Not All Plant-Based Diets Are Equal: Korean Study Links Diet Quality to Gut Bacteria and Heart Health

Describing a diet as "plant-based" turns out to say very little about how healthy it actually is. A cross-sectional study of 2,388 Korean adults published in the European Journal of Nutrition found that the quality of a plant-based diet, not just its plant content, determines whether it lowers or raises the risk of obesity, elevated blood glucose, and unfavorable cholesterol levels. The findings point to the gut microbiome as a likely messenger between what people eat and what happens to their metabolic health.

Healthful vs. Unhealthful Plant-Based Diets: How Risk Diverges

The researchers scored participants using three indices: an overall plant-based diet index (PDI), a healthful PDI (hPDI) that rewards whole grains, fruits, and vegetables, and an unhealthful PDI (uPDI) that scores higher when refined grains and added sugars dominate the plate, even within an otherwise plant-forward pattern. The distinction matters enormously. Higher hPDI scores were associated with a 28% reduction in obesity risk (OR=0.72; 95% CI, 0.57 to 0.91). Higher uPDI scores, by contrast, were associated with a 23% increased risk of elevated fasting glucose (OR=1.23; 95% CI, 1.00 to 1.52) and a 35% increased risk of low HDL-C levels (OR=1.35; 95% CI, 1.07 to 1.70). The same broad food category, plants, produced opposite metabolic signals depending on which plants dominated the meal. Research on plant-based diet cardiometabolic benefits has consistently shown meaningful risk reductions for coronary heart disease and type 2 diabetes, but those benefits appear contingent on diet quality in ways population-level summaries can obscure.

Why Korean Dietary Patterns Make This Study Timely

Korea offers a particularly informative setting for this question. Traditional Korean diets center on vegetables, fermented foods, and whole grains, but urbanization and processed food availability have reshaped eating habits in recent decades. Diet quality among Korean adults showed only modest improvement between 2007 and 2018, with younger cohorts tending toward poorer quality diets and worse cardiometabolic profiles. Against that backdrop, distinguishing the metabolic effects of a traditional plant-heavy pattern from a modern processed-but-plant-labeled one becomes a public health priority, not just an academic distinction. The Shin et al. cohort, recruited from two university hospitals between 2017 and 2019, reflects that contemporary dietary landscape.

Roseburia, Escherichia, and the Microbial Fingerprint of Diet Quality

The study profiled gut microbiota from all 2,388 participants using 16S rRNA gene sequencing and examined both community-wide diversity and specific bacterial genera. Higher hPDI scores were associated with greater alpha-diversity, a standard measure of microbial richness, while higher uPDI scores corresponded to lower diversity. Two genera illustrate the pattern clearly. Roseburia, a butyrate-producing bacterium, was positively associated with hPDI scores and fruit intake and inversely associated with BMI. Butyrate is a short-chain fatty acid that feeds the gut lining and carries anti-inflammatory properties. Escherichia/Shigella, a genus associated with dysbiosis, showed positive associations with both uPDI scores and fasting glucose levels, suggesting it may partly mediate the link between refined-carbohydrate-heavy plant diets and blood sugar dysregulation. Veillonella tracked with uPDI and is flagged by the authors as a possible marker of unhealthy dietary patterns. On the favorable side, Blautia was positively associated with HDL-C and dairy intake, while Caproiciproducens was inversely associated with BMI. A large population study linking diet-microbiome associations to cardiometabolic health has similarly found that plant-based eating patterns correlate with microbially mediated improvements in metabolic markers, with red meat consumption showing the opposite relationship.

The team also tested whether combining dietary index scores with genus-level microbiome data improved the ability to predict cardiometabolic risk. Using Random Forest machine learning models, they found that integrated diet-plus-microbiome models outperformed diet-only models for predicting obesity and elevated fasting glucose. Predictive performance for low HDL-C remained modest across all model types, suggesting that cholesterol risk carries more complexity than diet and microbiota alone can capture.

Limitations That Temper the Findings

The cross-sectional design is the most important caveat. Because diet, microbiota, and metabolic measures were all taken at a single point in time, the study cannot establish whether dietary patterns shaped the microbiome, whether the microbiome independently drove cardiometabolic risk, or whether both are downstream of some unmeasured third variable. Substantial gaps remain about the precise mechanisms by which gut microbiota contribute to cardiometabolic disease pathogenesis, particularly around causal pathways, and this study does not close those gaps. The 16S rRNA sequencing method also provides limited taxonomic resolution and no direct read on microbial function; what bacteria are present is not the same as what they are doing metabolically. The Random Forest models were not validated in an independent cohort, which limits how confidently predictive performance numbers can be generalized. Residual confounding from unreported medications or pre-existing conditions is also possible, and the researchers acknowledge incomplete covariate data.

What the Microbiome-Diet Connection Suggests for Future Research

Despite those constraints, the study contributes a concrete framework for thinking about plant-based diet quality in a non-European population, a group historically underrepresented in diet-microbiome research. The genus-level findings, particularly the Roseburia-hPDI link and the Escherichia/Shigella-uPDI link, generate testable hypotheses for longitudinal and intervention studies. The authors suggest the gut microbiome may serve as a more integrative biological readout of dietary patterns than self-reported questionnaires alone, an idea worth pursuing in clinical contexts where objective biomarkers of diet quality are scarce. For now, the clearest practical signal from the data is straightforward: within plant-based eating patterns, whole grains, fruits, and vegetables produce a microbial and metabolic profile that looks meaningfully different from one built on refined carbohydrates and added sugars.

Evidence rating

EMERGING — 86/100

Caveats:

• Cross-sectional design: the study shows associations, not cause and effect — diet, microbiome, and metabolic risk were all measured at the same point in time.
• Findings are from a Korean adult cohort and may not generalise to other ethnic or dietary contexts.
• 16S rRNA sequencing provides limited taxonomic and functional resolution; genus-level assignments should be interpreted cautiously.
• Machine learning models were not validated in an independent cohort, so predictive performance estimates may be optimistic.
• Funding was not disclosed by the authors.

Large cross-sectional cohort (n=2,388). Demonstrates associations between plant-based diet quality, gut microbiota composition, and cardiometabolic risk markers, but cannot establish causal direction or temporal sequence. Genus-level findings are consistent with prior mechanistic literature but have not yet been independently replicated in this specific combined dietary-microbiome framework.

References

1. Ji-Hee Shin, Eun-Ji Song, Mi Young Lim, Hye-Jo Choi, Jihye Kim, Young-Do Nam (2026). Plant-based diet quality and gut microbiota in relation to cardiometabolic risk in Korean adults. European Journal of Nutrition. https://doi.org/10.1007/s00394-026-03991-z
2. Hana Kahleova, Susan Levin, Neal Barnard (2017). Cardio-Metabolic Benefits of Plant-Based Diets. Nutrients. https://doi.org/10.3390/nu9080848
3. Garam Jo, Dahyun Park, Juhee Lee, et al. (2022). Trends in Diet Quality and Cardiometabolic Risk Factors Among Korean Adults, 2007-2018. JAMA Network Open. https://doi.org/10.1001/jamanetworkopen.2022.18297
4. Diet-microbiome associations and their impact on cardiometabolic health: a population-based study. Nature. https://www.nature.com/articles/s41591-023-02685-6
5. Role of the Gut Microbiome in Cardiometabolic Diseases – Frontiers. frontiersin.org. https://www.frontiersin.org/articles/10.3389/fmicb.2023.1190473/full