Spurious associations in oral epidemiological research: the case of dental flossing and obesity


The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs
Philippe P. Hujoel
University of Washington,
Box 357475
Seattle, WA 98195
E-mail: hujoel@u.washington.edu


Background: Individuals with increased oral health awareness may also have increased general health awareness, and vice versa. Such associations between oral and general health awareness has the potential to induce spurious associations in oral epidemiological research.

Objective: To assess the extent to which oral self-care patterns and general health awareness are confounded, we investigated the association between flossing and obesity, two lifestyle factors that are unlikely to be causally related.

Methods: A cross-sectional study of 1497 individuals presenting for an initial periodontal exam by the specialist. Self-reported flossing behaviors and body mass index (BMI) categories were related using logistic regression models.

Results: After adjustment for confounding variables, lack of daily flossing was associated in a dose-dependent way with morbid obesity (odds ratio (OR), 20.3; 95% confidence interval (CI), 2.7–154.0), obesity (OR, 2.1; 95% CI, 1.5–2.9), and being overweight (OR, 1.7; 95% CI, 1.3–2.2). When restricting to never smokers, a significant relationship between obesity and lack of flossing remained.

Conclusion: The strong associations between two causally unrelated oral and general lifestyle characteristics indicate that simplistic epidemiologic methodology is unlikely to provide insights into causal mechanisms of oral diseases or oral-systemic relationships.