Metabolomic Data May Support the Development of Personalized Nutrition Approaches to Type 2 Diabetes Management

Abstract

Diet is a major modifiable risk factor for type 2 diabetes (T2D), and suboptimal diets continue to contribute substantially to the population burden of T2D. Personalized nutrition proponents argue that specializing recommendations to interindividual differences in diet–health relationships will yield reductions in disease risk. However, how to personalize diet, and to whom, to reduce T2D risk remains unclear. Metabolites offer promise in this respect, as they capture (in part) dietary intake after the processes of digestion, processing, and absorption. The incorporation of metabolite data into diet–health studies therefore offers the opportunity to examine how the effects of food on the metabolome differ between individuals, and the extent that these differences give rise to differential diet–health associations. Ultimately, such studies hold promise for identifying personalized nutrition strategies to reduce the population-level burden of T2D.

BACKGROUND

Diet is one of the major modifiable risk factors for type 2 diabetes (T2D), which is the seventh leading cause of death in the United States.¹ In 2019, 35.6% of deaths from T2D in the United States were attributable to suboptimal nutritional intake,² highlighting the need to develop more effective dietary strategies for T2D prevention and management. Significant inter- and intra- individual differences in the postprandial responses of glucose, insulin, lipids, and markers of inflammation to a standardized intake are well documented,^3,4 arising (in part) from differences in the processes of food digestion, processing, and absorption, which give rise to differences in gastric emptying, endogenous glucose output, and nutrient availability and absorption.^5–7 However, current national guidelines, such as the US Department of Agriculture’s (USDA’s) Dietary Guidelines for Americans, take a global approach to dietary strategies for health promotion, offering recommendations that are only tailored to 1 of 5 life stages: “infancy,” “children ages 12 to 36 months,” “children and adolescents,” “women capable of becoming pregnant,” “pregnant and lactating women,” “individuals ages 50 years and older,” and “older adults.”⁸ The lack of more personalized advice, tailored to the metabolic response of different individuals, or the same individual at different times, is a missed opportunity to deliver the most efficacious advice for T2D prevention.

Dietary guidelines may have retained a “one-size-fits-all” approach because of a lack of rigorous, empirical data that capture how the biological milieu arising from dietary intake, which can provide health benefits or risks, differs between people. This lack of knowledge has made it difficult to know (1) which groups of people show differential dietary responses and (2) when individuals’ dietary responses change, the groups and times that may warrant tailored dietary strategies for the most efficacious health promotion. At the phenotypic level, only glycemic status has emerged as a robust moderator of diet–glycemia associations, with monounsaturated fatty acids (MUFAs) and MUFA-containing foods supporting glucose homeostasis in those with normoglycemia only.⁹ Investigations into other characteristics, such as gender, geographical ancestry,^10,11 and body mass index, have been less conclusive, with a lack of findings between different studies that converge on the same conclusions.¹² At the genotypic level, both candidate gene and genome-wide approaches have failed to yield robust, replicable interactions between dietary intake and T2D.¹³ Thus, to date, we know very little about which characteristics moderate the effects of diet on glycemia, in addition to how the biological effects of diet differ by these characteristics, and therefore cannot design and test the effects of personalized dietary approaches in T2D.

A ROLE FOR METABOLOMICS IN PRECISION NUTRITION

After ingestion, food is catabolized into smaller compounds, molecules that can be comprehensively characterized by current metabolomic technologies. By capturing the molecules that arise from food digestion, processing, and absorption, metabolites can provide a biological profile that can be interrogated for inter- and intra-individual differences in the bioactive compounds arising from dietary intake. Their status as an intermediate molecule between food ingestion and health outcomes may explain why diet-related metabolites show associations with health that are orders of magnitude stronger than those of self-reported dietary intake with health.^14–16 For example, in 1 study, unprocessed red meat intake explained less than 1% of the variance in C-reactive protein (CRP), but a metabolomic biomarker, a form of metabolic biomarkers that is derived from metabolomic assays, explained over 5%.¹⁴ Similarly, a biomarker of avocado intake was associated with a 6% reduction in T2D incidence over 18 years, while avocado intake itself was only associated with a nonsignificant reduction of approximately 1%.¹⁵ In another cohort, 3 dietary pattern–specific metabolite scores (representing the aggregated levels of metabolites associated with each of the Healthy Eating Index, the Dietary Approaches to Stop Hypertension [DASH] diet, and a Mediterranean-style diet) explained 31%–43% of the variance in each of 5 measures of insulin and glucose homeostasis, compared to less than 3% for the respective dietary intakes.¹⁶

Stronger metabolite–glycemia vs diet–glycemia relationships likely reflect many factors, including improved measurement precision for metabolite vs dietary intake data, and the ability of metabolites to capture multiple other dietary and environmental exposures associated with the biomarkers of interest. However, other objective measures of dietary intake, such as those from direct observation or wearable sensors, do not increase the strength of diet–health associations to the same magnitude, and studies have failed to identify other dietary/environmental factors that are significantly associated with the metabolomic biomarkers under investigation.^14–16 Therefore, the ability of metabolites to capture dietary intake after the processes of digestion, processing, and absorption remains a lead candidate to explain this pattern of findings, whereby the associations of diet and metabolites with health are stronger than expected from the associations of dietary intake with health, and illustrate that metabolite data provide information on how the physiological effects of food can differ between individuals.

REMAINING RESEARCH NEEDS

While metabolomic biomarkers of food intake can provide unique and personalized insights into the effects of food on our metabolism, their direct translation to clinical and public health settings is in its early stages for numerous reasons. First, because personalized responses to dietary intake, by nature, differ between individuals, identifying metabolomic biomarkers of individual foods that are broadly applicable across most individuals is challenging. As a result, individual metabolomic biomarkers of intake have broadly failed to replicate across studies,¹⁶ and a clear research need exists for identifying a robust set of metabolites that index the same food, or dietary pattern, across multiple different populations.

Second, there are technical issues specific to metabolomic data that need to be more thoroughly understood. A regulatory framework for translating metabolomic biomarker information into dietary advice must be developed for the incorporation of metabolomic biomarkers in public health nutrition and done with the trust of clinicians and public health professionals. Any such framework will need to include optimal protocols for the metabolome assessments, but stability and turnover rates vary across the spectrum of molecules captured by metabolomic assays.¹⁷ Furthermore, the extent to which various metabolites reflect short-term vs long-term intake is not well understood. A better understanding of how assay timing affects the relationship of the metabolome to dietary intake is necessary for developing standardized protocols that support the translation of metabolomic findings into dietary advice.

Last, the price of metabolomic assays is high, and the value of any information they yield needs to be more firmly established. The human body is best conceptualized as a biological system, with the effects of diet on the metabolome only captured comprehensively via a combination of costly analytical instruments, each priced in the hundreds of thousands of dollars.¹⁸ To date, findings are mixed on whether personalizing dietary advice by any strategy improves adherence¹⁹ or health outcomes²⁰ compared with more global recommendations. There is a need for studies on the “real world” impact of personalized nutrition advice on health that cover a broad range of population and health outcomes, evaluate these strategies in real-world settings, and use “gold standard” scientific techniques, such as control groups. Evidence that metabolomic biomarkers have a significant and positive impact on public health must be strong and consistent before public health stakeholders are likely to commit to the financial investment.

Thus, while metabolomic data show promise in explaining how, and potentially why, the effects of diet differ between individuals—the foundation of personalized dietary strategies—it will likely be some time before their potential is operationalized in public precision nutrition initiatives.

CONCLUSION

Poor diet quality continues to make a substantial contribution to rates of T2D, a relationship that is not expected to change. The well-documented differences in how people metabolize their food, and the effects of this on nutrient availability and absorption, suggest that shifting from population-level (“one-size-fits-all”) to individual-level (“personalized”) diets is likely to help harness the full potential of diet for T2D management/prevention. However, our ability to develop such approaches is hampered by a lack of knowledge regarding how, and when, and for whom, the metabolic response to dietary intake differs. Metabolomic data can provide a biological profile that, with the right study design, can be interrogated for associations with dietary intake, yielding information on differences in diet–metabolite associations. Emerging research has shown that personalizing dietary approaches through the incorporation of metabolomic (and other) information can improve glycemic outcomes, including in randomized controlled trials.^21,22 Although there were differences in terms of study content and duration between the diet groups, these studies support the notion that metabolomic data can inform the development of individualized dietary strategies, and that glycemic outcomes will be improved with such precision nutrition vs population-level dietary approaches. Research that uses metabolomic information is an important future direction for precision nutrition initiatives, and such efforts offer promise for improving our ability to prevent and manage chronic diseases of aging, such as T2D.

Author Contributions

A.C.W. and J.I.R. conceived the review idea. A.C.W. drafted the manuscript. D.J.L. and S.B. edited drafts of the manuscript. All authors read, reviewed, and edited the final draft and provided feedback prior to the manuscript submission.

Funding

A.C.W. and D.J.L. were supported, in part, by US Department of Agriculture/Agricultural Research Service (USDA/ARS) cooperative agreement #58–3092-5–001. The contents of this publication do not necessarily reflect the views or policies of the US Department of Agriculture, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government. J.I.R. was supported, in part, by National Institutes of Health grants from the National Institute of Diabetes and Digestive and Kidney Diseases (R01-DK109588, P30-DK063491) and from the National Center for Advancing Translational Sciences (UL1TR001420, UL1TR001881). The funding bodies played no role in manuscript preparation.

Conflicts of interest

A.C.W. has received funding from Hass Avocado Board and the Beef Checkoff. The other authors report no potential conflicts.