Treatment-resistant depression: Newer alternatives

As many as 50% of depressed patients do not achieve a 50% or greater reduction in severity of symptoms after an adequate antidepressant trial.¹ Moreover, among those who do respond to acute treatment, longer-term residual depressive symptoms are quite common. Persistent subsyndromal depressive symptoms are associated with impaired psychosocial functioning and increased risk of relapse. This possibility has reinforced an evolving consensus that full depressive remission rather than response is the proper goal of treatment.²

When your patient’s depressive symptoms fail to remit following an initial pharmacotherapy course, you must decide which of numerous possible next steps to pursue, in what combinations, and following what sequence. The range of strategies include the following:

1. Pursuing an extended trial with the initial agent using higher than usual dosages (e.g., fluoxetine 40 to 80 mg/qd);
2. Augmenting an antidepressant with an agent that offers no established antidepressant efficacy on its own, such as buspirone, lithium, thyroid, or estrogen;
3. Combining an antidepressant with another antidepressant, with another somatic therapy such as electroconvulsive treatment or phototherapy, or with some form of psychotherapy;
4. Switching to an antidepressant within the same class (e.g., from one selective serotonin reuptake inhibitor [SSRI] to another), or outside of class (from an SSRI to an atypical antidepressant), or switching to non-pharmacological somatic therapy or psychotherapy.³

A predictable lag exists between innovative clinical applications and the randomized, controlled trials (RCTs) designed to evaluate them. The result: Approaches that are thoroughly described in the literature involve medication regimens that are no longer first-line, such as lithium or thyroid augmentation of tricyclic antidepressants (TCAs), or the combination of TCAs with SSRIs. Conversely, approaches most widely used in current psychiatric practice (e.g., addition of bupropion to SSRIs) have received relatively little systematic attention.⁴

Further, while predictors of initial antidepressant response have been hard to come by, even less is known about predictors of antidepressant response after lack of response to previous antidepressants. The working hypotheses that clinicians use to decide which strategies to pursue, though plausible, are largely untested, whether based on characteristics of a patient’s initial response (e.g., non-response or partial response)⁵ or on side effects or comorbid diagnoses.

This review will describe new research and emerging strategies that address the common clinical problem of unremitted depression despite one or more adequate courses of antidepressant treatment.

A look at new augmentations and combinations

Adding a second agent is an appealing strategy when patients have tolerated an initial antidepressant without troublesome side effects or have shown partial response (≥ 25% and <50% symptom reduction) and/or when a second agent may serve an important additional goal such as treating a comorbid condition—attention-deficit disorder or smoking, for example—or a side effect (e.g., nausea or insomnia).

Box 1

The primary drawbacks to augmentation/combination strategies are an increased risk of drug interactions, cost, and potential decrement in adherence to treatment as the regimen’s complexity increases.

Thyroid augmentation Although thyroid augmentation of TCAs has been the subject of numerous RCTs, no controlled trials of thyroid augmentation of SSRIs or other newer-generation antidepressants are available.

Lithium augmentation Studies of lithium augmentation of SSRIs have yielded generally modest response rates with questionable durability.^6,7

Noradrenergic and/or dopaminergic agonist agents Open trials and case series in the treatment of patients with unremitted depression have supported combining SSRIs and venlafaxine with agents that possess primarily noradrenergic and/or dopaminergic agonist properties. These include bupropion,⁸ psychostimulants,⁹ and direct dopamine agonists, including pergolide and pramipexole.¹⁰

These combinations have the advantage of potentially ameliorating several common SSRI side effects, particularly sedation, sexual dysfunction, and putative SSRI-related apathy. In the case of psychostimulants and/or bupropion, the combinations may treat such common comorbidities as attention-deficit/hyperactivity disorder and smoking.

Modafinil The combination of the antinarcoleptic modafinil with SSRIs and other newer-generation antidepressants has also attracted interest. Its efficacy as an antidepressant adjunct and as a remedy for side effects must still be established, however.¹¹

Buspirone augmentation The augmentation of SSRIs and other antidepressants with the antianxiety azaperone buspirone is supported by impressive response rates in several open trials. But the only placebo-controlled trial evaluating buspirone augmentation in resistant depression failed to find a significant drug:placebo difference.¹² Still, its tolerability and anxiolytic efficacy, and potential for ameliorating sexual dysfunction in some patients, support its judicious use as an augmenting agent pending further study.

Pindolol Studies of this beta-agonist, 5HT-1A antagonist, as an antidepressant-augmenting combination have yielded promising results in some studies, including a negative 10-day, placebo-controlled trial.¹³ Some patients experience jitteriness or irritability on pindolol. Its role in treating resistant depression remains to be established.

Noradrenergic TCAs The combination of SSRIs with these agents is a good example of a strategy based on pairing complimentary antidepressant mechanisms. Double-blind, controlled trials of fluoxetine plus desipramine^6,7 have tempered the enthusiasm generated by earlier open studies, however. Response rates in these trials range from 25% to 30%, no higher than lithium augmentation and slightly lower than higher-dose fluoxetine.

Mirtazapine A more recent combination is mirtazapine with SSRIs or with high-dose venlafaxine.¹⁴ In some patients, this combination may attenuate sexual dysfunction, insomnia, and gastrointestinal side effects of SSRIs and venlafaxine by virtue of a 5-HT2, 5-HT3, and histamine receptor blockade by mirtazapine. This combination may also capitalize on the combined antidepressant effects of direct norepinephrine (NE) and serotonin (5-HT) reuptake inhibition as well as an alpha₂ adrenergic auto- and hetero-receptor blockade—facilitating presynaptic NE and 5HT release—and 5-HT2 receptor antagonism.

Antipsychotics The use of antipsychotics for nonpsychotic unipolar depression has been controversial. However, in the context of lower apparent risks of tardive dyskinesia with the newer, atypical antipsychotics, their use as antidepressant augmenting agents has been revisited. In one double-blind, placebo-controlled trial, the combination of fluoxetine plus olanzapine was more effective in a well-defined refractory depressed population than was either medication alone.¹⁵ (See Box 1 for a discussion of fluoxetine and folic acid.)

Ostroff and Nelson¹⁶ reported an extremely rapid (1 week) response after the addition of risperidone among 8 depressed individuals who had not responded to fluoxetine or paroxetine alone. The activating properties of ziprasadone for some patients, combined with its lower propensity for producing weight gain than other atypical antipsychotics, make it another potentially attractive candidate for antidepressant augmentation, though one that requires further study.

Anticonvulsant augmentation While controlled studies of anticonvulsant augmentation of newer-generation antidepressants are lacking, the efficacy of lamotrigine for treating bipolar depression¹⁷ has encouraged clinicians to combine the agent with antidepressants in unipolar depression. The potential sedative/anxiolytic effects of other anticonvulsants, including gabapentin and valproate, have also supported their use in resistant depression complicated by anxiety or irritability. Omega-3-fatty acids may work as mood-stabilizing substances when used for antidepressant augmentation.

Antidepressant switches: In class or out?

Switching from TCAs to SSRIs or vice versa following nonresponse has yielded generally high (50% to 60%) response rates.³ But the more common question now is whether to switch from an initial SSRI to another SSRI, to an SSRI-like agent with an additional noradrenergic mechanism (such as venlafaxine), or to an atypical antidepressant such as bupropion or mirtazapine.

Traditional teaching and regular psychiatric practice⁵ have favored a switch outside of class. More recently, however, a number of uncontrolled studies looking at response rates to a second SSRI following inadequate response to and/or tolerance of an initial SSRI have shown response rates of 40% to 75%.¹⁸ This coincides with response rates seen in studies involving crossover to antidepressants with differing mechanisms. Inferences from these studies are limited by, first, determining nonresponse retrospectively and, second, lumping together patients who did not remain on the first SSRI because of nonresponse with others who curtailed treatment because of intolerance.

Because of its dual 5-HT/NE reuptake inhibitory activity, particularly at higher doses, venlafaxine is a popular choice following failure of an initial SSRI. One-third of subjects with extremely resistant depression (3 or more trials of antidepressants) responded to venlafaxine.¹⁹ DeMontigny et al²⁰ also demonstrated its efficacy (with a 58% response rate) as a switching agent among resistant, depressed patients.

But comparative response rates for switching to venlafaxine from an SSRI vs. from other antidepressant classes have not been well delineated, nor have response rates for switching from SSRIs and other antidepressants to venlafaxine vs. to other non-SSRI agents.

Switches from SSRIs to nefazodone and mirtazapine or the selective NE reuptake inhibitor reboxetine have shown reasonably high response rates.³ Studies comparing these various strategies head to head are lacking. In the absence of data, reasonable guides for switching agents include projected tolerability and existence of comorbid conditions that may respond well to one agent over another.

Another issue is whether to cross-taper, that is, to lower the dose of the first agent while titrating up the next, or to discontinue the first before the second. When catastrophic drug interactions would result from overlap, such as the switch from an SSRI to a monoamine oxidase (MAO) inhibitor, a suitable wash-out period is necessary. This can also occur if the first antidepressant is believed to cause intolerable side effects.

In most other instances where no absolute contraindications exist, cross-tapering may be worthwhile despite theoretical risks such as serotonin syndrome when overlapping an SSRI and venlafaxine. Cross-tapering helps reduce the risk of side effects associated with abrupt discontinuation including nausea, myalgias, headache, and dysphoria, and may minimize the loss of antidepressant benefit from the initial agent before the impact of the second agent is realized.

Great hope for new knowledge with STAR*D

In recognition of the substantial direct and indirect costs, the morbidity and mortality associated with unremitted depression, and the considerable gaps in knowledge about optimal management at key decision points, the National Institute of Mental Health (NIMH) has launched the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study. This is a multisite, multistep, prospective, randomized clinical trial of nonpsychotic major depressive disorder.²¹

Box 2

Approximately 4,000 adult depressed outpatients without an established history of antidepressant resistance will be recruited across the country. In contrast to efficacy trials, this effectiveness trial is meant to more closely parallel real-world practice (Box 2).

Patients will be treated initially with citalopram for 12 weeks (Level 1), with general guidelines for titration based upon response and tolerability. It is anticipated that about 2,000 individuals will exhibit unremitting depression despite SSRI treatment and/or intolerance to the SSRI and will be eligible for randomization to subsequent “switching” or “augmentation” strategies; this forms the core of the study across 3 successive levels.

Level 2 randomized treatments include four distinctive switching options—sertraline, bupropion, venlafaxine, or cognitive therapy—as well as 3 augmentation options—citalopram plus bupropion, buspirone, or cognitive therapy.

Those whose depression fails to remit will be eligible for randomization at Level 3 to different switching options—mirtazapine vs. nortripyline, or augmentation options—lithium or thyroid—added to the primary Level 2 antidepressant.

Finally, at Level 4, patients who continue to have unremitted depression despite successive treatments will be eligible for random assignment to 1 of 2 switching options—tranylcypromine vs. the combination of mirtazapine and venlafaxine.

Two new somatic treatments

Repetitive Transcranial Magnetic Stimulation (rTMS) rTMS is used to induce electrical current in the brain without causing seizures. Using a rapidly alternating current with a hand-held electromagnet that generates about 2 tesla will, in turn, induce or interrupt current (depolarization) in cortical interneurons about 2 cm below the surface of the skull. rTMS delivers energy without the impedance of the skull and is thought to be excitatory at high frequencies (20 Hz) and inhibitory at lower frequencies (5 to 10 Hz). This theory remains to be rigorously tested.²²

Overall, the clinical effects reported in early studies on rTMS have been modest at best, while later studies have found either no difference from placebo or a robust effect. rTMS is a potential therapeutic tool not yet ready for widespread clinical use.²²

Vagus Nerve Stimulation The NeuroCybernetic Prosthesis System (Cyberonics, Inc., Houston, Tex) stimulates the left cervical vagus nerve for treatment of resistant partial-onset epileptic seizures.

An open treatment trial of vagus nerve stimulation (VNS) for treatment-resistant depression was done based on mood improvement observed in patients treated for seizure disorders. Positron-emission tomographic studies revealed activation of limbic structures and neurochemical and neuronal pathway findings.²³ The investigators reported improvement in a sample of 30 patients with major depressive disorder. Definite conclusions await controlled clinical trials and more work is needed to determine the ultimate mechanisms that might be antidepressant.^23,24

Related resources

• Amsterdam JD, Hornig M, Nierenberg AA. Treatment-resistant mood disorders. Cambridge, UK: Cambridge University Press; 2001.
• Stahl SM. Essential Psychopharmacology. Cambridge, UK: Cambridge University Press; 1996.
• STAR*D Web site: http://www.edc.gsph.pitt.edu/stard

Drug brand names

• Bupropion • Wellbutrin, Wellbutrin SR, Zybar
• Buspirone • BuSpar, BuSpar DIVIDOSE
• Citalopram • Celexa
• Fluoxetine • Prozac, Prozac Weekly
• Gabapentin • Neurontin
• Haloperidol • Haldol, Haldol Decanoate
• Lamotrigine • Lamictal
• Mirtazapine • Remeron, Remeron Soltab
• Modafinil • Provigil
• Nefazodone • Serzone
• Olanzapine • Zyprexa
• Paroxetine • Paxil
• Pergolide • Permax
• Pramipexole • Mirapex
• Reboxetine • Edronax
• Risperidone • Risperidal
• S-adenosyl-methionine • SAM-e
• Sertaline • Zoloft
• Tranycypromine • Nardil, Parnate, Marplan
• Valproate • Epival, Deprox, Alti-Valproic
• Venlafaxine • Effexor, Effexor X12
• Ziprasadone • Geodon

Disclosure

Dr. Alpert receives research funding, speaking honoraria or consultation fees from Organon, Forest Pharmaceuticals, GlaxoSmithKline, Pfizer Inc., and Pharmavite.

Dr. Nierenberg receives research funding, speaking honoraria, or consultation fees from Eli Lilly and Co., Wyeth-Ayerst Pharmaceuticals, GlaxoSmithKline, Janssen Pharmaceutica, Innapharma, BMS, Cyberonics, Lichtner, and Pfizer Inc.

As many as 50% of depressed patients do not achieve a 50% or greater reduction in severity of symptoms after an adequate antidepressant trial.¹ Moreover, among those who do respond to acute treatment, longer-term residual depressive symptoms are quite common. Persistent subsyndromal depressive symptoms are associated with impaired psychosocial functioning and increased risk of relapse. This possibility has reinforced an evolving consensus that full depressive remission rather than response is the proper goal of treatment.²

When your patient’s depressive symptoms fail to remit following an initial pharmacotherapy course, you must decide which of numerous possible next steps to pursue, in what combinations, and following what sequence. The range of strategies include the following:

1. Pursuing an extended trial with the initial agent using higher than usual dosages (e.g., fluoxetine 40 to 80 mg/qd);
2. Augmenting an antidepressant with an agent that offers no established antidepressant efficacy on its own, such as buspirone, lithium, thyroid, or estrogen;
3. Combining an antidepressant with another antidepressant, with another somatic therapy such as electroconvulsive treatment or phototherapy, or with some form of psychotherapy;
4. Switching to an antidepressant within the same class (e.g., from one selective serotonin reuptake inhibitor [SSRI] to another), or outside of class (from an SSRI to an atypical antidepressant), or switching to non-pharmacological somatic therapy or psychotherapy.³

A predictable lag exists between innovative clinical applications and the randomized, controlled trials (RCTs) designed to evaluate them. The result: Approaches that are thoroughly described in the literature involve medication regimens that are no longer first-line, such as lithium or thyroid augmentation of tricyclic antidepressants (TCAs), or the combination of TCAs with SSRIs. Conversely, approaches most widely used in current psychiatric practice (e.g., addition of bupropion to SSRIs) have received relatively little systematic attention.⁴

Further, while predictors of initial antidepressant response have been hard to come by, even less is known about predictors of antidepressant response after lack of response to previous antidepressants. The working hypotheses that clinicians use to decide which strategies to pursue, though plausible, are largely untested, whether based on characteristics of a patient’s initial response (e.g., non-response or partial response)⁵ or on side effects or comorbid diagnoses.

This review will describe new research and emerging strategies that address the common clinical problem of unremitted depression despite one or more adequate courses of antidepressant treatment.

A look at new augmentations and combinations

Adding a second agent is an appealing strategy when patients have tolerated an initial antidepressant without troublesome side effects or have shown partial response (≥ 25% and <50% symptom reduction) and/or when a second agent may serve an important additional goal such as treating a comorbid condition—attention-deficit disorder or smoking, for example—or a side effect (e.g., nausea or insomnia).

Box 1

The primary drawbacks to augmentation/combination strategies are an increased risk of drug interactions, cost, and potential decrement in adherence to treatment as the regimen’s complexity increases.

Thyroid augmentation Although thyroid augmentation of TCAs has been the subject of numerous RCTs, no controlled trials of thyroid augmentation of SSRIs or other newer-generation antidepressants are available.

Lithium augmentation Studies of lithium augmentation of SSRIs have yielded generally modest response rates with questionable durability.^6,7

Noradrenergic and/or dopaminergic agonist agents Open trials and case series in the treatment of patients with unremitted depression have supported combining SSRIs and venlafaxine with agents that possess primarily noradrenergic and/or dopaminergic agonist properties. These include bupropion,⁸ psychostimulants,⁹ and direct dopamine agonists, including pergolide and pramipexole.¹⁰

These combinations have the advantage of potentially ameliorating several common SSRI side effects, particularly sedation, sexual dysfunction, and putative SSRI-related apathy. In the case of psychostimulants and/or bupropion, the combinations may treat such common comorbidities as attention-deficit/hyperactivity disorder and smoking.

Modafinil The combination of the antinarcoleptic modafinil with SSRIs and other newer-generation antidepressants has also attracted interest. Its efficacy as an antidepressant adjunct and as a remedy for side effects must still be established, however.¹¹

Buspirone augmentation The augmentation of SSRIs and other antidepressants with the antianxiety azaperone buspirone is supported by impressive response rates in several open trials. But the only placebo-controlled trial evaluating buspirone augmentation in resistant depression failed to find a significant drug:placebo difference.¹² Still, its tolerability and anxiolytic efficacy, and potential for ameliorating sexual dysfunction in some patients, support its judicious use as an augmenting agent pending further study.

Pindolol Studies of this beta-agonist, 5HT-1A antagonist, as an antidepressant-augmenting combination have yielded promising results in some studies, including a negative 10-day, placebo-controlled trial.¹³ Some patients experience jitteriness or irritability on pindolol. Its role in treating resistant depression remains to be established.

Noradrenergic TCAs The combination of SSRIs with these agents is a good example of a strategy based on pairing complimentary antidepressant mechanisms. Double-blind, controlled trials of fluoxetine plus desipramine^6,7 have tempered the enthusiasm generated by earlier open studies, however. Response rates in these trials range from 25% to 30%, no higher than lithium augmentation and slightly lower than higher-dose fluoxetine.

Mirtazapine A more recent combination is mirtazapine with SSRIs or with high-dose venlafaxine.¹⁴ In some patients, this combination may attenuate sexual dysfunction, insomnia, and gastrointestinal side effects of SSRIs and venlafaxine by virtue of a 5-HT2, 5-HT3, and histamine receptor blockade by mirtazapine. This combination may also capitalize on the combined antidepressant effects of direct norepinephrine (NE) and serotonin (5-HT) reuptake inhibition as well as an alpha₂ adrenergic auto- and hetero-receptor blockade—facilitating presynaptic NE and 5HT release—and 5-HT2 receptor antagonism.

Antipsychotics The use of antipsychotics for nonpsychotic unipolar depression has been controversial. However, in the context of lower apparent risks of tardive dyskinesia with the newer, atypical antipsychotics, their use as antidepressant augmenting agents has been revisited. In one double-blind, placebo-controlled trial, the combination of fluoxetine plus olanzapine was more effective in a well-defined refractory depressed population than was either medication alone.¹⁵ (See Box 1 for a discussion of fluoxetine and folic acid.)

Ostroff and Nelson¹⁶ reported an extremely rapid (1 week) response after the addition of risperidone among 8 depressed individuals who had not responded to fluoxetine or paroxetine alone. The activating properties of ziprasadone for some patients, combined with its lower propensity for producing weight gain than other atypical antipsychotics, make it another potentially attractive candidate for antidepressant augmentation, though one that requires further study.

Anticonvulsant augmentation While controlled studies of anticonvulsant augmentation of newer-generation antidepressants are lacking, the efficacy of lamotrigine for treating bipolar depression¹⁷ has encouraged clinicians to combine the agent with antidepressants in unipolar depression. The potential sedative/anxiolytic effects of other anticonvulsants, including gabapentin and valproate, have also supported their use in resistant depression complicated by anxiety or irritability. Omega-3-fatty acids may work as mood-stabilizing substances when used for antidepressant augmentation.

Antidepressant switches: In class or out?

Switching from TCAs to SSRIs or vice versa following nonresponse has yielded generally high (50% to 60%) response rates.³ But the more common question now is whether to switch from an initial SSRI to another SSRI, to an SSRI-like agent with an additional noradrenergic mechanism (such as venlafaxine), or to an atypical antidepressant such as bupropion or mirtazapine.

Traditional teaching and regular psychiatric practice⁵ have favored a switch outside of class. More recently, however, a number of uncontrolled studies looking at response rates to a second SSRI following inadequate response to and/or tolerance of an initial SSRI have shown response rates of 40% to 75%.¹⁸ This coincides with response rates seen in studies involving crossover to antidepressants with differing mechanisms. Inferences from these studies are limited by, first, determining nonresponse retrospectively and, second, lumping together patients who did not remain on the first SSRI because of nonresponse with others who curtailed treatment because of intolerance.

Because of its dual 5-HT/NE reuptake inhibitory activity, particularly at higher doses, venlafaxine is a popular choice following failure of an initial SSRI. One-third of subjects with extremely resistant depression (3 or more trials of antidepressants) responded to venlafaxine.¹⁹ DeMontigny et al²⁰ also demonstrated its efficacy (with a 58% response rate) as a switching agent among resistant, depressed patients.

But comparative response rates for switching to venlafaxine from an SSRI vs. from other antidepressant classes have not been well delineated, nor have response rates for switching from SSRIs and other antidepressants to venlafaxine vs. to other non-SSRI agents.

Switches from SSRIs to nefazodone and mirtazapine or the selective NE reuptake inhibitor reboxetine have shown reasonably high response rates.³ Studies comparing these various strategies head to head are lacking. In the absence of data, reasonable guides for switching agents include projected tolerability and existence of comorbid conditions that may respond well to one agent over another.

Another issue is whether to cross-taper, that is, to lower the dose of the first agent while titrating up the next, or to discontinue the first before the second. When catastrophic drug interactions would result from overlap, such as the switch from an SSRI to a monoamine oxidase (MAO) inhibitor, a suitable wash-out period is necessary. This can also occur if the first antidepressant is believed to cause intolerable side effects.

In most other instances where no absolute contraindications exist, cross-tapering may be worthwhile despite theoretical risks such as serotonin syndrome when overlapping an SSRI and venlafaxine. Cross-tapering helps reduce the risk of side effects associated with abrupt discontinuation including nausea, myalgias, headache, and dysphoria, and may minimize the loss of antidepressant benefit from the initial agent before the impact of the second agent is realized.

Great hope for new knowledge with STAR*D

In recognition of the substantial direct and indirect costs, the morbidity and mortality associated with unremitted depression, and the considerable gaps in knowledge about optimal management at key decision points, the National Institute of Mental Health (NIMH) has launched the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study. This is a multisite, multistep, prospective, randomized clinical trial of nonpsychotic major depressive disorder.²¹

Box 2

Approximately 4,000 adult depressed outpatients without an established history of antidepressant resistance will be recruited across the country. In contrast to efficacy trials, this effectiveness trial is meant to more closely parallel real-world practice (Box 2).

Patients will be treated initially with citalopram for 12 weeks (Level 1), with general guidelines for titration based upon response and tolerability. It is anticipated that about 2,000 individuals will exhibit unremitting depression despite SSRI treatment and/or intolerance to the SSRI and will be eligible for randomization to subsequent “switching” or “augmentation” strategies; this forms the core of the study across 3 successive levels.

Level 2 randomized treatments include four distinctive switching options—sertraline, bupropion, venlafaxine, or cognitive therapy—as well as 3 augmentation options—citalopram plus bupropion, buspirone, or cognitive therapy.

Those whose depression fails to remit will be eligible for randomization at Level 3 to different switching options—mirtazapine vs. nortripyline, or augmentation options—lithium or thyroid—added to the primary Level 2 antidepressant.

Finally, at Level 4, patients who continue to have unremitted depression despite successive treatments will be eligible for random assignment to 1 of 2 switching options—tranylcypromine vs. the combination of mirtazapine and venlafaxine.

Two new somatic treatments

Repetitive Transcranial Magnetic Stimulation (rTMS) rTMS is used to induce electrical current in the brain without causing seizures. Using a rapidly alternating current with a hand-held electromagnet that generates about 2 tesla will, in turn, induce or interrupt current (depolarization) in cortical interneurons about 2 cm below the surface of the skull. rTMS delivers energy without the impedance of the skull and is thought to be excitatory at high frequencies (20 Hz) and inhibitory at lower frequencies (5 to 10 Hz). This theory remains to be rigorously tested.²²

Overall, the clinical effects reported in early studies on rTMS have been modest at best, while later studies have found either no difference from placebo or a robust effect. rTMS is a potential therapeutic tool not yet ready for widespread clinical use.²²

Vagus Nerve Stimulation The NeuroCybernetic Prosthesis System (Cyberonics, Inc., Houston, Tex) stimulates the left cervical vagus nerve for treatment of resistant partial-onset epileptic seizures.

An open treatment trial of vagus nerve stimulation (VNS) for treatment-resistant depression was done based on mood improvement observed in patients treated for seizure disorders. Positron-emission tomographic studies revealed activation of limbic structures and neurochemical and neuronal pathway findings.²³ The investigators reported improvement in a sample of 30 patients with major depressive disorder. Definite conclusions await controlled clinical trials and more work is needed to determine the ultimate mechanisms that might be antidepressant.^23,24

Related resources

• Amsterdam JD, Hornig M, Nierenberg AA. Treatment-resistant mood disorders. Cambridge, UK: Cambridge University Press; 2001.
• Stahl SM. Essential Psychopharmacology. Cambridge, UK: Cambridge University Press; 1996.
• STAR*D Web site: http://www.edc.gsph.pitt.edu/stard

Drug brand names

• Bupropion • Wellbutrin, Wellbutrin SR, Zybar
• Buspirone • BuSpar, BuSpar DIVIDOSE
• Citalopram • Celexa
• Fluoxetine • Prozac, Prozac Weekly
• Gabapentin • Neurontin
• Haloperidol • Haldol, Haldol Decanoate
• Lamotrigine • Lamictal
• Mirtazapine • Remeron, Remeron Soltab
• Modafinil • Provigil
• Nefazodone • Serzone
• Olanzapine • Zyprexa
• Paroxetine • Paxil
• Pergolide • Permax
• Pramipexole • Mirapex
• Reboxetine • Edronax
• Risperidone • Risperidal
• S-adenosyl-methionine • SAM-e
• Sertaline • Zoloft
• Tranycypromine • Nardil, Parnate, Marplan
• Valproate • Epival, Deprox, Alti-Valproic
• Venlafaxine • Effexor, Effexor X12
• Ziprasadone • Geodon

Disclosure

Dr. Alpert receives research funding, speaking honoraria or consultation fees from Organon, Forest Pharmaceuticals, GlaxoSmithKline, Pfizer Inc., and Pharmavite.

Dr. Nierenberg receives research funding, speaking honoraria, or consultation fees from Eli Lilly and Co., Wyeth-Ayerst Pharmaceuticals, GlaxoSmithKline, Janssen Pharmaceutica, Innapharma, BMS, Cyberonics, Lichtner, and Pfizer Inc.

As many as 50% of depressed patients do not achieve a 50% or greater reduction in severity of symptoms after an adequate antidepressant trial.¹ Moreover, among those who do respond to acute treatment, longer-term residual depressive symptoms are quite common. Persistent subsyndromal depressive symptoms are associated with impaired psychosocial functioning and increased risk of relapse. This possibility has reinforced an evolving consensus that full depressive remission rather than response is the proper goal of treatment.²

When your patient’s depressive symptoms fail to remit following an initial pharmacotherapy course, you must decide which of numerous possible next steps to pursue, in what combinations, and following what sequence. The range of strategies include the following:

1. Pursuing an extended trial with the initial agent using higher than usual dosages (e.g., fluoxetine 40 to 80 mg/qd);
2. Augmenting an antidepressant with an agent that offers no established antidepressant efficacy on its own, such as buspirone, lithium, thyroid, or estrogen;
3. Combining an antidepressant with another antidepressant, with another somatic therapy such as electroconvulsive treatment or phototherapy, or with some form of psychotherapy;
4. Switching to an antidepressant within the same class (e.g., from one selective serotonin reuptake inhibitor [SSRI] to another), or outside of class (from an SSRI to an atypical antidepressant), or switching to non-pharmacological somatic therapy or psychotherapy.³

A predictable lag exists between innovative clinical applications and the randomized, controlled trials (RCTs) designed to evaluate them. The result: Approaches that are thoroughly described in the literature involve medication regimens that are no longer first-line, such as lithium or thyroid augmentation of tricyclic antidepressants (TCAs), or the combination of TCAs with SSRIs. Conversely, approaches most widely used in current psychiatric practice (e.g., addition of bupropion to SSRIs) have received relatively little systematic attention.⁴

Further, while predictors of initial antidepressant response have been hard to come by, even less is known about predictors of antidepressant response after lack of response to previous antidepressants. The working hypotheses that clinicians use to decide which strategies to pursue, though plausible, are largely untested, whether based on characteristics of a patient’s initial response (e.g., non-response or partial response)⁵ or on side effects or comorbid diagnoses.

This review will describe new research and emerging strategies that address the common clinical problem of unremitted depression despite one or more adequate courses of antidepressant treatment.

A look at new augmentations and combinations

Adding a second agent is an appealing strategy when patients have tolerated an initial antidepressant without troublesome side effects or have shown partial response (≥ 25% and <50% symptom reduction) and/or when a second agent may serve an important additional goal such as treating a comorbid condition—attention-deficit disorder or smoking, for example—or a side effect (e.g., nausea or insomnia).

Box 1

The primary drawbacks to augmentation/combination strategies are an increased risk of drug interactions, cost, and potential decrement in adherence to treatment as the regimen’s complexity increases.

Thyroid augmentation Although thyroid augmentation of TCAs has been the subject of numerous RCTs, no controlled trials of thyroid augmentation of SSRIs or other newer-generation antidepressants are available.

Lithium augmentation Studies of lithium augmentation of SSRIs have yielded generally modest response rates with questionable durability.^6,7

Noradrenergic and/or dopaminergic agonist agents Open trials and case series in the treatment of patients with unremitted depression have supported combining SSRIs and venlafaxine with agents that possess primarily noradrenergic and/or dopaminergic agonist properties. These include bupropion,⁸ psychostimulants,⁹ and direct dopamine agonists, including pergolide and pramipexole.¹⁰

These combinations have the advantage of potentially ameliorating several common SSRI side effects, particularly sedation, sexual dysfunction, and putative SSRI-related apathy. In the case of psychostimulants and/or bupropion, the combinations may treat such common comorbidities as attention-deficit/hyperactivity disorder and smoking.

Modafinil The combination of the antinarcoleptic modafinil with SSRIs and other newer-generation antidepressants has also attracted interest. Its efficacy as an antidepressant adjunct and as a remedy for side effects must still be established, however.¹¹

Buspirone augmentation The augmentation of SSRIs and other antidepressants with the antianxiety azaperone buspirone is supported by impressive response rates in several open trials. But the only placebo-controlled trial evaluating buspirone augmentation in resistant depression failed to find a significant drug:placebo difference.¹² Still, its tolerability and anxiolytic efficacy, and potential for ameliorating sexual dysfunction in some patients, support its judicious use as an augmenting agent pending further study.

Pindolol Studies of this beta-agonist, 5HT-1A antagonist, as an antidepressant-augmenting combination have yielded promising results in some studies, including a negative 10-day, placebo-controlled trial.¹³ Some patients experience jitteriness or irritability on pindolol. Its role in treating resistant depression remains to be established.

Noradrenergic TCAs The combination of SSRIs with these agents is a good example of a strategy based on pairing complimentary antidepressant mechanisms. Double-blind, controlled trials of fluoxetine plus desipramine^6,7 have tempered the enthusiasm generated by earlier open studies, however. Response rates in these trials range from 25% to 30%, no higher than lithium augmentation and slightly lower than higher-dose fluoxetine.

Mirtazapine A more recent combination is mirtazapine with SSRIs or with high-dose venlafaxine.¹⁴ In some patients, this combination may attenuate sexual dysfunction, insomnia, and gastrointestinal side effects of SSRIs and venlafaxine by virtue of a 5-HT2, 5-HT3, and histamine receptor blockade by mirtazapine. This combination may also capitalize on the combined antidepressant effects of direct norepinephrine (NE) and serotonin (5-HT) reuptake inhibition as well as an alpha₂ adrenergic auto- and hetero-receptor blockade—facilitating presynaptic NE and 5HT release—and 5-HT2 receptor antagonism.

Antipsychotics The use of antipsychotics for nonpsychotic unipolar depression has been controversial. However, in the context of lower apparent risks of tardive dyskinesia with the newer, atypical antipsychotics, their use as antidepressant augmenting agents has been revisited. In one double-blind, placebo-controlled trial, the combination of fluoxetine plus olanzapine was more effective in a well-defined refractory depressed population than was either medication alone.¹⁵ (See Box 1 for a discussion of fluoxetine and folic acid.)

Ostroff and Nelson¹⁶ reported an extremely rapid (1 week) response after the addition of risperidone among 8 depressed individuals who had not responded to fluoxetine or paroxetine alone. The activating properties of ziprasadone for some patients, combined with its lower propensity for producing weight gain than other atypical antipsychotics, make it another potentially attractive candidate for antidepressant augmentation, though one that requires further study.

Anticonvulsant augmentation While controlled studies of anticonvulsant augmentation of newer-generation antidepressants are lacking, the efficacy of lamotrigine for treating bipolar depression¹⁷ has encouraged clinicians to combine the agent with antidepressants in unipolar depression. The potential sedative/anxiolytic effects of other anticonvulsants, including gabapentin and valproate, have also supported their use in resistant depression complicated by anxiety or irritability. Omega-3-fatty acids may work as mood-stabilizing substances when used for antidepressant augmentation.

Antidepressant switches: In class or out?

Switching from TCAs to SSRIs or vice versa following nonresponse has yielded generally high (50% to 60%) response rates.³ But the more common question now is whether to switch from an initial SSRI to another SSRI, to an SSRI-like agent with an additional noradrenergic mechanism (such as venlafaxine), or to an atypical antidepressant such as bupropion or mirtazapine.

Traditional teaching and regular psychiatric practice⁵ have favored a switch outside of class. More recently, however, a number of uncontrolled studies looking at response rates to a second SSRI following inadequate response to and/or tolerance of an initial SSRI have shown response rates of 40% to 75%.¹⁸ This coincides with response rates seen in studies involving crossover to antidepressants with differing mechanisms. Inferences from these studies are limited by, first, determining nonresponse retrospectively and, second, lumping together patients who did not remain on the first SSRI because of nonresponse with others who curtailed treatment because of intolerance.

Because of its dual 5-HT/NE reuptake inhibitory activity, particularly at higher doses, venlafaxine is a popular choice following failure of an initial SSRI. One-third of subjects with extremely resistant depression (3 or more trials of antidepressants) responded to venlafaxine.¹⁹ DeMontigny et al²⁰ also demonstrated its efficacy (with a 58% response rate) as a switching agent among resistant, depressed patients.

But comparative response rates for switching to venlafaxine from an SSRI vs. from other antidepressant classes have not been well delineated, nor have response rates for switching from SSRIs and other antidepressants to venlafaxine vs. to other non-SSRI agents.

Switches from SSRIs to nefazodone and mirtazapine or the selective NE reuptake inhibitor reboxetine have shown reasonably high response rates.³ Studies comparing these various strategies head to head are lacking. In the absence of data, reasonable guides for switching agents include projected tolerability and existence of comorbid conditions that may respond well to one agent over another.

Another issue is whether to cross-taper, that is, to lower the dose of the first agent while titrating up the next, or to discontinue the first before the second. When catastrophic drug interactions would result from overlap, such as the switch from an SSRI to a monoamine oxidase (MAO) inhibitor, a suitable wash-out period is necessary. This can also occur if the first antidepressant is believed to cause intolerable side effects.

In most other instances where no absolute contraindications exist, cross-tapering may be worthwhile despite theoretical risks such as serotonin syndrome when overlapping an SSRI and venlafaxine. Cross-tapering helps reduce the risk of side effects associated with abrupt discontinuation including nausea, myalgias, headache, and dysphoria, and may minimize the loss of antidepressant benefit from the initial agent before the impact of the second agent is realized.

Great hope for new knowledge with STAR*D

In recognition of the substantial direct and indirect costs, the morbidity and mortality associated with unremitted depression, and the considerable gaps in knowledge about optimal management at key decision points, the National Institute of Mental Health (NIMH) has launched the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study. This is a multisite, multistep, prospective, randomized clinical trial of nonpsychotic major depressive disorder.²¹

Box 2

Approximately 4,000 adult depressed outpatients without an established history of antidepressant resistance will be recruited across the country. In contrast to efficacy trials, this effectiveness trial is meant to more closely parallel real-world practice (Box 2).

Patients will be treated initially with citalopram for 12 weeks (Level 1), with general guidelines for titration based upon response and tolerability. It is anticipated that about 2,000 individuals will exhibit unremitting depression despite SSRI treatment and/or intolerance to the SSRI and will be eligible for randomization to subsequent “switching” or “augmentation” strategies; this forms the core of the study across 3 successive levels.

Level 2 randomized treatments include four distinctive switching options—sertraline, bupropion, venlafaxine, or cognitive therapy—as well as 3 augmentation options—citalopram plus bupropion, buspirone, or cognitive therapy.

Those whose depression fails to remit will be eligible for randomization at Level 3 to different switching options—mirtazapine vs. nortripyline, or augmentation options—lithium or thyroid—added to the primary Level 2 antidepressant.

Finally, at Level 4, patients who continue to have unremitted depression despite successive treatments will be eligible for random assignment to 1 of 2 switching options—tranylcypromine vs. the combination of mirtazapine and venlafaxine.

Two new somatic treatments

Repetitive Transcranial Magnetic Stimulation (rTMS) rTMS is used to induce electrical current in the brain without causing seizures. Using a rapidly alternating current with a hand-held electromagnet that generates about 2 tesla will, in turn, induce or interrupt current (depolarization) in cortical interneurons about 2 cm below the surface of the skull. rTMS delivers energy without the impedance of the skull and is thought to be excitatory at high frequencies (20 Hz) and inhibitory at lower frequencies (5 to 10 Hz). This theory remains to be rigorously tested.²²

Overall, the clinical effects reported in early studies on rTMS have been modest at best, while later studies have found either no difference from placebo or a robust effect. rTMS is a potential therapeutic tool not yet ready for widespread clinical use.²²

Vagus Nerve Stimulation The NeuroCybernetic Prosthesis System (Cyberonics, Inc., Houston, Tex) stimulates the left cervical vagus nerve for treatment of resistant partial-onset epileptic seizures.

An open treatment trial of vagus nerve stimulation (VNS) for treatment-resistant depression was done based on mood improvement observed in patients treated for seizure disorders. Positron-emission tomographic studies revealed activation of limbic structures and neurochemical and neuronal pathway findings.²³ The investigators reported improvement in a sample of 30 patients with major depressive disorder. Definite conclusions await controlled clinical trials and more work is needed to determine the ultimate mechanisms that might be antidepressant.^23,24

Related resources

• Amsterdam JD, Hornig M, Nierenberg AA. Treatment-resistant mood disorders. Cambridge, UK: Cambridge University Press; 2001.
• Stahl SM. Essential Psychopharmacology. Cambridge, UK: Cambridge University Press; 1996.
• STAR*D Web site: http://www.edc.gsph.pitt.edu/stard

Drug brand names

• Bupropion • Wellbutrin, Wellbutrin SR, Zybar
• Buspirone • BuSpar, BuSpar DIVIDOSE
• Citalopram • Celexa
• Fluoxetine • Prozac, Prozac Weekly
• Gabapentin • Neurontin
• Haloperidol • Haldol, Haldol Decanoate
• Lamotrigine • Lamictal
• Mirtazapine • Remeron, Remeron Soltab
• Modafinil • Provigil
• Nefazodone • Serzone
• Olanzapine • Zyprexa
• Paroxetine • Paxil
• Pergolide • Permax
• Pramipexole • Mirapex
• Reboxetine • Edronax
• Risperidone • Risperidal
• S-adenosyl-methionine • SAM-e
• Sertaline • Zoloft
• Tranycypromine • Nardil, Parnate, Marplan
• Valproate • Epival, Deprox, Alti-Valproic
• Venlafaxine • Effexor, Effexor X12
• Ziprasadone • Geodon

Disclosure

Dr. Alpert receives research funding, speaking honoraria or consultation fees from Organon, Forest Pharmaceuticals, GlaxoSmithKline, Pfizer Inc., and Pharmavite.

Dr. Nierenberg receives research funding, speaking honoraria, or consultation fees from Eli Lilly and Co., Wyeth-Ayerst Pharmaceuticals, GlaxoSmithKline, Janssen Pharmaceutica, Innapharma, BMS, Cyberonics, Lichtner, and Pfizer Inc.