A total of 272 patients with T2DM, all of Chinese Han ethnicity, who were hospitalized at the Department of Endocrinology, Shanghai East Hospital Affiliated to Tongji University, from November 2020 to December 2021 were included in our study. Patients were diagnosed with T2DM according to the 1999 World Health Organization (WHO) criteria. We excluded patients with type 1 diabetes mellitus or specific types of diabetes mellitus, acute complications of diabetes, liver or renal dysfunction, serious cardiac arrhythmias or acute cardiac insufficiency, diabetic foot ulcers, or history of acute myocardium infarction, percutaneous coronary intervention, cerebral infarction, tumor, or psychosis. We obtained written informed consent from all of the participants. The study was approved by the Human Research and Ethics Committee of Shanghai East Hospital and adhered to the tenets of the Declaration of Helsinki.

Participants’ demographic details, frequency of dental visits, education level, and medical history were carefully recorded. Medical histories mainly covered age, sex, smoking habits, current medications, body mass index (BMI), CVD, cerebral apoplexy, peripheral arterial disease (PAD), hypertension, and duration of diabetes. Patients’ medical details, including presence of hypertension, hyperlipidemia, and diabetes mellitus, were obtained from patient records. BMI was calculated as weight (kg) divided by the square of height (m). In terms of education level, patients were categorized as unschooled or as having attended secondary school, high school, or university. Diabetes of less than 10 years’ duration was regarded as short-duration diabetes, and 10 years’ duration or more as long-duration diabetes in this study. Additionally, we divided patients into two age groups: an older group (more than 60 years old) and a younger group (60 years old or younger).

All the included patients underwent a CAC CT scan, which was specified in the informed consent materials. The CAC scan of the heart was rapidly acquired, prospectively electrocardiogram-triggered and without contrast. The CAC score was quantified using the Agatston method, in which the area of calcified atherosclerosis (defined as an area of at least 1 mm² with a CT density >130 Hounsfield units [HU]) is multiplied by a density weighting factor and summed for the entire coronary artery tree using a CT dataset with 2.5 to 3.0 mm slice thickness. According to the recommendations of related guidelines, a CAC score of 0 is considered to represent no identifiable plaque burden; a CAC score of 1–10 to represent a minimal identifiable plaque burden; a CAC score of 11–100 to represent a mild atherosclerotic plaque burden; a CAC score of 101–400 to represent a moderate atherosclerotic plaque burden; and a CAC score of more than 400 to represent an extensive atherosclerotic plaque burden. Although a negative or extremely low calcium score (10 or lower) cannot be taken to fully exclude the presence of coronary atherosclerosis, it is consonant with the absence of a fixed (significant) coronary obstructive lesion (likelihood, 5 to 10% or lower) (35, 36). Since a CAC score of less than 100 indicates a low probability of coronary artery disease and coronary events (35–37), we chose 100 as cut-off point. Thus, all patients were assigned to one of two groups according to CAC score: a normal-to-mild group (0 ≤ score ≤ 100), or a moderate-to-severe group (≥101) (34).

Two trained examiners performed all oral examinations. All patients were examined with a dental explorer, mouth mirror, and (Williams and WHO) periodontal probe. They all underwent a clinical periodontal test and a panoramic radiograph was taken. The gingival index (GI) was measured according to the Silness and Loe method. Additionally, bleeding on probing (BOP), clinical attachment loss (CAL), calculus index (CI), probing depth (PD), and periodontal index (PI) were measured according to O’Leary’s index around the six Rumford index teeth (38). Periodontitis severity staging (39) was calculated accordingly: for each tooth, the CAL of the most severe site was recorded; a CAL of 1–2 mm (with no tooth loss due to periodontitis) was defined as Stage I, a CAL of 3–4 mm (with no tooth loss due to periodontitis) as Stage II (moderate), a CAL of ≥ 5 mm (with loss of ≤ 4 teeth due to periodontitis) as Stage III, and a CAL of ≥ 5 mm (with loss of ≥ 5 teeth due to periodontitis) as Stage IV (severe). All patients with periodontitis were categorized into three groups (40): a mild periodontitis group (Stage I), a moderate periodontitis group (Stage II), and a severe periodontitis group (Stage III-IV). Tooth loss was assessed via panoramic radiography readings. Extractions of third molars and extractions performed for orthodontic treatment were not considered to represent tooth loss, whereas extractions for the purpose of dental implant treatment were considered to represent tooth loss. All radiography readings and assessments were carried out by one dentist. Tooth loss was categorized as 0 to 9, 10 to 19, or 20 to 31 missing teeth (41).

Saliva and blood samples were obtained on the same day at the time of the patient’s first visit. After overnight fasting, venous blood samples were collected from all patients. Unstimulated whole buccal saliva was collected from individuals as previously described (42). Briefly, saliva was collected during the course of 3 min after an overnight fast and without prior oral hygiene measures. Saliva samples were immediately frozen at -80°C and stored for less than 3 months before biochemical analysis.

Levels of total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglycerides (TGs), fasting plasma glucose (FPG), glycosylated hemoglobin (HbA1c), and serum creatinine were measured using suitable laboratory techniques. ELISA assay kits supplied by Abcam (Cambridge, MA, USA) were used on saliva samples stored at −80°C to determine interleukin-6 (IL-6), matrix metalloproteinase-9 (MMP-9), and TNF-α levels.

Microbial DNA was extracted from saliva samples using the E.Z.N.A.® Soil DNA Kit (Omega Bio-tek, Norcross, GA, U.S.) according to the manufacturer’s protocols. The V1–V9 regions of the bacteria 16S ribosomal RNA gene were amplified by PCR (95°C for 2 min, followed by 27 cycles at 95°C for 30 s, 55°C for 30 s, and 72°C for 60 s and a final extension at 72°C for 5 min) using primers 27F 5′-AGRGTTYGATYMTGGCTCAG-3′ and 1492R 5′-RGYTACCTTGTTACGACTT-3′, where the barcode represents an eight-base sequence unique to each sample. PCR reactions were performed in triplicate using a 20 μL mixture containing 4 μL of 5 × FastPfu Buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of each primer (5 μM), 0.4 μL of FastPfu Polymerase, and 10 ng of template DNA. Amplicons were extracted from 2% agarose gels and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, U.S.) according to the manufacturer’s instructions.

SMRTbell libraries were prepared from the amplified DNA by blunt-ligation according to the manufacturer’s instructions (Pacific Biosciences). Purified SMRTbell libraries from the Zymo and HMP mock communities were sequenced on dedicated PacBio Sequel II 8M cells using the Sequencing Kit 2.0 chemistry. Purified SMRTbell libraries from the pooled and barcoded samples were sequenced on a single PacBio Sequel II cell. All amplicon sequencing was performed by Shanghai Biozeron Biotechnology Co. Ltd (Shanghai, China).

PacBio raw reads were processed using the SMRT Link Analysis software package, version 9.0, to obtain demultiplexed circular consensus sequence (CCS) reads with the following settings: minimum number of passes = 3, minimum predicted accuracy = 0.99. Raw reads were processed through SMRT Portal to filter sequences for length (<800 or >2500 bp) and quality. Sequences were further filtered by removing the barcode, primer sequences, chimeras, and sequences containing 10 consecutive identical bases. OTUs were clustered with 98.65% similarity cut-off using UPARSE (version 7.1; http://drive5.com/uparse/), and chimeric sequences were identified and removed using UCHIME. The phylogenetic affiliation of each 16S rRNA gene sequence was analyzed using RDP Classifier (http://rdp.cme.msu.edu/) against the silva (SSU132)16S rRNA database, with a confidence threshold of 70% (43). A rarefaction analysis based on Mothur v.1.21.1 (44) was conducted to reveal the diversity indices, including the Richness and Shannon diversity indices. A Mann–Whitney U test was performed to assess the statistical significance of differences in diversity indices between samples. Differences were considered significant at p < 0.05. For identification of biomarkers for highly dimensional colonic bacteria, LEfSe (linear discriminant analysis effect size) analysis was conducted (45). The Kruskal–Wallis rank-sum test was performed to examine the changes and dissimilarities among classes, followed by LDA analysis to determine the effect size for each of the distinctively abundant taxa (46).

SPSS 21.0 (IBM Corp., Armonk, NY, USA) was used for statistical analysis. Data are expressed in the form mean ± standard deviation (SD) for continuous variables and percentages (%) for categorical variables. Between-group differences were analyzed using a two-tailed Student’s t-test or non-parametric t-test. Categorical variables were compared using the chi-squared test or Fisher’s exact test. The association between the severity of periodontal disease or missing teeth and the presence of moderate-to-severe coronary atherosclerosis was analyzed via multivariate logistic regression, adjusted for main effects of all covariates, including age, sex, education level, smoking, LDL, HbA1c, BMI, duration of diabetes, and history of statin treatment. All p-values were two-tailed, and p < 0.05 was considered statistically significant.

184 patients (92 male and 92 female; mean age, 63 ± 10 years) were allocated to the normal-to-mild group, and 88 (54 male and 34 female; mean age, 65 ± 7 years) to the moderate-to-severe group. Demographical, medical, and biochemical data, including medications, dental habits, and education status, are presented in Table 1 . Patients in the moderate-to-severe group were older than those in the normal-to-mild group (P <0.001), and male subjects constituted the majority in the moderate-to-severe group (P <0.001). More subjects in the moderate-to-severe group were current smokers (P<0.01). The duration of diabetes was longer in the moderate-to-severe group than in the normal-to-mild group (P<0.001). However, there was no statistically significant difference in HbA1c levels between the two groups. Lower TC and LDL-C levels unexpectedly occurred in the moderate-to-severe group, with the differences compared to the normal-to-mild group being statistically significant (P<0.001). It is noteworthy that the rate of statin use at baseline was also higher in the moderate-to-severe group, which may be the reason for the lower TC and LDL-C levels. Regarding the salivary markers measured, IL-6 and TNF-α levels were significantly higher in the moderate-to-severe group than in the normal-to-mild group, whereas salivary MMP-9 levels did not differ significantly between the groups.

The periodontitis parameters of the subjects are shown in Table 2 . CAL values were much higher among patients in the moderate-to-severe coronary atherosclerosis group (P=0.05). Additionally, the results revealed that individuals in the moderate-to-severe group had more tooth loss compared to patients in the normal-to-mild group (P=0.016). In analyzing patients’ tooth counts, we observed that more subjects in the normal-to-mild group had more than 25 teeth and 11–24 teeth (P =0.0001, P =0.001, respectively). Severe periodontitis occurred more frequently among patients in the moderate-to-severe coronary atherosclerosis group, and this increase was statistically stronger in male patients (P=0.005). Among female patients, there was an increase in the proportion of moderate periodontitis among patients with moderate-to-severe coronary atherosclerosis, with the difference being statistically significant (P=0.027).

The results of univariate logistic regression analyses, with ORs and CIs, are listed in Table 3 (model 1). CAL (P=0.021) and having 10–19 missing teeth (p=0.01) were found to be highly associated with the presence of moderate-to-severe coronary atherosclerosis. The ORs calculated for CAL and for 10–19 missing teeth were 1.536 (95% CI: 1.066–1.214) and 4.106 (95% CI: 1.404–12.009), respectively. The results of multivariate logistic regression analysis are also presented in Table 3 (model 2). Adjusting for parameters such as age, sex, education level, smoking, LDL, HbA1c, BMI, duration of diabetes, and history of statin treatment in the multivariate logistic regression, we again observed a correlation with moderate-to-severe coronary atherosclerosis among patients with T2DM for CAL (OR: 1.231; 95% CI: 1.004–1.516, P = 0.047) and for having 10-19 missing teeth (OR: 1.604; 95% CI: 1.393–6.555, P=0.05). Unexpectedly, severity staging of periodontitis was not significantly associated with the presence of moderate-to-severe diabetic coronary atherosclerosis in either the unadjusted or the adjusted model.

When we compared the parameters of periodontitis between patients with different durations of diabetes, we found that CAL values were higher in patients with long-duration diabetes. However, when patients with short-duration diabetes and moderate-to-severe coronary AS were compared to patients with long-duration diabetes and normal-to-mild coronary AS, the inter-group difference in CAL values disappeared ( Table 4 ). Additionally, the results revealed that, among patients with short-duration diabetes, individuals in the moderate-to-severe group had more tooth loss compared to those in the normal-to-mild group. However, there was no relationship between tooth loss and extent of coronary AS among patients with long-duration diabetes ( Table 5 ).

The results showed that older patients had higher CAL values than younger patients. However, among younger patients with moderate-to-severe coronary AS, all the parameters of periodontitis and tooth loss were higher than in older patients with normal-to-mild coronary AS, with no statistically significant differences ( Table 6 ). Evaluating the results in younger patients with varying degrees of coronary AS, we observed that individuals with moderate-to-severe coronary AS had higher CAL and GI values, as well as more tooth loss, compared to individuals with normal-to-mild coronary AS. However, we did not observe this difference between the normal-to-mild and the moderate-to-severe group among older patients ( Table 7 ).

The CAC scores of patients in different severity staging groups for periodontitis are presented in Figure 1 . CAC scores were higher in the severe group (436.17 ± 162.424) than in the mild group (122.3 ± 69.394), with the difference reaching statistical significance (P=0.043).

Alpha diversity indices indicated significant differences in Shannon index and Richness between the two coronary AS groups (p = 0.037 and p = 0.05, respectively), indicating higher abundance and greater diversity in the normal-to-mild group ( Figures 2A, B ).

The predominant taxa identified in the dental plaque samples of oral microbiota from the normal-to-mild and moderate-to-severe coronary AS groups at different taxonomic levels are shown in Figure 3 . At the phylum level, Firmicutes, Proteobacteria, Bacteroidetes, and Actinobacteria accounted for the majority of the total sequences ( Figure 3A ). Levels of Actinobacteria were significantly lower in the moderate-to-severe coronary AS group (p =0.045, Mann–Whitney U test). At the class level, Bacteroidia, Betaproteobacteria, Bacilli, Clostridia, Alphaproteobacteria, Negativicutes, Chitinophagia, Gammaproteobacteria, Actinomycetia, and Fusobacteriia were the ten most common ( Figure 3B ), among which, levels of Actinomycetia were consistently more abundant in the normal-to-mild coronary AS group (p =0.035, Mann–Whitney U test). At the order level, Bacteroidales, Lactobacillales, and Neisseriales were predominant ( Figure 3C ). At the family level, the predominant families were Neisseriaceae, Streptococcaceae, Prevotellaceae, Chitinophagaceae, Veillonellaceae, and Porphyromonadaceae ( Figure 3D ), among which, levels of Porphyromonadaceae were more abundant in the moderate-to-severe coronary AS group (p =0.016, Mann–Whitney U test). At the genus level, Neisseria, Streptococcus, Prevotella, Veillonella, and Porphyromonas were predominant ( Figure 3E ). Porphyromonas levels were significantly higher in the moderate-to-severe coronary AS group (p =0.017, Mann–Whitney U test). Finally, at the species level, the predominant species were Neisseria subflava, Sediminibacterium magnilacihabitans, Streptococcus mitis_AV, Pseudoclostridium thermosuccinogenes, Veillonella dispar, and Porphyromonas gingivalis ( Figure 3F ). Among these, the abundance of Porphyromonas gingivalis differed significantly between the normal-to-mild and the moderate-to-severe coronary AS groups (p =0.042, Mann–Whitney U test). These results indicated distinctive differences in the bacterial communities of the groups with different degrees of coronary AS.

To identify the specific bacterial taxa of oral microbiota that were associated with the moderate-to-severe group, we compared the microbial composition of the two groups using the LEfSe method. A cladogram representative of the structure of the oral microbiota and the predominant bacteria is shown in Figure 4A , with indications of the largest differences in the taxa represented in the two communities. The levels of members of bacterial taxa belonging to Porphyromonadaceae, Porphyromonas, Porphyromonas gingivalis, and Neisseria_mucosa were increased in the moderate-to-severe coronary AS group, and those of Actinomycetia, Rothia_dentocariosa, and Blautia coccoides were enriched in the normal-to-mild group; thus, these could be used as biomarkers for discrimination between the groups. The LDA bar graph (log 10) for discrimination is presented in Figure 4B .