February 01, 2012
Blood Test Detects Mental Depression?

Not sure if you are feeling enough mental pain to be classified as depressed? A blood test can detect depression.

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.

"Traditionally, diagnosis of major depression and other mental disorders has been made based on patients' reported symptoms, but the accuracy of that process varies a great deal, often depending on the experience and resources of the clinician conducting the assessment," says George Papakostas, MD, of the MGH Department of Psychiatry, lead and corresponding author of the report. "Adding an objective biological test could improve diagnostic accuracy and may also help us track individual patients' response to treatment."

Use of a test to detect response to treatment isn't just of clinical value. It also has uses in the development of new ways to treat depression. If drug candidates, diet and life style changes can have their impact measured by shifts blood depression biomarkers then new methods of intervention can be tested more quickly and reliably.

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.

Can other mental states eventually be measured with blood tests? Imagine that blood biomarkers for highly productive mental states can be found. One could then start to try to control one's biochemistry to keep oneself in a more productive mental zone.

Share |      Randall Parker, 2012 February 01 10:47 PM  Brain Depression

Black Death said at February 2, 2012 6:58 AM:

The test panel is interesting. The predictive value of a positive result is 80% and the predictive value of a negative result is 92%. However, the study group had a disease prevalence of 46%. In a population with a lower disease prevalence, the relatively poor specificity (80%) will cause most positive results to be false positives.

Engineer-Poet said at February 2, 2012 12:54 PM:

That would only be an issue if the population as a whole was being tested, rather than people with symptoms suggesting depression.

Basil said at February 2, 2012 12:56 PM:

It sounds pretty scary. You go for a blood test and it comes back saying you're insane, so your assets are frozen and you're thrown in the loony bin.

Ben said at February 2, 2012 8:45 PM:

Tom Cruise finally has his answer!


AMac said at February 4, 2012 3:30 AM:

It seems to be pretty easy to get a multimarker panel test like this to the point where an optimistic peer-reviewed article is warranted. Unfortunately, most fail as clinical development efforts continue. Since failures typically aren't formally reported -- that's the stuff of chatter in poster sessions at meetings. The wider audience is left in a state of anticipation... the next step towards realization must be just around the corner!

Here's a long report from 2004 on a very promising test for ischemic stroke, being developed at that time by Daniel Laskowitz of Duke. CAP Today, "New biomarkers could deliver a strike at stroke". Commercial development was placed on hold, and isn't progressing (as far as I know (I could be wrong about that)).

Here are some of the issues. First, it's difficult using current technology to clearly identify people with "stroke" or "depression," defined as a unitary, biologically-meaningful and clinically-relevant condition. This task is easiest when comparing candidates who are likely sufferers from control cases who clearly are not (as in the report Randall cites). However, that's not a useful distinction for an actual test to make. The ER doc already knows that a cheerful, alert, oriented grandfather brought in by his daughters for a lacerated finger doesn't need to have depression excluded as a diagnosis.

These multimarker panels typically have a much harder time identifying likely diagnoses once you start studying people with clinically-relevant conditions. Laskowitz' panel foundered on making high-sensitivity, high-specificity distinctions among ischemic stroke, diabetic coma, septic shock, and a few other conditions -- the intake nurse or doctor's actual triage challenge.

Finally, ironically, the way the FDA has trials structured places a special burden on excellent new tests. Every "Gold Standard" current diagnostic has false-positive and false-negative rates. If your new test is better at correctly diagnosing these individuals, it gets punished. Since the predefined outcomes are comparisons to how the candidate does versus that Gold Standard.

I don't mean to pour cold water on this story. A reliable blood test for depression would be extremely useful, and the current explosion of knowledge means that the prospects for success are higher than they have ever been. It's just that it's a much longer road from "promising early result" to "marketed diagnostic" than would seem to be the case.

Randall Parker said at February 8, 2012 7:52 PM:


I wiped out all his posts on that web site I won't mention. I'll see how far I have to go. Next step I'll put on a filter against that URL if he posts it again.


Thanks for the learned commentary. For people who are unhealthy I can easily imagine how many of the markers for depression would turn up for other reasons. So a fingerprint for depression might also be a fingerprint for a dozen other disorders.

The number of signals needed to tell apart many different disorders is likely to be very high. That's especially the case when older people have multiple things going wrong with them.

Post a comment
Name (not anon or anonymous):
Email Address:
Remember info?

Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright