All data analyses were conducted under Anicom Insurance, Inc.’s data use policy and in accordance with the relevant privacy protection regulations. The use of the data was covered by the policyholders’ consent obtained at the time of insurance enrollment, which allowed the secondary use of anonymized claim information for research and statistical purposes. Individual policy numbers were used only to identify eligible contracts during data extraction and were removed before analysis; thus, no personal identifiable information was included in the research dataset. Ethical review by an institutional committee was not required for the secondary analysis of anonymized data.

The data used in this study were obtained from Anicom Insurance Inc.’s pet insurance claims database. This database contains information on insured dogs and cats, including species, breed, sex, age, policy inception date, and diagnostic terms for clinical conditions, at two hierarchical levels: major and minor. The major diagnostic category represents broad disease groups (e.g., diseases of the teeth and oral cavity, ocular diseases, and cardiac diseases), whereas the minor diagnostic term indicates specific conditions (e.g., periodontal disease, gingivitis, cataract, and hypertrophic cardiomyopathy). These diagnostic terms were entered by veterinarians based on clinical records, either by selection from a standardized list or by entering free-text descriptions. Because diagnoses were made in routine clinical practice across multiple veterinary hospitals, the specific diagnostic criteria for each condition may have varied among veterinarians and institutions. Claims related to preexisting conditions diagnosed before policy inception were excluded because such conditions were not covered under the insurance policy. The following policies were also excluded: (1) those without outpatient coverage, (2) those with a policy term other than 1 year, and (3) those who canceled immediately after enrollment due to a cooling-off request or similar reasons. Each year was treated as a single observation per animal. Analyses were conducted using claims data from policies commenced in 2013, 2018, and 2023. Animals with missing information regarding sex, age, or breed were excluded from the analysis. Sex was recorded as male or female only, and information on neuter status was not available. Crossbred animals, including crosses between different pure breeds as well as Japanese mixed-breed dogs and cats, were registered under the unified categories “mixed-breed dog” or “mixed-breed cat” and were excluded from our analyses. The numbers of breeds and animals included in the initial datasets were as follows: for dogs, 183 breeds (n = 410,506) in 2013, 178 (n = 551,871) in 2018, and 183 (n = 690,687) in 2023 (Supplementary Table S1); for cats, 51 breeds (n = 39,432) in 2013, 62 (n = 90,434) in 2018, and 68 (n = 186,605) in 2023 (Supplementary Table S2). Data from 2013 and 2018 were used to examine temporal trends. Detailed analyses were performed using the 2023 dataset, which contained the largest number of records.

For each animal, a binary indicator of periodontal disease-related morbidity was assigned based on the presence or absence of insurance claims related to periodontal conditions. Animals were classified as positive for periodontal disease if the minor diagnostic term in the claims record contained any of the following terms: “periodontal disease,” “gingivitis,” “alveolar pyorrhea,” “dental root abscess,” or “periapical abscess.” The search was conducted by partial text matching of these Japanese terms, encompassing both standardized diagnostic entries and free-text descriptions. Animals without any of these terms were considered negative. The number of animals identified for each diagnostic term is summarized in Supplementary Table S3.

For dogs, two breed-level categorical variables were included: body size and breed. Body size groups were classified into five categories based primarily on reference adult body weight for each breed: Toy (<4 kg), Small (4–<10 kg), Medium (10–<30 kg), Large (30–<45 kg), and Giant (≥45 kg). The reference body weight for each breed provided by the Japan Kennel Club (JKC). The breed groups were defined according to the 10-group classification of the Fédération Cynologique Internationale (FCI) as adopted by the JKC: G1, Sheepdogs and Cattledogs; G2, Pinscher and Schnauzer – Molossoid and Swiss Mountain and Cattledogs; G3, Terriers; G4, Dachshunds; G5, Spitz and primitive types; G6, Scent hounds and related breeds; G7, Pointing dogs; G8, Retrievers – Flushing dogs – Water Dogs; G9, Companion and Toy dogs; and G10, Sighthounds. Breeds for which group information could not be assigned were excluded from the group-based analyses. For cats, neither body size nor breed classification was applied, and the analyses were conducted only at the breed level.

The analyses were performed in four sequential phases based on generalized linear mixed models (GLMMs) with a binomial error distribution and a logit link. First, using policy records initiated in 2013, 2018, and 2023, temporal changes in age-specific periodontal disease claim rates were evaluated separately for dogs and cats. Second, data from 2023 were used to compare the overall morbidity risk and the magnitude of age-related effects (per-year change in the odds of a periodontal disease claim) between the two species. Third, within the 2023 dog dataset, baseline risk (probability of a periodontal disease claim at age zero) and age-related effects were compared among body size categories and breed groups (JKC/FCI classification) to examine whether body size or phylogenetic grouping contributed to the observed variation. Finally, breed-level models were fitted using the data from 2023 to quantify the baseline risk and age-related increase for each breed, evaluate the correlation between these two parameters, and identify breeds with distinctive characteristics in age-related periodontal disease risk.

The analyses were conducted in R (v4.3.3; R core Team, 2024) using GLMMs with a binomial error distribution and a logit link [lme4 (26)::glmer, Laplace approximation; optimizer = “bobyqa”]. To ensure stable estimates, breeds with fewer than 100 individuals were excluded from the analyses using the 2023 dataset (dogs: 81 breeds, n = 688,665; cats: 38 breeds, n = 185,782). Unless otherwise noted, Wald-type confidence intervals and tests were used for fixed effects, and results are presented as odds ratios (ORs) with 95% confidence intervals (CIs). Sex was coded as a binary factor (male, reference). Statistical significance was defined as p < 0.05 (two-sided) unless otherwise stated. Plots were generated using R packages ggplot2 (27), gridExtra (28), cowplot (29), and scales (30).

Using insurance policy data from 2013, 2018, and 2023, temporal changes in the occurrence of periodontal disease claims were evaluated separately for dogs and cats. The overall claim rates related to periodontal disease were as follows: dogs—2013: 2.95% (12,093/410,506), 2018: 4.13% (22,774/551,871), and 2023: 4.33% (29,884/690,687); cats—2013: 0.76% (300/39,432), 2018: 0.75% (677/90,434), and 2023: 1.05% (1,956/186,605). When limited to individuals that had at least one insurance claim (i.e., animals that received veterinary care during the policy year), the corresponding rates were: dogs—2013: 4.62% (12,093/261,644), 2018: 6.40% (22,774/355,771), and 2023: 6.55% (29,884/456,464); cats—2013: 1.59% (300/18,865), 2018: 1.66% (677/40,782), and 2023: 2.28% (1,956/85,795). Model comparisons indicated that GLMMs with a random intercept for breed provided a better fit than did GLMs without random effects in both species (dogs: AIC 505,620 vs. 517,048; cats: AIC 32,719 vs. 32,767). Therefore, subsequent analyses were based on the GLMMs to account for breed-level heterogeneity. In dogs, the fixed effects (age at policy inception, sex, and policy start year) were statistically significant (all p < 0.001). The estimated ORs with 95% CI were as follows: age per year, 1.156 (1.153–1.158); female vs. male, 1.083 (1.066–1.100); and policy year 2018 and 2023 vs. 2013, both approximately 1.13 (2018: 1.129, 2023: 1.131). Age-specific claim rates and the number of claim events at each age are illustrated in Figure 1A (dogs) and Figure 1B (cats). These results indicated that the risk of periodontal disease claims in dogs increased with age, was slightly higher in females, and has risen modestly in recent years compared to 2013. In cats, age and sex were also significant predictors (p < 0.001), whereas policy-year effects were significant only for 2023. The OR for age per year was 1.105 (1.095–1.115), and females had a lower risk (OR 0.852, 0.791–0.917). The 2018 policy year showed no significant difference compared with 2013, whereas the 2023 cohort exhibited a 1.238-fold higher risk (1.094–1.402). Thus, as in dogs, the risk of periodontal disease claims in cats increases with age and appears to be higher in the most recent period, although the effect of sex is in the opposite direction. The variance of the random intercept for breed was 0.49 [standard deviation (SD) = 0.70] in dogs and 0.05 (SD = 0.22) in cats, indicating that breed-level heterogeneity was more pronounced in dogs.

Using insurance policy data from 2023, we compared the risks of periodontal disease-related claims and age-related changes between dogs and cats. A GLMM was applied with age, species, and sex as fixed effects and breed as a random effect. All fixed effects were statistically significant (p < 0.001). In dogs, the age coefficient was 0.126 (OR = 1.134, 95% CI: 1.132–1.137), indicating that the risk of periodontal disease–related claims increased by approximately 13.4% per year of age (Figure 2). For cats, the age-by-species interaction term (age × cat) had a coefficient of −0.036, corresponding to an OR of 0.965 (95% CI: 0.954–0.976). This result indicated that the age-related increase in risk was approximately 3.5% per year smaller in cats than in dogs, and the difference in age effects between species was statistically significant (p < 0.001). The age effect for cats, calculated as the sum of the coefficients for age and the interaction term (0.126 + −0.036 = 0.090), corresponded to an OR of 1.094 (95% CI: 1.082–1.100) (Figure 2). Thus, although the risk of periodontal disease-related claims increased with age in both species, the rate of increase was more gradual in cats. Additionally, at baseline (age 0), cats had a significantly lower risk (OR = 0.538, 95% CI: 0.428–0.675). These results suggest that, in addition to the effect of age, cats have a lower initial risk of periodontal disease than do dogs.

Using the 2023 dataset, the breed-level risks and age-related effects of periodontal disease claims were evaluated. For both dogs and cats, a GLMM with a logistic link was applied, including age and sex as fixed effects, and breed as a random effect. Random intercepts and random slopes for age were specified to simultaneously estimate the breed-specific baseline risk and age-related effects.

In dogs, the variance of the random effects was 0.457 (SD = 0.676) for the intercept (baseline risk) and 0.00051 (SD = 0.0225) for the age slope, with a correlation coefficient of −0.34 between them. Thus, the baseline risk varied considerably among breeds, whereas the variation in the age-related effect was small, and the correlation between the two components was not statistically significant. A likelihood ratio test comparing the full model with one model, assuming zero correlation, showed no significant differences (p = 0.17). A two-dimensional plot of breed-specific baseline risk versus age-related effects is shown in Figure 5. Small-sized breeds generally exhibit higher baseline risks and lower age-related effects, although some breeds exhibit uniquely strong age-related effects or markedly lower baseline risks. Overall, the breeds could be broadly divided into those with a high baseline risk from an early age (e.g., companion and terrier-type breeds) and those showing a marked increase in risk with aging (e.g., working and hound-type breeds). Estimated breed-specific baseline risks and age-related effects, with 95% CIs for all the dog breeds, are shown in Supplementary Table S4.

In cats, the variance in the random effects was 0 (SD = 0.000) for the intercept and 0.00063 (SD = 0.0250) for the age slope, indicating minimal variation in the baseline risk among breeds and only a slight variation in age-related effects. Accordingly, a correlation between baseline risk and age-related effects could not be estimated. Based on the magnitude of age-related effects, five breeds with the highest and lowest age-related effects were identified (Supplementary Figure S1). The breed with the strongest age-related effect was the Siamese (OR per year = 1.144), whereas the breed with the weakest effect was the Russian Blue (OR per year = 1.056). The estimated age-related effects (OR per year) with 95% CIs for all the cat breeds are summarized in Table S5. Because all cat breeds share the same baseline risk, Table S5 does not include the baseline risk estimates.

This nationwide large-scale analysis based on Japanese insurance claims data identified a clear age-related increase in the risk of periodontal disease in both dogs and cats, capturing consistent age-related patterns across species and analytical levels.

In dogs, both the magnitudes of the age effect and baseline risk varied substantially among body size categories, breed groups, and individual breeds, whereas in cats, the variation in baseline risk among breeds was minimal.

Among insured animals that received veterinary care, the claim rates related to periodontal disease (dogs: 4.62–6.55%; cats: 1.59–2.28%) were lower than the prevalence estimates reported in primary-care studies [dogs—United States: 19.5% (6), United Kingdom: 12.5% (20); cats—United Kingdom: 15.2% (9)]. Although a lower prevalence of periodontal disease in the Japanese pet population cannot be completely excluded, the relatively low claim rates observed here are more likely to be attributable to the structural characteristics of the insurance data. Insurance claim records are based on veterinary clinical diagnoses but are generated only when the diagnosis or treatment is submitted for reimbursement. Consequently, cases in which only uncovered procedures (e.g., prophylactic scaling) were performed or mild gingivitis without treatment may not have been recorded. In a population of Japanese dogs, even among elderly individuals, gingivitis accounted for 66.7% of cases, whereas periodontitis accounted for only 24.0% (32). The fact that most animals diagnosed with periodontal disease present with gingivitis further supports the interpretation that insurance-based data likely underestimate the true population prevalence of periodontal disease. The veterinarians’ diagnostic awareness and sensitivity to periodontal disease may have also influenced these results. In both dogs and cats, the claims rate for periodontal disease in 2023 was significantly higher than that in 2013. As no major changes occurred in the insurance system during this period, this increase is unlikely to reflect a true nationwide surge in disease occurrence over only a few years but rather a heightened recognition and diagnostic sensitivity among veterinarians following increased scientific attention to periodontal disease. Furthermore, the higher prevalence of periodontal disease often reported in prospective clinical studies than in primary-care datasets is thought to result from lower diagnostic thresholds and the use of more rigorous oral examinations (20). As awareness of the importance of dental care continues to increase among veterinarians, insurance claim rates for periodontal diseases may increase in the future.

In the 2023 comparison between dogs and cats, both species showed age-related increases in risk; however, the baseline risk and magnitude of the effect of age were higher in dogs. The baseline risk in cats was approximately half that in dogs (0.54-fold lower), and the slope of the age effect differed by 3.5% (dogs: OR/year = 1.134; cats: OR/year = 1.094), indicating that the disparity in periodontal disease claims rates between the two species widened with age. In contrast, primary-care studies in the United Kingdom reported an annual prevalence of 12.5% in dogs and 15.2% in cats, suggesting a relatively higher prevalence in cats (9, 20). This discrepancy may reflect the lower veterinary visitation rates among cat owners in Japan. According to a 2024 national survey, the number of cats owned (approximately 9.16 million) will exceed that of dogs (approximately 6.80 million) (33). However, in the present dataset, the number of insured cats was less than one-third of insured dogs, indirectly suggesting a lower rate of insurance enrollment, implying that healthcare utilization is lower in cats than in dogs. Furthermore, cats often exhibit behavioral tendencies to conceal signs of illness or pain, which may reduce the likelihood of visiting veterinary clinics. These social and behavioral factors may have contributed to the underestimation of the baseline risk in cats. However, these factors alone do not sufficiently explain the milder effect of age on periodontal disease in cats than in dogs. Biological and dental structural differences are likely to play a major role. Compared with dogs, cats possess shorter shear-type dentition and chew for a shorter duration, resulting in less crowding and fewer occlusal fractures. Periodontal disease arises primarily from the accumulation of dental plaque (4, 24), and in humans, alterations in oral pH have been associated with disease progression (34, 35). Cats have adapted to eating small meals frequently throughout the day, reflecting their natural hunting and feeding behavior (36). Although direct evidence is lacking for cats, these feeding patterns may influence the dynamics of oral pH fluctuations, thereby affecting plaque accumulation. These biological characteristics may partially account for the milder effects of age observed in cats.

For dogs, detailed analyses were conducted across three hierarchical levels: body size, breed group, and individual breed. At the body size level, smaller breeds showed distinctly higher baseline risks than did larger breeds, whereas the slope of the age-related increase did not differ substantially among size categories. This finding agrees with previous reports that identified small breeds as particularly prone to periodontal disease (20, 22) and suggests that although small dogs are predisposed to developing the disease at younger ages, their subsequent rate of progression with aging is comparable to that of larger dogs. This predisposition in small and toy breeds may be explained by anatomical characteristics, including a reduced jaw size relative to tooth number, which results in increased dental crowding and greater plaque retention in the oral cavity. At the breed-group level, both the baseline risk and age-related effects differed significantly, revealing distinct patterns among the groups. These results indicate that body size mainly influences baseline risk, whereas breed group affects both baseline risk and the magnitude of age-related increase. At the individual breed level, a weak negative correlation (r = −0.34) was observed between baseline risk and age-related effects, although this was not statistically significant, suggesting that the two components may vary independently. In other words, breeds that are at a high risk of developing periodontal disease early in life do not necessarily exhibit faster progression with aging. This tendency could partly reflect clinical and epidemiological factors such as earlier intervention in high-risk breeds or survival bias among severely affected individuals. Overall, these findings suggest that the initiation and progression of periodontal disease in dogs is governed by distinct determinants. Anatomical factors, such as body size and oral conformation, may strongly affect baseline risk, whereas breed-group characteristics, immune function, and lifestyle factors may contribute to age-related increases. Accordingly, effective periodontal disease management in dogs should consider not only age and body size but also the specific breed-group characteristics and inherent predispositions of each breed. Moreover, the risk patterns identified in this study may have practical relevance depending on the level of awareness among pet owners regarding the importance of dental and oral diseases. For example, in breed groups with high baseline risk, emphasizing the importance of early intervention from a young age may be beneficial, whereas in groups with strong age-related effects, communicating the need for continuous lifelong dental care is likely to be important. Such tailored communication based on body size and breed-group-specific risk profiles may help insurers and veterinary clinics promote beneficial behavioral changes among owners and contribute to the improved prevention and management of periodontal disease.

The analysis of breed groups revealed four distinct epidemiological patterns in the age-related trajectory of periodontal disease risk based on the combination of baseline risk and magnitude of the age effect: (1) High baseline risk with mild-to-moderate age-related increase [companion dogs (9G) and terriers (3G)]; (2) Low-to-moderate baseline risk with relatively steep age-related increase [working dogs (2G), scenthounds (6G), and pointers and setters (7G)]; (3) Moderately high baseline risk with steep age-related increase [dachshunds (4G) and sighthounds (10G)]; and (4) Both baseline risk and age-related increase near the overall mean [herding and cattledogs (1G), spitz-type breeds (5G), and retriever-type breeds (8G)]. These patterns suggest that the onset and progression of periodontal disease are influenced by distinct combinations of morphological, behavioral, and genetic factors. For instance, companion and terrier groups, which mainly consist of small breeds, may have early onset periodontal disease due to dental crowding and smaller oral cavity volume but show relatively moderate progression with age. In contrast, large-breed groups, such as working dogs and hounds, may have a low risk at young ages but experience a steeper increase later in life, possibly owing to stronger occlusal forces, tooth wear, and other age-associated factors. Dachshunds may represent another high-risk group, in which long and crowded dentition, combined with breed-specific genetic factors, contribute to both a higher baseline risk and a steep age-related increase. Sighthounds were likewise classified into a pattern characterized by both a high baseline risk and a steep age-related increase. Therefore, these two distinct groups are expected to have a particularly high lifetime risk for periodontal disease. Greyhounds, which belong to the sighthound group, have repeatedly been reported as high-risk breeds, despite the general trend that small breeds tend to exhibit higher periodontal disease susceptibility (20, 37). This may indicate that the high-risk and strong age-effect pattern observed for the sighthound group was consistently detected because this group inherently reflects such breed-specific characteristics. These findings imply that even breeds traditionally considered to be at low risk may require dental care equivalent to that required by high-risk breeds if they exhibit steep age-related increases in risk.

However, several breeds deviated from general trends in their respective groups. For example, the French Bulldog, classified within the companion group (9G), which is generally characterized by high baseline and moderate age-related risk, showed an exceptionally low baseline risk (79th of 81 breeds) but a markedly high age-related effect (4th of 81 breeds). Similarly, the Siberian Husky, belonging to the spitz group (5G), showed a low baseline rank (70th out of 81 breeds) but one of the highest age-related effects (5th out of 81 breeds). Such inconsistencies indicate that morphological, genetic, or behavioral factors unique to each breed may modulate the risk of periodontal disease beyond group-level characteristics. The specific traits responsible for these differences are unclear. One possible approach to identify such determinants is to focus on breeds that share similar baseline risks but show markedly different age-related effects. For example, the Bichon Frise and Kaninchen Dachshund exhibited the two strongest age-related effects, whereas the Yorkshire Terrier and Papillon showed the weakest effects despite having comparable baseline risks. Such contrasting breeds may help elucidate the biological mechanisms underlying age-related periodontal deterioration, including connective tissue fragility, immune senescence, and age-associated changes in the oral microbiota, as well as the genetic and morphological traits that influence these processes.

In cats, differences were observed only in age-related effects, whereas baseline risk showed minimal variation among breeds. The annual age effect ranged from an OR of 1.144 in Siamese cats to 1.056 in Russian Blues, indicating a nearly 9% difference that could result in a substantial divergence in prevalence with advancing age. Girard et al. (18) and O’Neill et al. (9) reported higher risks in certain breeds, whereas Lommer and Verstraete (17) and Mestrinho et al. (23) reported no significant breed effects. In the present study, baseline risk did not differ among breeds, but a significant variation was detected in the age effect. In dogs, baseline risk is strongly influenced by body size and structural diversity among breeds; however, these factors may be less variable in cats or may have a smaller impact on periodontal risk. O’Neill et al. (9) also noted that breed-related differences in periodontal disease risk were smaller in cats than in dogs. Among breeds previously identified as predisposed to periodontal disease, Siamese, Somali, and Persian cats ranked 1st, 6th, and 7th, respectively, out of 38 breeds for the age effect, suggesting that their previously reported high risk may reflect stronger age-related increases. Conversely, Maine Coons and Bengals ranked 20th and 35th, respectively, indicating that these breeds may maintain intermediate or lower risks even at older ages. Brachycephalic breeds tend to exhibit relatively high age effects, which may be attributable to the craniofacial conformation or occlusal morphology, which increases mechanical stress on the dentition. In particular, Exotic Shorthair, Exotic, Himalayan, and Persian cats ranked 2nd, 3rd, 4th, and 7th, respectively, for the age effect, highlighting a consistent trend among short-headed breeds. These findings suggest that cat breeds with stronger age-related effects may experience an earlier onset or faster progression of periodontal disease. Therefore, these breeds would benefit from regular oral examinations and preventive dental care at a young age.

This study used a nationwide insurance claims database in Japan to comprehensively evaluate the risk of periodontal disease according to age, body size, breed group, and individual breeds. The use of large-scale, real-world clinical data is a major strength of this study; however, several limitations should be noted. First, the insurance claims data did not include information on disease stage or severity, making it impossible to distinguish between gingivitis and advanced periodontitis. Marshall et al. (21) reported that the progression of periodontal disease accelerates with age in Miniature Schnauzers, underscoring the need for longitudinal stage-specific analyses to assess disease progression. Second, environmental and behavioral factors such as housing conditions, feeding practices, and oral care habits were not considered. For example, soft diets have been associated with an increased risk of periodontal disease (6), and the presence or absence of home oral care has been correlated with oral health status in cats and dogs (38). These variables may influence the onset and progression of periodontal disease, and future studies should incorporate questionnaire-based or behavioral data to complement clinical information. Third, although sex was a significant factor in both dogs and cats, information on the neuter or spay status was not available. In cats, neutered cats were reported to exhibit higher odds of developing periodontal disease than were intact cats (9). Therefore, a potential bias related to reproductive status cannot be ruled out, and the observed sex effects may not accurately reflect true biological differences. In addition, a Japanese study demonstrated that the diversity of periodontal disease–associated bacteria increased with age in dogs (32). This finding suggests that age-related changes in oral microbiota and immune senescence may partially explain the differences in age-related effects observed among breed groups or individual breeds. Future studies that integrate longitudinal assessments of disease progression with microbiological and immunological analyses are essential for a more precise understanding of periodontal disease dynamics and risk stratification. Overall, the present findings highlight the importance of preventive dental care strategies tailored to age and breed characteristics and underscore the need for early diagnosis and individualized oral management in veterinary practice.