The Dog Aging Project is a community science project that aims to understand how genes, diet, lifestyle, and the environment influence aging and health outcomes in companion dogs (23, 24). Recruitment of owners in the DAP generally relies on various forms of online media or word of mouth and is ongoing with more participants every year (23). We used the DAP 2023 Curated Data Release version 1.0 (released on 9 February 2024) containing dog data from 47,444 owners that included the baseline Health and Life Experiences Survey (HLES) and the Annual Follow Up Surveys (AFUS) between 26 December 2019, and 31 December 2023 (25). Both surveys consist of multiple sections that collect information about dog demographic characteristics, physical activity, environment, dog behavior, diet, medications and preventatives, health status, and owner demographic characteristics. These sections are publicly available on GitHub (26). Data were collected and managed using REDCap (Research Electronic Data Capture) electronic data capture tools hosted through the DAP (27, 28).

Dogs were enrolled in the DAP after the baseline HLES was completed by the dog owners. The AFUS is given every year afterwards. The sections of HLES and AFUS used for cross-sectional and age-associated analyses included “dog demographics,” “physical activity,” “environment,” “dog behavior,” “diet,” “medications and preventatives,” “health status,” “owner contact,” and “owner demographics.” In the cross-sectional analysis, we used the most recently completed survey data of demographics and health conditions. For age-associated analysis, we used HLES and AFUS that were completed 9–15 months apart to assess the behaviors observed among dogs over time. In both analyses, we also filtered out surveys for dogs older than 18 years [which were considered outliners (29, 30)], as well as dogs with no age data. The survey variables examined in the cross-sectional and age-associated analyses are described in the following sections.

The frequency of CBD and hemp use in a dog was determined using owner responses in HLES and AFUS. “Frequent use” referred to owner-reported “non-prescription medication” or “daily supplement” use of CBD and hemp products. “Infrequent use” referred to CBD and hemp supplement use less than once a day. “No reported use” referred to no report of any CBD and hemp products. In the age-associated behavioral analysis, we only compared dogs with consistent use (i.e., frequent use over two consecutive surveys or more) to dogs with no use reported in any survey. Prior to July 2022, HLES had no check-box question regarding the use of CBD and hemp products, so only the free-response sections for “daily supplements,” “infrequent supplements” and “non-prescription medications” were used to capture CBD and hemp use based on the mentions of “CBD,” “hemp,” or “cannabidiol.” The AFUS offered a selection for CBD use as a “daily supplement” or “infrequent supplement,” but not under “non-prescription medications.” The CBD supplement selection and the three free-response descriptions that mentioned “CBD,” “hemp,” or “cannabidiol” were considered as CBD and hemp use for our analysis and “CBD” will refer to both CBD and hemp use hereafter. If a survey had both infrequent and frequent use marked, it was counted as “frequent use.”

All surveys after July 2022 include a comprehensive diet section, which has CBD as a specific supplement and asks about frequency with multiple choices for the frequency at which the supplements are given. Any CBD supplement use with one or more per day selected was considered frequent to ensure all survey results were consistent with the previous classifications. All other frequencies of indicated CBD supplement use (e.g., “weekly,” “monthly,” “less than monthly,” “other”) were counted as infrequent. Only one CBD use classification was used for each survey; if any two sections differed in the indicated frequency, the survey would be considered as indicating the higher level of use (i.e., “frequent use”).

Cannabidiol is not considered a controlled substance in the U. S. since 2018 (1, 2). In comparison, medical cannabis was only legalized in Washington, D. C. and 38 states as of December 31, 2023 (31). To test the hypothesis that the dog owners’ perception of medical cannabis use in humans affects their use of CBD on their dogs, we compared the prevalence of CBD use in dogs in U. S. jurisdictions with no legal access to cannabis (or with legalized CBD or low THC programs only) to those with legal access to medical cannabis. The four jurisdictions that enacted laws allowing medical cannabis during the DAP survey data collection in 2020–2023 were excluded from this analysis. The legal status of cannabis in each state was obtained from the National Conference of State Legislatures website (31).

We examined two categories of health conditions that owners could be giving their dogs CBD for as treatments. The first category was inflammatory conditions such as dermatitis, osteoarthritis, thrombocytopenia, hemolytic anemia, and inflammatory bowel disease. The second category consisted of other conditions for which CBD is frequently used as symptom management in dogs and humans: cancer (32, 33), epilepsy (7, 34), and dementia (35). We also included hip dysplasia since it could lead to secondary osteoarthritis and joint pain in dogs and humans (36, 37). In addition to the beneficial effects perceived by the owners, we included potential side effects of CBD such as clinical-sign–based gastrointestinal (GI) disorders – including chronic vomiting, chronic diarrhea, constipation, fecal incontinence, and bilious vomiting syndrome as outlined by Schmid et al. (38). We identified dogs affected by these conditions at any point in their lives for the analysis described below.

In the cross-sectional analysis, we first calculated the prevalence of CBD use in dogs with the above health conditions. Next, we determined if any of these conditions had a significant association with CBD use using logistic regression as follows:

logit (P(CBD use)) = β₀ + β₁*Dog Age + β₂*Spayed Female + β₃*Intact male + β₄*Neutered male + β₅*Dog weight + β₆*Owner’s annual income range + β₇*Owner’s maximum education level + β₈*Owner’s age range + β₉*Presence of specified health condition.

We included dog age, dog weight, and owner age, income, and education level as covariates. Dog sex and sterilization status were also included as a single covariate with four factors: intact female, spayed female, intact male, and neutered male. Exp (β₉) estimates the adjusted odds ratio summarizing the association between the specified health condition and CBD use. This model was run using the R Stats Package (Version 4.4.1) in RStudio (Version 2025.05.0 + 496).

Physical activity was calculated using a principal component analysis on three survey questions: (i) average daily time spent active reported in hours and minutes; (ii) average activity intensity level reported using a three-tier scale; and (iii) lifestyle activity level reported using a three-tier scale as described in Bray et al. (30) The principal component analysis results in a physical activity score that has a mean value of 0, with higher scores indicating more physical activity, and lower scores indicating less physical activity. Cognitive training of each dog was counted as a presence (1) vs. absence (0) variable with no distinction in levels of training. The history of cognitive training for each dog was determined based on the presence of primary and secondary activities that would necessitate cognitive training to some degree. Dogs who were determined to have a history of training included show dogs, working dogs, service dogs, and dogs participating in activities including agility training, field trials, hunting, obedience training, and search and rescue.

The presence of sensory and motion impairment was incorporated as a covariate in the age-associated analysis. Sensory impairments included any permanent conditions that greatly impacted sight or hearing, as well as one semi-permanent condition, cataracts, which was classified as a sensory impairment if there was no indication of a corrective surgery associated with it. Health conditions classified as motion impairments included neurological and orthopedic conditions that greatly impacted a dog’s mobility for their remaining lifetime. Conditions that met these standards for impairment are listed in Supplementary Table 1.

Dog behaviors were assessed using a modified version of the Canine Behavioral Assessment & Research Questionnaire (C-BARQ) (39, 40) which was completed by owners annually in the “dog behavior” section of HLES and AFUS. All 42 questions were categorized into one of the three main behaviors—aggression, fear, and agitation. The aggression score was based on questions about aggression and those about chasing wild animals. Fear was based on the questions about fear and behaviors when dogs were left alone such as restlessness, scratching, and whining. Agitation as a group consisted of all responses not classified as fear or aggression. Abnormal agitation covered a subset of the agitation behaviors considered to be abnormal. Additionally, 22 of the 42 questions were classified into three groups for secondary behavioral score calculation: (i) reactions to familiar circumstances; (ii) reactions to unfamiliar circumstances; and (iii) reactions to animals. All seven behavioral scores were calculated by averaging the score of questions in each category. The response to the questions was on a scale of 1–5, with 1 indicating the least intense behaviors and 5 the most. The full list of dog behavior questions, and their classification, can be found in Supplementary Table 2.

We used a linear mixed-effect model to assess dog behaviors over age, comparing dogs with consistent CBD use to dogs with no reported use. Age was used as the time variable in the model. We only included dogs with data from at least two consecutive surveys. Additionally, an approximate one-year gap must have separated all surveys to standardize response times. In addition to age, we included each dog’s sex and sterilization status, weight, physical activity score, history of cognitive training, sensory impairment, and motion impairment, as well as household characteristics (e.g., number of children, adults, other outdoor animals, and other indoor animals) as covariates in our model:

where β_j are the coefficients for the fixed effects, and γ_i0 and γ_i1 are the random intercept and slope for each individual dog.

The model was run using the R Package nlme (R Version 4.4.1; nlme Version 3.1–164) in RStudio (Version 2025.05.0 + 496). We deem normality not a particular concern given the large sample size in this study (41). We also did diagnostics with the residual plots for dog behaviors (data not shown), which raised no concerns for homoscedasticity. We used an alpha value of 0.05 without multiple test correction. The full model output is shown in Supplementary Table 3.

The University of Washington Institutional Review Board (IRB) deemed that recruitment of dog owners for the DAP and the administration and content of the HLES are human subjects research that qualifies for Category 2 exempt status (IRB ID no. 5988, effective 10/30/2018). No interactions between researchers and privately owned dogs occurred; therefore, the Institutional Animal Care and Use Committee oversight was not required.

In the most recent survey for all dogs, we found 2,592 dogs (or 5.5% out of 47,355 dogs) were reported to have infrequent or frequent CBD use. The mean (SD) age of dogs with frequent use was 10.2 years (4.3) compared to 8.1 (4.1) for dogs with infrequent use and 7.6 (4.3) for dogs with no reported use. A slightly higher proportion of male dogs (5.7%) were reported to use CBD than female dogs (5.2%). The dogs with CBD use were consistently older than the ones without across all weight classes, but the sex disparity was less consistent (Supplementary Table 4). The ratios of purebred and mixed breed dogs, the weight, the sterilization status of dogs, as well as the physical activity of dogs at their most recently recorded survey were comparable across different use levels (Table 1). Owner demographics were also comparable across different use levels (Supplementary Table 5).

For dogs in our dataset, the prevalence of CBD use in the 50 U. S. states and Washington, D. C. ranged from 1.6 to 9.7%, with Iowa having the lowest percentage and Hawaii having the highest (Figure 1). A total of 10,808 dogs lived in 12 jurisdictions where state residents had no legal access to cannabis or could only legally obtain CBD oil. 4.7 percent of the dogs in those jurisdictions were given CBD. In comparison, 34,404 dogs lived in 35 jurisdictions where state residents had legal access to medical cannabis before the beginning of the DAP surveys. 5.8 percent of the dogs in those jurisdictions were given CBD.

Combining the survey data of all dogs in every year, we found that 3,470 out of the 47,355 dogs, or 7.3%, were given CBD at any point. Of those, 2,759 reported frequent use. The prevalence of CBD use in each breed (if purebred) or primary breed (if mixed breed) were calculated based on any reported infrequent and frequent use throughout the surveys. The top 25 most common dog breeds in the DAP had a prevalence of CBD use ranging from 2.1 to 6.7% (Table 2). The highest CBD prevalence of use in breeds with >50 dogs surveyed was the purebred Leonberger at 10.3%, while purebred Scottish terrier, mixed black and tan coonhound, and mixed standard schnauzer had no reported CBD use (Supplementary Table 6).

The highest prevalence of CBD use was found to be associated with dogs with dementia at 18.2% (Table 3). The health condition with the highest number of dogs given CBD was osteoarthritis (12.5%), followed by cancer (10.0%). Of the 199 dogs using CBD that reported clinical-sign–based GI disorders, 54 reported chronic vomiting and 132 reported chronic diarrhea. After controlling for covariates, there were seven health conditions significantly associated with CBD use: dementia, osteoarthritis, epilepsy/seizures, hip dysplasia, cancer, chronic diarrhea, and clinical sign–based GI disorders (p < 0.05) (Table 3). There were only a small number of dogs (n < 60) affected with hemolytic anemia, thrombocytopenia, and clinical sign–based GI disorders other than chronic vomiting and diarrhea in our dataset, so the analysis for those conditions were excluded.

A total of 17,730 dogs had either consistent use (n = 331) or consistent non-use (n = 17,399) of CBD for at least 2 years and were included in age-associated analyses of seven behavioral and physical activity scores (Figure 2 and Supplementary Table 3). Significant interactions were found for CBD use with age in aggression (Figure 2A), reactions to unfamiliar circumstances (Figure 2F), and reactions to animals (Figure 2G). Dogs with CBD use were found to have differing behaviors in these categories at the outset, and the intensity of their behaviors decreased at a faster rate over time as compared to dogs with no use. No significant interactions were found for CBD use with age in the other four behaviors and physical activity.