Mayo Clin Proc 2019; 94(12):2399-2414; https://doi.org/10.1016/j.mayocp.2019.05.034
Khan TA, Mobushra T, Agarwal A, et al.
Objective
- To determine the association of total and added fructose-containing sugars on cardiovascular disease (CVD) incidence and mortality.
Background
- Much of the evidence supporting recommendations and policies to set upper-limits for added or free fructose-containing sugars derives from data on sugar-sweetened beverages (SSBs), the largest source of fructose-containing sugars in the United States and Canadian diets.
- Although some systematic reviews and meta-analyses of prospective cohort studies have shown adverse associations of SSBs with weight gain, diabetes, and hypertension, the same has not been seen for the fructose-containing sugars they contain.
- Whether the reported adverse associations of SSBs with CVD hold for different fructose-containing sugars independent of food form is unclear.
- The authors of this article conducted a systematic review and meta-analysis of prospective cohort studies using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system to determine the role of total or added fructose-containing sugars in the development of CVD.
Methods
- The Cochrane Handbook for Systematic Reviews of Interventions was followed to conduct this systematic review and meta-analysis. Results were reported according to the MOOSE (Meta-Analysis of Observational Studies in Epidemiology) and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.
- The authors conducted systematic searches in MEDLINE, EMBASE, and Cochrane Library databases January 1, 1980 through July 31, 2018 with restrictions for prospective cohort studies according to a pre-validated list. Targeted manual searches and reference lists of relevant articles served to supplement database searches.
- Prospective cohort studies in humans investigating the association between reported intakes of fructose-containing sugars (total sugars, fructose, sucrose, and added sugars) and CVD (incidence and mortality) in people free of the disease at baseline were included.
Findings
- Eligibility criteria were met by 24 prospective cohort comparisons (624,128 unique individuals;
11,856 CVD incidence cases and 12,224 CVD mortality cases).
- Total sugars, sucrose, and fructose were not associated with CVD incidence. There were no data available for the association of added sugars with CVD incidence.
- Total sugars and fructose were associated with increased CVD mortality, while sucrose was associated with decreased risk. There was no association found for added sugars.
- Dose-response analyses showed a beneficial linear dose-response gradient for sucrose and nonlinear dose-response thresholds for harm for total sugars (133 grams, 26% energy), fructose (58 grams, 11% energy) and added sugars (65 grams, 13% energy) in relation to CVD mortality (P<.05).
- It is important to note that the certainty of the evidence using GRADE was very low for CVD incidence and low for CVD mortality for all sugar types.
Conclusions
- The authors report that their findings suggest a higher-dose threshold for harmful association for added sugars compared with the current recommendations, implying that different food sources may differently contribute at different doses.
- More research is needed to assess whether the association of linear dose-response thresholds seen for SSBs in the literature hold across other important food sources of sugars such as grain and grain-based products, fruit and fruit products, and sweetened dairy and dairy products.
Points to Consider
- The prospective cohort studies included in this review were observational. Therefore, causality cannot be established and there is the possibility of measured and unmeasured residual confounding.
- Dietary consumption in the included studies was self-reported, and thereby subject to several bias including misreporting and underestimations.
- As noted in the publication, the overall quality and strength of the evidence included in this review ranged from “low” to “very low”. Weaknesses of the studies included inconsistency, indirectness, imprecision and publication bias.