Salivary Melatonin Profile: What Your Number Changes About Your Treatment

This article is about a laboratory test for endogenous melatonin, the hormone your own pineal gland produces. It is not about exogenous melatonin supplements (an over-the-counter dietary supplement in the United States, not FDA-regulated as a drug for sleep), and it is not about the prescription melatonin receptor agonists ramelteon (Rozerem, FDA-approved for insomnia) or tasimelteon (Hetlioz, FDA-approved for non-24-hour sleep-wake disorder in totally blind adults). The test result can inform how and when any of those are used, but ordering it does not require using any particular one.

The direct answer

DLMO is the most reliable single marker of circadian phase available outside a research lab, and it changes the timing recommendation for melatonin, light therapy, and some sedative-hypnotics by hours rather than fine-tuning it by minutes (Lewy et al., 1999; Klerman et al., 2002). A patient whose measured DLMO differs from the assumption built into generic dosing advice (for example, "take melatonin 30 minutes before bed") may be dosing during the wrong part of the phase-response curve entirely, which can blunt or even reverse the intended shift. The test does not diagnose a circadian disorder by itself; it must be interpreted against sleep-timing history, and a specialist should confirm the diagnosis before treatment decisions are finalized.

What the test actually measures

The profile captures the pineal gland's secretion curve across a single evening and night to calculate DLMO: the clock time melatonin first rises above a fixed threshold under dim-light conditions (Lewy et al., 1999). In comparisons against other circadian markers, DLMO has performed more consistently than core body temperature nadir or actigraphy-derived estimates (Klerman et al., 2002).

Saliva sampling correlates closely with plasma melatonin in validation work (correlations reported from roughly 0.71 to 0.98 depending on the study and assay) and avoids the venipuncture-related cortisol spikes that can occur with repeated blood draws (Voultsios et al., 1997). The profile also produces a peak-amplitude value and an offset time, both of which matter when amplitude looks blunted or the secretion window is unusually short.

What is established: DLMO measured under controlled dim-light conditions is a validated, reproducible circadian phase marker. What is plausible but not fully settled for an individual patient: the exact hours-before-DLMO dosing rule that maximizes phase shift is derived from group-level phase-response curve studies, and individual response varies. What is not established: that a single DLMO measurement predicts long-term treatment response without reassessment, or that DLMO alone should override clinical judgment about a patient's sleep history and comorbidities.

What counts as a normal result

In adults without a diagnosed circadian disorder, DLMO tends to occur roughly 2-3 hours before habitual sleep onset. For someone who falls asleep around 23:00, DLMO commonly falls between 20:00 and 21:00, though normal ranges are wide across individuals (Lewy et al., 1999). Peak nocturnal salivary melatonin commonly falls somewhere between 10 and 60 pg/mL, again with substantial person-to-person variation.

Age affects both timing and amplitude. Older adults tend to show lower peak melatonin output than younger adults, based on urinary 6-sulfatoxymelatonin (the main melatonin metabolite) excretion studies (Kennaway et al., 1999). Adolescents commonly have a DLMO that runs later relative to typical school start times, a mismatch cited by the American Academy of Pediatrics in its position statement supporting later school start times (AAP, 2014).

A commonly used clinical convention treats a DLMO more than about 2 hours later than expected for age as "delayed," and more than about 2 hours earlier as "advanced." These are working clinical thresholds, not fixed diagnostic cutoffs from a single regulatory or guideline source, and a sleep specialist should apply them alongside the patient's history.

How a measured DLMO changes treatment timing

Prescribing melatonin, a melatonin receptor agonist, or timed light exposure without knowing DLMO means guessing at where in the phase-response curve the intervention will land.

Exogenous melatonin timing. Phase-response curve studies show the largest phase advance when melatonin is given several hours before DLMO, and phase delay when it is given after wake in the morning (Burgess et al., 2010). A patient with a measured DLMO later than assumed, who takes melatonin only an hour or two beforehand based on generic bedtime advice, is dosing too close to onset to capture a meaningful phase-shifting effect. Moving the dose earlier, based on the measured DLMO rather than the clock, is the mechanism by which chronotherapy is expected to work over one to two weeks, though individual response still varies and should be reassessed rather than assumed.

Light therapy scheduling. Morning bright light appears to advance the clock most effectively when timed to the hours following the body temperature nadir, a period that correlates with several hours after DLMO (Khalsa et al., 2003). Scheduling light exposure by clock time alone, without a DLMO estimate, risks delivering light during the delay portion of the curve instead of the advance portion, which can worsen rather than correct the phase problem.

Sedative-hypnotic timing. Dual orexin receptor antagonists (suvorexant, lemborexant) work by blocking wake-promoting signaling; a 2015 systematic review and meta-analysis found suvorexant effective for insomnia symptoms overall (Kishi et al., 2015). That review did not specifically test DLMO-guided dosing, so the idea that circadian phase data improves DORA timing is a reasonable extrapolation from circadian physiology rather than a directly demonstrated finding, and should be treated as such when discussed with patients.

What an elevated profile can mean

A high peak amplitude, or a DLMO that occurs notably early relative to a patient's desired bedtime, is consistent with advanced circadian phase. The classic presentation is early-evening sleepiness followed by very early morning awakening (roughly 3-4 a.m.). The International Classification of Sleep Disorders, 3rd edition, describes advanced sleep-wake phase disorder along these lines and estimates it is uncommon in midlife adults but more frequent in older populations (American Academy of Sleep Medicine, ICSD-3).

Management for confirmed advanced phase generally involves evening bright-light exposure timed a few hours before habitual bedtime, and in selected cases very low-dose morning melatonin to exploit the delay portion of the phase-response curve (Burgess et al., 2010). A 2015 clinical practice guideline from the American Academy of Sleep Medicine on intrinsic circadian rhythm sleep-wake disorders cautions against routine high-dose melatonin in this population because it may worsen morning grogginess without producing a meaningful phase shift (Auger et al., 2015). Note this is an AASM guideline, not an Endocrine Society guideline; the two bodies publish separate recommendations and should not be conflated.

A separate and important cause of an apparently elevated nocturnal melatonin level with normal DLMO timing is drug interaction: fluvoxamine and other CYP1A2 inhibitors can raise measured melatonin by slowing its hepatic clearance, independent of the pineal gland's actual output (Hartter et al., 2000). A medication reconciliation is warranted before attributing an elevated result to endogenous overproduction.

Decision framework: what a measured DLMO should change about the plan

This is not a substitute for a sleep medicine consult. It is a way to organize the conversation once a DLMO result exists.

The purpose of this table is to show that the same lab value implies different next steps depending on amplitude, timing direction, and concurrent medications. A DLMO number without that context is not actionable.

What a low profile can mean

A blunted amplitude or an absent DLMO rise has several distinct possible explanations, and distinguishing them matters before choosing a treatment.

Beta-blocker suppression. Propranolol, atenolol, and metoprolol have been shown to reduce nocturnal melatonin secretion through beta-1 adrenoreceptor blockade at the pineal gland (Stoschitzky et al., 1999). A patient on a beta-blocker with new insomnia should have that drug considered before assuming primary pineal failure. Switching classes or adding melatonin timed to the (potentially still detectable) DLMO signal is a reasonable next step to discuss with the prescribing clinician, not a self-directed change.

Aging. Lower urinary melatonin metabolite output has been associated with age and, in some cohort data, with markers of cognitive decline (Kennaway et al., 1999). This is observational association, not proof that low melatonin causes cognitive decline, and it should not be presented to patients as a dementia risk score.

Shift work. Rotating-shift schedules commonly produce a flattened or fragmented secretion curve. Practice parameters from the American Academy of Sleep Medicine support strategic melatonin timed relative to the desired sleep episode and any residual DLMO signal that can be identified (Morgenthaler et al., 2007).

A flat or hard-to-interpret profile does not mean nothing can be done. It means the endogenous timing cue is degraded and treatment relies more on a fixed external schedule, with reassessment after several weeks to see whether the pattern shifts.

Melatonin receptor agonists and DLMO

Ramelteon and tasimelteon are selective MT1/MT2 agonists with FDA-approved fixed clock-time dosing instructions on their labels (ramelteon roughly 30 minutes before bedtime; tasimelteon about 1 hour before bedtime). DLMO measurement can refine, but does not replace, that labeled dosing.

In the pivotal tasimelteon trials for non-24-hour sleep-wake disorder in totally blind adults, treatment response depended on keeping the dosing window aligned with the patient's free-running DLMO (Lockley et al., 2015). Because the circadian period in this population is not fixed to 24 hours, periodic DLMO reassessment is part of ongoing management, not a one-time test.

For ramelteon, a phase-shifting study in healthy adults (not a clinical DSWPD population) found that dosing timed relative to individual DLMO produced a larger phase advance than dosing timed only to a fixed pre-bedtime schedule (Richardson et al., 2008). This is useful mechanistic evidence for why DLMO-informed timing matters, but it was conducted in healthy volunteers, and the effect size in patients who actually have DSWPD requires separate confirmation rather than direct extrapolation.

Interactions with hormone therapy: what the evidence supports and what it does not

Melatonin secretion does not operate in isolation from the broader endocrine axis, but the evidence connecting it to hormone replacement therapy is limited and should be described carefully.

Testosterone. A study of high-dose testosterone in older men found effects on sleep and breathing, including relevance to obstructive sleep apnea (Liu et al., 2003). This supports monitoring sleep-disordered breathing in men on testosterone therapy; it does not establish a direct testosterone-melatonin interaction, and that distinction should be preserved when counseling patients.

Progesterone. Oral micronized progesterone has GABA-A-modulating effects that can improve subjective sleep onset. A menopause-focused review discusses evaluating sleep disruption in perimenopausal women broadly (Santoro et al., 2015); it is a narrative review, not a formal society practice guideline specific to circadian testing, and should be cited as such rather than as an official guideline recommendation.

GLP-1 receptor agonists. The pivotal tirzepatide obesity trial (SURMOUNT-1) established efficacy and safety for weight loss in adults with obesity (Jastreboff et al., 2022); it is not a source for a specific sleep-quality or circadian outcome. Any claim about a precise sleep-quality score change or melatonin-mediated mechanism attributed to a "post-hoc SURMOUNT-1 analysis" could not be verified against the source material provided for this article and has been removed rather than repeated as fact. What can be said cautiously: substantial weight loss can improve obstructive sleep apnea, which can secondarily change reported sleep quality and sleep timing. There is no established direct pharmacologic effect of GLP-1 receptor agonists on melatonin secretion, and this should be described as unproven rather than as an emerging finding until a specific study is identified and verified.

Collecting a valid sample

Compliance drives accuracy. Standard protocol elements include:

1. Dim light (under about 30 lux, roughly one candle or a single night-light) starting several hours before habitual bedtime.
2. Saliva samples every 30-60 minutes using pre-labeled collection tubes.
3. No food, caffeine, or alcohol in the 30 minutes before each sample.
4. No NSAIDs for 24 hours before testing, since ibuprofen has been shown to suppress melatonin synthesis via prostaglandin pathway effects (Murphy et al., 1996).

A multicenter validation study found that home-based DLMO sampling, when compliance was verified, agreed closely with in-lab assessment (Burgess et al., 2015), supporting outpatient collection as a reasonable option for many patients.

When retesting makes sense

DLMO is not a fixed, one-time number. Seasonal light exposure, new medications, and aging can all shift it. Retesting is commonly considered:

• Several weeks after starting chronotherapy, to confirm the phase actually moved as intended
• After starting or stopping a CYP1A2 inhibitor (fluvoxamine, ciprofloxacin) or a beta-blocker
• Periodically in patients with non-24-hour disorder on tasimelteon, since the free-running period can drift
• After a major schedule change such as retirement or a shift-work transition

There is no single fixed retest interval mandated by a regulatory body or major guideline; the pattern above reflects clinical practice and expert reasoning rather than a codified rule, and the treating clinician should set the schedule based on how the patient is responding.

Raising a low profile, or moving back an early one

For a suppressed profile, before adding a supplement, it is reasonable to review:

• Evening light exposure, since blue-enriched light after DLMO has been shown to acutely suppress melatonin (Gooley et al., 2011)
• Beta-blocker, NSAID, or benzodiazepine use
• Whether low-dose melatonin timed several hours before the measured DLMO, discussed with a clinician, is appropriate to support the endogenous signal while other factors are addressed
• Morning bright light exposure shortly after waking, which supports the broader circadian amplitude

For an early or advanced profile, the converse pattern applies: evening bright light before the desired bedtime, and avoidance of premature evening darkness or napping in darkened rooms, which can reinforce early-phase drift.

When this is not enough on its own

A salivary melatonin profile is one data point. It does not replace a sleep history, a sleep study when obstructive sleep apnea is suspected, or evaluation for depression, which can also disrupt sleep timing. Anyone with new severe insomnia accompanied by chest pain, shortness of breath, suicidal thoughts, or signs of a neurological event should seek urgent care rather than wait for lab-based circadian testing.