Korean Cohort Data on the Traditional Diet and Metabolic Syndrome Complicate the 'Hansik Is Healthy' Story

In a 2013 analysis of 26,006 adults enrolled in the Korean Genome and Epidemiology Study, people whose rice intake was overwhelmingly white rice, with almost no mixed grains or legumes added to it, had higher rates of central obesity and abnormal fasting glucose than people who ate the same rice-centered structure but with beans or multiple grains mixed in, according to {Ahn et al., 2013} {BMC Public Health}. That finding sits awkwardly next to the popular claim, common in both Korean nationalist food writing and Western wellness coverage, that the traditional Korean diet is close to a nutritional gold standard because it is plant-heavy, fermented, and low in red meat by Western standards. Korean cohort epidemiology over the past fifteen years supports a narrower and less flattering claim: certain features nested inside the traditional pattern (grain diversity, vegetable variety, fish, moderate portions) associate with lower metabolic syndrome, type 2 diabetes, and cardiovascular risk, while the pattern’s rice-and-sodium-heavy core, isolated from those features, tracks flat or worse.

• Grains, vegetables, fish pattern Inversely associated with metabolic syndrome in a national KNHANES sample, the most consistently protective cluster identified.
• Balanced rice-and-kimchi pattern Lower blood pressure and triglycerides in men, lower blood pressure and metabolic syndrome in women, but also higher risk of low HDL cholesterol.
• White-rice-dominant, low-diversity pattern Higher central obesity and abnormal fasting glucose in KoGES; 45 percent higher metabolic syndrome odds at the highest quintile in KNHANES women.
• Multigrain, fruit, and nut Hansik pattern Inversely associated with metabolic syndrome risk in a 58,701-person hospital-based KoGES cohort.
• Meat, alcohol, and westernizing patterns Positively associated with elevated glucose, triglycerides, blood pressure, and incident type 2 diabetes across multiple cohorts.

Korean nutrition epidemiologists build dietary patterns statistically, not by asking whether food is “traditional”

None of the studies behind this discussion start from the premise that Hansik is a fixed, definable diet and then test whether people who eat it are healthier. They start from food frequency questionnaire data, usually collected as part of KoGES or the Korea National Health and Nutrition Examination Survey (KNHANES), and run factor analysis or reduced rank regression across dozens of food groups to see which foods statistically cluster together in how real people actually eat. Kim and Jo applied this approach to 9,850 KNHANES respondents and extracted four patterns from the data: a white rice and kimchi pattern, a meat and alcohol pattern, a high fat, sweets, and coffee pattern, and a grains, vegetables, and fish pattern, according to {Kim & Jo, 2011} {Journal of the American Dietetic Association}. Each pattern is a statistical description of covariance in food intake, not a menu anyone designed.

This method matters for how the findings should be read. “Traditional Korean diet” is not one thing in this literature; it is whatever combination of foods happens to load together in a given sample, and different studies using different food lists, different regions, and different survey years produce different-looking traditional clusters. Song and colleagues, working with a separate KoGES-linked sample, extracted a “Traditional” pattern defined by high rice and kimchi intake, a “Korean healthy” pattern that added noodles, bread, eggs, and milk to that base, and a “Meat and Alcohol” pattern dominated by processed meat and alcohol, according to {Song et al., 2012} {Nutrition, Metabolism and Cardiovascular Diseases}. The labels are useful shorthand, but they describe factor loadings on a questionnaire, not a coherent dietary philosophy, and that distinction is why the same broad “traditional” description can show up with opposite metabolic associations across studies depending on exactly which foods dominate the cluster in that particular sample.

The rice-and-kimchi pattern’s actual metabolic signature is mixed, not uniformly protective

Song and colleagues’ comparison is the clearest single test of the “traditional equals healthy” claim, because it pitted the Traditional pattern directly against the Meat and Alcohol pattern in the same cohort. Relative to the Traditional pattern, the Meat and Alcohol pattern carried a 33 percent higher risk of elevated blood glucose, a 21 percent higher risk of elevated triglycerides, and a 21 percent higher risk of elevated blood pressure, according to {Song et al., 2012} {Nutrition, Metabolism and Cardiovascular Diseases}. That is a real, meaningful protective signal for three of the five metabolic syndrome components. But the same analysis found the Traditional pattern carried a 23 percent higher risk of low HDL cholesterol compared with the other two patterns combined, a finding the authors attributed to the pattern’s high carbohydrate share crowding out dietary fat. A pattern that lowers glucose, triglyceride, and blood pressure risk while raising low-HDL risk is not a diet that reads as unambiguously protective; it is a diet with a specific, carbohydrate-driven tradeoff.

A separate national analysis, working from a different KNHANES sample, reached a comparable middle-ground conclusion under a different label. Oh and colleagues identified a “balanced Korean diet,” rice and kimchi supplemented with a wider variety of foods and macronutrients closer to recommended ratios, that associated with lower risk of elevated blood pressure and hypertriglyceridemia in men and lower risk of elevated blood pressure and metabolic syndrome in women, according to {Oh et al., 2013} {PLOS ONE}. Critically, the same study identified an “unbalanced Korean diet,” high in carbohydrate and sodium with comparatively little food variety, that associated with higher risk of metabolic syndrome and elevated blood pressure in women. Both patterns are built around the same two anchor foods, rice and kimchi. What separates them, in this cohort, is variety and macronutrient balance layered on top, not the anchor foods themselves.

A high-rice, low-diversity version of the pattern tracks with worse outcomes, not better ones

This is where the literature gets genuinely counter-intuitive relative to the popular narrative. When Ahn and colleagues classified the 26,006-person KoGES sample into four rice-eating groups, white rice only, rice with beans, rice with multiple grains, and mixed, the white rice-only group was the reference against which the others were compared, and it was not the best-performing group. Women in the rice-with-beans group had lower risk of central obesity and abnormal fasting glucose than the white rice group, and women in the rice-with-multigrain group had lower central obesity risk as well, according to {Ahn et al., 2013} {BMC Public Health}. In other words, inside a sample that is, by any conventional definition, eating a traditional Korean rice-based diet, the version of that diet with the least grain diversity performed the worst on two core metabolic syndrome components.

A national KNHANES analysis restricted specifically to white rice and kimchi intake, without the buffering effect of side-dish variety that shows up in Oh’s “balanced” pattern, found the same direction of effect at a larger magnitude. Women in the highest quintile of combined white rice and kimchi intake had 45 percent higher odds of metabolic syndrome than women in the lowest quintile, a multivariable-adjusted odds ratio of 1.45, along with elevated risk specifically for high blood pressure, according to {Kim, Ahn, & Son, 2018} {Korean Journal of Community Nutrition}. Kim and Jo’s four-pattern factor analysis is consistent with this: of the four patterns they extracted from national data, it was the grains, vegetables, and fish pattern, not the white rice and kimchi pattern, that showed the inverse association with metabolic syndrome, according to {Kim & Jo, 2011} {Journal of the American Dietetic Association}. Across three independent Korean samples using three different analytic methods, the specific combination of white rice as the dominant carbohydrate source with limited food-group diversity is the version of “traditional” that performs worst, not best, on metabolic syndrome risk factors.

Sodium is not a footnote to this diet’s health story; it is a structural feature of it

The traditional pattern’s core seasonings and preserved foods, soy sauce, doenjang, and salted, brined kimchi and jeotgal, are also its dominant sodium sources at the population level. Using three rounds of nationally representative 24-hour dietary recall data spanning 1998 to 2009, Lee, Duffey, and Popkin found that kimchi, salt, soy sauce, and soybean paste were the leading dietary sodium sources in South Korea, and that mean sodium intake held flat around 4.6 to 4.7 grams per day across the eleven-year window, roughly double the World Health Organization’s recommended ceiling, according to {Lee, Duffey, & Popkin, 2013} {Journal of Human Hypertension}. This is not a diet that happens to include a few salty condiments; sodium delivery is functionally built into how the pattern’s core foods, kimchi and soy-based sauces and pastes, are preserved and fermented in the first place.

That sodium load complicates every protective finding described above, because none of the cohort analyses can fully separate “eating rice with diverse vegetable side dishes” from “eating rice with diverse, and therefore also more heavily salted and fermented, vegetable side dishes.” Oh’s unbalanced Korean diet pattern, the one associated with higher metabolic syndrome and blood pressure risk in women, was explicitly characterized by high sodium alongside high carbohydrate and low variety, according to {Oh et al., 2013} {PLOS ONE}, which suggests sodium concentration, not carbohydrate share alone, is doing real work in that adverse association. A pattern where the beneficial vegetable diversity and the harmful sodium load travel together in the same fermented, salted, banchan-heavy dishes is not a pattern where “eat more of it” is a straightforwardly safe recommendation, even where the aggregate metabolic syndrome numbers look favorable.

Grain and vegetable diversity, not the traditional label, appears to carry the protective signal

The most recent large KoGES analysis on this question tried to isolate what inside the traditional pattern was doing the protective work, rather than testing “traditional” as a single block. Kim, Hur, Jang, and colleagues built a 14-component Korean diet index from the hospital-based KoGES cohort of 58,701 adults and found that a high score on a component combining multigrain rice, fruit, and nuts, alongside low intake of fried foods, was inversely associated with metabolic syndrome risk, according to {Kim et al., 2022} {Frontiers in Nutrition}. The components driving the protective association were specific: grain variety beyond plain white rice, fruit, nuts, and the absence of fried food, not “rice and kimchi” as an undifferentiated pairing.

Read against the white-rice-dominant findings from Ahn and from Kim, Ahn, and Son, this points to a coherent explanation across the literature rather than a contradiction. The metabolically favorable version of the traditional Korean pattern is the one with the most grain and vegetable diversity added on top of the rice-and-fermented-side-dish base; the metabolically unfavorable version is the one where white rice supplies most of the carbohydrate load with comparatively little added on top. Both versions would plausibly get called “the traditional Korean diet” by someone eating them, and both versions appear in these cohorts under overlapping labels, which is exactly why single-study summaries claiming Hansik is simply protective, or simply not, both oversimplify what a decade and a half of factor-analysis studies actually shows.

As Korean eating patterns shift toward meat, alcohol, and refined convenience foods, cardiometabolic risk rises with the shift

The comparison case for all of this is what happens as Korean diets move away from the rice-and-banchan structure toward more Westernized eating, and the cohort evidence here is more consistent than the traditional-pattern evidence. Lee, An, and Lee used reduced rank regression on 7,354 cardiovascular-disease-free KoGES participants followed for a median of twelve years and found that a dietary pattern built around eggs, fish, milk, and dairy, plus other reduced-rank-derived patterns capturing red meat and alcohol, showed distinct associations with incident cardiovascular disease across sexes, with the meat- and alcohol-heavy pattern carrying elevated risk in men, according to {Lee, An, & Lee, 2020} {American Journal of Clinical Nutrition}. Chung, Kim, and Kwock, following 3,335 KoGES women aged 40 to 69 for incident type 2 diabetes, found that a meat and fish pattern was positively associated with diabetes risk, with the association concentrated in participants over 60 and those who were already obese or had metabolic syndrome, according to {Chung, Kim, & Kwock, 2019} {Nutrients}. Neither of these westernizing patterns is competing against a single, well-defined “Hansik” baseline; they are competing against whichever version of the traditional pattern, diverse or rice-dominant, happens to characterize the rest of that cohort’s remaining variance. But the direction is consistent across both studies: heavier reliance on red meat and alcohol, layered on top of whatever base diet a person eats, tracks with worse glycemic and cardiovascular outcomes in Korean cohorts specifically, not just in the Western populations where that finding originated.

The overall picture that fifteen years of KoGES and KNHANES dietary pattern research supports is more specific, and more useful, than either a nationalist claim that Hansik is protective by nature or a dismissive claim that the sodium content cancels out any benefit. Grain and vegetable diversity, fish intake, and macronutrient balance associate with lower metabolic syndrome, and by extension lower downstream type 2 diabetes and cardiovascular risk, in study after study; a narrower, white-rice-dominant version of the same rice-and-kimchi foundation, with limited diversity added, associates with worse outcomes on central obesity, fasting glucose, and overall metabolic syndrome risk in at least three independent Korean cohorts; and the sodium delivered by the pattern’s signature fermented condiments and preserved side dishes remains a documented, unresolved cardiovascular cost that travels alongside the beneficial vegetable diversity rather than separate from it. None of this supports treating “traditional” as a health claim on its own. It supports treating grain diversity, vegetable variety, and sodium load as the three variables that actually move the metabolic needle inside a pattern that gets called Hansik whether it contains them or not.

Hannah Cho is the Nutrition Science Editor at Daily Bite Lab. She is a Registered Dietitian Nutritionist with a Master’s in Human Nutrition from Columbia University and a former clinical dietitian at Mount Sinai.