Smokeless Tobacco Use and Periodontal Disease in Rural Appalachia

Objective: To determine the relationship between the risk factors of smokeless tobacco (ST) use and periodontal disease by controlling for confounding factors. Methods: A total of 264 smokeless tobacco users and 116 non-users were recruited during one year in two of Ohio's rural Appalachian counties. One calibrated dental examiner obtained clinical data on each subject. Subjects also participated in a structured interview. Subjects were considered to have periodontal disease with a minimum average attachment loss of 1mm and an average gingival bleeding index score greater than 1. A total of 41 subjects (36 ST) had periodontal disease. Logistic regression was used to test the likelihood of having periodontal disease among smokeless tobacco users compared to non-users. The logistic regression model was adjusted for confounders and effect modifiers. The method of fractional polynomials was used to test the assumption of linearity in the logit. Results: Age (p<0.001) and plaque index (p<0.001) were the only covariates that confounded the relationship between periodontal disease and the risk factor of smokeless tobacco use in the logistic regression model. However, the method of fractional polynomials indicated that age was not linear in the logit, while plaque index was linear. After adjusting for age (dichotomized at 45 years) and plaque index, ST users had a 5.42 times (95% CI: 1.81 to 16.28) greater odds of experiencing periodontal disease as compared to individuals who had never used ST. The Hosmer-Lemeshow goodness of fit test indicated that the model was a good fit to the data (p-value = 0.9551). The area under the receiver operating characteristic curve was 0.7931, indicating that the model demonstrated good sensitivity discrimination. Conclusion: ST users have a great odds of having periodontal disease than non-users, after adjusting for age and plaque index. Supported in part by NIH grants: R01-DE013926, M01-RR00034 and T32-DE14320.