The Impact of Sugar Source on the Relationships Between Free Sugars Intake and Health: A Secondary Analysis.

BACKGROUND/

OBJECTIVES

This secondary and exploratory meta-analysis re-evaluated 30 randomized controlled trials on free and added sugars (FS) detailed in the European Food Safety Authority's (EFSA) report on the tolerable upper intake level for dietary sugars, focusing on the influence of food source (beverages, foods, or mixed) on cardiometabolic and anthropometric health.

METHODS

The EFSA's method of analyzing the relative FS intake (difference between treatment and comparator arms, Δ%Efs) was used, with further adjustment for the reported intake of all sources of FS and energy. The EFSA's "high vs. low" random-effects meta-analysis comparing groups with the highest and lowest FS intake was replicated, and additional exploratory dose-response meta-regressions (linear and non-linear) were performed, stratified by food source. Given the secondary and observational nature of the analysis, all source-stratified findings should be interpreted as hypothesis-generating, rather than causal.

RESULTS

There were no interactions between Δ%Efs and food source for any outcome, and within a source there were linearly positive and statistically significant regressions for body weight (mixed), low-density lipoprotein cholesterol (LDL-C, foods), and uric acid (beverages). Across 13 outcomes, Δ%Efs was positively and linearly related to greater fasting glucose, high-density lipoprotein cholesterol (HDL-C), and LDL-C, and non-linearly to body weight. However, the data were limited in their representation of FS intake at typical population levels, and there were insufficient data to investigate the effect of FS from foods on most anthropometric outcomes.

CONCLUSIONS

Meta-regressive dose-responses revealed little relationship between Δ%Efs from specific food sources and health outcomes, but such effects might be masked by confounding factors. Future trials that test realistic intakes of FS across diverse food matrices and account for dietary compensation would help to overcome limitations in the body of evidence.