Determination of Smoking and Obesity as Periodontitis Risks Using the Classification and Regression Tree Method

Background: A model that focuses on personal risk factors associated with poor lifestyle has been proposed for the etiology of generalized periodontitis. Numerous investigations have linked individual lifestyle‐related factors to periodontitis risk; however, a definite relationship among lifestyle‐related factors remains unclear. The objective of this study was to determine which lifestyle‐related factors demonstrated the greater impact on periodontitis risk. Methods: The association of lifestyle‐related factors, such as smoking status and obesity, with periodontitis was assessed in 372 Japanese workers via a self‐administered questionnaire. Smoking status and obesity were evaluated in terms of packyears and body mass index (BMI), respectively. Clinical periodontal examination included probing depth (PD). The effective impact on periodontitis risk was analyzed by the classification and regression tree (CART) method and multiple logistic regression analysis. Results: Simple logistic regression analyses revealed that factors such as age, gender, alcohol consumption, smoking status, BMI, and frequency of toothbrushing were associated with periodontitis. CART results demonstrated a significant correlation between periodontitis and pack‐years, BMI, and age; in contrast, alcohol consumption, gender, and toothbrushing frequency were not correlated with periodontitis. The strongest factor for periodontitis risk was pack‐years of smoking. Additionally, both pack‐years and BMI exhibited clear dose‐response relationships with periodontitis. These relationships were maintained despite adjustment for known confounding factors. Conclusions: Smoking displays the greatest impact on periodontitis among lifestyle‐related factors. Both smoking and obesity are independent risk indicators for periodontitis; moreover, these parameters exhibit a dose‐response relationship with respect to periodontitis risk. J Periodontol 2005;76:923‐928.