(PTFS06-02-23) “Healthy” Plant Diet Scores, Including Potatoes, and “Unhealthy” Plant Diet Scores, Including Potatoes, Were Both Sometimes Associated With Heart Disease Outcomes

Objectives: Various plant-based dietary indices (PDIs) have been tested for associations with coronary heart disease (CHD). Robustness of PDI associations is unclear because categories of foods vary based on many characteristics. In one approach, potatoes are uniquely categorized as a single ‘unhealthy’ group, while other categories of ‘healthy’ and ‘unhealthy’ plant-based foods include multiple plants. We tested the effect of reclassification of foods within this approach on associations between PDIs and CHD, including removing potatoes altogether or reclassifying them in various ways into ‘healthy’ or ‘unhealthy’ plant-based categories.

Methods: The scoring approach consists of 3 indices: overall PDI, healthful PDI (hPDI), and unhealthful PDI (uPDI). ‘Potatoes’ are classified as ‘unhealthy’ in the original approach. Using REasons for Geographic and Racial Differences in Stroke (REGARDS) data, we evaluated 5 alternative categorization schemes by moving all or some of the items in the ‘Potatoes’ group within or between ‘healthy’ and ‘less-healthy’ food groups, and estimated HRs and 95% CIs for each model consistent with the original approach.

Results: Using the original approach, we obtained statistically significant associations in some, but not all, PDIs stratified by sex with or without multivariable adjustment. Reconfiguration of potatoes in 5 different models did not have substantial influence on the associations between indices and CHD risk in multivariable- and age-adjusted models for women or men in any model. That is, moving all or some items of the ‘Potatoes’ group from ‘unhealthy’ to ‘healthy’, or removing ‘Potatoes’ altogether, did not make the associations much weaker or stronger compared to the original configuration.

Conclusions: Reclassification of PDI foods as ‘healthy’ and ‘unhealthy’ did not demonstrate substantial changes in associations to be expected from perturbing scores of ‘unhealthy’ plant foods with or without perturbing ‘healthy’ foods. Our results do not support the value of labeling individual foods as ‘unhealthy’ or ‘healthy’ within PDIs. Ongoing work involves evaluating influences from changing other PDI foods and categories.

Funding Sources: Funded by the Alliance for Potato Research and Education. Content and editorial authority remain with authors