Bi-directional association between type 2 diabetes and periodontitis in the nationally representative sample comprising Taiwanese adults from 2000 to 2015

doi:10.21203/rs.3.rs-2054852/v1

Objective

The two-way relationship between periodontitis and diabetes was advocated far past. However, the population-based bidirectional observation is still limited. Using the National Health Insurance Research Database of Taiwan, we estimate associations between pre-existing clinical periodontitis and incident type 2 diabetes mellitus (T2DM), and between pre-existing diabetes and incident periodontitis, respectively.

Methods

A total of 11,011 periodontitis patients who underwent periodontal therapy (Pt group) were recruited from 2000 to 2015. After matching by age, sex, and index date, 11,011 periodontitis patients who did not receive treatment (Pwt group) and 11,011 non-periodontitis controls were registered. The outcome of T2DM was traced. Conversely, 157,798 patients with T2DM and 157,798 non-diabetic controls were enrolled, whereas the development of periodontitis was traced. Cox proportional hazards model was performed.

Results

The patients with periodontitis tended to have a statistically high risk of having diabetes. The adjusted hazard ratio (aHR) was 1.94 [95% CI 1.49–2.63, p < 0.01] and 1.72 [95% CI 1.24–2.52, p < 0.01] for Pt and Pwt, respectively. Using Pwt group as a reference, patients in Pt group had a high risk of having diabetes [aHR, 1.17 (95% CI 1.04–1.26, p < 0.001)]. Conversely, patients with T2DM tended to have a high risk of periodontitis [aHR, 1.99 (95% CI 1.44–2.48, p < 0.01)]. However, a high risk was noticed for the outcome subgroup of Pt only [aHR, 2.08 (95% CI, 1.50–2.66), p < 0.001)].

Conclusions

The relationship between periodontitis and diabetes is bi-directional.

Diabetes

Database

National health programs

Periodontitis

Diabetes mellitus is a metabolic disease characterized by an increase in the blood glucose level. Periodontal disease is a bacteria-induced chronic inflammatory condition characterized by the breakdown of the tooth’s supporting tissues. Diabetes can be considered a pandemic, causing significant morbidity, mortality, and financial issues. In 2015, approximately 415 million individuals reportedly suffer from diabetes worldwide, and this number is expected to increase each year [1]. Meanwhile, periodontal disease is recognized as the most prevalent inflammatory disease, with approximately 796 million cases of severe periodontitis recorded worldwide in 2017 [2]. Links between diabetes and periodontitis have already been acknowledged. The majority of the studies have indicated that diabetes increases the risk of periodontal inflammation [3–5]; however, some studies have presented no such results [6, 7]. Clinical and epidemiological findings have suggested that periodontal infection contributes to a poorer glycemic control [8–10], whereas the benefits of periodontal therapy on glycemic conditions in periodontitis patients with diabetes remain ambiguous [11, 12]. The two-way relationship between periodontitis and diabetes was advocated far past [13]. However, the bi-directional observations were much limited [14–16]. Two cohort studies suggested the bi-directional [14, 15], but the third called into question the presence of a true bidirectional association [16].

In this present study, by using the nationally representative sample comprising Taiwanese adults, the development of type 2 diabetes mellitus (T2DM) in patients with periodontitis (direction 1), and 2) the development of periodontitis in patients with T2DM (direction 2) were tracked for 15 years.

Data sources

In this study, we used the data from the Longitudinal Health Insurance Database (LHID) in Taiwan (2000–2015), which is a subset database selected randomly from the National Health Insurance Research Database (NHIRD). The National Health Insurance (NHI) Program in Taiwan, which was launched in 1995, includes approximately 23 million beneficiaries or more than 99% of the entire population in Taiwan [17]. The NHI program covers all necessary medical care (including outpatient and inpatient), dental care, Chinese medicine, and prescription drugs. The LHID contains information on health service utilization for approximately one million beneficiaries who represent approximately 5% of the Taiwanese population. The NHIRD contains patient identification numbers, birthdays, sexes, ICD-9-CM diagnostic codes (up to five each), and outcomes.

Study design and sampled participants

Direction 1, that is, tracing T2DM in patients with periodontitis, comprised patients who were newly diagnosed with periodontitis from January 1, 2000, to December 31, 2015, according to the ICD-9-CM code 523.4. Each enrolled periodontitis patient was required to have made at least three dental visits with the 523.4 code being filed within the previous 1 year. The periodontitis patients were then categorized into two groups. The patients who had NHI order code for sub-gingival curettage/root planning (91006-91008C) or periodontal flap operation (91009B-91010B) were categorized into the treated periodontitis group (Pt group), whereas those without these codes were categorized into the periodontitis without therapy group (Pwt group)[18]. The exclusion criteria for the study were as follows: the patients with periodontitis from January 1, 1998, to December 31, 1999; the patients with T2DM from 1998 to 1999 or before their first visit during which periodontitis was diagnosed; the subjects aged < 40 years; and those who had insufficient medical information or withdrawal from the NHI program throughout the 15-year study period (Fig. 1). Among the 1,936,512 individuals, there were 11,011 periodontitis patients enrolled in the Pt group. After matching for age, sex, and index date, 11,011 periodontitis patients were included in the Pwt group (same exclusion criteria and one-fold propensity score) (Fig. 1). Additionally, another 11,011 dental patients who were not diagnosed with periodontitis were categorized into the non-periodontitis control group.

Direction 2, that is, tracking periodontitis in patients with T2DM who were taking ≥ 2 anti-diabetic medications, consisted of patients who were selected from the medical claim data according to the ICD-9-CM code 250. Patients who were diagnosed with T2DM or periodontitis from 1998 to 1999 and aged < 40 years were excluded. In total, 157,798 patients with T2DM were enrolled in the diabetes group, whereas 157,798 participants were recruited in the non-diabetes group. The outcome of periodontitis was then tracked. Besides, the outcome was subsequently categorized into the treated periodontitis subgroup (Pt) and the periodontitis without therapy subgroup (Pwt).

The covariates included gender, age, insurance premium (< 18,000, 18,000–34,999, and ≥ 35,000 NT$), urbanization level of residence (levels 1–4), and level of care (hospital center, regional, and local hospitals). The urbanization level of residence was based on the population and various other indicators. Briefly, the urbanization level 1 was defined as a population > 1,250,000 and specific designation, while the levels 2, 3, and 4 as populations between 500,000 and 1,249,999; 149,999 and 499,999; and < 149,999, respectively.

The baseline comorbidities included hypertension (ICD-9-CM codes: 401.1, 401.9, 402.10, 402.90, 404.10, 404.90, 405.1, and 405.9), hyperlipidemia (ICD-9 CM code: 272.x), coronary artery disease (CAD; ICD-9 CM code: 410–414), obesity (ICD-9 CM codes: 278.00–278.01), smoking (ICD-9-CM code: 305.1), chronic obstructive pulmonary disease (COPD; ICD-9-CM codes: 490–496), and alcoholism (ICD-9-CM codes: 303 and 305.0); additionally, the revised Charlson Comorbidity Index (CCI_R; CCI removed diabetes mellitus, hypertension, and CAD) was included.

Statistical analysis

Chi-square and Fisher’s exact tests were used to evaluate the differences between categorical variables, whereas the t-test and the one-way analysis of variance with Scheffe’s post hoc test were used for continuous variables. Multivariate Cox proportional hazards regression analysis was used to determine the risk of T2DM and periodontitis (directions 1 and 2, respectively). The results were presented as hazard ratios (HRs) with 95% confidence intervals (CIs). Sensitivity analysis was further used to exclude the diagnosis of dementia in the first 1 or 5 year(s) and to eliminate any potential protopathic bias. The difference in the risk of outcome disease between the study and reference groups was estimated using the Kaplan-Meier method and the log-rank test. A two-tailed p < 0.05 was considered statistically significant.

Direction 1: Periodontitis, a risk factor for diabetes

At baseline, significantly different characteristics were noted between groups (Supplementary Table S1). Differences were observed for all variables except gender and age. At the end of the follow-up, 772 (7.01%) and 658 (5.98%) patients with periodontitis developed diabetes in the Pt and Pwt groups, respectively, compared with 490 (4.45%) participants in the non-periodontitis control group (p < 0.001) (Fig. 1; and Table S2, Direction 1). The cumulative incidences for developing diabetes differed between groups (p < 0.001; Fig. 2a).

Risk factors for T2DM in patients with periodontitis

The patients with periodontitis in the Pt and Pwt groups tended to have a significantly increased risk of developing diabetes (Table 1, Direction 1). The adjusted HRs (aHRs) were 1.94 (95% CI, 1.49–2.63; p < 0.001) and 1.72 (95% CI, 1.23–2.53; p < 0.001) for the Pt and Pwt groups, respectively (Table 1). Besides, the participants who lived in places with high urbanization levels, who were treated in hospital centers/regional hospitals, and had hypertension, hyperlipidemia, CAD, obesity, smoking, alcoholism, and greater CCI_R values were associated with a higher risk of developing diabetes. Nevertheless, gender, insured premium, and comorbidity of chronic obstructive pulmonary disease among groups were not significantly associated with the risk of developing diabetes.

Table 1

Risks of periodontitis in DM groups and of DM in periodontitis groups determined by Cox regression
	Direction 1: Periodontitis, risk of DM		Direction 2: DM, risk of periodontitis
Variables	aHR	P	Periodontitis, overall		Pt Subgroup		Pwt Subgroup
	(95% CI)		aHR (95% CI)	P	aHR (95% CI)	P	aHR (95% CI)	P
Cohort, Non-P	Reference
Pt group	1.94 (1.49–2.63)	< 0.001
Pwt group	1.72 (1.23–2.52)	< 0.001
Cohort, Non-DM			Reference		Reference		Reference
DM			1.99 (1.44–2.48)	< 0.001	2.08 (1.50–2.66)	< 0.001	0.97 (0.38–1.57)		0.463
Male (F, ref)	0.96 (0.47–1.74)	0.570	1.11 (0.88–1.76)	0.250	1.11 (0.87–1.76)	0.253	1.12 (0.88–1.76)		0.263
Age (years)	1.22 (1.08–1.36)	< 0.001	1.10 (1.01–1.22)	0.044	1.11 (1.00–1.22)	0.047	1.11 (1.01–1.23)		0.042
IP (NTD10³)
< 18	Reference		Reference		Reference		Reference
18–35	1.27 (0.44–1.93)	0.580	1.14 (0.38–1.82)	0.530	1.15 (0.38–1.83 )	0.540	1.14 (0.37–1.80)		0.530
> 35	1.42 (0.85–1.97)	0.438	1.30 (0.62–1.94)	0.472	1.33 (0.69–1.95)	0.468	1.28 (0.56–1.93)		0.489
Urbanization
1 (Highest)	1.85 (1.24–2.24)	< 0.001	1.76 (1.26–2.18)	< 0.001	1.76 (1.25–2.24)	< 0.001	1.77 (1.28–2.27)		< 0.001
2	1.72 (1.18–2.06)	< 0.001	1.67 (1.23–2.11)	< 0.001	1.66 (1.21–2.12)	< 0.001	1.68 (1.25–2.12)		< 0.001
3	1.63 (1.16–2.02)	< 0.001	1.65 (1.18–2.06)	< 0.001	1.64 (1.15–2.05)	< 0.001	1.66 (1.19–2.08)		< 0.001
4 (Lowest)	Reference		Reference		Reference		Reference
Level of care
Center Hop.	2.40 (1.80–3.07)	< 0.001	2.24 (1.54–2.89)	< 0.001	2.26 (1.52–2.91)	< 0.001	2.23 (1.52–2.86)		< 0.001
Regional Hop.	1.96 (1.43–2.63)	< 0.001	1.82 (1.23–2.40)	< 0.001	1.80 (1.21–2.78)	< 0.001	1.84 (1.18–2.06)		< 0.001
Local Hop.	Reference		Reference		Reference		Reference
HTN	1.97 (1.57–2.45)	< 0.001	1.83 (1.46–2.35)	< 0.001	1.80 (1.45–2.36)	< 0.001	1.85 (1.46–2.35)		< 0.001
HLip	1.84 (1.41–2.18 )	< 0.001	1.73 (1.27–2.37)	< 0.001	1.75 (1.28–2.35)	< 0.001	1.71 (1.26–2.38)		< 0.001
CAD	1.75 (1.27–2.17 )	< 0.001	1.58 (1.33–2.10)	< 0.001	1.54 (1.24–2.14)	< 0.001	1.63 (1.43–2.07)		< 0.001
Obesity	4.87 (2.16–6.70)	< 0.001	1.95 (1.48–2.51)	< 0.001	1.94 (1.43–2.35)	< 0.001	1.95 (1.48–2.42)		< 0.001
Smoking	1.35 (1.11–1.51)	< 0.001	1.25 (1.20–1.64)	< 0.001	1.48 (1.21–1.67)	< 0.001	1.38 (1.19–1.61)		< 0.001
COPD	1.20 (0.86–1.48)	0.276	1.29 (0.41–1.79)	0.530	1.39 (0.59–1.98)	0.486	1.19 (0.24–1.61)		0.587
Alcoholism	1.53 (1.30–1.79)	< 0.001	1.85 (1.34–2.30)	< 0.001	1.85 (1.33–2.30)	< 0.001	1.85 (1.39–2.37)		< 0.001
CCI_R	1.25 (1.06–1.70 )	< 0.001	1.30 (1.10–1.52)	< 0.001	1.38 (1.10–1.29)	< 0.001	1.30 (1.10–1.69)		< 0.001
Abbreviation: aHR, adjusted hazard ratio; the adjusted variables listed in the table. CI, confident interval.
Pt, periodontitis and periodontal therapy received. Pwt, periodontitis and no-therapy received. Non-P, no periodontitis. DM, diabetes mellitus. Non-DM, without DM. IP, insured premium, 10³ NTD. HTN, hypertension. HLip, hyperlipidemia. CAD, coronary artery disease. COPD, chronic obstructive pulmonary disease. CCI_R, Charlson comorbidity index revised.

Sensitivity of periodontitis as a risk factor for T2DM

The rates of diabetes were 13.40 and 10.12 per 1,000 per 10³ person-years in the Pt and Pwt groups, respectively, which were significantly greater than that in the non-periodontitis group (6.31 per 1,000 person-years; Table 2, Direction 1). It was also observed that patients with periodontitis in the Pt and Pwt groups tended to have a significantly increased risk of developing diabetes, even though the data from the first 1 or 5 years were excluded. Moreover, the treated periodontitis patients had a high risk of having diabetes if the non-treated periodontitis patients were used as the reference, despite the data exclusion or non-exclusion (1.17, 95% CI, 1.04–1.26, p < 0.001 for the non-excluded data).

Table 2

Sensitivity of periodontitis or DM as the risk in the model.
	Events	Years	Rate (/10³ PY s)	aHR (95% CI)	aHR (95% CI)
Direction 1: Periodontitis, risk of DM
Pt group	772	57,601	13.40	1.94 (1.49–2.63)*	1.17(1.04–1.26)**
Pwt group	658	65,012	10.12	1.72 (1.24–2.52)*	Reference
Non-P control	490	77,656	6.31	Reference
Data in the first year excluded
Pt group	724	49,977	14.49	1.95 (1.49–2.64)*	1.18 (1.06–1.31)*
Pwt group	631	59,701	10.57	1.74 (1.24–2.54)*	Reference
Non-P control	465	70,987	6.55	Reference
Data in the first 5 years excluded
Pt group	509	32,402	15.71	1.96 (1.51–2.74)*	1.18 (1.07–1.35)*
Pwt group	482	43,501	11.08	1.75 (1.26–2.54)*	Reference
Non-P control	399	58,732	6.79	Reference
Direction 2: DM, risk of periodontitis
DM group	1,662	1,599,752	1.04	1.99 (1.44–2.48)*
Non-DM control	430	2,756,716	0.16	Reference
Data in the first year excluded
DM group	1,493	1,294,201	1.15	2.01 (1.47–2.57)*
Non-DM control	394	2,650,809	0.15	Reference
Data in the first 5 years excluded
DM group	1,313	1,076,506	1.22	2.06 (1.50–2.65)*
Non-DM control	323	1,918,852	0.17	Reference
Subgroup A: DM, risk of periodontitis with therapy received
DM group	1,579	1,598,722	0.99	2.08 (1.50–2.66)*
Non-DM control	370	1,678,127	0.22	Reference
Data in the first year excluded
DM group	1,422	1,287,843	1.10	2.10 (1.58–2.68)*
Non-DM control	339	1,370,244	0.25	Reference
Data in the first 5 years excluded
DM group	1,254	1,065,244	1.18	2.16 (1.59–2.75)*
Non-DM control	278	1,083,751	0.26	Reference
Subgroup B: DM, risk of periodontitis without therapy received
DM group	83	1,542,237	0.05	0.97 (0.38–1.57)
Non-DM control	60	1,078,589	0.06	Reference
Data in the first year excluded
DM group	71	1,291,022	0.05	0.97 (0.44–1.62)
Non-DM control	55	972,682	0.06	Reference
Data in the first 5 years excluded
DM group	59	1,142,575	0.05	0.98 (0.49–1.66)
Non-DM control	45	835,101	0.05	Reference
* and bold: significant difference at p < 0.001.Abbreviation: PYs, person-years. aHR:, adjusted hazard ratio and adjusted for the variables listed in Tables of 1 and 2. Pt, periodontitis with therapy received. Pwt, periodontitis without therapy received. Non-P, no periodontitis. DM, diabetes mellitus. Non-DM, without DM.

Direction 2: T2DM, a risk factor for developing periodontitis

The general characteristics of the diabetes and non-diabetes groups at the baseline were different, except for gender, age, or insurance premium (Supplementary Table S1). At the end of follow-up, 1,662 (1.05%) diabetic patients have reportedly developed periodontitis compared with 430 (0.27%) participants in the non-diabetic control group (p < 0.001) (Fig. 1; and Table S2, Direction 2). The difference was further noted in other variables except gender and insured premium. The cumulative incidence of periodontitis in the diabetes group was significantly higher than that in the non-diabetes control group (Fig. 2b1). As per the outcomes of periodontitis subgroups however, statistically different incidences between the groups of diabetes and non-diabetes were noticed for Pt, but not for Pwt subgroup (Fig. 2, b2, and b3).

Risks for developing periodontitis in patients with T2DM

Using the non-diabetes group as the reference, the aHR of developing periodontitis in the diabetes group was 1.99 (95% CI, 1.44–2.48; p < 0.001; Table 1, Direction 2). The elder patients who lived in areas of high urbanization levels were attended to in the center/regional hospital; moreover, hypertension, hyperlipidemia, CAD, obesity, smoking, alcoholism, and high CCI_R values tended to have a high risk of developing periodontitis. However, the statistical significance for an increased risk of developing the outcome of periodontitis in diabetic patients was observed for the Pt subgroup (aHRs: 2.08; 95% CI, 1.50–2.66; p < 0.001), but not for the Pwt subgroup (aHRs: 0.97; 95% CI, 0.38–1.57; p = 0.463).

Sensitivity of T2DM as a risk factor for periodontitis

The rate of having periodontitis in the diabetes group (1.04 per 10³ person-years) was determined to be significantly greater than that in the non-diabetes control group (0.16 per 10³ person-years; Table 2, Direction 2). The diabetic patients tended to have a significantly increased risk of developing periodontitis. After subgrouping the outcome of periodontitis, the rate of developing Pt was significantly greater in the diabetes group than that in the non-diabetes group (0.99 vs. 0.22 per 10³ person-years). The aHRs was 2.08 (95% CI, 1.50–2.66; p < 0.001). After the exclusion of the data from the first 1 or 5 years, the greater rate and aHR in the diabetes group than that in the non-diabetes group were repeated. For the Pwt subgroup, however, the rate in the diabetes group did not differ significantly from that in the non-diabetes group (0.05 and 0.06 per 10³ person-years, respectively). The aHRs of 0.97 (95% CI, 0.38–1.57; p = 0.462) was obtained. No difference was also noticed after excluding the data from the first 1 or 5 years.

This retrospective cohort study aimed to bi-directionally track the incidences of T2DM and periodontitis in aged adults (> 40 years old) who were chosen from the base population of Taiwan (Table 1, Fig. 1). As per our results, it was found that the risk of developing T2DM increased in the patients with periodontitis compared to those non-periodontitis control (Fig. 2a, Tables 1 and 2). In the same population, within the same observation period, the patients with T2DM also had an increased risk of developing periodontitis (significant greater risks noticed on the Pt subgroup only) (Fig. 2b, Tables 1 and 2). In brief, in this nationally representative sample, the bi-directional association between the two diseases was noticed during 15 years of tracking, similar to that previously reported community study [14]. In that community study, T2DM was defined as having a history from the questionnaire or fasting plasma glucose, and the periodontal status was assessed by the Community Periodontal Index of Treatment Needs (CPITN). However, the simple screening based on the fasting glucose alone might ineffectively detect an unacceptable number of subjects with glucose intolerance [19] Whether CPITN could accurately reflect the actual periodontal status of the sample population is also questioned [20]. Besides, the professional levels for those examiners were not clearly defined. In the present study, the accuracy and validity of the diagnoses in NHIRD in Taiwan were high [21]. Strategies to reduce the diagnosis bias were further selected [21], such as the triple diagnoses of periodontitis within the 1-year claim period. The enrolls in Pt group should provide full mouth charting/radiography. The subjects in the non-periodontitis control group should have enrollment in the Dentistry database of NHIRD to ensure dental examination and concerns. Similarly, the patients with diabetes were enrolled by the ICD-9-CM code of 250 and took more than two anti-diabetic medications.

Inflammation has been identified to be a known driver of insulin resistance [22]. Chronic dysregulation of the peripheral cytokine pool is a characteristic pre-diabetic feature [23]. It has been determined that circulating mediators such as C-reactive protein (CRP), tumor necrosis factor (TNF)-α, and interleukin (IL)-6, may be elevated in periodontal diseases and correlated with the clinical periodontal parameters [24]. A longitudinal study reported that the increase in HbA1c levels over a 5-year period in patients with periodontitis was highest in individuals with high levels of CRP [25]. The role of systemic inflammation as a mediator of the linkage between periodontitis and impaired fasting glucose was found using the mediation analysis [10]. Systemic oxidative stress is elevated in patients with diabetes and periodontitis [22]. It has been proposed that hyperactive neutrophils in the periodontium may be an important source of reactive oxygen species, which then activates the pro-inflammatory pathways and promotes insulin resistance in patients with periodontitis and diabetes [22]. Another study reported that the markers of lipid peroxidation in the gingival crevicular fluid were correlated with the clinical parameters of periodontitis and the levels of inflammatory mediators in diabetic patients [26].

The evidence of periodontal microorganisms having a direct impact on glycemic control or the diabetic state of the patient might still be lacking; however, a recent study in mice demonstrated the translocation of Porphyromonas gingivalis and gingipains to the pancreas, with alterations in the morphology of the islets in the pancreas [27]. Following non-surgical periodontal therapy, Porphyromonas gingivalis was detected more frequently in patients with increased HbA1c values compared with those with decreased values [28]. Furthermore, Porphyromonas gingivalis with type II fimbriae were detected only in patients with increased HbA1c levels.

Detailed pathogenesis of the high risk of developing periodontitis in diabetic patients remains under investigation; however, studies reported that the cytokine levels in the gingival cervical fluid, saliva, and/or gingival tissue of diabetic patients with periodontitis were altered when compared with those in systemically healthy patients [29]. The levels of TNF-α and IL-6 were significantly increased in diabetic mice compared with those in normal mice after the inoculation of Porphyromonas gingivalis [30]. The experiments with TNF-α inhibitors in the diabetic animal models of periodontitis indicated that these changes may be the secondary effects of TNF-α [31].

In addition, the advanced glycation end products (AGEs) in the gingival tissues of diabetic patients with periodontitis [32] have been significantly associated with the extent of periodontitis in diabetic individuals. Recently, hyperglycemia and AGEs have been found to increase the lipopolysaccharide-induced production of IL-6 in human gingival fibroblasts [33]. The interactions between AGEs and their receptors were significantly higher in the inflamed periodontal tissues of rats induced with hyperglycemia when compared with those in rats with normal blood glucose levels [34]. Treatment using soluble receptors of AGE decreased the levels of TNF-α, IL-6, and metalloproteinases in gingival tissues and suppressed alveolar bone loss in diabetic animals [35].

It is generally believed that the presence of diabetes has no significant effect on the composition of periodontal microorganisms. Recently, the findings of the shift in the subgingival microbiome suggest that diabetic patients are more susceptible to shifts in the subgingival microbiome toward dysbiosis [36]. Besides, Porphyromonas gingivalis-induced alveolar bone loss was noted to increase in diabetic mice when compared with non-diabetic controls; this was accompanied by the enhanced expression of the AGEs and their receptors [37].

This study has several limitations. First, like previous studies using the NHIRD on periodontal diseases, precise data on the severity of periodontitis were unavailable. Second, other factors such as genetic, psychosocial, and detailed environmental factors were excluded from the dataset. Third, the claims dataset did not include descriptions of the metabolic or diabetic state of the patient. Fourth, there were only a few patients in the Pwt subgroup. Therefore, it is advisable to interpret the results of this study with caution. However, the strengths of the study are as follows: First, this study included a large population, which allowed the results from this analysis to be generalized to the entire population of Taiwan. Second, because the NHIRD provides continued coverage; hence, we could follow the population trend for 15 years. Third, the accuracy and validity of the diagnoses in this study are high.

Using the data from the NHIRD of Taiwan, the relationship between periodontitis and diabetes was examined bi-directionally. At the end of the 15-year follow-up, the cumulative incidences of T2DM among the two periodontitis groups and non-periodontitis control group were determined to be statistically different (Direction 1). The aHRs were 1.94 and 1.72 for the Pt and Pwt groups, respectively. Besides, the risk of T2DM in the Pt group was greater than that in the Pwt group. In Direction 2, T2DM as a risk factor for periodontitis was tested. An increased risk of developing periodontitis was observed among patients with T2DM (aHR = 1.99, the development of periodontitis in the control participants without T2DM as the reference). However, the greater risk was observed only for the outcome of the treated periodontitis subgroup but not of the periodontitis without therapy subgroup. Based on the findings of this retrospective cohort study, we suggested that the association between diabetes and periodontitis was bi-directional.

Acknowledgments

We thank all the experts, the participating institutions, including Tri-Service General Hospital, National Defense Medical Center, and Taipei Tzu Chi Hospital.

Author Contribution

Wu-Chien Chien contributed to conception, design, data acquisition, analysis, interpretation, and drafted the manuscript; Chi-Hsiang Chung contributed to conception, design, data acquisition, analysis, interpretation, drafted and critically revised the manuscript. Earl Fu contributed to design, data acquisition, analysis, and interpretation for the data; Chia-Mao Cheng contributed to data acquisition and interpretation for the study; Hsiao-Pei Tu to data acquisition, analysis; Wei-Cheng Lee e contributed to data interpretation for the study; Wei-Liang Chen to data acquisition, analysis, and interpretation for the study; and Kuang-Chung Shih contributed to conception, design, data acquisition, interpretation, drafted and critically revised the manuscript. All authors provided final approval for the submission.

Funding

This study was partially supported by a research grant from Tri-Service General Hospital Research Foundation (TSGH-B-111018).

Compliance with ethical standards

Conflict of interests

The authors have no conflicts of interest to declare.

Ethics approval

Data access and ethical approval for this study were approved by the Institutional Review Board of the Taipei Tzu Chi Hospital (No. 09-W-043).

Ogurtsova K, da Rocha Fernandes JD, Huang Y, et al. IDF Diabetes Atlas: global estimates for the prevalence of diabetes for 2015 and 2040. Diabetes Res Clin Pract 128, 40-50 (2017). 10.1016/j.diabres.2017.03.024
Bernabe E, Marcenes W, Hernandez CR, et al. Global, regional, and national levels and trends in burden of oral conditions from 1990 to 2017: a systematic analysis for the global burden of disease 2017 study. J Dent Res 99, 362-373 (2020). 10.1177/0022034520908533
Andriankaja OM, Munoz-Torres FJ, Vivaldi-Oliver J, et al. Insulin resistance predicts the risk of gingival/periodontal inflammation. J Periodontol 89, 549-557 (2018). 10.1002/jper.17-0384
de Araujo Nobre M, Malo P Prevalence of periodontitis, dental caries, and peri-implant pathology and their relation with systemic status and smoking habits: results of an open-cohort study with 22009 patients in a private rehabilitation center. J Dent 67, 36-42 (2017). 10.1016/j.jdent.2017.07.013
Lee CY, Kuan YH, Tsai YF, et al. Correlation between diabetes mellitus and periodontitis in Taiwan: a nationwide cohort study. Diabetes Res Clin Pract 150, 245-252 (2019). 10.1016/j.diabres.2019.03.019
Aoyama N, Suzuki JI, Kobayashi N, et al. Japanese cardiovascular disease patients with diabetes mellitus suffer increased tooth loss in comparison to those without diabetes mellitus -a cross-sectional study. Intern Med 57, 777-782 (2018). 10.2169/internalmedicine.9578-17
Montero E, Carasol M, Fernandez-Meseguer A, et al. Prediabetes and diabetes prevalence in the Workers' Oral Health Study. Clin Oral Investig 23, 4233-4241 (2019). 10.1007/s00784-019-02875-3
Vadakkekuttical RJ, Kaushik PC, Mammen J, George JM Does periodontal inflammation affect glycosylated haemoglobin level in otherwise systemically healthy individuals? - a hospital based study. Singapore Dent J 38, 55-61 (2017). 10.1016/j.sdj.2017.08.002
Wernicke K, Zeissler S, Mooren FC, et al. Probing depth is an independent risk factor for HbA1c levels in diabetic patients under physical training: a cross-sectional pilot-study. BMC Oral Health 18, 46 (2018). 10.1186/s12903-018-0491-9
Torrungruang K, Ongphiphadhanakul B, Jitpakdeebordin S, Sarujikumjornwatana S Mediation analysis of systemic inflammation on the association between periodontitis and glycaemic status. J Clin Periodontol 45, 548-556 (2018). 10.1111/jcpe.12884
D'Aiuto F, Gkranias N, Bhowruth D, et al. Systemic effects of periodontitis treatment in patients with type 2 diabetes: a 12 month, single-centre, investigator-masked, randomised trial. Lancet Diabetes Endocrinol 6, 954-965 (2018). 10.1016/S2213-8587(18)30038-X
Vergnes JN, Canceill T, Vinel A, et al. The effects of periodontal treatment on diabetic patients: The DIAPERIO randomized controlled trial. J Clin Periodontol 45, 1150-1163 (2018). 10.1111/jcpe.13003
Grossi SG, Genco RJ Periodontal disease and diabetes mellitus: a two-way relationship. Ann Periodontol 3, 51-61 (1998). 10.1902/annals.1998.3.1.51
Chiu SY, Lai H, Yen AM, et al. Temporal sequence of the bidirectional relationship between hyperglycemia and periodontal disease: a community-based study of 5,885 Taiwanese aged 35-44 years (KCIS No. 32). Acta Diabetol 52, 123-131 (2015). 10.1007/s00592-014-0612-0
Morita I, Inagaki K, Nakamura F, et al. Relationship between periodontal status and levels of glycated hemoglobin. J Dent Res 91, 161-166 (2012). 10.1177/0022034511431583
Alshihayb TS, Kaye EA, Zhao Y, Leone CW, Heaton B A quantitative bias analysis to assess the impact of unmeasured confounding on associations between diabetes and periodontitis. J Clin Periodontol 48, 51-60 (2021). 10.1111/jcpe.13386
Ho Chan WS Taiwan's healthcare report 2010. EPMA J 1, 563-585 (2010). 10.1007/s13167-010-0056-8
Su SY, Chien WC, Chung CH, Su WF, Fu E Association of periodontitis with tinnitus: A population-based cohort study in Taiwan. J Clin Periodontol (2022). 10.1111/jcpe.13670
Sainaghi PP, Castello L, Limoncini AM, et al. Poor specificity of fasting plasma glucose cut-off values in ruling out glucose intolerance: the complementary usefulness of OGTT. Exp Clin Endocrinol Diabetes 115, 112-117 (2007). 10.1055/s-2007-949151
Bassani DG, da Silva CM, Oppermann RV Validity of the "Community Periodontal Index of Treatment Needs" (CPITN) for population periodontitis screening. Cad Saude Publica 22, 277-283 (2006). 10.1590/s0102-311x2006000200005
Hsieh CY, Su CC, Shao SC, et al. Taiwan's National Health Insurance Research Database: past and future. Clin Epidemiol 11, 349-358 (2019). 10.2147/CLEP.S196293
Bullon P, Morillo JM, Ramirez-Tortosa MC, et al. Metabolic syndrome and periodontitis: is oxidative stress a common link? J Dent Res 88, 503-518 (2009). 10.1177/0022034509337479
Kolb H, Mandrup-Poulsen T The global diabetes epidemic as a consequence of lifestyle-induced low-grade inflammation. Diabetologia 53, 10-20 (2010). 10.1007/s00125-009-1573-7
Paraskevas S, Huizinga JD, Loos BG A systematic review and meta-analyses on C-reactive protein in relation to periodontitis. J Clin Periodontol 35, 277-290 (2008). 10.1111/j.1600-051X.2007.01173.x
Demmer RT, Desvarieux M, Holtfreter B, et al. Periodontal status and A1C change: longitudinal results from the study of health in Pomerania (SHIP). Diabetes Care 33, 1037-1043 (2010). 10.2337/dc09-1778
Bastos AS, Graves DT, Loureiro AP, et al. Lipid peroxidation is associated with the severity of periodontal disease and local inflammatory markers in patients with type 2 diabetes. J Clin Endocrinol Metab 97, E1353-1362 (2012). 10.1210/jc.2011-3397
Ilievski V, Bhat UG, Suleiman-Ata S, et al. Oral application of a periodontal pathogen impacts SerpinE1 expression and pancreatic islet architecture in prediabetes. J Periodontal Res 52, 1032-1041 (2017). 10.1111/jre.12474
Makiura N, Ojima M, Kou Y, et al. Relationship of Porphyromonas gingivalis with glycemic level in patients with type 2 diabetes following periodontal treatment. Oral Microbiol Immunol 23, 348-351 (2008). 10.1111/j.1399-302X.2007.00426.x
Ribeiro FV, de Mendonca AC, Santos VR, et al. Cytokines and bone-related factors in systemically healthy patients with chronic periodontitis and patients with type 2 diabetes and chronic periodontitis. J Periodontol 82, 1187-1196 (2011). 10.1902/jop.2011.100643
Nishihara R, Sugano N, Takano M, et al. The effect of Porphyromonas gingivalis infection on cytokine levels in type 2 diabetic mice. J Periodontal Res 44, 305-310 (2009). 10.1111/j.1600-0765.2008.01130.x
Pacios S, Kang J, Galicia J, et al. Diabetes aggravates periodontitis by limiting repair through enhanced inflammation. FASEB J 26, 1423-1430 (2012). 10.1096/fj.11-196279
Schmidt AM, Weidman E, Lalla E, et al. Advanced glycation endproducts (AGEs) induce oxidant stress in the gingiva: a potential mechanism underlying accelerated periodontal disease associated with diabetes. J Periodontal Res 31, 508-515 (1996). 10.1111/j.1600-0765.1996.tb01417.x
Chiu HC, Fu MM, Yang TS, et al. Effect of high glucose, Porphyromonas gingivalis lipopolysaccharide and advanced glycation end-products on production of interleukin-6/-8 by gingival fibroblasts. J Periodontal Res 52, 268-276 (2017). 10.1111/jre.12391
Chang PC, Chien LY, Chong LY, Kuo YP, Hsiao JK Glycated matrix up-regulates inflammatory signaling similarly to Porphyromonas gingivalis lipopolysaccharide. J Periodontal Res 48, 184-193 (2013). 10.1111/j.1600-0765.2012.01519.x
Lalla E, Lamster IB, Feit M, et al. Blockade of RAGE suppresses periodontitis-associated bone loss in diabetic mice. J Clin Invest 105, 1117-1124 (2000). 10.1172/JCI8942
Shi B, Lux R, Klokkevold P, et al. The subgingival microbiome associated with periodontitis in type 2 diabetes mellitus. ISME J 14, 519-530 (2020). 10.1038/s41396-019-0544-3
Lalla E, Lamster IB, Feit M, Huang L, Schmidt AM A murine model of accelerated periodontal disease in diabetes. J Periodontal Res 33, 387-399 (1998). 10.1111/j.1600-0765.1998.tb02335.x

0TableSupplementarymaterials.docx

Bi-directional association between type 2 diabetes and periodontitis in the nationally representative sample comprising Taiwanese adults from 2000 to 2015

Status:

Version 1

Abstract

Objective

Methods

Results

Conclusions

Figures

Introduction

Materials And Methods

Data sources

Study design and sampled participants

Statistical analysis

Results

Discussion

Conclusions

Declarations

References

Supplementary Files

Status:

Version 1