Background: Periodontitis is a major cause of tooth loss globally. Risk factors include age, smoking, and diabetes. Intake of specific nutrients has been associated with periodontitis risk but there has been little research into the influence of overall diet, potentially more relevant when formulating dietary recommendations.
Objectives: We aimed to investigate potential associations between diet and periodontitis using novel statistical techniques for dietary pattern analysis.
Methods: Two 24-h dietary recalls and periodontal examination data from the cross-sectional US NHANES, 2009-2014 (n = 10,010), were used. Dietary patterns were extracted using treelet transformation, a data-driven hierarchical clustering and dimension reduction technique. Associations between each pattern [treelet component (TC)] and extent of periodontitis [proportion of sites with clinical attachment loss (CAL) ≥ 3 mm] were estimated using robust logistic quantile regression, adjusting for age, sex, ethnicity, education level, smoking, BMI, and diabetes.
Results: Eight TCs explained 21% of the variation in diet, 1 of which (TC1) was associated with CAL extent. High TC1 scores represented a diet rich in salad, fruit, vegetables, poultry and seafood, and plain water or tea to drink. There was a substantial negative gradient in CAL extent from the lowest to the highest decile of TC1 (median proportion of sites with CAL ≥ 3 mm: decile 1 = 19.1%, decile 10 = 8.1%; OR, decile 10 compared with decile 1: 0.67; 95% CI: 0.46, 0.99).
Conclusions: Most dietary patterns identified were not associated with periodontitis extent. One pattern, however, rich in salad, fruit, and vegetables and with plain water or tea to drink, was associated with lower CAL extent. Treelet transformation may be a useful approach for calculating dietary patterns in nutrition research.
Keywords: NHANES; diet; periodontitis; robust regression; treelet transformation.
© The Author(s) 2020. Published by Oxford University Press on behalf of American Society for Nutrition.