10). A basic evaluation should be done before any treatment is started; thereafter, if the patient does not respond to two properly conducted trials of treatment, the evaluation should be supplemented by a further diagnostic step to rule out pseudoresistance (Figure 1). Many of the treatment options require further diagnostic testing before they can be initiated, for reasons of safety (e.g., an ECG before treatment with selective serotonin reuptake inhibitors [SSRIs] or tricyclic antidepressants [TCAs]).

Many drugs can cause depression (eBox 2). Pharmacogenic depression should be suspected if there is a close correlation between the beginning of drug treatment and the onset of depressive manifestations; such suspicions should be confirmed, whenever possible, by a trial of drug cessation or substitution.

Pharmacotherapy

First step

Key elements in the successful treatment of depression are the following:

• adherence to a clear treatment strategy,
• treatment for an adequate period of time, and
• systematic evaluation of efficacy at predetermined time points.

Antidepressant drugs are the mainstay of pharmacotherapy for depression. About 30 drugs in this class are now approved for use in Germany. They all have a number of features in common: The rate of non-response is similar for all antidepressants. No antidepressant works immediately; as a clinical rule of thumb, one should wait three to four weeks from the beginning of treatment before judging whether a drug is effective, or up to six weeks in older patients (6). Nearly all antidepressants are thought to exert their clinical effect by increasing the intrasynaptic concentration of serotonin and/or norepinephrine in the central nervous system (CNS), through a variety of mechanisms (reuptake inhibition in the case of SSRIs, serotonin-norepinephrine reuptake inhibitors [SNRIs], and TCAs; inhibition of monoamine oxidase [MAO]; presynaptic α₂-receptor antagonism by mirtazapine and mianserin). Antidepressants vary considerably, however, in their side-effect profiles.

In the initial phase of treatment, an antidepressant should be selected on the basis of the patient’s history, the side-effect profile, potential interactions with other drugs the patient is taking, and the physician’s experience. It should be given in a standard dose. Suitable review articles should be consulted for information on standard dosages, side effects, and interactions (for example, References [6] and [11] are freely available).

For some antidepressants, notably TCAs and venlafaxine, the standard dose should not be given at the very beginning of treatment. A dose-escalation phase is needed to ensure tolerability and compliance, but it should be kept as short as possible. The latency period for the effect commences only when the standard dose has been reached. As soon as treatment with an antidepressant is begun, a particular date should be chosen by the physician and patient as the “decision day” on which they will jointly assess the efficacy of the drug (Figure 2). To make this assessment as reliable as possible, the patient’s depressive manifestations should be meticulously documented before the treatment is begun (psychopathological evaluation; standardized, observer based depression scales, e.g., the Hamilton Rating Scale for Depression [e4]; and self-assessment scales, e.g., the Beck Depression Inventory [e5]).

If the patient responds positively to the drug, the treatment is maintained as continuation therapy for six to twelve months after the remission of depressive manifestations. Clinical trials have clearly shown a very high risk of recurrence in this period; it is, therefore, recommended that antidepressant medication should be maintained for its entire duration, in an unchanged dose if possible (e6). An even longer, temporally open-ended administration of antidepressants is indicated for recurrence prophylaxis in patients who display a recurrent course of disease (as a rule of thumb, two or more episodes in five years, or three or more episodes overall).

If, on decision day, the response is judged to have been inadequate, drug treatment should be changed to a second-step strategy, as will be discussed in the next section.

Second step

Therapeutic Drug Monitoring (TDM)—For most antidepressants, adequate information is now available about the optimal range of the serum drug level for clinical efficacy. In case of non-response to any of these drugs, a serum level measurement is advisable, followed by upward or downward dose adjustment if indicated. A comprehensive overview of the recommended serum levels can be found in the consensus guidelines for therapeutic drug monitoring in psychiatry that have been issued by the German Working Group for Neuro-psychopharmacology and Pharmacopsychiatry (Arbeitsgemeinschaft für Neuropsychopharmakologie und Pharmakopsychiatrie, AGNP) (12).

On the other hand, the determination of serum levels is not recommended for the following antidepressants, either because inadequate information is available or for pharmacological reasons: agomelatine, bupropion, moclobemide, paroxetine, and tranylcypromine.

Serum levels should be measured at their trough, i.e., blood should be drawn just before the next dose is taken. If the serum level is not in the therapeutic range despite the administration of a standard dose, the possible reasons include non-compliance, (rarely) inadequate resorption, and genetic variability of the cytochrome P₄₅₀ system, by which most antidepressants are metabolized. Moreover, concomitantly administered drugs frequently cause either inhibition (e.g., by fluoxetine, fluvoxamine, and paroxetine) or stimulation (e.g., by carbamazepine); for further information, see the website www.drug-interactions.com.

Switching to another antidepressant—A switch to another antidepressant is perhaps the strategy that is most commonly used at present, despite a lack of scientific validation. The fact that all approved antidepressants were effective in placebo-controlled trials does not, in itself, justify switching to another one in case of non-response. A benefit from switching cannot be expected a priori, for at least two reasons: the approval trials did not specifically study patients who had failed to respond to a previously administered antidepressant, and the available antidepressants, numerous as they are, still tend to have very similar effects.

Any symptomatic improvement after switching to another antidepressant might simply be due to the longer duration of treatment. Thus, in a properly designed trial, non-responders to initial monotherapy should be randomized either to further treatment with the same antidepressant or to a switch to another one. This should be done in double-blind fashion.

A systematic, computer-assisted search yielded only three trials that were designed in this way. None of them showed the switching strategy to be superior to further treatment with the same antidepressant; neither did a meta-analysis taking all three trials into account (13). One of two further, recently published trials on this question showed no difference between the switching and continuation strategies (e7); in the other trial, the patients who were switched to a different antidepressant actually did significantly worse (e8).

There is thus a notable discrepancy between the frequency with which the switching strategy is applied and the meager scientific evidence for it. Further trials are needed; at least until more evidence is available, antidepressants should not be prescribed serially.

Lithium augmentation—This strategy was specifically developed for the treatment of refractory depression. Lithium can be added if the patient has not responded to at least one trial of antidepressant monotherapy of adequate duration (14–16). The purpose of this is to convert a non-response into a response. Ten randomized, double-blinded trials have shown that lithium augmentation is more than twice as likely to bring about a response than placebo augmentation (odds ratio 3.1, 95% confidence interval [CI] 1.8–5.4, number needed to treat [NNT] 5) (17). There is some evidence indicating that the addition of lithium to an antidepressant that has no serotoninergic component is unlikely to help (e9, e10). Lithium augmentation therapy should be given in a rapidly escalating dose, with due attention to the necessary preliminary tests and contraindications (6, 18); serum lithium levels should be checked regularly, with a therapeutic target range of 0.6–0.9 mmol/L. Once the serum level is in this range, the effect should be observed for two weeks; if there is no response, lithium treatment should be terminated (as long as there is no other indication for it). In the event of a response, the combination of lithium and the antidepressant should be given for a further six to twelve months of continuation therapy (19, 20).

The following are arguments in favor of lithium augmentation: that the antidepressant taken till now has been well tolerated and need not be discontinued, that prophylactic lithium treatment is indicated anyway because of a recurrent course, or that the patient is recurrently suicidal (21). The independent benefit of lithium in reducing suicidality has been consistently shown, not only in multiple analyses of published efficacy trials (22), but also in a prospective, controlled trial with the reduction of suicidality as its primary end point (e11).

Augmentation with a second-generation antipsychotic—Recent controlled trials have shown that augmentation of an antidepressant with a second-generation antipsychotic (alternative term: atypical neuroleptic) can convert a non-response to a response. A meta-analysis of such trials yielded an overall odds ratio of 1.69 (95% CI 1.46–1.95, NNT 9) compared to placebo (23). Quetiapine is the only drug approved for this indication in Germany. Weight gain, with the risk of metabolic syndrome, is a common side effect. In a randomized trial comparing lithium augmentation to quetiapine augmentation, sponsored by the manufacturer of quetiapine, the two drugs were found to be of comparable benefit (24).

High-dose antidepressant medication—It was concluded in a systematic review (25) that the principle “more is better” holds for some, but not all, antidepressants. Clinical trials have consistently shown high-dose treatment with SSRIs to be no more effective than treatment at the standard dose; there is thus no indication for dose escalation in the event of non-response. A SPECT study showed that treatment with paroxetine, an SSRI, at the standard dose of 20 mg/d already causes blockage of roughly 80% of serotonin transporters in the CNS, without any further increase in patients treated with a high dose of the drug (mean, 47 mg/d) (e12).

On the other hand, for TCA, tranylcypromine, and venlafaxine, doses in the high therapeutic range have been found to be more effective than doses in the low therapeutic range. Dose escalation would be a reasonable next step in case of non-response to one of these drugs at the standard dose (25), as long as the standard dose has been well tolerated and the patient has a positive attitude to the drug in question.

Combinations of two antidepressants—Randomized clinical trials demonstrating the greater efficacy of a combination of two antidepressants compared to a single antidepressant combined with placebo are available only for combinations that involve a reuptake inhibitor (i.e., an SSRI, SNRI, or TCA) paired with a presynaptic autoreceptor blocker (mianserin, mirtazapine, or trazodone) (26). Such combinations have been more extensively studied than others because drugs in these two classes given together can be expected, on theoretical grounds, to exert a synergistic effect. With only a single exception, nearly all trials have shown such combinations to be effective. They clearly can be recommended as a treatment strategy in case of non-response to monotherapy (6). Clinicians welcome such combinations, because mirtazapine, mianserin, and trazodone have a sedating effect (unlike the SSRIs) and can therefore be usefully given in the evening.

It is unclear from the available evidence whether each of the two drugs in such combinations should be given in the standard dose used for monotherapy, or whether lower doses might suffice. It is presumed, though not yet documented, that standard doses are more effective. It is important to note that combinations of MAO inhibitors with serotoninergic antidepressants are contraindicated and that fluoxetine, fluvoxamine, and paroxetine inhibit the cytochrome P₄₅₀ system.

Third step

Irreversible MAO inhibitors—Refractory depression is now the main indication for irreversible MAO inhibitors (in Germany: tranylcypromine), which are considered to be more effective against refractory depression than the reversible MAO inhibitor moclobemide (27). Although this conclusion was drawn mainly from older studies that do not meet current methodological standards, most experts still consider it correct. As a result, treatment with an irreversible MAO inhibitor occupies a step all by itself in current algorithms for the treatment of depression.

Irreversible MAO inhibitors are no longer drugs of first choice, however, because patients who take them must adhere to a low-tyramine diet to obviate the risk of a hypertensive crisis (28). Experience has shown that patients generally adhere to a low-tyramine diet without difficulty, and the diet has even been said to have an independently beneficial placebo effect. Because of the dose-dependence of the response to tranylcypromine in patients with refractory depression, the target dose is usually set in the high range, at roughly 60 mg/d (the dose should be escalated slowly, with the drug being given in the morning and at midday) (29). Irreversible MAO inhibitors must not be given in combination with serotoninergic drugs such as SSRIs, SNRIs, or clomipramine, or too soon after these drugs have been discontinued.

Electroconvulsive therapy (ECT)—ECT is the most effective treatment for refractory depression, with a response rate of 50% to 85% in general and 50% to 75% among non-responders to antidepressant medication (30). Contemporary ECT has minimal risks and side effects. General anesthesia and pharmacological muscle relaxation, administered by an anesthesiologist, prevent tonic-clonic muscle activity and the injuries that might result; ventilation with oxygen prevents cerebral hypoxia; and cerebral electrical stimulation with intermittent square-wave pulses reduces the quantity of energy required for provoking a grand mal seizure (31).

Cognitive impairment, mainly involving short-term memory, is seen relatively often after ECT, but it is transient and no longer demonstrable four days after the end of treatment (e13). The main clinical drawback is the high rate of early recurrence after an initially successful course of ECT—up to 80% within six months. This figure applies if no drug treatment is given after ECT; it can be markedly lowered by antidepressant medication. The best clinical trial on this subject showed that a combination of nortriptyline and lithium is significantly better for this purpose than nortriptyline alone, which is, in turn, better than placebo (e14).

Treatment algorithm for depression

With so many treatment options available, the question arises in what order they should be used. A scientifically grounded treatment algorithm would be desirable, as the indiscriminate serial administration of multiple drugs in inadequate doses is now considered an important cause of treatment resistance (or pseudoresistance) (32). Better results ought to be obtainable with a treatment algorithm in which the improvement of depressive manifestations is assessed in standardized fashion at predetermined intervals and the findings of the assessment then determine whether the treatment should be changed (i.e., whether the algorithm should proceed to the next step). Such algorithms are already in existence for pharmacotherapy and ECT but, in principle, would be useful for psychotherapy as well. The algorithm of the German S3 Guideline / National Disease Management Guideline Unipolar Depression is shown in Figure 3 (6).

Large-scale randomized controlled trials have repeatedly shown that treatment according to an algorithm brings about a response more often and more rapidly than conventional treatment at the discretion of the treating physician (33, 34). In one trial, the patients in the algorithm group achieved remission in a mean of 7.0 weeks, compared to 12.3 weeks in the control group (33). They also took less medication and underwent less frequent changes in their treatment strategy. It can be concluded that the use of algorithms prevents unnecessary scattershot therapies and premature switching of treatment measures (33).

Psychotherapy

Controlled trials have shown that interpersonal psychotherapy (IPT), behavioral therapy, and cognitive therapy are effective treatments for non-refractory and non-chronic depression; moreover, large-scale observational studies and clinical trials in groups of patients with mixed diagnoses provide evidence of efficacy for psychoanalysis and psychodynamic short-term therapy. The potential efficacy of psychotherapeutic methods in the treatment of refractory depression has been very little studied, but some types of psychotherapy were specifically developed to treat chronified depression. Psychotherapy seems to have the advantage of longer persistence of its beneficial effect after the end of treatment, compared to drug treatment (35, e15). Psychotherapy improves compliance with drug treatment and thus leads to better efficacy when these two modes of treatment are used in combination.

Cognitive behavioral therapy (CBT)

The aim of cognitive behavioral therapy (CBT) is to make patients aware of their illness-promoting, depressive-negative modes of perception and automatic thoughts so that they can rid themselves of them. CBT is based on the assumption that dysfunctional cognitions disturb the patient’s emotions and behavior. Important treatment strategies for chronified depression include overcoming the patient’s social withdrawal, supposed helplessness, and lack of positive reinforcement. CBT is intended to correct these dysfunctional cognitions. A ten-phase CBT manual has been specifically devised for the treatment of chronic depression, in order to do justice to the special features of this patient group (36).

Interpersonal psychotherapy (IPT)

Interpersonal psychotherapy is a type of short-term therapy that was developed specifically for the treatment of acutely depressed patients, based on the idea that depression typically arises in a particular interpersonal or psychosocial context. IPT is therefore oriented toward current problems in the patient’s life, usually in one or more of four problem areas: role changes, interpersonal conflicts, social deficits, and grief (37).

Adaptations of IPT for chronic depression are now being developed and tested. It should be noted that depression, once it becomes chronic, is less closely connected to current life events than acute depression. Patients who have been depressed for years often have negative expectations of treatment, a poor social network, and chronified interpersonal conflicts; these are the areas in which IPT for chronic depression might usefully intervene.

Psychodynamic psychotherapy

In psychodynamic psychotherapy, the therapist–patient relationship gives the patient an opportunity to emotionally re-experience unfavorable relationship experiences of the past in a sheltered setting. At the beginning of treatment, refractory depression is addressed with what can be thought of as “environmental protection measures”: the empathic attitude that is so important to depressed patients, along with an emotional resonance that patients can appreciate, are best assured by a psychotherapeutic attitude of acceptance with constant attention to countertransference on the part of the therapist.

Person-centered therapy (PCT)

PCT, which was developed on the basis of humanistic psychology, proceeds from the assumption that anyone seeking help from a psychotherapist already has the capacity to heal him- or herself. The goal of psychotherapy is, therefore, to create favorable conditions that enable self-healing. The therapist must understand what the patient is feeling (empathy), and the therapist’s attitude toward the patient must be characterized by genuineness (congruence) and acceptance (unconditional positive regard).

Cognitive behavioral analysis system of psychotherapy (CBASP)

The only psychotherapeutic method developed specifically for the treatment of chronic depression is rather awkwardly known as CBASP (38). It is a manual-based method that incorporates behavioral, cognitive, and interpersonal strategies and takes account of the psychological trauma to which many chronically depressed patients were exposed early in life. CBASP is intended to help the patient become aware of the consequences of his or her own behavior, acquire empathy, learn problem-solving skills and coping strategies, and benefit from an interpersonal process for the healing of prior trauma. The main techniques for achieving these goals in CBASP are so-called situational analysis and special interpersonal strategies. Large-scale randomized trials have yielded both positive (39) and negative (40) evidence concerning the putative efficacy of CBASP.

Overview

Treatment resistant depression is not an untreatable disease. Considering the wide diversity of depressive manifestations, disease courses, and chronification factors, no single method or small battery of methods can be expected to overcome treatment resistance and chronification in every case. Rather, what is needed is the structured application of a variety of treatment options that have been validated by the available scientific evidence, with drug treatment and psychotherapy playing equally important roles.

Conflict of interest statement

Prof. Bschor has received reimbursement of meeting participation fees and of travel and accommodation expenses from Lundbeck and AstraZeneca. He has received honoraria for lectures at continuing medical education events from Lilly, BMS, esparma (Aristo), Servier, AstraZeneca, Sanofi, and Lundbeck.
Prof. Bschor is a regular member of the Drug Commission of the German Medical Association and also the spokesman for the psychiatric working group within it. He is a board member of IGSLI and President of the Berlin Society for Psychiatry and Neurology.

Prof. Bauer has served as a paid consultant for Alkermes, AstraZeneca, BristolMyersSquibb, Ferrer Internacional, Janssen, Lilly, Lundbeck, Otsuka, Servier, and Takeda. He has received honoraria from UNI MED Verlag, Elsevier, Kohlhammer, and Springer for authorship of publications relating to the topic of this article. He has received reimbursement of meeting participation fees from Servier. He has received payment for the preparation of continuing medical education events from AstraZeneca, BristolMyers Squibb, Ferrer Internacional, Lilly, Lundbeck, Otsuka, Pfizer, and Servier. He has also received payment for carrying out clinical trials on behalf of Lilly, Servier, and AstraZeneca.

Dr Adli has served as a paid consultant for Lundbeck, esparma (Aristo), and Merz and has received reimbursement of meeting participation fees and of travel and accommodation expenses from Lundbeck and Servier. He has received payment for the preparation of continuing medical education events from Lundbeck, Servier, and esparma (Aristo).

Manuscript submitted on 30 May 2014, revised version accepted on
13 August 2014.

Translated from the original German by Ethan Taub, M.D.

Corresponding author
Prof. Dr. med. Tom Bschor
Abteilung für Psychiatrie, Schlosspark-Klinik
Heubnerweg 2, D-14059 Berlin, Germany
bschor@schlosspark-klinik.de

@For eReferences please refer to:
www.aerzteblatt-international.de/ref4514

eSupplement, eBoxes:
www.aerzteblatt-international.de/14m0766

Box 1

Definitions: chronic/treatment resistant depression, dysthymia

Box 2

Potential causes of pseudoresistance to treatment

Figure 1

Steps in an algorithm for the diagnostic evaluation of depression

Figure 2

Time course of treatment with an antidepressant

Figure 3

Algorithm: drug treatment for refractory depression

eBox 1

Risk factors for a chronic course of depression

eBox 2

Drugs that can cause depression