Is switching part of the TRD concept?

How is the question of switching between antidepressants crucial in defining the concept of TRD? For decades, since the 1970s, the scientific literature on the subject has abounded with proposals to define TRD. In the early days, extraordinarily complex definitions were proposed, all very elaborated and clever but impractical and often impossible to apply in the clinic. This first wave of definitions of TRD was also of little use for possible clinical trials aimed at evaluating the efficacy of a treatment or therapeutic strategy in TRD. It was a time, moreover, when there were no studies of this type and no recognition of a treatment in this indication. The basic principle of this first generation of definitions was to consider a depression as resistant to treatment only if the patient had been treated by multiple drugs or biological therapies, with the imperative to have received what was considered at the time as the most powerful treatments. At that time tricyclic antidepressants (TCA), monoamine oxidase inhibitors (MAOI), lithium, and electro convulsive therapy (ECT) were among the treatments to be incorporated in any TRD definition but all of them in different sequences and with various durations based on empirical assumptions. The main thing was therefore to consider TRD only if the depressive episode had benefited from these treatments.

Switching in staging models of TRD

Subsequently, more pragmatic proposals appeared, it was the era of staging models to better characterize TRD. These models are clinically extremely useful and have significantly contributed to a better understanding of the different stages of resistance to treatment in depression ( ). They give more space to the notion of “switching” regardless of the bias on relative effectiveness among the options used. Although particularly useful in the clinic, these models are however little or not used in clinical trials. They have therefore never contributed to demonstrating the efficacy of a treatment or a therapeutic strategy in TRD. Moreover, they are not considered by the regulatory authorities in charge of recognizing a specific indication in TRD, neither in Europe nor in the United States.

Switching antidepressant and clinical trials in TRD

The concept of switching antidepressant finally gained in importance when it came to a definition of TRD that could be used in clinical trials to demonstrate efficacy in TRD as a specific recognized indication. Lately, in response to the need to validate treatment strategies or specific medication in TRD, regulatory authorities both in Europe and in the United States elaborated their own recommendations to be used in clinical trials. Things were then significantly simplified, and for practical reasons TRD was defined as nonresponse to two consecutive trials of two different antidepressants at an adequate dose and duration. In this vision of TRD, what matters most is the switch in treatment. When, how and with which other antidepressant to make this switch?

At first, it was required to change antidepressant class. That is, using an antidepressant with a different mechanism of action than the one that did not work on the first attempt. This requirement was mainly based on clinical practice. It is more common to switch antidepressant classes in the hope of having a better chance of success. This practice, although theoretically meaningful, has never really been demonstrated through robust clinical data. The European Medical Agency (EMA) Guideline on clinical investigation of medicinal products in the treatment of depression consider monotherapy in patients with TRD as a separate claim and proposes clinical trial design and definition of TRD. TRD is considered, when treatment with at least two different antidepressant agents prescribed in adequate dosages for adequate duration and with adequate affirmation of treatment adherence showed lack of clinically meaningful improvement ( ). This pragmatic definition differs from the complex available staging models but is mainly intended to be used within clinical trials as a reference to characterize patients included based on the number and type of previous treatments. This definition differs from previous versions where two products of different pharmacological classes were requested. This important revision was in line with the data showing no advantage in favor of switching to a different class of antidepressant ( ; ; ; ).

The efficacy of Esketamine nasal spray, the newly approved medication in TRD was investigated using this definition of TRD ( ). As the study definition of TRD did not require the second antidepressant to be from a different class, a patient who had received two consecutive SSRIs or SNRIs was considered treatment resistant.

Diagnostic and clinical reassessment

The issue of diagnosis is fundamental and could be the basis of the vast majority of difficult-to-treat clinical situations. At this stage of our knowledge in the diagnosis of depression, it is likely that conventional antidepressant treatments are not sufficient to target the full heterogeneity of clinical manifestations of major depression. Most antidepressants have limited action on certain symptomatic dimensions of depression without affecting the overall clinical manifestations. Furthermore, it is possible that our diagnostic reading grid, based mostly on subjective assessments, does not identify the symptomatic dimensions that would be the real target of antidepressants. Every clinician must remain humble and consider the limits of antidepressant treatments in a double constraint. The part of resistance to treatment related to treatment choice and the part of “resistance” associated with the inability to target certain dimensions of psychopathology. This clinical posture allows a more comprehensive analysis of therapeutic choices that will not focus on treatment alone with an antidepressant molecule. It is undeniable that the combination of an antidepressant with, for example, psychotherapy or an anxiolytic drug treatment will be more likely to be effective. Acting on all dimensions of symptomatology, but also on personal and psychosocial elements is essential before evoking a real resistance to treatment.

Some practical and clinical considerations should be checked before switching antidepressant. These prerequisites are essential to minimize the risk of aggravation in case of inadequate discontinuation of the first antidepressant. It is also important to consider clinical dimensions that will increase the chances of success of the second antidepressant. In case of insufficient response to a treatment, it is advisable to seek for the possible causes of ineffectiveness, such as a possible noncompliance, the occurrence of undesirable side effects which reduce the chance of response, drug interactions or a possible psychiatric or somatic comorbidity. Because of the wide variety of forms of depression, care should be taken to ensure that the patients have received the antidepressant that best matches their clinical condition. After treatment failure, it is necessary to reevaluate in a detailed and rigorous manner the clinical parameters which could better guide the treatment. Clinicians should carefully consider factors associated with or predictive of TRD. Some clinical variables have been associated with poorer treatment outcomes and resistance to treatment. Identifying these variables must be a prerequisite for any therapeutic choice. A therapeutic choice based on the meticulous analysis of the clinical dimensions associated with resistance depression avoids long wanderings before realizing the inadequacy of a treatment. Personality disorders, melancholic features, severity, anxious depression, and/or comorbidity with anxiety disorders are among the most frequent clinical factors associated with TRD. Evaluating TRD must include clarification of the diagnosis and potential risk factors for TRD in addition to evaluation of treatment adherence, presence of comorbidities, medication dose and duration ( ).

Optimize dose and duration before switching

In the absence of a demonstrated therapeutic response, it is essential to optimize the ongoing treatment before considering switching. A simple option is to increase the dose of the antidepressant to reach an effective or the maximal tolerated therapeutic dose. Patients should be informed of potential side effects and should be informed of potential change of antidepressant. The clinical effect of an antidepressant usually appears within the first 2 weeks of treatment, but in many cases significant response may appear later, and it is recommended to wait 4–6 weeks before judging the effectiveness of a treatment ( ).

Initial improvement in the first weeks of antidepressant treatment may be early predictor of remission. It is difficult to find consensus in the literature about the optimal timing of a switch. Although the adequate duration of the acute phase of treatment is 4–8 weeks, some evidence suggests that one should not wait that long to switch to another treatment if the patient’s clinical condition has not significantly improved. Many clinicians wait for 6–12 weeks before considering a switch. This attitude of inertia despite the lack of response can have serious consequences on treatment compliance and leaves the patient with a significant degree of symptoms over a long period of time, thus increasing the risk of complications (e.g., suicidal behavior) and chronicity. A literature search reviewing the evidence (randomized controlled trials (RCTs) that examine early switching strategies) for the predictive value of early improvement at 2–4 weeks to predict final antidepressant response concluded that lack of early improvement (e.g., < 20% reduction in a depression scale score) at 2–4 weeks can be an accurate predictor to identify eventual nonresponders ( ). Only about one in five patients with lack of improvement at 4 weeks will have a response by 8 weeks. Based on these data, the authors recommend earlier assessment for improvement. If there is no indication of early improvement at 2–4 weeks after starting an antidepressant, a change in management can be considered. A metaanalysis of antidepressant naturalistic studies (nine studies; 6185 patients) investigating whether a partial decrease in depressive symptoms by week 4 was associated with response and remission by weeks 6–14 showed that such an early improvement is not a reliable outcome predictor in TRD ( ).

Optimal switching conditions and switching rules

If the decision is made to change the treatment, the switch of antidepressant must be carried out under optimal conditions and there are some important rules to follow in how to make this change. Switching from one antidepressant to another must be a carefully considered decision and the patient must be informed on potential consequences. The most frequent problems are related to possible withdrawal symptoms due to the discontinuation of the first molecule and undesirable effects in connection with possible drug interactions with the new antidepressant. It is common to wrongly attribute to the introduction of a new treatment side effects which are in fact linked to the discontinuation of the previous treatment. The switching phase is also more at risk to a resurgence of anxiety, depressive symptoms, or relapse.

There are rules for safely switching from one molecule to another, these depend on the pharmacokinetic and pharmacodynamic characteristics of the first and second molecule. The half-life of molecules, their metabolism by cytochromes, their specific side effects are all criteria to be considered. The characteristics related to the patient should also be noted: history of adverse effects or discontinuation syndrome with a given molecule, age, pregnancy, renal or hepatic failure, severity of depression, combined effects of two treatments.

Switching methods

Several strategies are available for switching between antidepressants ( ; ; ). The conservative or progressive switch allows for a gradual reduction of the first antidepressant. The new antidepressant is started after a drug-free period of two to five half-lives of the first molecule. Variants of this approach can be applied by reducing the length or missing the wash-out period. The new antidepressant is started according to its dose recommendation or gradually starting at a low dose. In the direct switch method, the first antidepressant is simply stopped and the second one is directly initiated at the usual therapeutic dose. The likelihood of discontinuation symptoms and drug interactions is higher and will mainly depend on the half-life of the first antidepressant. The cross-taper switch technique is frequently used and consists in gradually reducing the dose of the first antidepressant while introducing the second one at low dose. In this method, the patient is taking two antidepressants simultaneously, significantly increasing the risk of adverse effects. In a switch scenario where two antidepressants with the same or partially similar mechanism of action, a direct switch may be considered with minimal risks. For example, switching from an SSRI to another, switching from an SSRI to an SNRI, or from and SNRI to another SNRI.

Each method will have its advantages and disadvantages depending on the drug context, the individual parameters of the patient and the severity of the symptoms. Beyond the proposed method, the experience of the physician will be essential as well as the quality of the follow-up during this sensitive period. It is often necessary to reassess the original plan and adapt the method based on the outcome. This requires in-depth knowledge of the antidepressant pharmacopeia including the mechanism of action, metabolism, pharmacokinetics, interactions, and side effect profile of each molecule.

Pharmacokinetics and switching strategies in TRD

Genetic factors may interfere with treatment response to antidepressants. Genetic variation in drug-metabolizing genes may influence drug response and may predispose patients to TRD. These are for the most part highly penetrant and monogenic. More than 90 genetic variants in the CYP2D6 isoenzyme have been identified ( ). CYP2D6 is one of the cytochromes P450 isoenzyme responsible for metabolizing most of the antidepressants ( ). CYP1A2, CYP2B6, CYP2C9, CYP2C19, and CYP3A4 are also among cytochrome P450 isoenzymes involved in antidepressants metabolism. Variants within these genes define different group of individuals based on enzyme activity: poor metabolizers (PMs), intermediate metabolizers (IMs), extensive metabolizers (EMs), and ultrarapid metabolizers (UMs). Clinically, UMs are less likely to show positive response to treatment.

General consideration

Due to the high number of possible antidepressants change scenarios from an ineffective first choice, it is almost impossible to find controlled comparative trials in the literature for each possibility. Although switching is the most preferred strategy, little evidence exists to guide the clinician’s decision to switch at the level of class or mechanism of action but also at the level of which molecule within a class. Because of this reality, the effectiveness of different switching possibilities is considered comparable, whether from one class to a different class or from one molecule to another. The choice is essentially based on the clinician’s experience.

Contribution of the STAR*D study

A systematic review of studies evaluating the efficacy of changing antidepressant medication after failure with an SSRI showed that 50% of patients who failed to respond to one SSRI responded to a second ( ). This relatively optimistic result is in contradiction with other studies, notably the STAR*D (Sequenced Treatment Alternatives to Relieve Depression) study, a large study funded by the NIMH (National Institute of Mental Health) evaluating the treatment of depression, showing a significantly lower response rate (around 27%) in patients treated with sertraline after failure with citalopram ( ). The main objective of the STAR*D program is to determine the most effective treatment if the first-line antidepressant does not induce a significant response. Following ineffective treatment with a SSRI, there is little evidence on the effectiveness of a medication change strategy. In the STAR*D study, 2876 severely depressed patients undergoing psychiatric and general medical care were treated with citalopram as first-line therapy. At the end of the study, approximately 30% of patients were considered to be in remission after 6–7 weeks. In a second stage, 727 patients who did not respond to citalopram were treated with either bupropion 400 mg, sertraline 200 mg or venlafaxine 375 mg ( ). This change in treatment has enabled one in four patients to achieve complete remission. The interest of these three medications seemed quite equivalent.

In level 3 of STAR*D, switching to mirtazapine monotherapy ( n = 114) was compared to nortriptyline monotherapy ( n = 121) in a group of 235 patients who did not respond to citalopram and a switch to another antidepressant or an augmentation with another pharmacological option ( ). There was no significant difference in response and in remission rates. In another study, 150 patients who failed to respond to two antidepressants were switched to mirtazapine, venlafaxine, or paroxetine. There was no significant difference in terms of remission rates ( ).

Efficacy of switching: Other data

Clinically, there is a clear trend to replace the SSRI with an antidepressant with a different mechanism of action such as tricyclics, antidepressants with a mixed serotonergic and noradrenergic action (venlafaxine, duloxetine, mirtazapine), products with an effect on the dopaminergic and/or noradrenergic system (bupropion, reboxetine), and MAOIs. In a metaanalysis of five randomized clinical studies, switching from one SSRI to another class of antidepressant (venlafaxine, mirtazapine, and bupropion) was associated with a higher rate of remission than switching to another SSRI ( ). Response rates obtained by replacing an SSRI with a combination product ranged from 28% to 50%. Switching to bupropion resulted in a response in 26%–35% of cases. Depending on the studies, switching from an SSRI to a tricyclic gave response rates between 16.5% and 48.5%. In two open studies, nonresponders to SSRI were treated with Venlafaxine with response rates of 58% and 87% ( ; ). In another study, one third of 84 patients suffering from a highly resistant major depression were considered to be full or partial responders after 12 weeks of venlafaxine treatment ( ). In a double-blind study, response (52% vs 33%) and remission (42% vs 20%) rates were higher with venlafaxine than paroxetine in patients resistant to two antidepressant treatments ( ). The switch to mirtazapine did not appear to be an interesting option as it has been demonstrated in two studies ( ).

The use of MAOIs in cases of resistant depression has been widespread in clinical practice in relation to a perception of this class of antidepressant as particularly effective in difficult cases. Some studies have shown a high response rate with phenelzine or tranylcypromine in patients with resistance to tricyclics ( ; ). However, the risks associated with their prescription (hypertensive crises), their scarcity and the need for a period of 1–2 weeks of SSRI withdrawal before their introduction explain the lack of interest in this type of strategy.