Table 6 Cross-cutting study design considerations

• Future research should use experimental conditions that approximate realistic intake levels and chemical forms.

• There is a need to improve dietary assessment tools and methodology because of ongoing concerns regarding systematic biases and random errors in dietary surveys or dietary assessment methods. The inadequacy of current methods that can potentially invalidate the scientific findings has been discussed elsewhere [33–36].

• Controlled feeding studies cannot answer population questions.

• Statisticians or researchers with adequate statistical expertise should be involved in a study to ensure that appropriate statistical methods are used to control for confounders and to adjust for random measurement errors in dietary intake assessments.

• The effect of fructose intake on blood lipids has been well established [8, 37–39]; therefore there is no research need for this question. This is more of a biological mechanism question than a public health question.

• Diet quality could be an intermediate outcome for short- or long-term (more than 1 year) clinical outcomes.

• Diet quality and meal patterns may modify the response to dietary sugars and may need to be considered as a parameter for all studies.

• There are currently no biomarkers for sugars intake and appetite that are considered “gold standards.”

• Body composition can change independent of weight and should be measured whenever possible.

• Current tools to measure adiposity are either invasive or very costly.

• Exercise, fitness, and levels of other forms of physical activity will likely modify many responses to dietary sugars and should be considered in all studies.

• Physical activity must be measured objectively (e.g., accelerometry based activity monitors) in clinical trials and observational studies.

• All domains of physical activity (e.g., occupational, household, transport, leisure-time, exercise, etc.) must be measured on the population level. The measurement of one or two domains (e.g., leisure-time and exercise) is inadequate.

• Individual genetics may modulate effects of sugars, resulting in heterogeneity of results

• There might be epigenetic and microbiome modulation of effects.

• There is a potential for the effects of sugars on the gut microbiome to mediate biological effects of sugars.

• There is evidence for the hypothesis that circadian timing of intake modulates biological effect.