The study was approved by the Institutional Animal Care and Use Committee (IACUC; approval protocol MD-21-16,661-2), and a signed informed consent was obtained from all dog owners or legal guardians. Six dogs suffering from naturally-occurring osteoarthritis at least in one joint were recruited to this study. Following an orthopedic examination, osteoarthritis was confirmed radiographically, and complete blood count and biochemistry panel were performed before initiation of the study. Exclusion criteria included dogs that were younger than 2-or older than 15-years, orthopedic surgeon recommendation for any joint surgery, undergoing a surgical procedure 3 months before intervention, or suspected liver disease. For ethical reasons, all dogs continued receiving analgesics and joint supplements that were prescribed prior to recruitment.

Liposomal-CBD formulation (CBD Liposome Platform Technology; LPT) was obtained from Innocan Pharma^™ (Israel). According to the product certificate of analysis, the Liposomal-CBD was prepared under strict aseptic conditions. Prior to use samples were submitted to Hy-Labs (Rehovot, Israel), a certified and accredited laboratory by the Israeli Ministry of Health and FDA, to confirm the formulation was sterile and below the approved limit of endotoxins. The results of these tests met the requirements of extra-vascular administered drugs in people.

The liposomal-CBD formulation was composed of synthetic CBD (Purisys LLC., Athens, GA, United States; not considered a controlled substance) that was loaded at a concentration of 50 mg/mL into hydrogenated soy phosphatidylcholine (HSPC) liposomes (Lipoid GmbH, Ludwigshafen, Germany).

The injection was performed between the shoulders, after hair clipping and aseptic skin preparation. Liposomal-CBD was injected subcutaneously at a dose of 5 mg/kg (0.1 mL/kg) using a 21-gauge, 1-inch needle at the prepared skin area.

Power analysis was not performed, as due to safety reasons the number of participants was limited to 6 dogs by the IACUC. Statistical analysis was performed using Stata/SE statistical software version 15.0 (StataCorp, College Station, TX, United States). Because sample size was small, descriptive statistics are expressed as median (range as minimum-maximum). Data analysis was performed with repeated measures mixed-effects with random intercept at the dog level. All values at time points following intervention were compared with baseline. Additionally, the association between CBD plasma concentrations and CBPI and iVAS scores were tested using mixed-effects linear regression. A p-value <0.05 was considered significant.

Three spayed female and three male (1 neutered, 2 intact) dogs with a median age of 12 (9–14) years old and body weight of 34 (26–58) kg were recruited to the study and completed the 6 weeks monitoring period. Dogs’ signalments, joints affected, osteoarthritic supplements and routine oral analgesics are presented in Table 1.

CBD plasma concentrations were observed throughout the 6-weeks monitoring period, including at the 6-week time point (Figure 1; Table 2). The plasma profile obtained showed a gradual increase in CBD up to the maximal CBD plasma concentration (C_max), and then a decrease starting in most dogs (4/6) at one week following injection. In dog number 1 the increase and the decrease were very gradual, and in dog number 6 the decline started earlier, after 2-days from injection (Figure 1). Calculated pharmacokinetic data and CBD plasma concentrations at 3- and 6-weeks following injection are presented in Table 2.

Dogs had significantly improved CBPI pain scores compared with baseline at weeks 2–3 (p = 0.011 and 0.031, respectively), improved CBPI function scores at weeks 2 and 6 (p = 0.004 and 0.026, respectively), improved CBPI total scores at weeks 2–3 (p = 0.001 and 0.028, respectively) and borderline improvement at week 6 (p = 0.052), and improved CBPI quality of life at weeks 2–3 (p = 0.046 for both weeks; Table 3). iVAS pain scores were significantly improved at 1–3 weeks (p < 0.001) and at 4 weeks following injection (p = 0.034; Table 3). The improvement in pain scores was significantly associated with the pharmacokinetic profile obtained; total CBPI at weeks 1–6 (p < 0.001 to p = 0.039, coefficients −0.249 to −4.399) and iVAS at weeks 1 (p = 0.008, coefficient −0.018), 2 and 3 (p < 0.001, coefficients −0.09 and −0.326, respectively).

Collar activity scores were significantly increased on weeks 5–6 (p = 0.012 and 0.027, respectively). Position scores, calories expedite, and sleep scores did not change significantly from baseline recordings (Table 3).

HR decreased significantly from baseline at 6 h (p = 0.010), 4 days (p = 0.010), 4 weeks (p = 0.022), and 6 weeks from injection (p = 0.017). f_R decreased significantly from baseline at 2 h (p = 0.018), 2 and 4 days (p = 0.005–0.008), and 5–6 weeks (p = 0.003–0.008). MAP decreased significantly from baseline at 4 days from injection (p = 0.048), and no difference was observed in RT throughout the study period (Table 4).

Most median hematology and biochemistry values were within reference ranges at all measurement times, although, some parameters were changed significantly from baseline. White blood cells (WBCs), neutrophils and monocytes increased significantly from baseline at 2 days from injection (p < 0.001). These increases were mainly attributed to a dog that developed phlebitis around the intravenous catheter. At 1 week after injection WBCs increased in 3/6 dogs (in the reference range) and decreased in 1/6 dogs with no overall significant change. Eosinophils increased significantly at 4 weeks (p = 0.009). At 1 week, a significant decrease was observed in hematocrit (p = 0.046), packed cell volume (PCV; p = 0.006), mean corpuscular volume (MCV; p = 0.017) and reticulocytes (p = 0.031). Platelets decreased significantly at 2 days (p = 0.046) and increased significantly at 1 week (p < 0.001), and plateletcrit increased significantly at 1 week (p = 0.007; Table 5).

Clinical pathology assessment of blood smears revealed mature non-toxic neutrophils at baseline in all dogs. A mild number of neutrophils became bands with mild toxic appearance in 3 different dogs: at 2 days (1 dog that developed phlebitis associated with intravenous catheter positioning), at 1 week (1 dog) and at 4 weeks (1 dog). Mild number of reactive monocytes was observed at baseline in 5/6 dogs, which were absent at the 4-week assessment in 4 dogs and sustained in one of these dogs. Mild–moderate number of atypical granular lymphocytes was observed at baseline and throughout the monitoring period in 5/6 dogs. Although none of the dogs was anemic, occasional polychromasia was observed in 5/6 dogs at baseline and at the following assessments. The dog that did not show polychromasia had mild poikilocytosis at baseline, then mild spherocytosis and mild poikilocytosis at 1-week, which were not observed on the 4-week assessment.

Alkaline phosphatase (ALP) did not change significantly during the study, however, one dog (dog number 5) showed high ALP value at baseline, which was further increased at the 4-week measurement. Another dog (dog number 4) had ALP elevation only at the 2-day measurement, during an elevated HR event. Gamma-glutamyltransferase (GGT) decreased significantly from baseline at 1 and 4 weeks from injection (p = 0.002 and 0.015, respectively). Total bilirubin increased significantly at 2 days (p < 0.001). Albumin decreased significantly at 1 week (p = 0.008). Total protein decreased significantly at all time points (p = 0.001, p < 0.01 and p = 0.004, respectively). Creatinine decreased significantly from baseline at 1 and 4 weeks from injection (p = 0.004 and 0.044, respectively). When dog 2, who had a kidney disease, was excluded from the creatinine analysis, creatinine was still decreased significantly at 1 week (p = 0.001). Calcium and potassium decreased significantly at 2 days (p < 0.001 and p = 0.010, respectively). CO₂ decreased significantly at 2 days (p < 0.001) and increased significantly at 1 week (p = 0.010; Table 5).

Local response (minor, non-painful swelling at the injection site) was observed in 5/6 dogs. The swelling was resolved (i.e., absorbed completely) within 3–6 days following appearance without any treatment (Table 6). One dog had an increased HR to 140–160 beats per minute starting approximately 36 h after injection, which resolved without treatment a day later. An echocardiogram revealed sinus tachycardia. Another dog developed a fever, which was attributed to phlebitis around an intravenous catheter that was left for 24 h for blood sampling. The catheter was removed, oral antibiotics was initiated, and the fever was resolved within 12 h.

At the time of manuscript submission, one of the dogs died naturally more than a year following injection at the age of 15 years, and two dogs were euthanized 5- and 7-months following injection due to deterioration in their disease condition (Table 6).

Results from the present study suggest that a single subcutaneous liposomal-CBD administration provides long-term (i.e., several weeks) CBD plasma concentration and analgesia. Liposomal delivery systems provide a slow release of various encapsulated drugs (26, 28). Additionally, many liposomal-based formulations improve the therapeutic window of drugs and therefore reduce their toxicity (33). The use of liposomes as a delivery system for CBD in the present study, indeed provided slow drug release during the tested period, as was shown by the time it took to reach C_max (T_max; 2–14 days) and by the long half-life (7.7–42.6 days) (Table 2). Compared with various oral CBD-containing formulations administering a single 2 mg/kg dose in dogs, the median/mean C_max, T_max, and half-life were 102.3 ng/mL, 1.5 h, and 4.2 h, respectively (n = 4) (17); 213 ng/mL, 2.1 h, and 2.5 h, respectively (n = 6) (11); 301 ng/mL, 1.4 h, and 1.0 h, respectively, (22); 226 ng/mL, 2.5 h, and 3.8 h, respectively (34).

In people, bioavailability of CBD is very low (6%–10%) and depends on fasting conditions (21). In dogs, bioavailability is considered better than in people, and reported to be 13%–70% depending on the formulation used (23, 35, 36). First-pass liver metabolism is believed to be the primary reason for the low bioavailability of oral CBD (21, 37). Therefore, alternative routes of delivery, such as via mucosal absorption that would bypass the liver are of interest. A recent study investigated the pharmacokinetics of a single 1 mg/kg pure CBD in oil formulation via oral transmucosal (OTM) administration or orally (6 dogs per route). Mean C_max and T_max for OTM and oral routes were 200.3 ng/mL and 1.9 h, and 206.8 ng/mL and 2.2 h, respectively. Half-life was 2.6 h with both routes (37). Interestingly, there was no difference in pharmacokinetic parameters between administration routes, suggesting that absorption via oral mucosa was not optimal or that most of the drug was actually swallowed (37). CBD administration was also investigated via nasal mucosa (mean dose of 1.7 mg/kg) or intrarectally using suppositories (mean dose of 8.3 mg/kg) compared with oral route (mean dose of 8.3 mg/kg). Following rectal administration CBD levels were below the limit of quantification. Mean C_max and T_max for nasal and oral routes were 28 ng/mL and 0.5 h, and 217 ng/mL and 3.5 h, respectively. Terminal elimination half-life was 7.0 and 15.7 h, respectively (38). According to these studies, CBD administered via mucosal sites was inferior compared with oral administration in dogs, although more studies using different CBD formulations are required for conclusion. This is strengthened by a study in dogs with naturally occurring osteoarthritis reporting a significant improvement following OTM CBD compared with control dogs (19).

Administration of Sativex^® (phytocannabinoid-based) sublingual spray was investigated in healthy young beagles, using an approximate dose of 0.5 mg/kg. Following a single dose, mean C_max and T_max of CBD were 10.5 ng/mL and 2 h, respectively (39). It should be noted that blood was sampled from the jugular vein in the Sativex^® study, which may have resulted in a biased overestimation of CBD plasma concentrations, because the jugular sampling site was reported to affect concentration of drugs administered via the oral mucosal route (40).

Transdermal administration was also investigated in two studies; (i) one study administered CBD-infused transdermal cream applied to the pinnae, which was compared with two oral formulations (CBD-infused oil or microencapsulated oil beads). These formulations were tested at 5 mg/kg twice daily in young healthy beagles (n = 10 per treatment). Following a single dose, mean C_max and half-life reached 625.3 ng/mL and 3.3 h (infused oil), 346.3 ng/mL and 1.6 h (oil beads), and 74.3 ng/mL (transdermal cream), respectively. The half-life of the CBD-infused transdermal cream could not be determined due to lack of elimination phase (23). (ii) The second study administered a transdermal low-THC Cannabis sativa extract 4 mg/kg rubbed into the pinnae twice daily for two weeks in six healthy young beagles. Mean C_max was 12.8 and 10.6 ng/mL after 7- and 14-days of administration. The authors concluded that CBD absorption via the transdermal route was generally poor (41).

In the present study C_max was lower compared with CBD plasma/serum concentrations at steady-state following 2–6 weeks oral CBD administration in dogs; 60–125 ng/mL (34), 80–160 ng/mL (23), 5–860 (median 311) ng/mL (20), and 53–201 ng/mL (12). This difference in C_max could be the effect of the relatively lower dose used with the prolong-release liposomal formulation, which was based on the reported dose tested intravenously (36). In retrospect a higher dose could have been tested. On the other hand, in many studies of oral CBD in dogs, C_max among individuals was extremely variable, with some dogs reaching only 10th of CBD plasma concentrations of other dogs in the same study using the same formulation (20, 24, 35, 37). Reduced variability among dogs in the present study suggests a more uniform drug absorption across dogs. Subcutaneous injected CBD has the benefit of direct absorption and bypassing the high extraction ratio of CBD by the liver compared with the oral route (21). Furthermore, when evaluating prolong-release formulations, the area under the curve (AUC) is the most important assessment tool, as it presents the total drug exposure over time (28). When normalized to dose, the AUC following liposomal-CBD administration in the present study (2,351 ng·h/mL/mg/kg; Table 2) was higher in comparison to long-term/steady-state oral CBD administration; 241–480 ng·h/ml/mg/kg after 28 days, once a day 1–12 mg/kg (12), 346–588 ng·h/ml/mg/kg after cannabis herbal extract containing 1:20 THC:CBD at 2–10 mg/kg (11), or 328–423 ng·h/ml/mg/kg after 2 mg/kg twice daily for 2 weeks of three different forms of hemp extract (34). Therefore, it suggests that the exposure to CBD using the liposomal formulation is superior to the oral route.

CBD is known to have anti-inflammatory and anti-nociceptive effects (42–44) and was described in the past few years as an efficacious analgesic in dogs suffering from osteoarthritis (17–20). The therapeutic efficacy reported in the present study is similar to previous studies with pain reduction and improved function in all dogs. The endocannabinoid system plays an important role in afferent and efferent nociceptive pathways (45). CBD is considered to exhibit its anti-inflammatory properties and analgesia via cannabinoid receptor 2 (CB2) as an inverse agonist and as an inhibitor of the reuptake of the endocannabinoid anandamide (15, 45, 46). Additionally, CBD was reported to interact with many other receptors and channels that are involved in nociception, such as activation of serotonin receptors (5-HT_1A), activation of transient receptor potential channels, vanilloid subfamily (TRPV1), inhibition of tumor necrosis factor-alpha (TNF-α), and inhibition of adenosine transporters (15, 45, 47). Furthermore, CB2 receptors expression is upregulated during inflammation in the affected tissue, as occurs in an osteoarthritic or rheumatoid joint. Therefore, treatment with cannabinoids activates CB2 receptors, and results in inhibition of cytokine production, decrease in leukocyte infiltration, reduction in bone destruction, and pain relief (45).

Unfortunately, the plasma CBD dose–response curve in dogs is still unknown. In the present study, a significant improvement in CBPI and iVAS pain scores was observed up to 3–4 weeks from injection, which corresponded to a median CBD concentration of 2.8 ng/mL. This may suggest that at this CBD plasma concentration there is still an analgesic effect, although, it is possible that the positive effect is also attributed to the overall high exposure observed.

Mobility in dogs can be affected by osteoarthritic pain, as previously reported (6, 32, 48). Therefore, the use of activity monitoring collars was chosen in order to provide an objective activity measurement. PetPace is a non-invasive monitoring collar that allows continuous monitoring of activity, position, certain vital signs, and sleep quality, and showed an excellent correlation with real-time variables (31, 49, 50). Recently, PetPace collar was suggested as a monitoring device to detect osteoarthritic pain, as it detected a significantly lower overall and high intensity activity levels in arthritic dogs when compared to healthy dogs (32). In the present study increased activity was observed 5–6 weeks following intervention, which was delayed from improvement in pain scoring evaluations, and CBD plasma levels. Factors other than pain can play a role in the pattern of dogs’ daily activity, such as owner activities, car rides, or environmental conditions (rain/extreme heat). Therefore, activity data from the collar, including data from the present study, should be interpreted with caution.

Although some of the blood work values changed significantly from baseline during the monitoring period, most changes were not clinically important, as values were kept within the reference range. WBCs increased in some of the dogs, but were not above the reference range, except the dog who had phlebitis. The increase in WBCs can be explained by a mild response of the immune system to injection of foreign materials (51), and it suits the local response observed at the injection site. The authors are not aware of published studies evaluating the effect of other liposomal formulations on WBC count administered subcutaneously in dogs. Epidurally administered liposomal-morphine in dogs did not show a systemic elevation of WBCs, but WBC count in the CSF was higher in the liposomal-morphine group (17 ± 18 cells/mm³) versus the liposomal vehicle group (2 ± 1 cells/mm³), with a value of <20 as the normal range (52). Hematocrit decreased a week post injection, but it was mild with no clinical importance. ALP was reported to significantly increase from baseline following long-term (weeks to months) administration of oral CBD in dogs, which was thought to result from induction of liver CYP isoenzymes (22, 24). However, a recent study reported that increased ALP correlated with significant elevation in bone-specific ALP, suggesting that the rise in total ALP can be partly attributed to osteoblastic activity (13). In the present study ALP increased in two dogs (33%);one of them had increased levels at baseline, and the other dog had an increase only at the 2-day measurement. Albumin level decreased during the present study, although in the reference range. A recent study investigating long-term CBD administration in dogs reported that albumin decreased gradually and reached significant difference at 6-months from initiation of the CBD administration. But the albumin values were still within the reference range (13). Albumin level may be decreased due to effects on the liver, but no other changes related to liver function were observed. Other effects of the liposomal-CBD, such as proteinuria or inflammation, may have resulted in decreased albumin and should be further investigated.

The minimal local swelling at the injection site was not diagnosed further, because it was minor, did not require a medical intervention, and was self-limiting. A different liposomal formulation (Exparel, DepoFoam Bupivacaine; made of phospholipids, cholesterol, and triglycerides) was reported previously to produce local response at the injection site in dogs. That study used experimental dogs and described the formation of granulomatous inflammation following multiple injections, characterized by an increased number of multinucleated giant cells and vacuolated macrophages. The authors of the Exparel study considered the local response as a normal response to the liposomes and non-adverse (51).

In recent years CBD has gained popularity in the veterinary market (13). However, products’ label can be misleading as many “CBD” products are actually extracts or enriched extracts from Cannabis sativa, and therefore they contain varying amounts of CBD in addition to many other chemically complex cannabis ingredients. A recent study reported that of 29 CBD products for dogs the total median CBD concentrations of their label claim was 93% (0%–154%) of claims (53). Valid CBD label-claims require rigorous analytical characterization and regulation (53). The FDA has published a guidance explaining that CBD products that are marketed without a prescription are not approved and may put users at risk (54, 55). Compared with cannabis-based products, synthetic CBD, which is FDA approved with a drug master file, provides a true THC and other cannabinoids-free product. The use of synthetic CBD as the active pharmaceutical ingredient of the liposomal-CBD formulation, can provide a reliable desired effect repeatedly.

Limitations to this study include the small sample size, and the non-blinded study design, which could have introduced bias to the owner and veterinary evaluations. We calculated the bioavailability based on a study reporting intravenous CBD administration from 1988 (36), which may not be an accurate calculation, but no other study reporting intravenous CBD in dogs is available in the literature. Most of the dogs in this study were geriatric, which potentially can affect the absorption and elimination of the CBD, and younger animals may have different pharmacokinetic profile following liposomal-CBD. Although, this may also be a strength of this study, as some of the dogs had concurrent disease states and/or were receiving routine medications, and this is usually the population of dogs that can benefit from CBD treatment.