topical cbd oil for dogs pubmed

Feeding Cannabidiol (CBD)-Containing Treats Did Not Affect Canine Daily Voluntary Activity

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.


Growing public interest in the use of cannabidiol (CBD) for companion animals has amplified the need to elucidate potential impacts. The purpose of this investigation was to determine the influence of CBD on the daily activity of adult dogs. Twenty-four dogs (18.0 ± 3.4 kg, 9 months−4 years old) of various mixed breeds were utilized in a randomized complete block design with treatments targeted at 0 and 2.5 mg (LOW) and at 5.0 mg (HIGH) CBD/kg body weight (BW) per day split between two treats administered after twice-daily exercise (0700–0900 and 1,700–1,900 h). Four hours each day [1,000–1,200 h (a.m.) and 1,330–1,530 h (p.m.)] were designated as times when no people entered the kennels, with 2 h designated as Quiet time and the other 2 h as Music time, when calming music played over speakers. Quiet and Music sessions were randomly allotted to daily a.m. or p.m. times. Activity monitors were fitted to dogs’ collars for continuous collection of activity data. Data were collected over a 14-day baseline period to establish the activity patterns and block dogs by activity level (high or low) before randomly assigning dogs within each block to treatments. After 7 days of treatment acclimation, activity data were collected for 14 days. Data were examined for differences using the MIXED procedure in SAS including effects of treatment, day, session (Quiet or Music), time of day (a.m. or p.m.), and accompanying interactions. CBD (LOW and HIGH) did not alter the total daily activity points (P = 0.985) or activity duration (P = 0.882). CBD tended (P = 0.071) to reduce total daily scratching compared with the control. Dogs were more active in p.m. sessions than in a.m. sessions (P < 0.001). During the p.m. session, dogs receiving HIGH tended (P = 0.091) to be less active than the control (CON). During the a.m. and p.m. sessions, CBD reduced scratching compared with CON (P = 0.030). CBD did not affect the activity duration during exercise periods (P = 0.143). These results indicate that, when supplemented with up to 4.5 mg CBD/kg BW/day, CBD does not impact the daily activity of adult dogs, but may exert an antipruritic effect.


Pet owners and caretakers are increasingly interested in monitoring their animals’ behavior and activity as indicators of health and well-being. While several subjective measures like the Canine Brief Pain Inventory and the Hudson Visual Analog Scale are available for use (1, 2), the ability to measure activity through objective, non-invasive means such as with accelerometers is a potentially preferable tool that can provide an impartial measure of animal activity (3–6). The use of accelerometers, kinesiology, and gait analysis are becoming popular methods by which to evaluate the health status of an animal as well as response to treatment. Several triaxial accelerometers have been validated for the measurement of canine activity and can be easily attached to a collar or harness for home use (7–10). They have been used to evaluate the effectiveness of treatments for osteoarthritis and pruritic behaviors (11–14), the effects of exercise and rest on the voluntary activity of active sled dogs (15, 16), and to predict rest in dogs and sleep in humans (17–19).

Normal activity of healthy dogs is influenced by many factors, including breed, age, degree of socialization, and amount of exercise (20–22). Additionally, canine activity may be negatively influenced by factors such as disease, chronic illnesses like osteoarthritis, or behavioral issues such as anxiety (11, 23, 24). There are also certain circumstances where canine activity needs to be reduced as a result of normal activity in high-energy dogs, pruritic behaviors like scratching, or anxious behaviors like pacing and destruction. Activity may also need to be prevented or reduced following an illness, medical treatment, or surgical procedure (13, 24). In such instances, many turn to medications like sedatives or antidepressants that have been shown to reduce canine activity (25). However, some pet owners may be hesitant to turn to such medications due to potential side effects, cost, or personal bias against their use (26, 27). Instead, they often investigate alternatives to conventional medications, such as cannabidiol (CBD).

Cannabidiol is one of over 100 known cannabinoids produced in the glandular trichomes of Cannabis sativa. There has been considerable interest in the use of CBD for both humans and companion animals due to its reported benefits, such as analgesia, anti-inflammatory, anxiolytic, and sedative effects (28–30). The analgesic effect of CBD has been documented in rodent and human models (31–33), and the use of oral and transmucosal CBD oil formulations increased the Canine Brief Pain Inventory (CBPI) and Hudson scores in dogs with osteoarthritis, suggesting an increase in activity and comfort with CBD use (34, 35). However, despite evidence of an anxiolytic effect of CBD in both rodents and humans with doses ranging from 2.5 to 10 mg/kg (36–38), a recent report failed to demonstrate an anxiolytic effect of treats containing 1.4 mg CBD/kg body weight (BW) in dogs exposed to a noise-induced fear response test (39). Other effects attributed to CBD, such as sedative effects, are thought to be biphasic. Larger doses have been shown to exert sedative effects in both rats and humans, whereas low doses of CBD may increase wakefulness (40–42). While the effect of CBD on sedation has not been specifically investigated in a canine model, a preliminary investigation of the safety of escalating CBD doses in 20 healthy dogs reported mild constitutional adverse events recorded for dogs receiving 1.7–64.7 mg/kg CBD oil, which included both lethargy and hyperesthesia (43). A similar investigation into the safety of a 1:20 Δ 9 -tetrahydrocannabinol (THC)/CBD herbal extract reported mild neurological adverse events, like ataxia and delayed hopping, after single and multiple oral doses of 2 and 5 mg/kg CBD extract (44). While adverse events in both studies were mild and rare, they do highlight the potential of CBD to cause undesirable side effects as well as the need for continuing research evaluating the safety and efficacy of CBD use in dogs.

Despite the lack of scientific evidence demonstrating the safety and efficacy of CBD use in dogs, a recent survey of over 1,000 dog owners recruited on social media showed that almost 80% of the owners surveyed had purchased hemp or marijuana products for their dogs to provide pain relief, relieve anxiety, aid with sleep, and treat other health conditions. Many also indicated that they believed hemp products were more effective than conventional medications (26, 27). The study population included owners of both healthy and diseased animals as well as owners that either had or had not ever purchased hemp products for their dogs; however, this is likely an overestimation of the overall hemp use in companion animals due to the surveys being shared primarily within social media groups dedicated to cannabis use in pets. Nevertheless, these surveys provide insight into the overwhelmingly favorable perceptions of pet owners on the safety and efficacy of CBD use in companion animals. Due to this interest in the use of CBD in companion animals, there is a critical need for further evaluation of CBD use in dogs and its potential effects on canine activity. Thus, the objective of the current study was to determine the impact of CBD on the daily activity of healthy adult dogs with the underlying hypothesis that CBD would reduce the overall daily activity of dogs compared with the control. This hypothesis was tested using triaxial accelerometers to measure the activity of dogs receiving two levels of CBD administration compared to a control.

Materials and Methods

This study was approved by the Lincoln Memorial University (LMU) institutional animal care and use committee (protocol 1911-RES) before the start of the study. All housing and husbandry were provided in accordance with the Animal Welfare Act, the Guide for the Care and Use of Laboratory Animals (8th ed.), and all applicable LMU protocols.

Subjects and Housing

Thirty neutered adult dogs (15 males, 15 females, 9 months to 4 years old, 17.6 ± 3.4 kg) of various mixed breeds, including terrier, hound, Bassett, shepherd, border collie, husky, cur, lab, boxer, and pug mixes, were received at the LMU DeBusk Veterinary Teaching Center (DVTC) from a local shelter for inclusion in this study. The shelter was asked to provide dogs weighing 16 ± 4 kg. Additionally, the shelter was informed and gave consent for the use of the dogs for research purposes before their arrival. Before beginning the experiment, each dog had a complete blood count (CBC) and serum chemistry analysis (IDEXX Laboratories, Inc., Westbrook, ME) performed, along with a physical evaluation by a veterinarian and a fecal examination to rule out any underlying disease that might preclude enrollment. Dogs were excluded if they demonstrated serious behavioral issues, such as extreme fear or human aggression that would endanger research personnel, were severely emaciated or obese, classified as a body condition score <2 or >4 on a five-point scale (where one is emaciated and five is obese), or if the initial evaluations revealed an underlying disease that required more than routine treatments, such as heartworm infection, metabolic or infectious disease, and mobility issues. Three dogs were excluded due to positive heartworm tests and another three dogs excluded for behavioral concerns. The remaining 24 dogs (12 males, 12 females, 9 months to 4 years old, 18.0 ± 3.4 kg) were selected for inclusion in the study. The dogs were individually housed in 1.2 × 1.8-m cages within one of two dog kennels at the LMU DVTC for the duration of the study.

Diets and Treatments

Dogs were fed Purina Pro Plan EN Gastroenteric Fiber Balance Dry Dog Food (Nestle Purina Inc., St. Louis, MO) to meet the daily metabolizable energy requirements of neutered adult dogs at maintenance, calculated as (70 × BW 0.75 ) × 1.6, and split into two meals per day fed between 0700 and 0900 h and between 1,700 and 1,900 h each day. Dogs were weighed and body condition scored (five-point scale) weekly and the diets adjusted accordingly. Treatments were arranged in a randomized complete block design and consisted of 0 (placebo treats, CON), 34.0 ± 1.16 (LOW), or 75.6 ± 5.86 (HIGH) mg CBD/day. CBD is a constituent of a proprietary industrial hemp extract (AgTech Scientific, Paris, KY) that was incorporated into treats and administered in the form of two treats daily, each containing half the daily dose. Both control and CBD treats were composed of the following ingredients: chicken, chicken liver, Asian carp, catfish, and—in the case of CBD treats—industrial hemp extract. While CBD was the primary constituent of the industrial hemp extract, trace THC was present in both LOW and HIGH treatments (1.1 ± 0.37 and 2.9 ± 0.22 mg THC/day, respectively). Treats were formulated to target CBD at doses of 2.5 and 5.0 mg/kg BW/day for LOW and HIGH treatments, respectively, based on an estimation that dogs would weigh an average of 16 kg. The LOW dose was selected based on previous literature that utilized a similar dose in dogs to assess single-dose pharmacokinetics of CBD and to evaluate its potential to alleviate pain in dogs with osteoarthritis (35). That dose was then doubled to achieve the HIGH dosage. However, based on the mean BW of the dogs included in the study and analysis of the treats, the mean doses of CBD were 1.8 and 4.5 mg CBD/kg BW/day for the LOW and HIGH treatments, respectively. Treats were offered solely as a reward upon kennel reentry following twice-daily exercise, which occurred within 30 min of meals.

Experimental Design and Data Collection

Upon completion of intake exams, the dogs underwent a 7-day acclimation period for adjustment to the environment, diet, collars, and daily routine ( Table 1 ). Kennels were maintained on a 12-h light schedule. Dogs received two 15-min exercise periods each day, with the morning exercise occurring between 0700 and 0900 h and the evening exercise occurring between 1,700 and 1,900 h. During the exercise periods, dogs that were aggressive toward other dogs were individually hand-walked by research personnel; all other dogs were allowed to exercise freely in playgroups of two to four dogs in one of two adjacent grassy enclosures. The numbers of dogs being hand-walked and those in playgroups were balanced across all treatments. Four hours each day—from 1,000 to 1,200 h (a.m.) and from 1,330 to 1,530 h (p.m.)—were designated as times when no people were allowed to enter the kennels. Two of those 4 h were designated as Quiet time and the other 2 h as Music time, when calming music was played over speakers in each kennel. Quiet and Music sessions were randomly allotted to either a.m. or p.m. time each day. All dogs started receiving control treats (0 mg CBD) twice daily as a reward for kennel reentry after the twice-daily exercise.

Table 1

Schedule of events for monitoring activity in dogs receiving cannabidiol-containing treats.

Day of study Event Data collection Treats
−2 and −1 Intake and initial health exams None Control
1 to 7 Acclimation None Control
8 to 21 Baseline period activity collection Vetrax activated Control
22 to 28 Treatment adaptation None Treatment
29 to 43 Treatment period activity collection Vetrax activated Treatment

After the acclimation period, Vetrax® activity sensors (AgLogica Holdings, Norcross, GA) were fitted to dogs’ collars using the attachment provided by the manufacturer and placed ventral to the mandible. These triaxial accelerometers were used for the continuous collection of activity variables—activity points, activity duration (in minutes), duration of no activity (in hours), duration of resting (in hours), running duration (in minutes), walking duration (in minutes), scratching duration (in seconds), head shaking duration (in seconds), and sleep quality ( Table 2 ). Data collected by the sensors were automatically uploaded to the Vetrax® server via Wi-Fi once an hour for behavior algorithm processing, which has been previously validated (8). Except for a weekly consistent 2- to 3-h charging period, sensors remained on the dogs at all times. Before the start of the experiment, the data were collected over a 14-day baseline period to block dogs by mean daily activity—high (mean = 118.6 min, range = 88.6–157.5 min) or low (mean = 59.3 min, range = 30.2–85.1 min)—before stratifying dogs by age, weight, and sex and randomly assigning dogs within each block to treatments. Dogs were stratified by treatment and sex, evenly distributed between the two kennels, and adapted to treatments for 7 days before another 14-day collection of activity via Vetrax® sensors ( Table 1 ).

Table 2

Activity variables measured by Vetrax® activity sensors (AgLogica Holdings, Inc., Norcross, GA).

Variable Definition
Activity points Total activity of dogs weighted by each individual activity (e.g., running worth more points than walking), calculated using a proprietary algorithm
Activity (min) Duration of total activity including running and walking
No activity (h) Duration of complete inactivity
Resting (h) Duration of time not actively walking or running, but not completely inactive
Running (min) Duration of running
Walking (min) Duration of walking
Scratching (s) Time spent scratching
Head shaking (s) Time spent shaking head
Sleep Scale of sleep quality measured using a proprietary algorithm based on absence of nighttime disturbance; scaled 0 – 100, with 100 being undisturbed sleep

Consumption of food and treats, consistency of stool, frequency of elimination, subjective assessment of activity during exercise, mucus membrane color, and other indicators of general health status were monitored twice daily by research personnel. Evidence of any adverse event—defined as any symptom occurrence that would not be expected in normal dogs—was also monitored. However, no adverse events were observed in any dogs following the administration of CBD treats during this study.

Statistical Analysis

Based on variations in activity and behaviors reported in previous work using these sensors (8, 13), it was calculated that n = 8 dogs/treatment was sufficient to detect a 25% change with a 16% coefficient of variation (CV) (45). Activity monitors for two of the dogs in the control group (one in the high-activity block and one in the low-activity block) spontaneously stopped transmitting halfway through the treatment period, and activity data from the last 7 days of the experiment for those two dogs were lost.

The normality of the residuals was tested using the UNIVARIATE procedure in SAS (SAS Institute, Cary, NC). In instances where the data did not meet normality assumptions, statistical analysis was performed on transformed data. However, the data were then back-transformed for reporting purposes. The standard errors of the back-transformed data were calculated from the confidence limits of the transformed data as follows: SEM = (back-transformed upper limit – back-transformed lower limit)/3.92. The denominator relates to the Z-value of a 95% confidence interval (±1.96). For the baseline period, activity duration, running, scratching, and head shaking were not normally distributed and were log-transformed for statistical analysis. Activity points and walking were not normally distributed and were transformed into the square root for statistical analysis.

During the baseline period, dogs allotted to CBD treatments tended (P = 0.061) to run more than the control; thus, the mean duration of running from the baseline period was utilized as a covariate for the duration of running in the treatment period. All other variables were similar across treatments in the baseline period. For overall daily activity during the treatment period, running, scratching, and head shaking were not normally distributed and were log-transformed for statistical analysis. For Quiet and Music session activity periods, all variables, except for No Activity and Resting, were not normally distributed and were log-transformed for statistical analysis. For exercise activity periods, activity points, activity duration, and scratching were not normally distributed and were log-transformed for statistical analysis, whereas running and head shaking were not normally distributed and were transformed to the cube root for statistical analysis.

From the treatment period, overall daily activity and activity during the exercise periods (0700–0900 and 1,700–1,900 h) were analyzed using the MIXED procedure in SAS including the fixed effects of treatment, day, and the treatment by day interaction. Dog nested within the activity block (high or low) was included as a random effect and day was included as a repeated measure with dog nested within treatment as the subject. Activity during the Quiet and Music sessions was analyzed using the MIXED procedure in SAS including the fixed effects of treatment, day, session (Quiet or Music), time of day (a.m. or p.m.), and all accompanying interactions. Dog nested within the activity block was again included as a random effect. Time (a.m. or p.m.) was included as a repeated measure with dog nested within treatment as the subject. Treatment effects are described as the contrast between CON and both CBD treatments and the contrast between LOW and HIGH CBD treatments. The results are presented as the mean ± SE. Effects were considered significant when P ≤ 0.05 and considered a tendency when P < 0.10.


Total Daily Activity

CBD did not alter the total activity points, activity duration, no activity, resting, running, walking, head shaking, or sleep quality compared to CON (P > 0.05; Table 3 ). However, CBD tended to reduce scratching compared with CON (P = 0.071), but was not different between the LOW and HIGH treatments (P = 0.209). The level of CBD inclusion (LOW vs. HIGH) did not affect any variables measured (P > 0.05). With the exceptions of activity duration, running, walking, and scratching, all variables were affected by day of treatment (P < 0.05), but there were no treatment by day interactions (P > 0.05).

Table 3

Effect of treatment (TRT), day, and TRT * day interaction on total daily activity variables collected via Vetrax® activity sensors (AgLogica Technology, Norcross, GA).

Treatment SE b P-value
Variable a Control (CON) 1.8 mg CBD/kg BW/day (LOW) 4.5 mg CBD/kg BW/day (HIGH) CON vs. CBD LOW vs. HIGH Day TRT*Day
Activity points 56,834 57,079 56,634 4,952.5 0.985 0.950 0.017 0.857
Activity (min) 80.1 82.7 82.0 11.90 0.882 0.966 0.421 0.773
No activity (h) 13.6 13.9 14.3 0.58 0.496 0.652 0.002 0.394
Resting (h) 8.8 8.5 8.1 0.43 0.364 0.513 <0.001 0.389
Running (min) 5.8 5.6 6.9 0.36 0.612 0.172 0.172 0.447
Walking (min) 76.2 76.5 71.8 10.94 0.878 0.760 0.531 0.749
Scratching (s) 69.6 35.8 51.9 7.33 0.071 0.209 0.162 0.485
Head shaking (s) 32.4 27.3 42.4 5.49 0.878 0.229 0.006 0.194
Sleep quality 77.2 76.1 75.1 2.21 0.500 0.722 0.029 0.679

Treatment effects are shown as the contrast between control (CON) and both CBD treatments (CBD) and the contrast between CBD treatments (LOW vs. HIGH).

Quiet and Music Session Activity

Overall, dogs were more active in the p.m. sessions than in the a.m., with all variables affected by the time of day (P < 0.001; Table 4 ). Activity points, activity duration, running, walking, and resting increased in the p.m. compared to the a.m. (P < 0.001), while the duration of No Activity decreased in the p.m. compared to a.m. (P < 0.001). During these sessions, the Music session tended to reduce activity points (P = 0.055) and running (P = 0.098) compared to the Quiet session. The Music session reduced activity duration (P = 0.002), walking (P < 0.001), and resting (P = 0.045) while increasing the duration of no activity (P = 0.014) compared to the Quiet session.

Table 4

Effect of treatment (TRT), day, session (Quiet or Music), time of day (a.m. or p.m.), and all relevant interactions on activity variables collected via Vetrax® activity sensors (AgLogica Technology, Norcross, GA) at 1,000–1,200 h (a.m.) and 1,330–1,530 h (p.m.) each day.

Treatment P-value
Variable a Control (CON) 1.8 mg CBD/kg BW/day (LOW) 4.5 mg CBD/kg BW/day (HIGH) SE b CON vs. CBD LOW vs. HIGH Time of day Session Session*time TRT*time TRT*session Day
Activity points 2,884 2,509 2,308 148.6 0.152 0.550 <0.001 0.055 0.014 0.287 0.465 <0.001
Activity (min) 1.23 0.87 0.85 0.144 0.204 0.937 <0.001 0.002 0.076 0.079 0.257 <0.001
No activity (h) 1.21 1.25 1.34 0.059 0.273 0.327 <0.001 0.014 <0.001 0.013 0.443 0.002
Resting (h) 0.71 0.70 0.62 0.050 0.384 0.234 <0.001 0.045 <0.001 0.006 0.203 0.002
Running (min) 0.04 0.03 0.04 0.006 0.602 0.544 <0.001 0.098 0.078 0.369 0.834 0.223
Walking (min) 1.29 0.87 0.82 0.148 0.138 0.857 <0.001 <0.001 0.084 0.060 0.217 <0.001
Scratching (s) 4.18 2.29 2.62 0.413 0.030 0.612 <0.001 0.930 0.274 0.326 0.886 0.730
Head shaking (s) 1.78 1.65 1.82 0.128 0.882 0.600 <0.001 0.926 0.305 0.319 0.212 0.375

Treatment effects are shown as the contrast between control (CON) and both CBD treatments (CBD) and the contrast between CBD treatments (LOW vs. HIGH).

Session by time interactions were observed for activity points, no activity, and resting (P < 0.05; Table 4 ), and a trend for session by time interactions was observed for activity duration, running, and walking (P = 0.076, 0.078, and 0.084, respectively). The type of session (Quiet or Music) did not alter activity points or duration of activity during the a.m. session (P = 0.502 and 0.522, respectively). When the Quiet session was allotted to the p.m., however, activity points (P = 0.002), duration of activity (P < 0.001), resting (P < 0.001), running (P < 0.001), and walking (P < 0.001) were increased compared to when the Music session was allotted to the p.m. The durations of No Activity were similar between the Quiet and Music sessions when allotted to the a.m. (P = 0.230), but the duration was increased in the Music session compared to the Quiet session when allotted to the p.m. (P < 0.001).

Activity points, running, and head shaking were unaffected by treatment and all treatment interactions during the Quiet and Music sessions (P > 0.05; Table 4 ). Scratching was reduced by CBD during the Quiet and Music sessions compared to CON (P = 0.030), but the level of CBD inclusion did not affect time spent scratching (P = 0.612).

A treatment by time interaction was observed for No Activity and Resting (P = 0.013 and 0.006, respectively; Table 4 ), and a trend for a treatment by time interaction was observed for activity and walking duration (P = 0.079 and 0.060, respectively). Regardless of Quiet or Music session, activity durations were similar across treatments in the a.m. (P > 0.05), but dogs receiving HIGH CBD tended (1.44 ± 0.172 min, P = 0.091) to be less active than CON (2.64 ± 0.324 min) in the p.m. and tended (1.37 ± 0.162 min, P = 0.059) to walk less than CON (2.60 ± 0.326 min) in the p.m. Similarly, the duration of No Activity was unaffected by treatment in the a.m. (P > 0.05), but in the p.m. tended to increase in the HIGH CBD treatment (1.13 ± 0.060 h) compared to both CON (0.95 ± 0.063 h) and LOW (0.97 ± 0.060 h) treatments (P = 0.054 and 0.068, respectively). Conversely, resting duration increased in the LOW treatment (0.96 ± 0.051 h) compared to the HIGH treatment (0.80 ± 0.051 h) in the p.m. (P = 0.038), but similar across all other time points and treatments (P > 0.05). There were no treatment by session nor treatment by session by time interactions (P > 0.05) for any variables measured. All activity variables were affected by day of treatment period (P < 0.05), but there were no treatment by day interactions (P > 0.05).

Exercise Activity

Neither CBD treatment nor inclusion level affected any variables measured during the exercise periods (P > 0.05; Table 5 ). Day of treatment period tended (P = 0.066) to affect scratching and affected head shaking (P = 0.003), but no other variables were impacted by day of treatment (P > 0.05). Additionally, there were no treatment by day interactions (P > 0.05).

Table 5

Effect of treatment (TRT), day, and TRT * day interactions on activity parameters collected via Vetrax® activity sensors (AgLogica Technology, Norcross, GA) during the two periods of daily exercise, which included all data from 0700–0900 to 1,700–1,900 h each day.

Treatment P-value
Variable a Control (CON) 1.8 mg/kg BW/day CBD (LOW) 4.5 mg/kg BW/day CBD (HIGH) SE b CON vs. CBD LOW vs. HIGH Day TRT*day
Activity points 21,736 25,735 26,122 2,096.8 0.143 0.910 0.117 0.283
Activity (min) 37.66 48.09 48.07 3.213 0.143 0.998 0.528 0.305
No activity (h) 0.82 0.75 0.77 0.132 0.708 0.899 0.359 0.842
Resting (h) 2.47 2.24 2.31 0.116 0.312 0.838 0.338 0.847
Running (min) 4.9 4.5 5.3 0.31 0.940 0.298 0.258 0.531
Walking (min) 35.0 43.8 40.7 0.12 0.207 0.446 0.958 0.576
Scratching (s) 21.1 14.1 20.3 2.53 0.442 0.267 0.066 0.875
Head shaking (s) 17.8 14.7 25.8 0.27 0.682 0.167 0.003 0.273

Treatment effects are shown as the contrast between control (CON) and both CBD treatments (CBD) and the contrast between CBD treatments (LOW vs. HIGH).


Triaxial accelerometer sensors were used in this study to determine the effect of daily CBD dosing on activity in healthy adult dogs by measuring daily activity, pruritic behaviors, and an assessment of rest and sleep quality. The objective of this study was to evaluate the impact of CBD on the daily activity of healthy adult dogs with the hypothesis that CBD would reduce the overall daily activity compared to the control. However, the results showed that oral CBD administration did not alter the overall daily activity of healthy adult dogs. The lack of effect on overall daily activity and sleep quality was unexpected based on previous reports of the sedative and hypnogenic effects of CBD in rodent, human, and canine models. In humans and rats, CBD doses ranging from

2 to 40 mg/kg BW/day have been reported to induce sedative effects, improve sleep quality, and increase total sleep time (41, 46, 47). However, more recent work has reported CBD to have no influence on the sleep cycle in humans (48), and others argue that CBD by itself does not produce sedative effects but rather modulates the sedative effect of Δ 9 -tetrahydrocannabinol (THC), even if THC is only present in minute amounts (42, 49, 50).

The potential for the sedative effect of CBD to be caused by the presence of THC may be supported by an escalating dose study in dogs where placebo, CBD-predominant, THC-predominant, and CBD/THC combination oils were administered to dogs to evaluate the occurrence and severity of adverse events after administration (43). Doses for the CBD-predominant oil started at 1.7 mg CBD/kg BW/day and were incrementally increased to a maximum of 64.7 mg CBD/kg BW/day over 30 days. Lethargy was reported with the CBD-predominant oil formulation. However, that oil was not THC-free; it was reported to contain 0.7 mg/ml THC (43). The industrial hemp extract included in the CBD treats used in this experiment contained a similar THC content to the oil reported in Vaughn et al. (43), but did not produce a similar effect. The reason for these conflicting results remains unclear. These differences could be due to the difference in animals utilized for the study—shelter vs. research-bred dogs—or the different modes of delivery—eating a treat vs. oral gavage of an oil. There have been reports of variations in the pharmacokinetics of CBD depending on the mode of delivery. In one experiment, CBD-infused oil demonstrated an increased maximum plasma CBD concentration compared to the same dose administered as microencapsulated oil beads and a CBD-infused transdermal cream (51). Other reports using similar doses of oral CBD oil and chews showed an increased maximum plasma CBD concentration when administered as a chew compared to an oil; however, this has yet to be investigated in a single, controlled experiment (35, 52). Additionally, the dogs used in Vaughn et al. (43) fasted before the administration of CBD oil, whereas the dogs in the current experiment consumed CBD treats within 30 min of a meal. It has been suggested that administering cannabinoids with a fat meal increased bioavailability (53). Since the CBD used in Vaughn et al. (43) was mixed in a lipid-based formulation, it is unclear whether these differences in methodology would lead to the difference in the sedative effects observed between their report and the current study. Additional investigation using THC-free CBD is needed to evaluate the potential for CBD to exert a sedative effect in dogs.

While there was no observed effect on the overall daily activity with CBD treatment, it tended to influence activity during different times of the day. The dogs in the current study were more active in the p.m. than in the a.m. regardless of treatment and type of session. Playing calming music in the kennels (Music session) did reduce activity compared to when no music was played (Quiet session), which supports previous work showing that playing music can reduce stress and increase relaxed behaviors in kenneled dogs (54–56). This effect, however, appears to be independent of the effect of CBD as there was no interaction between treatment and session nor a treatment by session by time interaction. The tendency for dogs in the HIGH CBD treatment to be less active than CON dogs in the p.m. may indicate that CBD exerted some sedative or calming effect on the dogs. However, this potential sedative effect was expected to be observed in the a.m., as previous pharmacokinetic reports have shown a half-life for CBD of 1–4 h (35, 51, 52, 57). As this effect was not observed during the a.m. sessions, exercise periods, or overall daily activity, these collective results do not support a sedative or calming effect of CBD in dogs. Thus, the claim that CBD exerts a sedative or calming effect in dogs remains unsubstantiated, but further investigation may provide clarification of these results.

In the present study, dogs were necessarily regimented into a strict schedule of daily activities. It is possible that, in a setting where dogs were entirely free to choose their activities, such as a home, the outcome could have been different. The strict, consistent schedule of the kennel environment did not allow for much activity outside of the scheduled exercise periods, which may have prevented normally high-energy dogs from being as active as they could be with consistent free access to more space. Conversely, shelter environments have been shown to increase activity in dogs compared to a home environment and may prevent dogs from resting due to increased stress (58, 59). This may have artificially increased activity in dogs that would have otherwise been less active. As a result, it may be preferable to evaluate the effect of these treatments in familiar environments that have more space for dogs to exhibit normal activity and rest behaviors. Additionally, the small sample size and the use of healthy adult dogs were limitations of this study. The dogs included in this study exhibited high variability in voluntary activity despite being blocked by baseline activity and having no known mobility or behavioral issues. These limitations may preclude the extrapolation of these results to other canine populations. Since CBD is often used to increase comfort and activity in dogs with mobility issues like osteoarthritis or to decrease the activity of anxious or hyperactive dogs (27), future work should evaluate voluntary activity in animals with mobility or behavioral issues like osteoarthritis or anxiety.

The results from this study suggest a potential antipruritic effect of CBD. Phytocannabinoids like CBD act on the body through the endocannabinoid system (ECS), which is a signaling system including endocannabinoids like anandamide and 2-arachidonylglycerol, their receptors, and regulatory enzymes (60). The ECS helps regulate metabolic homeostasis, thermoregulation, epidermal homeostasis, and more (61, 62). While CBD has little to no affinity for CB1 and CB2 ECS receptors, it is a known agonist for the transient receptor potential vanilloid family of receptors (TRPV1-4), which are known ECS receptors widely expressed in the skin and play a role in itch sensation (61, 63–65). As TRPV1 is rapidly desensitized after activation, it is thought that CBD may exert antipruritic effects by keeping TRPV1 desensitized, thus preventing neuronal activation by irritants (66–68). Additionally, CBD has been shown to be an antagonist for transient receptor potential melastatin 8 (TRPM8) receptors (67, 69). In the skin, TRPM8 is responsible for environmental cold detection and has been suggested to contribute to the perception of pain and itch, which may indicate that it is another target for the potential antipruritic effect of CBD (64, 70). The antipruritic effect of cannabinoids has been observed in humans (71–73), but this is the first report of a potential antipruritic effect of CBD in dogs as a reduction in scratching duration was observed in dogs. While this experiment was not designed to assess the antipruritic effect of CBD, these results may suggest a potential for CBD to be beneficial in the treatment of skin conditions and pruritic behaviors in dogs. To investigate this potential effect, it would be beneficial for future work to specifically examine the effect of CBD in dogs with skin issues such as allergies, atopic dermatitis, or unexplained pruritus.


The results of the current study indicate that when supplemented with up to 4.5 mg/kg BW/day, CBD does not impact the overall daily activity of adult dogs. Total daily activity including duration of the activity, sleep quality, and resting were unaffected by CBD. Similarly, activity during the exercise periods was also unaffected by CBD. During the Quiet and Music session periods, 4.5 mg CBD/kg BW/day tended to reduce activity in dogs compared to both 1.8 mg CBD/kg BW/day and CON, but this did not translate to an overall daily effect. Playing classical music in the kennels reduced activity compared to having no music played, but did not alter the response to CBD. CBD reduced total daily scratching as well as scratching during the Quiet and Music sessions, which may indicate a possible antipruritic effect. Future work examining the effect of CBD on activity is warranted, particularly in dogs with mobility and behavioral issues like osteoarthritis and anxiety. Additionally, the potential antipruritic effect of CBD should be investigated using dogs with dermatological issues like skin allergies or atopic dermatitis.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics Statement

The animal study was reviewed and approved by Lincoln Memorial University institutional animal care and use committee (protocol 1911-RES). Written informed consent was obtained from the owners for the participation of their animals in this study.

Author Contributions

DH, EM, KM, and SK-M contributed to the conception and design of the study. EM, SK-M, DS, and JG facilitated data collection. EV and EM performed statistical analysis. EM wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.


We thank Lincoln Memorial University students K. Athey, H. Barnhart, L. Calvin, K. Dubois, J. Gauldin, S. Swears, M. Kight, M. Mendoza, J. Steen, S. Swears, and K. Williams for their assistance in caring for the dogs and facilitating data collection.


Funding. The authors declare that this study received funding from AgTech Scientific, Paris, KY. The funder was not involved in the study design, collection, analysis, interpretation of data, the writing of this article, or the venue for publication.

Prevalence and Characteristics of Cannabis-induced Toxicoses in Pets: Results from a Survey of Veterinarians in North America

Cannabis legalization in North America has coincided with an increase in reports of cannabis-induced toxicosis in pets, but the magnitude of this problem, as well as outcomes of these incidents remain unknown. Therefore, we examined the frequency, diagnostic criteria, clinical signs, and prognoses of cannabis toxicoses in pets in North America. We conducted an online survey between January, 2021 and April, 2021 targeting veterinarians practicing in Canada and the United States (US). Out of the 251 study participants, 191 practiced in Canada. Cannabis toxicosis was most commonly reported in dogs (n=226 veterinarians), and the number of toxicosis cases increased significantly in Canada (p<0.0001) and the US (p=0.002) after October, 2018. Frequently reported clinical signs of cannabis toxicosis included: urinary incontinence (n=195), disorientation (n=182), ataxia (n=178), lethargy (n=150), hyperesthesia (n=134), and bradycardia (n=112). Edibles were most commonly suspected to be the cause of toxicosis (n=116). The most common route of exposure was ingestion (n=135), while the most cited reason was ingestion while unattended (n=135). Cannabis toxicosis was mostly diagnosed using supportive clinical signs (n=229), the most common treatment was outpatient monitoring (n=182), and pets were most often treated as out-patients (n=103). The legalization of cannabis use in Canada and the US is likely an important factor associated with the increased cannabis toxicosis cases in pets; however, the legal status may also increase reporting. The medicinal use of cannabis by pet-owners for pets may also contribute to a portion of the reported toxicoses. Most pets that experienced cannabis toxicosis recovered completely, suggesting that most cannabis toxicoses do not result in long-term ill effects. Even though some deaths (n=16) were reported in association with cannabis toxicosis, the presence of confounders such as toxins, and underlying conditions cannot be ruled out, emphasizing the need for rigorous controlled laboratory studies to investigate this important issue.


With the widespread legislative changes legalizing cannabis across most of North America, cannabis has become the object of considerable public health and policy discussions (1). The increased accessibility to cannabis has prompted an increased interest for its therapeutic value in human and, more recently, veterinary medicine (2). In fact, the sales of cannabis products for pets have increased by 1000% between 2016 and 2017 and a survey found that 79.8% of Canadians have previously bought cannabis products for their dog(s) (3). Although research is ongoing, there are only a handful of published studies that examine the clinical use of cannabis in veterinary medicine, and even fewer have examined basic pharmacokinetic and toxicology data (4). Due to this, the education of veterinarians and pet owners is hindered, resulting in intentional or accidental cannabis exposure of pets without proper oversight or knowledge. In fact, a study in Colorado found a strong correlation between the number of registered medical cannabis cardholders and cases of cannabis toxicosis in dogs, with a 4-fold increase in reported cases between 2005 and 2010 (5). Additionally, over the past 6 years, there was a 448% increase in reports of cannabis poisoning cases in companion animals in the United States (USA) and Canada (6). The Animal Poison Control Center has also reported a 765% increase in calls regarding pets ingesting cannabis in 2019 compared to the previous year (ASPCA, 2019). In Canada, as expected, the total number of cases reported are fewer than in the USA, but they have been increasing since 2018 according to the Canadian Veterinary Medical Association (CVMA, 2019).

Taken together, these data suggest that cases of cannabis toxicosis in companion animals are on the rise, warranting further investigations into these incidents. The aim of our study was to gather relevant information from veterinarians in clinical practice regarding cannabis toxicoses. More specifically, our objectives were: a) to examine veterinarian-reported trends in the frequency of cannabis toxicoses pre- and post-legalization October 2018 (date of legalization in Canada); b) characterize diagnostic criteria used for cannabis toxicoses; c) characterize the clinical signs and prognoses of veterinary-reported cannabis toxicoses; d) identify any evidence supporting the lethality of cannabis in companion animals.


To assess cannabis toxicosis in companion animals in both Canada and USA, we designed an online survey using Qualtrics (Provo, Utah, USA). Participants were practicing veterinarians in either Canada or the USA, who were treating pets, and had been presented with cases of cannabis toxicosis. The duration of the survey was from 28 th January, 2021 to 30 th April, 2021. Before launching the survey, it was pre-tested by the authors and their colleagues, and feedback on the content of the survey was obtained and incorporated into the final version. The Canadian Association of Veterinary Cannabinoid Medicine, Canadian Veterinary Medical Association, Alberta Veterinary Medical Association, Nova Scotia Veterinary Medical Association, Newfoundland and Labrador Veterinary Medical Association, and the Ontario Veterinary College supported with recruitment of participants by distributing the survey to their members. Participants were also encouraged to distribute the survey to their colleagues. The link to the survey was distributed via websites, regular e-newsletters, and magazines of the aforementioned associations. All the data were collected anonymously in Qualtrics. The need for approval by the Institutional Ethical Review Board was waived due to the nature of the survey questions being only about their veterinary practice, and no personal data were collected. Additionally, the data collected were anonymized and participants were informed on the first page of the survey questionnaire that completing the survey implied consent to participate. Except for questions related to consent to participate and eligibility to participate in the survey, participants could opt not to answer any question, and could decide whether to complete the survey or not. Details of the online survey can be found in S1 File. We initially intended to compare the trends in toxicosis cases in Canada and the US; however, the number of US participants was too low for such a comparison to be made. Therefore, while we have described data obtained from veterinarians practicing in the USA in some instances, we have chosen to describe combined data from both Canada and the USA in cases where no differences were observed between the two countries.

Statistical analysis

At the end of the survey, the data were exported from Qualtrics to Microsoft Excel for analysis. All data were analyzed using libraries in Python (version 3.0). Since the data obtained were categorical, and therefore not normally distributed, non-parametric tests were used for analysis. Unless otherwise stated, the results reported represent the number of participants that responded to a specific question. Descriptive statistics were performed in Python. Comparison of the number of cannabis toxicosis cases pre- and post-legalization was performed using the Wilcoxon-signed rank test in Python. The number of pets treated according to hospital setting and practice type was analyzed using the Kruskal Wallis test followed by the Dunn’s post-hoc test in Python. The Chi-squared Goodness of fit test was used to compare the frequencies of each clinical sign, and for clinical signs with significantly different frequencies, a post-hoc binomial pairwise test was conducted. Differences were considered statistically significant when the p-value was less than 0.05. Graphs were plotted using GraphPad Prism version 6.01 (GraphPad Software Inc., La Jolla CA, USA).



Out of the 251 participants who began the survey, a total of 222 participants completed it (Figure 1A); 7 participants were excluded, as they did not meet the eligibility criteria. Most of the veterinarians practiced in Canada (n=191; Figure 1B) with the majority practicing in the province of Ontario (n=108; Figure 1C). Most veterinarians worked in urban areas (Figure 1D) and practiced general medicine (Figure 1E).

A. Graph showing the number of participants who completed the survey and those who did not. B. Graph showing the countries in which the study participants practice. C. A pie chart showing the distribution of participants who practiced in Canada according to provinces. D. Graph showing the hospital setting in which participants practice. E. Graph showing the type of medicine practiced by participants.

Veterinarians reported that cannabis toxicoses were most often observed in dogs, followed by cats, iguanas, ferrets, horses, and cockatoos (Figure 2A – inset). The number of toxicosis cases reported among all surveyed veterinarians was significantly higher after October 2018 (p<0.0001; Figure 2A). This trend was observed both in Canada (p<0.0001; Figure 2B) and in the US (p=0.002; Figure 2C). In all hospital settings and practice types, the numbers of cannabis toxicosis cases were reported to be higher post-legalization (all p≤0.0001; Figure 2D-I).

A. All reported cannabis toxicosis cases pre- and post-legalization. Inset: Species in which cannabis toxicosis was observed and the number of participants who reported them. B. Reported cannabis toxicosis cases pre- and post-legalization in Canada. C. Reported cannabis toxicosis cases before and after October 2018 in the US. D. Reported cannabis toxicosis cases pre- and post-legalization in urban settings. E. Reported cannabis toxicosis cases pre- and post-legalization in sub-urban settings. F. Reported cannabis toxicosis cases pre- and post-legalization in rural settings. G. Reported cannabis toxicosis cases pre- and post-legalization by participants who practice emergency medicine. H. Reported cannabis toxicosis cases pre- and post-legalization by participants who practice general medicine. I. Reported cannabis toxicosis cases pre- and post-legalization by participants who practice other types of medicine. Vets: Number of veterinarians who reported being presented with cannabis toxicosis in a particular species.

Changes in reports of toxicosis before and after 2018

At the individual level, most veterinarians reported no difference in the number of cannabis toxicosis cases annually pre- and post-2018 (Figure 3A-C). However, almost all veterinarians who reported changes between the two periods reported an increase in the number of cases they observed, in both Canada (Figure 3B) and the US (Figure 3C).

A. Graph showing changes in cannabis toxicosis case numbers reported by all participants. Insets: Pie chart showing number of participants who reported equal number of cases pre- and post-legalization (no change) and those who reported different numbers of cases pre- and post-legalization (change) B. Changes in cannabis toxicosis case numbers reported by participants in Canada. C. Changes in cannabis toxicosis case numbers reported by participants in the US. Increase: participants who reported increases in numbers of cannabis toxicosis cases pre- and post-legalization; Decrease: participants who reported decreases in numbers of cannabis toxicosis cases pre- and post-legalization. Inset pie chart: No change: participants who reported equal numbers of cannabis toxicosis cases pre- and post-legalization; Change: participants who reported different numbers of cannabis toxicosis cases pre- and post-legalization.

Commonly observed clinical signs

The clinical signs that veterinarians reported to have observed most commonly (in decreasing order) were: urinary incontinence, disorientation, ataxia, lethargy, hyperesthesia, bradycardia, stupor/obtundation, and twitching (Table 1). A small number of veterinarians reported witnessing other signs including head bobbing and hyperthermia. The Chi square Goodness of Fit test and the post-hoc binomial pairwise test revealed that urinary incontinence, disorientation, ataxia, lethargy, hyperesthesia, and bradycardia were the clinical signs that occurred most frequently (Table 1). Interestingly, except for bradycardia, all these clinical signs were reported to be usually severe.

Products and routes of exposure

As shown in Figure 4A, the products that often led to cannabis toxicosis in pets were edibles and dried cannabis. Other products reported by veterinarians to cause cannabis toxicosis were discarded joint butts, human feces, cannabis-infused butter/oil, and compost (Figure 4B). Most veterinarians (n=105/196) reported that they or the pet owner did not know the source of cannabis exposure. However, among those who reported the sources of cannabis products that led to cannabis toxicosis (n=101/196), most (n=34/196) reported that they were obtained from government regulated producers, followed by home cultivated plants (n=29/196), and the black market (n=28/196) (results not shown). The most common route of exposure (Figure 4C) was ingestion, and ingestion while unattended was the most cited reason for exposure (Figure 4D).

A. Graph showing the products that were reported to cause cannabis toxicosis by the study participants. B. Graph showing other products that were reported to cause cannabis toxicosis by the study participants. C. Graph showing the route of exposure to the products that caused cannabis toxicosis. D. Graph showing the reasons for exposure to the products that caused cannabis toxicosis.

Diagnosis and Treatment

Pets that presented at veterinary hospitals were diagnosed with cannabis toxicosis based on supportive clinical signs, a history of possible/known exposure, and/or the use of over-the-counter urine drug tests (Figure 5A). Following diagnosis, the most common treatments included: outpatient monitoring and supportive care, administration of intravenous fluids, in-hospital monitoring only, administration of activated charcoal, induction of emesis, administration of anti-emetics, thermal support (warming/cooling), and blood pressure monitoring (Figure 5B). Animals were usually treated either as outpatients or they were hospitalized for less than 24 hours (Figure 5C). Most participants reported that all clinical signs resolved following cannabis exposure, except for a few pets that reportedly died in association with cannabis toxicosis (n=16 animals). The cost of treatment for majority of the cases was less than CAD$ 500 (Figure 5D).

A. Graph showing methods used by participants to diagnose cannabis toxicosis. B. Graph showing various treatments for cannabis toxicosis reported by participants. C. Graph showing treatment duration following cannabis toxicosis. D. Graph showing cost of treatment for cannabis toxicosis. E. Graph showing the number of participants who either reported deaths or no deaths associated with cannabis toxicosis in pets. F. Graph showing the causes of deaths reportedly associated with cannabis toxicosis.

Even though most of the veterinarians reported no deaths (211/221), 10/221 veterinarians reported a total of 16 deaths believed to be attributable to cannabis toxicosis (Figure 5E). Other than euthanasia (n=2), the causes of death reported to be associated with cannabis exposures were aspiration pneumonia (n=5), respiratory arrest (n=3), uncontrolled seizures (n=2), coma (n=2), and pancreatitis (n=1) (Figure 5F).


Cannabis toxicosis was frequently reported in dogs, and in both Canada and the US, the number of cannabis toxicosis cases increased significantly after October 2018 (which coincided with legalization in Canada, but not the US). Additionally, of those who reported a change (85/211), nearly all (82/85) reported an increase in the number of cases.. Among the reported clinical signs of cannabis toxicosis (primarily observed in dogs, therefore clinical signs we report herein are likely biased towards canine-specific presentations), urinary incontinence, ataxia, disorientation, bradycardia, hyperesthesia, and lethargy were most common. The product which often caused cannabis toxicosis was edibles, and the most common route of exposure was via oral ingestion, with the most common reason being ingestion while unattended. Diagnosis was frequently based on the presence of supportive clinical signs, and the most common treatment was outpatient monitoring, which lasted for less than 48 hours. Except for a few patients that were reported to have died in association with cannabis exposure, all patients recovered completely after treatment, with a total treatment cost less than CAD $500.

Similar to other studies (7), the pets that were treated most often by the veterinarians in our sample were cats and dogs. This is consistent with a recent survey which revealed that there were 7.7 million dogs and 8.1 million cats in Canadian households (8). In our study, cannabis toxicoses were frequently observed in dogs compared to cats, similar to that previously reported (7). Consistent with previous work, participants also reported cannabis toxicoses in other companion animal species such as horses and iguanas (7, 9), and also in previously unreported species such as pet cockatoos and ferrets.

Similar to a number of previous studies (5, 10), we observed an overall increase in the number of cannabis toxicosis cases after October 2018, even though our analysis at the participant-level revealed that majority of the participants reported equal case numbers pre- and post-legalization. This could be because participants did not report the actual case numbers but rather selected among predetermined numeric ranges. As such, small increases that were within the same range, would have been reported as “no change.” The increase in case numbers could be due to any combination of the following factors: 1. legalization of cannabis for medical and recreational use in Canada; 2. increased reporting by pet owners due to legalization; and 3. increased awareness of veterinarians about cannabis toxicoses (5, 10). Moreover, it is important to consider the pharmacology of cannabinoids in cannabis, and how they may have contributed to the findings noted above.

Pharmacokinetics and pharmacodynamics of THC

The two major cannabinoids in cannabis are delta-9-tetrahydrocannabinol (THC) and its isomer, cannabidiol (CBD) (11). THC is the main psychoactive cannabinoid that produces euphoria but can also have intoxicating effects (11, 12). Meanwhile, CBD is considered non-psychoactive despite its anxiolytic, antipsychotic, and anti-inflammatory effects (13-15). Despite the difficulty in determining the exact time to onset of clinical signs following cannabis toxicosis since owners only seek medical attention upon the appearance of clinical signs; in dogs, the onset of clinical signs ranges from within minutes post-inhalation (16) to several hours post-ingestion (17, 18). This delay in onset may be due to the long biological half-life of THC in dogs due to adipose tissue storage, even though the THC plasma half-life is relatively short (16). The delayed onset of clinical signs may also be due to the time it takes for THC to undergo first-pass liver metabolism post-ingestion (17). When compared to humans, dogs seem to have similar oral absorption, but a much longer duration and wider range of clinical signs. Dogs produce the additional THC metabolites 8-OH-Δ 9 -THC and 11-OH-THC, which may contribute to the additional clinical signs observed only in dogs (16, 19). 11-OH-THC is an active metabolite that may be produced in larger quantities after cannabis ingestion following first pass metabolism.

Given that dogs were the species in which cannabis toxicosis was reported most frequently in our study, and that majority of the signs of cannabis toxicosis were neurological, similar to what others have reported in dogs, (18), the subsequent discussion will predominantly focus on these neurological signs. Among the clinical signs of cannabis toxicosis reported in our study, the most common was urinary incontinence. Expression of cannabinoid receptors has been demonstrated in the bladders of humans, rats, and mice (20-22). Even though the mechanism by which THC regulates bladder contractility in vivo is unclear, it was previously shown that in mice, the administration of THC in bladder tissue inhibited electrically-evoked contractions of the bladder (21), which could lead to urinary incontinence. This was also demonstrated in rats using a CB1 receptor agonist (20). Cannabinoid receptors have not yet been confirmed in the dog bladder; however, if cannabinoid receptor expression is conserved, bladder hyperactivity through increases in contraction could be contributing to the reported incontinence.

Cannabis toxicosis also resulted in bradycardia. In rats, the bradycardic effect of THC on the heart involves CB1-like cannabinoid receptors (23). Although a previous study has shown that bradycardia may result from the effects of THC directly on catecholamine receptors (including adrenergic receptors) in the heart (24), another study in cats concluded that THC decreases central adrenergic neuronal activity, leading to decreased sympathetic tone, and subsequently causing bradycardia (25). Schmid, Schwartz (26) reported the presence of the endocannabinoids, anandamide and 2-arachidonoylglycerol in the rat heart, while both CB1 and CB2 receptors were detected in the myocardia of rats, mice, and guinea pigs (27-29). CB1 mRNA was also detected in the human heart (30, 31). All the aforementioned studies suggest that the bradycardic effects of THC may be mediated by several types of receptors, including cannabinoid receptors, in the heart.

Due to the lipophilic nature of THC, it is easily taken up by highly-perfused organs like the brain (32), which may explain the neurological signs of cannabis toxicosis. Ataxia, a common neurological symptom of cannabis toxicosis in both humans and animals, refers to the lack of coordination during movement. In the brain, the region responsible for coordination is the cerebellum (33), which contains a higher number of CB1 receptors in dogs as compared to humans (34). Patel and Hillard (35) showed that intraperitoneal THC administration in a mouse caused motor deficits including ataxia. They subsequently proposed that even though THC inhibits both excitatory and inhibitory synapses in the cerebellum, its main mechanism of action is the inhibition of the inhibitory synapses (basket cell/Purkinje cell), leading to the disinhibition of Purkinje cells. Firing of these GABAergic cells inhibits deep cerebellar nuclei cells, thus resulting in ataxia. A similar mechanism may occur in dogs, but further investigation is required.

Previous studies in chronic fatigue syndrome patients revealed that brain regions implicated include the basal ganglia, anterior cingulate, and frontal, temporal, and parietal regions (36, 37). These regions also overlap with brain regions implicated in motivation (38). Interestingly, these regions in humans and animals contain CB1 receptors (34, 39), which can be modulated by THC. This may explain how cannabis toxicosis can cause lethargy. Disorientation is the loss of a sense of direction and mental confusion. Frontal and temporal cortices, regions shown to contain CB1 receptors (34, 39, 40), are involved during mental orientation in space and time; the disorientation exhibited by pets during cannabis toxicosis may involve THC modulation of CB1 receptors in similar regions of the brain (41).

Products that caused cannabis toxicosis and the routes and reasons for exposure

In our study, edibles were the most common cannabis product that resulted in toxicosis, which is not surprising since they are the most common form of cannabis products purchased for dogs (3); however it is difficult to ascertain from our findings whether these edible products were purchased for human or animal consumption. Pets are often exposed to homemade or commercial edible goods, which are typically made using THC butter (5). In our study, and previous studies (9, 18), plant materials, including dried and fresh green cannabis, was another common product that led to cannabis toxicosis. The least common cannabis toxicosis-causing products were topical cannabis products, capsulated cannabis products, and tablets containing cannabinoids.

The most common source of cannabis toxicosis-causing products reported by veterinarians was government-regulated producers, followed by home-cultivated plants. A few pet owners reported that they obtained the products from the black market, however, this might be susceptible to under-reporting. In our study, ingestion was reported as the most common route of exposure. Compared to inhalation (42), which was the second most common route of exposure reported in our study, ingestion of edibles made with THC butter has been reported to result in more severe clinical signs and a higher risk for cannabis toxicosis in animals, since a majority of animals presenting with moderate to severe clinical signs of toxicosis had ingested some form of cannabis product including edible goods (5, 9, 18). Furthermore, the presence of other toxins (e.g., chocolate) in the edible product may have contributed to clinical illness, and may explain some of the deaths reported by veterinarians in our study and in another retrospective study (5). Even though less common, the ingestion of synthetic cannabinoids also leads to more severe clinical signs (43), and is known to be lethal in dogs (44).

The most commonly stated reason for pet exposure to cannabis was via oral ingestion while unattended, which was also reported in a previous study (42), followed by intentional administration for recreation (given to pets for fun?), or as a medical treatment. Our findings suggest that pet owners would have to put measures in place to prevent pets from accessing cannabis products including restricting cannabis to hard-to-access areas of the house, putting their cannabis products in pet-proof containers, and monitoring pets when cannabis-based products may be accessible. Some pet owners stated that cannabis toxicosis occurred following medical treatment which may be a result of unintended over-administration of these drugs due to the delay in manifestation of their effects. A small number of participants reported that some pets, specifically dogs, were exposed while being walked.

Diagnosis of cannabis toxicosis

In our study, the most common diagnostic method was the use of supportive clinical signs, along with a history of possible/known exposure, and/or the use of over-the-counter urine drug test kits. The key to appropriate treatment and successful recovery is accurate diagnosis, based on clinical signs, and accurate medical history from pet owners. Pet owners may be inclined to withhold information from veterinarians regarding accidental exposures to drugs (18) for fear of legal consequences. Therefore, veterinarians must encourage owners to provide complete histories when possible (17).

Many veterinarians in our study reported diagnosing cannabis toxicosis using urine drug test kits. The use of urine test kits in dogs may be unreliable based on interactions with other drugs, since patients on nonsteroidal anti-inflammatory drugs could have false-positive results (45). The incidence of false-negative results using the human urine drug test kit is also a concern. False negatives may occur if the urine sample is tested too soon after exposure (5), if the urine sample is not handled appropriately leading to the THC binding to the rubber stoppers and glass containers (5), if the patient consumed synthetic cannabinoids (46), or if the patient has diluted urine (47). In dogs, false negatives can also occur since THC is metabolized into 8-OH-Δ 9 -THC, which may not be detected by the human urine drug test kits (48) since they were not designed to detect this compound.

Treatments for cannabis toxicosis

In our study, the treatment method used most frequently was outpatient monitoring, followed by the administration of intravenous fluids, activated charcoal, and anti-emetics. Intravenous fluids can be administered as a form of supportive care (16, 43) to prevent both dehydration (i.e., from vomiting) and hypothermia (17) during cannabis toxicosis. Activated charcoal is often administered to prevent further absorption of the ingested material in the stomach and aid in decontamination (16). This method was recommended in previous studies for many of the dogs that experienced cannabis toxicosis and for all the iguanas (5, 9, 18). Induction of emesis is commonly performed in dogs, cats, and iguanas as an initial treatment if a toxic dose was ingested within 15-30 minutes or a significant amount of plant material remains in the stomach (5, 9, 16, 18, 42, 43, 49). It is safest to perform this procedure if the patient is still asymptomatic and with a normal mentation, to decrease the risk of aspiration (16). Emesis should never be induced if the animal is extremely agitated, severely depressed, or unresponsive (17). The administration of intravenous intra-lipids was a treatment method reported by several participants in our study. This method was reportedly used to treat a Boxer dog that ingested synthetic cannabinoids and was also used in a dog that died during treatment after ingesting THC butter, even though this method can have adverse effects such as leading to serum lipemia (5, 43). However, it may be useful for patients that are unresponsive to conventional treatments (16).

Treatment duration and recovery time following cannabis toxicosis depends on the severity of the toxicosis, which is dependent on the dose of THC (or other cannabinoids), quantity of cannabis or cannabis products consumed, and the route of exposure. In our study, veterinarians reported that most animals were treated as outpatients, while the remaining patients were hospitalized for less than 48 hours. This is not surprising since pets usually recover within 72 hours after cannabis toxicosis (7, 16). A wide range of recovery times have been reported in the literature, but they appear to vary between species (9, 18, 42, 43, 49).

Potential lethality of cannabis requires further investigation

Although most of the cannabis toxicosis cases in companion animals made a full recovery, 10 veterinarians cited death as an outcome for 16 cases. The details surrounding each case were not captured, thus we cannot be certain that exposure to cannabis directly resulted in mortality, or that the presence of other toxins found in edible products (e.g., chocolate, xylitol), or other underlying medical conditions contributed to the fatalities. In certain cases, it appears that cannabis was unlikely to be the primary cause of death, such as with aspiration pneumonia. In other cases, it may be possible that cannabis may have resulted in death directly, for example cases that report coma, uncontrolled seizures, or respiratory arrest as the primary clinical signs. These clinical signs are consistent with the mechanism of lethality in rats as reported by Thompson, Rosenkrantz (50), but the lethality of cannabis in dogs has not yet been confirmed. Previous research aiming to determine the lethal dose of cannabis in dogs was unable to determine a lethal dose (administering up to 9000 mg/kg orally), and this issue has been the subject of controversy in the veterinary field, with several sources misreferencing this original scientific study (51, 52). Previous field reports claiming that cannabis resulted in the deaths of animals arrived at this diagnosis through exclusion of other diagnoses (5), and thus do not represent strong scientific evidence; further basic research is needed to determine the potential lethality of non-synthetic cannabis in dogs and other pets, and its mechanism, if applicable. The suspected cases documented here, however, provide some guidance regarding this research gap; small and/or young animals may be more likely to be exposed to a higher apparent dosage, particularly for cannabis edibles, and due to their small body mass, could theoretically be more likely to succumb to an overdose and associated central nervous system depression, as was seen in rats in the lethality study (50).

Regardless of lethality, aggressive treatment of young and/or small animals is warranted in most cases, since the dosage may be unknown, and decontamination with emetics, IV fluids, and activated charcoal is considered a relatively safe treatment course. Naloxone infusions may also be considered in severe cannabis toxicoses cases, since there is some clinical evidence from human medicine that this opioid antagonist is effective in treating cannabis overdoses, because it also binds to endocannabinoid receptors (53).


The aggregate data collected by this veterinarian-based survey are prone to several biases. Since the survey was voluntary, a selection bias could have skewed the data; participants from states or provinces where cannabis is legal for recreational use may be more likely to see or report cases of cannabis toxicosis in animals compared with participants practicing in states or provinces where cannabis remains illegal for recreational use in humans. Furthermore, this survey data may be prone to recall bias, as veterinarians may not accurately remember the details of previous cases. Most importantly, the type of data collected here represents subjective aggregated data concerning cannabis toxicosis cases seen and reported by veterinarians; thus, raw numerical data concerning individual animals was not captured here. Consequently, the data presented herein should be interpreted with caution, and are, in some cases, inevitably vague and imprecise, particularly for the types and frequencies of clinical signs. Additionally, our data may also be prone to misclassification bias because of the lack of highly sensitive and specific diagnostic tests to confirm cannabis intoxication in animals. Thus, most of the diagnoses were made based on clinical signs along with a history of possible or known exposure. The latter requires veterinarians to rely on the history reported by pet owners, which may not always be completely honest due to the stigma which continues to surround cannabis, despite legalization.


Based on our veterinarian-reported survey data, the incidence of cannabis toxicoses in companion animals (primarily dogs) appears to have increased following legalization of cannabis for recreational purposes in Canada in October 2018. Although several factors may account for this apparent increase in cannabis toxicosis cases, the increased availability of cannabis products for humans is likely an important factor, since most of the toxicoses reported here resulted from inadvertent exposures; however, edibles were not legalized in Canada until October 2019, even though edibles were reported as the most common source of exposure in our study. The lack of veterinary oversight regarding the medicinal use of cannabis for animals in Canada also remains problematic and may also be contributing to a certain portion of these reported toxicoses, as many pet owners attempt to self-medicate their animals with these products (some of which are from the black market). Most of the cannabis toxicoses in animals appear to be benign; most cases resulted in mild to moderate clinical signs (most commonly, lethargy, disorientation, urinary incontinence, ataxia, and hyperesthesia), were treated as outpatients, and nearly all animals were reported to have fully recovered. Although several veterinarians in our survey reported deaths in association with cannabis exposure, rigorous controlled laboratory studies are needed to investigate this important and controversial issue, to eliminate or control for the presence of confounders such as other toxins (e.g., illicit drugs, chocolate, xylitol), other underlying disease processes, or causes of death secondary to cannabis ingestion (e.g., aspiration pneumonia). Finally, the use of clinical history and over-the-counter urine drug tests, although routinely used to diagnose cannabis toxicity cases in clinical practice, may be prone to false positive or false negative test results. There is a need for more sensitive and specific diagnostic tests to diagnose cannabis toxicities, whether to support aggressive decontamination procedures in high-risk patients, or to differentiate between non-synthetic cannabis (lethality unknown) and synthetic cannabis (known to be lethal in dogs; Hanasono, Sullivan (44)). As the burgeoning field of medicinal cannabis use in humans and animals continues to grow, fundamental research into the pharmacokinetics, pharmacodynamics, and potential lethality of cannabis in different animal species is also needed to address outstanding research gaps.


This research was funded by a Natural Sciences and Engineering Research Council Alliance Grant (ALLRP 549529 to JYK; and a MITACS Accelerate Fellowship (IT27597 to RQA and JYK; in partnership with Avicanna Inc.


Dr. Urban is an employee of Avicanna Inc., during which time she has received stock options. Avicanna Inc. did not influence the design, conduct or interpretation of the data derived from this study.