Association of Radiographically Diagnosed Apical Periodontitis and Cardiovascular Disease: a Hospital Records–Based Study Gregory K

Clinical Research

Association of Radiographically Diagnosed Apical Periodontitis and Cardiovascular Disease: A Hospital Records–based Study Gregory K. An, DDS, MPH,* Douglas E. Morse, DDS, PhD,† Marc Kunin, DDS, MA,* Robert S. Goldberger, DDS,* and Walter J. Psoter, DDS, PhD†

Abstract Introduction: Numerous studies have demonstrated an he link between oral and systemic health has been debated for more than a century. association between oral health status and systemic dis- TIn the early 1900s the focal infection theory gained popularity and evoked fear eases. However, reports examining apical periodontitis among its subscribers. The theory suggested that many systemic illnesses were conse- (AP) and cardiovascular disease (CVD) are few. This quences of focal infections originating in the mouth (1, 2). Poor science and lack of study investigates whether an association exists be- evidence-based practice led to harmful consequences to patients. At the peak of its tween AP and CVD. Methods: The present study was popularity, edentulous therapy became the primary preventive measure (3). The theory a pair-matched, cross-sectional design that used medical was discredited by the mid-1900s; however, interest in the oral and systemic connection and dental chart review. The AP group (n = 182) was persisted. By the beginning of the twenty-first century, publications on the topic had defined as subjects with radiographic AP, and the non- grown exponentially. The first peer-reviewed study investigating oral health and cardio- AP group (n = 182) was defined as subjects without vascular disease (CVD) was published in 1989 (4); by 2010, the rate of publication on any radiographic AP. Samples for both groups were the subject matter had increased to more than 160 peer-reviewed articles per year and pair-matched by age and gender. Diagnosis for CVD, hy- more than 500 total articles since 1989 (5). Although the periodontal disease and CVD percholesterolemia, hypertension, and diabetes were link has received the most attention, a growing number of studies have examined other identified by using International Classification of associations between various oral health factors such as tooth loss, xerostomia, and Diseases, Ninth Revision, Clinical Modification and caries and systemic factors including vascular disease, diabetes, aspiration pneumonia, collected from electronic medical records. Documenta- and preterm birth. Few studies have examined the association between apical periodon- tion of alcohol use, smoking, race, and body mass index titis (AP) and CVD. within the electronic medical records was also collected. AP is an inflammatory process of endodontic origin usually occurring at or near Presence or absence of AP, missing teeth, teeth with root the apex of the tooth root. The biological explanatory model for the AP and CVD rela- canal treatment, caries experience, and history of peri- tionship resembles the well-established mechanistic evidence that exists for the periodontal disease were collected from the electronic odontal disease and CVD relationship (6–9), and both AP and periodontal disease dental records. Analysis was performed by using Pearson share similar bacterial flora, primarily gram-negative anaerobes, and similar destruc- c2, the paired t test, and conditional multivariate logistic tive inflammatory reactions (10, 11). Bacteria and inflammation have also been regression. Results: AP was significantly associated implicated in platelet aggregation, atherosclerosis, and the progression of CVD (7, 9). with CVD, hypercholesterolemia, race, missing teeth, The World Health Organization defines CVD as a group of disorders of the heart caries experience, and number of root canal treatments and blood vessels that include coronary heart disease, cerebrovascular disease, periph- in our bivariate analysis. Our final adjusted conditional eral arterial disease, rheumatic heart disease, congenital heart disease, deep vein logistic regression model showed statistically significant thrombosis, and pulmonary embolism (12). The goal of the current study was to inves- positive associations between AP and CVD (odds ratio, tigate an association between CVD and AP. In addition, other known covariates were 5.3; 95% confidence interval, 1.5–18.4). Conclusions: analyzed for their relationship with AP. Subjects with AP were more likely to have CVD than subjects without AP by 5.3-fold. However, further research Materials and Methods is needed to elucidate temporality and reinforce associ- The present study used a pair-matched, cross-sectional study design with data as- ation between CVD and AP. (J Endod 2016;42:916–920) certained through chart review and examination of dental radiographs. AP was defined radiographically by a periapical radiolucency exceeding twice the width of the normal Key Words periodontal ligament space (13–17), and patients with at least 1 tooth exhibiting Apical periodontitis, cardiovascular disease, endodon- radiographic evidence of apical pathology were designated as AP cases. All tics, root canal treatment, systemic disease measurements were recorded by 1 observer. Radiographic AP appearance has been

From the *Department of Endodontics, New York University Lutheran Dental Medicine, Brooklyn, New York; and †Department of Epidemiology and Health Promo- tion, College of Dentistry, New York University, New York, New York. Address requests for reprints to Dr Gregory K. An, Woodside Endodontics, 1690 Woodside Road, Suite 209, Redwood City, CA 94061. E-mail address: gregan@post. harvard.edu 0099-2399/$ - see front matter Copyright ª 2016 American Association of Endodontists. http://dx.doi.org/10.1016/j.joen.2016.03.011

916 An et al. JOE — Volume 42, Number 6, June 2016 Clinical Research validated and correlated to histologic findings of periapical disease), 420-429 (other forms of heart disease): 420 and 423 (peri- inflammation (18). The non-AP group was composed of persons carditis), 422 and 429 (myocarditis), 421 and 424 (endocarditis), 424 without radiographic evidence of apical pathology; these individuals (valvular heart disease); 430-438 (cerebrovascular diseases): 435 were pair-matched (1:1) to AP cases on age (Æ10 years) and gender. (transient cerebral ischemia), 434 (ischemic stroke); 440-448 (emb- The study was conducted at New York University-Lutheran Med- olisms and thrombosis that fall under diseases of the arteries, arterioles, ical Center (NYU-LMC), a tertiary care teaching hospital and its asso- capillaries), and 450-459 (diseases of veins). Other diseases found in ciated sites in Brooklyn, NY. Electronic medical and dental records the literature to be associated with AP and CVD were also recorded: 272 spanning 56 months from July 7, 2008 to February 28, 2013 were re- (hypercholesterolemia), 401.9 (hypertension), and 250 (diabetes mel- viewed. Patient data were gathered by using Veterans Health Informa- litus II). The observer responsible for recording medical status data was tion Systems and Technology Architecture electronic medical records blinded to AP status. (EMRs) and Dentrix (Henry Schein Practice Solutions, American Fork, Descriptive analysis was completed by using Pearson c2, paired t UT) Enterprise RT 4.0 electronic dental records (EDRs). The hospi- test, and unadjusted conditional logistic regression. AP status was re- tal’s EMRs and EDRs were integrated and shared patient demographic gressed on CVD and selected covariates in a multivariable, conditional information. logistic model. The most parsimonious model was then computed. IBM Patients from EDRs were selected on the basis of random SPSS Statistics 20.0 (Chicago, IL) was used for all analyses. encounter dates within the endodontic and general practice residency practices; dates were randomized by using the Random Calendar Date Generator (19). We included patients who were 30 years of age or older Results during the review period and had no less than 3 encounters recorded in Data from 364 patients were included in the analysis, 182 AP and both the EMRs and EDRs. For study inclusion, EDR charts were required an equal number of age-matched, gender-matched non-AP patients. to have a complete patient examination and treatment plan as well as a Characteristics of patients in the AP and non-AP groups are pre- full mouth set of digital radiographs. Subjects with less than 10 teeth sented in Table 2. Because of the pair-matched design used, the AP present were excluded. During the period reviewed, all radiographs and non-AP groups were similar on age (mean, 49 years) and gender within the hospital system were taken with Schick by Sirona CDR (73% female) distribution. Bivariate analysis revealed statistically signif- 2000 sensors. Sirona digital imaging software, CDR Dicom Version icant, positive relationships between the presence of AP and each of the 3.5 was bridged to the Dentrix software. following: CVD, hypercholesterolemia, race/ethnicity, missing teeth, An a priori power analysis indicated that a sample of 172 individ- number of RCTs, and caries experience. uals in each group (AP and non-AP) would achieve a power of 80% to A subgroup analysis of the AP group (N = 182) in Table 3 demon- detect an odds ratio (OR) of 2.0 or more when the prevalence of expo- strated no statistically significant association between CVD and number sure among the control group was 20%, assuming an alpha level of 0.05 of teeth with AP. However, among the AP group, a statistically significant and using logistic regression. association between CVD and number of teeth with existing RCTs was The independent variables collected from EDRs and EMRs for found. When a multivariable model was used, neither findings were sta- each patient were age of patient at date of observed radiograph(s), tistically significant. gender, race, alcohol use, smoking history, body mass index (BMI) (obesity measure), history of periodontal disease indicated by treatment or treatment plan, number of teeth with existing root canal treat- TABLE 2. Characteristics of Persons in AP and Non-AP Groups ment (RCT), number of missing teeth (calculated by subtracting n (%) number of present teeth from 32), and caries experience indicated P by teeth with existing restorations or caries. Among the AP group, the Variable AP Non-AP value* number of teeth with AP was measured. Gender On the basis of patient-specific EMRs, International Classification Male 49 (26.9) 49 (26.9) — Female 133 (73.1) 133 (73.1) — of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes Race/ethnicity were reviewed to identify individuals with CVD-related systemic diseases White 34 (18.7) 13 (7.1) Referent (12). Table 1 illustrates the range of ICD-9-CM codes used: 414 (cor- Hispanic 113 (62.1) 145 (79.7) .001 onary artery disease), 427 (arrhythmias), 745-746 (congenital heart Asian 14 (7.7) 3 (1.6) .707 Black 13 (7.1) 7 (3.8) .281 defects), 390-398 (rheumatic heart disease), 425 (cardiomyopathy), Other 8 (4.4) 14 (7.7) .013 413 (angina pectoris), 786.50 (unstable angina), 410 (myocardial CVD 58 (31.9) 19 (10.4) <.001 infarction), 411 (ischemic heart disease), 415-417 (pulmonary heart Hypercholesterolemia 89 (48.9) 53 (29.1) <.001 Hypertension 61 (33.5) 51 (28.0) .199 Diabetes mellitus II 34 (18.7) 26 (14.3) .230 TABLE 1. ICD-9-CM Codes Abstracted from EMRs Smoking 21 (11.5) 23 (12.6) .453 Periodontal disease 79 (43.4) 69 (37.9) .276 Diagnosis ICD code Mean (SD) Mean (SD) Coronary heart disease* 410-414, 420-429 Cerebrovascular disease* 430-438 Age (y) 49.1 (12.4) 49.0 (12.3) .896 2 Peripheral arterial disease* 440-448 BMI (kg/m ) 28.7 (5.8) 28.8 (4.7) .834 Rheumatic heart disease* 390-398 Alcohol use (servings 0.7 (2.4) 0.5 (2.4) .606 Congenital heart disease* 745-746 per week) Deep vein thrombosis/pulmonary 451-459 Number of RCTs 2.7 (2.0) 0.6 (1.2) <.001 embolism* Missing teeth 6.8 (5.0) 5.8 (5.0) .020 Hypercholesterolemia 272 Caries experience 12.1 (4.6) 10.4 (5.1) <.001 Hypertension 401.9 Diabetes mellitus II 250 AP, apical periodontitis; CVD, cardiovascular disease; RCT, root canal treatment; SD, standard deviation. Bold font indicates statistical significance (P < .05). *ICD codes classified for CVD identified in italics. *Unadjusted P values based on paired t tests or unadjusted conditional logistic regression.

JOE — Volume 42, Number 6, June 2016 Apical Periodontitis and Cardiovascular Disease 917 Clinical Research

TABLE 3. Subgroup Analysis of AP Group (n = 182) limitations; patients are subject to recall bias and may often be unaware of their current and past dental and medical status (26, 27). Mean (SD) Consequently, inaccuracies may lead to misclassification. A more Variable CVD (n = 58) Non-CVD (n = 124) P value* objective and reliable method of identifying medical disease is the No. of teeth 1.8 (1.2) 1.6 (0.8) .225 use of diagnostic codes such as ICD-9-CM found in EMRs (28).To with AP our knowledge, this is the first study to investigate an AP-CVD associa- No. of RCTs 2.4 (2.2) 1.5 (1.8) .001 tion by reviewing ICD diagnostic coding to establish medical diagnoses CVD, cardiovascular disease; RCT, root canal treatment; SD, standard deviation. and minimize misclassification of CVD and other diagnoses. Bold font indicates statistical significance (P < .05). After adjusting for the potential effects of multiple covariates, our *Unadjusted P values based on paired t tests. model estimated 5.3-fold higher odds of having at least 1 AP among persons with CVD, relative to those without CVD. Our findings reinforce pre- vious findings of an AP-CVD association (24, 25, 29–31); however, they Table 4 shows the results of the full conditional logistic regression contradict statistically non-significant findings by Frisk et al (32),whose model. Statistically significant associations were found between the sample was relatively homogenous, consisting of only Swedish women presence of AP and each of the following: CVD, hypertension, number (n = 1056). In another study of all male subjects, investigators found of RCTs, and race/ethnicity. a significant association between AP and CVD among young men (29). When adjusted for race/ethnicity and hypertension, our final con- Petersen et al (33) also reported that men with AP had a greater risk ditional multivariable logistic regression model showed statistically sig- for CVD. Because age and gender are potential effect modifiers of the nificant positive associations between AP and both CVD (OR, 5.3; 95% AP-CVD relationship, those variables may have played a role in the con- confidence interval [CI], 1.5–18.4), and number of RCTs (OR, 3.4; flicting results across studies. We controlled for confounding by these 95% CI, 2.3–4.9). variables through our pair-matched design and use of conditional logistic regression; however, matching on these variables precluded our ability to Discussion assess the influence of age and gender as effect modifiers. The biological plausibility for the association between periodontal As expected, our final model also showed a significant association disease and CVD has been well-detailed in the literature (6–11).More between AP status and number of RCTs. A systematic review reported a recently, studies have emerged outlining the biological mechanisms and significantly higher proportion of AP in teeth with RCTs than in teeth supporting the plausibility of an AP and CVD relationship (20, 21). without RCTs, 36% and 2%, respectively (34). Our study found that Various studies have demonstrated an independent association between for each existing RCT there was a 3.4-fold increase in having AP, a sig- AP and known inflammatory markers implicated in CVD (20–22). nificant association between these 2 variables. These inflammatory markers include inflammatory cytokines and RCT has commonly been used as a surrogate for AP or pulpal enzymes, as well as C-reactive protein, matrix metalloproteinases, asym- inflammation in studies investigating the link between endodontic- metric dimethylarginine, and reactive oxygen species. Cotti et al (23) related inflammation and CVD (24, 25). However, use of RCT as a reported that compared with non-AP controls, AP subjects had greater proxy has limitations. RCT and AP have several obvious differences; levels of proinflammatory markers and endothelial cell activation, an RCTs could present with or without AP, and if AP is present, it likely indication of acute CVD events and atherosclerosis. Although contribu- indicates a process of healing rather than an indication of ongoing tions have been made to better understand the biological link between inflammation, depending on when the RCT was completed and the AP and CVD, studies of association have been far from conclusive. quality of RCT (35, 36). Most studies have been cross-sectional in Commonly, AP and CVD status are gathered from patient self- design and therefore unable to differentiate healing from non-healing report and/or surveillance data (24, 25). Self-reported data present cases. Also, RCTs may have been performed for reasons other than

TABLE 4. Associations between Independent Variables and AP Variable OR, unadjusted (95% CI)* OR, adjusted (95% CI)† P value, adjusted† CVD‡ 6.57 (2.97–14.55) 5.81 (1.06–31.77) .042 Hypercholesterolemia‡ 2.80 (1.68–4.67) 1.34 (0.36–4.96) .657 Hypertension‡ 0.71 (0.43–1.19) 0.31 (0.10–0.92) .035 Diabetes mellitus II‡ 1.44 (0.79–2.63) 1.12 (0.34–3.63) .853 Smoking‡ (5+ cigarettes/day) 0.89 (0.45–1.74) 1.00 (0.23–4.32) .999 Periodontal disease‡ 1.27 (0.83–1.95) 1.32 (0.55–3.13) .534 BMI 1.00 (0.96–1.04) 1.03 (0.95–1.12) .498 Alcohol 0.98 (0.90–1.07) 0.98 (0.77–1.23) .838 No. of RCTs 3.07 (2.24–4.21) 3.53 (2.28–5.48) <.001 Missing teeth 1.06 (1.01–1.12) 1.10 (0.99–1.22) .090 Caries experience 1.09 (1.03–1.14) 1.00 (0.91–1.09) .940 Race White 1.0, referent 1.0, referent Referent Hispanic 0.26 (0.12–0.58) 0.16 (0.040–0.63) .008 Asian 1.32 (0.31–5.65) 8.09 (0.39–166.64) .175 Black 0.52 (0.16–1.72) 0.24 (0.030–1.67) .148 Other 0.23 (0.07–0.73) 0.03 (0.00–0.36) .005

BMI, body mass index; CI, conﬁdence interval; CVD, cardiovascular disease; OR, odds ratio; RCT, root canal treatment. Non-AP patients were matched to AP cases on age (Æ10 years) and sex. Bold font indicates statistical signiﬁcance (P < .05). *Unadjusted conditional logistic regression. †Adjusted conditional logistic regression. ‡Categorical variable (yes or no).

918 An et al. JOE — Volume 42, Number 6, June 2016 Clinical Research infection and inflammation such as prophylaxis in trauma cases. There- control. Well-controlled diabetics are less likely than uncontrolled di- fore, RCT and AP may have differing effects on CVD. Petersen et al (33) abetics to be at risk for oral infections (44, 45). A sufficient demonstrated that AP teeth without RCT were significantly associated proportion of our diabetic subjects may have been well-controlled, with CVD, whereas AP teeth with RCT were not. Our study showed the thereby contributing to a diluted effect. number of completed RCTs was statistically significant in the final Additional variables that did not show associations with AP were model. The correlation between RCT and AP would likely be higher BMI, periodontal disease, and alcohol use. In agreement with a previ- in our sample compared with findings from other studies because all ous study (46) we also did not find any significant association with AP our subjects were patients with multiple visits with completed examina- and hypertension in our final model. tions; therefore, most patients would have already had AP identified and There are several limitations to this study that warrant discus- consequently treated with RCT. Although it is intuitively obvious that the sion. The World Health Organization’s definition of CVD and its ICD number of RCTs would be a predictor of AP as found in our results, we diagnostic coding was used to identify CVD. The broad definition of also found that the number of RCTs was independently associated with CVD includes conditions beyond those that are infectious and inflam- both AP and CVD. matory in nature. Future studies should investigate the association of We explored the exposure-response relationship between both AP with specific CVD conditions with infectious or inflammatory number of teeth with AP and CVD and number of teeth with existing mechanisms. RCTs and CVD. The subgroup analysis of subjects with AP (n = 182) Certain characteristics of patients seen at NYU-LMC may have demonstrated that those with CVD were statistically more likely to differed considerably from the population at large. For instance, there have a greater number of teeth with existing RCTs than those without could be differences in culture, socioeconomic status, behavioral char- CVD. Subjects with CVD also had greater number of teeth with AP acteristics, health perceptions and literacy, as well as overall health sta- than those without CVD, but the association was not statistically signif- tus. More than 70% of our subjects were Hispanic, and in our sample, icant. The null finding may partially be explained by the underreporting we found that Hispanic patients and patients of ‘‘other’’ race were less of true AP in our sample. For instance, teeth with AP may have already likely than whites to have AP. Also, most NYU-LMC patients were of low been treated at time of record. Also, the method of radiographic detec- socioeconomic status, and nearly all were on Medicaid or subsidized tion of AP excludes inflamed teeth with AP not detectable by radio- free care. Many spoke little to no English and often used interpretive graphs. Last, our sample size was reduced by half to conduct the services. Furthermore, patients like those in our sample seeking hospi- subanalysis because all subjects without AP, the non-AP group, were tal care were likely to have more illness with greater morbidity. There- excluded. A larger sample may yield more conclusive results. Therefore, fore, generalizability of our findings to larger diverse populations or the exposure-response relationship should be further explored. other races may be limited. Further investigation into the influences Caries, the primary cause of endodontic infections, and missing of these variables including race and ethnicity is warranted. teeth have been shown to be associated with AP (37, 38).In Another limitation was the use of two-dimensional (2D) digital ra- agreement, we also found caries and missing teeth to be significantly diographs for AP identification. Interpretation of 2D radiographs can associated with AP in our unadjusted descriptive analysis. However, lead to underdiagnosis (47). Unlike chronic AP, most acute AP is radio- when both variables were included in our multivariable conditional graphically undetectable and unrecorded. Also, AP must reach a suffi- logistic regression, neither variable was significantly associated with cient threshold of bone destruction to be detectable; a classic study by AP, possibly because of multicollinearity because missing teeth and Bender (48) demonstrated that at least 30% mineral loss must occur carious teeth have statistically significant correlations with the before radiographic detection. Furthermore, the size of radiolucencies number of RCTs. on a periapical film is smaller than the actual size of bone destruction, Smoking has been found to have a strong association with AP (39). resulting in greater difficulty in identifying smaller AP (49). The lack of Surprisingly, we found no significant association in our sample. detection of acute AP cases and cases with radiolucencies near or below Possible explanations could stem from differences in reporting method- the detectable threshold leads to underestimation of the true AP preva- ology, definition of smoking, and culture. In EMRs, smoking and lence. Underdiagnosis may lead to an underestimation of the effect; alcohol were self-reported and, along with BMI, were the only medical therefore, the true association of AP and CVD may actually be greater. variables not assigned ICD-9-CM codes. Smoking in EMRs was recorded A review by Patel et al (50) showed that cone-beam computed tomog- as daily smoking, specifically smoking more than 5 cigarettes per day raphy (CBCT) was far superior in detecting true AP than 2D digital ra- (yes/no). In contrast, some studies define smoking as having ever diographs. However, CBCT should be used with caution for AP smoked. Variation in the definition of smoking, especially in the identification. Pope et al (51) demonstrated that the use of CBCT ill-defined ‘‘light’’ or ‘‘intermittent’’ smoking groups, may lead to showed a wide variation of periodontal ligament space size in normal misclassification (40). Also, self-reporting of smoking varies according healthy pulps. CBCT has strong sensitivity, the ability to correctly diag- to cultural, psychological, and cognitive factors of individuals (41). The nose AP, but relatively weak specificity, the ability to correctly diagnose study by Lopez-Lopez et al (39) was conducted in Spain where arguably normal periapical tissues. Therefore, there may be greater potential for smoking is more socially accepted than in the United States. Social misclassification of healthy pulps. stigma around smoking may have contributed to underreporting of The social and economic burden of CVD is high. According to the smoking in our study. Random misclassification and underreporting American Heart Association, the prevalence of CVD in the United States would have favored a null finding such as ours. is more than 36%, with total annual direct costs of more than $200 An association between diabetes and AP has been shown in previ- billion (52). A review of multiple community-based studies showed a ous studies (42, 43); however, our study did not find a significant wide range of AP prevalence from 14% to 70% of participants (53). association. The statistically non-significant finding may be explained As tooth retention increases, it is possible that the burden of AP will in- by our method of classifying diabetes; we used diagnostic codes rather crease, and consequently, CVD will increase. A systematic review of AP than directly measuring glycemic control through hemoglobin A1c or prevalence showed approximately 5% of all teeth had AP, or broadly 1 blood glucose levels. All subjects in ours study were repeat patients AP per patient (34). In view of the growing epidemic of CVD and the of at least 3 past visits; therefore, subjects diagnosed with diabetes high rates of AP, further studies should investigate possible causal re- were more likely to have undergone treatment and be under glycemic lationships through prospective interventions.

JOE — Volume 42, Number 6, June 2016 Apical Periodontitis and Cardiovascular Disease 919 Clinical Research

Conclusion 23. Cotti E, Dessi C, Piras A, et al. Association of endodontic infection with detection of an initial lesion to the cardiovascular system. J Endod 2011;37:1624–9. Within the limits of this investigation, we found that patients pre- 24. Caplan DJ, Pankow JS, Offenbacher S, Beck JD. The relationship between self- senting with AP were 5.3 times more likely to have CVD than patients reported history of endodontic therapy and coronary heart disease in the ARICS. without AP. Future studies should explore possible relationships be- J Am Dent Assoc 2009;140:1004–12. tween AP and specific CVD conditions, especially those with infectious 25. Joshipura KJ, Pitiphat W, Hung HC, et al. Pulpal inflammation and incidence of coronary heart disease. J Endod 2006;32:99–103. or inflammatory mechanisms. Also, methods should be used to minimize 26. Peacock ME, Carson RE. Frequency of self-reported medical condition in peri- misclassification of AP, CVD, and their risk factors. There is a need for odontal patients. J Periodontol 1995;66:1004–7. well-powered randomized studies to clarify temporality and replicate a 27. Thompson KS, Yonke ML, Rapley JW, et al. Relationship between a self-reported definitive association between specific CVD conditions and AP. health questionnaire and laboratory tests at initial office visits. J Periodontol 1999;70:1153–7. 28. Smith B, Chu LK, Smith TC, et al. Challenges of self-reported medical conditions and Acknowledgments electronic medical records among members of a large military cohort. BMC Med The authors deny any conflicts of interest related to this study. Res Methodol 2008;8:37. 29. Caplan DJ, Chasen JB, Krall EA, et al. Lesions of endodontic origin and risk of coronary heart disease. J Dent Res 2006;85:996–1000. References 30. Murray CA, Saunders WP. Root canal treatment and general health: a review of the 1. Miller W. The human mouth as a focus of infection. Dental Cosmos 1891;33:689–713. literature. Int Endod J 2000;33:1–18. 2. Billings F. Focal infection: its broader application in the etiology of disease. JAMA 31. Costa TH, de Figueiredo Neto JA, de Oliveira AE, et al. Association between chronic 1914;63:899–903. apical periodontitis and coronary artery disease. J Endod 2014;40:164–7. 3. Pallasch TJ, Wahl MJ. Focal infection: new age or ancient history? Endod Top 2003; 32. Frisk F, Hakeberrg M, Ahqwist M, Bengtsson C. Endodontic variables and coronary 4:32–45. heart disease. Acta Odontol Scand 2003;61:257–62. 4. Mattila KJ, Nieminen MS, Valtonen VV, et al. Association between dental health and 33. Petersen J, Glabl EM, Nasseri P, et al. The association of chronic apical periodontitis acute myocardial infarction. BMJ 1989;298:779–81. and endodontic therapy with atherosclerosis. Clin Oral Invest 2014;18:1813–23. 5. Lockhart PB, Bolger AF, Papapanou PN, et al. Periodontal disease and atheroscle- 34. Pak JG, Fayazi S, White SN. Prevalence of periapical radiolucency and root canal rotic vascular disease: does the evidence support an independent association? A sci- treatment: a systematic review of cross-sectional studies. J Endod 2012;38: entific statement from the American Heart Association. Circulation 2012;125: 1170–6. 2520–44. 35. De Chevigny C, Dao TT, Basrani BR, et al. Treatment outcome in endodontics: the 6. Willershausen B, Willershausen I, Zahorka D, et al. Association between chronic Toronto study—phase 4: initial treatment. J Endod 2008;34:258–63. dental infection and acute myocardial infection. J Endod 2009;35:626–30. 36. De Chevigny C, Dao TT, Basrani BR, et al. Treatment outcome in endodontics: the 7. Kebschull M, Demmer RT, Papapanou PN. ‘‘Gum bug, leave my heart alone!’’: epide- Toronto study—phases 2 and 4: orthograde retreatment. J Endod 2008;34:131–7. miologic and mechanistic evidence linking periodontal infections and atheroscle- 37. Frisk F, Hakeberg M. Socio-economic risk indicators for apical periodontitis. Acta rosis. J Dent Res 2010;89:879–902. Odontol Scand 2006;64:123–8. 8. Humphrey LL, Fu R, Buckley DI, et al. Periodontal disease and coronary heart dis- 38. Zhong Y, Garcia R, Kaye EK, et al. Association of endodontic involvement with tooth ease incidence: a systematic review and meta-analysis. J Gen Intern Med 2008;23: loss in the veterans affairs dental longitudinal study. J Endod 2010;36:1943–9. 2079–86. 39. Lopez-Lopez J, Jane-Salas E, Martın-Gonzalez J, et al. Tobacco smoking and radio- 9. Genco R, Offenbacher S, Beck J. Periodontal disease and cardiovascular disease: graphic periapical status: a retrospective case-control study. J Endod 2012;38:584–8. epidemiology and possible mechanisms. J Am Dent Assoc 2002;133:14S–22. 40. Ryan H, Trosclair A, Gfroerer J. Adult current smoking: differences in definitions 10. Ahmed GM, El-Baz AA, Hashem AA, Shalaan AK. Expression levels of matrix and prevalence estimates—NHIS and NSDUH, 2008. J Environ Public Health metalloproteinase-9 and gram-negative bacteria in symptomatic and asymptomatic 2012;2012:918368. http://dx.doi.org/10.1155/2012/918368. Epub 2012 May 9. periapical lesions. J Endod 2013;39:444–8. 41. Scott DA, Palmer RM, Stapleton JA. Validation of smoking status in clinical research 11. Noiri Y, Li L, Ebisu S. The localization of periodontal-disease-associated bacteria in into inflammatory periodontal disease. J Clin Periodontol 2001;28:715–22. human periodontal pockets. J Dent Res 2001;80:1930–4. 42. Segura-Egea JJ, Castellanos-Cosano L, Machuca G, et al. Diabetes mellitus, periapical 12. Cardiovascular diseases (CVDs). Fact sheet No. 317. Updated January 2015. World inflammation and endodontic treatment outcome. Med Oral Patol Oral Cir Bucal Health Organization. Available at: http://www.who.int/mediacentre/factsheets/fs317/ 2012;17:e356–61. en/. Accessed September 28, 2015. 43. Lopez-Lopez J, Jane-Salas E, Estrugo-Devesa A, et al. Periapical and endodontic sta- 13. Lofthag-Hansen S, Huumonen S, Grondahl K, Grondahl HG. Limited cone-beam CT tus of type 2 diabetic patients in Catalonia, Spain: a cross-sectional study. J Endod and intraoral radiography for the diagnosis of periapical pathology. Oral Surg Oral 2011;37:598–601. Med Oral Pathol Oral Radiol Endod 2007;103:114–9. 44. Bender IB, Seltzer S, Freedland J. The relationship of systemic diseases to endodon- 14. Low MTL, Dula KD, Burgin W, von Arx T. Comparison of periapical radiography and tic failures and treatment procedures. Oral Surg Oral Med Oral Pathol 1963;16: limited cone-beam tomography in posterior maxillary teeth referred for apical sur- 1102–5. gery. J Endod 2008;34:557–62. 45. Measley BL, Oates TW. Diabetes mellitus and periodontal diseases. J Periodontol 15. Bornstein MM, Lauber R, Sendi P, von Arx T. Comparison of periapical radiography 2006;77:1289–303. and limited cone-beam computed tomography in mandibular molars for analysis of 46. Segura-Egea JJ, Jimenez-Moreno E, Calvo-Monroy C, et al. Hypertension and dental anatomical landmarks before apical surgery. J Endod 2011;37:151–7. periapical condition. J Endod 2010;36:1800–4. 16. Patel S, Wilson R, Dawood A, Mannocci F. The detection of periapical pathosis using 47. Shoha RR, Dowson J, Richards AG. Radiographic interpretation of experimentally periapical radiography and cone beam computed tomography: part 1—pre-oper- produced bone lesions. Oral Surg Oral Med Oral Pathol Oral Radiol Endod ative status. Int Endod J 2012;45:702–10. 1974;38:294–303. 17. Patel S, Wilson R, Dawood F, et al. The detection of periapical pathosis using digital 48. Bender IB. Factors influencing the radiographic appearance of bony lesions. periapical radiography and cone beam computed tomography: part 2—a 1-year J Endod 1997;23:5–14. post-treatment follow-up. Int Endod J 2012;45:711–23. 49. Tanomaru-Filho M, Jorge EG, Duarte MAH, et al. Comparative radiographic and his- 18. Ørstavik D, Kerekes K, Eriksen HM. The periapical index: a scoring system for radio- tological analyses of periapical lesion development. Oral Surg Oral Med Oral Pathol graphic assessment of apical periodontitis. Endod Dent Traumatol 1986;2:20–34. Oral Radiol Endod 2009;107:442–7. 19. Random Date Generator. Random.org. Available at: https://www.random.org/ 50. Patel S, Dawood A, Pitt Ford TP, Whaites E. The potential applications of cone beam calendar-dates/. Accessed September 28, 2015. computed tomography in the management of endodontic problems. Int Endod J 20. Gomes MS, Blattner TC, Filho MS, et al. Can apical periodontitis modify systemic 2007;40:818–30. levels of inflammatory markers? a systematic review and meta-analysis. J Endod 51. Pope O, Sathorn C, Parashos P. A comparative investigation of cone-beam computed 2013;39:1205–17. tomography and periapical radiography in the diagnosis of a healthy periapex. 21. Inchingolo F, Marrelli M, Annibali S, et al. Influence of endodontic treatment on J Endod 2014;40:360–5. systemic oxidative stress. Int J Med Sci 2014;11:1–6. 52. Roger VL, Go AS, Lloyd-Jones DM, et al. Executive summary: heart disease and 22. Warner RL, Bhagavathula N, Nerusu KC, et al. Matrix metalloproteinases in acute stroke statistics—2012 update: a report from the American Heart Association. Cir- inflammation: induction of MMP-3 and MMP-9 in fibroblasts and epithelial cells culation 2012;125:188–97. following exposure to pro-inflammatory mediators in vitro. Exp Mol Pathol 53. Caplan DJ. Epidemiologic issues in studies of association between apical periodon- 2004;76:189–95. titis and systemic health. Endod Top 2004;8:15–35.

920 An et al. JOE — Volume 42, Number 6, June 2016