Abstract
Background: Diabetes, obesity, gout, hypertension; these are all global epidemics that have been on the rise in recent years. Hard evidence has been produced showing an association of these diseases with the intake of free sugars. The World Health Organisation (WHO) recommends reducing free sugars intakes to less than 10% of total energy intake per day. The current population intakes of this nutrient in New Zealand are unknown as an objective, cost-effective, time-efficient tool to estimate free sugars consumption has not been developed thus far.
Objective: To examine the association between free sugars intake estimated by a brief food frequency questionnaire (FFQ) and free sugars intake estimated by δ13C measurements in red blood cells (RBCs).
Method: We aimed to recruit 120 participants to provide adequate power, however, only 77 individuals aged 20-70 years old were recruited from the Dunedin population in this ten-week observational study. Participants were required to attend a baseline and end-of-study clinic appointment in which anthropometric data, a blood sample and blood pressure were taken and a short, whole-diet FFQ was completed. In weeks 4 and 7, participants were asked to complete an online FFQ specific to free sugars intakes over the previous month. Reliability of the free sugars-specific FFQ was evaluated by comparing sugars estimates from two administrations of the FFQ taken one month apart. Associations between various sugars estimates from the FFQ and δ13C measurements in RBCs were examined using multiple regression analysis, cross-classification, test-retest reliability and by calculating Spearmans correlation coefficients.
Results: Over half of the participants were female (66%) and 58% were of New Zealand European descent. There were no Pacific Islanders and 22% were from countries in Europe, indicating that our population may not be representative of the entire New Zealand population. The mean difference of sugars intakes was significantly lower in the second administration of the FFQ compared to the first for all sugars. The repeatability of the FFQ was weak to moderate as Spearmans correlation coefficients ranged from 0.39-0.53 for total sugars, free sugars, added sugars, sucrose and SSBs. The Spearmans correlation coefficient for δ13C was 0.84, indicating a strong repeatability of the biomarker. Classification into the same tertile was low; ranging from 18.7% for added sugars to 34.7% for Sugar-Sweetened Beverages. Sugars estimates from both instruments were not at all related (Spearmans correlation coefficients for all were <0.2). There was no association between the isotope ratio and β-coefficient plus R2 value for each sugar intake prediction even after adjusting for age, sex and body fat percentage (and including δ15N as a marker of protein intakes).
Conclusion: The lack of association found indicates that either the FFQ or the biomarker, or both, are unable to provide accurate estimates of an individuals’ dietary sugars intakes. Further research is needed to validate both the δ13C biomarker and the FFQ against a more reliable measure of dietary intake such as a 7-day weighed diet record.