Saturday, October 31, 2015

UW 3rd Annual Update - Treatment Resistant Depression

There were two presentations relevant to depression that were given at the UW conference this year.  The first was from Karen Dineen Wagner, MD, PhD from the University of Texas Medical Branch in Galveston, Texas.  Her message was a mix of the old and the new.  The old is the state of pharmacology of depressed children seems to have changed very little over the past 20 years.  This seem largely due to the fact that there have been few successful antidepressant trials in children.  This has led to the state where there are only two FDA approved medications fluoxetine and escitalopram based on a total of 4 clinical trials.  She  showed an additional 14 clinical trials of typical antidepressants including 3 that were positive for citalopram and sertraline but an additional negative study for the FDA approved medication escitalopram.  The difficulty in many of these trials is a high placebo response rate in the trials (40% greater than in adult clinical trials).  She recommended an informed consent approach explaining to the parents any time an off label approach was being used and the rationale for using any medication based approach.  She also recommended starting with the FDA approved medications for pediatric depression.

Her suggested approach to depression in children and adolescents is to start out with an FDA approved SSRI plus cognitive behavior therapy (CBT).  This is the most evidence based approach with the evidence rapidly disappearing at subsequent levels where the usual augmentation and substitution steps that are typically used in adults were suggested.  The Treatment for Adolescents with Depression (TADS) study was presented with the recovery rates for fluoxetine, fluoxetine + CBT, and CBT alone at 12, 18, and 36 weeks were presented.  The fluoxetine + CBT arm had superior results at 12 and 18 weeks but at 36 weeks the recovery rates were similar at 86% versus 81%.  Those are good results for any antidepressant trial and the placebo response rate in this study was more similar to the adult placebo response rate.  The results of this study were presented as a rationale for using antidepressants in adolescents with severe depression and/or suicidal ideation since the response rate for fluoxetine + CBT were faster than fluoxetine or CBT alone at 12 and 18 weeks and essentially the same at 18 and 36 weeks.

The issue of strategies for addressing SSRI resistant depression were presented in the form of a previous trial where 334 adolescents with SSRI treatment failures were randomized to a different SSRI or venlafaxine or SSRI + CBT or venlafaxine + CBT.  The trial done by Brent, et al showed that there was no difference in response rates switching to another SSRI or venlafaxine but switching antidepressants and adding CBT produced superior results.  Sides effects were greater for the venlafaxine arm with a slight increase in diastolic blood pressure and heart rate and a four fold increase in skin rashes - a complication that I have rarely seen in adults.  The overall impression was that CBT was the most effective intervention for adolescent depression but I am sure that most psychiatrists in the crowd were left wondering: "If I can't find CBT therapists for my adult patients with depression - what are the odds I can find them for my adolescent patients?  To me that has always been the critical shortage in psychiatry - not the number of people who can prescribe medications.

Others trials of medical interventions (omega-3 fatty acids, ECT, TMS, bright light therapy), psychotherapies (Interpersonal Therapy(IPT), family based IPT), and exercise were sparse.  Computer-based CBT has always been an underutilized modality and it showed that there were similar response rates between treatment-as-usual and an interactive fantasy based CBT called SPARX (Smart, Positive, Active, Realistic, X-factor thoughts).  In the game the child chooses an avatar and the goal is to restore balance in a fantasy world dominated by GNATS (Gloomy Negative Automatic Thoughts).  The SPARX game is available free online to residents of New Zealand.  New Zealand and Australia have been pioneers in the area of online CBT.  To find resources just Google "SPARX virtual therapy for depression".

Paul Holtzheimer, MD provided the adult perspective in the topic Management of Treatment Resistant Depression in Adults.  He made the epidemiological point that treatment resistant depression (TRD) is present in 10-33% of patients with major depressive disorder and in the U.S. that is about 1-3% of the population.  He had a fairly comprehensive agenda covering pharmacotherapy and augmentation strategies, electroconvulsive therapy, more recent non-invasive electromagnetic therapies and deep brain stimulation.  There was nothing new on the medication front.  After reviewing the basic medication groups, he suggested that the newest antidepressants offered no advantage over earlier medication.  He suggested that monoamine oxidase inhibitors (MAOIs) were being underutilized as a treatment for depression unresponsive to standard agents.  In the moderated discussion Ned Kalin, MD - the head of the department of psychiatry at the University of Wisconsin agreed.  The speaker said that he typically used phenelzine and tranylcypromine.  I personally have not prescribed either of these agents in some time.  I recall using them in situations where the person has treatment resistant depression and did not have any responders.  In those situations, response rates tend to be low anyway.  The other problem is that you have to think that your chronically depressed patient is going to be motivated and cognitively intact enough to adhere to the necessary diet, report what could be significant side effects and not try to kill themselves with the medication.  During the discussion there was a report of one patient who decided to eat high tyramine content food (prohibited on this diet due to a the risk of a hypertensive reaction) - have a stroke and die.  The patient in this case did have a stroke but did not die.  I personally know of situations where strokes have occurred, so this strategy is not without risk.

The augmentation strategies discussed were right out of STAR*D with the exception of using atypical antipsychotics with antidepressants.  Dr. Holtzheimer said that this was probably the most common augmentation strategy and the risks were discussed.  He and Dr. Kalin were advocates of augmentation with lithium and triiodothyronine (T3).  There were three slides on STAR*D showing cumulative remission and remission rates across all levels of care.  Those rates were 33% with initial monotherapy and 66% after 4 treatments and as expected less remission rates at each level of treatment change.  Dr. Holtzheimer made the point that the current rates of remission with medication and psychotherapy have really not changed since the 1950s and that makes electroconvulsive therapy (ECT) the most effective antidepressant treatment with a 50-75% remission rates and a >50% relapse rate in the first 6 months.  He touched on novel pharmacological agents categorized by neurotransmitter, neuroendocrine, or immunological systems.  He did not say much about ketamine (there is an intranasal preparation in clinical trials right now) but did mention that there is a IL-6 (cytokine) antibody trial going on right now.

He moved on to talk about more invasive therapies.  He presented a graphic that was a drawing by Papez.  To anyone trained in neuroanatomy around the time I was in medical school, many anatomy professors would present a saggital section of the brain and refer to the limbic structures as the Papez circuit.  At first I thought the drawing had a surprising amount of detail for a 1937 publication but then I went to the original article online (AMA web site) and found that the original drawing was not used.  The 1937 drawing had the surface anatomy correct but no tracts.  Papez mentions the amygdala three times in the last few paragraphs of his article but does not label it in the drawing.  Dr. Holtzheimer used this slide as a prelude to an article by Mayberg (3) providing a rational for deep brain stimulation as treatment for depression.  I plan to come up with a separate post in this technology based on several sources but right now there are a number of centers looking a deep brain stimulation for depression and addiction.  Dr. Holtzheimer briefly commented on transmagnetic stimulation (TMS).  There are apparently 4 FDA approved devices, use is expanding and insurance reimbursement is expanding.  He said it was 50% effective for treatment resistant depression.  I am highly skeptical of that number based on the people I see, but I also realize that I am seeing a highly treatment resistant with multiple comorbidities.  Seizure risk was listed as the most significant side effect.

Vagus nerve stimulation (VNS) has been around for about a decade.  I have seen a few of these patients and never referred anyone for placement of this device.  There is limited third part reimbursement and in my opinion waning enthusiasm for this technology.  The last time I interviewed a person with VNS, their speech quality changed every time the stimulator was active.  That is a significant side effect and I don't know if that has been addressed with current technology.   Transcranial direct current stimulation (tDCS), transcutaneous vagus nerve stimulation, and cranial electrical stimulation were all listed as having limited data.

Deep brain stimulation (DBS) was clearly the main focus of Dr. Holtzheimer's presentation.  The first article suggesting that it may be effective for obsessive compulsive disorder (OCD) was in the Lancet in 1999.  Based on that research DBS of the anterior internal capsule is an FDA approved indication for DBS.  An open label study suggested that it may also be effective for TRD and there were no adverse effects or neuropsychological effects.  Three additional pilot studies of DBS to the nucleus accumbens suggested that it may be useful for TRD and features of TRD like anxiety and anhedonia.  Since then there have been two randomized controlled trials of DBS to the ventral striatum subcallosal cingulate gyrus (SCC).  The first study (ventral striatum) was negative and the second (SCC) was stopped after a futility analysis.

The overall conclusion had to be that TRD was still a common and disabling condition.  The mainstays of treatment at this point are still the medications and ECT that we have had throughout my career.  My experience is that I can help most people get well, but there are significant obstacles even to standard care.  Every lecturer here emphasized the utility of cognitive behavioral therapy.  Like most psychiatrists, I can do cognitive behavioral therapy but by myself I can't meet the demand.  The people responsible for mental health policy and insurance standards certainly do not want to fund the recommended research courses of CBT for chronic depression.  There is no distinction for TRD versus non-TRD depression and no differential resource allocation.  That leaves most patients with TRD and non-TRD depression looking for "prescribers" who can see them for 10-30 minute appointments to get advice on how to recover and try various prescriptions.  None of the available care matches what top researchers recommend in these CME seminars, in articles, or in books.

We could do a lot better trying to live up to that standard while additional diagnostic and treatment strategies are developed.          

George Dawson, MD, DFAPA


1:  David Brent Adolescent depression references  

2:  Papez JW. A proposed mechanism of emotion. 1937. J Neuropsychiatry ClinNeurosci. 1995          Winter;7(1):103-12. PubMed PMID: 7711480.

3: Mayberg HS. Targeted electrode-based modulation of neural circuits for depression. J Clin Invest. 2009 Apr;119(4):717-25. doi: 10.1172/JCI38454. Review. PubMed PMID: 19339763


  1. In my experience a very high percentage of the people who do these studies, at least in adults, make almost no effort to differentiate dythymia from major depression by: 1) Not spending any time making certain that patients understand the pervasiveness and persistence critieria that differentiate the symptoms of the two disorders; and 2) Not taking a complete biopsychosocial history to distinguish psychological from limbic system factors. Furthermore, in Contract Research Organizations, experimenters get paid if they recruit a subject, and subjects get paid if they get recruited - giving a financial incentive for everyone to exaggerate symptoms in order to qualify.

    And of course, people with suicidal ideation, comorbid conditions, and significant personality pathology are excluded from studies.

    Garbage in, garbage out. CBT is worthless for someone in a melancholic depression, just as anti-depressants are usually worthless in clear-cut dysthymia.

    1. I agree with you on the melancholia. Some of my favorite neuroendocrine research was done by Mendlewicz. He would focus research only on people with the severest forms of depression. I have posted some of this here. There is a new article that looks at the naturalistic treatment of melancholia vs. atypical and typical forms of depression. The melancholia depression scores are much higher at baseline and the outcomes are worse in terms of remission.

      The Mayo group is collecting high quality data form both mood disorders and addictions. They are also developing an international collaborative database that is necessary for genetic/genomics studies.

    2. Which is why you have 50% placebo responses and 60% drug responses in industry sponsored SSRI trials.

      The high prevalence of TRD is in effect an admission that our drug treatments aren't very good.

      I agree with Dr. Allen it all starts with definitions. It's not coincidence that research went into a downward spiral in 1980.

  2. I think there should have been more in depth commentary on rTMS. The studies available vouching for it's efficacy continues to mount, and it is leaps and bounds more feasible than DBS. It does not require anesthesia or surgery or hospital based recovery facilities. The risk for seizure is near equivalent to general population risk. There are no cognitive side effects like in ECT. For rTMS as a research tool, it is highly versatile. Many studies have tinkered with dosing parameters including frequency, duration, and location and have found differences in effect. rTMS can also be image guided with MRI in order to localize the target within the millimetre. Some reviews have reported that the evidence is not compelling enough to influence clinical practice outcomes, but honestly, I have a hard time seeing how it could be less likely to work than your third or fourth medication trial.

    ECT is the most effective option for TRD so far, but there are also a variety of reasons why patients won't or cannot get ECT, especially dependent on local politics and financial factors. I recently talked to a clinical researcher who had seen DBS study protocols that did not exclude patients who did not have a trial of ECT. With the kind of evidence available for ECT, can depression really be considered treatment resistant without a trial of ECT?

    1. You make excellent points. I intend to have a separate post in the future on TRD. It is disappointing that in an era of intermediate phenotypes, endophenotypes, and GWAS scanning for markers in the heterogeneity of depression that there is no consensus on simple definitions. I also have concerns about including patients in DBS protocols who do not meet strict criteria. My concern about TMS is the fact that it is like QEEG of two decades ago, it is basically a research protocol that many patients are paying a great deal out of pocket for. You are also correct about ECT availability and the relevant factors but at this point there is much more data and experience, but if you read 10 different definitions of TRD - not one will include a trial of ECT.

    2. My take on that lack of definitions on heterogeneity is similar to Dr. Allen's - all of the fancy biological research is not being complemented by good old fashioned clinical phenotyping.To me, there are certain concepts that have been validated time and time again that I cannot figure out why do not get (re) adopted into practice and research - in particular melancholia and psychotic depression.

      I think TMS has advanced further than you may be familiar with. It is an FDA and and Health Canada approved treatment, which I would assume takes it a step above a research protocol. Again, even though some argue that the effect size is less than satisfying, I think it is worth pursuing instead of going down an assembly line of medications.

      (see: and

      Also, some studies have emerged to guide nuances of treatment. TMS is probably more effective for non-psychotic depression, and ECT for psychotic depression. TMS is probably more effective for less severe treatment resistant depression. (see and

      I look forward to your future post on TRD. If you are going to include discussion about DBS, consider the emerging literature on deep TMS. This targets centres deeper into the brain, which to me sounds similar to DBS. (see