Psychopharmacology in Dogs: What the Trials Show and What They Do Not

1. A Question That Is Usually Asked Backwards

1.1 Medication or Training

Few topics in dog behavior are discussed with as much certainty and as little reference to data as psychoactive medication. One camp treats it as chemical restraint that papers over training failures; the other treats it as the missing piece that will finally fix the dog. Both positions share an assumption that the evidence does not support: that medication and behavioral work are alternatives competing for the same job.

The controlled canine trials were mostly designed the other way around. Several deliberately excluded behavior modification in order to isolate the drug effect, and their authors point to combined treatment as the better clinical option. The useful question is therefore not whether medication works, but what it has been shown to change, in which conditions, on what time course, and what it leaves untouched.

1.2 What This Article Is and Is Not

This is an evidence review, not a dosing guide. Psychoactive drugs are prescription-only, and their selection, dosing, monitoring and discontinuation belong to a veterinarian. Nothing here is a recommendation for an individual dog, and dosages appear only where they define what a trial tested.

1.3 How to Read the Evidence

The canine drug literature is unusually well-designed compared with the rest of applied canine research — several randomized, placebo-controlled, multicenter trials exist — and unusually narrow. It covers three indications in any depth, its largest home-based trials rely heavily on owner-reported outcomes, and comes largely from studies funded by the companies that market the products. Each of those features is discussed below rather than treated as a footnote (how outcomes are defined and measured).

2. What Has Actually Been Tested in Dogs

2.1 Separation Anxiety: Fluoxetine

The clearest canine evidence concerns separation-related problems. In a multicenter randomized, placebo-controlled trial, fluoxetine at 1–2 mg/kg daily for six weeks improved overall severity scores in more dogs than placebo, without any behavior modification program; the difference reached significance at some weekly assessments and not others (Landsberg et al., 2008) (what else is known about separation problems).

2.2 Separation Anxiety: Clomipramine

The tricyclic antidepressant clomipramine was tested earlier in a prospective, randomized, double-blind, placebo-controlled, parallel-group multicenter trial. Ninety-five dogs diagnosed with separation anxiety — destruction, house soiling or vocalization in the owner's absence, together with behavior suggestive of hyper-attachment — were randomized to standard-dose clomipramine, low-dose clomipramine or placebo for two to three months (King et al., 2000). All dogs received behavioral therapy as well, so unlike the fluoxetine trial above this study tested the addition of a drug to a behavioral treatment background.

Not every controlled trial produced the same result. In a second study, 49 dogs with separation-related problems were randomized to placebo, low-dose or standard-dose clomipramine, and all of them received behavioral therapy. Clomipramine had a sustained suppressive effect on general activity and a more modest effect on attachment-related contact seeking, but the separation-related signs themselves did not improve significantly more than under placebo, while the behavioral therapy itself was associated with improvement (Podberscek, Hsu & Serpell, 1999).

King et al. (2000) is also the source of the most useful follow-up data in this field, discussed in chapter 5, because the same dogs were tracked after treatment ended (King et al., 2004).

2.3 Compulsive Disorder: Fluoxetine

Controlled work on compulsive behavior predates the modern trial: clomipramine was tested against placebo in dogs diagnosed with canine compulsive disorder in the late 1990s (Hewson, Luescher, Parent, Conlon & Ball, 1998), and an earlier randomized, double-blind, placebo-controlled trial examined fluoxetine in acral lick dermatitis (Wynchank & Berk, 1998), which was then widely treated as a compulsive disorder and would today prompt a closer look at dermatological and painful causes. The cleanest modern trial is more recent. Dogs diagnosed with a compulsive disorder by board-certified behaviorists received fluoxetine at 1–2 mg/kg daily or placebo for 42 days, with owners blinded and no behavior modification program included. Severity ratings improved more often under fluoxetine than under placebo at two, four and six weeks, while the diary-based measures — mean number of compulsive episodes per day and duration of the longest daily episode — showed no significant difference between the groups (Irimajiri et al., 2009).

That split is worth sitting with. Different outcome measures told different stories in the same trial, which is a reason to define in advance what improvement is supposed to mean rather than to assume that a treatment success means the behavior has gone away (what compulsive behavior in dogs is and is not).

2.4 Situational Fear: Three Approved Products

Noise aversion is the indication with products licensed specifically for it, and its registration trials are among the largest in canine behavioral medicine. On New Year's Eve, 182 dogs with a history of firework-associated fear received dexmedetomidine oromucosal gel or placebo up to five times as needed; an excellent or good treatment effect was reported for 72 percent of the treated dogs against 37 percent on placebo (Korpivaara, Laapas, Huhtinen, Schöning & Overall, 2017). In a second trial, 238 dogs received imepitoin at 30 mg/kg twice daily or placebo, started two days before New Year's Eve and continued for three days; owner-reported fear and anxiety behaviors were significantly lower under imepitoin, and a higher proportion of owners reported a good or excellent overall effect (Engel, Müller, Klee, Francke & Mills, 2019).

A smaller follow-up study looked at what happens over repeated events: across 22 dogs and ten noise events, the probability that a dog required treatment with the dexmedetomidine gel fell with successive events rather than rising (Gruen et al., 2020). That study was small and had no placebo group.

A third option has since been added. Tasipimidine, an alpha-2 adrenoceptor agonist given as an oral solution roughly an hour before the expected trigger, has been authorized in the EU since 2021 for the short-term alleviation of situational anxiety and fear triggered by noise or owner departure (European Medicines Agency, 2021), and on 6 May 2026 became the first product approved by the US Food and Drug Administration for both noise aversion and separation anxiety in dogs, on the basis of controlled field studies for each indication: 160 client-owned dogs for noise aversion and an eight-week study enrolling 224 client-owned dogs for separation anxiety (U.S. Food and Drug Administration, 2026a, 2026b).

These are short-term, situational treatments rather than daily medication, which makes them a different clinical proposition from the antidepressants above; imepitoin, however, is started before the expected event rather than at the moment of it (the wider evidence on noise sensitivity).

2.5 What Owners Report

Alongside the trials sits survey data. In a large owner survey on firework fear, prescription medication pooled across drugs was rated effective by 68.9 percent of the owners who had used it — similar to counterconditioning at over 70 percent and relaxation training at 69 percent (Riemer, 2020). Pooled owner ratings cannot tell us which drug did what, and owners who medicate are not a random sample; the figure shows perceived effectiveness, not comparative efficacy.

2.6 Where the Evidence Is Smaller

Much of what is prescribed is used off-label, and off-label does not mean unstudied. Trazodone appears in a retrospective series of 56 cases (Gruen & Sherman, 2008), and in a randomized, double-blind, placebo-controlled crossover trial a single dose of 9–12 mg/kg given 90 minutes before transport reduced behavioral and physiological signs of stress during veterinary visits in 20 dogs (Kim et al., 2022). Gabapentin has controlled canine data as well: in a double-blind, placebo-controlled crossover trial, a single dose of 25–30 mg/kg given at least 90 minutes before a storm reduced fear responses in 18 dogs, with ataxia as the most frequent adverse effect (Bleuer-Elsner, Medam & Masson, 2021). A second double-blind, placebo-controlled crossover trial gave 50 mg/kg two hours before a veterinary visit to 22 dogs and produced mixed results, with most measured outcomes not differing from placebo (Stollar et al., 2022). In a 2026 road-travel model, 36 laboratory Beagles were randomized to trazodone, gabapentin or placebo about two hours before the drive: trazodone reduced behavioral and physiological indicators of travel-induced anxiety, while gabapentin at the tested dose and timing did not differ significantly from placebo (Landsberg et al., 2026).

These studies are small and tied to specific situations and regimens, so they establish neither general anxiolytic efficacy nor a general ranking between the drugs, and neither has a registration-scale field trial for noise aversion comparable to the larger regulatory studies described above. For aggression, the presentation with the highest stakes, controlled evidence is particularly sparse, dated and inconsistent: a small crossover study reported a reduction in what was then called dominance aggression under fluoxetine (Dodman et al., 1996), while in a later randomized, double-blind, placebo-controlled trial in 28 dogs clomipramine was no more effective than placebo (White, Neilson, Hart & Cliff, 1999). Both used older diagnostic concepts that do not transfer straightforwardly to current practice, and nothing approaching the separation-anxiety trials exists (how aggression cases are assessed).

This is not an argument against those drugs. It is an argument for knowing which side of the line a given recommendation sits on, and for saying so to the owner.

2.7 What Was Actually Tested

It is worth being precise about the regimens behind the headline results, because "fluoxetine works" is a claim about a specific dose over a specific period. The separation anxiety trial used 1–2 mg/kg once daily for six weeks (Landsberg et al., 2008); the compulsive disorder trial used the same dose range for 42 days (Irimajiri et al., 2009). The clomipramine trial compared 1 to under 2 mg/kg twice daily with 0.5 to under 1 mg/kg twice daily and with placebo, over two to three months (King et al., 2000).

2.8 What Is Actually Prescribed

Trial evidence and prescribing practice differ. In a retrospective analysis of 32,468,046 canine medical records from US primary care hospitals between 2010 and 2020, trazodone was prescribed in 1.33 percent of records, clomipramine in 0.03 percent and fluoxetine in 0.02 percent. Behavior problem labels rose tenfold, to 10.2 percent in 2020. Among dogs carrying such a label, 8.4 percent received trazodone, 0.15 percent clomipramine and 0.14 percent fluoxetine — and about 70 percent of the dogs prescribed one of these drugs had no behavior problem label recorded (Weng, Morrison, Topdjian & Ogata, 2025).

Trazodone was prescribed far more often than fluoxetine or clomipramine, despite the larger controlled evidence base for the latter two in chronic behavioral indications, and most prescriptions were not documented against one of the behavior labels examined. The database cannot show whether an individual prescription was appropriate or why a drug was chosen, and situational uses of trazodone may not always be captured by those labels; the data document a difference between prescribing frequency and the controlled chronic-behavior literature, not the reason for it. Record-keeping is not clinical reasoning, and a missing label does not prove a missing diagnosis.

3. How to Read These Trials

3.1 Owner Report Is the Outcome

In the major home-based trials, the primary outcome is what the owner reported: severity scores, counts of episodes, global impressions. That is not a flaw unique to this field — the behaviors happen in homes, often when nobody is watching — but it has consequences. Owners who believe their dog is being treated report improvement, which is why a placebo arm is indispensable, and why a substantial caregiver placebo effect has been documented in canine clinical research, including owner and veterinarian assessments in osteoarthritis, which is one reason subjective outcomes need placebo controls (Conzemius & Evans, 2012).

3.2 Funding and Design

The fluoxetine trials for separation anxiety and compulsive disorder were both supported by the manufacturer, as is common in veterinary pharmacology. Industry funding does not invalidate a randomized, blinded, placebo-controlled design; it does mean the questions asked were the ones useful for registration — does this product beat placebo for this indication — rather than the ones clinicians argue about, such as which drug to choose.

3.3 Symptom Change Versus State Change

A drug that reduces a behavior has not necessarily made the dog feel better, and in an animal that cannot report its state that distinction is not academic. One canine study addressed it directly: dogs with separation-related problems treated with fluoxetine alongside a behavior plan shifted toward less pessimistic responding on a cognitive bias test (Karagiannis, Burman & Mills, 2015).

Five dogs in the treatment group, no untreated comparison group, and a combined intervention that cannot separate drug from plan — the study is a demonstration of method rather than evidence of efficacy. Its importance is that it shows the question is answerable with a measure that does not depend on the owner's impression (how judgment bias tests measure affective state).

3.4 What the Cross-Species Picture Adds

A systematic review and meta-analysis of pharmacological manipulations of judgment bias across species reached a qualified conclusion: the effects exist but are inconsistent across designs and drugs (Neville et al., 2020). Read together with the canine trials, the honest summary is that these drugs change something measurable, and that effect sizes and clinical relevance vary between studies, outcomes and substances.

3.5 What Counts as Improvement

The compulsive disorder trial makes the question unavoidable. Owner-rated severity improved more often under the drug while the diary measures showed no group difference (Irimajiri et al., 2009), so whether that trial counts as a success depends on which outcome is treated as the point of treatment.

The same choice has to be made in every individual case, preferably before treatment starts. Fewer episodes, shorter episodes, faster recovery, a dog that can eat or respond during a situation it previously could not, fewer injuries, a household that can manage — these are different targets, and a plan that does not name one will be judged on whichever impression is strongest after a fortnight.

4. What the Drugs Are Supposed to Do

4.1 The Standard Account

The usual explanation runs through neurotransmitters: SSRIs increase serotonin availability, tricyclics act on serotonin and noradrenaline, benzodiazepine-receptor drugs enhance GABAergic inhibition, and alpha-2 agonists dampen noradrenergic arousal. On that account, the effect is to lower the baseline level of fear or arousal enough that the dog can respond differently (how these systems are described in dogs).

4.2 Where That Account Comes From

Much of it comes from human and rodent pharmacology. The receptor work, the circuit models and the reason a drug was tried in dogs at all are extrapolations from other species, supported in dogs by pharmacokinetic work and by the fact that the drugs produce measurable behavioral effects in controlled trials, but not by evidence that a particular neural mechanism produces those effects.

That is a weaker foundation than the confident mechanistic explanations in circulation suggest, and the practical claims in this article do not depend on it. Fluoxetine reduced severity scores in two placebo-controlled canine trials regardless of whether the serotonin account of why is correct.

4.3 "Restoring the Capacity to Learn"

The most common clinical framing is that medication does not fix behavior but restores a state in which learning can happen. It is a reasonable and useful model: fear and high arousal do interfere with learning, and a dog that is less overwhelmed has more capacity to engage (how arousal is measured and why it matters for learning).

There is one piece of indirect canine support worth naming. In the noise-aversion work, the probability that a dog needed treatment fell across successive events (Gruen et al., 2020), which is at least consistent with the idea that experiencing an event without panicking changes what the dog learns from it. The study did not test that interpretation, the decrease was found across successive events overall rather than within repeated events of the same noise stimulus, and it concerns event-based medication rather than daily treatment.

It is also, in dogs, a proposal rather than a demonstrated mechanism — and the distinction matters. That adding a drug to a behavior plan can produce additional clinical benefit has been tested (see 7.2). That the benefit arises because the drug improves learning, or the rate of learning, has not: none of the clinical trials discussed here measured learning as an outcome. Presented as an established mechanism, the framing promises something the evidence has not delivered.

5. Time Course, Duration and Stopping

5.1 How Long Until Something Happens

In the compulsive disorder trial, severity differed from placebo at the two-week assessment and remained different at four and six weeks (Irimajiri et al., 2009). In the separation anxiety trial, the difference was significant at some weekly assessments but not at every one (Landsberg et al., 2008). For daily medication of this kind, a single early assessment is not enough to judge efficacy, and weekly differences in statistical significance are not the same as the drug working one week and failing the next. For situational products the time frame is entirely different: tasipimidine is given about an hour before the expected trigger (U.S. Food and Drug Administration, 2026a), and the dexmedetomidine gel is used around the event itself.

5.2 How Long to Continue

The daily-medication trials ran six weeks to three months. Some situational treatments were evaluated around acute noise events, while the tasipimidine separation-anxiety field study followed dogs for eight weeks (U.S. Food and Drug Administration, 2026b). None of these designs establishes an optimal duration for an individual dog. The clomipramine follow-up is the most informative canine data here, and its message concerns the manner of stopping rather than a duration.

5.3 What Happened After Stopping

The follow-up was an observational re-assessment rather than a randomized withdrawal study, and later treatment was not standardized. More than two weeks after therapy ended, separation anxiety worsened in 13 percent of dogs that had received standard-dose clomipramine, 15 percent of those on the low dose and 23 percent of those on placebo. The mean time to worsening was 37 weeks after standard-dose treatment, compared with 11 weeks after low-dose treatment and 11 weeks after placebo. The authors found no undesirable long-term effects of standard dosing and concluded that abruptly stopping a sub-optimal dose before the signs are well controlled is not advisable (King et al., 2004).

Two points follow. A minority of dogs worsened after treatment ended, in all three groups, which is enough to plan for without calling relapse the norm. And in that trial, standard dosing was followed by a longer interval before signs returned than low dosing was, which speaks against leaving a dog on the lower of the two tested clomipramine doses; the trial did not test starting low and titrating upward, did not establish what dose is adequate for an individual dog, and does not transfer automatically to other drugs. The study was small and did not randomize treatment duration, so it says nothing about how long a course should run.

5.4 Relapse Is Not Failure

The behavioral literature predicts the same thing from a different direction: behavior reduced through learning is not erased, and returns with context change, time and stress (why extinction is context-dependent and reversible). A relapse during or after tapering is information about load and context.

5.5 Tapering and What to Watch

Where a dog is doing well, the question of reducing or stopping comes up, usually from the household rather than the clinic. The canine data speak to it only indirectly: a minority of dogs worsened in all groups, and the standard dose tested was followed by a longer interval before signs returned than the low dose (King et al., 2004).

What follows for practice is modest. Reductions are a veterinary decision and are better made from a position of stability than in the middle of a difficult period; the behavioral measures that were used to judge improvement are the ones to keep running during and after a reduction; and it is worth deciding in advance what a relapse will trigger, so that the answer is not improvised during a bad week.

6. Adverse Effects and Monitoring

6.1 What the Trials Recorded

In the compulsive disorder trial, lethargy or a flat, quiet demeanor and decreased appetite were among the adverse effects recorded in the fluoxetine group, alongside less frequent events including vomiting (Irimajiri et al., 2009). Adverse-effect profiles are substance-specific rather than uniform: ataxia was the most frequent effect in the gabapentin storm trial (Bleuer-Elsner et al., 2021), while reduced activity, lethargy and gastrointestinal effects dominate the reports for the antidepressants — which is not the same as sedation. Many adverse events reported in the antidepressant trials were mild, but type, frequency and time course differ between substances, and product information also lists rarer serious effects, contraindications and interactions, which is one of the reasons the prescription is a veterinary decision rather than a formality.

6.2 The Effect Owners Notice and Misread

A quieter, less responsive dog in the first two weeks is frequently read as the drug working or as the dog being "drugged," depending on the owner's prior position. Neither reading is safe. Reduced activity or lethargy can occur with some of these drugs, particularly early in treatment, but profiles and time courses differ between substances, and fluoxetine is not a sedative. Such changes should not be equated with the intended anxiolytic effect.

6.3 What Belongs to the Veterinarian

Baseline screening, interactions, dose adjustment, monitoring intervals and the decision to change or stop a drug are veterinary decisions, as is the question of what an apparently paradoxical response — more agitation rather than less — means in a given dog. A trainer's contribution is observation and documentation.

7. Medication and Behavioral Work Together

7.1 Why the Trials Look the Way They Do

Two of the trials cited above deliberately excluded behavior modification so that the drug effect could be isolated (Landsberg et al., 2008; Irimajiri et al., 2009). This is good trial design and a poor model of practice: almost nobody treats a dog with fluoxetine and no plan. Those two trials answer the question of whether the drug does anything on its own, which is not the question a household faces.

7.2 What Is Known About the Combination

The combination has been tested. A multicenter randomized controlled trial evaluated fluoxetine at 1–2 mg/kg daily together with behavior management for canine separation anxiety and reported clinical efficacy and safety for that combination (Simpson et al., 2007), which is the basis of the product's licensed indication in conjunction with a behavior modification plan. What such a design shows is the added benefit of the drug on top of a behavior plan — not why that benefit occurs. The clomipramine trial had the same structure: all dogs received behavioral therapy, and the comparison was the drug added to it (King et al., 2000). The cognitive bias study also combined fluoxetine with a standardized plan and could not separate their contributions (Karagiannis et al., 2015).

7.3 A Reasonable Working Position

Medication has placebo-controlled canine evidence for several indications, concentrated in separation-related problems, compulsive behavior and noise aversion, and should be considered where severity, welfare or safety warrant it. Behavioral work addresses what the dog has learned and what the environment asks of it. Neither substitutes for the other, and the plan should say which problem each component solves (how graduated behavioral protocols are built).

7.4 Medical Causes Come First in Both Cases

Before either route, the medical question belongs on the table. In a review of 100 recent behavior cases, a conservative estimate put a painful component in roughly a third (Mills et al., 2020). Pain and anxiety can coexist, and medication for anxiety is not automatically the wrong choice when pain is present — but an untreated painful condition will limit what any behavioral plan achieves, so it has to be investigated and managed in parallel (how chronic pain changes behavior).

7.5 What the Household Has to Do Either Way

Neither route removes the daily work. The dog still has to be prevented from rehearsing the problem, the routine still has to be predictable, sleep and recovery still matter, and someone still has to keep the record that shows whether anything is changing.

8. The Trainer's Role

8.1 Recognizing When a Referral Is Warranted

Trainers do not diagnose and do not recommend specific drugs. What they can recognize is the pattern that warrants a veterinary behavioral consultation: distress that does not settle between sessions, panic rather than avoidance, self-injury, behavior that cannot be interrupted, a dog that cannot take food where it used to, no progress despite a reasonable plan, or a household at the end of its capacity.

8.2 What Not to Say

Two sentences cause avoidable damage. "You don't need medication, you just need to train properly" is a clinical judgment the trainer is not qualified to make, and "ask your vet for something to calm him down" invites a prescription without a diagnosis. The useful sentence is narrower: this is beyond what training alone can address, and a veterinary behavioral consultation is the next step.

8.3 What a Trainer Can Contribute to the Decision

A veterinarian seeing the dog for twenty minutes benefits from a structured record: what happens, how often, how long, in which situations, what was tried, what changed and what did not. Video, an episode log and an account of how consistently the plan was carried out beat an opinion about medication.

8.4 During Treatment

Once a dog is medicated, the behavioral work does not pause; its pacing changes. The same measures apply as before (what is known about accumulating load). The first weeks in particular should not be used to test how much the dog can now tolerate.

8.5 The First Eight Weeks

In the controlled combination trial, medication and behavior management ran together from the start (Simpson et al., 2007), and that is the sensible default. What changes in the first weeks is the level of difficulty rather than whether work happens at all.

A workable structure: in weeks one and two keep management tight and the routine stable, work at levels the dog handles, and record baseline measures and early adverse effects; from week three raise difficulty in small steps. No canine trial has compared different sequences, so this is a convention rather than a protocol.

9. Three Misconceptions

9.1 "It's Just Chemical Restraint"

Sedation and anxiolysis are not the same thing. The registration trials used fear- and anxiety-related behavior as the endpoint rather than sedation, and the dexmedetomidine gel specifically was studied at subsedative doses (Korpivaara et al., 2017). A drug that only flattened dogs would be a poor treatment and a good tranquilizer, and the distinction is one the prescribing veterinarian is making explicitly.

9.2 "It Changes the Dog's Personality"

The measured effects are reductions in severity scores and, in the one study that looked, a shift toward less pessimistic judgment (Karagiannis et al., 2015). The clinical trials reviewed here did not assess lasting changes in temperament as an outcome, and adverse effects such as lethargy, restlessness or changes in aggressive behavior can certainly change how a dog presents. Where owners report that the dog "isn't himself," that observation belongs with the veterinarian rather than being dismissed.

9.3 "It's a Last Resort"

No trial has compared early with late prescribing, so the case against this framing is clinical and welfare-based rather than experimental: a dog rehearsing panic while a sequence of alternatives is tried is suffering in the meantime.

10. Three Situations in Practice

10.1 Severe Separation Distress

Typical picture: panic within minutes of the owner leaving, destruction at exits, self-injury, vocalization throughout. This is one of the best-studied chronic behavioral indications in dogs (Landsberg et al., 2008; King et al., 2000; Podberscek et al., 1999), and one where waiting rarely helps: the dog rehearses panic every time it is left. Behavioral work targets absence tolerance in graduated steps, while management prevents full-blown panic episodes in the meantime.

10.2 Firework and Thunderstorm Season

Typical picture: a dog that panics predictably on a few nights a year. Here the situational products are relevant, and timing matters: a prescription requires a consultation that cannot be arranged on the afternoon of New Year's Eve. In the small uncontrolled follow-up study, the need for treatment fell across successive events rather than rising (Gruen et al., 2020). Behavioral work in the quiet months carries the rest.

10.3 Compulsive and Body-Directed Patterns

Typical picture: licking, spinning or fixating that is frequent and hard to interrupt. The medical work-up comes first here, because skin, orthopedic and gastrointestinal findings are common in these presentations (Mills et al., 2020). Where medication is used, the trial evidence predicts improvement in rated severity rather than disappearance of the behavior (Irimajiri et al., 2009).

11. Summary at a Glance

Fluoxetine has placebo-controlled evidence in separation anxiety — At 1–2 mg/kg daily for six weeks, without behavior modification, it improved overall severity scores in more dogs than placebo (Landsberg et al., 2008).

The clomipramine results are mixed — In 95 dogs receiving behavioral therapy, standard-dose clomipramine improved several separation-related signs more than placebo (King et al., 2000); in an earlier trial with 49 dogs, also on behavioral therapy, the core separation-related signs did not improve significantly more than under placebo (Podberscek et al., 1999).

Outcome measures disagreed in the compulsive disorder trial — Owner-rated severity improved more often under fluoxetine at two, four and six weeks, while diary-based episode counts and durations showed no group difference (Irimajiri et al., 2009).

Three products have regulatory approval for noise-related fear — Dexmedetomidine gel outperformed placebo in 182 dogs on New Year's Eve (Korpivaara et al., 2017) and imepitoin in 238 dogs (Engel et al., 2019); tasipimidine was FDA-approved in May 2026 for both noise aversion and separation anxiety, having been authorized in the EU since 2021 (U.S. Food and Drug Administration, 2026a; European Medicines Agency, 2021).

Stopping requires planning — After treatment ended, separation anxiety worsened in 13, 15 and 23 percent of the standard-dose, low-dose and placebo groups, with the longest mean time to worsening after standard dosing (King et al., 2004).

Adverse-effect profiles are substance-specific — Lethargy and decreased appetite were recorded in fluoxetine-treated dogs (Irimajiri et al., 2009) and ataxia in the gabapentin storm trial (Bleuer-Elsner et al., 2021); rarer serious effects and contraindications are listed in the product information.

Owners report similar perceived effectiveness for medication and training — In an uncontrolled owner survey, prescription drugs pooled together were rated effective by 68.9 percent, against over 70 percent for counterconditioning (Riemer, 2020).

Clinical improvement can coincide with a change in an independent measure of affective state — Dogs treated with fluoxetine plus a behavior plan moved toward less pessimistic judgment, in a five-dog demonstration without an untreated comparison group (Karagiannis et al., 2015).

Prescribing patterns differ from the trial literature — Across 32.5 million primary care records, trazodone was prescribed far more often than fluoxetine or clomipramine, and about 70 percent of dogs receiving one of these drugs had none of the examined behavior labels recorded (Weng et al., 2025).

Off-label does not mean unstudied — Trazodone reduced signs of stress during veterinary visits in 20 dogs (Kim et al., 2022) and gabapentin reduced fear responses during storms in 18 dogs (Bleuer-Elsner et al., 2021), both in placebo-controlled crossover designs too small to establish general efficacy.

12. Research Gaps and Critical Appraisal

The controlled evidence is concentrated, not confined. The largest placebo-controlled datasets concern separation-related problems, compulsive disorder and noise aversion. Smaller controlled studies cover several situational contexts, including veterinary visits and storm fear (Kim et al., 2022; Bleuer-Elsner et al., 2021). For generalized anxiety, a placebo-controlled trial of short-term imepitoin in 60 dogs with a diagnosed generalized anxiety disorder found no group difference in observed anxiety-related behavior after three days, while salivary cortisol fell under imepitoin and rose under placebo (Forster, Engel, Erhard & Bartels, 2020) — a physiological signal without a visible clinical one; for aggression, controlled evidence is exceptionally sparse, dated and inconsistent (Dodman et al., 1996; White et al., 1999). Limited is not the same as absent.

The mechanism behind the combination is untested. That adding fluoxetine to a behavior plan produces additional benefit has been trialed (Simpson et al., 2007). Whether it does so by improving learning, as the standard clinical account claims, was not measured in any of those trials.

The pivotal home-based trials rely on owner report. Severity scores and episode counts come from owners, with a documented caregiver placebo effect in the background (Conzemius & Evans, 2012); the newer situational studies add video coding and physiological measures (Kim et al., 2022). The single study using an independent affective measure had five treated dogs (Karagiannis et al., 2015).

Head-to-head comparisons are scarce. In 25 dogs with tail chasing, clomipramine and fluoxetine did not differ significantly from each other over twelve weeks; clomipramine outperformed placebo across all assessed intervals and fluoxetine in the later ones (Yalçın, 2010). No trial compares a daily antidepressant with a situational product for the same problem.

Long-term data are scarce. Most of the daily-medication trials reviewed here ran roughly six weeks to three months, while the situational studies covered single events or a few days. The clomipramine follow-up is the main source on what happens afterwards (King et al., 2004), and there is little on multi-year treatment.

Funding is concentrated. The two fluoxetine trials were manufacturer-supported, which shapes which questions were asked more than whether the results are trustworthy.

Much of the mechanism is borrowed. The neurotransmitter accounts come largely from human and rodent pharmacology, alongside canine pharmacokinetic work; the canine evidence is that the drugs change behavior, not why.

The "enables learning" claim is untested in these trials. It is a coherent model and the most common clinical justification, and none of the clinical treatment trials reviewed here measured learning rate or learning capacity as an outcome.

Individual variation is undescribed. Which dogs respond, which do not, and what predicts either is not answered by the available trials, which report group differences rather than response profiles.

Welfare outcomes are rarely measured. Most trials measure problem behavior. Whether the dog's overall state improved has been assessed with an independent measure in one small study (Karagiannis et al., 2015).

13. Conclusion

Psychoactive medication in dogs has a limited evidence base, but a broader one than three diagnoses. The largest placebo-controlled datasets concern separation-related problems, compulsive disorder and noise aversion: fluoxetine improved severity scores in separation anxiety without any behavior plan (Landsberg et al., 2008) and in compulsive disorder, where the diary-based episode counts and durations nonetheless showed no group difference (Irimajiri et al., 2009); clomipramine was tested against placebo in 95 dogs (King et al., 2000); and for situational fear associated with noise, controlled evidence supports dexmedetomidine gel (Korpivaara et al., 2017) and imepitoin (Engel et al., 2019), with tasipimidine FDA-approved since May 2026 for both noise aversion and separation anxiety and authorized in the EU since 2021 (U.S. Food and Drug Administration, 2026a; European Medicines Agency, 2021). Smaller controlled studies cover further situations, including veterinary visits (Kim et al., 2022) and storm fear (Bleuer-Elsner et al., 2021): off-label does not mean unstudied, but neither does a twenty-dog crossover establish general efficacy. The combination has been tested too — fluoxetine together with a behavior plan is the basis of the licensed indication (Simpson et al., 2007) — while the mechanism usually named for it has not: none of those trials measured learning as an outcome. Effects, onset, adverse-effect profiles and evidence quality differ substantially between substances and indications, which is why "does medication work in dogs" is the wrong question. After clomipramine was stopped, a minority of dogs worsened, and standard dosing was followed by a longer interval before signs returned than low dosing (King et al., 2004). The defensible conclusion is narrow. Medication can improve clinically relevant behavior in selected canine behavioral disorders and situational fears, sometimes in addition to behavioral work; how much of that benefit reflects changes in emotional state, arousal, learning or something else is unresolved, and findings from one indication do not transfer automatically to another. For practice that means a diagnosis rather than a symptom, a veterinarian rather than a recommendation from the training field, behavioral work continuing alongside, and progress judged on frequency, duration, recovery and what the dog can do.

Key Insights (Takeaways)

• Medication and training are not alternatives. Two trials excluded behavior modification to isolate the drug effect (Landsberg et al., 2008; Irimajiri et al., 2009), while the combination of fluoxetine with a behavior plan was trialed separately and is the basis of the licensed indication (Simpson et al., 2007).

• The largest placebo-controlled datasets concern separation-related problems (Landsberg et al., 2008; King et al., 2000), compulsive behavior (Hewson et al., 1998; Irimajiri et al., 2009) and noise aversion (Korpivaara et al., 2017; Engel et al., 2019), but controlled canine evidence is not confined to them: smaller trials cover veterinary-visit stress and storm fear (Kim et al., 2022; Bleuer-Elsner et al., 2021), and much more sparsely aggression.

• Improvement is not the same as disappearance. In the compulsive disorder trial, owner-rated severity improved while the diary-based episode counts and durations showed no group difference (Irimajiri et al., 2009) — a reminder to define in advance what improvement is supposed to mean.

• Clinical benefit from daily antidepressant treatment is judged over weeks, although group differences can appear earlier and the time course varies by outcome and dog; situational products are given around the event itself. Reduced activity early in treatment is not the anxiolytic effect the treatment aims at.

• Stopping deserves planning. A minority of dogs worsened after treatment ended — 13 to 23 percent across groups — and time to worsening was longest after standard dosing (King et al., 2004).

• The claim that medication "restores the capacity to learn" is a model, not a canine finding. The added benefit of the drug on a behavior plan has been trialed; none of those trials used learning as an outcome measure.

• The trainer's job is recognition, documentation and pacing: refer when distress exceeds what training can address, record what the veterinarian needs, and keep measuring frequency, duration and recovery rather than relying on impressions.

References

Bleuer-Elsner, S., Medam, T., & Masson, S. (2021). Effects of a single oral dose of gabapentin on storm phobia in dogs: A double-blind, placebo-controlled crossover trial. Veterinary Record, 189(7), e453. https://doi.org/10.1002/vetr.453

Conzemius, M. G., & Evans, R. B. (2012). Caregiver placebo effect for dogs with lameness from osteoarthritis. Journal of the American Veterinary Medical Association, 241(10), 1314–1319. https://doi.org/10.2460/javma.241.10.1314

Dodman, N. H., Donnelly, R., Shuster, L., Mertens, P., Rand, W., & Miczek, K. (1996). Use of fluoxetine to treat dominance aggression in dogs. Journal of the American Veterinary Medical Association, 209(9), 1585–1587. https://doi.org/10.2460/javma.1996.209.09.1585

Engel, O., Müller, H. W., Klee, R., Francke, B., & Mills, D. S. (2019). Effectiveness of imepitoin for the control of anxiety and fear associated with noise phobia in dogs. Journal of Veterinary Internal Medicine, 33(6), 2675–2684. https://doi.org/10.1111/jvim.15608

European Medicines Agency. (2021). Tessie (tasipimidine): European public assessment report and product information.

Forster, B., Engel, O., Erhard, M., & Bartels, A. (2020). Short-term imepitoin treatment reduces stress level in dogs with generalized anxiety disorder. Journal of Veterinary Behavior, 38, 67–73. https://doi.org/10.1016/j.jveb.2020.05.006

Gruen, M. E., & Sherman, B. L. (2008). Use of trazodone as an adjunctive agent in the treatment of canine anxiety disorders: 56 cases (1995–2007). Journal of the American Veterinary Medical Association, 233(12), 1902–1907. https://doi.org/10.2460/javma.233.12.1902

Gruen, M., Case, B. C., Robertson, J. B., Campbell, S., & Korpivaara, M. E. (2020). Evaluation of repeated dosing of a dexmedetomidine oromucosal gel for treatment of noise aversion in dogs over a series of noise events. Veterinary Record, 187(12), 489. https://doi.org/10.1136/vr.106046

Hewson, C. J., Luescher, U. A., Parent, J. M., Conlon, P. D., & Ball, R. O. (1998). Efficacy of clomipramine in the treatment of canine compulsive disorder. Journal of the American Veterinary Medical Association, 213(12), 1760–1766. https://doi.org/10.2460/javma.1998.213.12.1760

Irimajiri, M., Luescher, A. U., Douglass, G., Robertson-Plouch, C., Zimmermann, A., & Hozak, R. (2009). Randomized, controlled clinical trial of the efficacy of fluoxetine for treatment of compulsive disorders in dogs. Journal of the American Veterinary Medical Association, 235(6), 705–709. https://doi.org/10.2460/javma.235.6.705

Karagiannis, C. I., Burman, O. H. P., & Mills, D. S. (2015). Dogs with separation-related problems show a "less pessimistic" cognitive bias during treatment with fluoxetine (Reconcile™) and a behaviour modification plan. BMC Veterinary Research, 11, 80. https://doi.org/10.1186/s12917-015-0373-1

Kim, S. A., Borchardt, M. R., Lee, K., Stelow, E. A., & Bain, M. J. (2022). Effects of trazodone on behavioral and physiological signs of stress in dogs during veterinary visits: A randomized double-blind placebo-controlled crossover clinical trial. Journal of the American Veterinary Medical Association, 260(8), 876–883. https://doi.org/10.2460/javma.20.10.0547

King, J. N., Overall, K. L., Appleby, D., Simpson, B. S., Beata, C., Chaurand, J. P., Heath, S. E., Ross, C., Weiss, A. B., Muller, G., Bataille, B. G., Paris, T., Pageat, P., Brovedani, F., Garden, C., & Petit, S. (2004). Results of a follow-up investigation to a clinical trial testing the efficacy of clomipramine in the treatment of separation anxiety in dogs. Applied Animal Behaviour Science, 89(3–4), 233–242. https://doi.org/10.1016/j.applanim.2004.06.003

King, J. N., Simpson, B. S., Overall, K. L., Appleby, D., Pageat, P., Ross, C., Chaurand, J. P., Heath, S., Beata, C., Weiss, A. B., Muller, G., Paris, T., Bataille, B. G., Parker, J., Petit, S., & Wren, J. (2000). Treatment of separation anxiety in dogs with clomipramine: Results from a prospective, randomized, double-blind, placebo-controlled, parallel-group, multicenter clinical trial. Applied Animal Behaviour Science, 67(4), 255–275. https://doi.org/10.1016/S0168-1591(99)00127-6

Korpivaara, M., Laapas, K., Huhtinen, M., Schöning, B., & Overall, K. (2017). Dexmedetomidine oromucosal gel for noise-associated acute anxiety and fear in dogs – a randomised, double-blind, placebo-controlled clinical study. Veterinary Record, 180(14), 356. https://doi.org/10.1136/vr.104045

Landsberg, G. M., Melese, P., Sherman, B. L., Neilson, J. C., Zimmerman, A., & Clarke, T. P. (2008). Effectiveness of fluoxetine chewable tablets in the treatment of canine separation anxiety. Journal of Veterinary Behavior, 3(1), 12–19. https://doi.org/10.1016/j.jveb.2007.09.001

Landsberg, G., Mendes, J., Kosziwka, W., Keys, D. A., Gordon, A., King, A., Araujo, J. A., & de Rivera, C. (2026). Acute trazodone administration reduces behavioral and physiological indicators of travel-induced anxiety in dogs in a road travel model. Frontiers in Veterinary Science, 13, 1851849. https://doi.org/10.3389/fvets.2026.1851849

Mills, D. S., Demontigny-Bédard, I., Gruen, M., Klinck, M. P., McPeake, K. J., Barcelos, A. M., Hewison, L., Van Haevermaet, H., Denenberg, S., Hauser, H., Koch, C., Ballantyne, K., Wilson, C., Mathkari, C. V., Pounder, J., Garcia, E., Darder, P., Fatjó, J., & Levine, E. (2020). Pain and problem behavior in cats and dogs. Animals, 10(2), 318. https://doi.org/10.3390/ani10020318

Neville, V., Nakagawa, S., Zidar, J., Paul, E. S., Lagisz, M., Bateson, M., Løvlie, H., & Mendl, M. (2020). Pharmacological manipulations of judgement bias: A systematic review and meta-analysis. Neuroscience & Biobehavioral Reviews, 108, 269–286. https://doi.org/10.1016/j.neubiorev.2019.11.008

Podberscek, A. L., Hsu, Y., & Serpell, J. A. (1999). Evaluation of clomipramine as an adjunct to behavioural therapy in the treatment of separation-related problems in dogs. Veterinary Record, 145(13), 365–369. https://doi.org/10.1136/vr.145.13.365

Riemer, S. (2020). Effectiveness of treatments for firework fears in dogs. Journal of Veterinary Behavior, 37, 61–70. https://doi.org/10.1016/j.jveb.2020.04.005

Simpson, B. S., Landsberg, G. M., Reisner, I. R., Ciribassi, J. J., Horwitz, D., Houpt, K. A., Kroll, T. L., Luescher, A., Moffat, K. S., Douglass, G., Robertson-Plouch, C., Veenhuizen, M. F., Zimmerman, A., & Clark, T. P. (2007). Effects of Reconcile (fluoxetine) chewable tablets plus behavior management for canine separation anxiety. Veterinary Therapeutics, 8(1), 18–31.

Stollar, O. O., Moore, G. E., Mukhopadhyay, A., Gwin, W., & Ogata, N. (2022). Effects of a single dose of orally administered gabapentin in dogs during a veterinary visit: A double-blinded, placebo-controlled study. Journal of the American Veterinary Medical Association, 260(9), 1031–1040. https://doi.org/10.2460/javma.21.03.0167

U.S. Food and Drug Administration. (2026a, May 6). FDA approves first drug to treat noise aversion and separation anxiety in dogs (CVM Update).

U.S. Food and Drug Administration. (2026b). Freedom of Information Summary, original new animal drug application NADA 141-627, Tessie (tasipimidine oral solution).

Weng, H.-Y., Morrison, J. A., Topdjian, K., & Ogata, N. (2025). Real-world data on behavioral practices for dogs in primary care veterinary hospitals in the United States (2010–2020). Journal of the American Veterinary Medical Association, 263(7), 1–7. https://doi.org/10.2460/javma.24.06.0412

White, M. M., Neilson, J. C., Hart, B. L., & Cliff, K. D. (1999). Effects of clomipramine hydrochloride on dominance-related aggression in dogs. Journal of the American Veterinary Medical Association, 215(9), 1288–1291. https://doi.org/10.2460/javma.1999.215.09.1288

Wynchank, D., & Berk, M. (1998). Fluoxetine treatment of acral lick dermatitis in dogs: A placebo-controlled randomized double blind trial. Depression and Anxiety, 8(1), 21–23. https://doi.org/10.1002/(SICI)1520-6394(1998)8:1<21::AID-DA4>3.0.CO;2-8

Yalçın, E. (2010). Comparison of clomipramine and fluoxetine treatment of dogs with tail chasing. Tierärztliche Praxis Ausgabe K: Kleintiere/Heimtiere, 38(5), 295–299. https://doi.org/10.1055/s-0038-1622860