International Journal of Psychiatry in Clinical Practice

ISSN: 1365-1501 (Print) 1471-1788 (Online) Journal homepage: http://www.tandfonline.com/loi/ijpc20

Evidence-based pharmacotherapy of treatment-

resistant unipolar depression

Markus Dold & Siegfried Kasper

To cite this article: Markus Dold & Siegfried Kasper (2017) Evidence-based pharmacotherapy
of treatment-resistant unipolar depression, International Journal of Psychiatry in Clinical Practice,
21:1, 13-23, DOI: 10.1080/13651501.2016.1248852

To link to this article: http://dx.doi.org/10.1080/13651501.2016.1248852

Published online: 16 Nov 2016.

Submit your article to this journal

Article views: 541

View related articles

View Crossmark data

Citing articles: 3 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ijpc20

Download by: [Australian Catholic University] Date: 03 October 2017, At: 11:22
 INTERNATIONAL JOURNAL OF PSYCHIATRY IN CLINICAL PRACTICE, 2017
VOL. 21, NO. 1, 13–23
http://dx.doi.org/10.1080/13651501.2016.1248852

REVIEW ARTICLE

Evidence-based pharmacotherapy of treatment-resistant unipolar depression

Markus Dold and Siegfried Kasper
Department of Psychiatry and Psychotherapy, Medical University of Vienna, Vienna, Austria

ABSTRACT ARTICLE HISTORY

Treatment resistance to the antidepressive pharmacotherapy represents one of the most important clinical Received 30 December 2015
challenges in the pharmacological management of unipolar depression. In this review, we aimed to sum- Revised 21 July 2016
marise the evidence for various pharmacological treatment options in therapy-resistant unipolar depres- Accepted 12 October 2016
sion derived from clinical trials, systematic reviews, meta-analyses and treatment guidelines. The first
measure in case of insufficient response to the initial antidepressant monotherapy contains the debarment KEYWORDS
of ‘pseudo-resistance’, potentially caused by inadequate dose and treatment duration of the antidepres- Depression; treatment
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

sant, insufficient plasma levels, non-compliance of the patient regarding medication intake or relevant resistance; non-response;
psychiatric and/or somatic comorbidities. Applying a dose escalation of the current antidepressant cannot antidepressants; antipsy-
be generally recommended as evidence-based treatment option and the efficacy depends on the class of chotics; lithium
antidepressants. There is no compelling evidence for a switch to another, new antidepressant compound
after insufficient response to a previous antidepressant. The combination of two antidepressants should
be preferentially established with antidepressants characterised by different mechanisms of action (e.g.
reuptake inhibitors together with presynaptic autoreceptor inhibitors). At present, the most convincing
body of evidence exists for the augmentation of antidepressants with second-generation antipsychotic
drugs and lithium. Hence, both strategies are consistently advised by treatment guidelines as pharmaco-
logical first-line strategy in treatment-resistant depression.

Introduction unipolar depression derived from clinical trials, systematic reviews,

meta-analyses and treatment guidelines.
Major depressive disorder (MDD) is one of the most common
medical illnesses worldwide. The lifetime prevalence varied in sur-
veys between 11.2 and 16% (Kessler et al. 2003; Bromet et al. Definitions of treatment resistance
2011; Kovess-Masfety et al. 2013) and according to a review of 25
At present, no uniform definition of treatment resistance in the
prevalence studies, the median 12-month prevalence rate
pharmacotherapy of unipolar depression is available (Kasper &
amounts 6.9% (range: 1.0–10.1%) (Wittchen et al. 2011). In Europe, Montgomery 2013). The term treatment resistance is mostly used
approximately 30.3 million people suffer from unipolar depression in case of insufficient improvement of depressive symptoms after
(Wittchen et al. 2011) and following the World Health medication with antidepressant drugs in advised dose and dur-
Organization (WHO), MDD ranks fourth in the leading causes of ation. Historically, the degree of resistance was defined according
the global burden of disease (WHO 2001). It is estimated that up to the number of treatment failures with antidepressants of differ-
to 10% of the patients attempt suicide (Holma et al. 2014). ent pharmaceutical classes assuming implicitly a hierarchy of effi-
Antidepressant drugs are the well-established first-line medication cacy with respect to the various classes of antidepressants (Thase
in the pharmacotherapy of unipolar depression and their efficacy & Rush 1997). In their staging model suggesting five levels of
could be verified in a large number of clinical trials (Montgomery treatment-resistant depression, Thase and Rush (1997) regarded
et al. 2007; Cipriani et al. 2009; APA 2010; Bauer et al. 2013b; monoamine oxidase (MAO) inhibitors as more efficacious than tri-
Bauer et al. 2015). However, up to 60% of the patients with MDD cyclic antidepressants (TCAs) which are on the contrary considered
do not respond satisfactorily to an initial antidepressant trial to be superior to selective serotonin reuptake inhibitors (SSRIs).
(Souery et al. 1999; Souery et al. 2006; Kasper & Montgomery The highest level of resistance comprised a failure of a course of
2013). In the first phase of the STAR
D (Sequenced Treatment bilateral electroconvulsive therapy (ECT) (Thase & Rush 1997). The
Alternatives to Relieve Depression) study for instance, 53% of the Massachusetts General Hospital (MGH) staging method represents
participants (n ¼ 2876) were classified as non-responders to a a further meaningful tool to access treatment-resistance to anti-
14-week monotherapy with citalopram (Trivedi et al. 2006b). Thus, depressant pharmacotherapy (Petersen et al. 2005). In this model,
the need for further pharmacological strategies to achieve suffi- a score of resistance can be calculated based on the number of
cient treatment response arises very frequently in the clinical prac- trial failures and the attempts of treatment optimisation. In 1999,
tice and treatment resistance represents one of the most the European Group for the Study of Resistant Depression imple-
important clinical challenges in the pharmacological management mented a staging method for treatment-resistant unipolar depres-
of depression. In this article, we sought to summarise the evi- sion (Souery et al. 1999). Following their approach, the criteria for
dence for pharmacological treatment options in therapy-resistant treatment resistance are fulfilled if a patient is resistant to at least

CONTACT Siegfried Kasper, M.D., Professor and Chair [email protected] Department of Psychiatry and Psychotherapy, Medical University of
Vienna, W€ahringer G€
urtel 18-20 A-1090 Vienna, Austria
ß 2016 Informa UK Limited, trading as Taylor & Francis Group
 14 MARKUS DOLD AND SIEGFRIED KASPER

Table 1. Definitions of treatment response implemented by Kasper and Table 3. Clinical factors associated with treatment resistance in unipolar depres-
Akimova (2013). sion according to the European multicenter study (n ¼ 702) of Souery et al.
Treatment response Definition (2007).
Inadequate response Insufficient response to one therapy Significance
Treatment non-response Insufficient response to two therapies Clinical factors levela
Treatment refractory Insufficient response to ‘more’ treatment options Comorbid anxiety disorder p < .001
Chronic depression Depression over 2 years Comorbid panic disorder p < .001
The treatment options can comprise psychopharmacological interventions and Current suicidal risk p < .001
non-pharmacological strategies such as electroconvulsive therapy (ECT) or deep- Severity of the current episode p ¼ .001
brain stimulation. Number of hospitalisations p ¼ .003
Social phobia p ¼ .008
Recurrent episodes versus single episodes p ¼ .009
Early age at onset (<18 years) p ¼ .009
Melancholic features p ¼ .018
Table 2. Thresholds criteria for treatment response suggested by the World Non-response to first antidepressant treatment lifetime p ¼ .019
Federation of Societies of Biological Psychiatry (WFSBP) (Bauer et al. 2013b). Personality disorder (DSM-IV criteria) p ¼ .049
Non-response 25% decrease in symptom severity compared Treatment resistance was defined by a failure of at least 2 consecutive trials
with baseline with antidepressant drugs.
a
Partial response 26–49% decrease in symptom severity compared 2-step logistic regression model using non-resistance/resistance as a dependent
with baseline variable.
Response 50% decrease in symptom severity compared
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

with baseline
Response with residual Response with partial remission.
clinically meaningful to distinguish between non-response and
symptoms intolerability to the antidepressant medication, for example
Remission No symptoms or very few symptoms of minor caused by non-tolerable adverse effects. In the clinical routine
severity are still present, defined by an absolute care, an exact differentiation of these two conditions is often diffi-
scale score (e.g. HAM-D total score 7) and
cult because many adverse effects are dose related.
improvement in psychosocial and occupational
functioning
The symptom severity is commonly assessed by rating scales such as the Factors contributing to treatment resistance
Montgomery Åsberg Depression Rating Scale (MADRS) (Montgomery and Asberg
1979) or the Hamilton Depression Rating Scale (HAM-D) (Hamilton 1960). Souery et al. (2007) analysed 702 MDD patients participating in a
large multicenter study of the European Group for the Study of
two consecutive adequate antidepressant trials regardless from Resistant Depression and revealed the following clinical factors
the administered class of antidepressant. The different stages of that were associated with treatment resistance (defined as failure
resistance correspond to the number of failed antidepressant trials of at least two consecutive trials with antidepressants): comorbid
(Souery et al. 1999). The European medicines agency (EMA) anxiety disorders (in particular panic disorder and social phobia),
divides between treatment resistance (at least two treatment fail- comorbid personality disorder, suicidal risk, high degree of symp-
ures with different antidepressant drug classes) on the one hand tom severity, presence of melancholic features, more than one
and inadequate response (one treatment failure) on the other previous hospitalisation, recurrent episodes, non-response to the
hand. first antidepressant and an age at onset of 18 (Souery et al.
A practical approach for categorisation was proposed by 2007) (Table 3). The finding of significantly poorer treatment out-
Kasper and Akimova (2013). According to their definitions, inad- come for anxious depression could be replicated in a study of the
equate response is present in case of insufficient response to one German Algorithm project enrolling 429 inpatients with MDD
therapy, treatment non-response in case of insufficient symptom (Wiethoff et al. 2010). Balestri et al. (2016) investigated predictors
improvement to two therapies, and the term treatment refractory for a particularly high degree of treatment resistance by analyzing
should be used after insufficient response to at least three treat- socio-demographic and clinical variables in 98 subjects with insuf-
ment options (Table 1). The criterion for chronic depression is ful- ficient response to at least three antidepressants including venla-
filled when the depressive symptoms persist for more than 2 faxine and escitalopram. They could determine the features long
years (Kasper & Akimova 2013). duration and high severity of the current episode, outpatient sta-
tus, high suicidal risk level, first/second degree psychiatric antece-
dents and the occurrence of adverse effects during the treatment
Criteria for non-response
as clinical predictors. In terms of biological risk factors for treat-
In recent clinical trials, the thresholds criteria for the definition of ment resistance, a systematic review identified decreased
non-response suggested for example by the World Federation of gamma-aminobutyric acid (GABA) levels in the brain and a few
Societies of Biological Psychiatry (WFSBP) (Bauer et al. 2013b) are polymorphisms of gene coding for the serotonergic system as fac-
commonly applied. Following these, a 25% decrease in symptom tors contributing to treatment resistance (Bennabi et al. 2015).
severity compared with baseline is defined as non-response, a
26–49% decrease as partial response and a 50% decrease in
Operationalising treatment-resistance and debarment of
symptom severity is considered as response (Table 2). The symp-
pseudo-resistance
tom severity is commonly measured by rating scales such as the
Montgomery Åsberg Depression Rating Scale (MADRS) If a patient does not respond adequately to the first administered
(Montgomery & Asberg 1979) or the Hamilton Depression Rating antidepressant drug, then it should be ensured as first step within
Scale (HAM-D) (Hamilton 1960). According to the WFSBP, remis- an algorithm that non-response to the initial pharmacotherapy in
sion is defined by the presence of no or very few symptoms of the sense of the definition (adequate dose and duration) is pre-
minor severity (for example operationalised by a HAM-D total sent. An only alleged resistance (so-called pseudo-resistance)
score of 7) and improvement in psychosocial and occupational should be ruled out. In this case, the first treatment step contains
functioning (Bauer et al. 2013b). In this context, it appears the optimisation of the current pharmacotherapy, for example by
 INTERNATIONAL JOURNAL OF PSYCHIATRY IN CLINICAL PRACTICE 15

Table 4. ‘Checklist’ with potential reasons for an only alleged resistance to the initial antidepressant medication (‘pseudo-resistance’).
 Is the administered dose of the antidepressant adequate according to the recommendations of the guidelines of the psychiatric societies?
 Is the duration of the treatment sufficiently long (at least 2–3 weeks in the target dose)?
 Are the compliance and adherence of the patient concerning the medication intake sufficient? Can non-compliance be ruled out (e.g. by applying plasma level
determinations)?
 Are adequate drug plasma levels achieved and verified by therapeutic drug monitoring (TDM)? Can metabolisation abnormalities in the cytochrome P450 enzyme
system probably causing insufficient plasma levels below the therapeutic threshold (e.g. in case of ‘ultra-rapid metabolisers’) be ruled out?
 Is a clinical response maybe masked by the occurrence of adverse effects of the antidepressant medication?
 Are relevant psychiatric and somatic co-morbidities sufficiently considered and it is ensured that the depressive disorder is the primary diagnosis?
 Are psychosocial stressors probably associated with the depressive symptoms adequately taken into account?
The table highlights issues that should be considered when evaluating the presence of pseudo-resistance to the current antidepressant pharmacotherapy. If applic-
able, the first measure contains optimisation of the current trial (e.g. adjusting the antidepressant dose, prolonging the treatment duration, and/or implementing
arrangements to promote compliance of the patient).

adjusting the dose or duration of the medication and arrange- Table 5. Recommended doses for antidepressant drugs in the pharmacological
ments to promote compliance. When evaluating possible treatment of unipolar disorder according to the recommendations of the WFSBP
guidelines for the biological treatment of depressive disorders (Bauer et al.
pseudo-resistance, the following issues should be considered 2013b).
(Table 4):
Recommended Recommended Level of
1. Were the antidepressant dose sufficient and the treatment dur- starting dose target dose recommendation
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

ation adequate? Recommended doses for the most frequently Drug (in mg/d) (in mg/d) for TDM
prescribed antidepressant drugs in the management of MDD Melatonergic antidepressants
are displayed in Table 5 according to the WFSBP guidelines Agomelatine 25 25–50 4
for the biological treatment of unipolar depressive disorders Monoamine oxidase inhibitors (MAOIs)
Moclobemide 150 300–600 3
(Bauer et al. 2013b). It should be considered that older adults
Tranylcypromine 10 20–60 4
or patients with co-morbid somatic diseases (e.g. cardiovascu- Noradrenaline–dopamine reuptake inhibitors (NDRIs)
lar illnesses) might require lower doses. The onset of symp- Bupropion 150 150–450 3
tom improvement can vary considerably between individual Noradrenaline reuptake inhibitors (NARIs)
patients and current guidelines advise a minimum treatment Reboxetine 4–8 8–12 3
Noradrenergic and specific serotonergic antidepressants (NaSSAs)
duration of at least 2–3 weeks until in case of unsatisfactory Mirtazapine 15 15–45 2
response a change of the treatment strategy should be taken Other antidepressants
into account (NICE 2009; Bauer et al. 2013b). Mianserine 30 60–120 3
2. Were the compliance and adherence of the patient with regard Tianeptine 12.5 25–37.5
Vortioxetine 5–10 10–20
to medication intake sufficient? Non-compliance and non-
Selective serotonin reuptake inhibitors (SSRIs)
adherence are frequent reasons for treatment failures with Citalopram 20 20–60 2
antidepressive drugs. It can be assumed that a large number Escitalopram 10 10–20 2
of patients do not take the prescribed medication correctly. Fluoxetine 20 20–60 2
Especially, when patients received an inpatient treatment Fluvoxamine 50 100–200 2
Paroxetine 20 20–60 3
after they had failed to respond to an initial medication Sertraline 50 50–150 2
attempt in an outpatient setting, a therapeutic success often Serotonin antagonist and reuptake inhibitors (SARIs)
occurs without a fundamental change of the pharmacother- Trazodone 50–100 200–600 2
apy. This is probably only caused by the fact that the medica- Serotonin–noradrenaline reuptake inhibitors (SNRIs)
tion intake now takes place continuously under controlled Duloxetine 30–60 60–120 2
Milnacipran 50–100 100–200 2
conditions. Plasma level determinations can objectify compli- Venlafaxine 37.5–75 75–375 2
ance and adherence in terms of medication intake. Tricyclic antidepressants (TCAs)
3. Were adequate drug plasma levels achieved? At present, there Amitriptyline 25–50 100–300 1
is for most of the antidepressants no compelling evidence for Clomipramine 25–50 100–250 1
Desipramine 25–50 100–300 2
a clear relationship between drug concentrations in the blood Imipramine 25–50 100–300 1
and clinical response. Dose titration guided by therapeutic Nortriptyline 25–50 75–200 1
drug monitoring (TDM) is at best appropriate for TCAs Trimipramine 25–50 100–300 2
(Hiemke et al. 2011). Levels of recommendations for the use Older adults or patients with co-morbid somatic diseases such as cardiovascular
of TDM according to a consensus guideline (Hiemke et al. illnesses might require lower doses. The levels of recommendations to use
2011) are displayed in Table 5. However, plasma level meas- therapeutic drug monitoring (TDM) are based on the consensus guidelines for
TDM of the ‘Arbeitsgemeinschaft f€ ur Neuropsychopharmakologie und
urements might be valuable in the clinical routine to verify Pharmakopsychiatrie (AGNP)’ (Hiemke et al. 2011). The different levels indicate
compliance of the patient with respect to medication intake, that applying TDM is ‘strongly recommend’ (level 1), ‘recommended’ (level 2),
to uncover potential metabolisation abnormalities, or in case ‘useful (level 3)’ or ‘potentially useful (level 4)’.
of the emergence of pronounced adverse effects even at low
doses. The plasma drug concentration should be determined
in the steady-state condition which is attained for most of last medication intake (Hiemke et al. 2011). Applying TDM
the psychopharmacological compounds after approximately and genotyping, metabolisation abnormalities as well as
five elimination half-lives after drug initiation. This corre- insufficient compliance of the patient can be identified or
sponds to most antidepressants with the exception of fluox- excluded as reason for treatment failure. Polymorphisms in
etine to 1 week after achievement of the target maintenance the cytochrome P450 enzyme system, which is responsible
dose. The blood draw should be accomplished ideally before for the metabolisation of most psychotropic drugs, can be
ingestion of the morning dose which is usually 12–16 h (or detected. On the one hand, increased enzyme activity can
24 h if the drug is given once daily in the morning) after the cause an accelerated metabolisation of the drugs (‘ultra-rapid
 16 MARKUS DOLD AND SIEGFRIED KASPER

metaboliser’, about 1% of the population). On the other Dose increase

hand, reduced enzyme activity can cause a slower metaboli-
According to the treatment guidelines of the international psychi-
sation (‘poor metaboliser’, about 5% of the population). As a
atric societies, an increase of the antidepressant dose above the
consequence, drugs are eliminated much more slowly than in
officially approved dose range (dose escalation) cannot be recom-
the normal population. The compounds can, therefore, accu-
mended as general treatment option for the management of
mulate in the body even at very low doses and in some cases
toxic drug concentration levels can be achieved. The occur- treatment-resistant MDD (NICE 2009; APA 2010; Bauer et al.
rence of severe adverse effects even at low doses can be 2013b). However, the results of dose-finding studies suggest that
interpreted as a sign for possible pharmacokinetic problems. the efficacy of a dose increase depends first of all from the class
Moreover, TDM is beneficial to reveal potential critical drug of antidepressants. For instance, there are positive findings for a
interactions in case of pharmacological combination thera- dose-response relationship of TCAs (Adli et al. 2005) and this is
pies. Clinically relevant in the context of drug plasma levels is why TDM is ‘strongly recommended’ only for this class of antide-
the interaction between smoking and a medication with pressants in a TDM consensus statement (Hiemke et al. 2011).
drugs that are metabolised mainly by the cytochrome P450 Furthermore, positive evidence is available for the irreversible
isoenzyme 1A2 such as the antidepressant duloxetine. MAO inhibitor tranylcypromine indicating an increase of efficacy
Because smoking induces this isoenzyme, the clearance of in high dosages, probably due to an additional amphetaminergic
these compounds can be increased significantly. Therefore, effect (Amsterdam & Berwish 1989; Adli et al. 2008). In terms of
a dose escalation might be necessary to achieve the the serotonin-noradrenaline reuptake inhibitors (SNRIs), there are
positive trial results for venlafaxine (Adli et al. 2005), whereas
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

therapeutic window (the blood concentration in which a

drug is most effective and the occurrence of adverse effects duloxetine 120 mg/d was not more efficacious than 60 mg/d in an
is low). 8-week double-blind randomised controlled trial (RCT) (Brecht
4. Was the treatment response maybe masked by adverse effects et al. 2011). However, none of the above-mentioned clinical stud-
of the antidepressants medication? For instance, sedation can ies applied plasma level measurements to determine the adminis-
be misinterpreted as lassitude or akathisia may mimic psycho- tered dose of the antidepressant. In general, it must be critically
motor restlessness. considered that the judgments with respect to the dose–escala-
5. Were psychiatric and somatic co-morbidities sufficiently consid- tion strategies are primarily based on mainly TDM and compara-
ered? It should be ensured and re-evaluated that the depres- tive dose studies. These findings, however, cannot directly be
sive disorder is the primary diagnosis and that no relevant transferred without reservation to the clinically meaningful ques-
psychiatric or somatic co-morbidities were overseen. tion of whether increasing the dose is beneficial for subjects with
Particularly, the presence of psychiatric co-morbidities such as insufficient initial symptom improvement to standard-dose treat-
anxiety disorders, substance abuse, and personality disorders ment with the same antidepressant compound. To elucidate this,
should be taken into account, as these may contribute to RCTs comparing a dose increase directly to the continuation of
treatment resistance (Souery et al. 2007; Wiethoff et al. 2010). the standard-dose medication in patients with initial non-response
Similarly, somatic diseases such as thyroid diseases can cause to this standard-dose pharmacotherapy are needed. This meth-
depressive symptoms. If present, an adequate treatment of odological approach would reflect the common clinical practice
these co-morbidities should be established. Moreover, the where a dose increase is often employed as next treatment meas-
ability of some somatic drugs to provoke depressive symp- ure after a standard-dose trial of the same drug had failed. We
toms should be carefully regarded. found RCTs applying this study design for dose–escalation strat-
6. Were psychosocial stressors adequately considered? egies with the SSRIs fluoxetine (Dornseif et al. 1989), paroxetine
Psychosocial factors can cause a progression of depressive (Benkert et al. 1997; Ruhe et al. 2009) and sertraline (Schweizer
disorders and should, therefore, be recognised within the et al. 2001; Licht & Qvitzau 2002). None of these studies could
treatment course. Moreover, a secondary gain from the illness demonstrate significant superiority of the dose increase over con-
for the patient should be ruled out when assessing the tinuing standard-dose treatment. However, a recent patient-level,
response to antidepressant treatment. pooled analysis of acute-phase, placebo-controlled, fixed-dose tri-
als (n ¼ 2859) found contradictorily a statistically significant advan-
tage of SSRI high-dose medication compared to SSRI low-dose
Treatment options after non-response to the initial treatment suggesting a potential dose–response relationship of
antidepressant the SSRIs (Hieronymus et al. 2015). Ruhe et al. (2009) examined in
In case of non-response to the initial antidepressive drug and their paroxetine study additionally to the clinical response the
after debarment of pseudo-resistance, the question of the next serotonin transporter (SERT) occupancy by SPECT measurements.
therapeutic measures within an algorithm to achieve sufficient They could show that an adequate SERT occupancy rate (about
symptom improvement arises. Strategies that were commonly 80% in the midbrain) is achieved at doses of paroxetine 20 mg/d
applied in this case usually contain (1) increasing the dose of and increasing the dose up to 40 mg/d does, therefore, not result
the initial antidepressant, (2) switching to another, new anti- in higher receptor occupancy as the saturation of the serotonin
depressant, (3) combining two antidepressants, (4) augmenting transporters is already nearly complete. These findings are consist-
the antidepressant with other agents (e.g. second-generation ent with the [11C]DASB positron emission tomography study of
antipsychotics (SGAs) or lithium), and (5) combining the anti- Meyer et al. (2004) investigating the SERT occupancy of SSRIs/ven-
depressant with non-pharmacological biological therapies (e.g. lafaxine whereupon SERT occupancy of 80% occurs already at the
therapeutic sleep deprivation, light therapy, ECT) (Bauer et al. minimum therapeutic doses. However, the possibility of a non-lin-
2013b). Two strategies that are often used in clinical routine ear relationship between SERT occupancy and clinical response
care as second step contain a dose increase of the current cannot be ruled out.
administered antidepressant agent (dose escalation and high- Even though dose escalation cannot be advised generally and
dose treatment) and, second, switching to another, new anti- depends on the class of antidepressants, individual patients might
depressant drug. respond to high-dose or even off-label treatment. Applying TDM
 INTERNATIONAL JOURNAL OF PSYCHIATRY IN CLINICAL PRACTICE 17

arises the possibility to identify patients that could probably bene- of antidepressants (Papakostas et al. 2008). The authors found
fit from high-dose medication (Hiemke et al. 2011). For example, slight but significantly higher remission rates for the latter strat-
patients with polymorphisms in the cytochrome P450 enzyme sys- egy (28% for the across-class switch versus 23.5% for the within-
tem causing an accelerated elimination of drugs (‘ultra-rapid class switch). However, these findings indicate that altogether
metaboliser’) require probably higher doses of their antidepressant only a low proportion of patients benefit from switching the anti-
to achieve clinical response if the plasma drug concentration is depressant drug and that also a switch between the different
below the effective therapeutic range in a standard dose. SSRIs can be efficacious. Similarly, the results of the STAR
D study
can be interpreted in this sense. In this trial, citalopram non-res-
ponders achieved remission rates between 17.6% and 24.8% after
Switching the antidepressant drug
switching to bupropion, sertraline or venlafaxine without identify-
Although a switch of the currently administered antidepressant ing any significant differences between the different antidepres-
drug is an often employed step in case of non-response, the ques- sant substances (Rush et al. 2006). In a large European multicenter
tion of the effectiveness of switching from one antidepressant study (n ¼ 340), switching to a different subclass of an antidepres-
drug to a second one (each as monotherapy) could be answered sant (across-class switch) was not significantly superior to a
against the background of sparse data only with reservation. Even within-class switch when analyzing response and remission rates
though the efficacy of all officially approved antidepressant drugs (Souery et al. 2011a). Montgomery et al. (2014) compared in a
is verified in positive clinical trials, it must be considered that prospective, randomised, double-blind trial a switch to vortioxe-
these findings are derived from data with mainly non-resistant tine on the one hand and agomelatine on the other hand in
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

patients and these results cannot be transferred without reserva- depressive patients with inadequate response to prior SSRI/SNRI
tion to subjects who have no or inadequate response to previous monotherapy and revealed a significant superiority of vortioxetine
antidepressant medication. To ascertain the efficacy of switching over agomelatine in efficacy-related outcomes.
strategies adequately, RCTs are needed that randomised non-res- A so-called cross-over titration is generally recommended for
ponders to an antidepressant in either an intervention group in the accomplishment of a switch of the antidepressant drugs.
which the participants are switched to a new antidepressant drug Thereby, the dose of the first antidepressant is gradually
(switching group) or in a control group in which the medication is decreased while simultaneously the dose of the second one is
continued with the initial antidepressant drug without any dose gradually titrated up to its target dose. However, when establish-
adjustments. Bschor and Baethge (2010) accomplished a meta- ing a pharmacotherapy with the irreversible MAO inhibitor tranyl-
analysis of RCTs with such study design. Only three studies repre- cypromine, a washout period for serotonin reuptake inhibitors for
senting 395 patients could be incorporated and no significant dif- at least 2 weeks (5 weeks for fluoxetine) must be warranted
ferences between the pooled switching group and the before the first administration of the MAO inhibitor. It should be
continuation group could be determined, neither for response nor considered that the termination of the medication with an anti-
for remission rates. Since the elaboration of this meta-analysis, the depressant compound can provoke withdrawal symptoms (Fava
double-blind RCT by Romera et al. (2012) was conducted using a et al. 2015) and these are often misinterpreted as adverse effects
double randomisation: Non-responders to an initial 4-week mono- of the newly dispensed antidepressant drug hampering the possi-
therapy with escitalopram (n ¼ 566) were randomised to either an bility to assess the effectiveness of the switch adequately.
early switch to duloxetine or continuing escitalopram for further
4 weeks in the control group. Participants of the control group
Combination strategies
without sufficient response even after altogether 8-week escitalo-
pram treatment were again randomised to a switch to duloxetine The term combination treatment describes the simultaneous
or continuing escitalopram. As main finding, no significant administration of at least two drugs of the same class such as two
between-group differences in terms of response rates were identi- antidepressants. Although this strategy is widely employed in daily
fied. In another open RCT, Souery et al. (2011b) compared in non- clinical practice (Dold et al. 2015), the evidence for this measure is
responders with citalopram or desipramine (n ¼ 189) a switch to rather sparse and depends first of all on the compounds used for
the respective other drug with the continuation of the first admin- concomitant medication. A meta-analysis compared in non-res-
istered agent and found significantly higher remission rates ponders with antidepressant monotherapy add-on treatment with
among the non-switched patients. a second antidepressant to continuation of the monotherapy. The
In conclusion, a switch of the antidepressant in case of non- authors could comprise only five trials (n ¼ 483) indicating insuffi-
response to the present antidepressant cannot be regarded as evi- cient evidence (Lopes Rocha et al. 2013). Furthermore, the find-
dence-based strategy. However, some patients appear to benefit ings were inconclusive and limited due to considerable
from a switch of medication. From a clinical point of view, a heterogeneity. The same also applies to a corresponding meta-
switching strategy should be preferably accomplished if there is analysis investigating the efficacy of combination strategies estab-
absolutely no response with the current antidepressant medica- lished from the beginning of the treatment (Rocha et al. 2012).
tion and/or severe adverse effects are present. Especially, in case With regard to initial combination strategies, two large RCTs
of partial response to the first antidepressant, other treatment should be taken into account: in the double-blind RCT of Blier
options appear more appropriate. et al. (2010) examining 105 MDD patients, combination treatments
Following theoretical pharmacological considerations, it seems of mirtazapine together with fluoxetine, venlafaxine and bupro-
auspicious to choose preferably as second, new antidepressant a pion were significantly more efficacious than fluoxetine monother-
compound with a different mechanism of action compared with apy in terms of remission rates, but not response rates.
the first dispensed agent. The WFSBP guidelines for instance Bupropion–escitalopram and venlafaxine–mirtazapine combina-
advise explicitly that switching from an SSRI to venlafaxine or tra- tions were not significantly superior to escitalopram monotherapy
nylcypromine appears justified within a treatment algorithm when analyzing efficacy outcomes in the single-blind randomised
(Bauer et al. 2013b). A meta-analysis comprising four clinical trials controlled Combining Medications to Enhance Depression
with altogether 1496 participants compared a switch from an SSRI Outcomes (CO-MED) study (n ¼ 665) (Rush et al. 2011). In the
with either a second course of an SSRI or a switch to another class STAR
D study, a combination of citalopram with bupropion in
 18 MARKUS DOLD AND SIEGFRIED KASPER

patients after a treatment failure with citalopram monotherapy placebo augmentation when analyzing response and remission
leads to a remission rate of 29.7% (Trivedi et al. 2006a). However, rates. Altogether, 44.2% of the patients in the pooled SGA group
the lack of a citalopram monotherapy continuation arm should be exhibited treatment response compared with 29.9% in the pla-
regarded in this context. cebo group. These positive findings for SGAs could be replicated
Considering all trial results evaluating combination treatments, by a number of further systematic reviews and meta-analyses
it can be recommended that concurrent medication with antide- (Spielmans et al. 2013; Wang et al. 2015; Zhou et al. 2015). The
pressants should be preferably established with reuptake inhibi- efficacy of aripiprazole especially for elderly people (>60 years)
tors such as SSRIs or SNRIs on the one hand and inhibitors of with treatment-resistant depression could be demonstrated in a
presynaptic autoreceptors such as noradrenergic and specific sero- double-blind RCT (n ¼ 181) using venlafaxine ER as ongoing anti-
tonergic antidepressants (NaSSAs) or serotonin antagonist and depressant medication (Lenze et al. 2015). In agreement with the
reuptake inhibitors (SARIs) (e.g. mirtazapine or trazodone) on the study results for other SGAs, a recent double-blind RCT not incor-
other hand. For these combinations, the most compelling evi- porated in the above-mentioned reviews found additionally sig-
dence is available and this is why the treatment guidelines recom- nificant efficacy for a ziprasidone augmentation of escitalopram
mend consistently especially this strategy (Bauer et al. 2013b; (Papakostas et al. 2015).
Bschor et al. 2014). When administering these subclasses of anti- Some SGAs received the official approval for the administration
depressants concomitantly, synergistic effects can be expected in unipolar depression by regulatory authorities. For example, que-
due to their complementary mechanisms of action. Furthermore, tiapine XR is licensed as adjunctive medication in addition to anti-
these combinations appear auspicious from a clinical viewpoint as depressants after non-response to prior antidepressant
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

presynaptic autoreceptor inhibitors are for instance, in contrast to monotherapy in the United States and the European Union. In the
SSRIs/SNRIs, characterised by meaningful sedating properties same way, aripiprazole has the regulatory approval for this indica-
(Bschor et al. 2014). However, a possible higher rate of drug inter- tion by the US Food and Drug Administration (FDA) as add-on
actions and adverse effects must be critically considered. In this treatment, and olanzapine has the FDA license in combination
context, a recent meta-analysis found a significantly higher occur- with fluoxetine. Other SGAs are currently not officially approved
rence of adverse effects for combinations of SSRIs together with for the management of depression. Recently published phase 3
NaSSAs or TCAs in comparison with monotherapy (Galling et al. studies suggest efficacy of the new drug application brexpiprazole
2015). in treatment-resistant MDD and the official approval for this indi-
Antidepressive drug combinations that are not recommended cation is sought for this substance (Thase et al. 2015).
by the guidelines due to rather negative trial results comprise for The recommended doses for SGAs in unipolar depression are
instance a simultaneous administration of TCAs together with lower than those commonly used in the management of schizo-
MAO inhibitors or SSRIs. In this context, it should be also high- phrenic disorders (Gardner et al. 2010; Dold & Leucht, 2014). For
lighted that a parallel treatment of MAO inhibitors with serotonin instance, the WFSBP recommends for aripiprazole a starting dose
reuptake inhibitors must not be conducted due to the risk for the of 2–5 mg/d and a maximum final dose of 15 mg/d. For quetiapine
occurrence of a serotonin syndrome. XR, the staring dose amounts 50 mg/d and the maximum final
dose 300 mg/d (Bauer et al. 2013b). In a small post-hoc study, the
dose of 39 non-responders to augmentation with aripiprazole
Augmentation strategies
2.5 mg/d was increased to 5 mg/d and only 12.8% of these
Augmentation treatment is defined as the concurrent medication patients achieved treatment response suggesting only a modest
of at least two drugs of different classes, for example the co- additional benefit of dose-increasing strategies (Mischoulon et al.
administration of an antidepressant with antipsychotic drugs, 2012). Pooled analyses of the SGA trials revealed a significant
mood stabilisers or benzodiazepines. The most frequently applied symptom improvement already during the first week of augmen-
augmentation strategies in the clinical routine care contain those tation treatment suggesting a very rapid onset of action (Bauer
with SGAs and lithium (Dold et al. 2015). et al. 2010; Nelson et al. 2012). This early reduction of the depres-
sive symptoms might be in part caused by beneficial effects on
sleep patterns of some SGAs (Bauer et al. 2010).
Augmentation with antipsychotic drugs
With regard to the risk profile of the SGA augmentation, the
Psychopharmacoepidemiological studies found consistently a sub- findings of the RCTs suggest that the emerging adverse effects in
stantial increase of SGA administrations in unipolar depression MDD patients are comparable to those occurring in the manage-
over the last years (Mohamed et al. 2009; Konstantinidis et al. ment of schizophrenia (e.g. akathisia with aripiprazole; or sedation
2012; Gerhard et al. 2014). For instance, a significant rise of the and weight gain with quetiapine). However, there is some evi-
proportion of MDD patients receiving SGAs from 12.8% in 2000 to dence that patients with unipolar depression are probably charac-
28.3% in 2007 was found in a drug safety programme analyzing terised by higher vulnerability for antipsychotic-induced adverse
1826 inpatients in German-speaking countries (Konstantinidis effects compared to schizophrenic patients (Kato & Chang 2013).
et al. 2012). Another survey carried out in the United States For instance, aripiprazole, which leads to minimal weight gain in
(n ¼ 7757) detected an increase of SGA prescriptions in outpa- schizophrenia patients, was shown to cause an increase in body
tients with non-psychotic depression from 4.6% in 1999/2000 to weight in patients with unipolar depression (Fava et al. 2009).
12.5% in 2009/2010 (Gerhard et al. 2014).
The efficacy of an augmentation of antidepressant drugs with
Augmentation with lithium
SGAs could be demonstrated in a large number of randomised
clinical trials and meta-analyses (Nelson & Papakostas 2009; There is a large body of evidence for the efficacy of an augmenta-
Kasper & Akimova 2013; Turner et al. 2014; Zhou et al. 2015). tion of antidepressants with lithium in treatment-resistant MDD
Nelson and Papakostas (2009) for instance found in their meta- based on systematic reviews and meta-analyses (Crossley & Bauer
analysis of 16 placebo-controlled randomised SGA augmentation 2007; Bauer et al. 2014; Nelson et al. 2014). A recent meta-analysis
trials representing 3480 patients significant superiority for adjunct- comparing lithium augmentation with placebo augmentation
ive aripiprazole, olanzapine, quetiapine and risperidone over could include nine RCTs representing 237 MDD patients
 INTERNATIONAL JOURNAL OF PSYCHIATRY IN CLINICAL PRACTICE 19

(Nelson et al. 2014). Altogether, the proportion of responders was serum concentration compared to control subjects not treated
46.1% in the pooled lithium group and 25.4% in the placebo with lithium (Shine et al. 2015). With regard to metabolic adverse
group. In the majority of the incorporated trials (N ¼ 7), lithium effects, McKnight et al. (2012) found in their meta-analysis that
was used as augmentor of TCAs, whereas SSRIs/SNRIs were dis- patients receiving lithium gained more weight than those receiv-
pensed as antidepressants in only three trials. However, the effect ing placebo, but lesser than those medicated with olanzapine. In
sizes were equivalent for both classes of antidepressants challeng- the aforementioned direct comparison trial with quetiapine XR,
ing the common assumption that lithium is preferably efficacious there was significantly greater weight gain under quetiapine than
in combination with TCAs. Limitations in terms of these findings lithium (Bauer et al. 2013a).
arise from the small total number of participants analyzed and
from the low methodological quality of many included individual
SGA versus lithium augmentation
trials. However, lithium is consistently recommended by treatment
guidelines as adjunctive medication in treatment-resistant MDD Taken together, an augmentation of antidepressant drugs with
(NICE 2009; APA 2010; Bauer et al. 2013b). SGAs as well as lithium is consistently recommended by treatment
The WFSBP guidelines recommend lithium serum target levels guidelines as first-line treatment option in treatment-resistant
between 0.6 and 0.8 mmol/l in the management of unipolar cases. For instance, the WFSBP guidelines for the biological treat-
depression. The dose of the ongoing antidepressant should ment of depressive disorders state that ‘augmentation with lith-
remain unchanged during the augmentation phase and lithium ium, quetiapine and aripiprazole are the best documented
should be administered for 2–4 weeks in the target dose before strategies at present’ (Bauer et al. 2013b). The effect sizes
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

the assessment of the patient’s response. In case of satisfactory obtained from meta-analyses are comparable with a pooled
symptom improvement, the add-on lithium medication should be response difference of 20.7% for lithium (Nelson et al. 2014) and
continued for at least 12 months (Bauer et al. 2013b). 14.4% for SGAs (Nelson & Papakostas 2009). However, it should be
There is a lack of sufficient data evaluating possible predictors taken into account that (1) the effect sizes for the SGAs are based
for a response to lithium in unipolar depression. In one small on a much larger sample size (N ¼ 16, n ¼ 3480) compared with
study (n ¼ 79), lithium was significantly more effective in MDD those for lithium (N ¼ 9, n ¼ 237) (Nelson & Papakostas 2009;
patients (1) with the final diagnosis of bipolar disorder, (2) with at Nelson et al. 2014) and (2) that the clinical trials investigating
least three previous MDD episodes and (3) with family history for SGAs are mainly characterised by a better methodological quality
MDD and/or bipolar disorder in first-degree relatives (Sugawara compared to those evaluating lithium. First of all, the different
et al. 2010). In summary, these findings suggest an association publication years might account for this divergence in terms of
between response to lithium and signs for bipolarity in MDD. the study quality.
Furthermore, especially patients with a high risk for suicidal To ascertain the comparative efficacy, acceptability and toler-
behaviour can benefit from adjunctive lithium as the suicide rate ability of the different augmenting strategies, direct comparison
under lithium was significantly reduced in a large number of RCTs trials (head-to-head trials) with high methodological quality are
and observational studies (Kessing et al. 2005; Cipriani et al. 2013; requested. At present, lithium augmentation was directly com-
Rombold et al. 2014; Otuyelu et al. 2015). This effect exists for uni- pared to quetiapine XR (as augmentation and as monotherapy)
polar as well as for bipolar disorder and the anti-suicidal proper- only in the open RCT of Bauer et al. (2013a) applying a non-infer-
ties of lithium appear to be independent from the antidepressive ior study design. Even if lithium was statistically not inferior to
response. quetiapine in this trial, the efficacy outcomes such as the mean
Nevertheless, augmentation with lithium is less frequently MADRS reduction and the response/remission rates were in favour
applied in the clinical routine care than for instance an augmenta- of an augmentation with quetiapine. Moreover, add-on quetiapine
tion with SGAs (Dold et al. 2015). Probably, the use of lithium in could separate significantly from lithium within the first study
the clinical practice is limited by the need of continuous plasma week suggesting a rapid onset of action of this compound in
level determinations to ensure the achievement of the therapeutic depressive disorders.
window and due to the anticipation of severe adverse effects. This From a clinical viewpoint, the augmentation of an antidepres-
assumption is mainly based on a finding of the STAR
D study sant with an antipsychotic drug is especially recommended in
whereby the occurrence of adverse effects and the attrition rate patients with psychotic depression whereas lithium should be
was significantly higher for lithium than for triiodthyronine (T3) preferably considered in MDD patients with signs for bipolarity or
(Nierenberg et al. 2006). However, this observation could not be with high risk for suicidal behaviour. Moreover, the different risk
corroborated by a RCT comparing lithium augmentation directly profiles of the different substances should be regarded when
to quetiapine XR (Bauer et al. 2013a). In this head-to-head trial, choosing an adequate augmentation strategy for an individual
both compounds were characterised by equivalent acceptability patient. As for the combination treatment, possible drug interac-
and tolerability. tions and additional adverse effects by the adjunctive compound
A large systematic review and meta-analysis of RCTs and obser- must be critically taken into account and weighed against the
vational studies comprising altogether 385 studies evaluated the potential advantages in terms of antidepressive efficacy.
risk profile of lithium (McKnight et al. 2012). The authors deter-
mined for lithium-treated patients a significantly increased risk for
Augmentation with other compounds
reduced urinary concentrating ability, hypothyroidism and hyper-
parathyroidism. Additionally, there was a non-significant trend for For the pharmacological management of treatment-resistant uni-
a diminished glomerular filtration rate indicating that lithium polar depression, various compounds were investigated as aug-
might increase the risk of renal failure. About 0.5% (18 of 3369) of mentors of antidepressants in double-blind RCTs. However, there
the whole study sample received renal replacement therapy com- is currently no convincing evidence for the efficacy of substances
pared with 0.2% in the general population. Furthermore, a retro- other than SGAs and lithium in this indication. Add-on treatment
spective analysis of laboratory data of 4678 lithium-treated with lamotrigine did not significantly differ from placebo augmen-
patients revealed a statistically significantly increased risk of stage tation in 3 double-blind RCTs (Barbosa et al. 2003; Santos et al.
three chronic kidney disease, hypothyroidism and enhanced total 2008; Barbee et al. 2011). In the study of Barbee et al. (2011),
 20 MARKUS DOLD AND SIEGFRIED KASPER

however, there was a trend suggesting that adjunctive medication AstraZeneca, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline,
with lamotrigine could be advantageous for a subgroup of Janssen, Lundbeck, Merck Sharp and Dome (MSD), Novartis,
patients suffering from very severe depressive symptoms. The effi- Organon, Pfizer, Schwabe, Sepracor, and Servier; and he has
cacy of thyroid hormone supplementation in treatment-resistant served on speakers’ bureaus for Angelini, AOP-Pharma,
depression was examined in a number of case series and clinical AstraZeneca, Bristol Myers-Squibb, Eli Lilly, Janssen, Lundbeck,
trials (Bauer et al. 2008). In summary, the findings for tri-iodthyro- Neuraxpharm, Pfizer, Pierre Fabre, Schwabe, Sepracor, Servier and
nine (T3) were inconsistent and the results of a meta-analysis Wyeth. Furthermore, Dr Kasper is chief editor of the International
were limited by a significant level of heterogeneity indicating Journal of Psychiatry in Clinical Practice.
inconclusive evidence (Aronson et al. 1996). In the STAR
D study
(level 3), T3 was compared with lithium and no significant
between-group differences with regard to the response rates References
could be observed. However, there was a significantly lower rate
Adli M, Baethge C, Heinz A, Langlitz N, Bauer M. 2005. Is dose
of adverse effects and premature study discontinuation under T3
escalation of antidepressants a rational strategy after a
treatment (Nierenberg et al. 2006). In terms of L-thyroxine (T4),
medium-dose treatment has failed? A systematic review. Eur
there are positive open studies and case series suggesting an effi-
Arch Psychiatry Clin Neurosci. 255:387–400.
cacy of administering T4 in supraphysiological doses, but an
Adli M, Pilhatsch M, Bauer M, Koberle U, Ricken R, Janssen G,
evaluation of this treatment option with adequate RCTs is cur-
Ulrich S, Bschor T. 2008. Safety of high-intensity treatment
rently missing (Connolly & Thase 2011). Augmentation of SSRIs
with the irreversible monoamine oxidase inhibitor tranylcypro-
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

with the partial 5-HT 1A receptor agonist buspirone could not sig-
mine in patients with treatment-resistant depression.
nificantly outperform placebo augmentation in two double-blind
Pharmacopsychiatry. 41:252–257.
RCTs (Landen et al. 1998; Appelberg et al. 2001). In the same way,
Amsterdam JD, Berwish NJ. 1989. High dose tranylcypromine ther-
the 5-HT1A/beta-adrenoceptor ligand pindolol was not signifi-
apy for refractory depression. Pharmacopsychiatry. 22:21–25.
cantly superior to placebo in a meta-analysis of 9 RCTs
APA. 2010. Practice guideline for the treatment of patients with
(Ballesteros & Callado 2004). However, it could be shown that the
major depressive disorder, 3rd ed. Arlington (VA): American
addition of pindolol to SSRIs leads to a faster clinical response.
Psychiatric Association (APA).
Appelberg BG, Syvalahti EK, Koskinen TE, Mehtonen OP,
Conclusion/key findings Muhonen TT, Naukkarinen HH. 2001. Patients with severe
 Augmentation of antidepressants with second-generation depression may benefit from buspirone augmentation of select-
antipsychotic drugs (SGAs) can be considered as well- ive serotonin reuptake inhibitors: results from a placebo-con-
established first-line medication in treatment-resistant uni- trolled, randomized, double-blind, placebo wash-in study. J Clin
polar depression. Quetiapine XR, aripiprazole and olanza- Psychiatry. 62:448–452.
pine (in combination with fluoxetine) are officially licensed Aronson R, Offman HJ, Joffe RT, Naylor CD. 1996. Triiodothyronine
for this indication in the United States, whereas quetiapine augmentation in the treatment of refractory depression.
XR also received the official approval for the European A meta-analysis. Arch Gen Psychiatry. 53:842–848.
Union. Balestri M, Calati R, Souery D, Kautzky A, Kasper S, Montgomery S,
 Even if augmentation with lithium is recommended by a Zohar J, Mendlewicz J, Serretti A. 2016. Socio-demographic and
number of treatment guidelines, it should be critically con- clinical predictors of treatment resistant depression: a prospect-
sidered that the evidence for this treatment option is not ive European multicenter study. J Affect Disord. 189:224–232.
comparable with the studied antipsychotic compounds Ballesteros J, Callado LF. 2004. Effectiveness of pindolol plus sero-
based on sound clinical trial characteristics and small sam- tonin uptake inhibitors in depression: a meta-analysis of early
ple sizes. Moreover, study findings suggest a later onset of and late outcomes from randomised controlled trials. J Affect
action compared with antipsychotic compounds. Disord. 79:137–147.
 If the depression is accompanied by psychotic symptoms, Barbee JG, Thompson TR, Jamhour NJ, Stewart JW, Conrad EJ,
SGAs should be preferred. Reimherr FW, Thompson PM, Shelton RC. 2011. A double-blind
 Establishing a combination treatment appears beneficial placebo-controlled trial of lamotrigine as an antidepressant
when two antidepressant compounds with a different augmentation agent in treatment-refractory unipolar depres-
receptor binding are concomitantly dispensed. sion. J Clin Psychiatry. 72:1405–1412.
 The efficacy of a dose escalation of the currently adminis- Barbosa L, Berk M, Vorster M. 2003. A double-blind, randomized,
tered antidepressant depends on the class of antidepres- placebo-controlled trial of augmentation with lamotrigine or
sants and can be recommended at best for tricyclic placebo in patients concomitantly treated with fluoxetine for
antidepressants (TCAs) and monoamine oxidase (MAO) resistant major depressive episodes. J Clin Psychiatry.
inhibitors. 64:403–407.
 Switching to another, new antidepressant cannot be gen- Bauer M, Adli M, Ricken R, Severus E, Pilhatsch M. 2014. Role of
erally advised as evidence-based treatment option. lithium augmentation in the management of major depressive
 Especially for partial responders, augmentation- and com- disorder. CNS Drugs. 28:331–342.
bination strategies should be preferred. Bauer M, Dell'Osso L, Kasper S, Pitchot W, Dencker Vansvik E,
Ko€hler J, Jørgensen L, Montgomery SA. 2013a. Extended-release
quetiapine fumarate (quetiapine XR) monotherapy and quetia-
Disclosure statement
pine XR or lithium as add-on to antidepressants in patients
Dr Dold has received a travel grant from Janssen-Cilag. Dr Kasper with treatment-resistant major depressive disorder. J Affect
has received grant/research support from Bristol Myers-Squibb, Eli Disord. 151:209–219.
Lilly, GlaxoSmithKline, Lundbeck, Organon, Pfizer, Sepracor and Bauer M, El-Khalili N, Datto C, Szamosi J, Eriksson H. 2010.
Servier; he has served as a consultant or on advisory boards for A pooled analysis of two randomised, placebo-controlled
 INTERNATIONAL JOURNAL OF PSYCHIATRY IN CLINICAL PRACTICE 21

studies of extended release quetiapine fumarate adjunctive to [in German]. Poster presented at the DGPPN congress, Berlin,
antidepressant therapy in patients with major depressive dis- Germany, 25–28.11.2015.
order. J Affect Disord. 127:19–30. Dold M, Leucht S. 2014. Pharmacotherapy of treatment-resistant
Bauer M, Goetz T, Glenn T, Whybrow PC. 2008. The thyroid-brain schizophrenia: a clinical perspective. Evid Based Ment Health.
interaction in thyroid disorders and mood disorders. 17:33–37.
J Neuroendocrinol. 20:1101–1114. Dornseif BE, Dunlop SR, Potvin JH, Wernicke JF. 1989. Effect of
Bauer M, Pfennig A, Severus E, Whybrow PC, Angst J, Moller HJ, dose escalation after low-dose fluoxetine therapy.
et al. 2013b. World Federation of Societies of Biological Psychopharmacol Bull. 25:71–79.
Psychiatry (WFSBP) guidelines for biological treatment of uni- Fava GA, Gatti A, Belaise C, Guidi J, Offidani E. 2015. Withdrawal
polar depressive disorders, Part 1: update 2013 on the acute symptoms after selective serotonin reuptake inhibitor discon-
and continuation treatment of unipolar depressive disorders. tinuation: a systematic review. Psychother Psychosom. 84:72–81.
World J Biol Psychiatry. 14:334–385. Fava M, Wisniewski SR, Thase ME, Baker RA, Tran QV, Pikalov A,
Bauer M, Severus E, Kohler S, Whybrow PC, Angst J, Moller HJ, Yang H, Marcus RN, Berman RM. 2009. Metabolic assessment of
et al. 2015. World Federation of Societies of Biological aripiprazole as adjunctive therapy in major depressive disorder:
Psychiatry (WFSBP) guidelines for biological treatment of uni- a pooled analysis of 2 studies. J Clin Psychopharmacol.
polar depressive disorders. Part 2: maintenance treatment of 29:362–367.
major depressive disorder-update 2015. World J Biol Psychiatry. Galling B, Calsina Ferrer A, Abi Zeid Daou M, Sangroula D, Hagi K,
16:76–95. Correll CU. 2015. Safety and tolerability of antidepressant co-
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

Benkert O, Szegedi A, Wetzel H, Staab HJ, Meister W, Philipp M. treatment in acute major depressive disorder: results from a
1997. Dose escalation vs. continued doses of paroxetine and systematic review and exploratory meta-analysis. Expert Opin
maprotiline: a prospective study in depressed out-patients with Drug Saf. 14:1587–1608.
inadequate treatment response. Acta Psychiatr Scand. Gardner DM, Murphy AL, O’Donnell H, Centorrino F, Baldessarini RJ.
95:288–296. 2010. International consensus study of antipsychotic dosing.
Bennabi D, Aouizerate B, El-Hage W, Doumy O, Moliere F, Courtet P, Am J Psychiatry. 167:686–693.
Nieto I, Bellivier F, Bubrovsky M, Vaiva G, et al. 2015. Risk factors Gerhard T, Akincigil A, Correll CU, Foglio NJ, Crystal S, Olfson M.
for treatment resistance in unipolar depression: a systematic 2014. National trends in second-generation antipsychotic aug-
review. J Affect Disord. 171:137–141. mentation for nonpsychotic depression. J Clin Psychiatry.
Blier P, Ward HE, Tremblay P, Laberge L, Hebert C, Bergeron R. 75:490–497.
2010. Combination of antidepressant medications from treat- Hamilton M. 1960. A rating scale for depression. J Neurol
ment initiation for major depressive disorder: a double-blind Neurosurg Psychiatry. 23:56–62.
randomized study. Am J Psychiatry. 167:281–288. Hiemke C, Baumann P, Bergemann N, Conca A, Dietmaier O,
Brecht S, Desaiah D, Marechal ES, Santini AM, Podhorna J, Guelfi JD. Egberts K, Fric M, Gerlach M, Greiner C, Gr€ under G, et al. 2011.
2011. Efficacy and safety of duloxetine 60 mg and 120 mg daily AGNP consensus guidelines for therapeutic drug monitoring in
in patients hospitalized for severe depression: a double-blind psychiatry: update 2011. Pharmacopsychiatry. 44:195–235.
randomized trial. J Clin Psychiatry. 72:1086–1094. Hieronymus F, Nilsson S, Eriksson H. 2015. A mega-analysis of
Bromet E, Andrade LH, Hwang I, Sampson NA, Alonso J, de fixed-dose trials reveals dose-dependency, a rapid onset of
Girolamo G, et al. 2011. Cross-national epidemiology of DSM-IV action and a respectable effect size for the symptom-reducing
major depressive episode. BMC Med. 9:90 effect of SSRIs in adult major depression. Poster presented at
Bschor T, Baethge C. 2010. No evidence for switching the anti- the IFMAD, Prague, Czech Republic, 02–04.12.2015.
depressant: systematic review and meta-analysis of RCTs of a Holma KM, Haukka J, Suominen K, Valtonen HM, Mantere O,
common therapeutic strategy. Acta Psychiatr Scand. Melartin TK, Sokero TP, Oquendo MA, Isomets€a ET. 2014.
121:174–179. Differences in incidence of suicide attempts between bipolar I
Bschor T, Bauer M, Adli M. 2014. Chronic and treatment resistant and II disorders and major depressive disorder. Bipolar Disord.
depression: diagnosis and stepwise therapy. Dtsch Arztebl Int. 16:652–661.
111:766–775. Kasper S, Akimova E. 2013. The role of atypical antipsychotics in
Cipriani A, Furukawa TA, Salanti G, Geddes JR, Higgins JP, inadequate-response and treatment-resistant depression. In:
Churchill R, Watanabe N, Nakagawa A, Omori IM, McGuire H, Kasper S, Montgomery S, editors. Treatment-resistant depres-
et al. 2009. Comparative efficacy and acceptability of 12 new- sion. Chichester: Wiley-Blackwell; p. 107–128.
generation antidepressants: a multiple-treatments meta-analysis. Kasper S, Montgomery S. 2013. Treatment-resistant depression.
Lancet. 373:746–758. Chichester: Wiley-Blackwell.
Cipriani A, Hawton K, Stockton S, Geddes JR. 2013. Lithium in the Kato M, Chang CM. 2013. Augmentation treatments with second-
prevention of suicide in mood disorders: updated systematic generation antipsychotics to antidepressants in treatment-resist-
review and meta-analysis. BMJ. 346:f3646. ant depression. CNS Drugs. 27:S11–S19.
Connolly KR, Thase ME. 2011. If at first you don't succeed: a review Kessing LV, Sondergard L, Kvist K, Andersen PK. 2005. Suicide risk
of the evidence for antidepressant augmentation, combination in patients treated with lithium. Arch Gen Psychiatry.
and switching strategies. Drugs. 71:43–64. 62:860–866.
Crossley NA, Bauer M. 2007. Acceleration and augmentation of Kessler RC, Berglund P, Demler O, Jin R, Koretz D, Merikangas KR,
antidepressants with lithium for depressive disorders: two Rush AJ, Walters EE, Wang PS; National Comorbidity Survey
meta-analyses of randomized, placebo-controlled trials. J Clin Replication. 2003. The epidemiology of major depressive dis-
Psychiatry. 68:935–940. order: results from the National Comorbidity Survey Replication
Dold M, Kautzky A, Bartova L, Souery D, Mendlewicz J, Serretti A, (NCS-R). JAMA. 289:3095–3105.
et al. 2015. Psychopharmakologische Behandlungsstrategien bei Konstantinidis A, Papageorgiou K, Grohmann R, Horvath A, Engel R,
unipolarer Depression an europ€aischen Universit€atskliniken – eine Kasper S. 2012. Increase of antipsychotic medication in
pharmakoepidemiologische multizentrische Querschnittsstudie depressive inpatients from 2000 to 2007: results from the
 22 MARKUS DOLD AND SIEGFRIED KASPER

AMSP international pharmacovigilance program. Int J disorder who showed minimal response to initial antidepressant
Neuropsychopharmacol. 15:449–457. therapy. Int Clin Psychopharmacol. 27:125–133.
Kovess-Masfety V, Alonso J, Angermeyer M, Bromet E, de Girolamo G, NICE. 2009. Depression in adults: the treatment and management
de Jonge P, Demyttenaere K, Florescu SE, Gruber MJ, Gureje O, of depression in adults. National Collaborating Centre for
et al. 2013. Irritable mood in adult major depressive disorder: Mental Health.
results from the world mental health surveys. Depress Nierenberg AA, Fava M, Trivedi MH, Wisniewski SR, Thase ME,
Anxiety. 30:395–406. McGrath PJ, Alpert JE, Warden D, Luther JF, Niederehe G, et al.
Landen M, Bjorling G, Agren H, Fahlen T. 1998. A randomized, 2006. A comparison of lithium and T(3) augmentation following
double-blind, placebo-controlled trial of buspirone in combin- two failed medication treatments for depression: a STAR
D
ation with an SSRI in patients with treatment-refractory depres- report. Am J Psychiatry. 163:1519–1530.
sion. J Clin Psychiatry. 59:664–668. Otuyelu E, Foldvari A, Szabo E, Sipos V, Edafiogho P, Szucs M,
Lenze EJ, Mulsant BH, Blumberger DM, Karp JF, Newcomer JW, Dome P, Rihmer Z, Sandor J. 2015. Antidepressant drugs and
Anderson SJ, et al. 2015. Efficacy, safety, and tolerability of aug- teenage suicide in Hungary: time trend and seasonality analysis.
mentation pharmacotherapy with aripiprazole for treatment-
Int J Psychiatry Clin Pract. 19:221–225.
resistant depression in late life: a randomised, double-blind,
Papakostas GI, Fava M, Baer L, Swee MB, Jaeger A, Bobo WV,
placebo-controlled trial. Lancet. 386:2404–2412.
Shelton RC. 2015. Ziprasidone augmentation of escitalopram for
Licht RW, Qvitzau S. 2002. Treatment strategies in patients with
major depressive disorder: efficacy results from a randomized,
major depression not responding to first-line sertraline treat-
double-blind, placebo-controlled study. Am J Psychiatry.
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

ment. A randomised study of extended duration of treatment,

dose increase or mianserin augmentation. Psychopharmacology 172:1251–1258.
(Berl). 161:143–151. Papakostas GI, Fava M, Thase ME. 2008. Treatment of SSRI-resistant
Lopes Rocha F, Fuzikawa C, Riera R, Ramos MG, Hara C. 2013. depression: a meta-analysis comparing within-versus across-
Antidepressant combination for major depression in incomplete class switches. Biol Psychiatry. 63:699–704.
responders – a systematic review. J Affect Disord. 144:1–6. Petersen T, Papakostas GI, Posternak MA, Kant A, Guyker WM,
McKnight RF, Adida M, Budge K, Stockton S, Goodwin GM, Geddes JR. Iosifescu DV, Yeung AS, Nierenberg AA, Fava M. 2005. Empirical
2012. Lithium toxicity profile: a systematic review and meta- testing of two models for staging antidepressant treatment
analysis. Lancet. 379:721–728. resistance. J Clin Psychopharmacol. 25:336–341.
Meyer JH, Wilson AA, Sagrati S, Hussey D, Carella A, Potter WZ, Rocha FL, Fuzikawa C, Riera R, Hara C. 2012. Combination of anti-
Ginovart N, Spencer EP, Cheok A, Houle S. 2004. Serotonin depressants in the treatment of major depressive disorder:
transporter occupancy of five selective serotonin reuptake a systematic review and meta-analysis. J Clin Psychopharmacol.
inhibitors at different doses: an [11C]DASB positron emission 32:278–281.
tomography study. Am J Psychiatry. 161:826–835. Rombold F, Lauterbach E, Felber W, Mueller-Oerlinghausen B,
Mischoulon D, Witte J, Levy M, Papakostas GI, Pet LR, Hsieh WH, Ahrens B, Bronisch T, Kilb B, Lewitzka U, Richter K, Broocks A,
Pencina MJ, Ward S, Pollack MH, Fava M. 2012. Efficacy of dose et al. 2014. Adjunctive lithium treatment in the prevention of
increase among nonresponders to low-dose aripiprazole aug- suicidal behavior in patients with depression and comorbid per-
mentation in patients with inadequate response to antidepres- sonality disorders. Int J Psychiatry Clin Pract. 18:300–303.
sant treatment: a randomized, double-blind, placebo-controlled, Romera I, Perez V, Menchon JM, Schacht A, Papen R, Neuhauser D,
efficacy trial. J Clin Psychiatry. 73:353–357. Abbar M, Svanborg P, Gilaberte I. 2012. Early switch strategy in
Mohamed S, Leslie DL, Rosenheck RA. 2009. Use of antipsychotics patients with major depressive disorder: a double-blind, random-
in the treatment of major depressive disorder in the U.S. ized study. J Clin Psychopharmacol. 32:479–486.
Department of Veterans Affairs. J Clin Psychiatry. 70:906–912. Ruhe HG, Booij J, v Weert HC, Reitsma JB, Franssen EJ, Michel MC,
Montgomery SA, Asberg M. 1979. A new depression scale Schene AH. 2009. Evidence why paroxetine dose escalation is
designed to be sensitive to change. Br J Psychiatry. not effective in major depressive disorder: a randomized con-
134:382–389. trolled trial with assessment of serotonin transporter occupancy.
Montgomery SA, Baldwin DS, Blier P, Fineberg NA, Kasper S, Lader Neuropsychopharmacology. 34:999–1010.
M, Lam RW, Lepine JP, Mo €ller HJ, Nutt DJ, et al. 2007. Which
Rush AJ, Trivedi MH, Stewart JW, Nierenberg AA, Fava M,
antidepressants have demonstrated superior efficacy? A review
Kurian BT, Warden D, Morris DW, Luther JF, Husain MM, et al.
of the evidence. Int Clin Psychopharmacol. 22:323–329.
2011. Combining medications to enhance depression outcomes
Montgomery SA, Nielsen RZ, Poulsen LH, H€aggstro €m L. 2014.
(CO-MED): acute and long-term outcomes of a single-blind
A randomised, double-blind study in adults with major depres-
randomized study. Am J Psychiatry. 168:689–701.
sive disorder with an inadequate response to a single course of
Rush AJ, Trivedi MH, Wisniewski SR, Stewart JW, Nierenberg AA,
selective serotonin reuptake inhibitor or serotonin–noradrena-
line reuptake inhibitor treatment switched to vortioxetine or Thase ME, Ritz L, Biggs MM, Warden D, Luther JF, et al. 2006.
agomelatine. Hum Psychopharmacol. 29:470–482. Bupropion-SR, sertraline, or venlafaxine-XR after failure of SSRIs
Nelson JC, Baumann P, Delucchi K, Joffe R, Katona C. 2014. A sys- for depression. N Engl J Med. 354:1231–1242.
tematic review and meta-analysis of lithium augmentation of Santos MA, Rocha FL, Hara C. 2008. Efficacy and safety of anti-
tricyclic and second generation antidepressants in major depressant augmentation with lamotrigine in patients with
depression. J Affect Disord. 168:269–275. treatment-resistant depression: a randomized, placebo-
Nelson JC, Papakostas GI. 2009. Atypical antipsychotic augmenta- controlled, double-blind study. Prim Care Companion J Clin
tion in major depressive disorder: a meta-analysis of placebo- Psychiatry. 10:187–190.
controlled randomized trials. Am J Psychiatry. 166:980–991. Schweizer E, Rynn M, Mandos LA, Demartinis N, Garcia-Espana F,
Nelson JC, Thase ME, Bellocchio EE, Rollin LM, Eudicone JM, Rickels K. 2001. The antidepressant effect of sertraline is not
McQuade RD, Marcus RN, Berman RM, Baker RA. 2012. Efficacy enhanced by dose titration: results from an outpatient clinical
of adjunctive aripiprazole in patients with major depressive trial. Int Clin Psychopharmacol. 16:137–143.
 INTERNATIONAL JOURNAL OF PSYCHIATRY IN CLINICAL PRACTICE 23

Shine B, McKnight RF, Leaver L, Geddes JR. 2015. Long-term placebo-controlled study in patients with inadequate response
effects of lithium on renal, thyroid, and parathyroid function: a to antidepressants. J Clin Psychiatry. 76:1224–1231.
retrospective analysis of laboratory data. Lancet. 386:461–468. Trivedi MH, Fava M, Wisniewski SR, Thase ME, Quitkin F, Warden D,
Souery D, Amsterdam J, de Montigny C, Lecrubier Y, Montgomery S, Ritz L, Nierenberg AA, Lebowitz BD, Biggs MM, et al. 2006a.
Lipp O, et al. 1999. Treatment resistant depression: Medication augmentation after the failure of SSRIs for depres-
methodological overview and operational criteria. Eur sion. N Engl J Med. 354:1243–1252.
Neuropsychopharmacol. 9:83–91. Trivedi MH, Rush AJ, Wisniewski SR, Nierenberg AA, Warden D,
Souery D, Oswald P, Massat I, Bailer U, Bollen J, Demyttenaere K, Ritz L, Norquist G, Howland RH, Lebowitz B, McGrath PJ, et al.
Kasper S, Lecrubier Y, Montgomery S, Serretti A, et al. 2007. 2006b. Evaluation of outcomes with citalopram for depression
Clinical factors associated with treatment resistance in major using measurement-based care in STAR
D: implications for clin-
depressive disorder: results from a European multicenter study. ical practice. Am J Psychiatry. 163:28–40.
J Clin Psychiatry. 68:1062–1070. Turner P, Kantaria R, Young AH. 2014. A systematic review and
Souery D, Papakostas GI, Trivedi MH. 2006. Treatment-resistant meta-analysis of the evidence base for add-on treatment for
depression. J Clin Psychiatry. 67:16–22. patients with major depressive disorder who have not
Souery D, Serretti A, Calati R, Oswald P, Massat I, Konstantinidis A, responded to antidepressant treatment: a European perspec-
Linotte S, Bollen J, Demyttenaere K, Kasper S, et al. 2011a. tive. J Psychopharmacol. 28:85–98.
Wang HR, Woo YS, Ahn HS, Ahn IM, Kim HJ, Bahk WM. 2015. Can
Switching antidepressant class does not improve response
atypical antipsychotic augmentation reduce subsequent treat-
or remission in treatment-resistant depression. J Clin
Downloaded by [Australian Catholic University] at 11:22 03 October 2017

ment failure more effectively among depressed patients with a

Psychopharmacol. 31:512–516.
higher degree of treatment resistance? A meta-analysis of
Souery D, Serretti A, Calati R, Oswald P, Massat I, Konstantinidis A,
randomized controlled trials. Int J Neuropsychopharmacol.
Linotte S, Kasper S, Montgomery S, Zohar J, Mendlewicz J.
18:pyv023.
2011b. Citalopram versus desipramine in treatment resistant
WHO. 2001. World Health Report 2001. Mental health – new
depression: effect of continuation or switching strategies: a understanding, new hope. Genf: World Health Organisation.
randomized open study. World J Biol Psychiatry. 12:364–375. Wiethoff K, Bauer M, Baghai TC, Moller HJ, Fisher R, Hollinde D,
Spielmans GI, Berman MI, Linardatos E, Rosenlicht NZ, Perry A, Kiermeir J, Hauth I, Laux G, Cordes J, et al. 2010. Prevalence
Tsai AC. 2013. Adjunctive atypical antipsychotic treatment for and treatment outcome in anxious versus nonanxious depres-
major depressive disorder: a meta-analysis of depression, quality sion: results from the German Algorithm Project. J Clin
of life, and safety outcomes. PLoS Med. 10:e1001403. Psychiatry. 71:1047–1054.
Sugawara H, Sakamoto K, Harada T, Ishigooka J. 2010. Predictors Wittchen HU, Jacobi F, Rehm J, Gustavsson A, Svensson M,
of efficacy in lithium augmentation for treatment-resistant Jonsson B, Olesen J, Allgulander C, Alonso J, Faravelli C, et al.
depression. J Affect Disord. 125:165–168. 2011. The size and burden of mental disorders and other disor-
Thase ME, Rush AJ. 1997. When at first you don’t succeed: sequen- ders of the brain in Europe 2010. Eur Neuropsychopharmacol.
tial strategies for antidepressant nonresponders. J Clin 21:655–679.
Psychiatry. 58:23–29. Zhou XP, Keitner GIP, Qin BM, Ravindran AVP, Bauer MP, Del
Thase ME, Youakim JM, Skuban A, Hobart M, Augustine C, Giovane CP, Zhao J, Liu Y, Fang Y, Zhang Y, Xie P. 2015.
Zhang P, McQuade RD, Carson WH, Nyilas M, Sanchez R, Atypical antipsychotic augmentation for treatment-resistant
Eriksson H. 2015. Efficacy and safety of adjunctive brexpiprazole depression: a systematic review and network meta-analysis. Int
2 mg in major depressive disorder: a phase 3, randomized, J Neuropsychopharmacol. 18:pyv060.