Soymilk: what it replaces, and what the carton was fortified with

Soymilk is the plant drink with a meta-analysis of randomised trials behind it, and those trials all test the same move, 500 mL of soymilk in place of 500 mL of cow's milk. Across 17 trials and 504 adults the swap lowered LDL cholesterol by 0.19 mmol/L and non-HDL cholesterol by 0.26 mmol/L, which the review itself calls a moderate reduction in non-HDL and a small though important one in LDL. The same pool puts systolic pressure 8.00 mmHg lower, with an interval running from 14.89 to 1.11 mmHg, so treat that one as a wide range rather than a number to plan around. Which carton you pick decides most of the rest. The 11 USDA records for soymilk run from 0 to 1.08 ug of vitamin B12 per 100 g, and that spread is between products rather than between samples of one product, since none of the 11 is a measured Foundation entry. The fortification line on the label decides whether this drink stands in for milk.

The numbers, per 100 g

How much is actually in it

• The 11 USDA soymilk records run from 0 to 1.08 ug of vitamin B12 per 100 g, so whether this drink replaces milk nutritionally is decided by the fortification line on the carton and not by the soy. — 11 USDA SR Legacy records for soymilk, reference record Soymilk (All flavors), enhanced, FDC 173765 (n = 11 records) Evidence E source
  Vitamin B12 0 to 1.08 ug per 100 g across the 11 records, niacin 0.222 to 3.292 mg across 9, vitamin E 0.01 to 2.52 mg across 9

What your body absorbs

• Which calcium salt the maker uses decides the answer: absorption from carbonate-fortified soymilk was 0.211 and matched cow's milk at 0.217, while tricalcium phosphate soymilk gave only 0.181. — 20 healthy women aged about 23, three-way crossover with dual stable isotopes, 250 mg calcium per test (n = 20) Evidence B source
  Fractional calcium absorption 0.211 for carbonate-fortified soymilk, 0.217 for cow's milk, 0.181 for tricalcium phosphate soymilk (P < 0.05 against the first two)
• In 12 osteopenic post-menopausal women an Australian phosphate-fortified soymilk matched cow's milk for hourly calcium absorption, 0.65 against 0.66. — 12 osteopenic post-menopausal women, mean age 56.7 years, randomised single-blind acute crossover (n = 12) Evidence B source
  Hourly fractional calcium absorption 0.65 from fortified soymilk and 0.66 from cow's milk, p > 0.05

What cooking and storage change

• Fermenting the same soymilk changes what it does: 100 ml a day of soymilk fermented with Lacticaseibacillus paracasei for 4 weeks raised faecal lactobacilli in 112 Japanese adults against the unfermented base. — 112 healthy Japanese adults with a low faecal Bifidobacterium count, 100 ml daily for 4 weeks, double-blind (n = 112) Evidence B source
  Higher counts and relative abundance of faecal lactobacilli against unfermented soymilk placebo, with greater increases in faecal acetate in the subgroup analysis

What it does to your health

• Pooling 17 randomised trials in 504 adults, swapping 500 mL of cow's milk for 500 mL of soymilk lowered non-HDL cholesterol by 0.26 mmol/L, systolic pressure by 8.0 mmHg and LDL cholesterol by 0.19 mmol/L. — 504 adults with a mix of health statuses in 17 randomised trials of at least 3 weeks (n = 504) Evidence A source
  Non-HDL-C -0.26 mmol/L (95% CI -0.43 to -0.10), systolic BP -8.00 mmHg (-14.89 to -1.11), diastolic BP -4.74 mmHg (-9.17 to -0.31), LDL-C -0.19 mmol/L (-0.29 to -0.09), CRP -0.82 mg/L (-1.26 to -0.37)
• Across 52 observational studies, people with the highest soymilk intake had 25% lower cancer risk overall (RR 0.75, 95% CI 0.60 to 0.93), an association and not a demonstrated effect. — 52 studies, 17 cohorts and 35 case-control, highest against lowest intake categories (n = 52 studies) Evidence C source
  Total cancer RR 0.75 (95% CI 0.60 to 0.93) for soymilk, 0.69 (0.60 to 0.80) for total soy products, 0.78 (0.70 to 0.86) for tofu
• In 240 Thai adults with LDL of 130 mg/dl or higher, soymilk enriched with 2 g of phytosterols and 10 g of inulin a day cut LDL by 10.0% in 8 weeks against 1.3% on plain soymilk. — 240 adults aged 18 and over with baseline LDL-c of 130 mg/dl or higher, 8 weeks, double-blind (n = 240) Evidence B source
  LDL-c -10.03% against -1.31% on standard soymilk (p < 0.001), total cholesterol -6.60% against -1.76%, no difference in triglycerides or HDL-c
• Two 28-day controlled feeding trials found plant sterol enriched soy beverages cut total and LDL cholesterol by 10% to 15% against 1% dairy milk, and measured the mechanism as 27% less cholesterol absorbed. — 33 normal to mildly high cholesterol adults in study 1 and 23 hypercholesterolemic adults in study 2, both 28-day crossovers, 1.95 g plant sterols a day (n = 56) Evidence B source
  Total and LDL cholesterol -10% and -13% with the low-fat beverage, -12% and -15% with the moderate-fat one, cholesterol absorption -27%
• Three servings a day of plain vanilla soymilk for 4 weeks changed no blood lipid in 32 postmenopausal women, the cleanest available test of soymilk without added sterols. — 32 postmenopausal women, 16 on soy and 16 on reduced-fat dairy milk, 4 weeks after a 2-week run-in (n = 32) Evidence B source
  No difference in HDL, LDL or triglycerides between groups (P = .45) or from baseline (P = .83), total cholesterol likewise unchanged
• The companion trial in 31 postmenopausal women found 4 weeks of soymilk moved none of six markers of inflammation or oxidative stress, with p values from .08 to .99. — 31 postmenopausal women, 16 on soy and 15 on reduced-fat dairy milk, 3 servings a day for 4 weeks (n = 31) Evidence B source
  No change in TNF-alpha, IL-1beta, IL-6, superoxide dismutase, glutathione peroxidase or COX-2 between groups
• Drinking soymilk with white bread pulled the glycemic index of that bread down from 100 to 77.5 in 17 healthy men, while the insulin response did not change. — 17 healthy Asian men with BMI 18.5 to 29, randomised crossover, four test occasions (n = 17) Evidence B source
  Glycemic index of white bread 83.1 with rice bran soymilk and 77.5 with sugar-free soymilk against 100 for bread alone (p < 0.05), insulin response unchanged

Safety, limits and interactions

• Among 89 people who had systemic allergic reactions to normally mild plant allergens, the offending food was liquid soymilk in 15% of them against 2% of controls. The authors say soy beverages should be avoided by patients hypersensitive to PR-10 proteins. — 89 patients with a documented systemic reaction to labile plant food allergens and 81 age-matched controls with oral allergy syndrome, multicentre (n = 170) Evidence C source
  Liquid offending food (soymilk) in 15% of patients against 2% of controls (P < .025), proton pump inhibitor use in 19% of patients against 5% of controls (P < .025)
• A 22-month-old on a soymilk-based restrictive diet developed goiter and overt hypothyroidism, which reversed once iodine was supplied and the diet widened. — One toddler aged 22 months with cow's milk protein allergy, on soymilk since 12 months of age, in an iodine-deficient state (n = 1) Evidence C source
  Goiter and overt hypothyroidism requiring temporary levothyroxine, fully reversed by iodine supplementation and diet diversification
• Emergency physicians reported acute pancreatitis after daily large ingestions of soymilk, with the soybean trypsin inhibitor as the proposed mechanism. — One adult woman presenting to an emergency department after daily large soymilk intake (n = 1) Evidence C source
  Acute pancreatitis attributed to the trypsin inhibitor removing negative feedback on pancreatic acinar cells

What people believe that the data does not support

• The same 17 trials used soymilk carrying 6.9 to 17.2 g of added sugar per 250 mL, and sweetened soymilk changed none of the 19 outcomes for the worse, which is the direct test of the claim that sweetened plant milk is the harmful part. — 504 adults, median daily dose 500 mL of soymilk with 22 g soy protein against 500 mL of cow's milk (n = 504) Evidence A source
  No meaningful effect modification by added sugars across 19 intermediate cardiometabolic outcomes
• EFSA reviewed isoflavone supplements for peri- and post-menopausal women and reported no evidence of harm to the breast, uterus or thyroid, which is the regulator's answer to the oldest fear about soy. — Peri- and post-menopausal women taking food supplements with isolated isoflavones Evidence E source
  No harmful effects found on mammary gland, uterus or thyroid gland at the levels typically found in such supplements

Who this works differently for

• In 100 adults with type 2 diabetes randomised for 6 weeks, the soymilk plus placebo arm ended with lower diastolic pressure, insulin and triglycerides and higher HDL-C. — 100 patients with type 2 diabetes aged 40 to 75 in Shiraz, Iran, four arms, 6 weeks, 92 completed (n = 100) Evidence B source
  Soymilk with placebo lowered diastolic blood pressure (p = 0.01), insulin (p = 0.006) and triglycerides (p = 0.001) and raised HDL-C (p = 0.03)

What is still unknown

• The dose-response in that same analysis does not hold together: 23 g of soymilk a day, about two tablespoons, is credited with 28% lower cancer risk while 54 g of soy products overall gets 11%. — Same 52 observational studies, dose-response models per food (n = 52 studies) Evidence C source
  23 g/day of soymilk credited with 28% lower cancer risk against 11% for 54 g/day of total soy products and 12% for 61 g/day of tofu

The bottom line

Plain soymilk and fortified soymilk are different products, and the trials show it. Three servings a day of plain vanilla soymilk for four weeks moved no blood lipid in 32 postmenopausal women, with no difference between groups (P = .45) or from baseline (P = .83). The companion trial in 31 women found none of six markers of inflammation or oxidative stress had shifted after the same four weeks, p values from .08 to .99. Both are small and short, so they rule out a large effect rather than any effect. Now add plant sterols. In 240 Thai adults starting from an LDL of 130 mg/dl or higher, soymilk carrying 2 g of phytosterols and 10 g of inulin a day cut LDL by 10.03 percent in eight weeks against 1.31 percent on standard soymilk (p < 0.001). Two 28-day crossover feeding trials in 56 adults, using 1.95 g of plant sterols a day, cut total and LDL cholesterol by 10 to 15 percent against 1 percent dairy milk and measured 27 percent less cholesterol absorbed. The sterol is doing that work. Calcium is the other thing a label hides. In 20 healthy women of about 23, a single 250 mg dose absorbed at a fraction of 0.211 from carbonate-fortified soymilk and 0.217 from cow's milk, while tricalcium phosphate soymilk managed only 0.181 (P < 0.05). In 12 osteopenic post-menopausal women, an Australian phosphate-fortified soymilk matched cow's milk for hourly absorption, 0.65 against 0.66, with a proprietary fortificant that is not the same salt. Most cartons never name the salt, so a shopper often cannot tell which number applies. Two other things change the drink itself. Soymilk fermented with Lacticaseibacillus paracasei, 100 ml a day for four weeks, raised faecal lactobacilli in 112 Japanese adults against the unfermented base used as placebo. Drinking sugar-free soymilk with white bread pulled the glycemic index of that bread from 100 down to 77.5 in 17 healthy men, which is a property of the meal on one test day rather than of a diet. The idea that the sugar in plant milk cancels the benefit has been tested. The soymilk in those 17 trials carried 6.9 to 17.2 g of added sugars per 250 mL, and added sugars did not meaningfully modify any of the 19 cardiometabolic outcomes. That is an absence of effect modification across trials rather than a trial built to compare sweetened with unsweetened, and the authors had no eligible study for the lactose against sucrose comparison they planned. The cancer numbers deserve more caution than they usually get. Fifty-two observational studies put the highest soymilk intake at 25 percent lower cancer risk overall (RR 0.75, 95% CI 0.60 to 0.93), an association from data where two thirds of the studies asked people with cancer to recall their old diet. The same analysis credits 23 g of soymilk a day, about two tablespoons, with 28 percent lower risk while 54 g of soy products overall gets 11 percent, and a bigger effect from a smaller amount of the same plant is a sign of confounding. Some people need a different answer. In 100 adults with type 2 diabetes in Iran, six weeks of soymilk lowered diastolic pressure (p = 0.01), insulin (p = 0.006) and triglycerides (p = 0.001) and raised HDL, though most of those p values describe change within a group of about 24 people rather than a difference between arms. People sensitised to the birch-related PR-10 proteins are the group with a real warning: among 89 patients who had systemic reactions to normally mild plant allergens, the offending food was liquid soymilk in 15 percent against 2 percent of controls, and those authors say soy beverages should be avoided by such patients. A 22-month-old kept on a restrictive soymilk diet in an iodine-deficient state developed goiter and hypothyroidism that reversed with iodine and a wider diet, which is one child and a warning about soy drinks used as infant milk replacements. One emergency department reported acute pancreatitis after daily large ingestions of soymilk in a single adult. Against the oldest fear, EFSA reviewed concentrated isoflavone supplements in peri- and post- menopausal women and found no evidence of harm to breast, uterus or thyroid, in concentrated supplements rather than in a drink. We found no soymilk trial in pregnancy and none in kidney disease.

Sources

1. soym-r2-18 · USDA FoodData Central SR Legacy, dataset
   Soymilk (All flavors), enhanced (FDC 173765, SR Legacy, Legumes and Legume Products): vitamin B12 1.08 ug per 100 g = 45% DV; across the 11 soymilk records vitamin B12 ranges 0 to 1.08 ug, niacin 0.222 to 3.292 mg over 9 records, vitamin E 0.01 to 2.52 mg over 9 records; Foundation samples 0; CoFID match none; FNDDS portion none
   USDA FoodData Central, SR Legacy, FDC 173765 · checked 2026-09-24
2. soym-r2-09 · controlled crossover trial, isotope absorption study
   Fractional Ca absorption in CCSM (0.211 +/- 0.057) did not differ from that of cow's milk (0.217 +/- 0.040), but both were higher (P < 0.05) than that of TCPSM (0.181 +/- 0.039).
   J Nutr, 2005 · checked 2026-09-24
3. soym-r2-10 · acute crossover absorption trial
   The mean hourly fractional calcium absorption from fortified soymilk was found to be comparable to that of cows' milk
   Asia Pac J Clin Nutr, 2010 · checked 2026-09-24
4. soym-r2-12 · randomised controlled trial, fermented against unfermented soymilk
   They consumed 100 ml FSM or placebo (unfermented soymilk base) once daily for 4 weeks...The FSM group had a significantly higher number and relative abundance of faecal lactobacilli compared with the placebo group.
   Benef Microbes, 2024 · checked 2026-09-24
5. soym-r2-01 · systematic review and meta-analysis of randomised trials
   The substitution of soymilk for cow's milk resulted in moderate reductions in non-HDL-C (mean difference, - 0.26 mmol/L [95% confidence interval, - 0.43 to - 0.10]), systolic blood pressure (- 8.00 mmHg [- 14.89 to - 1.11]), and diastolic blood pressure (- 4.74 mmHg [- 9.17 to - 0.31]); small important reductions in LDL-C (- 0.19 mmol/L [- 0.29 to - 0.09]) and c-reactive protein (CRP) (- 0.82 mg/L [- 1.26 to - 0.37]); and trivial increases in HDL-C (0.05 mmol/L [0.00 to 0.09]).
   BMC Med, 2024 · checked 2026-09-24
6. soym-r2-03 · systematic review and meta-analysis of cohort and case-control studies
   High consumption of total soy products (RR: 0.69; 95% CI: 0.60, 0.80), tofu (RR: 0.78; 95% CI: 0.70, 0.86), and soymilk (RR: 0.75; 95% CI: 0.60, 0.93) were associated with reduced total cancer risk.
   Nutrients, 2024 · checked 2026-09-24
7. soym-r2-05 · randomised controlled trial of a fortified beverage
   Subjects were randomly assigned into the study group that received 2 g/day of phytosterols and 10 g/day of inulin-enriched soymilk or into the control group that received standard soymilk...The LDL-c reduction was significantly better in the study group (-10.03%, (-37.07, 36.00) vs -1.31% (-53.40, 89.73), p < 0.001).
   Lipids Health Dis, 2015 · checked 2026-09-24
8. soym-r2-06 · randomised crossover trials of a fortified beverage
   In comparison with the 1% dairy milk control, the low-fat soy beverage reduced (P < 0.05) total and LDL-cholesterol by 10 and 13%, respectively...Consumption of the moderate-fat PS-enriched soymilk reduced (P < 0.05) cholesterol absorption by 27%, but did not alter cholesterol synthesis in comparison with 1% dairy milk.
   Lipids Health Dis, 2009 · checked 2026-09-24
9. soym-r2-07 · randomised controlled trial of unfortified soymilk
   After a 2-week run-in period, subjects were randomly assigned to consume three servings of vanilla soy (n = 16) or reduced-fat dairy (n = 16) milk per day for 4 weeks...Plasma high-density lipoprotein, LDL, and triglycerides were not significantly different between groups post-intervention (P = .45) or from baseline (P = .83).
   J Med Food, 2010 · checked 2026-09-24
10. soym-r2-08 · randomised controlled trial of unfortified soymilk
    Subjects were randomly assigned to consume 3 servings of vanilla soy (n = 16) or reduced fat dairy (n = 15) milk per day for 4 weeks...No significant differences were observed for any of the plasma inflammatory (TNF-alpha, P = .08; IL-1beta, P = .14; IL-6, P = .26) or oxidative stress (SOD, P = .68; GPx, P = .58; COX-2, P = .99) variables by dietary treatment group.
    Nutr Res, 2009 · checked 2026-09-24
11. soym-r2-11 · randomised crossover postprandial trial
    Taking the glycemic index (GI) value of white bread as a reference value of 100, the GI of white bread when co-ingested with rice bran soymilk (RBS) was 83.1 (±7.7) and sugar-free soymilk (SFS) was 77.5 (±10.1), both were lower than white bread (p < 0.05).
    Nutrients, 2018 · checked 2026-09-24
12. soym-r2-14 · retrospective multicentre study of allergic reactions
    The ingestion of the offending food in liquid form (soymilk) was frequent among patients (15%) but unusual among controls (2%; P < .025)...Soybean beverages represent a risk for PR-10 hypersensitive patients and should be avoided.
    Allergy, 2021 · checked 2026-09-24
13. soym-r2-15 · case report
    A 22-month-old toddler with a previous diagnosis of cow's milk protein allergy came to clinical attention for important goiter and overt hypothyroidism. Detailed dietary anamnesis revealed that he was on a restrictive dietary regimen based on soymilk since 12 months of age.
    Front Endocrinol (Lausanne), 2022 · checked 2026-09-24
14. soym-r2-16 · case report
    We present a case of daily, large ingestions of soymilk that likely led to acute pancreatitis. Soybean contains trypsin inhibitor that when ingested will reduce the activity of trypsin in the intestine.
    Am J Emerg Med, 2021 · checked 2026-09-24
15. soym-r2-02 · systematic review and meta-analysis of randomised trials
    Eligibility criteria were met by 17 trials (n = 504 adults with a range of health statuses), assessing the effect of a median daily dose of 500 mL of soymilk (22 g soy protein and 17.2 g or 6.9 g/250 mL added sugars) in substitution for 500 mL of cow's milk (24 g milk protein and 24 g or 12 g/250 mL total sugars as lactose) on 19 intermediate outcomes...There was no meaningful effect modification by added sugars across outcomes.
    BMC Med, 2024 · checked 2026-09-24
16. soym-r2-17 · opinion of a European agency panel
    The evidence reviewed does not suggest there are harmful effects on the three organs considered for this assessment – mammary gland, uterus and thyroid gland.
    EFSA news release, isoflavones in food supplements, 21 October 2015 · checked 2026-09-24
17. soym-r2-13 · randomised clinical trial in type 2 diabetes
    One hundred patients with T2DM (aged 40-75 years old) were randomly assigned into 4 groups (soymilk + probiotics supplement, soymilk + placebo, conventional milk + placebo, and probiotics supplement) for 6 weeks...Soymilk + placebo administration significantly decreased DBP (p = 0.01), insulin (p = 0.006), and TG (p = 0.001) levels and significantly increased HDL-C (p = 0.03) levels.
    BMC Endocr Disord, 2023 · checked 2026-09-24
18. soym-r2-04 · systematic review and meta-analysis of cohort and case-control studies
    A 54 g per day increment of total soy products reduced cancer risk by 11%, a 61 g per day increment of tofu reduced cancer risk by 12%, and a 23 g per day increment of soymilk reduced cancer risk by 28%, while none of the other soy products were associated with cancer risk.
    Nutrients, 2024 · checked 2026-09-24