Textured vegetable protein: what the lab tests show

Textured vegetable protein is soy flour pushed through an extruder under heat and pressure until it holds a chewy shape. In a laboratory test of three matched meals, in vitro digestibility was 21.05 percent for the textured soy meal, 30.72 percent for the meat meal and 20.17 percent for the same legume protein left untextured. That number comes out of a test tube. It is a laboratory proxy, and it does not measure how much protein a person absorbs. The meat meal and the textured meal could not be separated statistically (p = 0.0832), while the untextured one stayed lower (p = 0.0247). Three meals in a lab cannot show that two foods digest alike. All they show here is that no difference turned up. In the human part of the same trial, 25 healthy men ate 758 kJ at a free meal 150 minutes after a bolognaise made with the textured protein, against 957 kJ after the meat version. That is one meal, in young healthy men, and how hungry they said they felt did not differ between the meals. Two things this page cannot tell you. We hold no USDA composition record for this food in this cohort, so there are no nutrient numbers here and no way to show how far brands differ. And the isoflavone studies below gave people a single measured dose, 0.44 mg per kg of body weight, then read blood and urine for a day. That is a laboratory dose given once to healthy volunteers, and nobody eats that way.

What your body absorbs

• Tempeh raised peak serum daidzein and genistein more than textured vegetable protein at the same single dose of 0.44 mg isoflavones per kg body weight in 59 healthy adults. — 59 healthy adults: 21 premenopausal women, 17 postmenopausal women, 21 men (n = 59) Evidence B source
  Higher peak daidzein (P < 0.001) and genistein (P < 0.01) and higher AUC after tempeh than after TVP
• In women, urinary genistein recovery was higher after soya milk than after textured vegetable protein given as a single dose of 0.44 mg isoflavones per kg body weight. — 20 premenopausal women, 17 postmenopausal women and 20 men (n = 57) Evidence B source
  Higher urinary genistein recovery after soya milk than after TVP in women (P<0.05), no difference in men
• A diet supplying 40 g of dietary fibre a day lowered plasma genistein by 55% at 24 hours after a single soy dose compared with a 15 g fibre control diet in seven healthy women. — 7 healthy women, single dose of 0.9 mg isoflavones per kg body weight from tofu or TVP (n = 7) Evidence B source
  55% lower plasma genistein at 24 h (P < 0.05), 20% lower total urinary genistein (P < 0.03)

What cooking and storage change

• In a laboratory test of three meals, in vitro protein digestibility was 21.05% for the textured vegetable protein meal, 30.72% for the meat meal and 20.17% for the non-textured legume meal. — three test meals matched for weight, energy, macronutrients and fibre, analysed in vitro (n = 3 meals) Evidence D source
  IVPD 21.05% for TVP, 30.72% for meat, 20.17% for non-textured legume protein
• In vitro digestibility of the extruded plant protein meal was not significantly different from the meat meal on three test meals (p = 0.0832), while the non-textured legume meal was lower than the meat meal (p = 0.0247). — three test meals of a 25 man randomised crossover trial, analysed for in vitro protein digestibility (n = 3 meals) Evidence D source
  Meat against non-textured legume p = 0.0247, meat against TVP p = 0.0832

What it does to your health

• At a single free meal 150 minutes after a bolognaise made with textured vegetable protein, 25 healthy men ate 758 kJ against 957 kJ after the same meal made with meat, with no difference in the appetite they reported. — 25 healthy men, randomised three way crossover, meals matched for weight, energy, macronutrients and fibre (n = 25) Evidence B source
  758 kJ after TVP against 957 kJ after meat and 903 kJ after non-textured legume protein
• Over 12 weeks of resistive training, vastus lateralis cross-sectional area rose 4.2% on a textured vegetable protein diet and 6.0% on a beef diet in 21 men aged 65, a difference that did not reach significance in a trial of this size. — 21 men aged 65 plus or minus 5 years, 12 weeks of resistive training, 0.6 g protein per kg per day from TVP or from beef as the test share of a total intake of 1.03 to 1.17 g per kg per day (n = 21) Evidence B source
  4.2% against 6.0% increase in muscle cross-sectional area, no significant difference between groups

Safety, limits and interactions

• Textured vegetable protein fed for 6 to 7 weeks lowered HDL cholesterol (P=.004) and raised serum triglyceride (P=.010) in eight subjects living on a metabolic ward. — 8 subjects on a metabolic ward, three randomly ordered 6 to 7 week diet periods (n = 8) Evidence B source
  LDL cholesterol down at borderline significance (P=.058), HDL cholesterol down (P=.004), triglyceride up (P=.010)
• 60 g of textured vegetable protein a day, carrying 45 mg of conjugated isoflavones, lengthened the follicular phase (P < 0.01) and delayed peak progesterone in premenopausal women over one menstrual cycle. — 15 healthy nonvegetarian premenopausal women studied over 9 months in a metabolic suite, 6 of them on 60 g TVP per day (n = 15) Evidence B source
  Follicular phase lengthened (P < 0.01), mid-cycle LH and FSH peaks suppressed (P < 0.05 and P < 0.01), no effect from the isoflavone-extracted product

What people believe that the data does not support

• FDA proposed in 2017 to revoke the health claim that soy protein reduces the risk of coronary heart disease, saying the totality of the evidence no longer shows significant scientific agreement. — US food labelling, 21 CFR 101.82 Evidence E source
  Proposed revocation of an authorised health claim, still a proposal as of the retrieval date

Who this works differently for

• Only 30% of 59 adults fed textured vegetable protein, soy milk and tempeh were equol producers, and equol production did not differ by age or gender. — 59 healthy adults: 21 premenopausal women, 17 postmenopausal women, 21 men (n = 59) Evidence B source
  30% equol producers, no difference by age or gender

What is still unknown

• The trial that compared absorption from textured vegetable protein at 0.44 mg isoflavones per kg body weight states that higher intakes and the relative weight of age and food matrix remain undefined. — 20 premenopausal women, 17 postmenopausal women and 20 men, single bolus dose (n = 57) Evidence B source
  Authors name the open questions rather than answering them

The bottom line

If you are eating soy for the isoflavones, this is the least efficient of the soy foods tested. At the same 0.44 mg per kg dose, tempeh produced higher peak daidzein (P < 0.001) and genistein (P < 0.01) in the blood of 59 adults. In women, soya milk put more genistein into the urine than the textured product did (P < 0.05), with no such gap in men. What else is on the plate moves the number as much as the food does. A diet carrying 40 g of fibre a day cut plasma genistein by 55 percent at 24 hours against a 15 g fibre day, in seven women. Only 30 percent of those 59 adults produced equol at all, with no difference by age or sex. For building muscle, the one trial we hold found no difference, and it was too small to rule one out. Twenty one men around the age of 65 trained with weights for twelve weeks. Each took 0.6 g of protein per kg a day from textured vegetable protein or from beef, and that 0.6 g is the share coming from the test food alone. Everything they ate brought the total to 1.03 to 1.17 g per kg a day in both groups, which is an adequate intake. So this compares two sources of protein at a normal amount, and says nothing about eating less protein. Thigh muscle cross-sectional area rose 4.2 percent on the soy diet and 6.0 percent on the beef diet, and that gap did not reach statistical significance. Twenty one men is a small trial. It was never designed to test the two diets as equal, and a real difference of that size could have passed it unnoticed. Two findings cut against the food, and they stay on this page. Eight people living on a metabolic ward ate it for six to seven weeks. Their HDL cholesterol fell (P = .004) and their triglyceride rose (P = .010). Fifteen premenopausal women were studied over nine months in a metabolic suite, and six of them took 60 g a day, carrying 45 mg of conjugated isoflavones, for one cycle each. On that arm the follicular phase lengthened (P < 0.01) and the peak of progesterone came later. The same work fed a version with the isoflavones taken out and saw no such effect. Both studies are small and both fed people under close control, so how far either carries into ordinary eating is open. If you are pregnant, trying to conceive, or being treated for anything hormonal, that second finding is worth taking to the person treating you rather than deciding alone. The heart claim is weaker than the packet suggests. In 2017 the FDA proposed to revoke the authorised claim that soy protein reduces the risk of coronary heart disease, on the grounds that the totality of the evidence no longer shows significant scientific agreement. When we checked, it was still a proposal. The holes are worth naming plainly. With no USDA record we can tell you nothing about protein, iron or sodium per 100 g, and nothing about how much those vary between the dry granules, the chunks and the ready meals. Nothing in what we hold says what a normal portion of this is in a normal week, so every number above is tied to a study diet rather than to your dinner. The authors of the absorption trial write the same kind of thing about their own work: what happens at higher intakes, and how much of the difference belongs to age against the food itself, is undefined.

Sources

1. tvp-r4-01 · randomised crossover
   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
2. tvp-r4-02 · randomised crossover
   Twenty premenopausal women, seventeen post-menopausal women and twenty men received a defined single oral bolus dose (0.44 mg isoflavones/kg body weight) of soya milk, textured vegetable protein (TVP) or tempeh on three separate occasions...For women the urinary genistein recovery was higher following soya-milk consumption compared with TVP (P<0.05).
   Br J Nutr, 2004 · checked 2026-09-23
3. tvp-r4-03 · randomised crossover
   The fiber-rich diet produced 55% lower plasma genistein at 24 h after soy dosing (P < 0.05) and reduced total urinary genistein by 20% (P < 0.03).
   J Nutr, 1996 · checked 2026-09-23
4. tvp-r4-05 · in vitro analysis inside an RCT
   The IVPD was significantly higher in the Meat meal (30.72%) than in the Green meal (20.17%), with the TVP meal in between (21.05%).
   Nutrients, 2024 · checked 2026-09-23
5. tvp-r4-06 · in vitro analysis inside an RCT
   The IVPD was higher in the Meat meal than in the Green meal (p = 0.0247), though not different from the TVP meal (p = 0.0832).
   Nutrients, 2024 · checked 2026-09-23
6. tvp-r4-04 · randomised crossover
   The adlib_EI was significantly lower after the TVP meal (758 kJ) than after the Meat meal (957 kJ), with the Green meal in between (903 kJ).
   Nutrients, 2024 · checked 2026-09-23
7. tvp-r4-07 · randomised controlled trial
   With resistive training, cross-sectional muscle area of the vastus lateralis increased in both groups (4.2 +/- 3.0% and 6.0 +/- 2.6% for the LOV and BC groups, respectively) with no significant difference between groups.
   Am J Clin Nutr, 2002 · checked 2026-09-23
8. tvp-r4-08 · randomised controlled feeding trial
   TVP resulted in a borderline significant reduction in LDL cholesterol (P=.058) with a highly significant reduction in HDL cholesterol (P=.004) and an increase in serum triglyceride (P=.010).
   Metabolism, 1999 · checked 2026-09-23
9. tvp-r4-09 · randomised controlled feeding trial
   Follicular phase length was significantly (P < 0.01) increased and peak progesterone concentrations were delayed with 60 g TVP but no effects were observed with Arcon F.
   Br J Nutr, 1995 · checked 2026-09-23
10. tvp-r4-10 · regulatory 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
    US Food and Drug Administration, Federal Register, 2017 · checked 2026-09-23
11. tvp-r4-12 · randomised crossover
    Only 30% of the subjects were equol producers and no differences in equol production with age or gender were observed.
    J Nutr, 2006 · checked 2026-09-23
12. tvp-r4-11 · randomised crossover
    Further studies are required to examine the effect of higher intake and define the relative influence of these factors in elderly population groups.
    Br J Nutr, 2004 · checked 2026-09-23