Medical doctors recorded medical and smoking history. TA and BA were diagnosed by electrocardiogram. Subjects who were diagnosed and/or treated as diabetes mellitus (DM), hypertension (HT) and dyslipidemia (DL) were recorded. Peripheral blood samples were obtained and level of C-reactive protein (CRP) was determined.

Trained periodontists counted the number of residual teeth. Probing pocket depth (PPD), clinical attachment level (CAL) and bleeding on probing (BOP) at six points per tooth (buccal-mesial, mid-buccal, buccal-distal, lingual-mesial, mid-lingual and lingual-distal) on an upper right molar, an upper incisor, an upper left molar, a lower right molar, a lower incisor and a lower left molar were measured with a manual probe (PCP-UNC 15, Hu-Friedy, Chicago, IL, USA). When the representative tooth was missing, next tooth was used.

We obtained unstimulated saliva and used DNeasy Blood and Tissue kit (Qiagen, Tokyo, Japan) according to manufacturer’s instructions to extract bacterial DNA. Real-time polymerase chain reaction (PCR) was used to detect three major periodontal bacteria, Porphyromonas gingivalis (P. gingivalis), Aggregatibacter actinomycetemcomitans (A. actinomycetemcomitans) and Prevotella intermedia (P. intermedia). The real-time PCR was performed and specific primers for each bacterium were used as previously described [12].

Numerical data was presented as means ± standard deviation (SD). Student’s t-test was used to compare numerical values such as age, CRP, teeth number, PPD, CAL and BOP. Chi-square test was performed to compare dichotomous variables such as sex, smoker rate, positive rates of DM, HT and DL and bacterial detection rates. Wilcoxon test was used to compare bacterial counts, because bacterial counts were not normally distributed. We used JMP 9.0.3 (SAS Institute Inc., Cary, NC, USA) for all statistical analyses and considered values of p < 0.05 significant.

The characteristics of the subjects in the present study are shown in Table 1. There was no statistical difference of age, sex, smoking rate, HT, DL and CRP between the groups. Prevalence of DM was higher in the TA group compared with the BA group in 81–90 year-old patients, but not in 71–80 year-old patients.

The number of missing teeth was comparable between the two arrhythmia groups of the same age categories (Table 2). There was no statistical difference of mean PPD, CAL and BOP rate between the groups.

The counts of P. intermedia in the TA group increased in comparison to the BA group in 71–80 year-old patients (Table 3). The TA group also had an increased amount of P. gingivalis compared to the BA group in 81–90 year-old patients (Table 3). The counts of A. actinomycetemcomitans were comparable between the two groups. We also showed a positive rate of P. intermedia (Table 4). The TA group had an increased detection rate of compared to the BA group in 71–80 year-old patients.

It is well known that there is a relationship between periodontitis and CVD. However, there are few papers to show the association between arrhythmia and periodontitis. It is well recognized that ischemic stroke is frequently induced by TA. Clinically, a recent case-control study showed that high levels of periodontal clinical attachment loss, gingivitis and radiographic bone loss were independently associated with stroke, while the etiology of the stroke was not indicated [13]. On the other hand, Holm-Pedersen et al. demonstrated that there might be a link between active root caries and cardiac arrhythmias in elderly persons, while there was no association between periodontal disease and arrhythmia in the study [14]. In an animal model, Yu et al. showed the effect of periodontitis on susceptibility to atrial fibrillation, which is a major TA. Periodontitis was induced by tying 2–0 silk ligatures at the second premolar of mandibula in adult mongrel canines. Electrophysiologic evaluation was performed to assess atrial refractoriness and atrial fibrillation inducibility. They found that periodontitis induced inflammatory responses in atrial myocardium, which disturbed the structural and electrophysiologic properties of the atrium and facilitated atrial fibrillation [15]. Therefore, periodontitis may deteriorate TA such as atrial fibrillation.

There are a limited number of reports to evaluate the influence of systemic periodontal pathogen infections on TA and/or stroke. Hosomi et al. demonstrated a relationship between serum antibodies against periodontal pathogens and ischemic stroke. They evaluated patients with acute ischemic stroke and patients without previous stroke or stroke subtype. The results showed that the serum-antibody level of P. intermedia was significantly higher in athero-thrombotic stroke patients than in patients with no previous stroke. Serum antibody against P. gingivalis was also significantly associated with atrial fibrillation. They concluded that infection with specific periodontal bacteria was associated with stroke and atrial fibrillation [16]. Pussinen et al. also showed serological evidence that a chronic infection caused by P. gingivalis and A. actinomycetemcomitans was associated with stroke incidence in prospective case-control studies [17, 18], while the etiology of the stroke was not clarified. These previous studies suggested that periodontal pathogens such as P. gingivalis and P. intermedia may have some crucial effects on stroke and/or TA. The mechanisms connecting periodontal infection and TA are not clear, but long-term systemic exposure to periodontal pathogens may influence cardiac homeostasis.