Saliva versus blood versus urine hormone testing: compare the specimen, not just the hormone

Saliva, blood, and urine hormone tests are not interchangeable ways to measure “the same level.” Each matrix supports only the named analyte, collection protocol, assay, units, reference interval, and intended use validated for it. A result cannot be transferred from one matrix to another by default, and whether any measurement is clinically useful is a separate claim.²³

This is a specimen-method comparison, not a diagnosis guide or an interpretation of any hormone panel. Its goal is to make a test claim reproducible enough to audit.

Give every result a matrix passport

If a report cannot supply those fields, a colorful trend graph does not repair the missing context.

Blood: a defined concentration at a defined moment

“Blood test” still needs a specimen name. Serum is the liquid remaining after clotting; plasma is collected with an anticoagulant; whole blood retains cells. A method validated for one is not automatically validated for another. Binding proteins, sample preparation, hemolysis, lipemia, biotin or other interference, and the assay’s ability to distinguish structurally related compounds can matter depending on the analyte.

A blood draw is usually a time-point observation. That can be appropriate when the clinical question, collection conditions, and reference interval were designed around that time. It can be misleading when a hormone has a strong daily rhythm, pulsatile release, a meal-related response, or medication-timing effect and none of that context is recorded.

CMS’s CLIA interpretive guidance expects laboratories to establish or verify performance characteristics and appropriate reference intervals, identify specimen source when necessary, and report results with the information needed for interpretation.² CLIA certification concerns the laboratory’s testing operations. It does not mean that every ordered analyte, sample time, or wellness claim has demonstrated clinical utility.

Saliva: convenient collection with a demanding preanalytical record

Saliva can be collected outside a phlebotomy setting and can support serial sampling. A salivary result is meaningful only for the named analyte, collection protocol, assay, reference interval, and intended use validated for it; it cannot be assumed to reproduce a blood or urine result.

Saliva also contains much lower concentrations of many analytes than blood. Oral blood contamination, recent eating or drinking, toothbrushing, smoking, chewing, saliva flow, collection device, sample volume, pH, freezing, transport time, and storage conditions can alter a sample or an assay’s performance. A 2025 systematic review of salivary cortisol research found substantial variation in sampling periods, protocols, covariates, and analysis across 58 articles.⁵ A separate systematic review documented wide variation in collection devices and handling practices.⁶ Those reviews do not invalidate saliva; they show why a validated protocol is part of the test.

A useful saliva report therefore states:

• exact collection clock and whether it was one sample or a serial set;
• restrictions and deviations before collection;
• collection device, sample adequacy, storage, and shipment conditions;
• analytical method, lower measurement limit, interferences, and quality controls;
• same-matrix, same-method reference interval; and
• evidence for the claimed clinical use.

Without those details, a morning-to-night curve may be visually precise but scientifically under-specified.

Urine: an interval measurement that depends on a complete interval

A spot urine specimen captures one point and is affected by urine concentration. A timed collection attempts to integrate excretion over a stated window—often 24 hours for a particular clinical question. The total result depends on starting and stopping at the specified times, including every required void, recording total volume, using the correct container or preservative, and keeping the specimen under required conditions. Renal function and other physiological factors may also change how a circulating substance or metabolite appears in urine.

The Endocrine Society’s Cushing syndrome guideline offers a useful design lesson: it does not treat random serum, late-night saliva, and 24-hour urine as interchangeable. It connects selected tests to a defined diagnostic question, calls for repeated urine or saliva measurements, and recommends against broad use of random serum cortisol for that purpose.¹ That example should not be repurposed as a self-testing algorithm. It demonstrates that the matrix, clock, repeat plan, and population belong to the clinical pathway.

Reference intervals cannot travel alone

A reference interval describes results observed in a defined reference population under a defined measurement system. It is not necessarily a treatment target, a boundary between wellness and illness, or a personalized optimum. Moving a printed interval from serum to saliva, from one assay to another, or from adults collected in the morning to samples collected at night strips away its validation context.

CDC’s hormone standardization work illustrates the point: standardization and certification are tied to named analytes, serum specimens, and specified measurement procedures.⁴ A laboratory or product should identify the scope of any certification rather than implying that one performance badge covers every hormone, method, and matrix.

Before accepting a reference band, look for five matches:

1. same analyte and molecular form;
2. same specimen matrix;
3. same collection time or interval and relevant physiological state;
4. same or demonstrably harmonized assay method; and
5. a reference population relevant to the report’s stated use.

An interval may be statistically established and still lack evidence that acting on small differences improves a health outcome.

Separate four kinds of test claim

FDA tells consumers that direct-to-consumer tests differ in purpose and review status and explains the distinction between analytical and clinical validity.³ A test being available without an in-person visit, or a sample being processed by a certified laboratory, does not by itself establish the clinical claim printed on the sales page.

Audit reproducibility before debating a number

Reproducibility asks whether the measurement is stable enough for its purpose when the relevant conditions are repeated. A same-day duplicate can assess something different from day-to-day biological variation. A laboratory precision study can assess something different from home collection reliability. A report should identify the level it tested.

1. Freeze the specimen identity. Record analyte form, serum or plasma type, saliva device, or urine collection interval rather than writing only “hormone test.”
2. Reconstruct the collection clock. Capture date, time, interval completeness, required restrictions, medicines or supplements disclosed to the ordering clinician, and protocol deviations.
3. Name the measurement system. Record laboratory, assay platform, units, reportable range, detection limits, known interferences, and quality or certification scope.
4. Match the reference evidence. Verify that interval and comparison data use the same analyte, matrix, timing, method, and relevant population.
5. Demand the utility bridge. Ask what validated decision the result is intended to change and what evidence connects that action to a meaningful outcome.

The strongest comparison does not ask, “Which specimen is best?” It asks, “Which validated specimen-and-method protocol measures the right quantity, at the right time, for this precisely stated clinical or research question—and what evidence supports acting on it?”