Plant-based diets and metabolites
Principal Investigator
Name
Rachael Solomon
Degrees
Ph.D., M.P.H., R.D.
Institution
NIH
Position Title
Senior Investigator
Email
About this CDAS Project
Study
IDATA
(Learn more about this study)
Project ID
IDATA-91
Initial CDAS Request Approval
Apr 28, 2025
Title
Plant-based diets and metabolites
Summary
Background:
Greater adherence to diets of higher quality has been associated with reduced risk of multiple non-communicable disease outcomes and mortality (1). Several recent studies have examined whether the dietary quality of consuming a plant-based diet influenced risk by examining the association between overall plant-based diet index (PDI), as well as, healthful (hPDI) and unhealthful (uPDI) versions of a plant-based diet with disease outcomes and mortality (2, 3). While for all three of the plant-based diet indices, greater adherence limited animal foods, the uPDI included less whole grains, whole fruits, vegetables, nuts, legumes, vegetable oil, tea, and coffee while the hPDI included greater nutrient dense plant-based foods such as whole grains, whole fruits, vegetables, nuts, legumes and vegetable oils but less refined grains, fried potatoes and chips, fruit juice, sugar sweetened beverages, sweets, and desserts (3, 4). The overall PDI is similar to hPDI, however includes more refined grains, fried potatoes and chips, fruit juice, sugar sweetened beverages, sweets, and desserts. hPDI and/or PDI have been associated with reduced risk of diabetes, cardiovascular disease, dislipidemia, hypertension, stroke, respiratory infection, pancreatic cancer, and all-cause and cancer mortality, while uPDI was associated with either a higher risk or no association (2, 5-20). The proposed study will examine the association between PDI, hPDI and uPDI and metabolomic profiles within the iDATA.
Methods:
We will include participants with complete (n~718 participant):
Diet: DHQ, ASA41, 4DFR
Metabolomic data (MetabolonHD4)-urine and blood samples at baseline and 6 months
Biospecimens data (if available glucose, insulin, inflammatory biomarkers)
Urine markers (sodium, potassium, protein)
Doubly labeled water (DLW)
Anthropometry
Demographic information (age, sex, race-ethnic group, education)
Physical activity data for controlling for confounding
We will calculate the plant-based diet from dietary data (using My Pyramid and Food Pyramid equivalents, and variables that we create) and examine correlations between the 3 patterns with serum and urinary metabolites. We will also conduct sensitivity analyses correlating the different dietary patterns to urinary sodium, potassium, and protein. We will divide the sample in half (test and validation sample) and apply machine learning algorithms such as Least Absolute Shrinkage and Selection Operator (LASSO) regression or elastic net regression to the test set and replicate in the validation set to determine the AUC.
Greater adherence to diets of higher quality has been associated with reduced risk of multiple non-communicable disease outcomes and mortality (1). Several recent studies have examined whether the dietary quality of consuming a plant-based diet influenced risk by examining the association between overall plant-based diet index (PDI), as well as, healthful (hPDI) and unhealthful (uPDI) versions of a plant-based diet with disease outcomes and mortality (2, 3). While for all three of the plant-based diet indices, greater adherence limited animal foods, the uPDI included less whole grains, whole fruits, vegetables, nuts, legumes, vegetable oil, tea, and coffee while the hPDI included greater nutrient dense plant-based foods such as whole grains, whole fruits, vegetables, nuts, legumes and vegetable oils but less refined grains, fried potatoes and chips, fruit juice, sugar sweetened beverages, sweets, and desserts (3, 4). The overall PDI is similar to hPDI, however includes more refined grains, fried potatoes and chips, fruit juice, sugar sweetened beverages, sweets, and desserts. hPDI and/or PDI have been associated with reduced risk of diabetes, cardiovascular disease, dislipidemia, hypertension, stroke, respiratory infection, pancreatic cancer, and all-cause and cancer mortality, while uPDI was associated with either a higher risk or no association (2, 5-20). The proposed study will examine the association between PDI, hPDI and uPDI and metabolomic profiles within the iDATA.
Methods:
We will include participants with complete (n~718 participant):
Diet: DHQ, ASA41, 4DFR
Metabolomic data (MetabolonHD4)-urine and blood samples at baseline and 6 months
Biospecimens data (if available glucose, insulin, inflammatory biomarkers)
Urine markers (sodium, potassium, protein)
Doubly labeled water (DLW)
Anthropometry
Demographic information (age, sex, race-ethnic group, education)
Physical activity data for controlling for confounding
We will calculate the plant-based diet from dietary data (using My Pyramid and Food Pyramid equivalents, and variables that we create) and examine correlations between the 3 patterns with serum and urinary metabolites. We will also conduct sensitivity analyses correlating the different dietary patterns to urinary sodium, potassium, and protein. We will divide the sample in half (test and validation sample) and apply machine learning algorithms such as Least Absolute Shrinkage and Selection Operator (LASSO) regression or elastic net regression to the test set and replicate in the validation set to determine the AUC.
Aims
The primary aim of this study is to:
1. Determine whether PDI, hPDI and uPDI are associated with metabolomic profiles in a cross-sectional analysis of the iDATA metabolomic data.
We hypothesize that PDI, hPDI, and uPDI will be associated with different metabolomic profiles with PDI and hPDI having similar metabolite associations while the uPDI will have different metabolomic profiles compared to the other plant-based diets. We will also examine associations between the PDI, hPDI, and uPDI components.
Collaborators
Laura Klein Ph.D. Student Emory University Rollins School of Public Health
Terry Hartman, PhD, MPH, RD, Professor Emory University Rollins School of Public Health
Erikka Loftfield Cronin, Ph.D., M.P.H., DCEG, NCI