Strengthening the reporting of diet item details in feeding studies measuring the dietary metabolome: The DID-METAB core outcome set statement

Nutrition research and diet–disease relationships historically rely on self-reported data assessed via dietary assessment instruments such as 24-h dietary recalls, food records, food frequency questionnaires, etc.,¹ which are prone to inherent bias and errors.^{1, 2} While these methods provide detailed information on what, how much, and when individuals eat, involvement from dietitians or nutritionists can help to minimise errors.³ However, misreporting remains inherent and can lead to misinterpretation of diet–disease relationships.² Controlled human feeding studies provide known amounts of foods/beverages and aim to mitigate inherent biases associated with self-reported dietary assessment while observing individual responses and enhancing adherence; however, they are also highly resource-intensive. The reliability and accuracy of dietary assessment methods have been shown to be increased by substituting or complementing dietary assessment instruments with objective biomarkers of food intake.^4-8 Currently, there are few valid dietary biomarkers routinely applied, for example, 24-h urinary sodium for salt,⁹ plasma carotenoids for fruit and vegetables,¹⁰ proline betaine for citrus fruits¹¹; however, their application can be limited to a specific nutrient or food/food group.¹¹ Human feeding studies utilising metabolomics as an adjunct objective dietary assessment method are gaining traction.^12-14 However, the methodology of dietary feeding interventions can vary in their approach,¹⁵ making cross-comparison between studies and synthesising dietary evidence difficult (see Box 1). Beyond the discovery of metabolites identified from biospecimens for qualifying and quantifying dietary intake of specific foods, nutrients and/or dietary patterns, metabolomics may also reflect the impact of diets on endogenous metabolism, accounting for individual variation driven by factors such as genetics and gut microbiome composition. For example, metabolites derived from the gut microbiome^{16, 17} or produced through microbial conversion,^{18, 19} contribute to the diverse metabolic responses to dietary interventions.¹⁶ Therefore, metabolomics offers promise for future incorporation within precision and personalised nutrition interventions, ultimately advancing the broader field of nutrition research.¹⁶

While metabolomics is being rapidly integrated as a biological assessment technique in nutrition research,²⁰ it is still in its infancy and therefore improved quality of reporting is required to facilitate consistency, reproducibility of findings, and advancement of the field long-term.

We previously demonstrated that there is extensive variability in the reporting of dietary intervention methodologies (e.g. design, delivery, implementation and interpretation) currently used in human feeding studies measuring the metabolome.¹⁵ Commonly, insufficient detail is reported, hindering replication, which limits evidence synthesis in the field of metabolomics.¹⁵ For example, information about included/restricted foods, the timing of biospecimen collection in relation to dietary assessment instruments used, or methods used to account for the consumption of nonstudy foods. Detailed information on these items is vital for the interpretation of the metabolome data. While reporting guidelines exist for human intervention studies more broadly,^21-23 including the developing CONSORT-Nut,^{24, 25} a nutrition extension of the CONSORT statement, no reporting guidance currently considers the specific nuances in dietary intervention research in which the metabolome is also measured. Therefore, there is a need for formal consensus on the minimum core set of items required for reporting, along with examples and recommendations for reporting in research papers to guide researchers and the review process. The primary aim was to gain consensus on core diet item details (DID) and standard reporting recommendations for each DID (i.e. a core outcome set, COS) in human feeding studies measuring the metabolome. The secondary aim was to develop a reporting guideline for use by researchers conducting such studies when reporting information in papers, and to assist journal reviewers and editors when critically appraising the papers (see Box 2). The purpose of this paper (i.e. the DID-METAB Statement) is to provide a short overview of the development of the COS and reporting guideline, including the Delphi process, and present the final reporting guideline (i.e. DID-METAB Checklist) to support usability and dissemination.

The DID-METAB Statement was developed by the Precision and Personalised Nutrition (PPN) Team (JJAF, EDC, JS, MGM, TJ and CC) in consultation with the DID-METAB Delphi Working Group under the iterative process of an online Delphi. Development of the core outcome set using the Delphi process was conducted in accordance with the Core Outcome Set-STAndards for Development: The COS-STAD recommendations.²⁷ The development of the reporting guideline was based on guidelines for developers of health research reporting guidelines and modelled off similar efforts.^{22, 28-30} The PPN Team has collective expertise in human clinical and experimental research design, conduct and implementation of human feeding interventions, dietary assessment methodology, human biospecimen collection and analysis, and design and management of Delphi processes. The DID-METAB Delphi Working Group experts were identified based on their extensive experience and contributions to the field, such as peer-reviewed publications, involvement in key professional organisations, and their recognised expertise and contributions to the field of metabolomics and nutrition research. International experts were invited by the PPN Team and encompassed expertise across clinical and experimental trial design of dietary interventions, feeding study intervention implementation, nutritional metabolomics and/or diet-related biospecimen analyses and interpretation. The two-stage Delphi process comprised five survey rounds, which were implemented online using QuestionPro Survey Software (QuestionPro Inc., Austin Tx). The Delphi was conducted between February 2024 and July 2024 to gain consensus on a core set of DIDs, DID phrasing, reporting recommendations including examples, and acceptance of the final checklist. A total of 67 experts were invited, with 25 providing input in stage 1, and 22 experts retained throughout all three rounds of stage 2.

All DID-METAB Delphi Working Group experts agreed with the PPN Team's recommendation that the checklist should be used alongside existing tools (e.g. as an extension of item 5 in CONSORT 2010 Statement updated guidelines for reporting parallel group randomised trials,²³ or item 11 in SPIRIT 2013 Statement: Defining standard protocol items for clinical trials²⁶) and that relevant journals should recommend use of the DID-METAB Checklist for relevant studies. This study was approved by the University of Newcastle's Human Research Ethics Committee (H-2023-0405) and has been registered on the Core Outcome Measures in Effectiveness Trials (COMET Initiative) database (https://www.comet-initiative.org/Studies/Details/3292). The methodology for the development of the reporting guideline, including findings of the Delphi have been thoroughly reported in the more comprehensive Explanation and Elaboration report, available at: [https://advances.nutrition.org/].

The final list of DIDs (29 items), plus examples and recommendations are categorised across five domains: (1) Dietary Intervention—Modelling (items 1 through 8), (2) Dietary Intervention—Implementation (items 9 through 11), (3) Dietary Assessment (items 12 through 20), (4) Adherence and Compliance Monitoring (items 21 through 24) and (5) Bias (items 25 through 29). The recommendations are presented in a checklist (Table 1) to aid users in completing it. The COS recommendations within the DID-METAB Checklist are guidelines for reporting research and do not prescribe how to design feeding studies. Examples are provided for each DID within the checklist, including reporting recommendations for each DID. The hierarchy of reporting recommendations was based on a vote count of the experts' responses and synthesis of their commentary regarding the level of detail to be provided. Reporting recommendations labelled as ‘consider’ and ‘optional’ are nonmandatory reporting items. Those labelled as ‘consider’ guide users to include this detail if possible, as it will likely benefit other researchers/the field, whereas ‘optional’ means given this data may or may not be relevant to report in this manner for a particular study or it may not be of benefit to other researchers, it is not necessary to provide it. To assist the application of the recommendations, we encourage readers to access and utilise the Explanation and Elaboration report.³¹

The quality of reporting in published research describing details of dietary intervention methods (e.g. design, delivery, implementation and interpretation) used in human feeding studies measuring the metabolome is considered poor.¹⁵ Currently, reporting of dietary characteristics and compositions of implemented interventions are highly variable with some studies reporting diets only in terms of macronutrient composition, others reporting foods provided or nutrient targets, while some provide example meal plans and portion sizes.¹⁵ The variability spans across several features of dietary intervention methods used in feeding studies.¹⁵ Therefore, without detailed and replicable reporting of core information relating to dietary intervention methods, particularly those that concern the validity and interpretation of the metabolome, it remains challenging to replicate research or synthesise the evidence base.

The explicit aim of this COS was to improve the quality of reporting of feeding studies measuring the metabolome by identifying a minimum set of information to be reported in order to provide details about how the diet intervention was designed, delivered and interpreted. The DID-METAB statement was developed to standardise reporting, enhance the peer review process of papers and assist researchers in critically appraising and synthesising published articles. We recommend submitting the checklist as an additional file with the research article. The supporting Explanation and Elaboration report presents published examples of best practice reporting for each item as well as highlighting potential limitations of some approaches. The DID-METAB Statement can be used to enhance the design of feeding studies and ensure all aspects of feeding study interventions are adequately reported with sufficient detail and clarity.

The structured and formal consultation process, high response rate (88%), retention of all 22 international experts in the final three rounds of stage 2 across a broad range of research expertise totalling >200 years, and unanimous consensus on the final checklist are key strengths of the DID-METAB Statement. Implementation of the DID-METAB Statement in research will strengthen the evidence base on nutritional metabolomics and potential application to precision and personalised nutrition strategies.

To encourage dissemination and use of this standard for reporting, we have simultaneously submitted for publication the Explanation and Elaboration report in Advances in Nutrition. The Enhancing the QUAlity and Transparency Of health Research (EQUATOR) Network (www.equator-network.org) will assist in disseminating and promoting the downloadable DID-METAB statement. Announcements, updates, details for contacting the PPN Team and supporting information relating to the DID-METAB Statement, including the downloadable checklist, can be found at the DID-METAB website (https://australianeatingsurvey.com.au/did-metab-statement). We will continue to approach journals identified as being widely read by the medical and research community that are conducting relevant studies to endorse the use of the DID-METAB Statement. The DID-METAB Statement will be periodically reappraised by the PPN Team, and if necessary, modified and/or updated to reflect comments, criticisms and any new evidence.

In conclusion, we recommend that authors publishing articles on human feeding studies where metabolomic samples are collected include a completed checklist in their paper submissions to aid the editorial process, facilitate critical appraisal by the readers and contribute towards advancing the field of metabolomics (available at: https://australianeatingsurvey.com.au/did-metab-statement).

JJAF, EDC, JS, MGM, TJ and CC conceptualised and designed the research. JJAF and TJ collected and assembled the data. JJAF analysed and collated the data, presenting it to EDC, JS and MGM after each stage of the Delphi process to incorporate expert feedback. TJ provided administrative and technical support for the Delphi. JJAF wrote the paper. EDC, JS, MGM, TJ and CC were involved in the critical revision of the paper. All members of the DID-METAB Delphi Working Group were participants in the entire two-stage Delphi process and thus contributed to data collection, the development of the checklist, and reviewed this paper. JJAF had primary responsibility for final content. All authors have read and approved the final manuscript.

EH is supported by a Laureate Fellowship from the Australian Research Council; EH is a director of Melico Ltd. outside the scope of the submitted work with no financial contribution. FZM is supported by a Senior Medical Research Fellowship from the Sylvia and Charles Viertel Charitable Foundation, a National Heart Foundation Future Leader Fellowship (105663), and an NHMRC Emerging Leader Fellowship (GNT2017382). HMS is supported by an NHMRC Emerging Leader Fellowship (APP2018118). MHT is supported by the BBSRC Core Capability Grant BB/CCG2260/1 and its constituent project BBS/E/QU/23NB0006 (Food & Nutrition National Bioscience Research Infrastructure). MS is supported by a National Heart Foundation Postdoctoral Fellowship (106698). JJAF holds a separate part-time employment at Sanitarium Health Food Company, which had no input into the study and is not financially supporting or sponsoring any part of this study. All other authors have no conflict of interest to declare. All DID-METAB Delphi Working Group members were volunteers.