Wednesday, January 13, 2010

Antidepressants only effective in severest depression?

A recent article in JAMA by Fournier et al. is a meta-analysis of antidepressant treatment effects assessed in relation to depression severity. Here's the reference:

The results show that antidepressants work significantly well, compared to placebo, only for very severe depression (corresponding to Hamilton Depression Rating Scale scores of at least 25).

The analysis is quite well-done, and the results are also presented in a graphical form clearly showing a linear increase in antidepressant effect as baseline depression scores increase.

The authors observe that antidepressants are most commonly prescribed to people who have milder depressions--a population in which they show that medications arguably do not work.

Here are a few of my criticisms of this study:

1) the duration of each trial included in the meta-analysis was between 6 and 11 weeks. In my opinion, depressive disorders are long-term, highly recurrent problems, which have a natural period over at least 6-11 months, not 6-11 weeks. Treatments to address mood disorders of any severity require much longer durations. The short duration could cause a significant under-estimation of treatment effects.

2) the study, like many, looks at "depression alone." In most real-life situations, outside of a research study, individuals have several different problems, such as mild depression + social anxiety, or mild depression + panic attacks, etc. The presence of other symptoms, particularly anxiety symptoms, most likely would increase the likelihood of antidepressants helping.

3) Milder depressions, just like more severe depressions, may actually improve more consistently with a "second step" such as combination with psychotherapy, or combining two different antidepressants. The mildness of a medical syndrome does not necessarily mean that the effective treatments need only to be "mild."

4) Milder depressive syndromes may be more prone to misdiagnosis.

5) current "resolution" to measure treatment effects in depression is quite poor. "Depression" is a very broad category. An analogy could be considering "abdominal pain" to be a diagnostic category. If "abdominal pain" is the only category, and is simply rated on a severity scale (rather than subcategorized to obtain a precise diagnosis), and the treatment offered for "abdominal pain" is appendectomy, then we would probably see no difference in treatment effectiveness between appendectomy and placebo. This is because appendectomy is only effective to treat appendicitis (a subset of the abdominal pain population), and is either ineffective or harmful in treating abdominal pain patients without appendicitis (except, perhaps, for those patients who have a placebo improvement of psychosomatic or factitious abdominal pain, an improvement which they attribute to having surgery).

We currently do not have the science to subcategorize depression in a more clinically meaningful way (there are subcategorization schemes, but they don't have much relevance in terms of treatment).

But we do have a research method which could improve "resolution":
-instead of comparing two populations of depressed individuals, one group receiving antidepressant (or some other treatment), and the other receiving placebo (or some other alternative), the study design could instead be to offer every individual courses of placebo, alternating with antidepressant (or "treatment one" alternating with "treatment two"). Each course of treatment would have to last an adequate length of time. The analysis would aim to show whether there is a subset of individuals who respond to the antidepressant, or a subset of individuals who do better with placebo. The averaged results over the whole group might show that antidepressant effects do not differ from placebo (just like appendectomy might not differ from placebo in treating "abdominal pain"), but the individualized result could show that some individuals improve substantially with the antidepressant (just like appendectomy would save the lives of the small group of "abdominal pain" patients who have appendicitis).


In the meantime, though, I think it is reasonable to recognize that antidepressants are less consistently helpful when symptoms are less severe.


Anonymous said...


A couple of my own thoughts: (highly subjective and possibly incoherent)

1) I think depression manifests differently in individuals and cultures/populations. The exclusion of journal articles focusing on specific populations and in different languages yields relative salience to the North American/Western European classification/definition of depression.
2) Furthermore this analysis only looked at one single scale (rating) of depression, the HRSD. This may limit the study's external validity. (Especially when trying to extrapolate this information and applying it to severity ratings rising from other depression scales.)

Not to mention the actual limitations of the HRSD as with any other objective scales. However I won’t get into this topic because every scale or objective measure has it’s weaknesses and although researchers and clinicians constantly struggle to be objective, transference and bias is unavoidable (and perhaps could be encouraged in some practices.) I guess the best we can hope for is to minimized subjectivity and acknowledge the constraints and confines of our research.
On the other hand, I praise the substantial and beneficial research/work/time and effort sacrificed to create a semi-reliable test for depression. I acknowledge the difficulties from content and construct validity to inter-rater and test-retest reliability.

My inclination is antidepressant effects will vary depending on the form and content of the depression SCALE utilized. I wonder what the outcome would be if another meta-analysis was down using a different scale-?

3) Why did the analysis only include placebo-controlled studies? (Many studies compare one antidepressant to the antidepressant “gold standard”. These studies could have been used because the analysis only really needed a Hamilton score and the presence of an antidepressant treatment… correct?) This eliminated hundreds of other studies, which may have resulted in completely different results.

Please note: I am not disputing the findings of the study. The findings are relevant to the population explored. I am only commenting on the studies external validity.

4) Also I didn’t find anywhere in the article the exclusion criteria for undergoing psychotherapy concurrently with antidepressant drug therapy. How did they control for this variable?

Anonymous said...


5) A note on severity:
I think it is important to separate severity from distress. They are two independent variables. (However, I do acknowledge that there may be situations, which may arise where they may be similar via mediation or moderation through a common third variable.)

Severity is used as a dimension on which to classify depression in a very objective manor. (For example the higher the numerical score on an HRSD the more intense the severity. Note also--the cut off points between mild/moderate and severe are quite arbitrary.)

However, distress is subjective and highly individual. There may be individuals who have a high HRDS score but may not be very distressed (because they have good coping skills, a lot of family support or unfortunately may be used to functioning at a fairly high level of distress…etc) or there may be individuals who are very distressed but who only have a low HRDS score. My point here is I believe (although I could be wrong) that antidepressants are prescribed on the basis of subjective distress more often then the score on a depression scale. (Remember the Hamilton scale is subject to personal bias of the rater…such as acquiescence biases, moderacy and extremity biases, social desirability responding problems, and the reference group effect problems… to name a few.) Had the studies included Hamilton scales of more than one rater?

I guess this study would assume that the raters were accurately collecting collateral information from the patient’s family and the patient’s subjective experience. What about the rater’s (or “pharmacotherapist” in this case…by the way the title of “pharmacotherapist” makes me question credential) overall intuition?

It may also be important to think about the personal opinion on drug therapy, of the rater, affecting the Hamilton score. Professional opinions and experiences could bias the results.

I may also be important to think about the opinion of the patient toward drug therapy and the opinions from their significant others, which may cause the patients to minimize or exaggerate effects.
Think about the impact on the Hamilton scale if someone significant to a patient, told them that antidepressants are “miracle drugs”… Or visa versa…. and how this could change patient symptoms.

The patient’s belief in the positive effects of anti depressants could be responsible for this response alone. Another interesting study… How do positive or negative beliefs about antidepressant therapy in raters, patients or the significant others of patients impact Hamilton rating scores over a certain amount of time.

Anyway---Just because a score on a depression scale indicates that you are mildly depressed does not mean that your level of distress is mild, nor, as was mentioned, that the treatment should be mild.

Anonymous said...


6) Depression represents a disorder of equifinality. Yet the foundation of categorization somewhat limited to its physical manifestations at time of presentation…i.e.: behaviors or outward expressions of emotion NOT on cause. (Which is understandable. However, just because categorization is not focused on causality does NOT mean that treatment should also NOT be based on causation.)

One man who suffers from depression due to the sudden tragic death of his wife and three kids would (hopefully) be treated very differently than another person who has had underlying chronic depression his whole life and has made no long lasting relationships with anyone. Yet their scale depression rates may be the same. Is one going to benefit more from antidepressants than the other? Probably. Is one going to benefit more from psychotherapy? Probably.

The other issue I wanted to bring up was the common western, allopathic ideology that I keep observing…. The idea that because the brain is a physical entity and depression is a disorder in the brain, depression can be “fixed” (akin to a mechanical problem) by using another physical entity, such as antidepressants.

I can understand how this thinking has evolved. Just examine the plethora of other drugs we administer to other parts of the body and the quick positive results observed. In other areas of medical research this type of thinking has and will continue to save countless lives.

Research has helped us identify certain neurotransmitter levels that seem to be different in certain psychological states. It has also helped us identify certain brain regions or structures that seem to be different in size or activity in people suffering from various mental health problems.

However, we are miles away from understanding the brain. It is much more complex than any other organ. There is a huge gap in knowledge between the physicality of neurotransmission and the creation of abstract ideas, cognitions, feelings or emotions. How does the neurotransmission mechanism or circuit differ in--- thoughts associated with suicide versus -----thoughts associated with remembering to brush your teeth? I don’t know…

In this way antidepressants may be most effective at relieving the somatic and behavioral symptoms or observable behaviors because antidepressants are only trying to solve THIS problem. This problem being the functional, physical problem originating from the structural cell level and resulting in a behavioral change at the organismal level. Antidepressants are only trying to solve the definition of depression that we have created and the symptoms that we have focused on as problematic

From this single perspective (Observable behavioral change means depression relief) negates the value of cognitive and interpretive changes.

For example: The same perspective that cardiac research uses.
I.e. a certain calcium channel blocker will affect the action potential of the heart and therefore have a measurable effect on and ECG.
This perspective is fixed in the physical, tangible, and phenomenal world and although it has offered a lot of success for other medical research it may not be the POV or perspective needed to solve mental health problems.

If the questions in the HRSD were split into somatic, cognitive, emotional, and behavioral, I would be interested to see if antidepressants had an equal affect on each category of questions over the treatment course/interval..?

As I re-read what I just typed… I am not disputing the power of antidepressants. They can be very helpful for some people. Even for the people who don’t have a great response usually experience small benefits that enable them a certain amount of “breathing room” which could be encapsulated in time or energy… to find other effective ways of coping.

Anonymous said...

7) Lastly… I believe (well, am at least starting to…) that antidepressants don’t/won’t change the content of depression just the form/intensity/or context. Which, I won’t deny, is incredibly important because it may allow an individual to address the content of their depression.

This idea makes sense intuitively because scientist/ clinicians/practitioners don’t know the origin of thought and thought content or how one thought differs from another

This is where psychotherapy may be useful. Sort of a top down approach—which I will describe here…

The metaphor:
If we were to think of mood as a compass (i.e.: north could be good mood “up”, south could be “down” or low mood) and the little triangle pointer made up of the components of mood, this is how the triangle could be divided.

-The base would be made up individual neurons and their supportive glial cells. (astrocytes, oligodendrocytes, ependymal cells.. etc..)
-Then the next level above that would be small networks/neural connections or circuits made up a distinct subset of the layer below (akin to a neighborhood).
-The next level above that would be larger, more complex and complicated connections, spreading or expanding horizontally or vertically throughout the brain. Spanning different brain regions and structures (made up of a few select combinations of the smaller networks in the level below).
-And finally on the top would be the resultant thought, cognition, emotion..etc. derived from activity of one of those pathways or tracks.

Anti- depressants would favor a bottom up approach in depression and psychotherapy a top down approach.
As you can see from this metaphor, the neurons that get “tickled” by antidepressants on the bottom may or may not be involved in the specific track which is responsible for cognitive, emotional or thought change at a higher level of the pyramid. Therefore some neurons that we don’t necessarily need to expose to the antidepressant are exposed and side effects result--- almost like redundant treatment of some neurons due to the unspecificity of the drug.

However a top down approach… implored by psychotherapy can actually target the specific neurons in the specific track resulting in a specific thought process.

I guess it is almost like asking the easiest way to turn a sailboat through the wind? Should you tack or gybe?.... Would you swing the bow (the front tip, which is light and mobile) through the wind with relative ease and start a new course and direction… or would you try to swing the stern (the back of the boat with the heavy propeller and heavy dangerous boom of the mainsail) through the wind?... I would hope many of us would choose to tack!

Sometimes I think people forget that we DO change our brains in numerous ways every day, all the time. One professor made this apparent to me when he explain how he had changed my brain since the first class of semester because I could remember my professor’s name. Somehow I had changed or made a new neural circuit or connection in my brain and stored his name in some cellular connection. And better yet… he made me realize he had done this from 40 feet away at the front of a lecture had.

Anonymous said...


There IS scientific research in support of this concept. For example there are dynamic brain imaging techniques that show after a course or specific psychotherapy treatment activity levels in certain parts of the brain are changed. Not changed permanently but changed nonetheless. This way of thinking has to be practiced and protected. Just like the development of a new language has to be practiced or else you will forget it. Just like antidepressant therapy usually has to be continued until after the individual feels better. Just like weightlifting has to be continued on a maintenance schedule/ regime long after muscle acquisition in order to avoid atrophy. The great thing here is that the process of maintaining is usually not as hard as changing…. (Although I know some people would argue that maintenance is the hardest, which may also be true in certain circumstances.)

It is the idea of “use if or lose it” due to the plasticity or our brains and our bodies. And I am discovering the sooner you try to change things the better because plasticity does decline with age. (But don’t worry… engaging and using the plasticity at any age will itself stall the decline. So…. I guess… a good incentive to try something new everyday!

Referring back to my triangle analogy the top down treatment for depression (i.e.: psychotherapy) may be more efficient than anti-depressants….Possibly…? Using antidepressants you have to expose specific individual neurons with the right dose at the right time. Sort of a shotgun approach compared to psychotherapy. However, I do acknowledge if we were to compare current anti-depressant therapy to other psycho- active drugs in previous decades, our ability to increase receptor/cell specificity has increased drastically.

However psychotherapy allows you to target the intangible, the noumenal (i.e. the thought) and therefore directly changes the neural circuits and the individual neurons responsible for making that thought. (Here I guess you could say you would get “more bang for you buck”.)

8) Finally—and yes it is actually finally, I have one further analogy for psychotherapy verses antidepressants. Suppose you needed to guess the title for a certain novel which you had never read and all you knew was that the title had to be composed of 5 words and a total of 14 letters.
In addition, your friend knows the title but is not able to share any information except whether your guess is was right or wrong.

Would you:
1) Randomly go through each of the 5 words, trying every combination of the 26 English language letters in each letter position hoping that the words would make sense and somehow remarkably relate to the content in the book?

2) Read the book and then try to guess the title?

Option one would be analogous to antidepressant therapy. A solution based in form or mechanics, the phenomenal. Option two would be analogous to psychotherapy. Content focuses and based on the noumenal. And the friend… well I guess they could be the therapist but… I don’t think they have all the answers nor have they read “your book.”

This is a little drastic I acknowledge and I guess really only applies to individuals who have depression arising out of strenuous circumstances… but I just thought it was fun.

Anyway- that’s enough of my rambling. I don’t even know if any of it makes sense anymore!


Anonymous said...

Something to think about....

Severity of depression associated with strength of positive response in psychotherapy

Also-- something else to think about:

The above study: meta analysis that found a small positive association between pharmacotherapy and depression ratings over psychotherapy and depression ratings.

However then the same authors came out with this article:

"Comparing psychotherapy and pharmacotherapy for adult depression: adjusting for differential dropout rates"

Conclusion in abstract:

"After controlling for dropout from studies, we found no indication for a significant difference between pharmacotherapy and psychotherapy.

These analyses do not result in direct evidence that there is no significant difference, and it is possible that pharmacotherapy yields lower depression scores while the number of responders is equal in both treatments."

(Perhaps the above paragraph is just damage control..)

"It does, however, suggest that differences in effect sizes found in studies comparing the 2 treatments may be influenced by the lower dropout rates with psychotherapy. Future meta-analyses should examine dropout rates more carefully."

SO-- What about the drop out rates in this study?

Perhaps the worse the depression the more willing a person is going to be to cope with the side effects of the medications. Therefore the drop out rates will be lower. Therefore the effects will be more robust?

Or perhaps there is a bias towards viewing more severe depression as endogenously caused? Therefore more people who have "worse" depression are more likely to be given medication to begin with? Or more encouraged to stay on them?

Just things to ponder.

GK said...

The analogy here I can think of is of a person trying to move a heavy boulder to the top of hill. How can this be done? One strategy would be to push the boulder up the hill, or perhaps to enlist a group of helpers to share the load. Another would be to find a horse, hook it up to a wagon, load the boulder into a wagon, and lead the horse up the hill. Another might be to rent a helicopter to carry the boulder up.

Is any one technique better than another? Depends on how you look at it. Perhaps the experience of pushing the boulder up yourself would get you in better shape. Then again, it might injure your back. The horse or helicopter might introduce risk of certain types of accidents.

In general, it does not surprise me much that the various therapies for depression have similar effectiveness, if large meta-analytic studies are done. The value of these studies, in my opinion, is to persuade us not to be biased in choosing a particular therapy based on assumptions having to do with severity, etc.

But in practice, if psychotherapy (or medication) is given a good try, but doesn't work on its own, then I believe a different therapy should be offered, or a combination (I tend to suggest combination therapy in most cases, as there is a very modest but reasonable evidence base that combo therapy is slightly more effective than one type of therapy alone).