To propose an approach for investigation of the relationship between the frequency of carbohydrates intake and dental caries in real-life conditions.

The frequency of separate eating events (with the exclusion of lunch and dinner) and their sugars and starch content were assessed on the basis of diaries collected for seven consecutive days. The total number of decayed, missing and filled teeth (DMFT) was assessed through clinical examination. The interrelation between DMFT and frequency of eating events was analysed through Pearson correlation coefficient and stepwise forward linear regression analysis.

Italy.

One hundred and ninety-three students (males and females, mean age 16 years) of a secondary school.

Individual DMFT varied between 0 (24% of subjects) and 12, with a mean of 3. Once lunch and dinner were excluded, the mean number of separate eating events was 2.9. A statistically significant relationship was found between DMFT and eating frequency thus defined. Correlations were calculated considering only eating events containing a proportion of sugars or starch higher than different cut-off levels. In a stepwise multiple regression model for DMFT, the frequency of ‘high sugars and high starch events’ accounted for 8% of the DMFT variance. Overall frequency of separate eating events defined according to their content of both sugars and starch accounted for 18% of the DMFT variance. A stronger correlation was found among males only.

The precise characterisation of eating events in terms of their sugars and starch content allows us to explain a consistent percentage of the variability in DMFT.