Periodontal disease is characterized by inflammation and destruction of some or all of the tooth-supporting structures (periodontum), which include the gingiva (gingivitis), cementum, periodontal ligament, and alveolar bone (periodontitis). Periodontal disease is considered to be one of the most common health problems affecting dogs.1 The prevalence of periodontal disease in dogs substantially increases with age but decreases with increasing body weight.2 The pathogenesis of periodontal disease involves bacteria, primarily gramnegative motile anaerobic rods, that accumulate within the gingival sulcus. This causes inflammation of the gingiva and formation of periodontal pockets, which result in periodontitis when left untreated. This is accompanied by an inflammatory response to periodontal pathogens resulting in endotoxin formation that mediates local release of inflammatory cytokines. The American Veterinary Dental Society3 cautions pet owners, “Studies have shown that oral bacteria will be filtered out by the kidney and liver, and can cause microabscesses within these organs. This leads to a decrease in function of these vital organs over time. In addition, it has been suggested that these bacteria can become attached to the heart valves and cause a disease called endocarditis.”
Because the prevalence of periodontal disease in dogs is > 75%,4 any association of periodontal disease with systemic organ damage is important to canine health. Systemic disease in dogs with chronic periodontal disease has been attributed to bacteremia and absorption of bacterial toxins from the oral cavity.5 Systemic diseases suggested to be associated with periodontal disease in dogs include chronic bronchitis, pulmonary fibrosis, endocarditis, interstitial nephritis, glomerulonephritis, and hepatitis. One report6 of 45 dogs at necropsy found a statistical relationship between the extent of periodontal disease and histopathologic changes in the kidneys, myocardium (papillary muscle), and liver.
A review7 of numerous case-control and longitudinal studies in humans found a significant association between periodontitis and an increased risk of cardiovascular disease. Many of the studies contained large numbers of subjects and were adjusted for traditional host-related risk factors and potential confounders. The availability of millions of medical records in electronic format from Banfield the Pet Hospital allowed us to conduct a similar epidemiologic study. In the study reported here, we tested the hypothesis that an increase in the severity of periodontal disease in dogs is associated with a subsequent increase in the risk of cardiovascular-related events, such as endocarditis and cardiomyopathy, as well as markers of inflammation but that an increase in the severity of periodontal disease in dogs is not associated with a subsequent increase in the risk of common endocrine, urinary tract, and musculoskeletal conditions; cancer; or infectious diseases.
Materials and Methods
Data source—Banfield the Pet Hospital operates a national network of > 650 full-service primary-care veterinary hospitals in 44 states. There are approximately 80,000 dog and cat office visits each week at these hospitals. Banfield practices use proprietary softwarea to create electronic medical records that are uploaded nightly to a central data warehouse where they are stored.b All medical records of dogs from 2002 through 2006 were accessed and converted into data setsc for selection of periodontal and nonperiodontal cohorts for this study and subsequent statistical analyses.
Periodontal cohort selection—Dogs were selected for the periodontal cohort when they had a diagnostic code of periodontal disease stage 1, 2, or 3 during the years 2002 through 2006. First, all dogs with a diagnosis of periodontal stage 3 were selected and the date of their first diagnosis was recorded (entry date into study). Then, all dogs with a diagnosis of periodontal disease stage 2 were selected in the same manner, provided they had never had a diagnosis of periodontal disease stage 3. Finally, all dogs with a diagnosis of periodontal disease stage 1 were selected, provided they never had a diagnosis of periodontal disease stage 3 or 2. Dogs with periodontal disease were not included in the study when there were missing or inconsistent data (eg, birth date after death date) or the stage of periodontal disease was not recorded, they had no follow-up visits after the diagnosis of periodontal disease, or they were > 15 years old.
Veterinarians were provided with written descriptions and photographs to help them identify the clinical stages8 of periodontal disease in dogs. Stage 1 periodontal disease is characterized by acute gingival inflammation with no loss of gingival attachment. Stage 2 is characterized by chronic gingivitis for > 6 months and up to 25% attachment loss or alveolar bone loss; periodontal pockets may exist. Stage 3 periodontal disease is characterized by up to 50% attachment loss or alveolar bone loss (1 to 3 mm) with periodontal pockets and possible gingival recession and root exposure. At the time of this study, veterinarians were not asked to distinguish between stages 3 and 4 periodontal disease, and radiographs were not required to make a stage determination.
Nonperiodontal cohort selection—Dogs were selected for the nonperiodontal cohort for the years 2002 through 2006 when they never had a diagnosis of periodontal disease and never had any of the following examination findings coded in their medical record: gingival recession, gingivitis, swelling or inflammation of gums, infected pockets in gums, or receding gums. Nonperiodontal cohort dogs were frequency matched to periodontal cohort dogs on the basis of year of visit and age. A random sample of nonperiodontal dogs was selected from each calendar year such that the size of the nonperiodontal cohort was maximized and that within each 3-month age group, the proportion of nonperiodontal dogs matched the proportion of periodontal dogs with stage 1, 2, or 3 periodontal disease in each calendar year. This resulted in a nonperiodontal cohort that was equal in size to the total periodontal cohort. For dogs > 13 years old, 6-month age groups were used for frequency matching because of small sample sizes. After forming the potential nonperiodontal cohort in each age category, nonperiodontal cohort dogs were randomly selected, with each dog in an age–calendar year group having the same probability of being selected. Similar to dogs in the periodontal cohort, dogs in the nonperiodontal cohort were excluded when there were missing or inconsistent data, they had no follow-up visits after the selection date, or they were > 15 years old.
Noncardiovascular outcome events—Diseases related to various organ systems were evaluated for an association with periodontal disease. These included cruciate ligament rupture, hip dysplasia, aggressive behavior, urinary incontinence, borreliosis (ie, Lyme disease), hemangiosarcoma, mast cell tumor, hypothyroidism, and diabetes mellitus. These diseases were selected to serve as comparison outcome events because there is no published scientific evidence in the veterinary literature to associate the risk of their occurrence with the presence or severity of periodontal disease.
Cardiovascular and inflammatory outcome events—Cardiovascular-related events evaluated for an association with periodontal disease included both specific cardiac diagnoses and clinical examination findings suggestive of cardiovascular disease. The diagnoses included dilated cardiomyopathy, hypertrophic cardiomyopathy, endocarditis, mitral valve insufficiency, other disease involving a heart valve (tricuspid valve insufficiency and ruptured chordae tendinae), cardiac murmur, abnormal cardiac rhythm, and congestive heart failure. The examination findings included cardiac murmur, abnormal pulse (fast, weak, bounding, missed beats, and slow), cardiac arrhythmia (atrial fibrillation, atrial premature contraction, atrial standstill, atrial tachycardia, atrial-ventricular dissociation, sinus bradycardia, bundle-branch block, first- or second-degree heart block, junctional premature beats, junctional tachycardia, sinus tachycardia, ventricular fibrillation, ventricular premature complexes, and ventricular tachycardia), and abnormal capillary refill time (> 2 seconds). Cardiovascular disease–related clinical events, such as coughing, tiring easily, and fainting, were included in the analyses because they are manifestations of a failing heart. Biomarkers of inflammation in dogs were also evaluated for an association with periodontal disease. The inflammatory biomarkers included an elevated WBC count (> 17 × 103 cells/PL) and an increase in the percentage of monocytes in blood samples (> 10%). Inflammation was included as an outcome event because it has been proposed as the biological process that links periodontal and cardiovascular diseases in humans.9,10
Missing data on body weight—Body weight was considered an important variable for the data analyses, but it was recorded for only 71% of all visits. Therefore, body weight of dogs was estimated for all visits in which weight was not recorded. A transformation-interpolation–back transformation method was used for dogs with at least 1 recorded weight. The transformation function used was the mean weight of all dogs in the database. Extrapolation along the basic shape of the weight versus age curve was performed for dogs when the missing weight was before the first recorded weight or after the last recorded weight. Interpolation was not possible when weight was not recorded for a dog at any visit. For that situation, a weight estimation procedure based on the mean weight for dogs with the same age, breed, sex, and neuter status was used to estimate body weight. To assess the accuracy of this weight estimation procedure, a random sample of 1% of body weights in the canine study population was selected and assigned as missing. Then, the weight estimation procedure was used to estimate those assigned missing weights. By use of this method for 72,747 assigned missing weights, the actual and assigned missing weights differed by a mean absolute error of 1.5 kg (3.2 lb) and there was significant correlation (r2 = 0.97) between the actual and estimated missing weights.
Data analysis—The incidence per 100 dogs for noncardiovascular outcome events and cardiovascular-related outcome events was calculated for dogs without evidence of periodontal disease and separately for dogs with periodontal disease (stage 1, 2, or 3). Included in these calculations were dogs that had the event of interest after the date of entry into the study. Excluded from these calculations were dogs that had the event of interest prior to or on the date of entry into the study. In calculating the incidence rates, the total number of office visits and duration of follow-up evaluation were not used as control variables.
A Cox proportional hazard regression model was used to estimate the risk of noncardiovascular and cardiovascular outcome events among dogs in the periodontal cohort, compared with risk for similar events among dogs in the nonperiodontal cohort, as a function of the stage of periodontal disease (1, 2, or 3) over time. The model also controlled for the effect of potential confounding factors. The proportional hazard model is the most general of the regression models because it is not based on any assumptions concerning the nature or shape of the underlying survival function. For this model, it is assumed that the underlying hazard rate (rather than survival time) is a function of the independent variables (covariates).11 The hazard ratio derived from the hazard model in a longitudinal study is interpreted in a manner similar to that of the odds ratio when logistic regression is used in a case-control study. A hazard ratio > 1.0 indicates an increased risk of an event in a group of dogs after a diagnosis of periodontal disease, compared with the risk for dogs that do not have periodontal disease, whereas a hazard ratio < 1.0 indicates a decreased risk. Statistical significance (P < 0.05) is evident when the 95% confidence limits of the hazard ratio exclude 1.0. The same criteria that were used to select dogs to calculate incidence rates of outcome events were also used to select dogs to calculate the Cox proportional hazard.
The time variable was computed as the number of days from the date of entry into the study to the date of the outcome event of interest. Host-independent variables (age, weight, sex, neuter status, and breed) were determined at time of entry into the study. Also included as independent variables in all Cox models were the number of office visits after the study entry date and the number of months from entry into the study until the last office visit. The variable age was entered as follows: 0 to < 2, 2 to < 4, 4 to < 6, 6 to < 8, 8 to < 10, and ≥ 10 years. The variable weight was entered as 0.5 to < 4.5 kg (1 to < 10 lb), 4.5 to < 13.6 kg (10 to < 30 lb), 13.6 to < 22.7 kg (30 to < 50 lb), 22.7 to < 34.1 kg (50 to < 75 lb), 34.1 to < 45.5 kg (75 to < 100 lb), and ≥ 45.5 kg. The variables sex, neuter status, breed, number of office visits during the study, and number of months of follow-up monitoring were entered in all Cox regression models, regardless of their statistical significance, whereas age and weight were selected on the basis of statistical significance (P < 0.1) by use of the stepwise procedure. All office visits were included in the study from entry date through the last office visit; the last office visit was used as a right-censoring date for hazard modeling.
Survival curves were calculated and graphed on the basis of the fitted Cox models for dilated cardiomyopathy, hypertrophic cardiomyopathy, endocarditis, mitral valve insufficiency, abnormal rhythm, and elevated WBC count. The estimated survival probability was plotted against the number of days in the study.
Results
The study population consisted of 59,296 dogs with a history of periodontal disease, of which 23,043 had stage 1 disease, 20,732 had stage 2 disease, and 15,521 had stage 3 disease. These dogs were frequency matched on the basis of year of diagnosis and age to 59,296 dogs with no history of periodontal disease. Mean ± SE and median age and weight, sex, and neuter status of dogs in each cohort were calculated (Tables 1 and 2). Given the relatively large number of dogs within each group, values for differences between groups by age, weight, sex, and neuter status were not reported because even small and clinically irrelevant differences were likely to be statistically significant.
Age and body weight of dogs with and without periodontal disease.
| Variable | Nonperiodontal cohort (n = 59,296) | |||||||
|---|---|---|---|---|---|---|---|---|
| Stage 1 (n = 23,043) | Stage 2 (n = 20,732) | Stage 3 (n =15,521) | ||||||
| Mean ± SE | Median | Mean ± SE | Median | Mean ± SE | Median | Mean ± SE | Median | |
| Age (y) | 7.5 ± 0.02 | 7.2 | 6.1 ± 0.02 | 5.5 | 7.5 ± 0.02 | 7.3 | 9.4 ± 0.03 | 9.7 |
| Weight (kg)* | 20.6 ± 0.05 | 19.5 | 18.5 ± 0.09 | 14.5 | 16.8 ±0.09 | 12.1 | 12.4 ± 0.09 | 8.2 |
* To convert to pounds, multiple value by 2.2.
n = Total number of dogs in the category.
Sex of dogs with and without periodontal disease.
| Variable | Nonperiodontal cohort (n = 59,296) | Periodontal cohort | ||
|---|---|---|---|---|
| Stage 1 (n = 23,043) | Stage 2 (n = 20,732) | Stage 3 (n = 15,521) | ||
| Male | 8,404 (14.2) | 3,256 (14.1) | 2,430 (11.7) | 2,094 (13.5) |
| Male neutered | 20,440 (34.5) | 8,263 (35.9) | 7,685 (37.1) | 5,462 (35.2) |
| Female | 5,863 (9.9) | 2,364 (10.3) | 1,775 (8.6) | 1,461 (9.4) |
| Female spayed | 24,589 (41.5) | 9,160 (39.8) | 8,842 (42.7) | 6,504 (41.9) |
| Total | 59,296 (100.1*) | 23,043 (100.1*) | 20,732 (100.1*) | 15,521 (100.0) |
Values reported as number (%) of dogs.
* Values do not sum to 100% because of rounding.
The frequency of diseases not expected to be associated with a prior diagnosis of periodontal disease (ie, noncardiovascular events) was found to be similar for dogs with periodontal disease and dogs free of periodontal disease (Table 3). In addition, there was no consistent pattern indicating that the frequency of such noncardiovascular events increased as the severity of periodontal disease increased. By comparison, the frequency of cardiovascular-related events was generally higher for dogs with periodontal disease than for dogs without periodontal disease, and this frequency increased with an increasing severity of periodontal disease (Table 4). For example, the risk of endocarditis was approximately 6-fold higher for dogs with stage 3 periodontal disease than it was for dogs in the nonperiodontal cohort and dogs with stage 1 periodontal disease. A similar increase was detected for inflammatory biomarkers such as an elevated WBC count and an increase in the percentage of monocytes in blood samples.
Frequency of noncardiovascular events during the follow-up period for dogs with or without periodontal disease.
| Event | Nonperiodontal cohort (n = 59,296) | Periodontal cohort | ||
|---|---|---|---|---|
| Stage 1 (n = 23,043) | Stage 2 (n = 20,732) | Stage 3 (n = 15,521) | ||
| Cruciate ligament rupture | 229 (0.4) | 134 (0.6) | 118 (0.6) | 69 (0.4) |
| Hip dysplasia | 243 (0.4) | 146 (0.6) | 120 (0.6) | 72 (0.5) |
| Aggressive behavior | 394 (0.7) | 211 (0.9) | 154 (0.8) | 107 (0.7) |
| Urinary incontinence | 181 (0.3) | 102 (0.4) | 107 (0.5) | 100 (0.6) |
| Borreliosis | 98 (0.2) | 38 (0.2) | 29 (0.2) | 38 (0.2) |
| Hemangiosarcoma | 47 (0.1) | 17 (0.1) | 25 (0.1) | 16 (0.1) |
| Mast cell tumor | 105 (0.2) | 58 (0.3) | 52 (0.3) | 42 (0.3) |
| Hypothyroidism | 450 (0.8) | 245 (1.1) | 259 (1.3) | 182 (1.2) |
| Diabetes mellitus | 165 (0.3) | 75 (0.3) | 72 (0.4) | 72 (0.5) |
Values reported as number (%) of dogs.
See Table 1 for remainder of key.
Frequency of cardiovascular-related events during the follow-up period for dogs with or without periodontal disease.
| Event | Nonperiodontal cohort (n = 59,296) | Periodontal cohort | ||
|---|---|---|---|---|
| Stage 1 (n = 23,043) | Stage 2 (n = 20,732) | Stage 3 (n = 15,521) | ||
| Cardiac diagnoses | ||||
| Dilated cardiomyopathy | 37 (0.06) | 21 (0.09) | 24 (0.12) | 49 (0.32) |
| Hypertrophic cardiomyopathy | 10 (0.02) | 8 (0.03) | 14 (0.07) | 24 (0.15) |
| Endocarditis | 5 (0.01) | 3 (0.01) | 8 (0.04) | 23 (0.15) |
| Mitral valve insufficiency | 37 (0.06) | 29 (0.13) | 55 (0.27) | 87 (0.56) |
| Heartvalve disease | 39 (0.07) | 33 (0.14) | 58 (0.28) | 92 (0.59) |
| Cardiac murmur | 752 (1.27) | 411 (1.78) | 572 (2.76) | 735 (4.74) |
| Abnormal cardiac rhythm | 54 (0.09) | 54 (0.23) | 56 (0.27) | 85 (0.55) |
| Congestive heart failure | 257 (0.43) | 114 (0.49) | 226 (1.09) | 295 (1.90) |
| Cardiac examination findings | ||||
| Cardiac murmur | 6 (0.01) | 8 (0.03) | 7 (0.03) | 14 (0.09) |
| Abnormal pulse | 136 (0.23) | 82 (0.36) | 117 (0.56) | 132 (0.85) |
| Cardiac arrhythmia | 551 (0.93) | 317 (1.38) | 461 (2.22) | 552 (3.56) |
| Abnormal capillary refill time* | 95 (0.16) | 81 (0.35) | 75 (0.36) | 80 (0.52) |
| Other clinical signs | ||||
| Coughing | 1,208 (2.04) | 647 (2.81) | 715 (3.45) | 703 (4.53) |
| Tiring easily | 647 (1.09) | 316 (1.37) | 362 (1.75) | 405 (2.61) |
| Fainting | 53 (0.09) | 17 (0.07) | 44 (0.21) | 43 (0.28) |
| Inflammatory biomarkers | ||||
| Elevated WBC count | 2,520 (4.25) | 1,662 (7.21) | 1,896 (9.15) | 1,763 (11.36) |
| Elevated percentage of monocytes | 2,050 (3.46) | 1,622 (7.04) | 1,663 (8.02) | 1,405 (9.05) |
Values reported as number (%) of dogs.
* Defined as > 2 seconds.
See Table 1 for remainder of key.
Cox regression models were developed to examine the relationship between the severity of periodontal disease and the risk of events of interest while controlling for potential confounding effects of age, weight, sex, neuter status, whether a dog was a purebred or mixed-breed dog, duration of follow-up monitoring, and the number of office visits during the follow-up period. Of the noncardiovascular events evaluated, the hazard ratio was significantly increased in dogs with stage 1 or 3 periodontal disease for hip dysplasia and in dogs with stage 3 periodontal disease for urinary incontinence and Lyme disease (Table 5). However, for both hip dysplasia and Lyme disease, there was no evidence of an increasing hazard ratio with increasing severity of periodontal disease, compared with results for dogs with no history of periodontal disease. Even for dogs that developed urinary incontinence, the risk of such an event was only 36% higher for dogs with stage 3 periodontal disease, compared with the risk for dogs with no history of periodontal disease. Analysis of the pattern for these noncardiovascular events indicated that the risk was not associated with prior diagnosis or severity of periodontal disease and was more likely attributable to a type I statistical error.
Cox regression models for noncardiovascular events during the follow-up period for dogs with and without periodontal disease.
| Dependent variable | All | Nonperiodontal cohort (n = 59,296) | Periodontal cohort | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Stage 1 (n = 23,043) | Stage 2 (n = 20,732) | Stage 3 (n = 15,521) | |||||||||||||
| n | No. of dogs | No. of dogs | No. of dogs | Ratio | LCL | UCL | No. of dogs | Ratio | LCL | UCL | No. of dogs | Ratio | LCL | UCL | |
| Cruciate ligament rupture | 117,898 | 550 | 229 | 134 | 1.11 | 0.89 | 1.38 | 118 | 1.08 | 0.86 | 1.35 | 69 | 1.03 | 0.77 | 1.36 |
| Hip dysplasia | 117,666 | 580 | 242 | 146 | 1.31* | 1.06 | 1.61 | 120 | 1.26 | 1.00 | 1.57 | 72 | 1.40* | 1.06 | 1.83 |
| Aggressive behavior | 117,168 | 866 | 394 | 211 | 1.02 | 0.86 | 1.21 | 154 | 0.95 | 0.78 | 1.15 | 107 | 1.06 | 0.84 | 1.32 |
| Urinary incontinence | 117,858 | 486 | 180 | 102 | 1.20 | 0.94 | 1.54 | 105 | 1.15 | 0.90 | 1.47 | 99 | 1.36* | 1.05 | 1.76 |
| Borreliosis | 118,177 | 202 | 98 | 37 | 0.84 | 0 57 | 1.23 | 29 | 0.74 | 0 48 | 1.13 | 38 | 1.58* | 106 | 2.35 |
| Hemangiosarcoma | 118,307 | 105 | 47 | 17 | 1.08 | 0.61 | 1.90 | 25 | 1.35 | 0.82 | 2.22 | 16 | 1.14 | 0.63 | 2.06 |
| Mast cell tumor | 118,153 | 257 | 105 | 58 | 1.20 | 0.87 | 1.67 | 52 | 1.02 | 0.73 | 1.43 | 42 | 1.13 | 0.78 | 1.64 |
| Hypothyroidism | 117,049 | 1,135 | 449 | 245 | 1.08 | 0.92 | 1.26 | 259 | 1.11 | 0.95 | 1.30 | 182 | 1.11 | 0.93 | 1.33 |
| Diabetes mellitus | 117,989 | 381 | 163 | 75 | 0.94 | 0.71 | 1.24 | 71 | 0.74 | 0.56 | 1.00 | 72 | 0.83 | 0.62 | 1.10 |
The independentvariables in each model were stage of periodontal disease, neuter status, breed (mixed-breed dogs vs purebred dogs), sex, age, and body weight. Nonperiodontal cohort dogs were the referent group; therefore, the estimated hazard ratio (ratio) for this category was 1.00.
* Ratio was significant (P < 0.05) because the 95% confidence limits excluded 1.00.
LCL = Lower 95% confidence limit for the ratio. UCL= Upper 95% confidence limit for the ratio.
See Table 1 for remainder of key.
Similar multivariable Cox regression models were developed to evaluate the relationship between the severity of periodontal disease and cardiovascular-related events and also inflammatory biomarkers. For most cardiovascular-related events evaluated, the hazard ratios increased with an increasing severity of periodontal disease and were significantly increased (> 1.0) for dogs with a history of stage 3 periodontal disease (Table 6; Figure 1). The highest hazard ratio was for endocarditis (hazard ratio, 6.36) and for hypertrophic cardiomyopathy (hazard ratio, 3.96) for dogs with a prior diagnosis of stage 3 periodontal disease. A similar pattern of increasing hazard ratio with increasing severity of periodontal disease was detected for an elevated WBC count and an increase in the percentage of monocytes as well as for systemic signs often associated with a failing heart, such as coughing and tiring easily.
Cox regression models for cardiovascular-related events and inflammatory biomarkers during the follow-up period for dogs with or without periodontal disease.
| Dependent variable | All | Nonperiodontal cohort (n = 59,296) | Periodontal cohort | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Stage 1 (n = 23,043) | Stage 2 (n = 20,732) | Stage 3 (n = 15,521) | |||||||||||||
| n | No. of dogs | No. of dogs | No. of dogs | Ratio | LCL | UCL | No. of dogs | Ratio | LCL | UCL | No. of dogs | Ratio | LCL | UCL | |
| Cardiovascular diagnoses | |||||||||||||||
| Dilated cardiomyopathy | 118,232 | 131 | 37 | 21 | 1.13 | 0.66 | 1.95 | 24 | 1.10 | 0.65 | 1.85 | 49 | 2.44* | 1.55 | 3.83 |
| Hypertrophic cardiomyopathy | 118,300 | 56 | 10 | 8 | 1.43 | 0.56 | 3.67 | 14 | 2.17 | 0.95 | 4.97 | 24 | 3.96* | 1.82 | 8.61 |
| Endocarditis | 118,299 | 39 | 5 | 3 | 1.10 | 0.26 | 4.65 | 8 | 2.25 | 0.72 | 7.01 | 23 | 6.36* | 2.34 | 17.25 |
| Mitral valve insufficiency | 118,155 | 207 | 37 | 29 | 1.60 | 0.98 | 2.62 | 54 | 2.06* | 1.35 | 3.16 | 87 | 2.74* | 1.85 | 4.07 |
| Tricuspid valve insufficiency | 118,317 | 17 | 2 | 5 | 4.18 | 0.78 | 22.29 | 3 | 1.84 | 0.30 | 11.32 | 7 | 3.03 | 0.61 | 15.06 |
| Ruptured chordae tendinae | 118,328 | 7 | 2 | 0 | 0.00 | 0.00 | NA | 5 | 3.60 | 0.66 | 19.62 | 0 | 0.00 | 0.00 | NA |
| Cardiac murmur | 114,757 | 2,464 | 750 | 411 | 1.22* | 1.07 | 1.37 | 570 | 1.40* | 1.25 | 1.56 | 733 | 1.76* | 1.58 | 1.95 |
| Abnormal cardiac rhythm | 118,139 | 249 | 54 | 54 | 2.00* | 1.36 | 2.93 | 56 | 1.93* | 1.32 | 2.83 | 85 | 3.30* | 2.30 | 4.72 |
| Congestive heartfailure | 117,784 | 887 | 256 | 114 | 1.02 | 0.81 | 1.27 | 222 | 1.37* | 1.14 | 1.65 | 295 | 1.49* | 1.25 | 1.77 |
| Cardiovascular examination findings | |||||||||||||||
| Cardiac murmur | 118,296 | 35 | 6 | 8 | 2.54 | 0.87 | 7.45 | 7 | 1.78 | 0.59 | 5.40 | 14 | 3.34* | 1.23 | 9.08 |
| Abnormal pulse | 118,095 | 464 | 135 | 81 | 1.48* | 1.12 | 1.96 | 117 | 1.70* | 1.32 | 2.19 | 131 | 1.78* | 1.38 | 2.30 |
| Cardiac arrhythmia | 116,211 | 1,878 | 550 | 317 | 1.26* | 1.10 | 1.46 | 460 | 1.48* | 1.30 | 1.68 | 551 | 1.70* | 1.50 | 1.92 |
| Abnormal capillary refill time† | 118,222 | 327 | 94 | 80 | 2.09* | 1.54 | 2.84 | 75 | 1.65* | 1.21 | 2.25 | 78 | 1.68* | 1.23 | 2.31 |
| Other clinical signs | |||||||||||||||
| Coughing | 114,601 | 3,261 | 1,204 | 647 | 1.08 | 0.98 | 1.19 | 710 | 1.11* | 1.01 | 1.22 | 700 | 1.17* | 1.06 | 1.29 |
| Tiring easily | 117,338 | 1,725 | 644 | 315 | 1.19* | 1.04 | 1.37 | 362 | 1.13 | 1.00 | 1.29 | 404 | 1.33* | 1.17 | 1.52 |
| Fainting | 118,250 | 156 | 53 | 17 | 0.69 | 0.40 | 1.20 | 44 | 1.36 | 0.90 | 2.06 | 42 | 1.17 | 0.77 | 1.79 |
| Inflammatory biomarkers | |||||||||||||||
| Elevated WBC count | 114,029 | 7,814 | 2,507 | 1,661 | 1.37* | 1.29 | 1.47 | 1,893 | 1.58* | 1.48 | 1.68 | 1,753 | 1.84* | 1.73 | 1.96 |
| Increase in percentage of monocytes | 113,433 | 6,726 | 2,047 | 1,620 | 1.52* | 1.43 | 1.63 | 1,661 | 1.65* | 1.54 | 1.76 | 1,398 | 1.86* | 1.73 | 2.00 |
† Defined as > 2 seconds.
NA= Not applicable.
See Tables 1 and 5 for remainder of key.
Estimated survival curves for dilated cardiomyopathy (A), hypertrophic cardiomyopathy (B), endocarditis (C), mitral valve insufficiency (D), abnormal cardiac rhythm (E), and elevated WBC count (F) for dogs with periodontal disease (grade 1, dashed line; grade 2, dotted line; and grade 3, dotted and dashed line) and dogs without periodontal disease (solid line). The host-independent variables of age, weight, sex, neuter status, and breed were determined at the first office visit and were set to the mean values for all dogs. Also included as independent variables and set to their mean values were the number of office visits after the study entry date and the number of months from entry into the study until the last office visit. Notice that the scales on the x- and y-axes differ among the panels. * Day 0 was defined as the day of diagnosis of periodontal disease (periodontal cohort) or the day of entry into the study (nonperiodontal cohort).
Citation: Journal of the American Veterinary Medical Association 234, 4; 10.2460/javma.234.4.486
Discussion
Significant associations were detected between the severity of periodontal disease and a subsequent risk of cardiovascular-related conditions in this epidemiologic study of pet dogs. However, significant associations were not consistently detected between the severity of periodontal disease and the risk of a variety of common conditions, such as cruciate ligament rupture, hip dysplasia, urinary incontinence, Lyme disease, aggressive behavior, and several neoplastic and endocrine disorders. Three (hip dysplasia, urinary incontinence, and Lyme disease) of the 9 noncardiovascular-related diseases evaluated for an association with severity of periodontal disease were found to have a significantly increased frequency in dogs with a prior diagnosis of stage 3 periodontal disease. However, the lack of a monotonically increasing risk of these conditions with an increasing risk of periodontal disease suggests that these significant associations were chance findings that resulted from multiple comparisons and were not biological in nature. This is in contrast to the consistently increasing risk of most cardiovascular-related events associated with an increasing severity of periodontal disease.
A statistical association between periodontal disease and histologic lesions in multiple organs has been reported in a cross-sectional study6 of 45 dogs at necropsy; however, the temporal relationship between these conditions could not be evaluated. The strongest evidence for the systemic effects of periodontitis, including cardiovascular disease, comes from epidemiologic studies of humans. There have been at least 42 published observational studies12 in which investigators have described associations between oral conditions and cardiovascular diseases. Putative associations have also been described between periodontal disease and stroke, pneumonia, diabetes mellitus, and adverse outcomes for pregnancy.13 These associations between periodontal disease and heart disease in humans appear moderate in magnitude. Nonetheless, given the high prevalence of both periodontal disease and heart disease in humans and dogs, it is possible that routine use of prophylactic periodontal procedures could have profound health consequences, provided the detected associations between periodontal disease and heart disease are causal in nature.
The strongest association in the current study was between stage 3 periodontal disease and an increased risk of endocarditis. For example, the risk of endocarditis was approximately 6-fold higher for dogs with stage 3 periodontal disease, compared with the risk for dogs in the nonperiodontal cohort. Infective endocarditis is a microbial infection of a cardiac valve or surrounding cardiac tissue that may be caused by various microorganisms, including bacteria, fungi, and rickettsiae,14 and results in proliferative (vegetative) lesions and destruction of valvular tissue.15 Numerous microorganisms, including Corynebacterium spp, Erysipelothrix rhusiopathiae, and Streptococcus spp, have been isolated from dogs with endocarditis of the aortic valve.16 Aortic valve endocarditis caused by Bartonella clarridgeiae has been described in a dog,17 and a prospective study18 of infective endocarditis in dogs found that Bartonella spp were causative organisms. The 2 most common microorganisms associated with infective endocarditis in humans are Streptococcus viridans and Staphylococcus aureus, either of which may be normal commensals in the oral cavity. Other causative microorganisms for infective endocarditis include Actinobacillus actinomycetemcomitans, which is a putative periodontal pathogen. It is not known whether infective microbes colonize microscopic sterile lesions or directly invade normal epithelium. In humans, it has been reported19 that transient bacteremias frequently develop following periodontal procedures, and periodontal bacteria have been isolated in atheromatous plaques in patients with chronic periodontitis.20
Mitral valve endocarditis may also be a complication of hypertrophic cardiomyopathy in humans and affect the left ventricular aspect of the anterior mitral valve leaflet.21 This, in part, may explain the significant association detected in the present study between cardiomyopathy (both dilated and hypertrophic), mitral valve insufficiency, and cardiac murmurs and an increasing severity of periodontal disease. The primary causes of cardiomyopathy in dogs are unknown.
To evaluate the potential benefit of periodontal treatment for improving cardiovascular health, a 6-month study22 of 40 human patients with severe chronic generalized periodontitis who received intensive periodontal treatment (including the adjunctive use of locally delivered antimicrobials) or standard treatment consisting of subgingival scaling and root planning was conducted. The patients that received intensive treatment had a significant reduction in systemic inflammatory markers, reductions in systolic blood pressure, and improvements in lipid profiles, compared with results for patients that received standard treatment.22 Large, randomized, and controlled clinical trials have not yet been conducted to definitively determine whether periodontal treatments reduce overall cardiovascular risk.23
Little research has been conducted on the benefits of periodontal treatment on systemic health in dogs, despite estimates that approximately 75% or more of dogs will develop periodontitis to some degree by 4 years of age24 and the finding of a significant association between periodontal disease and histologic lesions in multiple organs of dogs at necropsy.6 A multicenter clinical study25 in which investigators evaluated bacterial cultures from 33 dogs with periodontitis resulted in 156 bacterial isolates, of which black-pigmented anaerobic bacteria were the most common. Porphyromonas denticanis, Porphyromonas gulae, and Porphyromonas salivosa were recovered from 76% of the dogs. In another study,26 Porphyromonas spp were isolated from the plaque of adult dogs but not from any oral sites of puppies and adolescent dogs. These and other similar findings have stimulated development and evaluation of a vaccine against Porphyromonas spp for use in dogs. When this vaccine was evaluated in a controlled clinical trial,d it resulted in a significant improvement in bone reactivity scores in dogs vaccinated and then challenged-exposed, compared with results for unvaccinated, challenged-exposed dogs. Vaccination is not necessarily the optimal approach for preventing periodontal disease in dogs. However, it can play an important role when used in conjunction with other forms of dental treatment, including professional dental prophylaxis and home preventive care.
To our knowledge, the present study is the largest ever conducted of periodontal disease in dogs, but it had several limitations. Some limitations were attributable to the inability to standardize patient examinations and can be categorized as related to adequacy of staging of the severity of periodontal disease (the exposure variable) and to the accuracy of diagnosis of cardiovascular-related events (the outcome variable). The staging of periodontal disease was clinical in nature and did not require radiographic confirmation or probing of gingival pocket depth. Veterinarians were provided with guidelines that consisted of color photographs and written descriptions to assist with determining the stage of periodontal disease. It was not possible in this study to compare the clinical stage with the examination findings recorded in the medical record because some veterinarians had assigned a periodontal stage to a dog but failed to record any examination findings, whereas other veterinarians recorded examination findings for a dog but failed to formally record the periodontal stage. Dogs were only included in the present study when a veterinarian had assigned a stage of periodontal disease to the patient. Because the staging of periodontal disease was always performed prior to the onset of cardiovascular events, any misclassifications with regard to the stages of periodontal disease were assumed to be unbiased and likely to force the estimated hazard ratios toward unity. The fact that a dose-response relationship was detected between increasing severity of periodontal diseases and cardiovascular events, such as endocarditis and cardiomyopathy, suggested that the periodontal scoring system used was robust. Similar potential sources of measurement error have been indicated in a review27 of longitudinal studies of periodontal disease in humans.
A second major limitation to the study reported here was that diagnostic criteria were not established a priori for cardiovascular-related events. Rather, recording of examination findings or diagnosis of specific cardiac conditions was left to the judgment of each veterinarian as part of primary-care practice. Misclassification with regard to recording of cardiac events, similar to determining the stage of periodontal disease, was assumed to be unbiased because the data used were obtained during routine primary care and not as part of this study.
The retrospective nature of this study was another limitation, particularly with respect to inflammatory biomarkers. Clinical signs of periodontal disease result from microorganisms interacting with the host's immune system and the subsequent inflammatory response. Therefore, it is likely that biological markers of these interactions would provide direct measures of the microbial load and host response that constitutes periodontal disease. The present study in dogs evaluated crude measures of chronic inflammation, such as WBC count and monocyte percentage. However, the retrospective nature of the study did not allow for more direct measures of host-pathogen interactions, such as antibody titers against known periodontal pathogens or concentrations of monocyte-derived cytokines, C-reactive protein, or plasma fibrinogen. Future veterinary studies should consider including such markers.
Another limitation to this study is that instructions to owners or the treatments dogs may have received following a diagnosis of periodontal disease were not considered when evaluating the subsequent risk of developing cardiovascular- or noncardiovascular-related events. At the time of this study, it was the policy of Banfield the Pet Hospital to provide client education regarding oral hygiene and home care following a diagnosis of periodontal disease. In addition, stage-specific protocols were suggested for the treatment of dogs with periodontal disease that included prophylaxis every 4 to 6 months as needed and systemic administration of antimicrobials for 2 to 3 weeks for dogs with stage 3 disease. It is possible that treatment for periodontal disease may mitigate the risk for developing subsequent cardiovascular-related events, and this hypothesis is being evaluated.
The study reported here revealed a significant relationship between periodontal disease and an increased risk of endocarditis and cardiomyopathy in pet dogs. Future studies should be conducted to confirm these associations and explore possible explanatory mechanisms. Veterinarians should educate pet owners about the potential health implications of periodontal disease and encourage them to practice home dental care, including a thorough oral examination as an essential part of a wellness program in companion animals throughout all life stages. From a comparative perspective, the great variability in risk of periodontal and cardiovascular disease across dog breeds suggests that dogs could be particularly useful in identifying candidate genes linking periodontal and cardiovascular diseases.28
References
- 1.↑
Lund EM, Armstrong PJ, Kirk CA, et al. Health status and population characteristics of dogs and cats examined at private veterinary practices in the United States. J Am Vet Med Assoc 1999;214:1336–1341.
- 2.↑
Harvey CE, Shofer FS, Laster L. Association of age and body weight with periodontal disease in North American dogs. J Vet Dent 1994;11:94–105.
- 3.↑
American Veterinary Dental Society. Periodontal disease. Available at: www.avds-online.org/info/periodontaldisease.html. Accessed Mar 1, 2008.
- 4.↑
Colmery B, Frost P. Periodontal disease (etiology and pathogenesis). Vet Clin North Am Small Anim Pract 1986;16:817–834.
- 5.↑
Eisner ER. Periodontal disease in pets: the pathogenesis of a preventable problem. Vet Med (Praha) 1989;84:97–104.
- 6.↑
DeBowes LJ, Mosier D, Logan E, et al. Association of periodontal disease and histologic lesions in multiple organs from 45 dogs. J Vet Dent 1996;13:57–60.
- 7.↑
Beck JD, Offenbacher S. The association between periodontal diseases and cardiovascular diseases: a state-of-the-science review. Ann Periodontol 2001;6:9–15.
- 9.
Paquette DW, Brodala N, Nicholas TC. Cardiovascular disease, inflammation, and periodontal infection. Periodontol 2000 2007;44:113–126.
- 10.
Raskin RE, Latimer KS, Tvedten H. Leukocyte disorders. In: Willard MD, Tvedten H, eds. Small animal clinical diagnosis by laboratory methods. 4th ed. St Louis: Elsevier Saunders Co, 2004;75.
- 11.↑
Allison PD. Estimating Cox regression models with PROC PHREG. In: Survival analysis using the SAS system: a practical guide. Cary, NC: SAS Institute Inc, 1995;111–184.
- 12.↑
Beck JD, Offenbacher S. Systemic effects of periodontitis: epidemiology of periodontal disease and cardiovascular disease. J Periodontol 2005;76:2089–2100.
- 14.↑
Genco RJ, Offenbacher S, Beck J, et al. Cardiovascular diseases and oral infections. In Rose LF, ed. Periodontal medicine. Boulder, Colo: NetLibrary Inc, 2000;63–82.
- 15.↑
Greene CE. Cardiovascular infections. In: Infectious diseases of the dog and cat. 3rd ed. St Louis: Saunders Elsevier Inc, 2006;855–856.
- 16.↑
Sisson D, Thomas WP. Endocarditis of the aortic valve in the dog. J Am Vet Med Assoc 1984;184:570–577.
- 17.↑
Chomel BB, MacDonald KA, Kasten RW, et al. Aortic valve endocarditis in a dog due to Bartonella clarridgeiae. J Clin Microbiol 2001;39:3548–3554.
- 18.↑
MacDonald KA, Chomel BB, Kittleson MD, et al. A prospective study of canine infective endocarditis in northern California (1999–2001): emergence of Bartonella as a prevalent etiologic agent. J Vet Intern Med 2004;18:56–64.
- 19.↑
Kinane DF, Riggio MP, Walker KF, et al. Bacteraemia following periodontal procedures. J Clin Periodontol 2005;32:708–713.
- 20.↑
Padilla C, Lobos O, Hubert E, et al. Periodontal pathogens in atheromatous plaques isolated from patients with chronic periodontitis. J Periodontal Res 2006;41:350–353.
- 21.↑
Morgan-Hughes G. Mitral valve endocarditis in hypertrophic cardiomyopathy. Internet J Cardiol [serial online] 2003;2 (1). Available at: www.ispub.com/ostia/index.php?xmlFilePath=journals/ijc/vol2n1/mitral.xml. Accessed Jan 13, 2009.
- 22.↑
D'Aiuto F, Parkar M, Nibali L, et al. Periodontal infections cause changes in traditional and novel cardiovascular risk factors: results from a randomized controlled clinical trial. Am Heart J 2006;151:977–984.
- 23.↑
Offenbacher S, Beck JD. A perspective on the potential cardioprotective benefits of periodontal therapy. Am Heart J 2005;149:950–954.
- 24.↑
Kyllar M, Witter K. Prevalence of dental disorders in pet dogs. Vet Med (Praha) 2005;50:496–505.
- 25.↑
Hardham J, Dreier K, Wong J, et al. Pigmented-anaerobic bacteria associated with canine periodontitis. Vet Microbiol 2005;106:119–128.
- 26.↑
Isogai H, Kosako Y, Benno Y, et al. Ecology of genus Porphyromonas in canine periodontal disease. Zentralbl Veterinarmed [B] 1999;46:467–473.
- 27.↑
Genco R, Offenbacher S, Beck J. Periodontal disease and cardiovascular disease epidemiology and possible mechanisms. J Am Dent Assoc 2002;133:14S–22S.
- 28.↑
Kornman KS, Duff GW. Candidate genes as potential links between periodontal and cardiovascular diseases. Ann Periodontol 2001;6:48–57.
