Depression Biomarker Tests: What Blood, Saliva and Metabolomic Studies Can Tell Us
Key takeaways
- Accurately measuring a molecule does not automatically validate its interpretation.
- Blood RNA, salivary cortisol, inflammatory markers and metabolomic models answer different questions.
- Independent validation is valuable, but classification accuracy alone does not prove clinical benefit.
- Symptoms and support needs remain important regardless of a normal or abnormal panel result.
A blood or saliva test for depression sounds appealing because it promises something concrete. A number may seem easier to explain than a collection of symptoms, and a personalized report can appear more decisive than a difficult clinical conversation.
Depression biomarker tests nevertheless need the same scrutiny as any diagnostic claim. What exactly is measured? What does the result predict? Who was studied? Most importantly, what evidence shows that acting on the result helps the person being tested?
This article reviews several types of primary research. It does not endorse a commercial panel or suggest that a laboratory result should replace an assessment of mood, functioning, treatment history and safety.
What is a biomarker being asked to do?
A biomarker is a measurable biological characteristic used for a defined purpose. That purpose matters. A marker associated with depression is not necessarily a marker that diagnoses depression, predicts its future course or selects the best treatment.
Consider four hypothetical reports. One describes an inflammatory value, another estimates a diagnostic probability, a third forecasts relapse and a fourth recommends a medicine. Even if they use the same blood sample, their claims require different evidence.
Before discussing accuracy, ask the provider to name the intended decision. A vague promise of insight is difficult to evaluate. A specific claim can be checked against the participant group, comparison and outcome in the study offered to support it.
Blood RNA research is a discovery example
In a 2014 primary-care study, Redei and colleagues compared blood transcript levels in 32 adults with major depressive disorder and 32 matched nondepressed controls. Nine markers differed at baseline. The researchers also explored patterns before and after cognitive behavioral therapy.
These results helped identify candidate biological signatures. The small study was not a demonstration that a routine blood panel could replace diagnosis across an unselected clinic population. Comparing selected cases and controls is different from testing everyone who arrives with overlapping symptoms.
The paper also found that some transcript differences remained after remission. That is an important reason to define the intended use: a pattern associated with vulnerability or diagnostic history is not necessarily a direct reading of how depressed someone is today.
Ask whether a marketed RNA test uses the same assay, specimen and algorithm as its cited study. A paper about one set of markers does not validate every service that uses the phrase gene expression.
Salivary cortisol is not a depression verdict
The cortisol awakening response describes a change in cortisol after waking. Collecting saliva is relatively convenient, but convenience does not settle predictive validity. Timing and the exact research question remain important when interpreting a study.
A prospective study by Carnegie and colleagues found no evidence that salivary cortisol measures predicted subsequent depression after adjustment. The authors concluded that any association, if present, was unlikely to have substantial clinical significance in that setting.
Other research has reported associations in different populations. For example, a study of 549 people with remitted major depression examined cortisol and recurrence. Predicting recurrence in people with a known history is a different task from diagnosing a first episode in someone buying a consumer test.
These studies cannot be reduced to high cortisol means depression or a normal result rules it out. Our cortisol and chronic stress article provides broader biological context. A hormone narrative should not displace the actual pattern of symptoms.
Inflammation research does not supply a universal treatment rule
C-reactive protein, including high-sensitivity CRP measurements, is used in research on inflammation and depression. It is not a direct readout of one brain circuit or an explanation of every person’s low mood.
A randomized infliximab trial in 60 people with treatment-resistant depression found no overall treatment advantage on depressive symptoms. The researchers observed a relationship between baseline inflammation and treatment response that motivated further investigation of subgroups.
This is not evidence to self-prescribe an anti-inflammatory treatment from a CRP value. The distinction between a promising treatment-marker interaction and a validated prescribing strategy is substantial. A strategy needs evidence that using the marker to guide care improves relevant outcomes with acceptable risks.
Our separate inflammation and anhedonia coverage examines reward-related mechanisms. The question here is narrower: whether a measured blood value has been shown to improve a particular clinical decision. A biologically interesting result does not automatically meet that standard.
New metabolomic models are more than a single molecule
Zhao and colleagues’ 2026 study analyzed serum and fecal metabolomes in first-episode depression and matched controls, with additional independent cohorts. A model using 34 metabolites reported receiver-operating-characteristic area-under-the-curve values of 0.82 and 0.80 in test and validation sets.
This is meaningful research into classification and associated biology. An AUC summarizes discrimination across possible thresholds; it is not the percentage probability that a particular person has depression, and it is not proof that the model improves treatment outcomes.
The independent cohorts strengthen the research compared with performance measured only on model-development data. Clinical use still requires asking whether the intended patients, assay, threshold and decision match the evidence. Do not substitute a generic microbiome report for the specific model studied.
Nor does a classification study establish that changing a metabolite or taking a probiotic treats depression. A predictor, a possible mechanism and a successful intervention are separate claims. Each deserves an appropriate test.
Three layers of evidence before a purchase
| Layer | Core question | Useful evidence |
|---|---|---|
| Measurement | Is the specimen analyzed reliably? | Clear assay methods and quality controls |
| Interpretation | Does the result identify the intended condition or outcome? | Validation in relevant independent patients |
| Clinical use | Does acting on it improve decisions or health? | Prospective evidence for the proposed strategy |
A test can perform well at the first layer while remaining uncertain at the next two. A polished report, many biomarkers or a sophisticated algorithm cannot bridge those gaps by themselves. More measured variables can mean more information to interpret, not automatically a better answer.
Ask what the clinician would do after a positive, negative or uncertain result. If every possible result leads to the same paid supplement package, request an explanation of how the test changed the decision. This is a practical evaluation question, not a judgment about a particular company.
What a clinical assessment contributes
The National Institute of Mental Health describes diagnosis through symptoms, duration and effects on daily life. It also explains that medicines and medical conditions can produce similar symptoms, and that examination, interview and appropriate laboratory tests can help assess those possibilities.
That is different from saying laboratory testing has no role. A clinician may order a test to investigate a specific alternative or accompanying condition. The purpose should be clear enough that the result has a meaningful interpretation.
Prepare a short account of what changed, how long it has lasted and which activities are affected. Include relevant medicines, supplements and previous treatment responses. You do not need an abnormal molecular result to make that conversation legitimate.
If you already have a commercial report, bring the original document and ask which findings are actionable, which are uncertain and which depend on evidence from a different population. Avoid changing prescribed treatment solely because a report uses a red or green label.
Read accuracy claims beyond the headline
A provider should say whether performance came from a development sample or genuinely independent patients. Ask whether the comparison included people with other conditions that resemble depression, rather than only clearly nondepressed controls.
Also ask what false positive and false negative results would mean in the intended setting. A confident label can affect further testing and treatment discussions. The relevant standard is useful performance for a specific decision, not simply whether a model separates two groups better than chance.
Cost and data handling belong in the same decision. Clarify who receives the results, what additional interpretation costs and whether recurring testing is being recommended. Ask what evidence supports repeating the measurement rather than assuming that tracking more often improves care.
Common questions about depression biomarker tests
Can a normal panel prove that my symptoms are not depression?
The studies reviewed here do not establish that use for an arbitrary commercial panel. A normal result should not dismiss persistent symptoms or replace an appropriate clinical assessment.
Does an elevated CRP value prove brain inflammation?
No. A systemic marker does not directly identify inflammation in a particular brain region or establish the cause of low mood. Its significance depends on the clinical question and context.
What would make a biomarker test more useful?
A clearly defined purpose, reliable measurement, independent validation in intended patients and evidence that using the result improves care. Until those elements support the particular claim, treat the test as information to discuss rather than a stand-alone verdict.
Sources and further reading
- Redei et al., 2014: blood transcript markers in primary-care depression
- Carnegie et al., 2014: cortisol awakening response and subsequent depression
- Hardeveld et al., 2014: cortisol and recurrence after remission
- Raison et al., 2013: randomized infliximab and inflammatory-marker study
- Zhao et al., 2026: metabolomic classification and independent cohorts
- NIMH: depression assessment and care
Continue reading
- Microplastics in Human Brains: What Autopsy Studies Reveal and Cannot Prove
- Long COVID and Cognition: Mechanism Clues, Rehabilitation Trials and Testing Limits
- Hearing Aids and Dementia Risk: What ACHIEVE and Newer Trials Actually Show
- Inflammation and Anhedonia: What Research Shows About Reward and Motivation

