The biggest mistake in DSM III was introducing the very broad and heterogeneous category 'Major Depressive Disorder'. This combined under one rubric what had previously been two seperate and quite different presentations: 1) severe, melancholic, delusional, or incapacitating depressions, and 2) reactive to stress, mild, and often transient depressions. The result is that many people get the label Major Depressive Disorder, even though their presentation isn't really 'Major', isn't really 'Depressive', isn't really 'Disorder'. Mild sadness in reaction to stress and disappointment is lumped together with the most severe suffering known to man.
Drug companies jumped on the opportunity to peddle a pill for every problem and misleadingly described all depressions as a chemical imbalance requiring a chemical solution. Treatment studies that previously showed clear superiority of medicine over placebo for severe depression showed little or no superiority with patients whose depression was mild or questionnable. And biological marker studies that showed promise in tagging severe depression came up empty with the watered down Major Depressive Disorder.
Critics of medication jumped on this to argue misleadingly that depression is a myth and/or that medication treatment for depression doesn't work.
Mark Kramer, MD PhD is the perfect person to explain what has happened. He has held senior positions both in academia and in the pharmaceutical industry. Currently, he is Chief of Biological Psychiatry at Thomas Jefferson University.
Dr Kramer writes: "Recently an FDA committee of external advisors rejected a potential new antidepressant "gepirone" on the grounds that it was not effective enough.
Surprisingly, the FDA's leadership reversed the down votes of its staff and external advisors and approved the drug anyway - on the basis of two positive studies, even though there had also been many other studies that were not positive. Gepirone's approval on "2 positives" annoys statisticians. Two positives, among many studies which were not, can be a chance finding.
The modern FDA database (≈1980-present) reveals that antidepressants are on-average more efficacious than placebo, but not robustly so.
This says more about how the studies have been done than about the efficacy of the meds. The original Tricyclic antidepressants (TCAs), while often annoying to take, were never weak or non-specific in their effects, nor has their efficacy progressively weakened.
What happened instead was a progressive change in the population being studied- originally, only the incapacitated and clearly depressed; increasingly as the years progressed, people with milder depressions, perhaps reactive to life stresses, disappointments, and demoralization.
The push to do more studies and to loosen the indications for antidepressant use resulted in their being studied in many people who may not need them at all and who have an extraordinarily high placebo response rate.
Especially after 1980, the demand for archetypical drug-responsive study patients outstripped their timely supply. Studies having trouble recruiting took in less sick patients, thus decreasing the average difference between drug and placebo. As monthly enrollment proceeds in most of today's studies, the drug-placebo difference and its statistical significance diminishes greatly after the first ≈1/3 (of say ≈150) patients per group are enrolled. Patients who are not really, or not very, depressed are pulled to get the study completed.
The patients who now feed clinical trials differ from those pre-1965: 1) much less depressed (not melancholic, not delusional, not suicidal, much less functional disability 2) diagnosis by checklist criteria, not by daily observation in a closed habitat with attention to clinical significance, and; 3) patients recruited by advertisements. The disturbing paradox is that many people in studies likely don't need pill treatment, while those who really do are often excluded as too sick, too suicidal, too noncompliant, too comorbid, or unwilling to volunteer.
In the late 1990's, Dr. Ghosh and I presented the first report concluding that placebo was about as effective as antidepressant in about 50% of the studies in the modern FDA dababase.
Yet, these data did not align with my own earlier experience and the findings in the literature that first generation antidepressants (TCA's) were markedly better than placebo in treating incapacitating depression.
Why were antidepressants suddenly so much less superior to placebo? Two possible explanations: TCA's are perhaps more effective than SSRI's and/or SSRI's are tested in progressively less sick patients.
Upon receiving imipramine, 40% of severely, endogenously, melancholically depressed inpatients returned to normal; 35% more exhibited clinically significant improvement; and the rest did not respond at all. These are very satisfying results in very sick patients who were not on a continuum with normal sadness and who have a very low placebo response rate.
But the advantage of medication vs placebo has progressively diminished over the years as studies shifted to less depressed patients receiving SSRI-like medications.
This does not mean, as some critics allege, that antidepressants don't work for depression. It means only that they are on-average necessary and specifically effective for the more disabling of depressions and that they barely beat placebo.
Biological psychiatry has been a casualty of the largely Pharma spread propaganda that pills are needed for, and have an answer to, all of life's problems. It has been undermined by the study of populations far too heterogeneous in signs, symptoms, and incapacitation to reveal robust and homogeneous biological markers and pathophysiologic mechanisms.
You'd think that with norepinephrine, serotonin, and dopamine theories now unsupported, and with the poor treatment response in questionably depressed patients, researchers would return to studying potential biomarkers in patients with easily confirmable depression who have a clearer response favoring active treatment. Despite their potential harms and inconvenience, but given their substantial benefits, perhaps we need to give the TCA's a second look from a research perspective.
So, in the end how are we to understand gepirone? Is it just an expensive placebo for people who generally don't need medication? The present efficacy data are not convincing. Most likely, it will be an easy drug for primary care docs to prescribe to quickly get patients out of the office, so that piles of paperwork can be attended to. As to its commercials: I can already hear soothing soundtracks playing behind animated cartoons of serotonin molecules locking joyfully into their appointed receptors."
Thanks so much, Dr Kramer, for sharing your wealth and breadth of experience. I have two thoughts that build on your remarks.
The FDA is meant to regulate the drug industry and protect the public from ineffective and unsafe drugs. It used to be a reasonably effective filter, approving only about one-third of drugs submitted to it.
All that has changed. FDA is now under tremendous pressure to speed review and approves just about every proposed drug, even when its own advisors recommend rejection. Gepirone and the useless/unsafe female Viagra are just the most explicit examples of Pharma now dominating the approval process.
The FDA has, in effect, become more a marketing tool of Pharma than a balanced regulator. Ineffective/unsafe drugs get an FDA legitimizing stamp of approval they most certainly do not deserve.
The new FDA director has unprecedented conflicts of interest based on a career of intimate ties with the drug industry. This can work out for better or for worse. The pessimistic view is that he will drag FDA even further into the Pharma trap. The optimistic view is that he will do the right thing because he is legacy sensitive and under special vigilance from Congress, the public, the media, and the medical profession. He certainly knows all the tricks, nooks, and crannies and could do a great job, should he choose to do so.
The next point seems too obvious to be stated, but nonetheless desperately needs stating. Only people who are clearly clinically depressed and clearly need antidepressants should be included in research studies and should be taking antidepressants in everyday clinical practice. Depression has been too carelessly diagnosed- encouraged by the loose DSM definition, by Pharma's desire to push product; by rushed doctors; and by people's hope for a quick fix for life's problems.
Warning: If you are already on an antidepressant, don't stop it abruptly based on anything said above. You may need the med If your depressions have been incapacitating, recurrent, or persistent. And even if you don't need it, withdrawal can be difficult and requires medical supervision.
If you feel depressed but are not yet on antidepressant, get yourself fully informed and get consultations before starting. Disabling depression almost always requires an antidepressant, mild usually does not (watchful waiting or psychotherapy are better first choices). The decision to start an antidepressant is a serious one that should not be made lightly after a brief evaluation with a primary care doctor.
Allen Frances is a professor emeritus at Duke University and was the chairman of the DSM-IV task force.