Application of survival analysis to carious lesion transitions in intervention trials

Objective: To demonstrate the usefulness of a survival-time regression model for the analysis of data from two 3-year trials of the caries-preventive effect of sugar-substituted chewing gums and fluoride toothpaste, carried out among 892 Lithuanian children. Methods: A caries onset was defined as a transition from sound to carious and a caries recovery was defined as a transition from carious to sound. The time at risk for each type of transition was calculated. Using an exponential survival-time regression model, the hazard ratios for the covariates experimental group (control, sugar substitute, fluoride), age, gender, surface type and posteruptive surface age was estimated. This analysis was repeated using two alternative definitions of the caries transitions. Results: The analyses confirmed that caries rates are higher in occlusal surfaces, and that posteruptive surface age influences caries rates. Moreover, it also confirmed that fluoride affects the outcome of ongoing caries activity more than the initiation of caries. Conclusions: Survival-time analysis of caries transitions allows for the extraction of much more information from caries trials than does the traditional DMF-based analysis, and traditional DMF incremental values may easily be derived from the models.