Diabetes as a prognostic factor in periodontitis

Background + Aims

• Periodontitis is a chronic inflammatory disease causing tissue destruction and tooth loss, often requiring lifelong management.
• Regular maintenance, termed supportive periodontal therapy (SPT), plays a crucial role in preventing disease progression.
• The World Workshop (2017) introduced a multidimensional classification for periodontitis, categorising cases into stages (severity) and grades (rate of progression).
• This study aims to evaluate how the regularity of maintenance visits impacts tooth loss related to periodontitis (TLP) based on staging and grading.
• It hypothesises that higher disease severity (stages III/IV) and rapid progression (grade C) increase susceptibility to tooth loss, especially with inconsistent maintenance.

Materials + Methods

• The study reviewed records from 442 patients (11,125 teeth) treated for periodontitis at the University of Michigan over at least 10 years.
• This study applied a time-dependent covariate approach to minimise bias from confounding factors. The analysis focused on how prior compliance influenced tooth loss, rather than future attendance patterns.
• Patients were stratified into:
  • Stages I-IV based on disease severity
  • Grades A-C based on progression risk
• Teeth were categorised as either test or control. 
  • Test teeth included those lost due to periodontitis during follow-up, with the preceding five years analysed. 
  • Control teeth were associated with random five-year intervals not linked to tooth loss. 
  • Each tooth contributed only one time period, and intervals were selected to ensure they did not overlap with any TLP events within the same patient. 
  • In cases where a patient experienced multiple TLP events, all were analysed. For patients with no recorded tooth loss, a random five-year control period was selected.
• Several maintenance-related variables were assessed, including the total number of visits over the five-year periods, variation in visit frequency (calculated using the coefficient of variation, CV), and visit patterns categorised as increasing (uphill), consistent (stable), decreasing (downhill), or irregular. 
  • The study introduced CV to periodontal research as a novel way to measure variability in maintenance visits.
  • Patients were grouped based on annual attendance gaps:
    • Those who attended every year (no gaps)
    • Those with one missed year (one gap)
    • Those with multiple missed years (multiple gaps).
• Multi-level logistic regression with generalised estimation equations (GEE) was used to account for the clustering of multiple teeth within patients. 
  • Predictor variables included both patient-level factors, such as age and disease classification, and tooth-specific factors, including probing pocket depth (PPD) and bleeding on probing (BoP). 
• A post hoc power analysis indicated that the study had 88% power to detect meaningful associations, based on the inclusion of 11,125 teeth.

Results

• The study analysed 11,125 teeth from 442 patients undergoing SPT) for a mean follow-up period of 22.7 years. 
• Patients were evenly split by gender, with a mean age of 47.5 years. 
• In total, 554 teeth (4.98%) were lost due to periodontitis in 205 patients. 
  • Tooth loss progression (TLP) was more prevalent in later stages and grades of periodontal disease, with 223 teeth lost within 10–20 years and 147 after 20 years.
• Logistic regression using generalised estimation equations (GEE) showed that TLP risk was significantly associated with stage III and IV disease, grade C classification, older age, and a higher coefficient of variation (CV) in maintenance visits. Each 1% increase in CV corresponded to a 1% increase in TLP risk. However, the number of maintenance visits (NMV) alone did not significantly affect TLP risk. Visit patterns (uphill, downhill) were not associated with TLP in stages II-IV or grades A-C. For stages I-II and grade A patients, analysis was limited by the small number of TLP events.
• Missing SPT visits also impacted TLP. The presence of multiple gaps (two or more years without visits) was associated with increased TLP risk in stages III-IV and grade C patients. Older age correlated with a higher likelihood of TLP, with the risk increasing as the number of missed visits accumulated.
• Subgroup analysis revealed that CV, age, and PPD after APT were significant predictors of TLP in stage III-IV and grade C patients. Adjusting the models for smoking and diabetes indicated that diabetic patients had a 4.69-fold higher risk of TLP compared to normoglycaemic patients.

Limitations