The effect of 2-year chewing-gum use on the caries rates of primary teeth was studied in a combined school and home program in a sample of 510 initially 6-year-old subjects with high caries experience, low availability of fluoride, and difficult access to dental care. The gum, formed into either sticks or pellets, comprised either xylitol, sorbitol, or mixtures thereof. The gum was chewed for 5 min under supervision five times a day during the school year, and for variable times during nonschool days. Seven groups were studied. One group received no gum; two xylitol gum groups received either pellet or stick gum as did, two sorbitol gum groups, and two groups received either of two types of xylitol/sorbitol pellet gum. The response variable was the development of a frank carious lesion detectable by physical loss of enamel and probable extension to the dentin for those surfaces of primary teeth that were not cavitated at baseline. Caries rates associated with the use of each of the gum types were compared to the caries rates in the no-gum group. The usage of all polyol gums resulted in a significant decrease of the caries onset rate (p < 0.05). The caries onset risk for a primary surface in the xylitol pellet and the sorbitol pellet groups was 35 and 44% of that in the no-gum group (relative risk, 0.35; 95% confidence interval, 0.21-0.59; relative risk, 0.44; 95% confidence interval, 0.30-0.63, respectively). The caries onset risk in the xylitol stick gum group was 53% of that in the no-gum group (relative risk, 0.53; 95% confidence interval, 0.39-0.72), which was marginally (p = 0.1520) lower than in the sorbitol stick gum group (relative risk, 0.70; 95% confidence interval, 0.52-0.94). The usage of both xylitol/sorbitol mixtures in pellet form was associated with a caries onset rate comparable with the usage of the xylitol stick gum. The largest caries risk reduction was observed in the group receiving xylitol pellet gum.