Effect of fructose instead of glucose or sucrose on cardiometabolic markers: a systematic review and metaanalysis of isoenergetic intervention trials

Context: Free, or added, sugars are considered important determinants in the pandemics of obesity and associated chronic diseases, and fructose has emerged as the sugar of main concern. Objective: The aim of this review was to assess the evidence of the effects of isoenergetic replacement of fructose or high-fructose corn syrup (HFCS) for glucose or sucrose on cardiometabolic markers in controlled dietary intervention trials. Data Sources: The electronic databases PubMed/MEDLINE, the Cochrane Library, and Embase were searched from 1980 to May 5, 2020. Study Selection: Studies were eligible if they measured at least one of the following outcomes: total cholesterol, low- and high-density lipoprotein cholesterol, triacylglycerols, apolipoprotein A1, apolipoprotein B, systolic blood pressure, diastolic blood pressure, fasting glucose, and body weight. Data Extraction: For each outcome, the mean values and the corresponding measure of dispersion were extracted after the intervention or control diet. Data Analysis: Fixed-effects and random-effects models were used to pool study-specific estimates. Between-study heterogeneity was assessed by the chi(2) test and the I-2 statistic and publication bias by the Egger test and funnel plots. Results: Twenty-five studies involving 1744 volunteers were identified. No significant effects were found when fructose or HFCS was substituted for glucose, except for a slight decrease in diastolic blood pressure when fructose was substituted for glucose. Similarly, no effects were found when fructose or HFCS was substituted for sucrose, except for a small increase, of uncertain clinical significance, of apolipoprotein B when HFCS was substituted for sucrose. Conclusions: Isoenergetic substitution of fructose or HFCS for glucose or sucrose has no significant effect on most of the cardiometabolic markers investigated; however, some results were affected by residual between-study heterogeneity and studies with high or unclear risk of bias.