Tempeh: the fermentation is measurable, the benefit is pooled

The strongest evidence on this page is not only about tempeh. A 2025 meta-analysis of 15 randomised trials of fermented soy foods found total cholesterol 5.0 mg/dL lower (CI 3.39 to 6.60), fasting glucose 6.39 mg/dL lower (CI 2.40 to 10.38), waist circumference 1.50 cm smaller and BMI 0.14 kg/m2 lower than in controls. That pool mixes tempeh with natto, miso and similar foods and cannot separate them, and the other metabolic markers did not move. What is specific to tempeh is what fermentation does to the soy before you eat it. At an identical isoflavone dose of 0.44 mg per kg of body weight, 59 healthy adults reached higher peak serum daidzein and genistein from tempeh than from textured vegetable protein, and the abstract gives no size for that gap. In 17 healthy men, nine days of 112 g of tempeh a day produced higher urinary recovery of both isoflavones than 125 g of unfermented soybean pieces, even though fermentation had lowered the isoflavone content of the food itself.

The numbers, per 100 g

How much is actually in it

• Tempeh carries 20.29 g protein, 0.56 mg copper and 0.358 mg riboflavin per 100 g in USDA data, while the British CoFID record gives 20.7 g, 0.67 mg and 0.48 mg, a 25% spread on riboflavin, and only 0.1 ug vitamin B12. — USDA SR Legacy record 174272 (Tempeh) and CoFID 2021 record 13-118 (Tempeh), Public Health England, OGL v3.0 (n = 2 reference records) Evidence E source
  Protein 20.29 vs 20.7 g (2%), copper 0.56 vs 0.67 mg (16%), riboflavin 0.358 vs 0.48 mg (25%); CoFID calcium 120 mg, iron 3.60 mg, vitamin B12 0.1 ug, energy 166 kcal, fat 6.4 g

What a real portion looks like

• A 307 kcal breakfast containing tempeh lowered plasma acyl-ghrelin by 35% at 30 minutes and kept it below baseline for 120 minutes in 13 young women with obesity, whereas an isocaloric unfermented soybean breakfast changed nothing, yet subjective appetite did not differ. — 13 Indonesian females aged 18-20 years with BMI 25-30, randomized counter-balanced crossover, single breakfast after overnight fast (n = 13) Evidence B source
  Acyl-ghrelin -35% at 30 min (P < 0.05); insulin response 30% greater at 120 min; arginine 24% higher at 120 min; no difference in VAS appetite

What your body absorbs

• At the same isoflavone dose of 0.44 mg per kg body weight, tempeh produced higher peak serum daidzein and genistein and larger areas under the curve than textured vegetable protein in 59 adults, because fermentation leaves the isoflavones mostly as aglycones. — 59 healthy adults (21 premenopausal women, 17 postmenopausal women, 21 men), random crossover of soy milk, textured vegetable protein and tempeh (n = 59) Evidence B source
  Tempeh vs TVP: higher peak daidzein (P < 0.001) and genistein (P < 0.01), higher AUC (P < 0.001 and P < 0.03); soy milk absorbed faster; 30% equol producers

What cooking and storage change

• For legumes boiled 15 to 20 minutes, drained and then fried, the USDA retention table keeps 70% of copper and riboflavin, 55% of folate, 60% of thiamin and 85% of iron and calcium, a group factor that is the only cooking-loss estimate available for tempeh. — USDA Table of Nutrient Retention Factors, Release 6 (2007), group Legumes and Legume Products, code 505, table dated Sep-75 (n = 1 group factor set) Evidence E source
  Retained: folate 55%, vitamin C 60%, thiamin 60%, niacin 65%, vitamin B6 65%, copper 70%, riboflavin 70%, potassium 75%, magnesium 80%, calcium 85%, iron 85%; protein is not in R6
• Nine days of 112 g tempeh a day gave 17 healthy men significantly higher urinary recovery of daidzein and genistein than 125 g of unfermented soybean pieces, even though fermentation had lowered the isoflavone content of the food. — 17 healthy men aged 20-40 years (22 enrolled), randomized crossover, two 9-day controlled feeding periods (n = 17) Evidence B source
  Recovery of daidzein and genistein significantly higher on tempeh than on soybean pieces (P < .002); total isoflavonoid and lignan excretion similar between the two; 5 of 17 were high equol excreters

What it does to your health

• Across 15 randomized trials, fermented soy products lowered total cholesterol by 5.0 mg/dL, fasting glucose by 6.39 mg/dL, waist circumference by 1.50 cm and BMI by 0.14 kg/m2 compared with controls, with other metabolic markers unchanged. — adults in 15 RCTs of fermented soy products (tempeh, natto, miso and similar), searched to March 2024 (n = 15 RCTs) Evidence A source
  WMD: BMI -0.14 kg/m2 (95% CI -0.28 to -0.01), waist -1.50 cm (-2.94 to -0.07), visceral fat -692 mm2 (-1012 to -373), fasting glucose -6.39 mg/dL (-10.38 to -2.40), total cholesterol -5.0 mg/dL (-6.60 to -3.39); other parameters not significant
• US labelling law still authorizes the claim that 25 g or more of soy protein a day, in a diet low in saturated fat and cholesterol, may reduce the risk of coronary heart disease. FDA proposed in 2017 to revoke that claim and the proposal still stands. A 100 g serving of tempeh carries about 20 g of protein. — foods labelled in the United States under 21 CFR 101.82; the daily intake level named in the regulation Evidence E source
  Regulatory threshold: 25 g/day of soy protein; the claim must use may or might and name the serving's contribution

Safety, limits and interactions

• Bongkrekic acid, the bacterial toxin first known from coconut tempeh, killed 2 of 33 people who ate contaminated wet rice noodles in Taipei in March 2024, through liver failure and cardiac arrest, while no soy tempeh outbreak appears in the corpus. — 33 patients poisoned in East Taipei, Taiwan, March 2024; two fatalities described (a 40-year-old woman and a 39-year-old man) (n = 33 (2 deaths)) Evidence C source
  Refractory hypotension, jaundice, acute hepatic failure, metabolic acidosis, coagulopathy; death on day 39 and after cardiac arrest on admission

What people believe that the data does not support

• Across 41 clinical studies with testosterone measured in 1,753 men, neither soy foods, soy protein nor isoflavone extracts changed total or free testosterone, estradiol or estrone at any dose or duration, so the claim that soy feminizes men is not supported. — adult men in 41 clinical studies published to April 2020; TT in 1753 men, FT in 752, E2 in 1000, E1 in 239, SHBG in 967 (n = 41 studies) Evidence A source
  No significant effect of soy protein or isoflavone intake on TT, FT, E2, E1 or SHBG under any statistical model; no effect by isoflavone dose or study duration

Who this works differently for

• Postmenopausal Malay women absorbed 36.9% of the calcium from a 206 g tempeh meal, no different from 34.3% from 114 g of milk, but because tempeh holds less calcium, four servings are needed to match the absorbed calcium of one 4-ounce glass of milk. — 20 postmenopausal Malay women, randomized crossover, 130 mg calcium per test meal, dual stable isotopes, 1-month washout (n = 20) Evidence B source
  Fractional calcium absorption 36.9 +/- 10.6% from tempeh vs 34.3 +/- 8.6% from milk, not significantly different
• Among iron-deficient pregnant women in 39 Indonesian villages, a weekly food supplement built on 600 g of tempeh plus guava, papaya and orange was associated with a smaller fall in ferritin (ratio 1.42, 95% CI 1.16 to 1.75) between mid-pregnancy and term than no intervention, while iron-replete women gained nothing. — 252 pregnant women at 12-20 weeks of gestation, cluster-randomized by village, supplementary food 6 days a week at home (n = 252) Evidence B source
  In Fe-deficient women: smaller decreases in Hb (1.02, 95% CI 0.98 to 1.07, P = 0.058), ferritin (1.42, 1.16 to 1.75, P = 0.046) and body iron (2.57, 1.71 to 3.43 mg/kg, P = 0.073); overall groups did not differ
• In 40 randomized trials of 3,285 postmenopausal women taking a median 75 mg a day of soy isoflavones for a median 24 weeks, endometrial thickness changed by -0.22 mm and estradiol by 1.61 pmol/L, neither significant, with high to moderate certainty that isoflavones are not estrogenic on these measures. — 3,285 postmenopausal women in 40 RCTs (52 comparisons) of at least 3 months, searched to August 2024, GRADE assessed (n = 40 RCTs, n = 3285) Evidence A source
  Endometrial thickness MD -0.22 mm (95% CI -0.45 to 0.01), vaginal maturation index 2.31 (-2.14 to 6.75), FSH -0.02 IU/L (-2.39 to 2.35), estradiol 1.61 pmol/L (-1.17 to 4.38); certainty high to moderate
• After a single 0.44 mg/kg isoflavone dose from soy milk, textured vegetable protein or tempeh in 57 adults, urinary equol excretion was higher after tempeh among equol producers, and urinary genistein recovery depended on sex and food matrix, with no food differences in men. — 20 premenopausal women, 17 postmenopausal women and 20 men, three single doses on separate occasions, four consecutive 24 h urines (n = 57) Evidence B source
  Tempeh raised urinary genistein recovery relative to soy milk in premenopausal women (P < 0.052); equol excretion higher after tempeh in equol producers (P < 0.01); no food differences in men

What is still unknown

• Since 31 October 2017 FDA has had a proposed rule on the books to revoke the soy protein and coronary heart disease health claim, stating that the totality of evidence no longer supports significant scientific agreement, and as of September 2026 it remains a proposed rule. — US food labelling; FDA review of the publicly available scientific evidence on soy protein and CHD Evidence E source
  Proposed revocation of 21 CFR 101.82; not finalized; the authorized claim continues to be usable meanwhile

The bottom line

Per 100 g, USDA puts tempeh at 20.29 g protein, 0.56 mg copper and 0.358 mg riboflavin, and the British CoFID record for the same food at 20.7 g, 0.67 mg and 0.48 mg. The riboflavin figures differ by a quarter, which is the size of the disagreement between two reference tables describing one product that is made in many ways. CoFID also lists 0.1 micrograms of vitamin B12 per 100 g. At that level tempeh is not a vitamin B12 food. Neither table carries a sample-to-sample spread for tempeh and there is no US survey portion weight for it, so a serving on this page means whatever a given study used. Nothing in this cohort measures what frying or steaming does to tempeh. The nearest thing is a USDA group factor for legumes boiled 15 to 20 minutes, drained and then fried, a 1975 row covering the whole legume group, which keeps 85 percent of iron and calcium, 70 percent of copper and riboflavin, 60 percent of thiamin and 55 percent of folate. Tempeh is usually fried or steamed without that long boil, so the percentages are borrowed from beans rather than measured on tempeh. Protein has no retention row at all. Two trials look at people rather than markers. Twenty postmenopausal Malay women absorbed 36.9 percent of the calcium from a 206 g tempeh meal and 34.3 percent from 114 g of milk, and the difference was not significant. Because tempeh holds less calcium to begin with, the authors calculate that four servings are needed to match the absorbed calcium of one 4-ounce glass of milk. In 39 Indonesian villages, 252 pregnant women were cluster-randomised to a weekly food package built on 600 g of tempeh with guava, papaya, orange and 30 g of meat. Among the iron-deficient women, ferritin fell less than with no intervention (ratio 1.42, CI 1.16 to 1.75), while iron-replete women gained nothing and the groups as a whole did not differ. The package was not tempeh alone, so the fruit and the meat carry part of whatever happened. A single 307 kcal tempeh breakfast lowered plasma acyl-ghrelin by 35 percent at 30 minutes in 13 Indonesian women aged 18 to 20 with a BMI of 25 to 30, where an isocaloric unfermented soybean breakfast did nothing. Their rated appetite was the same after both meals. Thirteen people, one meal, a hormone that moved and a hunger that did not. The fear that soy feminises men has been tested. Across 41 clinical studies, with total testosterone measured in 1,753 men and estradiol in 1,000, neither soy foods nor soy protein nor isoflavone extracts moved total or free testosterone, estradiol, estrone or SHBG under any statistical model, at any dose and over any duration. That paper notes industry affiliations for its authors. In 40 randomised trials of 3,285 postmenopausal women taking a median 75 mg of isoflavones a day for a median 24 weeks, endometrial thickness changed by minus 0.22 mm (CI minus 0.45 to 0.01) and estradiol by 1.61 pmol/L (CI minus 1.17 to 4.38), with the certainty of the evidence rated high to moderate. Most of that work used isolated protein and supplements at doses above what a serving of tempeh supplies, which limits how far it stretches to the food. On labels, US law still allows the statement that 25 g a day of soy protein, in a diet low in saturated fat and cholesterol, may reduce the risk of coronary heart disease, and 100 g of tempeh carries about 20 g of protein. FDA proposed on 31 October 2017 to revoke that claim, saying the evidence no longer supports significant scientific agreement, and in September 2026 the revocation is still a proposal. Both the claim and the doubt about it are current. The one safety card on this page is about a different food. Bongkrekic acid killed 2 of 33 people who ate contaminated wet rice noodles in Taipei in March 2024, through acute liver failure in one case and cardiac arrest in the other. That toxin is first known from coconut tempeh, so the link to the soy tempeh in your shop is historical rather than a recorded outbreak. We found no soy tempeh outbreak in this cohort and no data on interactions with medicines, so on drugs, on children and on pregnancy beyond the Indonesian trial above this page has nothing to tell you.

Sources

1. tmph-c5-01 · reference dataset
   FDC 174272 Tempeh: Protein 20.29 g | Copper 0.56 mg | Riboflavin 0.358 mg || CoFID 13-118 Tempeh: Energy 166 kcal | Fat 6.4 g | Protein 20.7 g | Calcium 120 mg | Copper 0.67 mg | Iron 3.60 mg | Riboflavin 0.48 mg | Vitamin B12 0.1 ug
   USDA FoodData Central SR Legacy, FDC 174272; CoFID 2021 (Public Health England, OGL v3.0), code 13-118 · checked 2026-09-23
2. tmph-c5-08 · randomized crossover trial
   While no change was observed after the non-fermented soybean meal, plasma acyl-ghrelin decreased by 35% at 30 min and remained below baseline until 120 min after the fermented soybean meal (P < 0.05)...No difference in subjective appetite was observed between the fermented soybean meal and the non-fermented soybean meal.
   Obes Res Clin Pract, 2021 · checked 2026-09-23
3. tmph-c5-04 · randomized crossover trial
   Consumption of tempeh (mainly isoflavone aglycon) resulted in higher serum peak levels of both daidzein (P < 0.001) and genistein (P < 0.01) and the associated area under the curve (P < 0.001 and P < 0.03, respectively) compared with textured vegetable protein (predominantly isoflavone glucosides).
   J Nutr, 2006 · checked 2026-09-23
4. tmph-c5-02 · reference dataset
   USDA-RETN-R6:505 LEGUMES,CKD 15/20MIN,BLD,DRND,FRIED (Legumes and Legume Products): folate 55% | vitamin C 60% | thiamin 60% | niacin 65% | vitamin B6 65% | copper 70% | riboflavin 70% | potassium 75% | magnesium 80% | calcium 85% | iron 85%
   USDA Table of Nutrient Retention Factors, Release 6 (2007), USDA-RETN-R6:505 · checked 2026-09-23
5. tmph-c5-05 · randomized crossover trial
   Fermented soy product (112 g tempeh) or unfermented soy (125 g soybean pieces) was consumed during each controlled feeding period...recovery of daidzein and genistein was significantly higher when subjects consumed the tempeh diet than when they consumed the soybean pieces diet (P < .002)...Fermentation of soy decreased the isoflavone content of the product fed but increased the urinary isoflavonoid recovery.
   J Am Diet Assoc, 1995 · checked 2026-09-23
6. tmph-c5-06 · meta-analysis of RCTs
   Results demonstrated significant beneficial effects of fermented soy on body mass index (WMD = -0.14 kg m-2, 95% CI: -0.28, -0.01, P = 0.039), waist circumference (WMD = -1.50 cm, 95% CI: -2.94, -0.07, P = 0.04), visceral fat (WMD = -692.17 mm2, 95% CI: -1011.58, -372.77, P < 0.001), fasting plasma glucose (WMD = -6.39 mg dL-1, 95% CI: -10.38, -2.40, P = 0.002), and total cholesterol (WMD = -5.0 mg dL-1, 95% CI: -6.60, -3.39, P < 0.001) compared with controls.
   Food Funct, 2025 · checked 2026-09-23
7. tmph-c5-11 · codified federal regulation
   The daily dietary intake level of soy protein that has been associated with reduced risk of coronary heart disease is 25 grams (g) or more per day of soy protein.
   21 CFR 101.82, Health claims: Soy protein and risk of coronary heart disease (CHD) · checked 2026-09-23
8. tmph-c5-14 · case report
   Bongkrekic acid (BKA) is a life-threatening toxin linked to foodborne illnesses. Herein, we report two fatalities among 33 patients who ingested BKA in March 2024, East Taipei, Taiwan.
   Taiwan J Obstet Gynecol, 2025 · checked 2026-09-23
9. tmph-c5-09 · meta-analysis of clinical trials
   A total of 41 studies were included in the analyses. TT and FT levels were measured in 1753 and 752 men, respectively; E2 and E1 levels were measured in 1000 and 239 men, respectively and SHBG was measured in 967 men. Regardless of the statistical model, no significant effects of soy protein or isoflavone intake on any of the outcomes measured were found.
   Reprod Toxicol, 2021 · checked 2026-09-23
10. tmph-c5-03 · randomized crossover trial
    Absorption of calcium from tempeh did not differ significantly from milk (36.9 +/- 10.6% vs. 34.3 +/- 8.6%, respectively). Due to differences in the calcium content of tempeh, four servings of this product would be needed to get the same amount of absorbed calcium as that obtained from a 4-ounce glass of milk.
    Int J Food Sci Nutr, 2010 · checked 2026-09-23
11. tmph-c5-07 · cluster-randomized trial
    The average weekly food provided comprised 600 g of tempeh, 30 g of meat, 350 g of guava, 300 g of papaya and 100 g of orange...In Fe-deficient women, consumption of an optimized diet was associated with smaller decreases in Hb (1·02 (95% CI 0·98, 1·07) g/l; P = 0·058), ferritin (1·42 (95% CI 1·16, 1·75) μg/l; P = 0·046) and body Fe (2·57 (95% CI 1·71, 3·43) mg/kg; P = 0·073) concentrations, compared with a state of no intervention.
    Public Health Nutr, 2011 · checked 2026-09-23
12. tmph-c5-10 · meta-analysis of RCTs
    We included 40 trials (52 trial comparisons, n = 3285) assessing the effect of a median reported dose of 75 mg/d of soy isoflavones in substitution for non-isoflavone controls over a median of 24 wk. Soy isoflavones had no statistically significant effect on any measure of estrogenicity; ET [mean difference, -0.22 mm (95% confidence interval, -0.45, 0.01 mm), PMD = 0.059], VMI [2.31 (-2.14, 6.75), PMD = 0.310], FSH [-0.02 IU/L (-2.39, 2.35 IU/L), PMD = 0.987], and estradiol [1.61 pmol/L (-1.17, 4.38 pmol/L), PMD = 0.256]. The certainty of evidence was high to moderate for all outcomes.
    Adv Nutr, 2025 · checked 2026-09-23
13. tmph-c5-13 · randomized crossover trial
    Tempeh consumption also resulted in an increased urinary genistein recovery relative to soya milk in premenopausal women (P<0.052). No differences in urinary genistein recoveries between soya foods were observed in the men...urinary equol excretion was higher (P<0.01) following tempeh ingestion among equol producers.
    Br J Nutr, 2004 · checked 2026-09-23
14. tmph-c5-12 · regulation notice (proposed rule)
    The Food and Drug Administration (FDA, the Agency, or we) is proposing to revoke its regulation authorizing the use of health claims on the relationship between soy protein and coronary heart disease on the label or in the labeling of foods...our tentative conclusion that such evidence does not support our previous determination that there is significant scientific agreement (SSA) among qualified experts for a health claim regarding the relationship between soy protein and reduced risk of coronary heart disease.
    Federal Register, 2017-23629, Food Labeling: Health Claims; Soy Protein and Coronary Heart Disease (proposed rule) · checked 2026-09-23