Impact of periodontitis on mortality

Background + Aims

• Periodontitis is associated with several systemic conditions, including diabetes and cardiovascular disease. Shared risk factors, inflammation and microbial pathways may contribute to these relationships, but associations with mortality do not establish causation.
• This prospective cohort study investigated associations between baseline periodontal status and subsequent all-cause and cause-specific mortality in US adults.

Materials + Methods

• The study used six National Health and Nutrition Examination Survey (NHANES) cycles from 1999–2014, linked to National Death Index mortality records through 31 December 2019. Follow-up began at the periodontal examination and continued until death or the end of follow-up.
• Of 24,958 eligible participants, 15,030 were included after exact matching on age and sex: 7,515 in the periodontal-disease group and 7,515 controls. Follow-up was approximately nine years.
• Trained examiners recorded clinical attachment loss and probing depths. Examination protocols differed across survey cycles, including partial- and full-mouth assessments. Periodontitis was classified using CDC/AAP surveillance definitions.
• The periodontal-disease group comprised 4,872 participants with mild/moderate disease and 2,643 with severe disease or edentulism (no natural teeth). Edentulous participants were combined with severe periodontitis to increase statistical power; tooth loss was not necessarily caused by periodontitis.
• Cox proportional hazards models assessed all-cause mortality. Fine–Gray models assessed cause-specific mortality while accounting for competing deaths from other causes.
• Models adjusted for demographic, socioeconomic and clinical factors, smoking, alcohol use, calorie intake and comorbidity history. Missing covariate data were handled using multiple imputation. Sensitivity analyses examined alternative cohorts, examination protocols, smoking measures and exclusion of edentulous participants.
• Results are reported as hazard ratios (HRs) or sub-distribution hazard ratios (SHRs), with 95% confidence intervals (CIs). These relative estimates are not percentage-point increases in an individual’s probability of death.

Results

• All-cause mortality was higher in the overall periodontal-disease group after adjustment: HR 1.22 (95% CI 1.12–1.31), corresponding to a 22% higher relative hazard.
• For mild/moderate periodontitis, the all-cause estimate was not statistically significant: HR 1.04 (95% CI 0.95–1.15). For severe periodontitis/edentulism, the adjusted HR was 1.37 (95% CI 1.26–1.50).
• Cardiovascular mortality was associated with the overall periodontal-disease group (SHR 1.21; 95% CI 1.03–1.41) and severe periodontitis/edentulism (SHR 1.38; 95% CI 1.16–1.64). The mild/moderate estimate was not significant (SHR 0.98; 95% CI 0.81–1.19).
• Diabetes-related mortality was not significantly associated with the overall periodontal-disease group after adjustment (SHR 1.25; 95% CI 0.83–1.89). The severe periodontitis/edentulous group had an adjusted SHR of 1.59 (95% CI 1.02–2.49), as reported in the detailed Results and Figure 4.
• Severe periodontitis/edentulism was associated with chronic lower respiratory mortality (SHR 1.62; 95% CI 1.07–2.45) and malignancy-related mortality (SHR 1.28; 95% CI 1.06–1.56). Neither association was statistically significant in the mild/moderate group.
• Influenza/pneumonia mortality showed a different pattern: the adjusted association was significant for mild/moderate periodontitis (SHR 2.34; 95% CI 1.25–4.37), but not severe periodontitis/edentulism (SHR 1.30; 95% CI 0.69–2.45).
• There was no statistically significant association between the overall periodontal-disease group and Alzheimer’s-related mortality (SHR 0.87; 95% CI 0.60–1.27) or renal mortality (SHR 1.26; 95% CI 0.76–2.09).
• Sensitivity analyses broadly supported the direction of the principal findings. Persistence after adjustment does not exclude residual confounding or demonstrate that periodontal disease caused the deaths.

Limitations

• This observational study cannot establish causation or determine whether periodontal treatment reduces mortality. Shared risk factors and unmeasured or incompletely measured confounders may explain part of the associations.
• Combining edentulism with severe periodontitis limits interpretation and comparison with conventional periodontal case definitions. Tooth loss may result from conditions other than periodontitis, although excluding edentulous participants produced a similar direction of effect in sensitivity analyses.
• Periodontal status was assessed at baseline only. Subsequent treatment, disease progression or new disease in controls was not captured. Comorbidities could also change during follow-up.
• Periodontal examination protocols varied across survey cycles, creating potential differences in disease ascertainment. Sensitivity analyses addressed this issue but cannot remove all measurement limitations.
• Medical histories and several other covariates were self-reported, introducing possible recall error and underreporting. Some cause-specific outcomes had few events and wide confidence intervals.
• The study used broad underlying-cause-of-death categories and did not distinguish individual cancer types or all specific cardiovascular diagnoses. Findings may not generalise to populations outside the US.

Conclusion

• Baseline periodontal disease, particularly the combined severe periodontitis/edentulous category, was associated with higher all-cause mortality and selected cause-specific mortality outcomes after adjustment. Mild/moderate periodontitis was not significantly associated with all-cause mortality.
• The findings support considering oral health within a person’s wider health assessment. They do not show that periodontal treatment extends life or prevents deaths; intervention studies are needed to investigate those questions.