Systematic review on risk of cardiovascular disease

• A majority of the included studies reported inadequate confounding adjustment and were deemed to have serious risk of bias.
• No eligible studies evaluated children aged <18 years.
• The association between digestible carbohydrate intake and cardiovascular outcomes was nonlinear and supported by low strength of evidence.
• When digestible carbohydrate intake was analyzed as the percentage of total energy intake, the risk of incident cardiovascular disease (CVD) significantly increased when carbohydrate intake exceeded 65% total energy intake, compared with the carbohydrate intake reference level of 50% total energy intake. The lowest risk was at a carbohydrate intake of 50% total energy intake. The risk of incident coronary heart disease increased starting at 45% total energy intake.
• When digestible carbohydrate intake was analyzed as grams per day, the risk of CVD significantly increased when carbohydrate intake exceeded 300 grams per day of carbohydrates, compared with the reference level of 300 grams of carbohydrates per day. The lowest risk was at a carbohydrate intake of 250 grams per day. The risk of incident coronary heart disease increased starting at 250 grams per day of carbohydrates.
• The risk of CVD-related mortality was U shaped and might be lowest with an intake 250–300 grams of carbohydrates per day.
• The risk of stroke was not significantly associated with carbohydrate intake and had a less defined dose-response relationship. The risk increased when carbohydrate intake exceeded 50% total energy intake.
• The nonlinear relationships were overall similar based on sex or geographic location but with variable carbohydrate intake ranges associated with the lowest risk.

Systematic review on risk of type 2 diabetes, growth, size, and body composition

• A majority of the included studies reported inadequate confounding adjustment and were deemed to have serious risk of bias.
• No eligible studies evaluated children aged <18 years.
• The association between carbohydrate intake and incident type 2 diabetes (T2D) was nonlinear and supported by low strength of evidence.
• When digestible carbohydrate intake was reported as the percentage of total energy intake, the risk of incident T2D gradually decreased from the lowest carbohydrate intake until 45% total energy intake, remained relatively flat between 45% and 55% total energy intake, and increased gradually from 55% total energy intake.
• When digestible carbohydrate intake was reported as grams per day, the risk of incident T2D gradually decreased from the lowest carbohydrate intake until 270 grams per day, remained relatively flat between 270 to 350 grams per day of carbohydrates, and increased gradually from 350 grams per day of carbohydrates.
• The evidence was insufficient to determine an association between carbohydrate intake and weight or body composition.
• The nonlinear associations were overall similar based on sex but with variable carbohydrate intake ranges associated with the lowest risk.
• Very few studies evaluated surrogate outcomes.

Background: Epidemiological studies have shown inconsistent findings regarding the effect of dietary digestible carbohydrate intake on the risk of cardiovascular disease and type 2 diabetes (T2D). Synthesis of such evidence is important for determining the Dietary Reference Intakes (DRI) for carbohydrates, which can have consequences on incidence and morbidity of chronic conditions.

Methods: Two systematic reviews were conducted, one addressing cardiovascular outcomes and the second addressing incidence of T2D, body weight and composition. We searched several databases from January, 2000 to October, 2023. We also conducted gray literature search and reference mining. Eligible studies evaluated the outcomes of interest in healthy individuals over 2 years old and isolated the effect of digestible carbohydrate intake from other macronutrients in grams per day or percent of total energy intake. Random-effects dose-response meta-analyses were conducted when feasible.

Results: The systematic review on cardiovascular outcome included 21 prospective cohort studies with 1,277,621 participants. The majority of the studies reported inadequate confounding adjustment (73%) and were deemed to have serious risks of bias (80%). No eligible studies evaluated children aged <18 years. The association between digestible carbohydrate intake and cardiovascular outcomes was nonlinear and was supported by low strength of evidence. When carbohydrate intake was analyzed as the percentage of total energy intake, the risk of incident cardiovascular disease significantly increased when carbohydrate intake exceeded 65% total energy intake, compared with the carbohydrate intake reference level of 50% total energy intake. The lowest risk was at a carbohydrate intake level of 50% total energy intake. The risk of incident coronary heart disease increased starting at a carbohydrate intake level of 45% total energy intake. When carbohydrate intake was analyzed as grams per day, the risk of incident cardiovascular disease significantly increased when carbohydrate intake exceeded 300 grams per day, compared with the carbohydrate intake reference level of 300 grams per day. The lowest risk was at a carbohydrate intake level of 250 grams per day. The risk of incident coronary heart disease increased starting at 250 grams per day of carbohydrates. The nonlinear relationships were overall similar based on sex or geographic location but with variable intake range associated with the lowest risk. Higher carbohydrate intake was associated with lower levels of high-density lipoprotein- cholesterol (HDL-C) and higher levels of triglycerides.

The systematic review on diabetes and body composition included 17 studies with 463,228 participants. The majority of the studies reported inadequate confounding adjustment (79%) and were deemed to have serious risks of bias (92%). No eligible studies evaluated children aged <18 years. The association between carbohydrate intake and incident T2D was nonlinear and was supported by low strength of evidence. Analyzing carbohydrate intake as a percentage of total energy intake showed a gradual reduction in the risk of incident T2D up to 45% total energy intake. The risk then plateaued between 45% and 55% total energy intake before rising with higher carbohydrate intake levels. Similarly, analyzing carbohydrate intake in grams per day revealed a gradually reduced risk up to 270 grams per day of carbohydrates, followed by a plateau between 270–350 grams per day of carbohydrates and increased risk after 350 grams per day of carbohydrates.

The evidence was insufficient to determine an association between carbohydrate intake and weight or body composition. The nonlinear relationships were overall similar based on sex but with variable intake range associated with the lowest risk. Very few studies evaluated intermediate outcomes.

Conclusion: Dose-response meta-analyses suggest a nonlinear relationship between the intake of digestible carbohydrates and cardiovascular disease and incident T2D. These associations appear to be U-shaped and suggest certain ranges of carbohydrate intake that were associated with the lowest risk. Such ranges can help in establishing future DRI for carbohydrates, which can have important consequences on incidence and morbidity of chronic conditions and public health.