Methods

This cross-sectional study utilized DAP²¹ data, including owner-reported survey responses and veterinarian-reported BCS. The participants included dogs and owners from across the US who were recruited, enrolled, and retained in the DAP's largest cohort (the DAP Pack) since the launch of the project in November 2019.

Dog Aging Project recruitment and data collection are ongoing. In brief, enrollment in the DAP Pack cohort requires human participants to review and digitally sign an informed owner consent form describing their participation in the DAP, to create a password-protected online portal, and to complete the Health and Life Experience Survey (HLES),²² which gathers comprehensive data, including dog demographic characteristics, environment, health, physical activity, behavior, diet, etc. Upon completion of HLES, participants are asked to upload a copy of their dog's veterinary electronic medical record (VEMR) obtained from their primary care veterinarian. The submission of VEMRs is strongly encouraged but is not a requirement for DAP enrollment nor for participation in the study reported here. Annually participants are invited to complete the Annual Follow-Up Survey (AFUS), which updates the information collected in HLES and includes additional survey instruments. The “Eating Behavior Survey,” further described below, is included in the AFUS. Within each survey instrument, all items require a response, and the instrument is not able to be submitted with any unanswered items. The survey instruments used by the DAP are freely available online.²²

The DORA survey¹⁴ was presented to DAP participants using the title “Eating Behavior Survey’’ to prevent participant response bias from use of the word “obesity.” The American English version of the survey, as provided by the questionnaire creators, was utilized, and the survey was presented in a digital format. Otherwise, all questions and domains remained unchanged, and the instrument is referred to as DORA in the manuscript presented here.

The DORA was designed by its authors to use 3 dog factors (food responsiveness and satiety, lack of pickiness, and interest in food) to create the dog FMS. The DORA was designed by its authors to derive the OMS from factors relating to the owner's perception of the condition of their dog's body (weight and fitness) and interventions used to control weight (eg, controlling meal portion size, restriction of treats, intentional exercise). Both calculated scores are expressed as percentages,^14,22 with higher scores indicating greater food motivation or more intensive owner management practices, respectively.

While DAP data collection is ongoing, periodic data releases are used to provide standard datasets that can be used by all researchers, and the 2021 Curated Data Release²³ was used here. Data from all participants who had completed initial HLES and 1 AFUS by December 31, 2021, were included. Data utilized from HLES included static dog demographic information, such as sex, owner-reported breed (including mixed breed), and American Kennel Club (AKC) breed group for purebred dogs. All other data used in this analysis, such as spay/castration status, owner-reported physical condition, social environment, and type of developed environment (rural, suburban, urban) as well as all DORA responses, were obtained from the AFUS. Owner-reported physical condition (overweight vs not overweight) was determined by participant response to the item within the diet section of the AFUS that reads, “Within the past year, has your dog been considered overweight?” with response variables “yes,” “no,” and “unknown.”

To assess owner agreement in reporting dog overweight with veterinarian assessment, BCS was obtained through review of VEMRs for a subset of dogs. A preliminary analysis was performed using the first 410 AFUS completions to estimate the frequency with which BCS within 12 months of DORA completion would be found in submitted VEMRs; this frequency was estimated to be 38.9%. Consultation with an epidemiologist determined that there was insufficient data to construct a rigorous power analysis for the number of BCS needed to assess veterinarian/owner agreement in the context of the DAP and that for purposes of subset analysis on this secondary aim of the study, an empiric target of 200 BCS was appropriate. To avoid bias in the creation of this subset of 200 BCS, study identification numbers for all dogs in the study population with submitted VEMRs were entered into a randomization table. Veterinary electronic medical records were manually reviewed in the order they appeared in the randomization table until a total of 200 BCS recorded within 12 months of the date of DORA completion were extracted. Body condition score was obtained from VEMRs by searching “BCS” and “body condition score” using the “Command + F” keyboard function. If the record was not digitally searchable, the record was read in its entirety to determine if a BCS was reported. If these methods did not return a BCS, the participant was excluded from the portion of the study utilizing BCS.

Body condition scores were reported on either a 5- or 9-point scale, and each score was initially assigned an interpretation as follows: BCS 5-point scale: 1 and 2 = underweight, 3 = ideal weight, 4 = overweight, and 5 = obese; BCS 9-point scale: 1, 2, and 3 = underweight, 4 to 5 = ideal weight, 6 to 7 = overweight, and 8 to 9 = obese. Any score containing a decimal of 0.5 and greater was rounded up to the next whole-number score. Numerical scores were required for inclusion; therefore, written descriptors of a dog's body condition, such as “ideal” or “obese,” were not included.

Results

A total of 15,720 participants completed at least some sections of the AFUS by December 31, 2021, and were included in the 2021 curated data release. A total of 14,125 respondents completed both the diet section (containing the question about dog overweight status) and the DORA section of the AFUS. Of these, 235 responded “unknown” to the question asking whether their dog had been considered overweight in the past year and were excluded from analyses related to owner-reported physical condition, leaving a total of 13,890 dogs for analysis of owner-reported physical condition.

Among the 15,720 participants who completed at least some sections of the AFUS, 10,540 participants (67%) provided a VEMR. A manual review of a total of 324 VEMRs was necessary to reach the target (n = 200) of BCS recorded within 12 months of completion of the diet and DORA sections of the AFUS. Two dogs with extracted BCS had responses of “unknown” to the question of whether the dog had been considered overweight in the past year; the owner-reported physical condition was not able to be compared to veterinarian-reported BCS for these 2 dogs.

The demographics of dogs for which BCS was extracted were similar to the overall study population, as were the medians (range) for the FMS and the OMS derived from the DORA instrument (Table 1). While there was agreement between owner-reported physical condition and veterinarian-reported BCS for a majority of the dogs with numeric BCS extracted from the VEMR (Table 2), disagreement was seen in 24% of cases, with owner-reported nonoverweight/veterinarian-reported overweight being the more common disagreement.

For 13,890 dogs with an owner-reported physical condition and DORA completion, there was a statistically significant (P < .001; OR, 1.06; 95% CI, 1.057 to 1.064) difference in OMS between owner-reported nonoverweight and overweight dogs when adjusting for dog age, sex, and weight. Specifically, owners of dogs in the overweight category had higher OMS (Figure 1). However, there was no significant difference in OMS between veterinarian-reported nonoverweight and overweight dogs (P = .78). Among dogs who were reported to be overweight by the veterinary BCS, owners who also reported that their dog was overweight had higher OMS than owners who did not report that their dog was overweight (P = .006; t = 2.84).

View Full Size

Owner-reported overweight physical condition is positively associated with owner management score (P < .001). Dogs with Dog Obesity Risk and Appetite (DORA) questionnaire completion and owner-reported body status (n = 13,890) are included; dogs whose owners responded “unknown” to the body status question (n = 235) were excluded. Boxes extend to the 25th and 75th percentiles, the central line shows the median, and whiskers extend to the maximum and minimum values, with outliers shown as single points.

Citation: American Journal of Veterinary Research 86, 5; 10.2460/ajvr.24.11.0358

• Download Figure
• Download figure as PowerPoint slide

There was a significant difference in dog FMS between owner-reported nonoverweight and overweight dogs, adjusting for age, sex, and weight (P < .001; OR, 1.017; 95% CI, 1.014 to 1.019) as shown in Figure 2. There was no significant difference in dog FMS between veterinarian-reported nonoverweight and overweight dogs (P = .89; OR, 1.0; 95% CI, 0.98 to 1.01).

View Full Size

Owner-reported overweight physical condition is positively associated with food motivation score (FMS; P = .016). Dogs with Dog Obesity Risk and Appetite (DORA) questionnaire completion and owner-reported body status (n = 13,890) are included; dogs whose owners responded “unknown” to the body status question (n = 235) were excluded. Boxes extend to the 25th and 75th percentiles, the central line shows the median, and whiskers extend to the maximum and minimum values, with outliers shown as single points. Odds of overweight status increase by 1.017 (95% CI, 1.014 to 1.019) per FMS percentage point when controlling for age, sex, and weight.

Citation: American Journal of Veterinary Research 86, 5; 10.2460/ajvr.24.11.0358

• Download Figure
• Download figure as PowerPoint slide

There was a small association of weight with FMS (P < .001) when controlling for age and sex in that larger dogs had FMS that were higher by 0.48/10 kg. There was a small but statistically significant association between the type of developed environment and FMS in that dogs living in an urban environment had higher FMS than those in a rural environment (P < .005). There was a statistically significant (P < .001) difference in mean FMS across AKC breed groups (Figure 3) when controlling for age, sex, weight, and type of developed environment. Specifically, when controlling for these 4 variables, FMS for dogs from the sporting group was higher by 10.2% (95% CI, 9.1% to 11.2%) than FMS for mixed-breed dogs, FMS for dogs from the hound group was higher by 5.3% (95% CI, 3.6% to 7.1%) than FMS for mixed-breed dogs, and FMS for dogs from the nonsporting group was lower by 4.1% (95% CI, 5.7% to 2.5%) than FMS for mixed-breed dogs. Finally, the number of dogs in the household was associated with FMS. After controlling for age, sex, and weight, dogs in multidog households had 3.56% (95% CI, 2.60% to 4.53%) higher FMS than dogs living in a single-dog household.

View Full Size

Dog FMS vary among American Kennel Club (AKC) breed groups. The FMS for 13,890 dogs with completed DORA are shown. Boxes extend to the 25th and 75th percentiles, the central line shows the median, and whiskers extend to the maximum and minimum values, with out­liers shown as single points. When controlling for age, sex, weight, and type of developed environment, FMS was higher (sporting, hound) or lower (nonsporting) for dogs from certain AKC breed groups compared to mixed-breed dogs as indicated by asterisks; *P < .001. FSS = Foundation stock service.

Citation: American Journal of Veterinary Research 86, 5; 10.2460/ajvr.24.11.0358

• Download Figure
• Download figure as PowerPoint slide

Discussion

This study demonstrated that dog FMS and OMS were each positively associated with overweight physical condition and that food motivation may be impacted by the type of developed environment, household environment, and breed group.

In this study, dogs with a higher FMS were more likely to be overweight, which is consistent with the data presented by the original DORA study.¹⁴ Higher FMS were also found in larger dogs and dogs from urban environments. American Kennel Club breed group was significantly associated with FMS, with the sporting breed group having the highest mean FMS across all breed groups. In our sample, the majority of dogs from the sporting group were Labrador Retrievers and Golden Retrievers, which are among the 3 most common breeds in the US.²⁵ While breed groups in the UKC and AKC are not identical, the original UK DORA study found that the gun dog group had the highest mean FMS,¹⁴ and the breeds represented in the UKC gun dog group overlap substantially with the AKC sporting dog group. Some evidence²⁶ suggests that these breeds are likely to eat indiscriminately (including food and nonfood objects) and to be strongly motivated by food. A specific pro-opiomelanocortin gene deletion that has been associated with greater food motivation and obesity in Labrador and Flat-Coated Retrievers²⁷ may be a contributing factor to indiscriminate eating behaviors. Alternatively, food motivation in sporting group dogs may also be influenced by exercise habits or use of food as a reward in training activities for their sports, which were not directly queried in this study.

Owners of overweight dogs also had heightened feeding management practices, including the use of controlled meal portion size, and/or restriction of treats, and/or intentional exercise for weight control. The DORA presents more than 1 item in each domain^14,22; thus, respondents with high OMS may engage in various components of diet and weight management. However, studies^28–30 have reported that interventions to promote dog weight loss are neither consistently followed nor consistently successful, and that owners may not understand the health impacts or specific action steps to take with overweight dogs, even if veterinarians believe they regularly convey that information.^31,32 Given our finding that dogs from certain AKC breed groups have higher food motivation, it may be that dogs from such breeds require even stricter owner management practices than dogs from other backgrounds to maintain an ideal body weight. Further research to identify the most successful weight management practices for dogs and the most effective communication strategies for veterinarians on the topic of weight management is needed.

Body condition score within a year of DORA completion was only identified in 61% of screened records. While this is an improvement over previous literature that reported 34% of medical records containing BCS,³³ a substantial proportion of records still do not contain a recent BCS. In our study, as in previous reports, owners underreported overweight physical condition in their dogs as compared to veterinarians. Owners may underreport overweight physical condition because they are not properly educated on the phenotypic appearance of excess fat deposition in dogs, especially those of various body conformations. While owner accuracy improves somewhat with training on and/or access to a BCS chart, many owners continue to underreport overweight in their dogs.^34–36 Alternatively, it may be that dog owners are reluctant to accept that their dog is overweight, similar to parents of overweight children, over half of whom underestimate their child's weight status.³⁷ Notably, a previous study³⁸ has reported that owners of dogs who were active and successful in various sports were better able to assess BCS than typical pet dog owners, which suggests that owner motivation may be a factor in learning to assess BCS.

While this study advances our understanding of the influence of dog food motivation and owner management on canine overweight, it has limitations. First, owner-reported physical condition was utilized as the determination of the dog's physical condition. Given that owners underreported overweight physical condition compared to veterinary assessment in the subset analysis reported here and in prior studies,^34–36 it is likely that overweight was underreported in this study, which may have affected our analyses. Additionally, the survey question in which owners were asked about dog physical condition was binary (overweight yes or no); it is possible that owners responded affirmatively to a binary question less frequently than they would have identified overweight on a graded scale, like a BCS. Also, the survey question did not clarify whether the assessment of the dog's weight status was solely the owner's opinion or whether that opinion reflected outside sources, such as their veterinarian. Another limitation is that the BCS extracted from medical records could have been recorded up to 1 year prior to AFUS submission. Some dogs’ BCS could have changed within this timeframe, which could account for some of the discrepancies noted between owner-reported physical condition and veterinarian-reported BCS. Body condition score was only evaluated for a small subset of the study population, creating the possibility of type II error in the lack of association with FMS. Numbers of dogs within breed groups were also unequal, which may have limited the ability to detect patterns in less well-represented breed groups. Finally, due to the small number of intact dogs in this study, the effect of neuter status was not analyzed.

Cross-sectional studies such as this one can identify associations between OMS, FMS, and BCS. Longitudinal analyses are required to clarify the temporal relationship among those factors. Longitudinal data being collected by the DAP from the dog-owner pairs reported here, and thousands more, will enable investigation into temporal relationships, such as whether heightened OMS is followed by a decline in BCS or whether an individual dog's FMS tends to change in a predictable way over the life course.

In conclusion, breed background, type of developed environment, and household composition may influence dog food motivation, and dogs with higher food motivation are more likely to be overweight. Additionally, owners of overweight dogs report higher levels of management of their dogs’ eating habits. A substantial proportion of the medical records reviewed did not include a recent BCS; increased adherence by veterinarians to reporting this feature of dog physical examination is needed. Longitudinal studies of changes in dog food motivation over time and the effectiveness of heightened owner management practices on preventing or improving overweight in their dogs are warranted.

Disclosures

The authors have nothing to disclose. No AI-assisted technologies were used in the composition of this manuscript.

Funding

This research is based on publicly available data collected by the Dog Aging Project, under U19 grant No. AG057377 (Principal Investigator: Daniel Promislow) from the National Institute on Aging, a part of the NIH, and by additional grants and private donations, including generous support from the Glenn Foundation for Medical Research, the Tiny Foundation Fund at Myriad Canada, and the WoodNext Foundation. Dr. Creevy, Dr. McNulty, and Dr. Ruple also received funding from U19 grant No. AG057377. Dr. Gartner was funded by NIH grant No. T35OD010991-16 and by the Texas A&M College of Veterinary Medicine and Biomedical Sciences. Dr. Hoffman is supported by NIH grant No. AG059920. These data are housed on the Terra platform at the Broad Institute of The Massachusetts Institute of Technology and Harvard.

ORCID

A. Ruple https://orcid.org/0000-0002-5223-0217

K. E. Creevy https://orcid.org/0000-0003-4169-374X