Tuesday, January 05, 2010

Do You Have Mild, Moderate, or Severe Depression? Here, Take This Placebo, er, Antidepressant

Yet another meta-analysis with the same damn result -- antidepressants for most cases of depression are placebos. This is in a paper with authors including Jay Amsterdam, Richard Shelton, and Jan Fawcett, who are not exactly cut from the Peter Breggin mold. This was based on six studies which compared antidepressant to placebo in patients who had a wide range of depression severity. Key results:
  • Mild to moderate depression: Effect size of d = .11, which is tiny (and was not statistically significant)
  • Severe depression: Effect size of d = .17, which is pretty darn small (and not statistically significant)
  • Very severe depression: Effect size of d = .47, which is moderate.

Hmmmm. Not looking so hot. Of course, anyone who has paid attention to the clinical trial literature on antidepressants over the past 10 years or so already knew this. But now it's in JAMA, so a wider audience may now pay attention. Or ignore it. Good marketing usually beats science, so maybe this won't make any difference.

Antidepressants for all but very severe depression: All the benefits of placebo plus the added bonus of side effects. Sign me up! To quote the authors: "What makes our findings surprising is the high level of depression symptom severity that appears to be required for clinically meaningful drug/placebo differences to emerge, particularly given the evidence that the majority of patients receiving ADM in clinical practice present with scores below these levels." In other words, most people who receive antidepressants would likely have done just as well on placebo (without the side effects).

A few other posts on the topic:
I've linked the abstract of the latest JAMA study here. Enjoy.


Anonymous said...

No problem. Pharma has the answer with antipsychotic drugs to augment the underperforming antidepressant drugs. Welcome to akathisia with Abilify!

Bernard Carroll said...

The same point was made by Jules Angst almost 20 years ago for imipramine. See Jules Angst, Human Psychopharm 1993;8:401. Keep in mind that the original antidepressant drugs were validated in patients with melancholia or endogenous depression, not patients with generic major depression a la DSM-III-IV. The newer antidepressant drugs are antidepressants lite.

CL Psych said...

And 20 years from now, someone will make the same point yet again. Learning from history? Nah, no way.

And yes, welcome to the land of akathisia. The next step... "If you're still having symptoms of depression after taking Abilify and an antidepressant, try [insert drug name here]..."

Joseph Arpaia, MD said...

Another take on this is that placebos are quite effective. (I think this comes from how the stress response is modulated when one expects things are going to get better.)

What about using extremely small doses of an antidepressant, for example 6.25mg of Zoloft per day, and telling the patient that this dose is likely to be as effective as a larger dose. Its not a lie, and since so many patients want to be taking something it should satisfy that desire with fewer side effects than a "therapeutic" dose.

CL Psych said...

Dr. Joe,
I agree that placebos often work wonders for depression. Exercise and counseling and likewise good stuff for many people.

I recall that a very low dose of Prozac is as effective as a "therapeutic" dose, though I don't have the reference handy.

The problem is that antidepressants often lose their effects upon discontinuation. So I'd rather see other options tried for mild-moderate depression prior to using meds. But maybe a very low dose of an antidepressant would be less apt to lose effect over time (maybe fewer discontinuation/withdrawal problems).

Ben A Alcobra said...

You mentioned science vs. marketing. You forgot funding. How was this study funded, and what lobbyists represent the funding agencies in Congress?

Anonymous said...

For what it's worth, I found a VERY low dose of an AD effective, and I'll forever wonder if it was a placebo effect. But if not, it's one anecdote suggesting "therapeutic doses" might not be necessary.

I can certainly tell you that a so called therapeutic dose of other ADs did not help me and engendered horrendous side effects. Tourette's-like spasms, anyone??

By the way, I know something was going on despite the low dose of Prozac (2.5 mg, where standard dose is 20 mg) because I did have mild side effects: deletion of sex drive, chronic eyelid twitch, and weight gain, thanks to massive carb cravings in the PM.

jsmd said...

What is the pt difference between responders and placebo on the Beck at the high end when people have a Beck over 25-27 and are called responders? It is not easy teasing this info out of the study. They spend a lot of time saying you do not get a statistical difference until this point but very little time saying what the real world difference is. This is just as important. Who cares if I am a "TX responder" and my Beck goes from a 27 to a 23.

Anonymous said...

Well past the "Era of Evidence Based Practices" we now continue to find that there was little evidence to suport them. Having taken numerous anti-depressants both old and new at dosages which were frequently well above those approved by the FDA it is disheartening to find that which I had experienced anecdotally is now being evidenced empirically. Those of us coping with mental illnesses were promised so much and continued to be promised more where so little has been delivered.

I can't help wonder if polypharmacy for depression is being driven by what was reported in the both the JAMA study and the earlier PLoS meta-analysis. Afterall, the solution in the land of symptom management for "biologically based brain diseases" is medication, subsequently increase the medication and then add one or more medications adjunctively. It would be a great paradigm if all those things which make life worth living could be measured in and replaced by milligram equivalents of medication.

Aussie Child Psychiatrist said...

As for placebo and wanting to take something - fish oil omega-3 supplements are what I prescribe. There is evidence of benefit above placebo (needs more replication but looking better than antidepressants despite lack of Big Pharma patent). There is some theoretical basis with depression relating to inflammation and anti-inflam effects of righting the Omega-6/Omega-3 pro-inflammatory imbalance due to modern diet.

Exercise (for BDNF stimulation and host of other benefits), sunshine (for Vit D and melatonin synthesis), sleep (and get them off computer late at night), social activity, diaphragmatic breathing inducing the relaxation response - all have supporting anti-depressant efficacy. Throw in some counselling/psychotherapy/family therapy (can be difficult but often the key given what attachment/social oriented species we are)/ liaison with school to stop the bullying etc - voila - most teens I see get better.

Reserve antidepressant medication for the rare very severe depression with significant neurovegetive symptoms, or severe OCD where CBT not enough.

Neuroskeptic said...

"I agree that placebos often work wonders for depression."

But these trials don't show that - the improvement in the placebo group might have happened anyway (depression gets better with time) or it could just be rating-scale improvement, not real improvement (rating scales suck).

Bruce said...

By putting "severe" in the heading, you are misinforming. SSRI antidepressants are effective for severe depression and are often the only thing that will pull someone out of suicidal despaire. So why do you include "severe" in your heading? Daniel Carlat has a more accurate view of this study on his blog. Please don't misinform us.

CL Psych said...

Thanks for all the comments. To Bruce...
I am actually not misinforming. Have you read the study? I simply took the authors' results and reproduced them with very slight alteration. They used the term "severe depression" and found a meager effect in that class of depression. Hence, it's in the heading. Maybe I didn't read the study carefully - give it a look and tell me where I'm wrong.

Please read the study in JAMA then let me know what you think.

Also, there is precious little evidence that antidepressants relieve suicidality.

I'm not saying that people who take antidepressants don't get better in many cases. I'm saying that the advantage over placebo is apparently not much to write home about for the good majority of depression cases.

Anonymous said...

Bruce: With all due respect, you are the one who is misinformed. Read the study. CL reported accurately. Carlat made the same mistake you made and was corrected by a poster. Have a nice day.

Allen said...

While people may have different views still good things should always be appreciated. Yours is a nice blog. Liked it!!!

Medical Supplies said...

Hi, it is a very good way of marketing. I agree with your effort, it’s a nice blog. Really liked it!!!!!!

iwinc123 said...

Thankyou for this website.There needs to be alot more information like this out there in the general public.
Keep up the good work.

Dr Joseph-that is outright deception.Your basically worming around thing with carefully chosen words so that you can make a deception but without it being an outright lie so you can say "but technically it isnt a lie because the small doses did work in studies"
It is NEVER ethical to give a patient a placebo without their knowledge and permission otherwise you are conducting deception practices,betraying your patient-doctor trust and quite frankly,i hope you get sued in very public lawsuits if you do this or any other doctor that does this so that it will be a warning and deterrant to others in the medical profession from indulging in deceptive practices.
It doesnt matter even if a small dose did work as effectively in clinical trials as a large dose due to the placebo effect-this is irrevelant.The fact remains that unless you disclose to your patient that the small dose was effective due to placebo you are being deceptive and a conartist.
You cant justify it either by saying that patients get helped by the placebo -because the fact remains "the end doesnt justify the means".That means it is not ok to do something unethical even if someone gets helped by it.

The only way that placebo is ethical to be used is if you say outright to a patient with no worming around words,half truths or leaving out information,that low dose medications have sometimes been shown to work on some people by placebo effect and would you be willing and interested to try this?
That way the patient knows that they are being offered a placebo and has the option to make an informed choice and give permission either yes they would like to try it or no they are not interested and want full strength active medication.
That way no one is being deceived.
Medicine must ALWAYS involve autonomy,honesty,transparency and "do no harm" and deceiving people causes alot of harm.
Please read the below links


Things such as exercise,fish oil etc are not placebo as these have been shown to have real physiological effects in the body and fish oil has omega 3 which has shown in studies to benefit brain function so its not an inactive placebo.
Exercise enhances brain circulation.
Even hugs or a soothing touch have shown in studies to cause physiological changes.
I feel its important to mention this as people often confuse anything non-pharmaceutical with being a placebo but we need to separate these words to make it clear that when we say placebo we mean something that isnt proven in studies and hasnt been proven and to not call things "placebo effect" which have been proven to cause hormonal,chemical or circulatory etc changes/effects & we can explain their mechanism of action eg: like exercise.

I really urge that we call things like exercise,hugs,vitamin supplementation,some herbal medicine,some forms of Psychology,fish oil
(and anything else that has some scientific evidence/ known mechanism of action behind it)

nonpharmaceutical treatment options or NPO & that we call things like inactive tablets with no active chemicals,Homeopathy(which hasnt been proven) and pharmaceuticals that are doses too low to produce physiological effects Placebos.

So there should be two group.
One being NPO(or a better title) and one being Placebo so that the two groups are distinguised and not interchangable and no confusion occurs.

moviedoc said...

I think telling a patient 6.25 mg of sertraline might treat depression would be a lie, but if you started them on a therapeutic dose, they got better (for whatever reason), and wanted to reduce the side effects, they could reduce the dose to what could be tolerated.

Are we forgetting that these data apply to large numbers of patients and do not tell us much about an individual. I can't send someone away when they want help and I have a tool that might work.

And it seems the same people who discredit pharmaceuticals have no problem endorsing St John's Wort and other "alternative" "treatments" where the placebo effect likely dominates.

Anonymous said...

All the studies in the meta-analysis used the Hamilton depression rating scales. Now, antidepressants have side-effects. This could be e.g. nausea, leading to reduced appetite and weight loss, sleep disturbances, etc.

One could say that antidepressants actually improve depressive symptomatology, and therefore decreased the HAM-D score, but also induces side-effects that fits the items on the HAM-D. So the rater will score side-effects as depressive symptoms, eventhough the patient's depressive condition is improved.

Therefore antidepressants work in depression(decreases HAM-D), but the mistake of labelling side-effects as depressive symptoms on HAM-D (increases HAM-D), result in a "no difference" result, even if this is not the case.

Just a thought...

Anonymous said...

I didn't get the benefit of the placebo effect, although I got lots of unpleasant side effects. I wonder if the reason the placebo effect didn't work is because I had read some about this issue prior to trying antidepressants & was already skeptical.

So, why did I end up trying them? My shrink told me the data about placebo effect was created by the scientologists/antipsychiatrist crew. I wish I hadn't taken what he said at face value. I would have saved myself a world of hurt.

The Girl In Orange said...

Very enlightening post. Unfortunately the future of the antidepressants looks pretty much as the past. Meds are given the light way, docs just take the word for it, they just don't know better.

Harry said...

Recommending placebos to patients suffering from anxiety would probably just create more problem 'cause the effect of it is, it will make the person feel he's being treated when the truth is, he is not. It will be deception in terms of the patient-doctor relationship. Therapy without placebo prescriptions would be a better alternative medicine. Depression of suffering patients would be severe if he finds out that he's being deceived. A natural remedy for depression is a much safer way for treating people with depression.