Linear response of mutans streptococci to increasing frequency of xylitol chewing gum use: a randomized controlled trial [ISRCTN43479664]

BACKGROUND Xylitol is a naturally occurring sugar substitute that has been shown to reduce the level of mutans streptococci in plaque and saliva and to reduce tooth decay. It has been suggested that the degree of reduction is dependent on both the amount and the frequency of xylitol consumption. For xylitol to be successfully and cost-effectively used in public health prevention strategies dosing and frequency guidelines should be established. This study determined the reduction in mutans streptococci levels in plaque and unstimulated saliva to increasing frequency of xylitol gum use at a fixed total daily dose of 10.32 g over five weeks. METHODS Participants (n = 132) were randomized to either active groups (10.32 g xylitol/day) or a placebo control (9.828 g sorbitol and 0.7 g maltitol/day). All groups chewed 12 pieces of gum per day. The control group chewed 4 times/day and active groups chewed xylitol gum at a frequency of 2 times/day, 3 times/day, or 4 times/day. The 12 gum pieces were evenly divided into the frequency assigned to each group. Plaque and unstimulated saliva samples were taken at baseline and five-weeks and were cultured on modified Mitis Salivarius agar for mutans streptococci enumeration. RESULTS There were no significant differences in mutans streptococci level among the groups at baseline. At five-weeks, mutans streptococci levels in plaque and unstimulated saliva showed a linear reduction with increasing frequency of xylitol chewing gum use at the constant daily dose. Although the difference observed for the group that chewed xylitol 2 times/day was consistent with the linear model, the difference was not significant. CONCLUSION There was a linear reduction in mutans streptococci levels in plaque and saliva with increasing frequency of xylitol gum use at a constant daily dose. Reduction at a consumption frequency of 2 times per day was small and consistent with the linear-response line but was not statistically significant.