The development and progression of carious lesions were monitored clinically and radiographically for 1 year in caries-free and high-caries children resident in non-fluoridated (NF) and fluoridated (F) areas. The dmfs/DMFS increment score and the transition score systems were used to monitor caries activity. Ninety-nine variables including exposure to fluoride, nutritional, dietary, salivary, oral hygiene, socioeconomic, and demographic factors were examined longitudinally.
The transition score system was found to be a more sensitive method of monitoring caries activity than the dmfs/DMFS increment score system. Bivariate and multivariate analyses showed that the absence of water fluoridation, past caries experience, lower levels of fluoride intake from diet and toothpaste, frequent use of syrup medications, lower socio-economic class, intake of cariogenic foods at bedtime, increased intake of niacin derived from fast foods, increased intake of fat derived from bakery products, fast foods, and snack foods, and increased intake of sucrose derived from bakery products were associated with high-caries activity. Factors identified to be associated with low-caries activity were parental assistance with brushing, higher levels of fluoride intake from diet and toothpaste, increased intake of maltose derived from flavoured milk drinks, increased intake of fat and sucrose from dairy products, increased intake of fluoride from food and drinks in the NF areas, increased intake of Vitamin B6 from cereals, and an increased flow rate of mixed saliva. The caries prediction models based on some of these factors produced a sensitivity level in excess of the recommended level of 0.75 but a specificity level below the recommended level of 0.85. Both the sensitivity and specificity levels were higher than those found in other caries risk assessment studies.
The high sensitivity level of the prediction model indicated that the combination of risk factors identified were suitable to predict which children would show high-caries activity in the future. However, low specificity levels meant that all the factors that predicted which children would remain caries-free were not identified. Longitudinal studies aimed at identifying risk factors only in that group of children with no caries at baseline are required if specificity of caries prediction models is to be higher.
