This is a huge question and virtually with no simple answer. Partially responsive patients often represent a mixture of biological, psychosocial, personality, addictions and even economic or legal problems. Drugs provide no real benefit then. However, generally speaking, I try to keep treatment simple and prefer multiacting drugs over combinations. What I always try to add are BDNF-related therapies (such as fitness training or yoga:-) and CBT elements.

Managing partial response or nonresponse: switching, augmentation, and combination strategies for major depressive disorder.

Papakostas GI1.

Abstract

Despite the multitude of agents approved for the treatment of major depressive disorder, approximately 50% of patients experience no response to treatment with a first-line antidepressant. Clinicians have 4 broad pharmacologic strategies to choose from for treating antidepressant nonresponders: increasing the dose of the antidepressant, switching to a different antidepressant, augmenting the treatment regimen with a nonantidepressant agent, and combining the original antidepressant with a second antidepressant. To date, the most comprehensively studied treatment strategy for nonresponse or partial response to antidepressants is augmentation with atypical antipsychotic agents, including aripiprazole, olanzapine, quetiapine, and risperidone. However, augmentation or combination with other agents such as mirtazapine, mianserin, and omega-3 fatty acids is also supported by considerable efficacy data. Lithium, desipramine, triiodothyronine, and modafinil have mixed data. While more studies are needed, agents such as bupropion, desipramine, mecamylamine, and testosterone look promising. Switching antidepressants, especially to the newer agents, including selective serotonin reuptake inhibitors, bupropion, mirtazapine, and venlafaxine, is also supported by considerable efficacy data. Clinicians should carefully reevaluate patients with major depressive disorder who are nonresponders to treatment, particularly those who have had several adequate trials. When choosing the best treatment strategy for antidepressant nonresponders, clinicians should take into account the efficacy and tolerability of treatment as well as patient preference and treatment history. Finally, the risk of potential loss of partial therapeutic benefit from the first-line antidepressant, as well as the risk of withdrawal symptoms, should be taken into account when considering switching antidepressants, while the risk of drug interactions and poor adherence should be taken into account when considering combination and augmentation treatments.

http://www.ncbi.nlm.nih.gov/pubmed/19922740

This is a huge question and virtually with no simple answer. Partially responsive patients often represent a mixture of biological, psychosocial, personality, addictions and even economic or legal problems. Drugs provide no real benefit then. However, generally speaking, I try to keep treatment simple and prefer multiacting drugs over combinations. What I always try to add are BDNF-related therapies (such as fitness training or yoga:-) and CBT elements.

My opinion is ...

When we have a response on the antidepressant then we continue with monotherapy (within first 2 weeks).

When we do not have a response within first 2 weeks then higher dose is recommended.

When you have partial response after 4-6 weeks then combine. When you do not have then switch (if highest dose was used).

Regards

This is one of the main questions about treatment.

The task is, effectively,  obtaining a result as good and as durable as possibile.

Dissapointed by therapeutic efficacy in the 80th, I have developped, as for other  intern illnesses, an algorithm leading to better results. Not taking in account patient failing to come back after the 2 first meetings, I reach 85 % excellent, 5 % good; 5% fair and under 5% total failure (did they take their medications ?). This result is as interresting as it takes into account disappearance of somatic features and psychic, not presently recongnized as depressive as well. 

This algorithm is detailed in my book in french : "La dépression et son traitement : aspects méconnus"Bernard MAROY  L'Harmatan Paris 2011. The english abstract is joined.

Basis is to start with an antidepressant, preferably tricyclic for cost sake. Dosage should be low, because the curve efficiency-dosage is bell-shaped, after a flat, infra-liminar part.  If not well tolerated (be carefull because of nocebo effect, so frequent among depressive people), it's going to be inefficient: stop it. Secondary effects are signs or herald inefficiency, except for dry mouth for amitryptiline. If well tolerated but inefficient after a delay according to emergency (one to 4 weeks), increase of 50%, untill positive or negative effect.

If no positive effect, change medication, preferably for an other family.

When effet starts, (after 2 d. to 4 w.) wait and check that every week, at least one sign improves an no one worsens. When optimum is reached, increases dosage. It's like for a rocket: ignite next stage as soon as previous one stops working. When optimal dose has been reached:3 possibilities. Perfect (i.e. absolutely no signs, no symptom, never better and no fluctuation): carry on at least 6 month for relatively recent depressions. Be aware of the possibility of a lowering of optimal dosage implying a lowering test for dosage. Dose decrease should be prudent and slow not to perform next tapering before effect of earlier has developped, and so one. In case of worsening increase of 50%, see, and wait at least 3 month before testing further tappering.

In fact, the most frequent result is a sub-optimal. If < 50%, except for very difficult cases, change medication, back to start.

Otherwise, add another medication, supposedly complementary, belonging to another family. I use frequently mianserine,venlafaxine, new antiepileptics or antipsychotics or, in difficult cases, moclobemide. As a matter of fact, effect of association fails to bee foreseable. It is not a reason not to try cautiously. Don't be afraid by reversible IMAO. In 30 y., I have never seen a serious effect. However, patient must be explained the possibility of accident and the physician must be contactable at any time. Start with a dose lower than if used alone and carry on.following above principles. Don't hesitate in difficult case, to use a third, even a forth medication.

In case o vanishing effect, a very frustrationg condition, add a stabilizer as Li, valpromide or carbamazepine.

The standards of practice discussed above fail to ask the most important question; why do antidepressant medications fail. Our research suggests that there are four primary reasons why medications fail: encephalopathy, epileptiform discharges (ED), focal slowing (FS), and spindling excessive beta (SEB).

Encephalopathy (low voltage slow EEG): If there is not enough metabolic support to the brain, medications have very limited effect.

Epileptiform discharges (ED): A recent study published in BMC Psychiatry, 1/06/2016 "Antidepressant use and risk of seizures in people aged 20 to 64 years: Cohort study using a primary care database" details the risk associated. We know that antidepressants lower seizure threshold and antipsychotics even more so. Our research suggests that those that develop epilepsy are just the tip of the iceberg of those who have ED and the medication approach to depression listed above make patients with ED brains worse.

Focal Slowing (FS): The reason the FDA approved TMS for refractory depression is because those patients have focal slowing in the left anterior temporal lobe area. Focal slowing can be from a head injury and the FDA has yet to approve a medication for the treatment of TBI because medication effect is global and focal.

Spindling Excessive Beta (SEB): European studies have found that SSRIs are contraindicated when SEB is present because of the increased risk of negative side effects.

When the first medication used to treat depression fails to work, consider that there may be a neurological abnormality that needs to be assessed with an EEG/qEEG study prior to adding on another medication.

Reference:

Swatzyna, R.J., Tarnow, J.D., Tannous, J., Schieszler, C., Pillai, V.J. & Kozlowski, G.P. (2014). EEG/QEEG Technology Identifies Neurobiomarkers Critical to Medication Selection and Treatment: A Preliminary Study. Journal of Psychology and Clinical Psychiatry.

Arns, M., Swatzyna, R.J., Gunkelman, J., & Olbrich, S. (2015). Sleep maintenance, spindling excessive beta and regulatory systems approach?  Neuropsychiatric Electrophysiology electronically published June 2015.

Swatzyna, R.J., Kozlowski, G.P. & Tarnow, J.D. (2015). Pharmaco-EEG: A Study of Individualized Medicine in Clinical Practice. Clinical EEG and Neuroscience.  Vol. 46(3) 192-196: DOI: 10.1177/15500594|4556|20

Interesting, but devoid of any practical  application.

I feel more comfortable with  try and test.

It's interesting to notice that various antidepressants belonging to the same family are devoid of cross-resistance.

We must admit that we are still failing to understand many parts of the problem. 

This is not a reason to invoke the first theory which comes, like neuro-transmitters.

We have used this information to assist in the selection of medication for the past 10 years. It has been our experience that this data has helped us avoid medication that will likely result in side effects and select medication targeted to the brain not symptoms. 

Well, there is no fast and hard rules!

Logically, you start with a monotherapy with SSRI/SNRI antidepressant as  first-line .

Adequate trial dictates full therapeutic dose (e.g. BNF doses) for adequate duration (6 weeks traditionally)

If : * no response at all (fist 2 weeks were muted)

* intolerable side-effect appears

so, I would suggest SWITCH/SHIFT- whether interclass or intraclass might be helpful unlike old assumptions of shift should be interclass!

On the other hand, if : * Partial response achieved then plateaued (for weeks)

* Residual symptoms persists (whether part of depression or meds)

then, I would OPTIMIZE/MAXIMIZE if not already done for the former and I would normally COMBINE/AUGMENT for the latter

Combination usually refers to antidepressants combos.

Augmentation traditionally refers to add-on drugs other than antidepressants to the treatment regimen.

But usually this distinction is of academic interest.

When I combine antidepressants, I try to combine agents that target more than one neurotransmitter at a time e.g. SSRI (sertraline) with NDRI (bupropion)

 Also, I try to select which neurotransmitter needs to be boosted to target specific symptom e.g. Anhedonia- needs to boost DA. fatigue needs NE and so forth.

When it comes to augmentation, I normally go for well-known combos.

It entails atypical antipsychotic (esp.  aripiprazole/quetiapine as it is FDA-approved)

, stimulants (esp. elderly frail age-group), Lithium, Cytomel (T3), dopaminomimetics, estrogen (postmenopausal women), buspirone, pindolol, folate, .... and literature abounds in treatment-resistant depression TRD strategies.

3 last point I would advise, of paramount clinical relevance:

1) when resistance or poor response is met - revise major Ds (diagnosis-disease/disorder comorbidities-duration-drug compliance,-dosage)  Always rule out/in BIPOLARITY

2) Do NOT undermine value of psychotherapy of good evidence-base esp. if high-quality and trained provider- e.g. CBT/IPT

3) Do NOT defer physical/somatic modalities as a last ditch! I mean neuromodulation like ECT can be life-saving and can be used anywhere along algorithm of treatment not as a last resort as some would erroneously claim.

The lesson we got from STAR-D and other large naturalistic studies with all these various modalities available to tackle depression is-

NEVER GIVE UP- in the battle of depression. Every further step brings about more response and with it more patients back to life.

Trust this might help.

A bit out-dated !

Nobody has, presently any belief in the neuro-transmitter theory. There modification is merely an epi-phenomenom.

However, In  agree completely with "never give up" and add "never statisfy of a partial result". "Use somatic signs like fatigue, sleep and memory disturbances, sensory abnormalities and, moreover allodynia" to monitor evolution..

Taking in account the present lack of any robust theory: be pragmatic.

Dr Naguy above gives a very common sensical mainstream psychiatry response that has an evidence base for most of his rec's.  I will only add - in keeping with a focus on confounding variables in TRD- that many treatments, including ECT, falter due to lack of recognition of Axis II factors that severely limit the response to somatic treatments.  I have seen many dysphoric patients with BPD pursue multiple aggressive treatments for unipolar or bipolar depression without success and with collusion by treating physicians who did not want to "label" the patient with a personality disorder ( which might be more treatable actually). 

I agree with Michael that BPD  often takes 10 years to diagnose based on recent studies. I approach this diagnosis from a neuropsychological perspective along with actuarial data I have gathered using the Luscher  Color test. I generally treat most of my patients with MTHFR supplements as a precaution and add in Omega-3's. If anxiety is a significant feature we add lithium carbonate 150-300mg/day and then titrate up from this platform.

The question is very interesting. I clarify my opinion and my experience in a recent free PMC article  (http://www.ncbi.nlm.nih.gov/pubmed/26425446). I hope to receve your comments to paper.

In my practice I treat  mixed bipolar depresssion with anticonvulsant and/or quetiapine, lithium carbonate could be an adjuntive therapy. In my (large) experience Omega-3 are ineffective in mixed and non-mixed bipolar depression.

Dear Edward

Interesting. I

n fact, it directs to the most frequently efficient switches.

Pragmatically, the solution remain try and test.

It does'nt address associations.

Antonio, I agree with you. I  have never noticed the slightest effect of omega 3 on depression.

There are many algorithms in this respect the most convenient one is the Star-d

you can review it on this link

http://www.ncbi.nlm.nih.gov/core/lw/2.0/html/tileshop_pmc/tileshop_pmc_inline.html?title=Click%20on%20image%20to%20zoom&p=PMC3&id=1764539_i1523-5998-8-5-291-f03.jpg

Thanks to all of you for a nice discussion, which has been seen on this very important topic.

1) Bipolar depression should be calculated

more than 10-20 % od depression are bipolar and therefore after 2-3 unsuccessful treatment with antidepressant with at least of 6-8 weeks of treatment (or even part response, e.g.10-20 % Hamilton-D-17) a use of anticonvulsant  and/or antipsychotic is very good option (especially this with efficacy proven for bipolar depression, e.g. quetiapine, lithium).

2) MAO-inhibitors 

Very few patients are treated with MAO-inhibitors, because of different things. In my point of view these drugs are underprescribed and are not first line according to the guidelines (e.g. APA guidelines), although MAO-inhibitors are very effective in several cases.

3) Wait, wait, wait...

Clinicians are usually not very good in waiting, although in some cases waiting is the best treatment option (also in the treatment with antipsychotics, e.g.clozapine). Usually many clinicians do not wait until 2 weeks of monotherapy but they switch very early in the treatment, which often means huge polypharmacy, which has not been proven in RCTs (DDIs should be also calculated)

4) Drug adherence

Drug adherence should be checked before prescribing polypharmacy. Often patients with mood disorders do not take their drugs regularly, although it it necessary (more than 50 % of patients discontinue within first 6 months)! This is very good reason why we should check patients' drug adherence.

5) DDIs

Many DDIs are important in term of drug efficacy and safety. Also many antidepressants have important DDIs and often these drugs are not appropriate in term of safety (more than 1/4 of patients within first 5 months), because of DDIs. Here a cooperation with clinical pharmacist is beneficial for patients.

Regards,

Matej

Never confuse real bipolar (rare) illness, with usual fluctuations of common "endogenous" depression. Never "pull the carpet under the feet" of the patient, enjoying, endly, a better state than ever.

Never, cut down antidepressant, unless PROVEN hypomania (never seen during 35 y. practice). Even worse, never prescribe in these conditions a chemical straightjacket. NEVER.

The key point why there are treatment resistant cases or, alternatively, that a working medication therapy fails with time, is that the SERT transporter actually not only transports back to vesicles part of the released serotonin but also the newly synthesized sertotonin located near on the side of the place of the synaptic cleft. Thus, as many practitioners have may already observed, an antidepressant works, or if it not works, it works for some time with an increase of dosage, then it again stops working, then a change or a combination (SSRI + NRI/DRI, indirect serotoninergic/noadrenergic agonism) works; drug-free periods could help; nonetheless, in the end, time gives its results: the real point is that when antidepressants are not working, it is because there is too little stored serotonin in the synaptic clefts and/or reduced subsequently serotonin release due to overly increased Serotonin reuptake inhibition. The only solution thus is to lower SERT inhibition by SSRI/SNRI and to increase serotonin synthesis by adding natural 5HTP + peripheral synthesis inhibitor for a probe period, then to increase SSRI/SNRI dosage again and to alternate this regime for the whole treatment period. Noradrenaline (alpha-1-Rs) activates raphe nuclei, as does dopamine (D2Rs) as a supplemental activator of the raphe nuclei. GABA-B and GABA-A decrease serotonin release of serotonin terminals but valproate, a GABA-ergic drug, is an indirect 5-HT1A agonist. Glutamate activates both activating and deactivating currents to the raphe nuclei with net effect increase in serotonin release. NMDA-antagonism, as autoreceptor antagonism, should in theory increase serotonin release via glutamate release increase (ketamine).