Over two thousand salivary rinse samples were collected during the 1986- 87 national survey of U.S. schoolchildren. Specimens collected with this technique were plated on three selective media: MSB, GSTB and TSY20B. Counts of mutans streptococci were determined on all three media. Mean DMFS were computed for all children aged 7 to 17 who had rinse samples. 1109 specimens had readable plate counts for mutans on all three media. In over one-fifth of the samples from each media, no CFU of mutans were detected. Mean mutans counts (CFU/ml of salivary rinse) were approximately 1.5 x 10-4 for MSB, 1.2 x 10-4 for TSY20B and 1.7 x 10-4 for GSTB. Repeated measures ANOVA indicated significant differences (p equals less than.0001) between counts on different media. However, correlations between media were high and differences were normally in the same direction with GSTB highest and TSY20B lowest. Associations with dental caries were measured for all three media. Mean CFU for those with zero DMFS were 30 to 40% lower than for children with 1 or more DMFS. Mean DMFS scores were 50% lower for groups with zero mutans counts versus those with any mutans for all three media. (approximately 2 DMFS for 0 mutans counts and 4 DMFS for those with mutans). Odds ratios for the association between mutans and caries were approximately 2.2 for all three media, i.e. children with mutans were about twice as likely to have caries as those with no mutans (p is less than .0001). Counts from all three media were able to differentiate between groups of children with or without dental caries reflecting the strong association between dental caries and mutans and making the salivary rinse a viable collection technique.
