Exploring the Bidirectional Relationship Between Periodontitis and Rheumatoid Arthritis in a Large Danish Cohort

Abstract

Objective: We investigated the bidirectional relationship between rheumatoid arthritis (RA) and periodontitis and their cross-sectional association using national administrative health care data.

Methods: The sample included 3,308,903 individuals aged 20 to 79 years who resided in Denmark in 2000 and had remained free of RA and periodontitis in the previous 10 years. RA and periodontitis were defined using diagnosis and treatment codes. Marginal structural survival models were employed to estimate the effects of RA on periodontitis incidence and vice versa from 2000 to 2017. Using a cross-sectional sample of 2,574,536 individuals from 2017, the association of periodontitis with RA was investigated using regression analyses and probabilistic quantitative bias analyses, simulating RA and periodontitis misclassification and unmeasured confounding of smoking.

Results: Between 2000 and 2017, 20,348 individuals developed RA and 740,799 developed periodontitis. The estimated hazard ratio (HR) for the effect of periodontitis on incident RA was 1.05 (95% confidence interval [CI] 0.88-1.25), resulting in a restricted mean survival time difference of one day. The HR for the effect of RA on incident periodontitis was 0.84 (95% CI 0.80-0.88), corresponding to a restricted mean survival time difference of 151 days. Cross-sectionally, the unadjusted prevalence ratio for the association was 1.15 (95% CI 1.11-1.19), whereas the estimate adjusted for measured and simulated confounding was practically null (0.99, 95% simulation interval 0.93-1.04).

Conclusion: These findings challenge previously reported bidirectional relationships between periodontitis and RA, pointing to potential residual confounding as an important link and prompting reconsideration of the biologic plausibility and clinical significance of these associations.

Figure 1

Flowchart of sample selection process. PD, periodontitis; RA, rheumatoid arthritis; obs, observations.

Figure 2

(A) Unadjusted probabilities of remaining free from RA by periodontitis status over the 18‐year follow‐up period from 2000 to 2017. (B) Adjusted probabilities of remaining RA free by periodontitis status over the 18‐year follow‐up period from 2000 to 2017. Note the scales of the y‐axes. RA, rheumatoid arthritis.

Figure 3

(A) Unadjusted probabilities of remaining periodontitis free by rheumatoid arthritis status over the 18‐year follow‐up period from 2000 to 2017. (B) Adjusted probabilities of remaining periodontitis free by rheumatoid arthritis status over the 18‐year follow‐up period from 2000 to 2017. Note the scales of the y‐axes.