Drugs That Distort the Salivary Melatonin Profile Test

The salivary melatonin profile is a timed series of saliva samples, usually collected every 30 to 60 minutes through an evening under dim light, used to calculate dim-light melatonin onset (DLMO). It is distinct from a single random serum melatonin level and from a 24-hour urinary 6-sulfatoxymelatonin collection, which reflects total daily melatonin metabolite output rather than onset timing. This article addresses only the salivary DLMO protocol.

Disambiguation for readers comparing tests: DLMO testing measures when melatonin rises in the evening, not how much total melatonin the body produces over 24 hours. A drug can distort DLMO timing (shifting it earlier or later) and separately distort amplitude (how high the peak reaches), and these are not the same error. A clinician reading a report needs to know which type of distortion is plausible for each medication the patient takes.

The question that matters more than "is my melatonin normal"

Most patients and even some referring clinicians ask whether a melatonin level is "high" or "low." For this specific test, the more useful question is: did anything the patient took, ate, or was exposed to in the prior 24 to 72 hours make this particular curve untrustworthy, independent of the patient's true circadian phase? A flat curve caused by a beta-blocker taken that evening looks identical on paper to a flat curve caused by genuine loss of circadian rhythmicity (as can occur in some Smith-Magenis syndrome or non-24-hour sleep-wake disorder cases). Only a medication and light-exposure audit, done before the test is scheduled, lets you tell the two apart.

What is established, what is plausible, and what is not established

Established from controlled human studies:

• Beta-blockers reduce nocturnal melatonin secretion, with non-selective and beta-1 selective agents both showing substantial suppression in controlled pharmacology studies (Stoschitzky et al.; Arendt et al.).
• NSAIDs, through prostaglandin pathway inhibition, reduce nocturnal melatonin output in healthy volunteers (Murphy et al.).
• Evening light exposure at ordinary room brightness suppresses melatonin onset and shortens melatonin duration (Gooley et al.).
• Caffeine taken several hours before habitual bedtime measurably delays circadian phase markers in controlled laboratory conditions (Burke et al.).
• Fluvoxamine, a strong CYP1A2 inhibitor, increases circulating melatonin by reducing its clearance (Hartter et al.).
• Over-the-counter melatonin supplements show substantial label-to-content variability, so a patient's report of "not taking melatonin" does not guarantee an interference-free result if they stopped a supplement only shortly before testing (Erland & Saxena).

Plausible but not well quantified in the literature provided here:

• The exact percentage suppression for many individual drugs (for example, specific benzodiazepines, atypical antipsychotics, or COX-2 selective NSAIDs) comes from small studies with limited generalizability. Treat any single-study percentage as an estimate, not a fixed conversion factor for an individual patient.
• Washout durations expressed in exact hours (below) are derived from general pharmacokinetic half-life reasoning, not from a validated pre-test washout trial for this specific assay. They should be treated as a starting point for a conversation with the ordering clinician, not a protocol to self-administer.

Not established:

• There is no validated correction formula that lets a lab "adjust for" a known interfering medication and still report a reliable DLMO. If a known interferent could not be washed out, the standard approach is to document it and consider repeat testing rather than to mathematically correct the result.

Beta-blockers: the largest single-drug effect

Melatonin synthesis in the pineal gland depends on norepinephrine acting on beta-1 adrenergic receptors, which activates arylalkylamine N-acetyltransferase (AANAT), the rate-limiting step in melatonin production. Blocking those receptors reduces melatonin output. Comparative pharmacology work found that atenolol produced a much larger reduction in nocturnal melatonin than nebivolol, a beta-1 selective agent with additional nitric oxide activity (Stoschitzky et al.). Earlier work using atenolol showed abolition of the normal 24-hour urinary melatonin metabolite rhythm (Arendt et al.).

The practical consequence: a patient on a beta-blocker who undergoes a salivary melatonin profile may show a flattened curve with no clearly identifiable DLMO threshold crossing. This can be mistaken for absent circadian rhythmicity when it may simply reflect pharmacological suppression. Any interpretation of a flat or low-amplitude curve should record current beta-blocker use before a circadian rhythm disorder is diagnosed.

Stopping a beta-blocker carries real cardiac risk in patients treated for arrhythmia, post-myocardial infarction secondary prevention, or heart failure. Any washout decision belongs to the prescribing physician, not to the sleep clinician ordering the test, and in many patients the safer choice is to interpret the result knowing suppression is likely rather than to hold a cardioprotective medication.

NSAIDs and prostaglandin-pathway interference

NSAIDs suppress melatonin through inhibition of cyclooxygenase (COX), which reduces prostaglandin signaling that normally supports pineal AANAT activity. A controlled trial found that a single evening dose of ibuprofen reduced overnight urinary melatonin metabolite excretion substantially compared with placebo (Murphy et al.). Older crossover work also reported reductions in plasma melatonin after ibuprofen and indomethacin, though this is a small, dated study and its exact effect size should be treated as illustrative rather than a number to rely on for an individual patient (Surrall et al., verification of exact effect size recommended).

There is not enough published data here to call COX-2 selective agents (celecoxib, meloxicam) safe alternatives for pre-test use. Acetaminophen is the more conservative choice for pain control in the 48 hours before collection, based on the mechanism (it does not meaningfully inhibit the prostaglandin pathway implicated in this interaction), though a dedicated trial confirming acetaminophen's neutrality on salivary melatonin was not part of the source material for this page.

Caffeine and the CYP1A2 pathway

Caffeine can shift circadian phase through adenosine receptor blockade and shares a metabolic pathway (CYP1A2) with melatonin. A randomized, placebo-controlled study found that caffeine consumed several hours before habitual bedtime delayed a circadian phase marker by an amount roughly half the effect of a few hours of bright light exposure, with the two exposures combined producing a larger delay than either alone (Burke et al.). The original study text is the appropriate source for the exact minute-level effect sizes; readers and reviewers should confirm the specific numbers against the published paper rather than a secondhand summary, and any attributed quotation from the study's authors should be verified against a citable interview or press statement before being republished, since none is included in the source material available for this draft.

Because the phase shift of interest in many circadian evaluations is itself only one to two hours, an evening caffeine dose that delays phase by even a fraction of that window can meaningfully distort a DLMO reading. Patients should be advised to avoid caffeine from all sources (coffee, tea, energy drinks, chocolate, some analgesics) for a conservative window before and during the collection evening.

Benzodiazepines, Z-drugs, and other GABAergic agents

Benzodiazepines have been reported to reduce nocturnal melatonin amplitude in small studies of agents such as alprazolam (McIntyre et al., verify exact effect size against full text). Data on Z-drugs (zolpidem, zopiclone, eszopiclone) are sparse; some evidence suggests a smaller effect than benzodiazepines because of more selective GABA-A receptor binding, but this is not well established. Gabapentin and pregabalin act on voltage-gated calcium channels rather than GABA receptors and do not have a clear documented effect on salivary melatonin in the literature reviewed for this page.

Antidepressants, antipsychotics, and serotonergic agents

Serotonin is the direct biochemical precursor of melatonin, so drugs that alter serotonergic signaling or its metabolism can plausibly change melatonin output.

Fluvoxamine is a well-documented CYP1A2 inhibitor and has been shown to substantially increase melatonin bioavailability when co-administered, by slowing its clearance rather than by increasing its synthesis (Hartter et al.). This can produce a falsely high or falsely early-appearing curve. Other SSRIs have weaker, less predictable effects on CYP1A2 and melatonin.

Tricyclic antidepressants and antipsychotics. Small studies of antidepressant effects on melatonin metabolites exist, but effect sizes vary between agents and studies, and the evidence for atypical antipsychotics is inconsistent (Carvalho et al., verify claims against full text before citing an exact percentage). Given the risks of destabilizing a psychiatric medication regimen, decisions about holding these drugs before testing should be made jointly by the ordering clinician and the prescribing psychiatrist, not unilaterally for the sake of a lab result.

Exogenous melatonin and melatonin-receptor agonists

Exogenous melatonin supplements and melatonin-receptor agonists (ramelteon, tasimelteon) will contaminate a salivary melatonin profile. Over-the-counter melatonin content is notoriously inconsistent with label claims: an analysis of commercial supplements found actual content ranging from well below to well above the labeled dose (Erland & Saxena). Ramelteon is not melatonin itself, but its active metabolite has weak melatonin receptor activity and has been reported to cross-react with some melatonin immunoassays, which can produce a falsely elevated reading; prescribers can review ramelteon's metabolism in the FDA label (Rozerem prescribing information). A conservative approach is to discontinue over-the-counter melatonin for at least a week and ramelteon for at least five days before testing, in coordination with the prescriber, given melatonin's short half-life but the unpredictable content of many supplement products.

Light, smoking, and alcohol

Light. DLMO testing depends on dim-light conditions, typically well under 30 lux. Controlled research shows that ordinary room lighting in the evening suppresses melatonin onset and shortens the duration of the melatonin signal compared with dim light (Gooley et al.). A smartphone screen checked during sample collection is a common, underappreciated source of this contamination.

Smoking. Tobacco smoke induces CYP1A2 activity, the enzyme pathway that clears melatonin, and chronic smokers have been reported to have lower peak melatonin than non-smokers of similar age. Asking a patient to quit smoking before a single lab test is not a realistic or appropriate clinical recommendation; documenting smoking status for the interpreting clinician is the practical step.

Alcohol. Evening alcohol intake has been shown in a dose-dependent, placebo-controlled crossover study to suppress melatonin secretion (Ekman et al.). Patients should avoid alcohol for at least 24 hours before and during collection.

Decision framework: is this DLMO result trustworthy?

This framework is a clinical reasoning aid, not a validated diagnostic algorithm. It is meant to structure the conversation between the ordering clinician, the patient, and (when relevant) the prescriber of an interfering medication. Washout hours below are derived from general half-life reasoning (approximately five half-lives to reach near-complete drug elimination), not from a dedicated pre-test washout trial for this assay, and should be confirmed with the prescriber before any medication is held.

Step 1: Identify hard-stop interferents. These have the strongest documented mechanistic and pharmacologic basis for distorting the curve:

• Beta-blockers (any agent, any dose)
• NSAIDs taken within the prior 24 to 48 hours
• Exogenous melatonin or melatonin-receptor agonists
• Fluvoxamine

If any of these is present and cannot be safely held, the result should be reported alongside that fact, and a repeat test after resolution should be discussed rather than treated as optional.

Step 2: Identify soft-caution interferents. These have weaker or more variable evidence, or effects that are plausible but not well quantified:

• Other benzodiazepines and Z-drugs
• Other SSRIs, tricyclics, and atypical antipsychotics
• Caffeine and alcohol within the prior 24 to 48 hours

Document these rather than automatically canceling the test; whether to reschedule depends on how much the pending clinical decision depends on precise phase timing.

Step 3: Check the environment, not just the medication list. A patient who properly held every interfering drug but sampled saliva under a bright bathroom light or checked a phone screen every 30 minutes has still produced an unreliable curve. Light and screen exposure should be reviewed with the same rigor as the medication list.

Step 4: If a hard-stop interferent could not be held, treat the result as directional, not diagnostic. For example, if a patient on a beta-blocker for arrhythmia shows a flat curve, that flat curve does not confirm absence of circadian rhythmicity. It confirms that beta-blockade was present, which is expected to suppress the signal regardless of the patient's true phase.

Illustrative washout starting points (confirm with prescriber before use):

Preparing a patient for collection

Two weeks before: Review every prescription, over-the-counter product, and supplement. Identify hard-stop and soft-caution interferents using the framework above, and route any medication-hold decision through the prescriber.

48 hours before: Stop alcohol. Minimize caffeine, ideally for 24 to 48 hours, and at minimum well before the evening collection window.

Collection evening: Keep ambient light dim (well under 30 lux) starting several hours before habitual bedtime. Avoid uncovered screens. Collect saliva at the scheduled intervals without eating, drinking, or brushing teeth in the 15 minutes beforehand. Label each sample with the exact clock time.

Interpretation: The performing lab reports raw concentrations at each timepoint; DLMO is the clinician's calculation of when the curve crosses the assay's defined threshold. If a known interferent could not be avoided, that fact belongs in the interpretation, not just in the chart.

What normal and abnormal results generally mean

Reference ranges vary by assay and laboratory, so exact cutoffs should always be confirmed against the performing lab's reported normal range rather than treated as universal. In general terms, daytime salivary melatonin is low, rising in the evening and peaking overnight, then declining toward daytime levels by morning. DLMO in adults with typical sleep timing is generally described as occurring a few hours before habitual sleep onset.

An unusually high result can reflect exogenous melatonin contamination, co-administration of a strong CYP1A2 inhibitor such as fluvoxamine, or, rarely, a pineal lesion with autonomous secretion, which would warrant further clinical evaluation rather than assumption. A low or flat curve can reflect beta-blocker or NSAID suppression, light contamination during collection, normal age-related decline in melatonin secretion reported in some but not all studies of older adults (Scholtens et al.), or a genuine circadian rhythm disorder. Distinguishing these requires the medication and behavior audit described above, not the number alone.

When to seek clinical evaluation rather than repeat testing

A flat or ambiguous melatonin curve is not an emergency, but persistent, severe difficulty falling asleep or waking at a socially functional time, especially with impact on work, school, or safety (for example, drowsy driving), warrants evaluation by a sleep medicine clinician rather than repeated home light and caffeine adjustments alone. Circadian rhythm disorders are diagnosed using history and validated tools alongside DLMO, not from a single lab value in isolation.