Association between periodontitis and diabetes

Background + Aims

• Prediabetes, a precursor to diabetes mellitus, is marked by elevated fasting plasma glucose levels below the diabetes threshold.
• Its prevalence varies based on diagnostic criteria and population characteristics, affecting nearly half of Asian and European Caucasian adults.
• Up to 70% of individuals with prediabetes progress to diabetes, leading to complications like cardiovascular disease, retinopathy, and neuropathy, linked to hyperglycaemia-induced biological abnormalities such as oxidative stress and impaired wound healing. These effects extend to periodontal tissues, with periodontitis recognised as the “sixth complication” of diabetes.
• Given the prevalence of prediabetes and periodontitis, studying the association between elevated fasting glucose and periodontal tissue destruction in non-diabetic individuals is critical. Prior cross-sectional studies report positive links, while limited longitudinal research confirms increased periodontal risks in dysglycaemic conditions.
• This study hypothesises a connection between fasting plasma glucose and periodontal status, examining changes in glucose levels and periodontal health over 15 years.

Materials + Methods

• The Northern Finland Birth Cohort 1966 study investigated the relationship between fasting plasma glucose (FPG) and periodontal health in a well-defined population.
• Data were drawn from 1,964 participants who underwent an oral health examination at age 46, with prior health information collected at age 31.
• The study excluded individuals with diagnosed diabetes or inflammatory conditions, focusing on never-smokers to minimise confounding factors, resulting in 647 eligible participants.
• Periodontal health was assessed through:
  • Clinical assessments by calibrated dentists, with rigorous quality checks for inter- and intra-examiner reliability.
    • Probing pocket depth (PPD)
    • Bleeding on probing (BOP)
  • Radiographic assessments via panoramic radiographs
    • Alveolar bone loss (BL)
• FPG levels were categorised into normoglycaemia, slightly elevated, and prediabetes, based on ADA criteria, at ages 31 and 46. Participants were grouped into stable or worsened glucose balance categories to assess long-term glucose changes.
• Data on socio-demographics, health behaviours, lipid profiles, and BMI were included as covariates.
• Statistical analyses employed Poisson regression models to estimate relative risks, adjusted for potential confounders, and examined associations between glucose levels and periodontal outcomes in the total population and highly educated subgroups.

Results

• Among 647 participants, 43% maintained FPG <5.0 mmol/l, while 47% had slightly elevated FPG (5.0–5.59 mmol/l), and 10% had prediabetes (5.6–6.9 mmol/l) at both ages 31 and 46.
• Among those with FPG <5.0 mmol/l at age 31, 52% progressed to slightly elevated FPG, and 16% developed prediabetes by age 46.
• Using two-class FPG categories, individuals with stable prediabetes (5.6–6.9 mmol/l) had weak or no significant associations with increased probing pocket depths (PPD ≥4 mm) or alveolar bone loss (BL ≥5 mm) compared to normoglycaemia.
• Using the three-class FPG categories:
  • Slightly elevated FPG at both ages was associated with more sites with PPD ≥4 mm (RR 1.7; 95% CI 1.3–2.2) in the total population and RR 3.3 (95% CI 2.4–4.5) in highly educated individuals.
  • Worsened glucose balance from normoglycaemia to prediabetes increased the risk of BL ≥5 mm (RR 1.3; 95% CI 1.0–1.6).
  • Individuals whose glucose balance worsened from normoglycaemia (<5.0 mmol/l) to prediabetes were at greater risk for BL ≥5 mm (RR 1.3; 95% CI 1.0–1.6).
  • Risks for PPD ≥4 mm increased significantly with worsening FPG levels, peaking at RR 2.8 (95% CI 2.0–3.8).
  • When adjusting for potential confounders, including HDL, triglycerides, and alcohol consumption, risk estimates remained largely unchanged, confirming a robust association between glucose dysregulation and periodontal deterioration.
  • Risk estimates were generally lower among highly educated participants compared to the total population, suggesting a potential modifying effect of socioeconomic factors.

Limitations