Habitual xylitol gum-chewing may have a long-term preventive effect by reducing the caries risk for several years after the habitual chewing has ended. The goal of this report was (1) to determine if sorbitol and sorbitol/xylitol mixtures provide a long-term benefit, and (2) to determine which teeth benefit most from two-year habitual gum-chewing - those erupting before, during, or after habitual gum-chewing. Children, on average 6 years old, chewed gums sweetened with xylitol, sorbitol, or xylitol/sorbitol mixtures. There was a "no-gum" control group. Five years after the two-year program of habitual gum-chewing ended, 288 children were re-examined. Compared with the no-gum group, sorbitol gums had no significant long-term effect (relative risk [RR], 0.65; 95% confidence interval [c.i.], 0.39 to 1.07; p < 0.18). Xylitol gum and, to a lesser extent, xylitol/sorbitol gum had a long-term preventive effect. During the 5 years after habitual gum-chewing ended, xylitol gums reduced the caries risk 59% (RR, 0.41; 95% c.i., 0.23 to 0.75; p < 0.0034). Xylitol-sorbitol gums reduced the caries risk 44% (RR, 0.56; 95% c.i., 0.36 to 0.89; p < 0.02). The long-term caries risk reduction associated with xylitol strongly depended on when teeth erupted (p < 0.02). Teeth that erupted after 1 year of gum-chewing or after the two-year habitual gum use ended had long-term caries risk reductions of 93% (p < 0.0054) and 88% (p < 0.0004), respectively. Teeth that erupted before the gum-chewing started had no significant long-term prevention (p < 0.30). We concluded that for long-term caries-preventive effects to be maximized, habitual xylitol gum-chewing should be started at least one year before permanent teeth erupt.