Pheromones, Supplements and L-Theanine for Dogs: What the Evidence Supports

1. A Shelf Full of Products and a Thin Literature

1.1 The Shape of the Problem

Calming products for dogs are a large and growing market: pheromone diffusers, collars and sprays, milk-protein capsules, amino-acid supplements, tea-leaf extracts, and now cannabidiol in every format. They are sold in veterinary practices, recommended by trainers, and used by owners who have been told something helps.

The evidence behind them is not absent. It is thin, uneven, context-dependent, and sometimes contradictory. That combination is harder to talk about than either "it works" or "it's snake oil", and it is the actual situation.

1.2 Two Arguments That Do Not Work

"It's natural, so it can't hurt." Whether something can hurt is an empirical question, and one of the studies below found a physiological marker moving in the wrong direction. More importantly, a product that does nothing is not harmless when it delays an intervention that would have worked, or when it is used instead of addressing a medical cause (how pain is missed in dogs).

"It worked for my dog." Almost every problem these products target fluctuates on its own. Fear of noise, separation distress and reactivity all have good weeks and bad weeks, and products get bought during bad weeks. Regression to the mean produces an apparent improvement with no active ingredient involved, and the owner who administers the product is usually also the person rating whether it helped.

1.3 What This Article Covers

This article sets out how to read a study on a calming product, then goes through what has actually been found for pheromones, alpha-casozepine, tryptophan, L-theanine and cannabidiol, and ends with what the pattern across all of them means in practice. It is a companion to the article on prescription medication, which is a different evidence situation entirely (psychopharmacology in dogs).

2. How to Read a Calming-Product Study

2.1 Within-Group Is Not Between-Group

This is the single most important distinction in this area, and it accounts for a large share of the gap between headlines and findings.

A within-group result asks whether the treated animals changed over time. A between-group result asks whether the treated animals changed more than the untreated ones. Only the second tells you the product did anything, because the first is compatible with the passage of time, habituation to the setting, and everything else that was happening anyway.

A study can report a statistically significant within-group improvement and a non-significant between-group difference. The honest summary of such a study is that it did not demonstrate an effect. The press release usually says otherwise, and one of the studies below is exactly this case.

2.2 Open-Label Means the Expectation Is in the Data

In an open-label study everyone knows who is getting the product. When the outcome is an owner's rating of their own dog's anxiety, the owner's expectation is part of what is being measured. This is not a criticism of owners; it is a property of the design, and it is why placebo control exists.

Several of the most-cited supplement studies are open-label. They are not worthless — they show that a product is tolerated and that owners will use it — but they cannot establish efficacy.

2.3 Who Rates the Outcome

Even in a blinded study, the choice of outcome measure matters. An owner questionnaire, a blinded observer scoring video, a heart-rate trace and a cortisol value can disagree with each other, and in several of the studies below they do. When they disagree, the honest report says so rather than leading with whichever measure moved (measuring stress in dogs).

2.4 Who Paid

Much of this literature is manufacturer-funded, which is normal in veterinary product research and is not by itself disqualifying. It is a reason to weight independent replication heavily, and independent replication is exactly what is scarce here.

2.5 Significant Is Not the Same as Meaningful

A statistically significant result says the difference is unlikely to be chance. It says nothing about whether the difference is large enough to matter to a dog or to an owner.

In this literature the two get conflated constantly, partly because effect sizes often go unreported. A shift in a cortisol value that reaches significance in a crossover design with fifty-four dogs may correspond to a change no handler would notice. The question to ask of any positive result here is not only whether it was significant but how big it was and on which measure — and for several of the studies below, the answer to the second part is "small, on one measure out of many."

2.6 The Placebo Effect Is Mainly Measured Through the Owner

Dogs do not have expectations about capsules. What a placebo-controlled design in this field controls for is not a response in the dog but a shift in the human: in what the owner notices, how they record it, and how they behave around the dog once they believe treatment has started.

That last part is not trivial. An owner who has just started a product they expect to work may be calmer at the door, may stop bracing on the lead, may stop apologizing in advance. Those are real changes in the dog's environment, and they would produce a real change in the dog without the product doing anything (emotional contagion).

So a positive open-label result is genuinely ambiguous rather than merely weak. Something may well have improved; the question is what caused it.

3. Pheromones

3.1 What Is Being Claimed

Dog-appeasing pheromone products are based on a substance derived from the skin secretions of lactating bitches, on the hypothesis that it conveys information associated with maternal care and security, although its functional relevance in adult dogs remains debated. It is sold as diffusers, collars, sprays and wipes.

The claim is not implausible on its face. Dogs get a great deal of information through the nose, and chemical signaling is a real channel (canine olfaction). The question is whether this particular preparation does what is claimed for it, and that is an empirical matter rather than a matter of mechanism.

3.2 The 2010 Systematic Review

Frank, Beauchamp, and Palestrini conducted a systematic review of prospective studies published between January 1998 and December 2008 on pheromone treatment of undesirable behavior in cats and dogs. They identified 14 reports, seven canine and seven feline, and assessed their quality.

Their finding: 11 of the 14 reports provided insufficient evidence, and one provided a lack of support, for the effectiveness of pheromones in treating undesirable behavior. One study showed adequate evidence that dog-appeasing pheromone could reduce fear or anxiety in a training context.

Only one of the fourteen included reports provided supporting evidence. That review drew published objections at the time, which is itself informative — this is a contested area rather than a settled one — but the underlying observation that the trial literature was small and methodologically weak was not seriously disputed.

3.3 Where It Stands Now

Wong and Govendir revisited the question for dogs over six months of age, appraising eight controlled trials: four randomized controlled trials and four non-randomized studies or trials with unclear allocation.

Their clinical bottom line is that the evidence for using dog-appeasing pheromone to manage stress behaviors associated with anxiety in dogs over six months remains weak. The overall strength of evidence was rated weak.

The picture by context is more informative than the overall rating. They found moderate evidence for reducing fear and anxiety from thunderstorm noise, weak evidence for reducing non-specific stress behaviors in hospitalized dogs, and mild evidence for reducing barking in shelter dogs and increasing relaxation behaviors.

And one detail with direct practical consequences: behavioral improvements typically occurred only during active exposure to the product, with minimal residual effects afterwards. Whatever it does, it does not appear to change anything once it is removed.

3.4 Shelters Specifically

Mooney appraised the shelter case separately and found two studies meeting criteria. The clinical bottom line is that it is not possible to determine whether pheromonotherapy reduces stress in dogs in shelter environments. Both studies showed reduced barking amplitude, which the appraisal declines to interpret as stress reduction without further research, and the evidence was characterized as weak to moderate because of limited sample sizes and short intervention periods.

The barking point deserves emphasis because it recurs across this literature. A quieter kennel is not the same as a calmer dog, and the two can come apart. Barking is one behavior among many and a poor proxy for an internal state (vocal communication in dogs, behavior is not emotion).

3.5 What Pheromones Add Up To

Fifteen years of appraisal produce a consistent verdict: weak overall, with the strongest signal in noise-related fear, no demonstrated carry-over after removal, and a repeated problem of outcome measures that are easy to record rather than informative.

That is not nothing. A moderate rating for thunderstorm noise is a real finding, and the products are cheap and, on the available data, considered well tolerated. It is also nowhere near what the marketing implies, and it does not support using a diffuser as the primary response to a dog with a fear problem (noise sensitivity).

4. Alpha-Casozepine

4.1 What It Is

Alpha-casozepine is a peptide derived from milk protein, sold for dogs and cats on the claim of an anxiolytic effect. It is among the most widely recommended of the behavioral supplements.

4.2 The Appraisal

Buckley reviewed the evidence and reached a clinical bottom line worth quoting for its specificity: there is currently no evidence to show that alpha-casozepine is effective as an anxiolytic when administered to dogs shortly — minutes to a few days — before exposure to an anxiety-provoking stressor.

The appraisal describes the available evidence as limited and weak, with low quality or high risk of bias, and confounding variables providing alternative explanations for the findings. Some weak evidence suggests possible benefit over the medium to longer term, but the quality of that evidence remains poor.

4.3 Why That Distinction Matters in Practice

The short-term use is the common one. An owner gives a capsule before a vet visit, a car journey, or New Year's Eve, on the assumption that it will take the edge off within a day or two. That is precisely the use for which the appraisal found no supporting evidence.

A product with possible medium-term benefit and no demonstrated acute effect is frequently used in a window for which there is no adequate evidence — which is a communication problem as much as an evidence problem.

5. Tryptophan

5.1 The Reasoning and the Reality

Tryptophan is the amino-acid precursor of serotonin, and the argument for supplementing it is that more precursor should mean more serotonin and therefore less anxiety. The reasoning is clean, which is part of why the idea has persisted. Whether the step from dietary precursor to central neurotransmitter availability works that way in a living dog is the question the reasoning skips.

5.2 The Appraisal

Sofyan appraised the question of whether supplementary dietary tryptophan reduces signs of anxiety in adult dogs. Two studies met the criteria: one randomized, double-blinded and placebo-controlled, the other single-blind and placebo-controlled. The evidence quality was rated moderate, which is better than most of what appears in this article.

The results diverged. The first found no overall significant influence of dietary tryptophan as an aid in reducing anxiety and fear-related behavior in anxious dogs. The second found reduced anxiety-related behaviors by owner observation, with no significant difference in urinary cortisol.

The conclusion: tryptophan as a supplementary dietary component is not enough as a sole treatment to assist in reducing anxiety in anxious adult dogs.

5.3 The Detail Worth Noticing

In the second study the owner-reported measure moved and the physiological measure did not. That dissociation appears repeatedly across this literature, and it has two possible readings: either the behavioral change was real and cortisol was too coarse to detect it, or the owners' ratings shifted without the dogs changing.

Neither reading can be ruled out from a single study, which is the honest position. What can be said is that a result resting entirely on owner report, in a design where owners knew what they were giving, is weaker than the same result with two measures agreeing (diet and behavior in dogs).

6. L-Theanine: A Case Study in Design

6.1 The Study

L-theanine, an amino acid derived from tea leaves, is sold for canine anxiety. The trial most often cited for it is worth going through in detail, not because it is bad work — it is clearly reported and its limitations are in its own title — but because it is the clearest available illustration of why design decides what a result means.

Pike, Horwitz, and Lobprise ran an open-label trial in client-owned dogs with a history of storm sensitivity. Dogs were aged one to eight years, healthy on physical examination and laboratory analysis, recruited from general practices, and not being treated for any chronic medical or behavioral disorder.

Owners completed a questionnaire covering eleven individual behavioral manifestations of storm sensitivity on a 0–5 scale for one initial storm. They then began giving the product and completed the same questionnaire for each of five subsequent storms, also rating storm severity and the time the dog took to return to its normal baseline state. Owners were given a standardized protocol of environment and behavior management; no other behavior modification was prescribed. Eighteen dogs completed the trial.

6.2 The Results Look Strong

Global anxiety scores decreased from baseline to exit with a p-value below 0.0001. Time to return to baseline decreased significantly as well, at p = 0.0063. Treatment success was reported for drooling in 83.33% of dogs, following people in 75%, panting in 76.47%, pacing in 78.57%, and hiding in 78.57%. Owner satisfaction with treatment was 94%, or 17 of 18.

The authors conclude that L-theanine can be an effective treatment for storm sensitivity, decreasing the severity of the dog's overall response and the time to return to baseline.

6.3 Why Those Numbers Cannot Carry That Conclusion

Read the design again and three alternative explanations appear, none of which the study can exclude.

There is no control group. Every comparison is a dog against its own earlier self. A p-value below 0.0001 describes how reliably the scores fell, not what made them fall. This is the within-group problem in its purest form: there is no between-group comparison to make, because there is no second group.

The dogs went through six storms in sequence. Repeated exposure to a feared stimulus, under a management protocol, is not a neutral background condition — it is an intervention. Whether repeated storm exposure produces habituation or sensitization in a given dog is exactly the open question elsewhere in this literature (habituation and sensitization), but a design in which every dog receives graded repeated exposure cannot then attribute improvement to a capsule.

Every dog also received a behavior and environment management protocol. That is a second co-administered intervention, given to all participants, with no arm that received it alone. If the management protocol worked, the results would look exactly like this.

Add the fact that all outcomes were rated by owners who knew their dog was being treated, and the 94% satisfaction figure — which measures the owner rather than the dog — and the honest summary is that eighteen dogs improved over six storms while receiving a supplement, a management protocol, and repeated exposure, and that the study cannot say which of the three did it.

6.4 What This Case Is Good For

Two things are worth taking from it, and neither is "L-theanine does not work".

The first is that a very small p-value tells you nothing about causation when there is nothing to compare against. Significance describes the reliability of a difference; the design decides what the difference can be attributed to. A reader who reacts to p < 0.0001 without reading the design has been given a number and no information.

The second is that this study is honest about itself. It says "open-label" in the title, reports its sample size plainly, and describes the management protocol. The problem arises downstream, where a carefully hedged trial becomes a product claim. That pattern — sound reporting, unsound citation — runs through most of this article (operationalizing behavior for measurement).

What L-theanine needs is the study that has not been done: a placebo-controlled, blinded trial in which one arm gets the management protocol alone.

7. Cannabidiol

7.1 Why CBD Needs Its Own Treatment

CBD is the newest entry and the one with the widest gap between market size and evidence. It is also the one where the three available dog studies contradict each other in an instructive way — so rather than summarizing, it is worth going through all three.

7.2 The Study Behind the Headlines

Corsetti and colleagues studied 24 shelter dogs, twelve assigned to treatment and twelve to control by alternating assignment. The treated group received extra virgin olive oil titrated to 5% CBD, one drop per two kilograms of body weight daily, with an increase for obese dogs; the control group received olive oil alone. Treatment ran for 45 days, with behavioral observation at baseline, day 15, day 45, and 15 days after the end of treatment.

The result that was widely reported: treated dogs showed reduced aggressive behavior toward humans following treatment.

The result that was not: the difference in the decrease of aggressive behavior between the two groups was not significant. The authors report this plainly, with the statistics, and also note that other stress-related behaviors — displacement activities and stereotypies — did not decrease. Their own framing is that the findings suggest it is worth doing more research.

This is the within-group versus between-group problem in its clearest form. Twenty-four dogs, assignment by alternation rather than randomization, a within-group change that reached significance and a between-group comparison that did not. A paper concluding that more research is warranted became a headline that CBD from Cannabis sativa could reduce aggression in dogs.

7.3 The Study That Went the Other Way

Marliani, Vaccari, Cavallini and colleagues ran a comparable design with a different outcome. Twenty shelter dogs, ten treated and ten control, 50 days of weight-based microdoses of CBD and CBG below 0.5 mg per kilogram twice daily, against MCT oil without cannabinoids. They ran temperament tests at baseline and at the end using four stimuli — an unfamiliar person, a novel object, a doll, and an unfamiliar dog — and measured hair cortisol.

No behavioral differences were found between the groups at either timepoint. And the CBD-treated dogs showed significantly elevated hair cortisol at the end of treatment, rising from 1.60 to 4.81 pg/mg.

Two caveats belong with that immediately. Hair cortisol is a long-window measure with known interpretive difficulties, and a rise is not straightforwardly bad (chronic stress and cortisol). And twenty dogs is a small sample. But a null behavioral result paired with a physiological marker moving in the unexpected direction is not what "natural and harmless" predicts, and it is the reason the safety question cannot be waved away.

The authors also note that some treated dogs with chronic pain or stereotypies showed reduced aggression and more social behavior, while the group-level analysis was non-significant. That is a hypothesis about who might respond rather than a finding, and the pain angle is worth holding onto (chronic pain and aggression).

7.4 The Best-Designed Study So Far

Flint, Weller, Hunt, and King ran the design the other two lacked: randomized, blinded, and crossover, so every dog served as its own control.

Fifty-four healthy adult dogs — Norfolk Terriers, Beagles and Labrador Retrievers, mean age 3.6 years — each received all four treatments in separate phases: placebo; CBD at 2 mg/kg; CBD at 4 mg/kg; and CBD at 2 mg/kg combined with a blend of L-tryptophan and alpha-casozepine. The stressor was a standardized ten-minute car journey. Measures included blood cortisol, plasma CBD, heart rate, heart-rate variability, activity, and video-coded stress behaviors: whining, lip-licking, postural change, panting.

One result reached significance: the CBD-plus-blend treatment produced a significantly smaller rise in cortisol from baseline to post-stress than placebo. Everything else — including both CBD doses on their own, and all the behavioral and cardiac measures — showed no significant difference. The authors note that large individual variation in CBD absorption limited the detection of treatment effects.

The honest reading: the best-controlled study in this area found a mild effect on one physiological measure, from a combination product, with the single-ingredient arms showing nothing. That is a real finding and a small one, and the authors title it accordingly.

7.5 What the Three Together Say

Twenty-four dogs with a failed between-group comparison. Twenty dogs with a behavioral null and a cortisol rise. Fifty-four dogs with one significant measure out of many, from a blend rather than from CBD alone.

Anyone claiming that CBD is established for canine anxiety is ahead of that. Anyone claiming it is useless is also ahead of it, because the best-controlled study identified a small effect on one physiological measure. What is actually established is that the effect, if there is one, is small enough that three studies with reasonable designs could not agree on it — and that absorption varies enough between individual dogs to complicate every one of them.

8. The Rest of the Shelf, and Why the Evidence Is Thin

8.1 What Is Not Covered Above

Several widely sold ingredients are not covered above, and their absence is itself the finding: for a number of them there is no structured evidence appraisal to report, and the primary studies that exist are frequently open-label, manufacturer-run, or both.

That includes various herbal blends, melatonin for noise fear, and most multi-ingredient "calming" chews, where the combination has typically never been tested as sold even when one component has been studied on its own.

The honest position on these is not that they fail. It is that nobody has run the test that would tell you. An owner buying one of them is running an uncontrolled experiment with a sample size of one, and should be told that rather than given a confidence the literature does not have (operationalizing behavior for measurement).

8.2 Why So Little Has Been Tested Properly

It is worth asking why a market this large rests on a literature this small, because the answer is structural rather than a matter of anyone cutting corners.

Many products marketed as supplements or complementary products are not required to demonstrate the same level of efficacy evidence as licensed veterinary medicines before being marketed. A manufacturer can bring a product to market on plausibility and safety, and has no obligation to run the trial that would settle whether it works. That is not an accusation; it is the incentive structure, and it predicts exactly the literature that exists — a scattering of small studies, several of them run or funded by the people selling the product, with little reason for anyone to fund a large independent replication.

The second reason is that an adequate trial here is expensive relative to the product. Detecting a small effect on a fluctuating behavioral outcome needs a large sample, blinding, a standardized stressor, and more than one outcome measure. The Flint study shows roughly what that costs in effort for fifty-four dogs; the effect sizes involved suggest that settling most of these questions would take considerably more.

8.3 Product Quality Is a Separate Question

Content variability is a known issue with unlicensed supplements generally, and it interacts with the evidence problem in an awkward way: a trial tests one preparation at one concentration, while the product an owner buys may differ. Where a study reports a dose, that dose is part of the finding, and a different product at a different concentration is not covered by it.

The practical consequence is narrow but real. If a product has evidence behind it, the evidence attaches to that preparation and that dose, not to the ingredient in general.

9. What the Comparison Should Be

9.1 The Alternatives Are Not All Equally Thin

A product with weak evidence looks better or worse depending on what it is being compared against. Compared with nothing, a cheap and apparently safe diffuser is defensible. Compared with the interventions that have actually been shown to change canine fear and anxiety, it is a long way down the list.

Systematic exposure-based work — graded exposure with counterconditioning — is one of the interventions with the strongest evidence base for fear and anxiety problems in dogs, and it is the one these products are most often used instead of (desensitization and counterconditioning). Where the problem is severe enough that a dog cannot engage with that work at all, prescribed medication is the evidence-based way to make it possible (psychopharmacology in dogs).

9.2 Environmental Change Is Cheaper Than a Supplement

Several of the changes with the clearest rationale cost nothing and are routinely skipped in favor of buying something.

Undisturbed rest is the most under-assessed variable in a household with a stressed dog (sleep and emotional processing). Giving a dog control over what happens to it — the ability to approach, to withdraw, to end an interaction — has a stronger evidential basis than anything in this article (controllability and predictability). Reducing the accumulation of manageable stressors into an unmanageable state is a scheduling decision, not a purchase (trigger stacking).

None of that is exciting and none of it is for sale, which is part of why it competes badly with a product on a shelf.

9.3 Handling and Husbandry

Where the trigger is veterinary care, grooming or handling, the alternative to a calming capsule before the appointment is a different appointment: cooperative-care preparation, an approach whose benefit has been examined in studies on cooperative handling (cooperative care). That is more work than a capsule and it addresses the actual problem.

9.4 Where an Adjunct Genuinely Fits

None of this makes these products pointless. The defensible use is narrow and worth stating: a product with at least some evidence, in the context that evidence covers, running alongside a behavioral plan, with a defined review point — and with everyone involved clear that it is not the plan.

Used that way, a pheromone diffuser during a firework season, on top of a management and exposure plan, is a reasonable thing to do. Used as the plan, it is a way of feeling that something is being done (welfare science and quality of life).

10. The Pattern Across All of Them

10.1 Six Recurring Features

Small samples. Twenty, twenty-four, fifty-four dogs. These are pilot-scale studies being read as answers.

Weak allocation. Alternating assignment, unclear allocation, and single-blinding appear repeatedly. Of the eight pheromone trials appraised by Wong and Govendir, half were not randomized or had unclear allocation.

Owner-reported outcomes. Frequently the measure that moves, and frequently the only one.

Measures that disagree. Owner report and cortisol going different ways in the tryptophan literature; behavior and cortisol going different ways in the CBD literature. When measures disagree, the effect is at best fragile.

Effects that stop with the product. The pheromone appraisal found minimal residual effect after removal. Nothing in this literature demonstrates a lasting change.

Very little independent replication. Which is what would settle most of the above.

10.2 What That Does Not Mean

It does not mean these products are frauds. Several have plausible mechanisms and at least one context with a moderate evidence rating. It does not mean an owner using one is being foolish. And it does not mean the field will stay here; the Flint study shows what an adequate design looks like, and there is no reason it cannot be repeated with larger samples.

What it does mean is that none of these products is in the same evidential category as an established behavioral intervention or a prescribed medication, and they should not be presented as if they were.

11. What Follows for Practice

11.1 Nothing on This Shelf Is a Treatment

The strongest finding in the entire area is a moderate evidence rating for one product in one context, with no carry-over. That is an adjunct at best. A dog with noise fear, separation distress or fear-based aggression needs a behavioral plan and, where indicated, veterinary assessment including medication — with a supplement, if used at all, sitting alongside that rather than instead of it (desensitization and counterconditioning, psychopharmacology in dogs).

11.2 Rule Out the Medical Cause First

This is where an ineffective product does its real damage. Six weeks of a calming supplement is six weeks in which an undiagnosed pain condition continues (pain masking, visceral pain). The order is: medical, then behavioral plan, then adjuncts.

11.3 If You Try One, Run It Properly

An owner who wants to try a product can at least get information out of it. Change one thing at a time. Define in advance what would count as improvement, in described behavior rather than impression — settles within so many minutes, eats after the trigger, stops pacing by a given time (behavioral assessment). Record for two weeks before starting. Set an end date. And expect the first week to look good regardless, because it usually does.

11.4 Match the Product to the Claim That Was Tested

Alpha-casozepine has no evidence for use in the minutes-to-days window before a stressor, which is how it is mostly used. Pheromone effects appear during exposure and not after. Neither of those is a reason never to use them; both are reasons to use them in the form and timing the evidence addresses rather than the form the packaging suggests.

11.5 Be Careful With the Word "Natural"

One of the CBD studies found hair cortisol rising in the treated group. Whatever that turns out to mean, it is a reminder that a compound with a biological effect can have unintended ones, and that "natural" is a marketing category rather than a pharmacological one. Dosing, interactions with prescribed medication, and product quality are veterinary questions.

11.6 What to Say to an Owner Who Asks

Accurately and without either overselling or sneering: some of these have modest evidence in specific situations, most have little, none has been shown to change a dog's behavior lastingly, and none replaces working on the problem. If they want to try one, help them set it up so the answer is informative. And if they are already using one and feel it helps, there is no reason to take it away — just do not let it stand in for the plan.

12. Summary at a Glance

Within-group change is not evidence of an effect. Only a between-group comparison tells you the product did something, and the most-cited CBD study failed that comparison while reporting the within-group change.

The 2010 pheromone review found one usable study out of fourteen. Eleven of fourteen reports provided insufficient evidence; one provided a lack of support.

The current pheromone verdict is still weak. Across eight controlled trials — four randomized, four not — the overall strength of evidence was rated weak, with moderate evidence only for thunderstorm noise.

Pheromone effects stop when the product does. Improvements occurred during active exposure, with minimal residual effect afterwards.

The shelter question is unresolved. Two studies, both showing reduced barking amplitude, which the appraisal declines to read as reduced stress.

Alpha-casozepine has no evidence for short-term use. Which is how most owners use it: minutes to a few days before a stressor.

Tryptophan is not enough on its own. Two studies of moderate quality, diverging, with owner report and urinary cortisol disagreeing in one of them.

A tiny p-value can mean nothing. The L-theanine trial reported p < 0.0001 for improvement in eighteen dogs — with no control group, six sequential storms, and a management protocol given to everyone.

The CBD studies contradict each other. A failed between-group comparison in 24 dogs; a behavioral null with rising hair cortisol in 20; one significant physiological measure out of many in 54.

The best-designed study found a mild effect from a blend, not from CBD alone. Randomized, blinded, crossover, 54 dogs — and both single-ingredient CBD arms showed nothing.

Individual absorption varies a lot. Which the authors identify as a limit on detecting any effect at all.

The regulatory category helps explain the literature. Many of these products are not required to show the same level of efficacy evidence as a licensed veterinary medicine before being marketed, so there is little incentive to fund the trial that would settle it.

Significant is not the same as noticeable. Effect sizes frequently go unreported, and a cortisol shift that reaches significance may correspond to a change no handler would see.

The placebo effect here sits in the owner. An owner who believes treatment has started behaves differently around the dog, which is a real change in the dog's environment.

Most of the shelf has not been tested as sold. Multi-ingredient calming products are rarely trialled in the combination on the label.

13. Research Gaps

Almost nothing has been independently replicated. The appraisals keep reaching the same conclusion for the same reason: single studies, often manufacturer-linked, rarely repeated by another group.

Sample sizes are pilot-scale. An effect small enough to need a large sample is exactly the effect these products are likely to have, and no study in this area is powered for it.

L-theanine has never been placebo-controlled in dogs. The trial that exists is open-label and single-arm; the missing study is one with an arm receiving the management protocol alone.

Combination products are untested as sold. The one positive CBD finding came from a blend, which means nobody knows which component did it or whether the combination matters.

Duration and carry-over are unstudied. Whether anything here produces change that outlasts administration has not been established for any product.

Responder subgroups are hypothesized, not identified. The suggestion that dogs with chronic pain or stereotypies respond differently is a post-hoc observation in a small sample, and it is the most interesting untested idea in this literature.

Safety has been assumed rather than examined. One study found a cortisol marker moving unexpectedly. Systematic safety data for long-term use of most of these products in dogs does not exist.

14. Conclusion

The evidence for calming products in dogs is not empty and it is not adequate. Fifteen years of structured appraisal have produced one moderate rating, for pheromones in noise-related fear, with no carry-over after removal. For alpha-casozepine there is currently no support for the short-term use immediately before a stressor that is often recommended. Tryptophan is not sufficient alone. Cannabidiol has three studies that disagree, the best of which found a mild effect from a combination product and nothing from CBD by itself.

The recurring problem is not that the products were tested and failed. It is that most of them have been tested in a way that could not have produced a clear answer either way: small samples, weak allocation, unblinded owners rating their own dogs, and outcome measures that disagree with each other when more than one is taken.

For practice that leads somewhere fairly simple. These are adjuncts with modest and mostly unproven effects, they belong after a medical workup and alongside a behavioral plan rather than instead of either, and the honest thing to tell an owner is what the evidence is rather than what the packaging says. A trainer who says "there is weak evidence for this in thunderstorm fear, it seems safe, and it will not fix the problem on its own" is being more useful than one who recommends it confidently — and more useful than one who dismisses it.

Key Insights

Ask whether the treated dogs improved more than the untreated ones. The most-cited CBD study answered no, and was reported as if it had answered yes.

The strongest result in the whole area is moderate evidence for one product in one context. Pheromones and thunderstorm noise.

Pheromone effects do not outlast the diffuser. Improvements occurred during exposure, with minimal residual effect.

Alpha-casozepine before the vet visit has no evidence behind it. The appraisal is specific about that window.

A quieter kennel is not a calmer dog. Reduced barking is the measure that is easy to take, not the one that answers the question.

When owner report and cortisol disagree, the effect is fragile. That dissociation shows up in both the tryptophan and CBD literatures.

"Natural" is not a safety claim. One CBD study found hair cortisol rising in the treated group.

The real cost of an ineffective product is delay. Six weeks on a supplement is six weeks an undiagnosed pain condition keeps running.

If you try one, define success first. Described behavior, a baseline period, one change at a time, an end date.

Read the design before the p-value. Significance describes how reliable a difference is; only the design says what caused it.

Compare it to the right alternative. Against nothing, a cheap diffuser is defensible. Against graded exposure, undisturbed rest and giving the dog control, it is a long way down the list.

Expect the first week to look good. Products get bought during bad weeks, and bad weeks end on their own.

References

Buckley, L. A. (2017). Is alpha-casozepine efficacious at reducing anxiety in dogs? Veterinary Evidence, 2(3), 67. https://doi.org/10.18849/ve.v2i3.67

Corsetti, S., Borruso, S., Malandrucco, L., Spallucci, V., Maragliano, L., Perino, R., D’Agostino, P., & Natoli, E. (2021). Cannabis sativa L. may reduce aggressive behaviour towards humans in shelter dogs. Scientific Reports, 11, 2773. https://doi.org/10.1038/s41598-021-82439-2

Flint, H. E., Weller, J. E., Hunt, A. B. G., & King, T. (2025). Treats containing cannabidiol, L-tryptophan and α-casozepine have a mild stress-reducing effect in dogs. Frontiers in Veterinary Science, 12, 1632868. https://doi.org/10.3389/fvets.2025.1632868

Frank, D., Beauchamp, G., & Palestrini, C. (2010). Systematic review of the use of pheromones for treatment of undesirable behavior in cats and dogs. Journal of the American Veterinary Medical Association, 236(12), 1308–1316. https://doi.org/10.2460/javma.236.12.1308

Marliani, G., Vaccari, L., Cavallini, D., et al. (2024). Assessing the effectiveness of cannabidiol additive supplementation on canine behavior and cortisol levels. Heliyon, 10(10), e31345. https://doi.org/10.1016/j.heliyon.2024.e31345

Mooney, T. F. (2020). It is unclear what impact pheromonotherapy has on stress in dogs in shelter environments. Veterinary Evidence, 5(2), 278. https://doi.org/10.18849/ve.v5i2.278

Pike, A. L., Horwitz, D. F., & Lobprise, H. (2015). An open-label prospective study of the use of l-theanine (Anxitane) in storm-sensitive client-owned dogs. Journal of Veterinary Behavior. https://doi.org/10.1016/j.jveb.2015.04.001

Sofyan, L. (2024). In adult dogs is supplementary tryptophan in the diet effective in reducing signs of anxiety? Veterinary Evidence, 9(4), 686. https://doi.org/10.18849/ve.v9i4.686

Wong, C. F., & Govendir, M. (2021). Can dog appeasing pheromone ameliorate stress behaviours associated with anxiety in mature domestic dogs? Veterinary Evidence, 6(4), 421. https://doi.org/10.18849/ve.v6i4.421