Is doenjang (Korean soybean paste) good for belly fat and blood pressure?

Traditional doenjang starts with cooked soybeans pressed into bricks called meju, which are hung to dry and inoculated with airborne molds (Aspergillus oryzae, Bacillus subtilis) and yeasts. The bricks are then submerged in salt brine for several months. The liquid that runs off becomes Korean soy sauce; the solids, mashed and aged for another six to twenty-four months, become doenjang. Commercial doenjang is faster — single-strain koji fermentation, much shorter aging — and produces a milder, less complex paste with a different bioactive profile than long-aged traditional doenjang⁷.

Three things happen during long fermentation that make doenjang nutritionally distinct from cooked soybeans:

• Isoflavone activation. Glycoside-bound isoflavones (genistin, daidzin) are cleaved by microbial β-glucosidases into the more bioavailable aglycone forms (genistein, daidzein), which is what most plasma isoflavone studies measure.
• Peptide generation. Bacillus proteases break soy protein into short peptides, some of which are angiotensin-converting enzyme (ACE) inhibitors that lower blood pressure. The dipeptide arginine-proline is one such fragment identified in fermented doenjang.
• Microbial succession. A complex consortium of Bacillus, Tetragenococcus, Lactobacillus, Pediococcus, and aspergilli colonize the paste. Some survive ingestion; others leave behind metabolites that shape the gut microbiota of the eater.

Aging time matters. Doenjang aged 24 months has roughly two to three times the antitumor activity of paste aged only 3–6 months in animal models, with parallel increases in natural killer cell activity and liver glutathione S-transferase⁵.

Visceral Fat and Body Weight

The clearest human signal is in body composition. In a 12-week double-blind RCT, 51 overweight adults received either 9.9 g/day of dried doenjang (about 2 tablespoons fresh-paste equivalent) or placebo. The doenjang group showed significant reductions in body weight and visceral fat area, with no change in total or subcutaneous fat — meaning the loss came specifically from the metabolically active fat depot around the organs¹. A follow-up trial stratifying by PPAR-γ2 C1431T genotype found that the visceral-fat benefit was concentrated in carriers of the T allele, while antioxidant effects (catalase activity, radical clearance) appeared mostly in C-allele carriers, hinting at gene-diet interaction².

Blood Pressure and Metabolic Syndrome

In a large cross-sectional Korean cohort of 58,701 adults, daily intake of jang (a category that includes doenjang and related fermented soybean pastes) was inversely associated with metabolic syndrome and its components — particularly waist circumference, body fat, and hypertension — even after adjusting for sodium intake. The protective association was stronger in men⁶. Mechanism studies in salt-fed rats suggest doenjang counteracts salt-induced hypertension partly by suppressing renin and increasing urinary sodium excretion, rather than simply offsetting salt with isoflavones⁴.

Gut Microbiome

In a mouse study, doenjang feeding reduced the Firmicutes-to-Bacteroidetes ratio, increased Akkermansia and Lactobacillus, suppressed Enterobacteriaceae and β-glucuronidase activity, and lowered circulating lipopolysaccharide — a metabolic-endotoxemia marker tied to insulin resistance and low-grade inflammation³. These shifts overlap with what's seen in Akkermansia muciniphila and butyrate-producer literature.

Practical Tips

• A typical Korean serving is 1–2 tablespoons per day, used in soups (doenjang jjigae), dipping sauces (ssamjang), or stews. Most clinical effects in trials come from this amount.
• Sodium awareness. Doenjang is salty (roughly 11–14% salt by weight in traditional paste). Replacing other salty seasonings with doenjang seems to be net beneficial in cohort data, but adding it on top of an already-high-sodium diet is not the same intervention.
• Look for traditional, long-aged doenjang in Korean grocery stores — labels may say jaeraesik (traditional) versus gaeryangsik (improved/commercial). Traditional pastes have outperformed commercial pastes head-to-head on lipid and menopausal markers⁷.
• Pregnant women, people on warfarin, and anyone with diagnosed soy allergy should consult a clinician before adding daily fermented soy.
• Heat handling: doenjang's Bacillus subtilis spores survive cooking, but most lactic acid bacteria don't. If probiotic bacteria matter to you, stir doenjang into already-finished soup off the heat. Isoflavones, peptides, and ACE-inhibitor activity are heat-stable.

See our fermented foods overview for context on what fermentation does to food, and miso for the closest Japanese analog.

Body Composition (Human RCTs)

Cha et al. (2012) randomized 51 overweight adults to 9.9 g/day dried doenjang or placebo for 12 weeks in a double-blind design. The doenjang group showed significant reductions in body weight (−1.0 kg, p<0.05) and visceral fat area on CT (p<0.05), with no significant change in total or subcutaneous fat areas, serum lipid profile, or self-reported dietary intake. Sample size is modest and the trial was industry-supported, but the visceral-fat selectivity is mechanistically interesting and has held up in follow-up work¹.

Cha et al. (2014) extended this in a 12-week RCT stratified by PPAR-γ2 C1431T genotype (60 enrolled, 51 completers). Visceral fat reduction was significant only in T-allele carriers, while antioxidant biomarkers (catalase, total radical clearance) responded mainly in C-allele carriers. Plasma free fatty acid, insulin, and HOMA-IR rose modestly in the doenjang arm, complicating the metabolic story. The genotype-stratified design is unusual and the subgroups are small, but the effect modification is biologically plausible given PPAR-γ's role in adipocyte biology².

Han et al. (2024) randomized 56 perimenopausal women across three doenjang variants (traditional high-dose, traditional low-dose, commercial). All three reduced the Kupperman menopausal symptom index, with the largest decrease in the traditional low-dose group. Traditional doenjang lowered LDL cholesterol; commercial did not. Microbiome analysis showed Bacteroidetes increases in traditional high-dose and commercial groups. The trial is small and short, but the head-to-head comparison of traditional versus commercial preparations is one of the few in the literature, and it consistently favors traditional fermentation⁷.

Blood Pressure and Metabolic Syndrome

Jeong et al. (2023) analyzed 58,701 adults (20,293 men; 38,408 women) from a Korean hospital-based cohort, classifying jang intake from food frequency data and adjusting for sodium and total energy. Daily jang intake was inversely associated with metabolic syndrome (especially waist circumference, body fat, and hypertension) in men, and with hypo-HDL cholesterolemia in women. Effects persisted after sodium adjustment, suggesting the association is not driven by salt substitution alone. As cross-sectional data, this cannot establish causation, but the size and direction are consistent with the smaller intervention work⁶.

Mun et al. (2019) fed Sprague-Dawley rats a high-salt diet (HS) or HS plus doenjang (HSD) for 5 weeks (n=6 per group). HSD significantly reduced systolic blood pressure versus HS alone, with increased urinary and fecal sodium and potassium excretion and reduced plasma renin. The renin finding parallels what nattokinase trials show in humans, suggesting renin-angiotensin-aldosterone system (RAAS) modulation is a shared mechanism across fermented soybean foods⁴.

Anticancer and Antimetastatic (Animal/In Vitro)

Jung et al. (2006) compared doenjang aged 3, 6, and 24 months in sarcoma-180 transplanted mice and a colon-26 lung metastasis model. Twenty-four-month doenjang produced two- to three-fold greater tumor inhibition than shorter-aged paste, increased splenic NK cell activity, and raised hepatic glutathione S-transferase. The mechanistic interpretation is that long aging concentrates aglycone isoflavones, melanoidins, and peptide breakdown products that drive Phase II detox enzyme induction⁵. This is animal evidence; doenjang's anticancer signal in humans is limited to ecological and case-control associations.

Gut Microbiome (Mouse Model)

Jang et al. (2014) administered doenjang to mice and observed reduced Firmicutes-to-Bacteroidetes ratio, increased Bifidobacterium, Akkermansia, Lactobacillus, and Odoribacter, decreased Enterobacteriaceae, suppressed bacterial β-glucuronidase, lowered circulating lipopolysaccharide, and increased IL-10 and PPAR-γ expression. The shifts are coherent with metabolic-syndrome-protective microbiome signatures from independent studies. Mouse-to-human translation is uncertain, but the direction overlaps with what the cohort and RCT data suggest at the phenotype level³.

Limitations and Open Questions

• Sodium load is real. Even traditional doenjang is around 11–14% salt by weight. The cohort signal that jang reduces hypertension risk is robust, but it depends on doenjang replacing — not supplementing — other sodium sources.
• Trial sizes are small. Most human RCTs have fewer than 60 completers and run 12 weeks or less. Long-term endpoint data on cardiovascular events, cancer, or all-cause mortality with doenjang specifically are not available; what we have is biomarker-level evidence and population correlations.
• Traditional vs. commercial matters. The two head-to-head comparisons (Han 2024; aging-time studies) both favor longer-fermented traditional preparations. Off-the-shelf supermarket doenjang in Korea is mostly gaeryangsik — short-aged single-strain product — and may not deliver equivalent effects.
• Genotype interaction. The PPAR-γ2 stratification result is one of very few examples of nutrigenomic effect modification for a fermented food. It needs replication before clinical use.

Overall evidence quality: Moderate for body composition (two small RCTs, consistent direction). Moderate-to-strong for the metabolic-syndrome association (very large cohort, sodium-adjusted, biologically plausible mechanism). Mechanistic for blood pressure and microbiome (animal models converging on RAAS and microbiome pathways). Preliminary for anticancer effects.