Date of Degree
PhD (Doctor of Philosophy)
Steven M. Levy
Dental caries is a common childhood disease and important health problem in the United States and throughout the world. Most studies that have assessed risk factors for dental caries focused on non-modifiable risk factors such as previous caries experience and socioeconomic status. It is also important to investigate modifiable risk factors that can be used in developing guidelines for risk assessment and prevention. The present dissertation assessed mainly the associations between dental caries and modifiable factors, including dietary factors, water fluoride levels and toothbrushing frequency in children, while adjusting for non-modifiable factors. Data were obtained from subjects who were participants in the Iowa Fluoride Study. Dietary data were collected using 3-day dietary diaries from 1.5 months to 8.5 years and detailed questionnaires from 9 years to 13 years. Dental caries examinations were conducted at about 5, 9 and 13 years of age. There are three main analyses.
The first analysis assessed risk factors for a 4 group primary dentition caries experience variable: the caries-free (reference group), the d1, the d2+f, and the d1d2+f groups. The dietary consumption frequencies (from ages 3 to 5 years) for the children in the 4 caries groups were compared using multivariable multinomial regression analyses. Lower consumption frequency of milk at meals and greater consumption frequency of pre-sweetened cereal at meals significantly increased the likelihood of being in the d1 group. Greater consumption frequency of regular soda pop at snacks significantly increased the likelihood of being in the d1d2+f group. Greater consumption frequency of added sugars at snacks significantly increased chance to be in the d2+f group and the d1d2+f group.
The second manuscript assessed risk factors for new mixed dentition cavitated caries determined based on surface-specific transitions from the primary to mixed dentition exams on 16 teeth using logistic regression analysis. Greater consumption frequency of processed starch at snacks significantly increased the likelihood of having new cavitated caries (p = 0.04 for the model excluding previous caries experience).
The third manuscript used negative binomial regression with the Generalized Linear Mixed Models procedure to assess separately the longitudinal associations of 1) new non-cavitated caries and 2) new cavitated caries with modifiable risk factors. Surface-specific counts of new non-cavitated caries and cavitated caries at each of the primary, mixed and permanent dentition examinations were used as outcome variables. Greater consumption frequency of 100% juice was significantly associated with fewer non-cavitated and fewer cavitated caries surfaces.
In this study, some factors were associated with caries at one age only, while others were associated with caries across childhood. Consumption of foods or beverages at meals generally decreased their cariogenicity. Previous caries experience is strongly associated with other independent variables in the regression models that examined risk factors for new cavitated caries. Thus modifiable factors that usually have weaker associations with caries might not be retained in the models due to collinearity issues. Future researchers are encouraged to present results both ways so that scientific communities can best interpret the complex results. Also, repeated measures analysis might be more appropriate for variables that are common in all age groups, such as toothbrushing frequency and fluoride exposures. More studies of the complex relationships between diet and caries are needed, including additional studies that place more emphasis on investigation of modifiable risk factors for both non-cavitated and cavitated caries.
Copyright 2010 Oitip Chankanka