A step toward useful biomarkers for depression? Wouldn't that be great?
That was my reaction when I got word of a paper just out in the journal Molecular Psychiatry. It found that in a small pilot study, a test of nine biological measures — or "biomarkers" — in a patient's blood, including levels of stress hormones and inflammation, could determine whether they were depressed with nearly 92% accuracy. (See a full list of biomarkers and details of the results below.)
You may scoff at my excitement. Who needs biomarkers? After all, you might say, how hard is it to tell if someone is depressed? They're miserable. They can't get out of bed or can't sleep or can't eat. Not rocket science.
My response: No, that's not rocket science. But if there's one thing psychiatry desperately needs and has been seeking for decades, it's the biological markers of mental illness — the kind of reliable, lab-testable signatures that so many physical diseases have, and that could be used for better diagnoses and perhaps even to help decide which treatments to prescribe. The new study calls biomarker tests "the 'holy grail' of Psychiatry."
The paper's findings are just a first step and need to be replicated and confirmed among more subjects before there's even any question of using them in the clinic. But I asked to speak with the study's lead author, Dr. George Papakostas of Massachusetts General Hospital, because the findings made me wonder this: What does it mean that you can pretty much nail major depression using nine biomarkers? And that those markers measure oddly disparate processes like inflammation and the birth of new neurons and stress hormones and metabolism? Do we have, in effect, a sort of a recipe for depression, and if so, what does it tell us about what the darkness of depression is?
Dr. Papakostas is an associate professor of psychiatry at Harvard Medical School and director of treatment-resistant depression studies at Mass. General. First, his important disclaimer about the state of the research: Yes, he said, the paper's findings are promising but this was only an initial pilot study, so people "should get excited but should not panic."
It's good that his team "knows what they do not know," he said, and is well aware of how their findings should not be translated: "What this shouldn't do is invalidate patients' experiences, or invalidate clinicians' impressions," he said. "What this shouldn’t do is, if a patient shows up with all the signs and symptoms of depression needing treatment, but the test is negative, then the insurance company says, 'We won’t pay.' Or the patient is disabled from depression, is receiving disability benefits, and they’re cut off because it’s a negative test."
Potential positive uses of the test, he said, include perhaps distinguishing between depression and bipolar disorder; helping physicians who are on the fence about a diagnosis and can't quickly get a second opinion; and confirming the diagnosis for the relatively rare patients who "feel the need to see things in a test. They say, 'I understand I have depression and of course I see the need for treatment, but I just wish there were something biologically you could show me that removes any doubt that it's just me being sensitive.'"
[module align="right" width="half" type="pull-quote"]'What this shouldn't do is invalidate patients' experiences, or invalidate clinicians' impressions.'[/module]
Now for my question: The paper describes decades of frustration in the search for useful depression biomarkers, spikes of hope that petered into disappointment. Researchers thought the the key might be molecules known as catecholamines, which include that famed brain chemical of pleasure and addiction, dopamine. But no. Might it be cortisol, which is linked to stress? No, or perhaps only in some patients. Or brain-derived neurotrophic factor, which has been linked to the birth of new neurons in the brain known as neurogenesis? Another no.
None of those leads has panned out individually, but the nine-measure test incorporates all of them. So is depression, then, some sort of convergence of chemical badness?
"I think this really reinforces the idea that depression is mult-factorial, and that there are several common final pathways, but that you can’t really speak of one major risk factor for depression. It’s basically a minor contribution from many different things — perhaps with the exception of chronic inflammation, where it's actually a moderate contribution.
It really speaks to the fact that it’s very difficult to map a certain process to the outcome of depression, just because of the many factors that are required."
And yet, I said, it all manifests itself as one mood. So is depression ultimately one big overarching thing, or should we think of it as more like a bunch of chunks of different processes?
At this point, it's looking more like chunks, Dr. Papakostas replied. The nine-test paper "essentially confirms individual leads in these areas that were pointed out previously. And also, the fact that these individual markers made it into the test suggests kind of an independent role as opposed to their being confounding. Otherwise, they wouldn't have been in the model because the model makes sure that if two markers are mapping onto the same thing then the stronger one wins.
"It suggests that these preliminary leads that researchers were talking about over the last several decades were in fact valid. And they seem to have a kind of an independent or semi-independent role, again pointing to the multi-factorial nature of depression. Because if you had things from just inflammation, you'd say it's only inflammation; or just neurogenesis, you'd say it's only a pruning of synapses. But all of these things are maintaining their independence.
So it may be that what’s part of the process is that all these things are going wrong. It needs chronic inflammation plus pruning of synapses plus hormonal dysregulation. It may be you have to have a combination or some level of all of these to get depression, as opposed to just one."
For the brain-chemical experts and Wikipedia-lovers out there, here are the nine biomarkers measured in the study:
- alpha1 antitrypsin
- apolipoprotein CIII
- brain-derived neurotrophic factor
- epidermal growth factor
- soluble tumor necrosis factor alpha receptor type II
And here are some details from the Mass. General press release. Financial note: The study was financed by Ridge Diagnostics Corp.
BOSTON – The initial assessment of a blood test to help diagnose major depressive disorder indicates it may become a useful clinical tool. In a paper published in the journal Molecular Psychiatry, a team including Massachusetts General Hospital (MGH) researchers reports that a test analyzing levels of nine biomarkers accurately distinguished patients diagnosed with depression from control participants without significant false-positive results...
The test developed by Ridge Diagnostics measures levels of nine biomarkers associated with factors such as inflammation, the development and maintenance of neurons, and the interaction between brain structures involved with stress response and other key functions. Those measurements are combined using a specific formula to produce a figure called the MDDScore – a number from 1 to 100 indicating the percentage likelihood that the individual has major depression. In clinical use the MDDScore would range from 1 to 10.
The initial pilot phase of the study enrolled 36 adults who had been diagnosed with major depression at the MGH, Vanderbilt University or Cambridge Health Alliance in Cambridge, Mass., along with 43 control participants from St. Elizabeth's Hospital in Brighton, Mass. MDDScores for 33 of the 36 patients indicated the presence of depression, while only 8 of the 43 controls had a positive test result. The average score for patients was 85, while the average for controls was 33. A second replication phase enrolled an additional 34 patients from the MGH and Vanderbilt, 31 of whom had a positive MDDScore result. Combining both groups indicated that the test could accurately diagnose major depression with a sensitivity of about 90 percent and a specificity of 80 percent.
"It can be difficult to convince patients of the need for treatment based on the sort of questionnaire now used to rank their reported symptoms," says Bilello. "We expect that the biological basis of this test may provide patients with insight into their depression as a treatable disease rather than a source of self-doubt and stigma. As we accumulate additional data on the MDDScore and perform further studies, we hope it will be useful for predicting treatment response and helping to select the best therapies."
Papakostas adds, "Determining the true utility of this test will require following this small research study with larger trials in clinical settings. But these results are already providing us with intriguing new hints on how powerfully factors such as inflammation – which we are learning has a major role in many serious medical issues – contribute to depression."
This program aired on February 6, 2012. The audio for this program is not available.