Effect of Fructose on Established Lipid Targets: A Systematic Review and Meta-Analysis of Controlled Feeding Trials

Abstract

Background: Debate over the role of fructose in mediating cardiovascular risk remains active. To update the evidence on the effect of fructose on established therapeutic lipid targets for cardiovascular disease (low-density lipoprotein cholesterol [LDL]-C, apolipoprotein B, non-high-density lipoprotein cholesterol [HDL-C]), and metabolic syndrome (triglycerides and HDL-C), we conducted a systematic review and meta-analysis of controlled feeding trials.

Methods and results: MEDLINE, EMBASE, CINHAL, and the Cochrane Library were searched through July 7, 2015 for controlled feeding trials with follow-up ≥7 days, which investigated the effect of oral fructose compared to a control carbohydrate on lipids (LDL-C, apolipoprotein B, non-HDL-C, triglycerides, and HDL-C) in participants of all health backgrounds. Two independent reviewers extracted relevant data. Data were pooled using random effects models and expressed as mean difference with 95% CI. Interstudy heterogeneity was assessed (Cochran Q statistic) and quantified (I(2) statistic). Eligibility criteria were met by 51 isocaloric trials (n=943), in which fructose was provided in isocaloric exchange for other carbohydrates, and 8 hypercaloric trials (n=125), in which fructose supplemented control diets with excess calories compared to the control diets alone without the excess calories. Fructose had no effect on LDL-C, non-HDL-C, apolipoprotein B, triglycerides, or HDL-C in isocaloric trials. However, in hypercaloric trials, fructose increased apolipoprotein B (n=2 trials; mean difference = 0.18 mmol/L; 95% CI: 0.05, 0.30; P=0.005) and triglycerides (n=8 trials; mean difference = 0.26 mmol/L; 95% CI: 0.11, 0.41; P<0.001). The study is limited by small sample sizes, limited follow-up, and low quality scores of the included trials.

Conclusions: Pooled analyses showed that fructose only had an adverse effect on established lipid targets when added to existing diets so as to provide excess calories (+21% to 35% energy). When isocalorically exchanged for other carbohydrates, fructose had no adverse effects on blood lipids. More trials that are larger, longer, and higher quality are required.

Clinical trials registration: URL: https://www.clinicaltrials.gov/. Unique Identifier: NCT01363791.

Keywords: lipids; meta‐analysis; nutrition.

© 2015 The Authors. Published on behalf of the American Heart Association, Inc., by Wiley Blackwell.

Figures

Figure 1

Flow of the literature.

Figure 2

Forest plots of the effect of fructose on LDL-C in isocaloric feeding trials. Pooled effect estimates are shown as diamonds. Data are expressed as MD with 95% CI using generic inverse variance random effects models. Interstudy heterogeneity was tested the Cochran Q statistic at a significance level of P<0.10 and quantified by the I2 statistic, where I2≥50% is considered to be evidence of substantial heterogeneity and ≥75% considerable heterogeneity. Any CHO denotes any carbohydrate comparator. HTG indicates hypertriglyceridemic; LDL-C, low density lipoprotein; MD, mean difference.

Figure 3

Forest plots of the effect of fructose on apo B in isocaloric feeding trials. Pooled effect estimates are shown as diamonds. Data are expressed as MD with 95% CI using generic inverse variance random effects models. Interstudy heterogeneity was tested by the Cochran Q statistic at a significance level of P<0.10 and quantified by the I2 statistic, where I2≥50% is considered to be evidence of substantial heterogeneity and ≥75% considerable heterogeneity. Any CHO denotes any carbohydrate comparator. apo B indicates apolipoprotein B; HTG, hypertriglyceridemic; MD, mean difference.

Figure 4

Forest plots of the effect of fructose on non-HDL-C in isocaloric feeding trials. Pooled effect estimates are shown as diamonds. Data are expressed as MD with 95% CI using generic inverse variance random effects models. Interstudy heterogeneity was tested by the Cochran Q statistic at a significance level of P<0.10 and quantified by the I2 statistic, where I2≥50% is considered to be evidence of substantial heterogeneity and ≥75% considerable heterogeneity. Any CHO denotes any carbohydrate comparator. HDL-C indicates high density lipoprotein; HTG, hypertriglyceridemic; MD, mean difference; T2DM, type 2 diabetes mellitus.

Figure 5

Forest plots of the effect of fructose on triglycerides in isocaloric feeding trials. Pooled effect estimates are shown as diamonds. Data are expressed as MD with 95% CI using generic inverse variance random effects models. Inter-study heterogeneity was tested by the Cochran Q statistic at a significance level of P<0.10 and quantified by the I2 statistic, where I2≥50% is considered to be evidence of substantial heterogeneity and ≥75% considerable heterogeneity. A, B refers to study A and study B (two separate trials) within the same report. Any CHO denotes any carbohydrate comparator. HTG indicates hypertriglyceridemic; MD, mean difference.

Figure 6

Forest plots of the effect of fructose on HDL-C in isocaloric feeding trials. Pooled effect estimates are shown as diamonds. Data are expressed as MD with 95% CI using generic inverse variance random effects models. Interstudy heterogeneity was tested by the Cochran Q statistic at a significance level of P<0.10 and quantified by the I2 statistic, where I2≥50% is considered to be evidence of substantial heterogeneity and ≥75% considerable heterogeneity. Any CHO denotes any carbohydrate comparator. HDL-C indicates high density lipoprotein; HTG, hypertriglyceridemic; MD, mean difference.

Figure 7

Forest plots of the effect of fructose on apo B in hypercaloric feeding trials. Pooled effect estimates are shown as diamonds. Data are expressed as MD with 95% CI using generic inverse variance random effects models. Interstudy heterogeneity was tested by the Cochran Q statistic at a significance level of P<0.10 and quantified by the I2 statistic, where I2≥50% is considered to be evidence of substantial heterogeneity and ≥75% considerable heterogeneity. Any CHO denotes any carbohydrate comparator. apo B indicates apolipoprotein B; MD, mean difference.

Figure 8

Forest plots of the effect of fructose on triglycerides in hypercaloric feeding trials. Pooled effect estimates are shown as diamonds. Data are expressed as mean difference with 95% CI using generic inverse variance random effects models. Interstudy heterogeneity was tested by the Cochran Q statistic at a significance level of P<0.10 and quantified by the I2 statistic, where I2≥50% is considered to be evidence of substantial heterogeneity and ≥75% considerable heterogeneity. Any CHO denotes any carbohydrate comparator.