In the 30 years since Prozac first hit the U.S. market, antidepressants have revolutionized modern treatments for major depressive disorder, while screening for depression has remained largely unchanged. To diagnose depression, healthcare providers still rely primarily on patient responses to questions based on criteria from the Diagnostic and Statistical Manual of Mental Disorders. Even after a recent update, these criteria have lower diagnostic reliability than those for almost any other psychiatric disorder. This is a serious shortcoming given that, according to the World Health Organization, depression is the number one cause of disability worldwide, and also in light of the recent recommendation from the U.S. Preventive Services Task Force that primary care providers should screen all adults for this condition.
Could a biomarker-based assay improve diagnosis and help lighten the global burden of depression? A string of promising but ultimately unsuccessful studies on single biomarkers for depression has left many psychiatrists highly skeptical about the feasibility of such a test. Other experts, however, believe that a recent shift toward studying multi-biomarker panels that reflect depression’s true complexity could become an unexpected breakthrough for certain patient populations.
When Statistical Powers Combine
Researchers have found numerous biomarkers associated with depression, but the statistical significance of each of these in isolation has not been strong enough to make a diagnosis, according to Richard Shelton, MD, Charles B. Ireland Professor and vice chair of research in the department of psychiatry at the University of Alabama at Birmingham School of Medicine. By combining the results from measuring many of these analytes, however, scientists are beginning to create tests that could potentially wield real predictive power.
Shelton collaborated with one of the first research teams in the field to move away from single biomarkers and pioneer a probabilistic approach in a pair of studies led by John Bilello, PhD, chief scientific officer of Ridge Diagnostics, and George Papakostas, MD, associate professor of psychiatry at Harvard Medical School. In a combined pilot (n = 79) and replication (n = 77) study and a larger follow-up confirmation study (n = 154), this research team showed that a panel of nine serum biomarkers and a prediction algorithm could differentiate between depressed and non-depressed individuals (Mol Psychiatry 2013;18:332–9 and J Clin Psychiatry 2015;76:e199–206).
Known as MDDScore, the test demonstrated a sensitivity and specificity both >90%, and holds the distinction of being one of the few biomarker-based assays for depression whose effectiveness has been validated in a replication study. The nine biomarkers it measures are associated with changes observed during major depression that take place in the neurotrophic, metabolic, inflammatory, and hypothalamic-pituitary-adrenal axis pathways. Using the pattern of these biomarkers’ concentrations as well as body-mass index, the algorithm then calculates a score on a scale of 1 to 9 that is adjusted for sex indicating the probability that a patient has depression.
Ridge Diagnostics currently offers MDDScore through its CLIA-certified laboratory. The company also continues to collect data on the test’s performance in the more than 4,000 patients it has been used for so far. However, according to Bilello, because the test is used to aid clinicians in making a diagnosis and not as a definitive diagnostic, the company has chosen not to seek Food and Drug Administration (FDA) approval for now. “Eventually we might want to do that, but it wasn’t as important as getting the test out there to healthcare professionals who needed the support,” said Bilello.
The Challenge of Comorbid Conditions
As depression researchers embrace multianalyte panels, they still face hurdles on the road to developing an FDA approved test that clinical labs across the country could use. The most significant of these, according to Shelton, is creating a test that performs well, not just in research samples, but in depressed patients seen in clinical settings. Most biomarker investigations study patients who are depressed but otherwise healthy. In reality, however, patients with major depressive disorder often have comorbid conditions that could interfere with test results.
“Patients have elevations of certain inflammatory biomarkers in depression, but they also have elevation of those same inflammatory biomarkers in heart disease, diabetes, and connective diseases like lupus,” Shelton said. “So researchers and clinicians can’t really take those as primary biomarkers.”
While both published studies on MDDScore were conducted in somewhat restricted patient populations, Bilello said that Ridge Diagnostics has gone on to evaluate the test in patients with comorbidities such as chronic pain and intends to further validate it in populations with other conditions such as HIV infection.
Shelton’s own work on depression biomarkers at the University of Alabama has involved broadly inclusive populations. As he explains it, however, designing studies with representative patient samples can present a catch-22. Many times, these depression biomarker studies are performed by small biotech companies that don’t have a lot of money and only get one shot at success. If their first attempt to validate a test fails because they used a broad patient sample, they probably won’t get the funding for additional studies.
Notes From the Field
If and when FDA approves a test for depression, the ultimate question is, will it be useful? By looking at how healthcare providers have received MDDScore, one can get a glimpse of how a more widespread test might be implemented in clinical practice.
Interestingly, healthcare providers report that MDDScore has helped with a surprising patient management issue that is not unusual in primary care. When patients are clearly depressed but in denial about it, clinicians have discovered that an MDDScore result indicative of depression can help convince patients to accept treatment. “By the time somebody gets kicked to see me, they have already committed to coming to see a psychiatrist,” Shelton said. “But in the primary care environment, because of the stigma associated with mental illness, some people just do not want to accept the diagnosis.”
Though MDDScore has proven useful in primary care, Shelton does not use it and feels a biomarker-based test would not add to his practice as a psychiatrist. In contrast, other psychiatrists have told Bilello that when patients want to stop their medication prematurely, they’ve found that a high MDDScore result can dissuade them. Bilello also envisions that MDDScore will help identify depression in adolescents and the elderly, two populations that might not communicate their internal states as accurately as adults in their prime.
Critics have charged that at $800 a pop, MDDScore isn’t cost-effective. Speaking about biomarker-based tests for depression in general, however, Ramin Parsey, MD, PhD, begged to differ on the claim that they are not worth their potentially hefty price tags. “People say, ‘well, we would never spend $1,000 for a depression test,’” said Parsey, who is chair of the department of psychiatry and director of PET Research at Stony Brook University School of Medicine in Stony Brook, New York. “But we spend thousands of dollars on certain tests for cancer. And the truth is that depression is just as devastating if you look at the statistics.”
As for how laboratory medicine professionals might be affected when tests for depression and other mood disorders lift off, Mark Frye, MD, said they likely will play a central role in ensuring that results are accurate. “Clinical labs are going to be critical in this process,” said Frye, who is chair of the department of psychiatry and psychology at Mayo Clinic in Rochester, Minnesota. “We are really going to rely on the lab to recognize the importance of standardization of the assays.”
The Holy Grail of Mood Disorder Testing
Beyond improving depression diagnosis, biomarker-based tests also could distinguish patients with major depressive disorder from those who have bipolar disorder and are experiencing a depressive episode—something that experts describe as the Holy Grail of mood disorder testing. The great need for such a test stems from the fact that bipolar disorder is treated with mood stabilizers, a completely different class of medications from antidepressants. If antidepressants are given to a bipolar patient, they can actually make the illness worse.
While bipolar disorder biomarkers have proved as elusive as those for depression, a recent study (n = 288) by researchers from the Mayo Clinic in Rochester, Minnesota used proteomic profiling to identify six inflammatory and immune-mediated serum proteins that distinguish bipolar I/II and major depression patients from healthy individuals (Transl Psychiatry 2015;5:e689). Three of these proteins differentiated bipolar I patients, not just from healthy controls, but also from the patients with major depression.
Mark Frye, MD, the paper’s first author and chair of the department of psychiatry and psychology at Mayo, wrote that these preliminary results “support the possibility of developing a diagnostic test using the discovered biomarkers,” though further investigation is needed to validate them. “Our conclusion from this work is very intriguing,” Frye told CLN. “What we need to do now are larger studies and replications and those studies are now going forward.”