Cannabis (THC) Toxicosis in Dogs: Emergency Management

Bottom line

Cannabis (THC) toxicosis is a high-morbidity, very-low-mortality neurologic intoxication: treat it as a supportive-care problem, not a poisoning you can reverse with a single antidote. The hallmark presentation is obtundation with static ataxia, urinary incontinence (dribbling urine), and hyperesthesia/startle to stimuli, often with mydriasis [1][2]. Manage airway and thermoregulation, give IV fluids and antiemetics, confine to a quiet space, and reserve intravenous lipid emulsion (ILE) for severe/refractory obtundation given THC's extreme lipophilicity [1][3]. Prognosis is excellent with nursing care — the danger is a concentrated edible dose, aspiration, or a co-ingested toxin (chocolate, xylitol, grapes/raisins), not the THC itself [2][4].

Clinical facts

• Toxic (signs) dose: mild signs from ~0.3–0.5 mg/kg THC; moderate signs at >2–3 mg/kg [1].
• Onset: commonly 30–90 min after ingestion, and delayed several hours with edibles or a fasted stomach [6].
• Duration: typically 12–24 h; can persist up to 72 h with large or edible doses (fat-store release, biphasic kinetics) [1][2].
• Minimum lethal oral dose: No median lethal oral dose has been established; the probable lethal dose exceeds 3–9 g/kg of THC-dominant plant material [1]. Experimental oral THC doses of 3–9 g/kg (3,000–9,000 mg/kg) were non-lethal with recovery inside 24 h, and deaths are extremely rare and cluster around concentrated products — the sentinel report is two dogs that died after eating baked goods made with medical-grade THC butter [1][5].
• Mortality: essentially nil with supportive care — all 223 dogs in the largest recent series survived, and only 22 required hospitalization [2].

Toxicokinetics — why THC lingers and why charcoal is repeated

THC is highly lipophilic (log P ~6–7), rapidly distributed to a large volume of tissue, and released slowly back from fat stores, which is why the biologic half-life stretches to ~30 h and signs outlast serum levels [1][6]. It is metabolized in the liver by cytochrome P450 and eliminated predominantly in feces (~85%) with only ~15% in urine [1].

Clinically, the load-bearing fact is enterohepatic recirculation: hepatically excreted THC and metabolites are reabsorbed from the gut, prolonging and re-intensifying signs. This is the rationale for multi-dose activated charcoal — repeating charcoal interrupts the recycling loop rather than just binding the initial meal [6]. The lipophilicity is also the mechanistic basis for ILE as a lipid sink in severe cases [3][6]. As of 2026, edibles drive the majority of exposures reported to poison centers, and because a product meant to be split into many servings is often eaten whole, the effective dose per kilogram can be extreme [2][4].

Clinical signs — the neurologic hallmark and where the textbook is wrong on heart rate

Lead with the neuro exam. In the largest retrospective (223 dogs, 2024), the dominant findings were ataxia (88%), hyperesthesia (75%), lethargy (63%), and urinary incontinence (46%), with vomiting in ~26% [2]. Static ataxia (a swaying, "sea-legs" stance) plus urine dribbling plus exaggerated startle to sound/touch is close to pathognomonic in the right history [1][2]. Mydriasis, hypersalivation, tremors, disorientation, and nystagmus are common; roughly a quarter of dogs are stimulated/agitated rather than depressed [1]. Severe cases progress to stupor or coma, and — with obtundation, hypersalivation and vomiting — aspiration is the principal life threat. Seizures are rare [1].

One nuance worth correcting at the point of care: classic teaching (and older APCC/case data) emphasizes bradycardia, hypothermia, and hypotension [6], and these do occur — but the 2024 series actually found tachycardia (37%), hyperthermia (23%), and hypertension (61%) to be the more common vital-sign abnormalities, with mild hyperkalemia and ionized hypercalcemia also seen [2]. Do not assume a bradycardic, hypothermic picture; measure and treat the vitals in front of you.

Diagnosis — clinical, and do not trust a negative urine stick

Diagnose on history plus the neurologic signature; confirmatory testing is rarely needed to treat [1][2]. The key trap: over-the-counter human urine THC immunoassays are insensitive and unreliable in dogs — canine urine often lacks detectable concentrations of the metabolite the human test targets, so false negatives are common. In the Meola series, dogs with confirmed ingestion tested negative on urine drug screening; a negative stick does not rule out cannabis toxicosis [5][1]. A positive test confirms exposure but does not quantify dose or exclude a co-ingested toxin. When history is absent, treat the clinical picture and call a poison center [4].

Decontamination — get the timing and the airway right

Emesis is the exception, not the rule. Induce vomiting only if ingestion was very recent (within ~30–60 min) and the dog is still asymptomatic — because THC is antiemetic (emesis often fails) and, once neuro signs appear, obtundation makes emesis a direct aspiration hazard [1][6]. Because apomorphine can deepen CNS depression/sedation in an already-obtunded patient, emesis is appropriate only for a recently-ingested, still-asymptomatic dog; never induce emesis in an obtunded or seizing dog.

Activated charcoal is the more useful tool. Give 1–4 g/kg as an aqueous slurry, and — because of enterohepatic recirculation — repeat every ~8 h through the first 24 h in appropriate patients [6]. The overriding caveat is airway protection: in an obtunded dog, secure the airway (and consider NG delivery) before charcoal, or defer it — an aspirated charcoal slurry is worse than the THC. Charcoal is generally reserved for larger or edible ingestions rather than mild plant exposures [2].

Supportive treatment and intravenous lipid emulsion (ILE)

Supportive, symptomatic care is the mainstay and is usually sufficient. Core measures [1][6]:

• Thermoregulation — actively rewarm the hypothermic dog with warm IV fluids/external heat; conversely cool the ~23% who present hyperthermic [1][2].
• IV fluids — support perfusion and hasten elimination (e.g. balanced crystalloid at maintenance-to-resuscitation rates per status) [1].
• Antiemetics — maropitant (1 mg/kg) or ondansetron (0.1–0.2 mg/kg) to reduce vomiting/aspiration risk [1].
• Nursing — confine to a quiet, dimly lit, padded space to blunt hyperesthesia; keep the patient clean and dry, and manage the incontinent bladder [1].
• Bradycardia — atropine (0.02–0.04 mg/kg IV) only if bradycardia is hemodynamically significant [1].
• Agitation — light sedation only if the dog is truly agitated: butorphanol (0.1–0.4 mg/kg) or a benzodiazepine (diazepam 0.25–0.5 mg/kg). Avoid oversedation, which compounds obtundation [1].

Intravenous lipid emulsion (20%) is an off-label adjunct reserved for severe or refractory obtundation/coma, exploiting THC's lipophilicity as a lipid sink [3][6]. A representative protocol: 1.5 mL/kg IV bolus over ~1 min, then 0.25 mL/kg/min CRI for 30–60 min; the bolus may be repeated in non-lipemic patients (some protocols up to ~7 mL/kg total, or the bolus every 4–6 h up to 24 h / 5 doses) [1][6]. In a 2016–2020 cohort of 313 dogs and 100 cats — which included 20 THC poisonings — ILE produced a positive effect in ~74–75% of dogs, but ~4% worsened and ~6% had adverse events (transient obtundation/coma, bradycardia, hyperthermia, GI signs, respiratory distress), so ILE is an adjunct for severe cases, not a routine reflex [3]. As of 2026, flumazenil is an emerging investigational adjunct: a small (n=17) uncontrolled 2024 study reported that 0.01 mg/kg IV transiently improved gait and mentation at 30 min, but the effect is short-lived, the mechanism is unclear, and it should not displace supportive care [7].

The edible trap — screen for and treat the co-toxin

The most dangerous THC case is often not the THC. Because most ingestions are edibles, actively history-take and examine for concurrent toxins and treat each on its own protocol [2][4]:

• Chocolate/methylxanthines — tachycardia, hypertension, tremors, seizures; factor cocoa type and dose into your risk, decontaminate, and monitor ECG/BP.
• Xylitol — hypoglycemia (±acute hepatic injury) in dogs; check and serially monitor blood glucose, supplement dextrose, and monitor liver values.
• Grapes/raisins — idiosyncratic acute kidney injury; decontaminate and monitor renal values with IV fluid support.
• High-fat baked goods/butter — GI upset and pancreatitis risk, and the vehicle that delivers the concentrated, potentially lethal THC dose [5].

Cross-check the sibling references on xylitol toxicosis and, for the therapeutic-vs-toxic cannabinoid distinction, cannabidiol in canine epilepsy. If in doubt about a mixed edible, consult ASPCA Animal Poison Control Center (APCC) or Pet Poison Helpline early — they can quantify the methylxanthine/xylitol/THC load and guide decontamination [4].

Monitoring and prognosis

Monitor mentation, temperature, heart rate and rhythm, blood pressure, and hydration; watch obtunded dogs continuously for aspiration, and check glucose and (with a suspicious edible) renal and hepatic values [1][2]. Most dogs are managed as outpatients or with brief hospitalization and recover fully within 12–72 h [1][2]. Prognosis is excellent with supportive care; it becomes guarded only with a massive/concentrated (edible or butter) dose, aspiration pneumonia, or a serious co-ingested toxin [1][5]. Discharge once the dog is ambulatory, continent, normothermic, and eating, and counsel the owner on secure storage — recurrence is a storage problem.

Frequently Asked Questions

What is the lethal dose of THC in dogs?

No median lethal oral dose has been established; the probable lethal dose exceeds 3–9 g/kg of THC-dominant plant material [1]. Experimental oral THC doses of 3–9 g/kg (3,000–9,000 mg/kg) were non-lethal with recovery inside 24 h [1]. Fatalities are extremely rare and cluster around concentrated products: the sentinel cases are two dogs that died after eating baked goods made with medical-grade THC butter [1][5]. Practically, plant-material ingestions are almost never lethal, but a concentrated edible or butter can deliver a dangerous dose per kilogram, so dose the exposure by product potency, not weight of "weed."

Does activated charcoal help in marijuana toxicosis?

Yes, in the right patient. Because THC undergoes enterohepatic recirculation, activated charcoal (1–4 g/kg) and — importantly — repeated dosing every ~8 h through the first 24 h can interrupt the recycling loop and shorten signs [6]. It is most worthwhile for larger or edible ingestions. The critical caveat is the airway: in an obtunded dog, secure the airway (consider NG administration) or defer charcoal, because an aspirated charcoal slurry causes worse harm than the THC itself [6][2].

When is intravenous lipid emulsion indicated for THC toxicosis?

Reserve 20% ILE for severe or refractory cases — marked obtundation, stupor, or coma not responding to supportive care — because THC's high lipophilicity makes it a good lipid-sink target [1][3]. A representative protocol is 1.5 mL/kg IV bolus, then 0.25 mL/kg/min for 30–60 min, with the bolus repeatable in non-lipemic patients [1][6]. It is off-label and an adjunct, not a routine reflex: in a 313-dog/100-cat cohort ILE helped ~74–75% of dogs but ~4% worsened and ~6% had adverse events, so use it for the sick patient and monitor for transient CNS depression, bradycardia, hyperthermia, and GI or respiratory reactions [3].

Are urine THC tests reliable in dogs?

No — over-the-counter human urine THC immunoassays are insensitive and unreliable in dogs, and a negative result does not rule out cannabis toxicosis. Canine urine frequently lacks a detectable concentration of the metabolite the human test targets, so false negatives are common; in the Meola series, dogs with confirmed ingestion screened negative [5][1]. Diagnose on history plus the neurologic signature (static ataxia, urine dribbling, hyperesthesia, obtundation); a positive test confirms exposure but neither quantifies dose nor excludes a co-ingested toxin [2].

How do you manage an edible ingestion with chocolate or xylitol?

Treat it as a multi-toxin case. Manage the THC supportively, then screen for and treat each co-toxin on its own protocol: chocolate/methylxanthines (ECG/BP monitoring, decontamination, control tremors/seizures), xylitol (serial blood glucose with dextrose supplementation and liver-value monitoring for hepatic injury), and grapes/raisins (IV fluids and renal-value monitoring for acute kidney injury) [2][4]. High-fat butter/baked goods add pancreatitis risk and are the vehicle for the most concentrated — occasionally lethal — THC dose [5]. Call ASPCA APCC or Pet Poison Helpline early to quantify the combined load and cross-check the xylitol-toxicosis reference.

Should I make a THC-intoxicated dog vomit?

Only if ingestion was very recent (within ~30–60 min) and the dog is still asymptomatic [1][6]. THC is antiemetic, so emesis often fails, and once neuro signs appear, obtundation makes induced emesis a serious aspiration hazard — do not induce vomiting in an obtunded, ataxic, or seizing dog [1]. Because apomorphine can deepen CNS depression/sedation in an already-obtunded patient, reserve emesis for a recently-ingested, still-asymptomatic dog. When emesis is off the table, shift to activated charcoal (with airway protection) and supportive care.

How long do signs last and what is the prognosis?

Signs usually appear within 30–90 min (delayed several hours with edibles) and resolve within 12–72 h as THC redistributes out of fat stores [1][2]. Prognosis is excellent with supportive care — every one of the 223 dogs in the largest recent series survived, most without hospitalization [2]. It is guarded only with a massive or concentrated (edible/butter) dose, aspiration pneumonia, or a serious co-ingested toxin [1][5]. Discharge once the dog is ambulatory, continent, normothermic, and eating.

Is there an antidote or reversal agent for THC in dogs?

There is no established antidote — care is supportive, and that is almost always sufficient [1]. Flumazenil is an emerging investigational adjunct: a small (n=17) uncontrolled 2024 study reported that 0.01 mg/kg IV transiently improved gait and mentation at 30 min, but the effect is brief, the mechanism is unclear, and it has not been shown to change outcome [7]. For genuinely severe obtundation, ILE remains the better-supported adjunct [1][3]. Do not delay thermoregulation, fluids, airway protection, and nursing while chasing a "reversal."