Sec. 11
< 10 - Use of Antidepressants With Other Medications Table of Contents 12 - References >
Outpatient Management of Depression
11 - When the Patient Does Not Respond Well to Initial Antidepressant Therapy

Regrettably, no single antidepressant will produce an adequate response in every patient with clinical depression. Despite using all of the information summarized in the preceding chapter to make the best possible treatment selection, some patients will not respond because they do not tolerate the medication well enough to remain on it for an adequate period of time to respond. Others will simply have a form of the illness which is not responsive to the mechanism of action of the antidepressant. This chapter will present some simple and useful approaches to take when faced with such an outcome.

Some of the recommendations in this chapter will be to add a second medication either as an "antidote" for a treatment limiting adverse effect or as a means of boosting efficacy in a patient who has experienced an incomplete response. In essence, these approaches are planned drug-drug interactions and thus are the opposite side of the issue discussed in Chapter 10. Although such therapeutic drug-drug interactions can be quite clinically helpful when properly employed, they should only be used when monodrug therapy has failed to produce an acceptable response. The goal is always to keep treatment as simple and safe as possible.

Treatment Options When Faced With Adverse Effects

The practitioner can use the pharmacological profiles of the various antidepressants discussed in Chapters 6 through 8 to optimally match the antidepressant to the patient. Nevertheless, adverse effects will occur despite the most careful selection. When the effect is serious and/or severe, the clinician can refer to Table 6.4 to determine what mechanism likely mediates the effect, and then use Table 6.3 to select an antidepressant which does not have that mechanism of action.

In many cases, the adverse effect will be a nuisance, but will not warrant a medication switch. Often, these adverse effects are dose dependent and a dose reduction may alleviate the problem. Moreover, tolerance occurs for many of the acute adverse effects of the antidepressants probably as a result of receptor down regulation. For this reason, the dose can be gradually re-escalated, if needed for optimal efficacy, without the adverse effect necessarily recurring. Treatment of the most common nuisance adverse effects of the various antidepressants (Tables 6.7 and 6.8) are discussed below.

  • Anorexia

This effect is most commonly mediated by indirect serotonin (5-hydroxytryptophan [5-HT]) agonism, probably of the 5-HT2C receptor, produced by the inhibition of serotonin uptake.238 It occurs in a dose-dependent fashion with all serotonin selective reuptake inhibitors (SSRIs) and venlafaxine. Tolerance commonly develops.103,171 Such tolerance is the reason these antidepressants were not able to show sustained weight loss when they were being tested as weight-reduction agents. If the anorexia persists and the clinician and patient do not want to switch to a different type of antidepressant, then the addition of a 5-HT2C antagonists such as mirtazapine or cyproheptadine can be helpful.

  • Confusion

This effect can occur as a result of several mechanisms of action; principally, the blockade of the muscarinic acetylcholine, histamine-1, and 5-HT2A receptors.83,193,220 It is dose dependent. Tolerance frequently does not develop. There is no good antidote.

  • Constipation

This effect is most commonly seen with antidepressants that block the muscarinic acetylcholine receptor, but may also occur with norepinephrine uptake inhibitors. It is dose dependent. Tolerance frequently does not develop when it is caused by muscarinic acetylcholine receptor blockade. Bulk forming stool softeners can be used to minimize the problem.

  • Diaphoresis

This effect is most commonly seen with antidepressants that inhibit norepinephrine uptake. It is frequently not dose dependent. Tolerance often does not occur. Beta blockers can be helpful in some cases.

  • Diarrhea

Loose or frequent stools is a better descriptor. This effect occurs in a dose-dependent fashion with all SSRIs and venlafaxine. Tolerance commonly develops. Drugs which slow gut motility can be helpful, such as lomotil or anticholinergics.

  • Dizziness

This term is often used to describe two different effects: one associated with decreased peripheral vascular resistance as a result of alpha-1-adrenergic receptor blockade and one due to direct or indirect 5-HT1A stimulation.83 The former is clinically more important than the latter because it is more likely to be associated with loss of balance leading to falls, and is more likely to be persistent. There is no good antidote besides dose reduction. When only tricyclic antidepressants (TCAs) were available, clinicians would recommend increasing the salt intake and the use of thromboembolic disease hose; these steps can be helpful in some cases. Falls can be minimized by instructing patients to slowly change from lying to sitting and then to standing to allow accommodation to occur.

The dizziness caused by 5-HT1A stimulation is most likely centrally mediated. It can occur early in treatment with antidepressants which inhibit serotonin uptake (eg, SSRIs and venlafaxine) and/or block 5-HT2A receptors (eg, nefazodone and mirtazapine) as well as drugs which directly stimulate the 5-HT1A receptor (eg, buspirone). There are no cardiovascular changes associated with it. Balance and coordination are subjectively, but not objectively, affected. Tolerance commonly develops within days to a week. While it is not medically serious, it can be disturbing if the patient has not been forewarned about this potential adverse effect.

  • Drowsiness

This effect can occur as a result of several mechanisms of action, principally the blockade of the histamine-1 and 5-HT2A receptors and serotonin reuptake inhibition. It is generally more severe, persistent, and dose dependent when caused by histamine-1 receptor blockade. When the pharmacokinetics of the drugs permit (ie, half-life of approximately 1 day), the majority, if not all, of the dose can be given at night which can help with sleep and reduce, if not eliminate, daytime sedation.

As discussed in Chapter 8, there are some reasons to believe that this effect of mirtazapine may follow a curvilinear dose-response curve (ie, the effect diminishes with higher doses). The presumed mechanism is the offsetting arousal effect that can occur with the alpha-2-adrenergic receptor blockade produced by higher doses of mirtazapine (Figure 6.4).

  • Dry Mouth

This effect is most commonly seen with antidepressants that block the muscarinic acetylcholine receptor, but may also occur with norepinephrine uptake inhibitors. It is dose dependent. Tolerance frequently does not develop when caused by muscarinic acetylcholine receptor blockade. While it is frequently considered a trivial or nuisance adverse effect, it can result in increased tooth decay and gum disease due to the loss of the bacteriostatic effects of saliva. Patients should be advised to practice excellent dental hygiene, including brushing after every meal, avoiding snacks (particularly those which are sweet), and flossing. Chewing sugarless gum can be helpful.

  • Fatigue

This term is frequently used interchangeably with drowsiness and tiredness in clinical trial reporting. However, these terms are not necessarily synonymous. Fatigue can be a persistent symptom of an incompletely treated depressive episode. In other instances, it can be due to the type of disturbed sleep characteristically produced by serotonin reuptake inhibitors (SRIs) (ie, decreased rapid eye movement [REM], sleep and a shift from stage IV deep, restorative sleep to light stage I sleep).9,159,217,226,230 Paradoxically, patients with decreased REM sleep may report an increase in vivid dreaming since they are more likely to awaken during or near an REM interval as a result of the shift to light stage I sleep. For this type of daytime fatigue, the use of a 5-HT2A blocker such as trazodone can be helpful or the use of a short-lived benzodiazepine such as lorazepam to restore more normal sleep efficiency and restorative value.155

  • Hyperphagia

Although this effect was thought to be due to histamine-1 receptor blockade, it is more likely the result of 5-HT2C blockade which can be produced by antidepressants such as mirtazapine.83 Dose reduction is usually not helpful. Tolerance develops less frequently than with other adverse effects of antidepressants (eg, nausea). There are no truly effective antidotes. Instead, switching to another antidepressant is frequently necessary.

  • Insomnia

Initial insomnia can be either a persistent depressive symptom or an adverse effect of a more stimulating antidepressant such as a norepinephrine and/or dopamine reuptake inhibitor (eg, desipramine or bupropion, respectively). In the latter case, the approach is to move more of the dose to earlier in the day. Nevertheless, bupropion still must be given on a divided schedule as discussed in Chapter 8. Middle or late insomnia is more likely due to the disruptive effects that serotonin agonism can have on sleep architecture and is treated as discussed above under the section on fatigue.159,217,226,230

  • Nausea

This effect is most commonly mediated by indirect 5-HT3 agonism resulting from the inhibition of serotonin uptake.83 It occurs in a dose-dependent fashion with all SSRIs and venlafaxine. Tolerance typically develops. Cisapride can be used to block this effect when necessary,23 but it should not be used with antidepressants that produce substantial inhibition of the drug metabolizing cytochrome P450 (CYP) enzyme 3A (ie, fluvoxamine, nefazodone and high doses of fluoxetine) (Table 6.10).

  • Nervousness

This effect can occur as a result of the inhibition of the uptake pumps for norepinephrine and to a lesser extent serotonin. It is dose dependent. Tolerance may not develop. Beta blockers can be helpful in some patients. Benzodiazepines may also be used, but carry a modest risk for abuse and/or dependence.

  • Sexual Dysfunction

This effect is most commonly mediated by indirect serotonin agonism produced by the inhibition of serotonin uptake. It occurs in a dose-dependent fashion with all SSRIs and venlafaxine.145,147 Unlike most of the other adverse effects discussed here, this one typically occurs only after several weeks of treatment. Tolerance frequently does not develop.

Although many possible antidotes have been tried, no single effective approach has emerged. All of the following have been tried with varying degrees of success: buspirone (Buspar), bupropion (Wellbutrin), cyproheptadine (Periactin), yohimbine (Aphrodyne, Erex, Yocon, Yohimex), sildenafil (Viagra), and topically applied 1% testosterone creams for anorgasmia in women. The best antidote may be dopamine agonism produced by the addition of bupropion or methylphenidate. However, caution must be used with the dose of bupropion, particularly when used in conjunction with antidepressants that produce substantial inhibition of the drug metabolizing CYP enzymes (Table 6.10) as discussed in Chapter 6.

  • Vision Disturbance

This term is used to describe two different effects. The first is visual trails most likely caused by 5-HT receptor blockade.83 The second is impaired accommodation caused by the blockade of the muscarinic acetylcholine receptor. Visual trails are also referred to as "after images." Nefazodone is the antidepressant which is most likely to cause visual trails (Table 6.8). While it is not serious, some patients may be alarmed by its occurrence if not prewarned. Tolerance may develop in the case of both visual trails and accommodation. There are no established antidotes for either adverse effect.

Special notes: Dose-dependent adverse effects produced by serotonin reuptake inhibition can occur later in treatment and persist longer with fluoxetine than with the other SSRIs due to the long half-lives of the parent drug and its active metabolite, norfluoxetine.168 The clinician should keep this issue in mind when managing patients on this SSRI.

There are other, less frequent adverse effects and also more complicated management approaches than those discussed here. However, they are beyond what many primary-care practitioners would use in their practice. In such cases, referral to a psychiatrist who specializes in more complicated drug management may be the best course of action.

Partial Response

Partial response means that there has been at least a 25% improvement in the patient's symptomatology. While the patient is better, they are still symptomatic to a clinically significant extent. The clinician can quantitate the severity of the episode either by clinical interview or by the use of a patient self-reporting scale such as the Zung Depression Self-Report Rating Scale (Figure 9.2). The improvement in the depressive syndrome may be in one or more of the following areas:

  • Mood
  • Sleep
  • Appetite
  • Energy
  • Sex drive
  • Interest
  • Concentration/attention.

The areas that have improved are often dependent on the class of antidepressant used. For example, SRIs may produce more improvement in irritability, interest, and concentration/attention, while TCAs, nefazodone, and mirtazapine will often produce more improvement in sleep and appetite.103

  • Dose Adjustment

If the depressive episode has had a partial response but not a full remission after a 4-week trial, then a dose increase with all of the antidepressants except TCAs is a reasonable approach. This recommendation even holds true for the SSRIs despite their flat dose-response curve because the patient may be a rapid metabolizer and hence may develop plasma drug levels too low for an optimal response.170 The patient should receive a 4-week trial of this higher dose, assuming that his/her condition is improved sufficiently to warrant this disciplined approach and that s/he is agreeable to it. In the case of TCAs, the dose should have already been optimized based on therapeutic drug monitoring as discussed in Chapter 8. If the patient has had a partial response after 4 weeks of treatment with a TCA at therapeutic plasma drug levels, then a dose change is not warranted.191 Instead, the patient should be maintained on this dose for an additional 2 weeks to assess whether there will be any further response to this drug.

  • Augmentation Strategies

Another approach is augmentation strategies.244 The concept behind such strategies is a planned drug-drug interaction (Chapter 10). Unfortunately, the amount of systematic and controlled data supporting the usefulness and long-term safety of these approaches are modest. The data that exist are not as robust as initially suggested by case reports and anecdotal clinical literature. Nevertheless, such strategies may be beneficial and are frequently tried in psychiatric practices when dealing with patients in whom other approaches have not worked.

The potential advantages of such strategies include:

  • Conversion of partial responders to responders without having to start a new medication trial, and thus hopefully saving time and thus reducing patient suffering
  • At times, lower doses of one or both agents may be sufficient, minimizing the likelihood of adverse effects
  • The second drug may also treat a comorbid condition (eg, subclinical thyroid dysfunction).

The potential disadvantages stem from the fact that two agents are being used rather than one, and this increases the possibility of:

  • Adverse effects
  • Noncompliance
  • Expense.

The most frequent augmentation strategies involve the addition of lithium, thyroid hormone, or psychostimulants to antidepressants. More recently, agents such as pindolol and buspirone have also been tried.


The addition of lithium is perhaps the best studied and most popular augmentation strategy.97,152,212 Improvement is often reported within a few days to weeks of adding lithium. The dose of lithium used in this strategy (ie, 450 to 600 mg/day) is generally lower than that used for the treatment of acute mania or prophylaxis of bipolar disorder.231 Similarly, plasma lithium levels are lower in the range of 0.4 to 0.8 mEq/L. Added expense is minimal.

Thyroid Supplementation

There is a large body of literature outlining a strong association between clinical thyroid disease and psychiatric syndromes, especially mood disorders.68 Over the past 25 years, evidence has accumulated suggesting T3 augmentation of partial or nonresponders with major depression can be a useful clinical strategy.106 T3 (dosage range of 25 to 75 mg/day) is generally the preferred agent in the literature rather than a comparable dose of T4. The literature suggests that 50% to 60% of patients will benefit. One carefully controlled study comparing thyroid versus lithium augmentation found them to be equally efficacious.106

Crucial questions of biochemical and/or clinical predictors, dose-effect relationship, what constitutes an adequate trial, duration of treatment and long-term adverse effects have not been answered. However, the ease of use, the modest risk of untoward adverse drug-hormone interactions, and the administration of doses that are unlikely to effect the hypothalamic pituitary axis are factors that make this a useful strategy. Added expense is also minimal.

Psychostimulants/Dopamine Agonists
(methylphenidate and d-amphetamine)

Like bupropion, these drugs may be useful as monodrug therapy for certain patients with apathy and/or clinical depression associated with cerebrovascular accidents, Parkinson's disease and cancer. Their advantages include:

  • Relative safety
  • Rapidity of onset.

Their adverse effects can include:

  • Anxiety
  • Agitation
  • Rarely psychosis
  • Possible worsening of hypertension
  • Abuse/addictive potential in some patients.

These drugs may also be useful as augmenting agents in the treatment of depressive disorders.30,103 Dopamine agonists such as bromocriptine and pergolide have also been used as augmenting agents, but there is even less data supporting the efficacy and safety of such a combination. Added expense is more of an issue than with lithium and thyroid supplementation, but is still relatively modest.


Pindolol is a beta blocker and a 5-HT1A receptor antagonist which has recently been studied as a potentiating agent.27,149 Some open-label and controlled studies have reported a more rapid response and perhaps greater efficacy when pindolol is added to an SRI to treat patients with major depression. The presumed mechanism involves interruption in the short-loop negative feedback system, allowing for an increase in synaptic concentrations of 5-HT. Added expense is modest.


Buspirone is a partial 5-HT1A agonist marketed in the 1980s as an anxiolytic agent. There is some evidence suggesting that it may have weak intrinsic antidepressant properties. It has recently been promoted as possibly being useful as an augmenting strategy, particularly with SRIs.235 Neither the time of onset of action nor the optimal dose for augmentation has been rigorously established. When used as monodrug therapy for anxiety disorder, the dose is usually 15 to 30 mg/day and the time to onset is several weeks. Advantages of this augmentation strategy include its safety, the possible anxiolytic effects in the absence of abuse potential, and ease of use. Its principal disadvantage beyond limited data is cost. It is the most expensive of the augmentation strategies discussed in this section. Given that buspirone is still a patented drug, there will hopefully be more rigorous clinical trials to determine the relative merits of using this agent as an antidepressant augmentor.

Failed Trial

By definition, this term means that there has been less than a 25% change in depressive symptoms. In such cases, approaches other than those discussed above are generally recommended. The issue of what to do in such cases is arguably the greatest research need in antidepressant pharmacotherapy, but is disappointingly sparse.98,157,215,246 The existing studies have serious limitations such as small numbers of subjects and design problems, including no blinding, no random assignment, and no appropriate control group.

  • Alternative Monodrug Therapy

Some clinicians will try multiple members of the same class (eg, multiple sequential trials of different SSRIs) in patients who have not experienced adequate benefit. An argument can be made from a pharmacokinetic standpoint that a trial of a second SSRI might be a reasonable approach. Unfortunately, the studies that have been done to support this practice are all seriously flawed and thus of dubious merit.37,105,182,242 Certainly, the more rational approach after two SSRIs have failed would be switching to an antidepressant with a different presumed mechanism of action (Table 6.2).

The best data in this area come from double-blind, crossover studies which show that norepinephrine selective reuptake inhibitors (NSRIs) will work in approximately 50% of patients who have not benefited from an adequate trial of an SSRI and vice versa.103,245 Another approach is to escalate the dose of a drug with more than one antidepressant mechanism of action (eg, venlafaxine).156,177 For further discussion of this matter, refer to Chapters 6 through 8.

  • Copharmacy With Antidepressants

When two or more trials of monodrug therapy with antidepressants from different classes have failed to produce even a partial response, the most common approach is to use two chemically unrelated antidepressants in combination.245 This strategy is based more on clinical empiricism and a theoretical understanding of relevant neurotransmitter systems than on rigorous clinical trials. The earliest example of this strategy was the combined use of a TCA and an MAOI.103 With the introduction of safer antidepressants, the use of this specific combination has decreased due to concerns about toxicity. However, the use of other combinations of antidepressants has increased in terms of both frequency and the variety of combinations used. The most popular is the combined use of an SSRI (eg, sertraline) with an NSRI (eg, desipramine) or with a dopamine and norepinephrine reuptake inhibitor (eg, bupropion).153,258

The following points should be kept in mind when considering this option:


  • They are planned drug-drug interactions (Chapter 10) with the potential for increased adverse effects, the likely need to reduce the dose, and the need for even more careful monitoring, including therapeutic drug monitoring (TDM).
  • These strategies should be attempted in a systematic and well documented manner; patients should be well informed.
  • Certain combinations must be avoided such as an SSRI plus MAOI due to the potentially fatal serotonin syndrome.

Specific Combination Treatments for Special Patients

Some patients with specific types of clinical depression may need specific combination treatment strategies, such as:

  • Antidepressant plus anxiolytic
  • Antidepressant plus antipsychotic
  • Antidepressant plus mood stabilizer.

  • Antidepressant Plus Anxiolytic

There are patients with prominent anxiety and depressive symptoms (ie, mixed anxiety and depression).190 Ideally, monodrug therapy, preferably an antidepressant with proven efficacy in anxiety disorders, should be tried first (Table 8.2). When using this approach, it is advisable to start with a low dose of the antidepressant because a variety of antidepressants can cause nervousness (Chapter 6), and anxious patients may be particularly vulnerable to this adverse effect. If the patient experiences an increase in anxiety despite starting with a low dose, the combined use of the antidepressant plus anxiolytic (eg, clonazepam) can be used. After the antidepressant has had time to exert its full antidepressant effects (2 to 4 weeks), an attempt can be made to taper and stop the anxiolytic medication. There is a small percentage of patients in whom anxiety symptoms will recur when the anxiolytic is stopped, even though they have a full resolution of their depressive symptoms. Such patients may need indefinite combined treatment with an antidepressant and anxiolytic, but the goal would be to have such combined treatment be the exception rather than the rule.235

  • Antidepressants Plus Antipsychotics

There are also patients who have both depressive and psychotic symptoms (eg, mood congruent hallucinations or delusions). Such a presentation may lead to a false positive diagnosis of schizophrenia which, in turn, can result in treatment with antipsychotics alone. That may alleviate the psychotic symptoms but not address the underlying mood disorder. Conversely, there are patients with subtle forms of delusions such as a nihilistic delusion ("I have never been any good" or "I am financially ruined") which may be missed. When treated with an antidepressant alone, these patients often do not improve. A better strategy for such patients is combined treatment with an antidepressant and an antipsychotic or alternatively electroconvulsive therapy.103

  • Antidepressants Plus Mood Stabilizers

As discussed in Chapter 1, patients with bipolar (ie, manic-depressive) disorder may present for the first time clinically in the depressed phase of their illness. If there is a past personal (or family) history of unequivocal hypomania or mania, treating with an antidepressant alone may increase the likelihood of a manic episode. In such patients, the use of a mood stabilizer (eg, lithium, valproate) alone or in combination with an antidepressant is a preferable course of action. In fact, there is some evidence that lithium can be effective as monodrug therapy for mild depressive episodes.

Electroconvulsive Therapy

Electroconvulsive therapy (ECT) remains the single most effective treatment for clinical depression and remains an important treatment option for the severely depressed patient or the patient who has not benefited from antidepressant therapy.103,216 A recently developed and still investigational treatment is repetitive transcranial magnetic stimulation (rTMS) which involves the depolarization of neurons in a localized area of the brain by applying a powerful magnetic field in rapid flux.22


< 10 - Use of Antidepressants With Other Medications Table of Contents 12 - References >

Copyright and Disclaimer

©2010, Sheldon H. Preskorn, M.D.
site design by CyberKansas Technologies
Questions or Comments about the site?