Antidepressant medications have been grouped as follows: 1)
TCAs, which for the purposes of this guideline also include the
tetracyclic antidepressant medication maprotiline; 2) SSRIs, which
include fluoxetine, sertraline, paroxetine, fluvoxamine, citalopram,
and escitalopram; 3) SNRIs, which include venlafaxine, desvenlafaxine,
and duloxetine; 4) other antidepressant medications, including bupropion,
nefazodone, trazodone, and mirtazapine; and 5) MAOIs, which include
phenelzine, tranylcypromine, isocarboxazid, and the transdermal
formulation of selegiline. Although some studies have suggested
superiority of one mechanism of action over another, there are no
replicable or robust findings to establish a clinically meaningful
difference. For most patients, the effectiveness of antidepressant
medications is generally comparable between classes and within classes
of medications. Response rates in clinical trials typically range
from 50% to 75% of patients, with some evidence
suggesting greater efficacy relative to placebo in individuals with
severe depressive symptoms as compared with those with mild to moderate symptoms
(71–73). Although remission rates are less
robust and selective publication of positive studies could affect
the apparent effectiveness of treatment (74, 75),
these factors do not appear specific to particular medications or
Nevertheless, antidepressant medications do differ in their
potential to cause particular side effects such as adverse sexual
effects, sedation, or weight gain. Therefore, the initial selection
of an antidepressant medication will largely be based on the tolerability,
safety, and cost of the medication, as well as patient preference
and history of prior medication treatment (Table 3). Other factors
include the medication half-life and potential for drug interactions
related to properties such as plasma protein binding or metabolism through
the cytochrome P450 system (Tables 4 and 5). On the basis of these
considerations, the following medications are optimal for most patients:
SSRIs, SNRIs, mirtazapine, and bupropion. Table 6 provides the starting
and usual doses of medications that have been shown to be effective for
treating major depressive disorder.
Table 3. Factors to Consider in Choosing an Antidepressant
| Add to My POL
Table 3. Factors to Consider in Choosing an Antidepressant
Nature of prior response to medication
Relative efficacy and effectiveness
Safety, tolerability, and anticipated side effects
Co-occurring psychiatric or general medical conditions
Potential drug interactions
Table 4. Cytochrome P450 Enzyme Metabolism of
| Add to My POL
Table 4. Cytochrome P450 Enzyme Metabolism of
Table 5. Cytochrome P450 Enzyme Inhibition by
| Add to My POL
Table 5. Cytochrome P450 Enzyme Inhibition by
Table 6. Dosing of Medications Shown To Be Effective
in Treating Major Depressive Disordera
| Add to My POL
Table 6. Dosing of Medications Shown To Be Effective
in Treating Major Depressive Disordera
|Generic Name||Starting Dose (mg/day)b||Usual Dose
Selective serotonin reuptake inhibitorsd
Dopamine norepinephrine reuptake inhibitord
Serotonin norepinephrine reuptake inhibitorsd
Tricyclics and tetracyclics
Monoamine oxidase inhibitors (MAOIs)
MAO B selective inhibitor
A selective inhibitor
In choosing an antidepressant medication, many psychiatrists
also consider the family history of response to particular medications;
however, the impact of this factor on the likelihood of the patient's
response to these medications is unclear. Nevertheless, the medication
experiences of others close to the patient do influence the patient's
belief about particular medications and pharmacotherapy in general.
The presence of co-occurring psychiatric or general medical
conditions can be a significant factor influencing the choice of
an antidepressant medication. For example, TCAs are generally not
optimal in patients with cardiovascular conditions, cardiac conduction
defects, closed angle glaucoma, urinary retention, significant prostatic
hypertrophy, or eating disorders with significant malnutrition or
purging. In older adults and others with malnutrition, autonomic
disorders (e.g., diabetic neuropathy, Parkinson's disease),
or low blood pressure, TCAs may exacerbate hypotension and orthostasis, resulting
in syncope or falls. Selective serotonin reuptake inhibitors and SNRIs
may be inappropriate for patients who are experiencing sexual dysfunction.
Patients who are receiving tamoxifen for breast cancer or other
indications should generally be treated with an antidepressant (e.g.,
citalopram, escitalopram, venlafaxine, desvenlafaxine) that has
minimal effect on metabolism through the cytochrome P450 2D6 isoenzyme,
because reduced metabolism of tamoxifen through CYP 2D6 is likely
to be associated with lower levels of tamoxifen's active
metabolite (76–79) with the possibility of poorer patient
outcomes (80, 81).
Because of the need for dietary restrictions and the potential
for serious side effects and drug interactions, use of MAOIs is
generally limited to patients who do not respond to other treatments.
MAOIs may be particularly effective for patients with major depressive
disorder with atypical features, although many psychiatrists prefer
to prescribe SSRIs for such patients because of SSRIs' greater safety
and tolerability and more favorable adverse effect profile.
a. Efficacy of antidepressant
1. Selective serotonin reuptake inhibitors
Selective serotonin reuptake inhibitors currently available
include fluoxetine, sertraline, paroxetine, fluvoxamine, citalopram,
and escitalopram. A large body of literature supports the superiority
of SSRIs compared with placebo in the treatment of major depressive
disorder. In more than 10 systematic reviews and meta-analyses,
the effectiveness of SSRIs has been compared with that of other antidepressant
medications, mainly TCAs. The SSRIs have demonstrated comparable
efficacy to the TCAs (84–88),
even when anxiety symptoms are considered (85, 87–90).
Although a few analyses suggest small advantages of SNRIs
over SSRIs in rates of remission (91), a preponderance
of the data finds no significant evidence of the superiority of any
other class or agents over SSRIs (84, 89, 90, 92–95).
One meta-analysis suggests a slight superiority of escitalopram compared
with other SSRIs and venlafaxine (93), and another
found significantly greater efficacy for escitalopram, sertraline,
venlafaxine, and mirtazapine as compared with duloxetine, fluoxetine,
fluvoxamine, and paroxetine (96), but other studies
show no differences in efficacy among individual SSRIs (84, 93–95, 97, 98).
2. Serotonin norepinephrine reuptake inhibitors
The serotonin norepinephrine reuptake inhibitors currently available
are venlafaxine, desvenlafaxine (the principal metabolite of venlafaxine),
and duloxetine. An immediate-release form of venlafaxine is available,
but most clinicians prefer the extended-release formulation because
it is approved for once-daily dosing and may be less often associated
with reported withdrawal symptoms.
Each of these medications is efficacious (i.e., superior to placebo
in controlled studies and meta-analyses) (95, 99), and
venlafaxine (75–150 mg/day) and duloxetine (60 mg/day)
showed comparable efficacy in a pair of trials (100). For
venlafaxine and perhaps desvenlafaxine, clinically significant norepinephrine
reuptake inhibition may not be achieved for the average patient
at lower therapeutic doses, although desvenlafaxine has a much greater
bioavailability, resulting in a lower effective dose. In individual
studies, venlafaxine and duloxetine are generally as effective as
SSRIs (see the meta-analyses of Nemeroff et al.  and
Thase et al.  for tabulated
summaries of individual study results from the more than 40 relevant
randomized controlled trials). Results of comparative studies of
desvenlafaxine are not known at this time. Relative to SSRIs, some
analyses of pooled data sets have suggested a small advantage for
SNRIs (91), which might afford clinically modest benefits
for patients with more severe depression (102) or for
patients who have not responded to prior trials of SSRIs (103).
However, other meta-analyses have shown equivalent efficacy for
SSRIs and SNRIs (95), whereas some have shown superiority of
individual medications but no clearcut medication class effects
(96). Relative to TCAs, venlafaxine's efficacy
is comparable (91, 104, 105), whereas
the more recently introduced duloxetine and desvenlafaxine have
not been systematically compared with TCAs.
3. Other antidepressant medications
A number of other antidepressant medications differ structurally or
in their pharmacological action from medications in the categories
just described and are included here.
Although bupropion is classified as a norepinephrine and dopamine
reuptake inhibitor, the latter effect is relatively weak, and its
mechanism of action remains unclear (106). There are
three formulations of bupropion: immediate release, sustained release,
and extended release. Bupropion is distinct from most antidepressants
in not having an indication for the treatment of any primary anxiety
disorder, and it may be less well tolerated than other antidepressants
among patients with significant anxiety. In addition, a meta-analysis
showed that SSRIs were modestly superior to bupropion for a subset
of patients with major depressive disorder and anxiety (107).
For individuals with low to moderate levels of anxiety, the same
meta-analysis showed that the efficacy of bupropion in treating
major depressive disorder was roughly comparable to that of the
SSRIs (107). Results of another meta-analysis suggested
that bupropion may be more likely to improve symptoms of fatigue
and sleepiness than at least some of the SSRIs (108).
Bupropion may be a good choice for patients who have a goal of quitting
smoking as it has U.S. Food and Drug Administration (FDA) approval for
this indication, reduces desire for nicotine, and doubles rates
of smoking cessation (109, 110). Patients
typically experience minimal weight gain or even weight loss on
bupropion (111), and for this reason it may be an appropriate
antidepressant for patients who are overweight or obese.
Mirtazapine is thought to work through noradrenergic and serotonergic
mechanisms, although this tetracyclic compound is not a reuptake
inhibitor (112). Mirtazapine has comparable efficacy
to SSRIs (113).
Trazodone is the oldest medication from this group that is still
in wide use. Although trazodone is an effective antidepressant,
relative to placebo (105, 114, 115),
in contemporary practice it is much more likely to be used in lower
doses as a sedative-hypnotic than as an antidepressant. Despite widespread
use of trazodone as a hypnotic, few data support its use for this
Nefazodone has an analogous structure to trazodone but somewhat
different pharmacological properties. In comparative trials versus
SSRIs, nefazodone showed comparable efficacy and overall tolerability
4. Tricyclic antidepressants
Tricyclic antidepressants (amitriptyline, nortriptyline, protriptyline,
imipramine, desipramine, doxepin, and trimipramine) are effective
treatments for major depressive disorder and have comparable efficacy
to other classes of antidepressants, including SSRIs, SNRIs, and
MAOIs (85, 105). The efficacy of subclasses
of tricyclics (e.g., secondary amines or tertiary amines) appears
to be comparable. TCAs may be particularly effective in certain
populations, such as in hospitalized patients (117, 118).
Conventional wisdom is that this advantage is explained by the superiority
of TCAs (versus SSRIs) among the subset of patients with melancholia or
more severe depression, because such a specific advantage has not
been consistently documented in studies of less severely ill outpatients
(85, 105, 118).
5. Monoamine oxidase inhibitors
MAOIs currently used as antidepressants include phenelzine,
tranylcypromine, isocarboxazid, moclobemide, and the transdermally
delivered formulation of selegiline.
MAOIs have comparable efficacy
to other antidepressants for outpatients with major depressive disorder
and may be appropriate for patients with major depressive disorder who
have not responded to safer and more easily used treatments (119, 120).
In fact, the role of MAOIs in major depressive disorder is now almost
exclusively reserved for patients who have not responded to at least
several other pharmacotherapies. Studies have demonstrated the effectiveness
of MAOIs in patients who have not responded to other antidepressant
medications, particularly TCAs (119). However, the effectiveness
of MAOIs relative to other strategies for treatment-resistant patients
in contemporary practice remains unclear, particularly for patients
who have not responded to multiple sequential trials with SSRIs
and SNRIs (121).
MAOIs have been shown to be particularly effective in treating
depressed patients with atypical features, so psychiatrists should
consider using these medications for patients with symptoms such
as reactive moods, reverse neurovegetative symptoms, and sensitivity
to rejection (119, 120, 122). There
do not appear to be any significant differences in efficacy among
the older MAOIs (119), although there are important individual
differences in responsiveness, and these medications are not interchangeable.
There are no comparative studies of the newer transdermal (skin
patch) formulation of selegiline; its efficacy has only been established
relative to placebo (123–125),
and clinical experience is limited.
b. Side effects of
The severity of side effects from antidepressant medications in
clinical trials has been assessed both through the frequency of
reported side effects and through the frequency of treatment dropout.
The likelihood of different side effects varies among classes of
antidepressant medications, among subclasses, and among individual
agents. In addition, most newer antidepressants are better tolerated
than TCAs (84–88, 97, 117, 126)
and safer in overdose (127, 128).
When side effects occur during treatment with an antidepressant,
an initial strategy is to lower the dose of the antidepressant or
to change to an antidepressant that is not associated with that
side effect. As an example, bupropion can be used if patients encounter
sexual side effects with an SSRI medication. When lowering the dose
or discontinuing the medication is not effective, additional strategies may
be considered. These additional strategies are described in Table
7, which also lists prominent and clinically relevant side effects
associated with particular medication classes.
Table 7. Potential Treatments for Side Effects
of Antidepressant Medications
| Add to My POL
Table 7. Potential Treatments for Side Effects
of Antidepressant Medications
Associated With Effect||Treatmenta|
Avoid in patients with cardiac instability or ischemia. Attend
to interactions with anti-arrhythmics.
Monitor blood pressure. Keep dose as low as possible.
Seek emergency treatment. If hypertension is severe, intravenous
antihypertensive agents (e.g., labetalol, sodium nitroprusside)
may be required.
Increase in cholesterol
Add a statin.
TCAs, trazodone, nefazodone, MAOIs
Add fludrocortisone. Add salt to diet.
Encourage adequate hydration. Add bulk laxative.
Evaluate for other possible contributors to delirium.
TCAs, SNRIs, bupropion
Suggest use of sugarless gum or candy.
Add pilocarpine eye drops.
SSRIs, SNRIs, bupropion
Assess for other etiologies (e.g., caffeinism, bruxism, migraine,
Bupropion, TCAs, amoxapine
Assess for other etiologies, and add anticonvulsant medication,
if clinically indicated.
Arousal, erectile dysfunction
TCAs, SSRIs, SNRIs
Add sildenafil, tadalafil, buspirone, or bupropion.
TCAs, SSRIs, venlafaxine, desvenlafaxine, MAOIs
Add sildenafil, tadalafil, buspirone, or bupropion.
Obtain emergency urological evaluation.
SSRIs, SNRIs, bupropion
Administer in the morning.
Add a beta-blocker or benzodiazepine.
Obtain dental consultation, if clinically indicated
TCAs, some SSRIs, SNRIs
Add an 1-adrenergic
antagonist (e.g., terazosin), central 2-adrenergic
agonist (e.g., clonidine), or anticholinergic agent (e.g., benztropine).
Monitor blood pressure for evidence of hypotension
or orthostasis; assess for sedation, blurred vision, or confusion;
modify environment to reduce risk.
Gastrointestinal (GI) bleeding
Identify whether concomitant medications may affect clotting.
Provide education about and monitor for clinical evidence
of hepatic dysfunction. Obtain hepatic function tests, if clinically
SSRIs, SNRIs, bupropion
Use morning dosing. Add a sedative-hypnotic at bedtime. Add
melatonin. Provide CBT or education in sleep hygiene.
SSRIs, SNRIs, bupropion
Administer after food or in divided doses.
If clinically indicated, obtain bone density monitoring
and add specific treatment to reduce bone loss (e.g., calcium and
vitamin D supplements, bisphosphonates, selective estrogen receptor
TCAs, trazodone, nefazodone, mirtazapine
Use bedtime dosing. Add modafinil or methylphenidate.
Severe serotonin syndrome
Obtain emergency evaluation. Consider admission to
a critical care unit.
SSRIs, mirtazapine, TCAs, MAOIs
Encourage exercise. Obtain input from dietician. If
changing antidepressants, consider a secondary amine (if a TCA is required)
or other antidepressant with fewer weight issues (e.g., bupropion).
Serotonin syndrome, as the
name implies, is presumed to result from high levels of serotonin
in the brain. Although it can occur with administration of one or
more serotonergic medications, it is most severe when an MAOI is
coadministered with another serotonergic medication. Consequently,
great care must be taken when changing patients from another antidepressant
medication to an MAOI and from an MAOI to other antidepressant medications
because of the persistent effects of discontinued medications. A
washout period is essential before and after using an MAOI. If the
psychiatrist chooses to discontinue a monoamine-uptake-blocking
antidepressant medication (e.g., SSRI, SNRI, TCA) and substitute an
MAOI, toxic interactions can best be avoided by allowing at least
a 2-week washout period between medication trials (Table 8). The
long half-life of the SSRI fluoxetine and its metabolites necessitates
a 5- to 6-week washout period or longer before the use of an MAOI.
Additional information about serotonin syndrome with specific medication
classes can be found in Sections II.B.2.b.1.g. and II.B.2.b.5.b.
Table 8. Required Washout Times Between Antidepressant
| Add to My POL
Table 8. Required Washout Times Between Antidepressant
Drug with long-half-life metabolites (e.g., fluoxetine)
Drug without long-half-life metabolites (e.g., TCAs,
paroxetine, fluvoxamine, venlafaxine)
Because knowledge of potential drug-drug interactions is frequently
changing, it is useful to consult a frequently updated drug information
database before selecting an antidepressant in a patient taking
1. Selective serotonin reuptake inhibitors
SSRIs have comparable tolerability overall, but the specific medications
differ somewhat in their side effect profiles, which may guide selection
of an agent for an individual patient. Pharmacokinetic issues, including
half-life and effect on the CYP-450 enzyme system, are additional
considerations in the choice of an SSRI.
SSRIs commonly cause nausea, vomiting, and diarrhea (98). These
adverse events are generally dose dependent and tend to dissipate
over the first few weeks of treatment. In some patients, however,
SSRIs sometimes precipitate or exacerbate restlessness, agitation,
and sleep disturbancesside effects that often attenuate with
time. Anxiety may be minimized by introducing the agent at a low
dose. Akathisia has also been reported in patients taking SSRIs
(129) and may contribute to reported restlessness or
activation. If akathisia does occur, a beta-blocker or benzodiazepine
can be tried to reduce symptoms. Insomnia can be treated by using
sleep hygiene techniques or CBT as a first approach or by adding
a sedative-hypnotic medication or trazodone. Some have found melatonin
to be helpful in treating SSRI-induced insomnia.
Although loss of erectile or ejaculatory function in men and loss
of libido and anorgasmia in both sexes may be complications of virtually
any antidepressant medication, these side effects appear to be more
common with SSRIs. The psychiatrist should ascertain whether the
reported sexual dysfunction is a result of the antidepressant medication,
the underlying major depressive disorder, a co-occurring medical
disorder, a disturbance in a relationship, or a need for education
about sexual functioning. If sexual dysfunction is determined to
be a side effect of the antidepressant medication, a number of strategies
are available, including continuing treatment to assess whether
the dysfunction will disappear with time, lowering the dose, discontinuing
the antidepressant, or substituting another antidepressant such
as bupropion (130). Specific pharmacological treatments
that can be added for arousal difficulties, erectile dysfunction,
or orgasm dysfunction include buspirone (131), bupropion
(132), sildenafil (133), and tadalafil
(134). Other phosphodiesterase inhibitors may be also
useful in treating sexual side effects, and a variety of other medications
have been used with anecdotal success (135, 136).
Selective serotonin reuptake inhibitors can initially exacerbate
both migraine headaches and tension headaches. These effects tend
to be transient and improve within the first few weeks of treatment.
With continued treatment, SSRIs may actually help prevent and treat
migraine headaches (137, 138). Selective
serotonin reuptake inhibitors have also been associated with extrapyramidal
side effects, including akathisia, dystonia, parkinsonism, and tardive
dyskinesia (139, 140). The incidence of
such side effects is very low with SSRIs but may be higher in older
patients, especially those with Parkinson's disease.
Selective serotonin reuptake
inhibitors, like other antidepressive agents, have been associated
with an increased risk of falls. In studies of nursing home residents,
SSRI use has been associated with an approximately twofold increase in
the risk of a fall (141, 142). An even
greater risk of falls in patients who were taking SSRIs was noted
in a community-based cohort study (143). Meta-analyses
have also documented an increased risk of falls in patients treated
with antidepressive agents, in general (144, 145).
The implications of this increase in fall risk are complicated by
the decrease in bone density that has been noted in depressed patients
(146) and in patients treated with SSRIs (147, 148).
An increase in the risk of hip fractures has also been noted (149).
Rarely, SSRI use has been associated with bradycardia, which could also
contribute to syncope and falls (150). In all patients,
including those treated with SSRIs, fall risk may be increased in
individuals receiving benzodiazepines or other hypnotic agents (144, 145, 151)
and in those receiving multiple medications (144, 145).
Systematically reviewing patients' medication regimens
may help to eliminate medications that may no longer be needed,
although such interventions have not been found to alter fall risk,
per se (152). Inquiring about a history of falls in
the past year and assessing for abnormalities in gait and balance
can also help in identifying patients at particular risk for falling
Weight gain, at times substantial, occurs in some patients taking
SSRIs (154). Patients who take paroxetine have a higher
incidence of weight gain than those who take other SSRIs (98, 155).
Fluoxetine causes an initial reduction in weight, which tends to
normalize with continued treatment (156).
Use of SSRIs has been associated with the rare development of
a syndrome caused by an excess of central nervous system serotonergic
activity. Features of serotonin syndrome include abdominal pain,
diarrhea, flushing, sweating, hyperthermia, lethargy, mental status
changes, tremor and myoclonus, rhabdomyolysis, renal failure, cardiovascular shock,
and possibly death (157). Although serotonin syndrome
can occur rarely with the use of SSRIs alone, it is usually associated
with the simultaneous use of multiple serotonergic agents and is
most severe when SSRIs are given together with MAOIs. Consequently,
when an SSRI is being changed to an MAOI or vice versa, particular
attention must be give to the duration of time between treatments (Section
II.B.2.b) to avoid precipitating a potentially lethal serotonin
syndrome. Serotonin syndrome has also been reported when SSRIs are used
in combination with tramadol, high-dose triptans, or the antibiotic
linezolid, which also has some ability to inhibit MAO (158, 159).
The potential for drug-drug interactions differs significantly across
the SSRIs. Selective serotonin reuptake inhibitors have variable
effects on hepatic microsomal enzymes and therefore cause both increases
and decreases in the blood levels of other medications. For example,
when SSRIs that strongly inhibit the CYP 2D6 isoenzyme (e.g., paroxetine, fluoxetine)
are administered concomitantly with tamoxifen, the metabolism of
tamoxifen to its active metabolite is reduced (76–79),
resulting in a potential decrease in its efficacy in preventing
breast cancer relapse (80, 81). Interaction
with other drugs was higher for fluoxetine, fluvoxamine, and paroxetine
than for sertraline, citalopram, and escitalopram (98, 160, 161).
As described above, there can be a potentially lethal interaction
between SSRIs and MAOIs: the serotonin syndrome. At least five half-lives
should elapse between the time an SSRI is stopped and an MAOI is
started; for fluoxetine discontinuation, this means waiting approximately
5–6 weeks before starting an MAOI; for discontinuation
of other SSRIs, approximately 2 weeks should pass before starting
an MAOI (162). A 2-week waiting period has been suggested
after discontinuing an MAOI before starting an SSRI or another MAOI
i. Discontinuation syndrome
Selective serotonin reuptake
inhibitors generally should not be abruptly discontinued after extended
therapy and, whenever possible, should be tapered over several weeks
to minimize discontinuation-emergent symptoms. Clinical experience
and a few controlled studies suggest that among the SSRIs, discontinuation-emergent
symptoms are more likely with paroxetine than sertraline, citalopram,
or escitalopram and least likely to occur with fluoxetine (due to
the long elimination half-life of its primary metabolite, norfluoxetine) (163, 164).
Discontinuation-emergent symptoms include both flu-like experiences such
as nausea, headache, light-headedness, chills, and body aches, and
neurological symptoms such as paresthesias, insomnia, and "electric
shock-like" phenomena. These symptoms typically resolve
without specific treatment over 1–2 weeks. However, some
patients do experience more protracted discontinuation syndromes,
particularly those treated with paroxetine, and may require a slower downward
titration regimen. Another strategy is to change to a brief course
of fluoxetine, e.g., 10 mg for 1–2 weeks, and then taper
and discontinue the fluoxetine (165).
2. Serotonin norepinephrine reuptake inhibitors
The most common side effects of the SNRIs (venlafaxine, desvenlafaxine,
and duloxetine) are similar to those seen with SSRIs, including
nausea and vomiting, sexual dysfunction, and activation; like the
side effects seen with SSRIs, those with SNRIs can attenuate with
continued use. The SNRIs also are more likely to be associated with
side effects that reflect noradrenergic activity, including increased
pulse rate, dilated pupils, dry mouth, excessive sweating, and constipation.
Although all three SNRIs carry the warning of increased blood pressure,
this risk is greater during therapy with venlafaxine at doses above
150 mg/day (166) than with duloxetine at doses
of 60–120 mg/day (167) or desvenlafaxine
at doses of 50–100 mg/day (168). Because
this blood pressure increase is dose-related, SNRI-induced hypertension
may respond to dose reduction. In the absence of a reduction in
hypertension, a different antidepressant medication may be considered.
Alternatively, in a patient with well-controlled depressive symptoms,
it may be preferable to add an antihypertensive agent rather than
risk a depressive relapse or recurrence with medication tapering.
As with the SSRIs, abrupt discontinuation of SNRIs should be avoided
whenever possible. Discontinuation symptoms, which are sometimes
protracted, are more likely to occur with venlafaxine (and, by implication, desvenlafaxine)
than duloxetine (100) and may necessitate a slower
downward titration regimen or change to fluoxetine. As described
above (Section II.B.2.b), there can be a potentially lethal interaction
between SNRIs and MAOIs: the serotonin syndrome.
3. Other antidepressant medications
Bupropion differs from other modern antidepressants by its lack
of direct effects on serotonergic neurotransmission and, as a consequence,
a virtual lack of sexual side effects (169). Neurologic
side effects with bupropion include headaches, tremors, and seizures
(106). The risk of seizures is minimized by avoiding
high doses (e.g., using no more than 450 mg/day), avoiding
rapid titration, using divided dosing schedules for the immediate-release
and sustained-release formulations, and avoiding use of bupropion in
patients with risk factors for seizures. Bupropion should also not
be used in patients who have had anorexia nervosa or bulimia nervosa because
of elevated risk of seizures (170). The risk of seizures
may also be increased by the concomitant use of inhibitors of CYP
2B6 (e.g., desipramine, sertraline, paroxetine, fluoxetine) due to
the resulting increase in bupropion blood levels. Bupropion has
been associated with a low risk of psychotic symptoms, including delusions
and hallucinations. It should therefore be used cautiously in patients
with psychotic disorders. Other side effects with bupropion include agitation,
jitteriness, mild cognitive dysfunction, insomnia, and gastrointestinal
The most common side effects of mirtazapine include dry mouth,
sedation, and weight gain. For this reason, mirtazapine is often
given at night and may be chosen for depressed patients with initial
insomnia and weight loss. Although these side effects tend to occur
early in the treatment course and may attenuate with continued use, the
weight gain associated with mirtazapine is greater than that with
other non-TCA, non-MAOI antidepressants (95) and may
make it a less attractive choice for some patients. Mirtazapine
increases serum cholesterol levels in some patients (171).
Although several patients treated with mirtazapine were observed
to have agranulocytosis in early studies, subsequent clinical experience
has not confirmed an elevated risk (172).
The most common side effect with trazodone is sedation. Because
the sedation associated with trazodone is greater than that with
other non-TCA, non-MAOI antidepressants (95), this
can be an advantage in patients with initial insomnia (173).
Trazodone can also cause cardiovascular side effects, including
orthostasis, particularly among elderly patients or those with preexisting
heart disease. Use of trazodone has also been associated with life-threatening ventricular
arrhythmias in several case reports (173). Trazodone
also can cause sexual side effects, including erectile dysfunction
in men; in rare instances, priapism occurs, which might require surgical
correction (174, 175).
Side effects with nefazodone include dry mouth, nausea, constipation,
orthostasis, and visual alterations (176). Sedation
is also common and may necessitate a gradual titration of nefazodone.
However, in patients with insomnia, the sedating properties of nefazodone
can be helpful in improving sleep (177). There appears
to be a low incidence of treatment-emergent sexual dysfunction (178, 179)
with nefazodone and, unlike trazodone, it has not been associated
with priapism. Nefazodone has also been associated with rare but potentially
fatal liver failure (180, 181), which
has limited its use in recent years. Drug-drug interactions can
also be problematic as nefazodone inhibits hepatic microsomal enzymes and
can raise levels of concurrently administered medications such as
certain antihistamines, benzodiazepines, and digoxin.
4. Tricyclic antidepressants
a. Cardiovascular effects
Cardiovascular effects, including arrhythmias, can be problematic
with TCA treatment. Consequently, a pretreatment ECG is indicated
for patients with significant cardiac risk factors and patients
older than age 50 years. Follow-up ECGs may also be indicated to
identify the development of conduction changes, typically during
the early phase of TCA use (182). Tricyclic antidepressants
act similarly to class Ia antiarrhythmic agents such as quinidine,
disopyramide, and procainamide, which increase the threshold for
excitation by depressing fast sodium channels, prolong cardiac cell
action potentials through actions on potassium channels, and prolong
cardiac refractoriness through actions on both types of channels
(183). As a result, combinations of TCAs with other class
I antiarrhythmic agents can exert additive toxic effects on cardiac
conduction; patients with ventricular arrhythmias taking another
class I antiarrhythmic agent who require TCA therapy should be under
careful medical supervision. Individuals with prolonged QT intervals,
whether preexistent or medication induced, are predisposed to develop
ventricular tachycardia (184). Even patients with normal
pretreatment ECG results may develop atrioventricular block with
TCAs that reverts to normal after discontinuation of antidepressant medication
treatment (185). Because of these effects on cardiac
conduction, TCAs (like other class Ia antiarrhythmic agents) may
carry an increased risk of serious cardiac adverse effects, including
mortality (186–189). In addition, fatal arrhythmias
can occur in the context of TCA overdose (190, 191).
In addition to causing arrhythmias,
TCAs can cause a number of other cardiovascular side effects, including
tachycardia (through muscarinic cholinergic blockade and -adrenergic blockade)
or orthostatic hypotension (through -adrenergic blockade).
Side effects such as orthostatic hypotension may in turn lead to
events such as dizziness, falls, or fractures, which are of particular
concern in elderly patients (192). Of the TCAs, nortriptyline
may be less likely to contribute to orthostatic blood pressure changes (185). Preexisting
orthostasis, antihypertensive treatment, dehydration, and salt depletion, whether
voluntary or a result of diuretic treatment, may contribute to symptomatic
orthostatic hypotension with TCAs. If there is no medical contraindication,
patients with symptomatic orthostatic hypotension should maintain
adequate fluid intake and be cautioned against extreme dietary salt
b. Anticholinergic side effects
All TCAs have antimuscarinic effects; tertiary amine tricyclic antidepressants
produce the most anticholinergic side effects, whereas the secondary
amines desipramine and nortriptyline have less antimuscarinic activity
(193). The most common consequences of muscarinic blockade
are dry mouth, impaired ability to focus vision at close range,
constipation, urinary hesitation, tachycardia, and sexual dysfunction. Although
patients can develop some degree of tolerance to anticholinergic
side effects, these symptoms may require treatment if they cause
substantial dysfunction or interfere with adherence. Impaired visual
accommodation may be counteracted through the use of pilocarpine
eye drops. Dry mouth may be counteracted by advising the patient
to use sugarless gum or candy and ensuring adequate hydration. Constipation
can be managed by adequate hydration and the use of bulk laxatives.
Antidepressant medications with anticholinergic side effects should
be avoided in patients with cognitive impairment, narrow-angle glaucoma,
or prostatic hypertrophy. Tricyclic antidepressants can impair memory
and concentration and even precipitate anticholinergic delirium, particularly
in patients who are elderly, medically compromised, or taking other
anticholinergic medicines. Such toxic confusional states may signal
the presence of high TCA blood levels and can improve with lowering
of the dose (194).
Tricyclic antidepressants also have affinity for histaminergic receptors
and produce varying degrees of sedation. In general, tertiary amines
cause greater sedation, while secondary amines cause less (193).
Sedation often attenuates in the first weeks of treatment, and patients
experiencing only minor difficulty from this side effect should
be encouraged to allow some time to pass before changing antidepressant
medications. Patients with major depressive disorder with insomnia may
benefit from sedation when their medication is given as a single
dose before bedtime.
Tricyclic antidepressants can cause weight gain, possibly through
their histaminergic properties and/or blockade of 5-HT2 receptors
(195). The degree of weight gain appears to vary by
agent (e.g., greater weight gain with amitriptyline and less with
desipramine), is often dose dependent, and is potentially reversible
with cessation of TCA therapy. Regular monitoring of weight permits
early detection of weight gain and can allow the treating clinician
and patient to determine whether a management plan to minimize or
forestall further weight gain is clinically indicated.
Tricyclic antidepressants can cause myoclonus (196).
Since this may be a sign of toxicity, the clinician may wish to
check the blood level (if available) to ensure that it is not excessive. If
the level is nontoxic and myoclonus is not bothersome to the patient,
the agent may be continued without a change in dose. If the myoclonus
is problematic and the blood level is within the recommended range,
the patient may be treated with clonazepam at a dose of 0.25 mg
t.i.d. Alternatively, the antidepressant medication may be changed.
In overdoses, TCAs can cause seizures. Some vulnerable patients
may experience seizures even on therapeutic doses of a TCAespecially
clomipramine and maprotiline (197). Amoxapine, a dibenzoxazepine-derivative
tricyclic antidepressant, also produces seizures in overdose and
has active metabolites that block dopamine receptors, conferring
a risk of extrapyramidal side effects and tardive dyskinesia (198).
Use of TCAs has been associated with an increased risk of
falls in a number of studies and meta-analyses, and the relative
risk of falling appears comparable to that with SSRI treatment (141, 144, 145).
Although systematic reviews show a relatively minor effect of orthostatic
hypotension on fall risk, TCAs may contribute to orthostasis and
falls in individual patients (153). If orthostatic
hypotension is prominent or associated with gait or balance problems,
it may require further evaluation and treatment to minimize the
likelihood of falls (199). Other aspects of fall risk
with TCAs are similar to those that have already been described for
patients treated with SSRIs (Section II.B.2.b.1.e). Other causes
of falls include bradycardia, cardiac arrhythmia, a seizure, or
g. Medication interactions
A number of medications that inhibit, induce, or are metabolized
by hepatic microsomal enzymes can interact with TCAs (200).
For example, medications that induce CYP 3A4 such as carbamazepine
or barbiturates will cause a decrease in serum levels of TCAs. Drugs
such as the antipsychotic medication perphenazine or SSRIs such
as fluoxetine or paroxetine can inhibit metabolism via CYP 2D6,
resulting in a reduced clearance and increased levels of TCAs. Tricyclic antidepressants
can also alter the pharmacokinetics or pharmacodynamics of other
medications; for example, TCAs can cause a lowering of valproate
levels and reduce the activity of clonidine. Therefore, adjustments
in medication doses may be necessary when TCAs are administered
concomitantly with other drugs for which there is an interaction.
The ability to obtain meaningful antidepressant blood levels to
guide dosing is an advantage with several of the TCAs (e.g., nortriptyline, amitriptyline,
desipramine, imipramine) (201). Potentially dangerous
interactions, including hypertensive crises and serotonin syndrome,
can develop when TCAs are administered with MAOIs (see Sections
II.B.2.b and II.B.2.b.5.b), norepinephrine, or epinephrine.
5. Monoamine oxidase inhibitors
A hypertensive crisis can occur when a patient taking an MAOI
ingests large amounts of tyramine or other vasoactive amines in
foods or medications (202). This reaction is characterized
by the acute onset of severe headache, nausea, neck stiffness, palpitations,
profuse perspiration, and confusion and can possibly lead to stroke
and death (119). Dietary restrictions include avoiding
foods such as aged cheeses or meats, fermented products, yeast extracts,
fava or broad beans, red wine, draft beers, and overripe or spoiled
foods (202, 203). A number of medications
including norepinephrine reuptake blocking drugs (e.g., SNRIs, TCAs),
sympathomimetic vasoconstrictive agents, and over-the-counter decongestants
can also produce a hypertensive crisis when used in combination
with MAOIs (202, 204). Individuals with
asthma who receive MAOIs should be cautioned regarding interactions
with sympathomimetic bronchodilators, although other antiasthma
agents appear to be safe. Stimulants may be added to MAOIs, but only
with caution and in selected individuals with treatment-resistant
symptoms (205, 206).
At low doses (6 mg/24 hours), selegiline differs
from the older MAOIs in selectively blocking MAO B. In addition, the
transdermal delivery of selegiline bypasses enzyme inhibition in
the gut and first-pass metabolism in the liver. As a result, a low-tyramine
diet is not needed when selegiline is prescribed at the minimum
therapeutic dose. However, few safety data are available at higher
doses at which selegiline becomes nonselective and inhibits both MAO
A and MAO B. Consequently, a low-tyramine diet is needed when doses
of 9 mg/24 hours and higher are prescribed as with other
MAOIs (207, 208). Moclobemide, which is
available in Canada but not the United States, differs from the
above MAOIs in binding reversibly to MAO and makes dietary restrictions
unnecessary with moclobemide. The potential for drug-drug interactions
with selegiline and moclobemide has not been fully studied, but
caution suggests that the same drug interactions should be considered
as when prescribing the older, nonselective, irreversible MAOIs.
Although some clinicians continue to recommend that patients
carry nifedipine as a self-administered antidote (e.g., 10 mg by
mouth at the first sign of a hypertensive crisis ),
this practice has not been approved by the FDA, and there are concerns
about both the safety and efficacy of this strategy, which can produce
dangerous hypotension (210). Definitive treatment of
hypertensive crises usually involves intravenous administration
of an antihypertensive agent (e.g., labetalol, sodium nitroprusside)
in an emergency department setting.
As discussed previously in Section II.B.2.b.1.g, serotonin
syndrome is caused by excess CNS serotonergic activity and is characterized
by abdominal pain, diarrhea, flushing, sweating, hyperthermia, lethargy,
mental status changes, tremor and myoclonus, rhabdomyolysis, renal
failure, cardiovascular shock, and possibly death. Serotonin syndrome
most commonly occurs when MAOIs (including reversible inhibitors
of monoamine oxidase and selegiline) are taken in close proximity
to other serotonergic agents, such as buspirone or antidepressants
(157, 204, 211). Consequently,
when patients are being changed from an SSRI other than fluoxetine
or an SNRI to an MAOI, a waiting period of at least 2 weeks is needed
between the discontinuation of one medication and the initiation of
the other. When changing from fluoxetine to an MAOI, a waiting period
of at least 5 weeks is needed before the MAOI is started (Table
8). Other medications that have been reported to produce serotonin
syndrome when used in conjunction with MAOIs include synthetic opioids (e.g.,
dextromethorphan, meperidine, tramadol, propoxyphene, methadone),
nonantidepressant tricyclic compounds (e.g., carbamazepine, cyclobenzaprine),
sibutramine, and over-the-counter cold products such as chlorpheniramine
c. Cardiovascular effects
Orthostatic hypotension is commonly seen during MAOI treatment.
Possible treatments for this side effect include adding dietary
salt to increase intravascular volume, or use of the mineralocorticoid
fludrocortisone. Use of MAOIs can also be associated with the development
of peripheral edema, which may be helped by the use of support stockings.
Weight gain is also commonly seen in patients treated with nonselective
MAOIs. Although clinical experience is limited, results of one 52-week
study suggested that treatment with transdermal selegiline may not
be associated with an increased risk of weight gain (212).
Sexual side effects seen with MAOI therapy include anorgasmia,
decreased libido, and erectile or ejaculatory dysfunction. Sexual
side effects may diminish over time or with reductions in MAOI doses.
The transdermal formulation of selegiline appears to have a relatively
low risk of sexual side effects (213).
Treatment with MAOIs can also be accompanied by headaches
and insomnia; these side effects may diminish over time with continued
use. Other neurological effects seen with MAOIs include sedation,
myoclonic jerks, paresthesias, intense daytime drowsiness, and,
rarely, peripheral neuropathy.
Improvement with pharmacotherapy can be observed as early
as the first 1–2 weeks of treatment, and improvement continues
up to 12 weeks. Many patients may show partial improvement as early
as the end of the first week (214–216). Others
achieve improvement within the first 2–4 weeks (217–220).
In short-term efficacy trials, all antidepressant medications appear
to require at least 4–6 weeks to achieve maximum therapeutic
effects (221, 222). There is also evidence
for continued accrual of benefit for an additional 4–6
weeks (223). Furthermore, longer time to therapeutic
effect has been seen with studies conducted in "real world" settings
(224), as well as in studies of patients with more
chronic illness (225, 226) or patients
with major depressive disorder complicated with co-occurring medical and/or
Axis I disorders (224, 227).
Once an antidepressant medication has been selected, it can
be started at doses suggested in Table 6. Initial doses should be
incrementally raised as tolerated until a therapeutic dose is reached
or the patient achieves remission, provided there has been at least
some improvement in symptoms in the initial weeks of treatment (217–220).
For patients who exhibit a partial response to treatment, doses
of antidepressant medications should be maximized, side effects
permitting, before changing to a different antidepressant medication.
In some instances, due to factors such as rapid metabolism of medication
(228, 229), patients may require doses
above those noted in FDA labeling. Patients who have achieved some
improvement during the initial weeks of treatment should be encouraged
to continue taking antidepressant medication for a total of at least
4–8 weeks. If at least moderate improvement is not observed
with maximally tolerated doses after 4–8 weeks of treatment,
reappraisal and adjustment of the pharmacotherapy should be considered.
Patients with no improvement in the initial weeks of treatment generally
need an earlier adjustment of treatment. For these patients, the psychiatrist
should consider changing to another antidepressant rather than increasing
the dose of the medication. For some antidepressant medications,
the exact relationships between doses and major depressive disorder
symptom response have not been rigorously investigated with fixed-dose studies,
and minimum effective doses have not been clearly established; moreover,
for other antidepressant medications, some studies have failed to
show dose-response relationships (230, 231).
Therefore, the initial doses and usual adult doses in Table 6 are
intended to serve as general guidelines, and actual doses may vary
from individual to individual.
Titration of the dose to full therapeutic doses generally
can be accomplished over the initial week(s) of treatment but may vary
depending on the development of side effects, the patient's
age, and the presence of co-occurring medical and psychiatric conditions.
In general, patients who are older, are medically compromised, or
have decreased ability to metabolize and clear antidepressant medications
will require lower doses. In such patients, reduction of initial
and therapeutic doses to 50% of usual adult doses is often
recommended, and dose escalations should be made at a slower rate
than for younger and healthier adults. Doses will also be affected
by the side effect profile of medications and the patient's
ability to tolerate these side effects. Medication doses should
also be tailored to individual patients depending on the potential
for pharmacokinetic alterations and drug-drug interactions.
Patients who have started taking an antidepressant medication
should be carefully and systematically monitored to assess their
response to treatment, the emergence of side effects, their clinical
condition, safety, and adherence to treatment. Use of clinician-
and patient-rated scales can facilitate such assessments (see Section
II.A.8). Factors to consider when determining the frequency of treatment
visits include the severity of illness, the patient's cooperation
with treatment, the availability of social supports, the presence
of co-occurring general medical illnesses, and the progression of symptom
change. Visits should also be frequent enough to monitor and address
suicide risk and to actively promote treatment adherence, since
attrition from treatment continues to be a major hurdle in maximizing
outcomes. Patients in clinical trials appear to benefit from monitoring
once a week or more. This frequency of monitoring enhances adherence rates
and likely helps patients avoid the demoralization that may occur
before the onset of beneficial effects (216). In the recently
completed STAR*D ("Sequenced Treatment Alternatives
to Relieve Depression") trial, up to six visits were recommended
during the first 12 weeks (acute phase) of measurement-based treatment
at each of the four treatment steps (40). In clinical
practice, the frequency of monitoring during the acute phase of
pharmacotherapy may vary and can be as often as multiple times per
week in more complex circumstances. The method of monitoring (e.g.,
face-to-face visits, telephone contact, or contact with another
clinician knowledgeable about the patient) may vary depending on
the clinical context and the treatment modality.
Although for most patients, monitoring of antidepressant blood
levels is not necessary, it may be useful for those taking TCAs.
For some medications, particularly nortriptyline, amitriptyline,
desipramine, and imipramine, blood drug levels correlate with both
efficacy and side effects (201, 232, 233). When
such medications are given, obtaining blood drug levels can be particularly
informative when patients have not responded to treatment with an
adequate dose of antidepressant medication for an adequate duration;
when patients are particularly vulnerable to the toxic effects of
a medication and require the lowest possible effective dose; when
there are concerns about patient adherence; and when there is concern that
drug-drug interactions are adversely affecting antidepressant medication
levels. In time, genetic testing may help guide selection or dosing
of antidepressants, but data are currently insufficient to justify
the cost of such tests (229).
Some antidepressant medications, especially TCAs, can cause
significant morbidity and mortality in overdose (190, 191).
Ingestion of a 10-day supply of a tricyclic agent administered at
a dose of 200 mg/day is often lethal. Early on in treatment,
it is prudent to dispense only small quantities of such antidepressant
medications and keep in mind the possibility that patients can hoard
medications over time. Alternatively, in patients who are suicidal,
it may be preferable to employ agents that are safer in overdose
such as the SSRIs, bupropion, or mirtazapine.
3. Other somatic
Electroconvulsive therapy has the highest rates of response
and remission of any form of antidepressant treatment, with 70%–90% of
those treated showing improvement (234–236).
Although the remission rate with ECT appears to be lower when it is
used in community settings than when it is used in clinical trials
(237), the proportion of patients with major depressive
disorder who respond to ECT is still greater than the proportion who
respond to antidepressant medication. In addition, ECT has been
associated with significant improvements in health-related quality
of life (238). Consequently, ECT should be considered
for patients with severe major depressive disorder that is not responsive
to psychotherapeutic and/or pharmacological interventions,
particularly those with significant functional impairment who have
not responded to numerous medication trials (239).
Electroconvulsive therapy may be particularly beneficial and can
be considered as a first-line treatment option for severe major
depressive disorder when it is coupled with psychotic features (240, 241),
catatonia (239, 242), suicide risk (243),
or food refusal leading to nutritional compromise, as well as in
other situations when a particularly rapid antidepressant response
is required (240), such as with severely ill inpatients
(239). Electroconvulsive therapy is also indicated
as a first-line treatment for patients who have previously shown
a positive response to this treatment modality or who prefer it (239).
1. Side effects of electroconvulsive therapy
Electroconvulsive therapy is a very safe treatment, and there are
no absolute contraindications to its use (239). Risks
of morbidity and mortality, in general, do not exceed those associated
with anesthesia alone (239, 244, 245).
However, the presence of some medical conditions may necessitate
modifications in anesthesia or ECT administration.
Electroconvulsive therapy may have cardiovascular side effects,
mediated by changes in the autonomic nervous system with the initial
stimulus and subsequent seizure activity (239). More
specifically, ECT typically causes a transient rise in heart rate
and blood pressure, with associated increases in cardiac workload
and intracranial pressure. These effects can be managed by optimizing
blood pressure control prior to ECT and administering antihypertensive
agents (e.g., short-acting beta-blockers or calcium channel blockers)
at the time of ECT (239). Arrhythmias, which are usually
transient, can also occur in conjunction with ECT and can be managed
with usual antiarrhythmic therapies if they do not resolve spontaneously
can also be associated with cognitive effects, the most common of
which is a period of confusion following the ECT and associated
anesthesia that generally lasts between 30 and 60 minutes (246).
Electroconvulsive therapy also is associated with anterograde amnesia, which
typically resolves soon after the last ECT treatment (247).
Some degree of retrograde amnesia, particularly for recent memories,
may continue for a longer period of time after the end of the ECT
course (247) but is less pronounced for autobiographical
memories than for impersonal memories (248). These
cognitive effects of ECT are related to electrode placement, stimulus dosage,
age, and premorbid cognitive status (249–253).
Retrograde amnesia also improves over time, typically resolving
within 6 months (248, 252), although some
patients report incomplete recovery of memories, particularly for
events around the time of the treatment (247, 254). Rarely,
patients report more pervasive and persistent cognitive disruption,
the basis of which is uncertain (252, 255).
For many individuals, however, subjective memory (256)
and quality of life (238) is improved following ECT
with the resolution of the major depressive episode and its associated deficits
in memory or executive functioning (257, 258).
2. Implementation of electroconvulsive therapy
The evaluation preceding ECT consists of a psychiatric history
and examination to verify that ECT is indicated, a general medical
evaluation (including medical history and physical examination with
cognitive assessment, vital signs, and any specifically indicated
laboratory, radiologic, or imaging studies) to define factors that
may influence the risk of ECT, and an anesthesia evaluation to identify
and address the nature and extent of anesthetic risk and the need
for modification of medications or anesthetic technique (239).
This evaluation should include a summary of treatment indications,
treatment risks, and a suggestion of any indicated additional evaluative
procedures, alterations in treatment, or modifications in ECT technique
(239). In assessing indications for caution (e.g.,
recent myocardial infarction, cardiac arrhythmias, intracranial
space-occupying lesions), the relative risks and benefits should
be carefully weighed in collaboration with an anesthesiologist,
a general medical physician, and other specialists, as necessary.
Once completed, the pre-ECT evaluation will serve as the basis for
a specific, individualized discussion of the risks and benefits of
ECT relative to other therapeutic options as part of the informed
consent process. With the patient's permission, it is helpful
to educate the patient's family about ECT and involve them
in discussions relating to consent.
An additional aspect of decision-making prior to ECT relates
to the use of psychotropic medications during the ECT course. There
is growing use of ECT combined with antidepressant medication. Although
data supporting this practice are still few, it does not appear
to increase side effects and may augment response (259, 260).
An additional goal of combination treatment is to minimize the risk
of relapse between the end of the ECT course and the attainment of
full antidepressant effectiveness. Antipsychotic medications are typically
continued during the ECT course (239, 261),
although most data on this practice come from studies of patients
with schizophrenia who are receiving ECT. The safety of combining
lithium and ECT has been questioned, although there are conflicting
data (239, 262). Medications that have
anticonvulsant properties are often discontinued or given at decreased
doses during the ECT course to minimize effects on seizure induction
(239, 261). With benzodiazepines, there
is some evidence that concurrent use may diminish ECT effectiveness,
particularly when right unilateral electrode placement is used (263).
Electroconvulsive therapy may be administered either unilaterally
or bilaterally (using a bitemporal or bifrontal electrode placement).
Compared with patients who receive bilateral treatment, most patients
who receive right unilateral electrode placement with low stimulus
intensities experience fewer cognitive effects but less therapeutic
benefit (253). Stimuli of higher intensity (i.e., 500% above
seizure threshold) are associated with antidepressant effects more
comparable to those seen with bilateral electrode placements, although
such stimulus intensities are not always achievable with existing
ECT devices (264). Regardless of what electrode placement
is chosen, stimulus dosing should be individualized and stimulus
parameters adjusted to induce an adequate generalized seizure, which
is typically at least 20 seconds or greater in motor duration and
30 seconds in EEG duration (239). Failure to induce
an adequate seizure should be followed immediately by restimulation
at higher energies until an adequate seizure is elicited.
is typically administered 2–3 times/week; less
frequent administration has been associated with less cognitive
impairment but also a longer lag in the onset of action (265).
In clinical practice, the need for ECT to be administered at this
frequency could produce logistical barriers for some patients who
would either require hospitalization or transportation after ECT
sessions. The acute course of ECT treatment typically consists of
six to 12 treatments and generally does not exceed 20 treatments
(239, 266). It is important that treatment
continue until symptoms have remitted or clearly reached a plateau,
since relapse rates appear to be greater and overall prognosis worse
if ECT is discontinued prematurely (237). Use of a
formal rating scale may be helpful in assessing symptom response
as well as the cognitive side effects of treatment, permitting adjustments
in the treatment parameters or frequency (239, 267).
For more detail on the administration of ECT, see APA's The
Practice of Electroconvulsive Therapy: Recommendations for Treatment,
Training, and Privileging (A Task
Force Report of the American Psychiatric Association) (239).
b. Transcranial magnetic
Transcranial magnetic stimulation (TMS) uses a specifically designed
magnetic coil that is placed in contact with the head to generate
rapidly alternating magnetic-resonance-imaging-strength magnetic
fields and produce electrical stimulation of superficial cortical
neurons. Based on the results of a multisite randomized sham-controlled
clinical trial of high-frequency TMS over the left dorsolateral
prefrontal cortex (268), TMS was cleared by the FDA
in 2008 for use in individuals with major depressive disorder who
have not had a satisfactory response to at least one antidepressant
trial in the current episode of illness. However, another large
randomized sham-controlled trial of TMS added to antidepressant pharmacotherapy
showed no significant benefit of left dorsolateral prefrontal cortex
TMS (269). In comparisons of actual TMS versus sham
TMS, most (270–272) but not all (273)
recent meta-analyses have found relatively small to moderate benefits
of TMS in terms of clinical response. Although the primary studies
used in these meta-analyses are highly overlapping and the variability
in TMS stimulus parameters and treatment paradigms complicates the
interpretation of research findings, these meta-analyses also support the
use of high-frequency TMS over the left dorsolateral prefrontal
cortex. Lesser degrees of treatment resistance may be associated
with a better acute response to TMS (274).
In comparison with ECT, TMS has been found in randomized studies
to be either less effective than ECT (275) or comparable
in efficacy to ECT (276–278), but in the latter studies
TMS was more effective and ECT was less effective than is typically
seen in clinical trials. A fewer number of studies have compared
cognitive effects of TMS and ECT. One randomized trial found no significant
difference between TMS and non-dominant unilateral ECT on performance
on neuropsychological tests at 2 and at 4 weeks of treatment (276),
although a small open-label trial reported a greater degree of memory
difficulties with ECT than with TMS shortly after the treatment
Across all studies, TMS was well tolerated and was associated with
low rates of treatment dropout (270, 280).
Transient scalp discomfort and headaches were the most commonly
reported side effects (280).
In clinical practice, the need for daily TMS could produce logistical
barriers for some patients.
c. Vagus nerve stimulation
Vagus nerve stimulation is approved for use in patients with treatment-resistant
depression on the basis of its potential benefit with long-term
treatment. There is no indication for the use of VNS in acute phase
treatment of depression, as data showed no evidence for acute efficacy
(281, 282). Further information on the
use of VNS as an adjunct to other antidepressive treatments is provided
in Section II.B.7.c.
There has been considerable research on time-limited psychotherapies
for major depressive disorder, although the number of studies is
smaller than for pharmacotherapies. Most research has focused on
individual, in-person, outpatient treatment, in part based on the
needs and constraints of research methods. However, research has
also begun to explore psychotherapies in differing formats, including
groups, over the telephone, and with computer assistance.
When psychotherapy is part of the treatment plan, it must be
integrated with psychiatric management (Section II.A) and any other
treatments (e.g., pharmacotherapy) that are being provided. Clinical
considerations and other patient factors should be considered in
determining the nature and intensity of psychotherapy. Typically
psychotherapy is given in an ambulatory setting, although some psychotherapies
might benefit depressed inpatients, given adequate lengths of stay
and courses of treatment (283–285). Like pharmacotherapy,
the effectiveness of psychotherapy will vary with the skill and training
of the therapist. Patient factors, such as the nature and duration
of depressive symptoms, beliefs and attitudes toward psychotherapy,
and early life experiences (e.g., history of trauma) also affect
treatment response to psychotherapy. Psychotherapy is particularly
useful in addressing the psychosocial stressors and psychological
factors that have an impact on the development or maintenance of
Cognitive-behavioral therapy, interpersonal psychotherapy (IPT),
and behavioral psychotherapies (e.g., behavioral activation) have
demonstrated acute efficacy in treating major depressive disorder.
There is less evidence for other psychotherapies. However, one meta-analysis
found no large differences in long-term efficacy between any of
the major psychotherapies, including dynamic psychotherapy, for
mild and moderate depression (286). In terms of longer
term outcomes, psychotherapy is generally found to have more prolonged
effects than pharmacotherapy after cessation of active treatment.
In particular, IPT and CBT have shown lasting benefits in maintaining
remission (287–289). These time-limited treatments
are essentially equipotent with antidepressant medications for outpatients
with mild to moderate acute depression but probably should be used
in conjunction with medication for severe or melancholic major depressive
disorder. Some research has suggested patient and illness characteristics
that might predict differential benefits of CBT over IPT, and vice versa,
for patients with major depressive disorder (e.g., see reference 290),
but such preliminary findings require replication.
Cognitive-behavioral therapy and IPT appear less effective
than pharmacotherapy for chronic depression, at least as acute monotherapy
(291–296). Nonetheless, in patients who respond
to medication, psychotherapy may foster the development of social
skills and confidence after years of depression-related impairments
Psychotherapy carries its own "side effects." A
psychotherapy that requires considerable time or patience may be poorly
tolerated. The work of psychotherapy itself may generate anxiety
or other strong feelings, which may be difficult for patients to
manage. An indirect measure of the relative side effects and tolerability
of psychotherapy can be obtained from the dropout rates in clinical
trials; however, many other factors can also affect these rates
(e.g., other burdens of the research trial, specific features of
the clinical management provided, logistical barriers in attending
appointments). Depending on what can reasonably be expected with
the given type of psychotherapy, the psychiatrist should consider
a change in the intensity or type of psychotherapy and/or
addition or change to medication if psychotherapy for major depressive
disorder has not resulted in significant improvement in 4–8
a. Specific psychotherapies
1. Cognitive and behavioral therapies
In the treatment of depressed patients, psychotherapies that focus
primarily on aspects of cognitive patterns and those that emphasize
behavioral techniques can be used alone, but are generally used
in combination. Cognitive-behavioral therapy combines cognitive
psychotherapy with behavioral therapy and maintains that irrational beliefs
and distorted attitudes toward the self, the environment, and the
future perpetuate depressive affects and compromise functioning.
The goal of CBT is to reduce depressive symptoms by challenging
and reversing these beliefs and attitudes and encouraging patients to
change their maladaptive preconceptions and behaviors in real life
Cognitive-behavioral therapy is an effective treatment for major
depressive disorder. In meta-analyses, CBT has generally surpassed
control conditions in efficacy and has had equal efficacy compared
with other empirically supported psychotherapies (i.e., IPT and
behavior therapy) (299). Studies comparing the effectiveness
of CBT with pharmacotherapy, however, are methodologically challenging to
conduct, and results are inconsistent (296, 300).
Also unclear is whether CBT is less effective for patients with
more severe depressive symptoms.
Behavior therapy for major depressive disorder is based on
theoretical models drawn from behavior theory (301)
and social learning theory (302). Behavioral activation
is a newly articulated behavioral intervention with some positive
preliminary results that merit further study (288, 303).
Specific behavior therapy techniques include activity scheduling (304, 305),
self-control therapy (306), social skills training (307),
and problem solving (308). Behavior therapy involves graded
homework, scheduling of enjoyable activities, and minimizing unpleasant
activities (309). Behavior therapy has demonstrated
efficacy, at times superior to cognitive therapy, in treating major
depressive disorder (310).
2. Interpersonal psychotherapy
The focus of IPT is on current life changes, including losses, role
disputes and role transitions, social isolation, deficits in social
skills, and other interpersonal factors that may interact with the
development of the depressive episode (311, 312). In
IPT the goal is to intervene by identifying the current trigger
of the depressive episode, facilitating mourning in the case of
bereavement, promoting recognition of related affects, resolving
role disputes and role transitions, and building social skills to
improve relationships and to acquire needed social supports. In
IPT, major depressive disorder is defined as a medical illness,
and the illness, rather than the patient, is blamed for the symptoms.
Interpersonal psychotherapy's medical model makes it highly
compatible with pharmacotherapy in combined treatment.
Interpersonal psychotherapy is an efficacious treatment for
major depressive disorder (296, 313).
Studies have shown efficacy of this treatment in depressed primary
care patients and patients with more severe depression (311).
The efficacy of IPT has been demonstrated for adolescents, pregnant
and postpartum women, and geriatric patients (311).
Interpersonal psychotherapy can also be used as a monthly maintenance
therapy to prevent relapse (289, 314, 315).
Some studies have also suggested possible subgroups in whom IPT may
show differential efficacy, specifically among HIV-positive patients
(316) and patients who have co-occurring obsessive
personality traits and who are single and not living with others
(317). Furthermore, for patients with severe life events,
IPT may have advantages over therapies that do not focus on such
3. Psychodynamic psychotherapy
The term "psychodynamic psychotherapy" encompasses
a range of brief to long-term psychotherapeutic interventions (318–320).
These interventions derive from psychodynamic theories about the
etiology of psychological vulnerability, personality development,
and symptom formation as shaped by development and conflict occurring
during the life cycle from earliest childhood forward (321–325).
Some of these theories focus on conflicts related to guilt, shame,
interpersonal relationships, the management of anxiety, and repressed
or unacceptable impulses. Others address developmental psychological
deficits produced by inadequacies or problems in the relationship
between the child and emotional caretakers, resulting in problems
of self-esteem, sense of psychological cohesiveness, and emotional
self-regulation (323, 326–330).
Psychodynamic psychotherapy may be brief but usually has a
longer duration than other psychotherapies, and its aims extend
beyond immediate symptom relief. These goals are to modify underlying
psychological conflicts and deficits, which increase the patient's
vulnerability to depressive affect and the development of major
depressive disorder. Psychodynamic psychotherapy is therefore broader than
most other psychotherapies, encompassing both current and past problems in
interpersonal relationships, self-esteem, and developmental conflicts
associated with anxiety, guilt, or shame. Time-limited, structured
psychodynamic psychotherapy may focus more on understanding the
psychological basis of the presenting symptoms or on a selected
underlying conflict. Sometimes a goal of psychodynamic psychotherapy,
brief or extended, may be to help the patient accept or adhere to
necessary pharmacotherapy (331).
Although psychodynamic psychotherapy is often used in clinical
practice, its efficacy in the acute phase of major depressive disorder
remains less well studied in controlled trials than the efficacy
in this phase of some other forms of psychotherapy. This research
is reviewed in Part B, Section V.B.3.
4. Problem-solving therapy
Problem-solving therapy is a manual-guided, brief treatment lasting
six to 12 sessions. This therapy, often administered by nurses or
social workers, has been used to prevent depression in elderly and/or
medically ill patients, and it has also been used to treat patients
with relatively mild depressive symptoms. The approach combines elements
of cognitive therapy (addressing negative assessments of situations)
and IPT (focal problem solving). Some studies have reported modest
improvement in patients with mild depressive symptoms. Although
problem solving therapy has had limited testing for patients with
major depressive disorder, it may have a role in targeted patient
populations with mild depression (332–335).
5. Marital therapy and family therapy
Marital and family problems are common in the course of mood
disorders, and comprehensive treatment often demands assessing and
addressing these problems. Marital and family problems may be the
consequence of major depressive disorder but may also increase vulnerability
to developing major depressive disorder or retard recovery from
it (336–339). A number of marital and family
therapies have been shown to be effective in the treatment of depression. Techniques
include behavioral approaches (338), problem-focused
approaches (340), and strategic marital therapy (341, 342).
Family therapy has also been found to be helpful in the treatment
of more severe forms of depression in conjunction with medications
and hospitalization (343).
Group psychotherapy is widely practiced, but research on its application
to major depressive disorder is limited. Specific types having some
data to support their efficacy include CBT (344–347)
and IPT (348–351). Meta-analyses of the relative effectiveness
of psychotherapeutic approaches conducted in group format versus
individual format have not involved patients with rigorously defined
major depressive disorder (352–355).
On the basis of a very limited controlled study, supportive group
therapy has been suggested to have utility in the treatment of major
depressive disorder. In a study of depressed outpatients, a mutual
support group and group CBT were found to be equally effective in
reducing depressive symptoms (346). In a study of HIV-positive
patients with mild to moderate major depressive disorder, structured
supportive group therapy plus placebo yielded similar decreases
in depressive symptoms to structured group therapy plus fluoxetine
(356). Individuals experiencing stressors such as bereavement
or chronic illness may benefit from contact with others facing similar
Medication maintenance support groups may also offer benefits,
although data from controlled trials for patients with major depressive
disorder are lacking. Such groups inform the patient and family
members about prognosis and medication issues, providing a psychoeducational
forum that contextualizes a chronic mental illness in a medical
The efficacy of self-help groups led by lay members (357) in
the treatment of major depressive disorder has not been well studied.
However, one investigation of group therapies found that a higher
proportion of depressed outpatients had remission following treatment
in groups led by professionals than had remission following participation
in groups led by nonprofessionals (346). Further study
is needed on the possibility that self-help groups may serve a useful
role in enhancing the support network and self-esteem of participating patients
with major depressive disorder and their families.
Overall, group therapy has some evidence to support its use
as well as the potential advantage of lowered cost, inasmuch as
one or two therapists can treat a larger number of patients simultaneously.
This advantage needs to be weighed against the difficulties in assembling
the group, the lesser intensity of focus patients receive relative
to individual psychotherapy, and potentially adverse effects from
interactions with other group members.
It can be useful to establish an expected duration of psychotherapy
at the start of treatment. Communicating this expectation may help
mobilize the patient and focus treatment goals, yet there are few
data available on the optimal duration of specific depression-focused
psychotherapies. In many trials, CBT and IPT have been delivered
in approximately 12–16 weekly sessions. In a subanalysis
of one clinical trial, CBT delivered in 16 weeks was more effective
than CBT delivered in 8 weeks for those with severe major depressive
disorder (358). Moreover, evidence suggests benefit
from monthly continuation phase treatment with IPT in reducing the
probability of relapse (314). In addition, patients
with chronic, treatment-resistant depression may require long-term
The optimal frequency of psychotherapy has not been rigorously
studied in controlled trials. The psychiatrist should consider multiple
factors when determining the frequency for individual patients,
including the specific type and goals of the psychotherapy, the
frequency necessary to create and maintain a therapeutic relationship,
the frequency of visits required to ensure treatment adherence,
and the frequency necessary to monitor and address suicide risk
and other safety concerns. Time-limited brief psychotherapies may mobilize
many depressed patients to more rapid improvement. The severity
of illness, the patient's cooperation with treatment, availability
of social supports, cost, geographic accessibility, and presence
of co-occurring general medical problems may also influence visit
frequency. The frequency of outpatient visits during the acute phase
is generally weekly but may vary based on these factors. Some experienced
clinicians find that sessions are needed at least twice weekly,
at least initially, for patients with moderate to severe depression.
Regardless of the type of depression-focused psychotherapy
that is selected, the clinician should carefully and systematically
monitor the patient's response to treatment, which can be
facilitated by the use of clinician- and patient-rated scales at regular
intervals (see Section II.A.8). If 4–8 weeks of treatment
do not yield at least moderate improvement (20% diminution
in symptoms), the clinician should thoroughly review and reappraise
the treatment plan.
c. Combining psychotherapy
Several meta-analyses of studies of the combination of psychotherapy
and pharmacotherapy for patients with major depressive disorder
have documented a modest advantage for the combination as compared
with one or the other modality alone (359–361).
Particularly large additive effects have been observed in individual
studies of patients with chronic depression (362),
patients with severe recurrent depression (359), and
hospitalized patients (285). Combined treatment might
therefore be considered a treatment of first choice for patients
with major depressive disorder with more severe, chronic, or complex
presentations. Combining family therapy with pharmacotherapy has
also been found to improve posthospital care for depressed patients
Dual treatment combines the unique advantages of each therapeutic
modality: while pharmacotherapy may provide earlier symptomatic
relief, psychotherapy yields broader and longer lasting improvement
(363). Psychotherapy can also be used to address issues
that arise during pharmacotherapy, such as decreased adherence.
However, the advantage of routinely combining interventions may
be modest for patients with less severe depressive symptoms (359).
There are no empirical data from clinical trials to help guide
the selection of particular antidepressant medications and particular
models of psychotherapeutic approaches for individuals who will
receive the combination of both modalities. In general, the same
issues that influence these decisions when choosing a monotherapy
will apply, and the same doses of antidepressant medication and
the same frequency and course of psychotherapy should be used for
patients receiving combination modality treatments as are used for
patients receiving them as a monotherapy.
Results from a series of recent studies provide indirect evidence
that for patients who have had only a partial response to pharmacotherapy,
adding a course of CBT may be an effective strategy for preventing
relapse (363–368). During 12 weeks of treatment
in the STAR*D study, cognitive therapy was as effective
as either augmenting with bupropion or buspirone or changing antidepressants
to bupropion, sertraline, or venlafaxine. However, patients who
did not respond to an initial course of citalopram were less likely
to accept cognitive therapy as a change or augmentation option than
they were to accept a different medication option (369),
perhaps due to the nature of the study design.
and alternative treatments
As defined by the National Center for Complementary and Alternative
Medicine, complementary and alternative medicine is "a
group of diverse medical and health care systems, practices, and
products that are not presently considered to be part of conventional
medicine." As the definitions are usually applied, "complementary" therapies
are used conjunctively with conventional medicine, "alternative" therapies
are used in place of conventional medicine, and "integrative" medicine makes
use of all therapies appropriate to an individual patient's
The use of integrative therapies is increasingly common, although
training and comfort with complementary and alternative modalities
vary greatly by practitioner. Many patients do not spontaneously
disclose use of complementary or alternative treatments to health
care professionals, so it is particularly important that direct
inquiry about such treatments be part of routine health care questions. At
this time, there are several modalities that have modest evidence
for antidepressant efficacy and deserve further study. Some of these modalities
can be recommended with enthusiasm for their general health benefits;
however, patients should be informed that evidence for their antidepressant
efficacy as monotherapy is limited or absent.
St. John's wort is a plant widely used to treat depressive
symptoms. Overall, studies of St. John's wort show greater
consensus and support for benefits in mild to moderate major depressive
disorder, as compared with less consistent findings in patients
with more severe symptoms. One review of 14 short-term, double-blind
trials conducted in outpatients with mild to moderate symptoms of
major depressive disorder concluded that St. John's wort
in doses of 300 mg/day and 1,800 mg/day had efficacy
that was superior to placebo (105). St. John's
wort had generally comparable efficacy and fewer side effects than
low-dose TCA treatment (e.g., 30–150 mg/day of amitriptyline)
(105), but doses at the low end of this range would
not be expected to produce therapeutic benefits. However, in the
two largest controlled studies conducted in the United States (370, 371),
effects of St. John's wort did not differ from placebo,
which somewhat limits confidence in the magnitude of the antidepressant
actions of St. John's wort. In addition, preparations of
St. John's wort are not regulated by the FDA as a drug
and lack standardization of their ingredients, composition, and
potency. Based on the evidence cited, St. John's wort would
not meet the FDA's minimum requirements to be declared
an effective antidepressant and is not recommended for general use
in treating depression.
Another important consideration with St. John's wort
is the potential for drug-drug interactions (372–374).
St. John's wort appears to induce the metabolism of drugs
via CYP 3A4, reducing the efficacy of medications, including antiretroviral
medications, immunosuppressants (including cyclosporine), antineoplastic
agents, anticoagulants (including warfarin), oral contraceptives,
and hormone replacement therapy (373, 374).
Unwanted pregnancies have been reported with concomitant St. John's
wort and oral contraceptive use (373, 375, 376),
and rejection of transplanted organs has been observed when St.
John's wort is taken concurrently with cyclosporin (374).
The significant decreases in antiretroviral medication levels with
concomitant St. John's wort use suggest that these medications
will be less effective in treating HIV infection (374). Effects
of St. John's wort on P-glycoprotein have also been observed,
altering the pharmacokinetics and pharmacodynamics of medications such
as digoxin that are transported by this route (374).
Apart from affecting blood levels of nonpsychiatric medications, the
safety and efficacy of the combined use of St. John's wort with
other antidepressant medications is not known. The combined use
of St. John's wort with MAOIs is contraindicated.
S-adenosyl methionine is
a naturally occurring molecule. In humans, it is concentrated in
the liver and the brain and serves as a methyl donor in the synthesis
of biologically active compounds such as phospholipids, catecholamines,
and the neurotransmitters dopamine and serotonin (377).
Cerebrospinal fluid levels of SAMe are lower in individuals with severe
major depressive disorder, compared with control subjects (378),
and treatment with SAMe increases CSF SAMe and 5-hydroxyindoleacetic
acid levels (379). S-adenosyl
methionine is available for both parenteral and oral administration
Some data support the efficacy and tolerability of SAMe in patients
with major depressive disorder. Oral, intravenous, and intramuscular
formulations have been assessed and appear efficacious in at least
pilot studies (381–383). Like St. John's
wort, SAMe is not regulated by the FDA and lacks standardization
of its composition and potency. S-adenosyl methionine
has been compared with TCAs and has been reported to have greater
efficacy while being more tolerable (381). It is unclear
at present how SAMe compares to SSRIs in efficacy and cost-efficiency.
While some data support the use of SAMe as monotherapy and as augmentation
therapy, the data at this time, as with St. John's wort,
are insufficient to make a recommendation for its use in the treatment
of major depressive disorder.
Most studies of omega-3 fatty acids for major depressive disorder
have been adjunctive studies, in which patients were already receiving
antidepressant medications but still met the criteria for major
depressive disorder. Studies vary in the omega-3 fatty acids used
(eicosapentaenoic acid [EPA], docosahexaenoic
acid [DHA], or the combination), and doses and
durations of study trials have also varied. It is difficult to interpret
the literature on this treatment given the heterogeneity in study
design and outcomes.
Omega-3 fatty acids are generally recommended as an adjunctive
therapy for mood disorders, as health benefits, including those
for cardiovascular health, are well established, and individuals
with psychiatric disorders may be at greater risk for obesity, metabolic
problems, and other health problems than the general population
(384, 385). More evidence is required
to establish a definitive role in the acute treatment of major depressive
disorder. Doses of 1–9 grams have been studied in mood
disorders, with a majority of evidence supporting use of lower doses.
Adjunctive EPA or the combination of EPA and DHA (the combination found
in most commercially available brands) appears most useful, with
less evidence for DHA alone for the treatment of major depressive
disorder. Further data are needed to ascertain the role of omega-3
fatty acids as monotherapy for major depressive disorder.
Folate has been primarily assessed as a predictor of antidepressant
medication response and as an adjunctive treatment. Low folate blood
levels have been associated with lack of response and slower response
to fluoxetine for major depressive disorder (386, 387),
and higher folate levels at treatment baseline appear associated
with better response to antidepressants (388). Folate
has been studied as an adjunctive treatment compared with placebo
in addition to fluoxetine, with significantly greater improvement
in those receiving folate, especially among female patients (389).
Folate is a low-risk intervention with general health benefits.
Folate protects against neural tube defects in early pregnancy.
In general, 0.4–1 mg of folate is recommended for women
of reproductive age. Considering the modest evidence that supports
folate as an augmentation strategy and its attractive risk-benefit
profile, folate can be recommended as a reasonable adjunctive strategy
for major depressive disorder that carries little risk and may decrease
birth defects in the case of pregnancy. Data are inadequate to suggest
efficacy as a monotherapy.
Bright light therapy appears effective for seasonal affective disorder
and nonseasonal major depressive disorder, as demonstrated in generally
short-term, placebo-controlled trials (390–394),
although some studies have methodological limitations (395).
The mechanism of action for light therapy is not clear but appears
to involve the serotonergic neurotransmitter system (396, 397).
There is some evidence that light therapy may hasten the response
to treatment with antidepressant medication (398).
Open-label data also support light therapy for patients with major
depressive disorder that has not responded to antidepressant medication
(399). Greater intensity of light is associated with
efficacy (400). Light therapy also may augment the
antidepressant benefits of partial sleep deprivation (401, 402).
Monitoring for mania and hypomania may be appropriate with initiation
of light therapy, as hypomania has been reported (392).
However, in general bright light therapy is a low-risk and low-cost
option for treatment.
Acupuncture is a treatment modality that is part of traditional Chinese
medicine. Its efficacy is somewhat difficult to assess, as much
of the research is published in Asian languages and overlooked in
typical literature searches. In addition, there is significant variation
in the acupuncture techniques used as well as limited descriptions
of methodology and diagnosis (403). One randomized
trial showed comparable benefits of electroacupuncture and amitriptyline
(404), and another small randomized trial in depressed women
showed benefits of acupuncture relative to a sham control (405).
However, a subsequent larger study did not replicate these results
(406), and a recent meta-analysis concluded that acupuncture
was not associated with any benefits in treating major depression in
terms of response or remission rates (407). Assuming
needles are properly sterilized, there do not appear to be substantial
risks of acupuncture treatment. However, based on current evidence,
acupuncture is not recommended in the treatment of major depressive
7. Strategies to
address incomplete response
The psychiatrist should consider a change in treatment for patients
who have not fully responded to an adequate acute phase treatment
over a sufficient time, generally 4–8 weeks. The treatment
plan can be revised by implementing one of several therapeutic options,
including optimizing the initial treatment, changing to a different
treatment, and combining treatments. These options are outlined
in Figure 2 and described in more detail below.
Figure 2. Assessment of Treatment
Tolerability and Adequacy of Response
Following any change in treatment, the patient should continue
to be closely monitored. If there is not at least a moderate improvement
in major depressive disorder symptoms after an additional 4–8
weeks of treatment, the psychiatrist should conduct another thorough
review. This reappraisal should include the following: verifying
the patient's diagnosis and adherence; uncovering and addressing
clinical factors that may be preventing improvement, such as the
presence of co-occurring general medical conditions or psychiatric
conditions (e.g., alcohol or substance abuse); evaluating for potential drug-drug
interactions; obtaining collateral information from those involved
with the patient; and uncovering and addressing psychosocial, psychological,
and personality factors that may be impeding recovery (Table 9).
If no new information is uncovered to explain the patient's
lack of adequate response, other treatment options should be considered,
including ECT and a consultation from an expert in mood disorders.
Despite optimal treatment, some patients may continue to have chronic
depressive symptoms. For these patients, the psychiatrist should
add a disease management component to the overall treatment plan.
This component involves setting realistic expectations, improving
functioning, and developing self-management skills (415, 416).
a. Maximizing initial
For patients who have not fully responded to treatment for depression,
an initial strategy is to optimize the intensity of psychotherapy
or maximize the dose of medication, especially if the upper limit
of the antidepressant dose has not been reached. Decisions about
pharmacotherapy will involve a balancing of efficacy, side effects,
and medication adherence. Dose escalation and management of side
effects at critical decision points are essential in order to avoid
premature discontinuation of the chosen antidepressant medication
and to maximize the dose and duration of the antidepressant therapy
Because of pharmacokinetic and pharmacodynamic differences
among individuals, some patients may require doses higher than those
approved by the FDA to achieve adequate blood levels of a medication
and receive therapeutic benefits. Patients who have had their dose
increased should be monitored for increased severity of side effects;
dose increases should be considered only for patients who do not
have significant or intolerable side effects while taking the medication.
Frequent follow-up contact (either in person or via the phone) may
be necessary to address symptoms, side effects, and patient adherence
in order to personalize treatment to the specific clinical needs
of the patient. When available and clinically meaningful, therapeutic
ranges for blood levels of antidepressant medications are useful
in optimizing medication dosing (201, 232, 233).
Individual differences are common in the time to response and
the tolerability of treatments. For patients who have shown a partial
response to treatment, particularly those with features of personality
disorders and prominent psychosocial stressors, extending the antidepressant
medication trial (e.g., by 4–8 weeks) may allow up to one-third
of patients to respond more fully (417–419).
In patients who are receiving psychotherapy, similar principles
apply in terms of monitoring and adjusting treatment in the context
of nonresponse or difficulty tolerating psychotherapy (331).
Factors to be considered include the frequency of sessions, the
type of psychotherapy being used, the quality of the therapeutic
alliance, and the possible need for medications in lieu of or in
addition to psychotherapy. Whereas increasing the frequency of therapy
sessions is a reasonable approach to nonresponse, this approach
is based on clinical wisdom and has not been systematically studied.
b. Changing to other
Changing to a different non-MAOI antidepressant medication
is a common strategy for patients with treatment-resistant major
depressive disorder, especially those who have not shown at least
partial response to the initial medication regimen. Although there
are no specific patient characteristics that predict which medication
to choose (420), results from STAR*D suggest
that changing to a second-step treatment results in additional remission
rates of about 25%, and further changes are associated
with continued remission, albeit at lower rates (about 13%–14%).
Treatment can be changed to a non-MAOI antidepressant medication
from the same pharmacological class (e.g., from one SSRI to another
SSRI) or to one from a different class (e.g., from an SSRI to a
TCA) (369, 421–422). Adding or
changing to a depression-focused psychotherapy should also be considered
for patients with major depressive disorder who do not respond fully
to medication treatment. Other strategies for patients who do not
respond adequately to pharmacotherapy include changing to an MAOI
after allowing sufficient time between medications to avoid hazardous
interactions (see Table 8). Transcranial magnetic stimulation could
also be an option, as it appears to be safe and well tolerated (270, 280).
In addition, it has shown small to moderate benefits in most (268, 270–272)
but not all (269, 273
trials and recent meta-analyses. Recent randomized trials suggest
that quetiapine monotherapy also produces a greater reduction in depressive
symptoms than placebo (423, 424), with
comparable efficacy to duloxetine (424), although the
potential side effects of second-generation antipsychotic treatment
need to be taken into consideration. ECT remains the most effective
therapy for patients with treatment-resistant symptoms (239, 425),
although results of clinical trials differ on whether patients with
medication-resistant symptoms have responses to ECT that are comparable to
those of patients without documented medication resistance (426–428).
c. Augmenting and
Pharmacotherapy can be combined with a depression-focused
psychotherapy, both as an initial treatment plan, and as a strategy
to address nonresponse to treatment in one modality or the other.
See Section II.B.4.c above for further information about combining
pharmacotherapy and psychotherapy.
Antidepressant medications can be augmented with another non-MAOI
antidepressant or with other, nonantidepressant agents. The addition
of a second non-MAOI antidepressant may be helpful, particularly
for patients who have had a partial response to antidepressant monotherapy
(429). One option is to add a second non-MAOI antidepressant
medication from a different pharmacological class, taking care to
avoid drug-drug interactions. Another option is to add an adjunctive,
nonantidepressant medicationsuch as lithium, thyroid hormone,
an anticonvulsant, a psychostimulant, or a second-generation (atypical)
antipsychotic. More information about these strategies is given
later in this section.
Some limited evidence and
clinical experience support the addition of bupropion to an SSRI.
This combination is generally well tolerated, although bupropion,
a moderately potent inhibitor of CYP 2D6, increases blood levels
of some SSRIs (430). In one study, combined treatment
with bupropion and an SSRI resulted in better outcomes than either
therapy alone (431). Another commonly used strategy
is the combination of mirtazapine and an SSRI or venlafaxine. Generally,
mirtazapine 15–30 mg at bedtime is added to the incompletely
effective antidepressant and titrated up to 45 mg/day on
the basis of response and tolerability (432).
For patients with pronounced anxiety or persistent insomnia not
adequately relieved by an SSRI or SNRI, adjunctive use of anxiolytic
and sedative-hypnotic medications is common (433, 434).
These include benzodiazepines such as clonazepam (435)
and selective GABA agonists such as zolpidem (436) and
eszopiclone (437). Buspirone has also been used adjunctively
in anxious individuals (429). Although adjunctive therapy
of anxiety or insomnia can hasten symptomatic relief, there is no
evidence of sustained benefit, and some patients have difficulty
stopping the anxiolytic or hypnotic medication (438, 439).
Lithium, thyroid hormone, and stimulants are sometimes combined
with antidepressants to augment response. Lithium is the most extensively
studied of these adjuncts (440–443) and may
also reduce the long-term risk of suicide (444). The
interval before full response to adjunctive lithium is said to be
in the range of several days to 6 weeks. The blood level required
to enhance the effects of antidepressants still has not been confirmed.
If effective and well tolerated, lithium should be continued at
least for the duration of acute treatment and perhaps beyond the
acute phase for purposes of relapse prevention.
Thyroid hormone supplementation, even in euthyroid patients,
may increase the effectiveness of antidepressant medication treatment,
whether used as an augmentation agent (445, 446)
or in combination with an antidepressant from the outset of therapy
(447). The dose typically used for this purpose is
25 mcg/day of triiodothyronine, increased to 50 mcg/day
if the response is inadequate after about a week. The duration of
treatment required has not been well studied.
Second-generation antipsychotic medications may increase the
rates of response or remission of depressive symptoms in patients
who typically have not responded to more than two medication trials
(448), even when psychotic symptoms are not present.
Generally, in clinical practice, lower doses are used for antidepressant
augmentation than for treatment of psychosis. For example, the combination
of olanzapine and fluoxetine has been extensively studied (449–452)
and is typically initiated with 6 mg of olanzapine and 25 mg of
fluoxetine daily and titrated upward as tolerated to a maximum of 18
mg of olanzapine and 75 mg of fluoxetine daily. Aripiprazole has
received FDA approval for augmentation of antidepressive agents
and is typically initiated at 2.5–5 mg/day and titrated
upward as tolerated to a maximum of 15 mg/day (453). With
quetiapine, doses of 25 to 400 mg/day have been used, with
benefits for depressive symptoms found in some (454, 455)
but not all (456) clinical trials. Risperidone augmentation,
in doses of up to 3 mg daily (457, 458)
also appears to improve the response to antidepressant agents. In
most of these trials, the onset of the effect of second-generation
antipsychotic augmentation has been rapid, although the magnitude
of the advantage relative to placebo has been relatively modest.
In the only two trials to utilize active comparison groups, the combination
of olanzapine and fluoxetine was not significantly more effective
at study endpoint than continued therapy with nortriptyline (450)
or venlafaxine (451). Naturalistic follow-up data also
suggest that long-term weight gain can be problematic for many patients
receiving second-generation antipsychotic augmentation therapy,
particularly with the olanzapine-fluoxetine combination (459).
In addition, a recent meta-analysis suggests that the rate at which
SGA augmentation is discontinued is nearly four-fold greater than
study discontinuation among subjects randomly assigned to placebo (448).
When compared with other strategies for antidepressant nonresponders,
augmentation with a second-generation antipsychotic carries disadvantages:
the high cost of many agents, the significant risk of weight gain
and other metabolic complications (e.g., dyslipidemia, hypertriglyceridemia,
glucose dysregulation, diabetes mellitus), and potential risk of hyperprolactinemia,
tardive dyskinesia, neuroleptic malignant syndrome, and QTc prolongation.
Thus, the advantages and disadvantages of antipsychotic medications
should be considered when choosing this augmentation strategy. In
addition, when augmentation with a second-generation antipsychotic is effective,
it is uncertain how long augmentation therapy should be maintained.
Many clinicians find that augmentation of antidepressants with
low doses of stimulants such as methylphenidate or dextroamphetamine
may help ameliorate otherwise suboptimally responsive depression
(460–462), although not all clinical trials
have shown benefits from this strategy (463). More
recently, the novel compound modafinil has shown modest benefit
when combined with SSRIs, related to specific effects on residual
symptoms such as fatigue and hypersomnolence (464–467).
Although there are no clear guidelines regarding the length of time
stimulants or modafinil should be coadministered, in one extension
study the effects of modafinil were maintained across 12 weeks of additional
therapy (468). Physicians prescribing modafinil for
this off-label use should become familiar with rare but dangerous
cutaneous reactions to it, including reported instances of Stevens-Johnson
syndrome, toxic epidermal necrolysis, drug rash with eosinophilia
and systemic symptoms (469), and cytochrome P450 interactions.
Modafinil can also induce CYP 3A4 and render contraceptive medications and
other medications metabolized through this route ineffective.
Although their use in this context has not been extensively evaluated,
anticonvulsants such as carbamazepine, valproic acid, and lamotrigine
may offer some benefit in the treatment of medication-resistant
major depressive disorder (121, 429, 470–473).
A rarely used strategy is the combined use of a TCA or trazodone
and an MAOI (474, 475). The combination
of a TCA and an MAOI has been used for more than three decades for
treatment of some of the most treatment-resistant depressive disorders;
however, the risk of drug-drug interactions necessitates careful
monitoring (119). Of particular concern with these
combinations is the serotonin syndrome, characterized by delirium,
hyperthermia, hyperreflexia, myoclonus, and, rarely, death (see
Section II.B.2.b.5.b). Use of an MAOI in combination with a TCA
and related antidepressants should probably not be considered until
other pharmacological strategies for patients with treatment-resistant
illness have been exhausted; psychiatrists and patients choosing
to use the combination of an MAOI and a TCA should be well acquainted
with the potential hazards and carefully weigh the relative risks
and benefits of such a strategy.
Vagus nerve stimulation was approved for use in patients whose
symptoms have not responded to at least four adequate trials of
antidepressant medications and/or ECT. Acute benefits were
not observed in the sham-controlled portion of the VNS trial (282).
When compared with a parallel treatment-as-usual arm, the long-term
(1–2 year) open-label extension showed small (476, 477)
but persistent (478) improvements in symptoms with
VNS that could be clinically significant for some patients. Other
open-label studies have also shown some benefit when VNS is used
simultaneously with pharmacotherapy (479–481).
As with any surgical device implantation, there is a small risk
of postsurgical infection (482). A majority of individuals experience
hoarseness or voice alteration during stimulation, and coughing,
dyspnea, and neck discomfort are common (281, 481)
but generally are tolerable to patients (282, 479). Patients
also need to be informed of the implications of having an implanted
VNS device for future medical care (482). For example,
with a VNS device in place, brain MRI requires the use of a special
send-receive coil. The VNS device may affect the operation of other
implanted devices such as cardiac pacemakers or defibrillators and
other procedures such as diathermy, and whole body or radiofrequency
receive-only MRI are contraindicated. VNS is also contraindicated
in the presence of bilateral or left cervical vagotomy.
Relative to other antidepressive treatments, the role of VNS
remains a subject of debate. However, it could be considered as
an option for patients with substantial symptoms that have not responded
to repeated trials of antidepressant treatment.