Association between periodontitis and the prevalence and prognosis of prediabetes: a population-based study

The NHANES (RRID:SCR_013201) is continuous survey research that provides population estimates related to the nutrition and health of adults and children in America. The survey employs a stratified, multistage probability design to recruit a representative sample of the American population. Data were obtained via personal structured interviews at home, health examinations at a mobile examination center, and specimen analyses in the laboratory. NHANES 1999–2004 and 2009–2014 were approved by The National Center for Health Statistics (NCHS) Ethics Review Board and conducted in accordance with the Helsinki Declaration of 1975, as revised in 2013. Informed consent was obtained from all participants. The acquisition and analysis of data was consistent with NHANES research requirements.

Prediabetes was diagnosed by one of the following four principles: (1) Self-reported prediabetes: participants answer “Yes” to the question “Have you ever been told prediabetes/IGT/IFG/borderline diabetes/blood sugar higher than normal but not high enough to be called diabetes in diabetes questionnaire”. (2) HbA1c: 5.7%–7.0%. (3) Fasting plasma glucose: 5.6–7.0 mmol/L. (4) oral glucose tolerance test: 7.8–11.0 mmol/L [10].

Periodontitis was defined as clinical attachment loss (CAL; the difference between pocket depth and gingival recession) ≥ 3 mm. This has been recommended for epidemiological studies as an extent measure of periodontitis and is henceforth referred to as CAL extent. The severity of periodontitis was defined by following Centers for Disease Control and Prevention in partnership with the American Academy of Periodontology (CDC–AAP) case definitions for surveillance of periodontitis [11].

We used the data from the NHANES 1999–2004 and 2009–2014 cycles. First, 15390 patients with clear periodontal, prediabetes and other covariable data were included to explore the association between periodontitis status and the risk of prediabetes. Furthermore, we enrolled 4518 participants with prediabetes of 15390 sample to further explore the association between periodontitis status and all-cause mortality (Fig. 1).

The outcome of association analysis was diagnosed with prediabetes. The outcome of survival analysis was all-cause mortality with periodontitis or not among patients diagnosed with prediabetes. All-cause mortality was ascertained by linkage to National Death Index records through 2018. ICD-10 was used to determine the causes of death.

Standardized questionnaires obtained information on sociodemographic characteristics, smoking status, and alcohol consumption. Never smokers were classified as those who reported smoking < 100 cigarettes during their lifetime. Those who smoked > 100 cigarettes in their lifetime were considered current smokers, and those who smoked > 100 cigarettes and had quit smoking were considered former smokers. Nondrinkers were defined as consuming no drinks per year. Heavy alcohol use was defined as ≥ 3 drinks per day for females, ≥ 4 drinks per day for males, or binge drinking [≥ 4 drinks on the same occasion for females, ≥ 5 drinks on the same occasion for males] on 5 or more days per month. Moderate alcohol use was defined as ≥ 2 drinks per day for females, ≥ 3 drinks per day for males, or binge drinking ≥ 2 days per month. Other frequencies of alcohol consumption were defined as mild alcohol use. Body mass index (BMI) was calculated from weight/height² (kg/m²). Diabetes was defined as follows: (1) doctor diagnosed you as having diabetes; (2) HbA1c > 7%; (3) fasting glucose ≥ 7.0 mmol/L; (4) random blood glucose ≥ 11.1 mmol/L, (5) two-hour oral glucose tolerance test (OGTT) blood glucose ≥ 11.1 mmol/L; and (5) use of diabetes medication or insulin. Hyperlipidemia was defined as follows: (1) triglyceride ≥ 150 mg/dL, (2) cholesterol ≥ 200 mg/dL, (3) low-density lipoprotein (LDL)-cholesterol ≥ 130 mg/dL, (4) high-density lipoprotein (HDL)-cholesterol < 40 mg/dL (male); 50 mg/dL (female), and (5) use of hyperlipidemia medication. Coronary heart disease (CHD) was defined as self-reported CHD. Homeostasis model assessment-insulin resistance (HOMA-IR) was calculated by the following formula: HOMA-IR = [fasting plasma glucose (mU/L) × fasting plasma glucose (mmol/L)]/22.5. Homeostasis model assessment-insulin sensitivity (HOMA-IS) was estimated by the following formula: HOMA-IS = 22.5/[fasting plasma glucose (mU/L) × fasting plasma glucose (mmol/L)]. In addition, strict laboratory analyses were performed, including the assessment of glucose, insulin, HbA1c, triglyceride, HDL, LDL, and C-reactive protein (CRP) levels at baseline. Further details of these measurements were documented in the NHANES Laboratory Medical Technologists Procedures Manual.

All analyses incorporated sample weights, strata, and primary sampling units to produce accurate national estimates. Sample characteristics are reported as the means (SEs) for normally distributed continuous variables, medians (interquartile ranges) for nonnormally distributed continuous variables, and percentages for categorical variables. Weighted multivariate logistic regression models were performed to estimate the OR and 95% confidence interval associated with periodontitis status and prediabetes with the first group as a reference. Weighted Cox proportional hazards regression was used to estimate hazard ratios (HRs) and 95% CIs of all-cause mortality in relation to periodontitis. Person-time was calculated as the interval between the NHANES interview date and the date of death or the end of the follow-up (31-December-2018), whichever occurred first. We constructed three statistical models. Model 1: no covariates were adjusted. Model 2: Age (continuous), gender (male or female), race/ethnicity (Mexican American, other Hispanic, non-Hispanic white, non-Hispanic black, or other Race), and education level (less than high school, high school, more than high school) were adjusted. Model 3: age (continuous), gender (male or female) and race/ethnicity (Mexican American, other Hispanic, non-Hispanic white, non-Hispanic black, or other Race), education level (less than high school, high school, more than high school), smoking status (never, former, or current), alcohol use (never, former, mild, moderate. severe), BMI (continuous), hypertension (no or yes), hyperlipidemia (yes or no), and CHD (yes or no) were adjusted. Any individuals with missing covariate data were not included in the analyses in Model 2 and Model 3 (listwise deletion). Furthermore, stratified analyses were conducted to examine whether the detected association differed by age, sex, and body mass index. All analyses were performed using R software (4.1.0). Two-sided P < 0.05 was considered statistically significant. The Bonferroni correction threshold was used to account for multiple comparisons.

In this study, 15390 participants were enrolled in our analysis. There were 5033 participants with periodontitis and 10357 participants without periodontitis. Compared to participants without periodontitis, participants with periodontitis were more likely to be female, older, have a higher education level, hypertension, hyperlipidemia, CHD, be former alcohol user, smoke, have a higher BMI and glucose/insulin/HbA1c/HOMA-IR, have prediabetes, and have diabetes (Table 1).

We conducted weighted multivariate logistic regression models to explore the association between periodontitis status and prediabetes (Table 2). Compared to nonperiodontitis, periodontitis was positively associated with prediabetes in all models (OR = 2.37, 95% CI 2.16–2.60; 1.45, 95% CI 1.30–1.61; 1.47, 95% CI 1.31–1.63). The p value of all models was < 0.001. We further stratified periodontitis into three statuses according to the CDC–AAP case definitions. In Model 1, all periodontitis statuses were significantly associated with prediabetes (OR = 1.42, 95% CI 1.15–1.77; 2.39, 95% CI 2.15–2.66; 2.74, 95% CI 2.26–3.32). The p value of all periodontitis statuses equaled 0.001. After adjusting for covariables, moderate and severe periodontitis were positively associated with prediabetes in Model 2 (OR = 1.46, 95% CI 1.30–1.65, 1.57, 95% CI 1.27–1.94) and Model 3 (OR = 1.46, 95% CI 1.29–1.65; 1.62, 95% CI 1.31–2.01). The p value of moderate and severe periodontitis was < 0.001 in Models 2 and 3. Furthermore, a significant association remained between periodontitis and prediabetes in most subgroups (Additional file 1: Table S1).

A total of 4518 participants with prediabetes were included to further explore the correlation between periodontitis status and diabetes mortality by weighted Cox proportional hazards regression. The characteristics of these participants are shown in Additional file 1: Table S2. The average follow-up time was 128.41 ± 1.67 months (Additional file 1: Table S2). Compared to the nonperiodontitis group, participants with periodontitis were pivotally correlated with mortality of prediabetes in all models (HR = 2.41, 95% Cl 1.98–2.94, p < 0.001; 1.47, 95% Cl 1.20–1.80, p < 0.001; 1.28, 95% Cl 1.04–1.57, p = 0.02). After stratifying the status of periodontitis, moderate and severe periodontitis were positively correlated with mortality of prediabetes (HR = 2.42, 95% Cl 1.95–3.01, p < 0.001; 1.806, 95% Cl 1.19–2.74, p = 0.005). After further adjustments for age, sex, race, education level, alcohol use, BMI, hypertension, CHD, moderate periodontitis was significantly associated with mortality of prediabetes in Model 2 (HR = 1.45, 95% CI 1.17–1.80, p < 0.001) and Model 3 (HR = 1.25, 95% CI 1.00–1.55, p = 0.047) (Table 3). The Kaplan–Meier survival curve presented similar results (Fig. 2). In the subgroup analysis, older age, female sex, high education level, mild alcohol use, former smoking, hypertension, hyperlipidemia, and higher BMI were positively correlated with all-cause mortality of prediabetes (Additional file 1: Table S3).