Problems With Relevance, Certainty, and Misinterpretation of Empirical Evidence in the Umbrella Review “Health Effects of Drinking 100% Juice”

We read with interest the recent umbrella review by Beckett et al.¹ The authors suggest there are more potential health benefits than risks associated with 100% juice consumption and argue the evidence does not support the exclusion of 100% juice from dietary guidelines. We do not consider these conclusions to be methodologically sound or to appropriately reflect the totality of the data reported.

First, the health outcomes reported are not those most relevant for juice. Concerns around mortality, cardiovascular disease, or other included outcomes are not driving the exclusion of juice from dietary guidelines in countries that have recently done so. As the authors acknowledge, this has been prompted by the contribution of juice to intake of free sugars, acid, and energy for weight and oral health, as well as their unfavorable direct comparison to whole fruits.¹ The inclusion of irrelevant outcomes and exclusion of important outcomes (eg, dental erosion) limit the review’s use for informing dietary guidelines.

This focus on less relevant outcomes may, at least in part, be linked to study inclusion/exclusion criteria. The authors concede in the discussion that definitions, descriptions, and interpretations (including in self-reported intake) of 100% juice vary, making this distinction difficult to robustly maintain in practice. Decisions made in managing this issue could have led to the exclusion of useful studies evaluating relevant outcomes as well as the inclusion of immaterial or ostensibly ineligible studies. This potential remains difficult to assess.

Second, we have concerns about the interpretation of the evidence presented. High and unexplained heterogeneity for many outcomes (shown by high I² values) imply that these outcomes might vary significantly for different types of juices (eg, due to pulp and nutrition content, fruit type, processing, and storage) and/or for different populations (most likely by age and nutritional status). The article fails to report this significant limitation. Although the evidence is rated almost exclusively as very low to low certainty (GRADE [Grading of Recommendations Assessment, Development, and Evaluation]), the implications of this assessment are not reflected in their conclusion; limited or very little confidence in the findings would be more appropriate to highlight in their subsequent synthesis. The broad generalizations made in drawing conclusions from the empirical data obfuscate important nuances in nutrition science.

Third, the authors conclude that consumption at “low to moderate levels… can provide exposure to beneficial nutrients and bioactive compounds, without the risks associated with excess consumption of free sugars and calories” and “100% juice can be incorporated into a healthy, balanced diet”. These statements are not supported by their study.

Finally, we have significant concerns given the demonstrated risk that food industry involvement in nutrition research may bias research outcomes.^2–5 The juice and horticultural industries clearly have vested interest in ensuring that dietary guidelines support juice consumption. Relevant industry interests were reported in over half of the studies included in the umbrella review (8/15), as well as more comprehensively by the authors of this review in other recent related work.⁶

In conclusion, while the authors assert that the limited, weak evidence included in this paper does not itself warrant excluding or limiting 100% juice in dietary guidelines, it is equally arguable that nothing they have presented refutes other precautionary reasons behind why many countries worldwide are doing just that.

Author Contributions

All authors conceptualized, wrote, and revised the paper and have approved the final version.

Funding

No funding or in-kind contribution was received for this work. D.M. is supported by an Australian National Health and Medical Research Council (NHMRC) Fellowship held by A.J., and within the past 5 years has received support from the World Health Organization (WHO), Australian Government Medical Research Future Fund, Australian Government Research Training Program, Australian Department of Health and Aged Care, NHMRC Centre of Research Excellence in Food Retail Environments for Health, University of New South Wales, and Public Health Association of Australia. A.J. is supported by an NHMRC Fellowship, and within the past 5 years has received support from the WHO, Australian Government Department of Health and Aged Care, Victorian Health Promotion Foundation (VicHealth), United Nations Children's Fund (UNICEF), and the Gates Foundation for work on matters of public health nutrition policy. None of these funding sources were involved in the conception, design, performance, and/or approval of this work. R.S. and S.B. have no potentially relevant funding sources to declare.

Conflicts of Interest

D.M., R.S., and A.J. regularly engage, on a professional and personal level, in advocacy on matters relevant to public health nutrition. S.B. has no potentially relevant interests to declare.

REFERENCES