LAUSR

BACKGROUND This study aimed to examine whether modifiable risk factors can predict tooth loss over 10 years and estimate population attributable risk (PAR) for a combination of modifiable factors. METHODS This longitudinal study included 1,466 participants who underwent dental examinations in 2007 and 2017 and were aged 40-79 years at baseline. Periodontal conditions were assessed using the 2018 periodontal classification. Incident tooth loss was defined as ≥4 teeth lost over a 10-year period. We calculated the partial PAR (pPAR%) for tooth loss to estimate the combined effect of modifiable risk factors. RESULTS Incidence of tooth loss was 17.5%. Directed acyclic graphs were used to identify risk factors for tooth loss. A logistic regression model showed that baseline periodontitis, dental caries experience, no regular dental visit, periodontal treatment, smoking, and obesity were associated with tooth loss after adjusting for covariates; pPAR% was 55.5% (95% confidence interval: 31.1-73.0%) in periodontitis stage III-IV and 87.6% (50.4-97.4%) in the combination of all factors, respectively. The sex-stratified analysis showed that smoking and no regular dental visit in men and obesity in women were identified as potential risk factors for tooth loss. CONCLUSIONS Modifiable factors accounted for most cases of incident tooth loss. Risk factors for tooth loss might differ by sex, suggesting that the appropriate approach for preventing tooth loss base on sex. This article is protected by copyright. All rights reserved.