Milk or soy milk: what changes when you swap the carton

A meta-analysis of 17 randomized trials in 504 adults tested exactly this swap: about 17 fl oz (500 mL) of soymilk a day in place of 17 fl oz (500 mL) of cow's milk, for at least three weeks. Non-HDL cholesterol fell by 10 mg/dL (95% CI 3.9 to 17; metric: 0.26 mmol/L, 0.10 to 0.43), a result the reviewers rated high certainty. Systolic blood pressure fell by 8 mm Hg, with an interval running from 1 to 15 mm Hg, and they rated that one only moderate because the interval is so wide. Within those trials, whether the soymilk had added sugar made no meaningful difference to any of the 19 outcomes. That is not evidence that added sugar is harmless: all 19 were blood markers and pressure, not heart attacks, and the reviewers found no trial that set lactose against added sugar directly. On the label the two cups sit close. A 244 g cup of whole milk gives about 149 calories, 8 g protein and 300 mg calcium. A cup of unsweetened soy milk gives about 93 calories, 8.7 g protein and 246 mg calcium, calcium the maker adds, so it varies by brand.

How much is actually in it

• A 244 g cup of whole milk gives about 149 calories, 8 g protein and 300 mg calcium, while a cup of unsweetened soy milk gives about 93 calories, 8.7 g protein and 246 mg calcium. — USDA survey foods Milk, whole (FNDDS 11111000) and Soy milk, unsweetened (FNDDS 11320100) Evidence E source
  Per 100 g whole milk vs unsweetened soy milk: energy 61 vs 38 calories, protein 3.27 vs 3.55 g, calcium 123 vs 101 mg, total sugars 4.81 vs 0.56 g, saturated fat 1.86 vs 0.314 g, vitamin D 1.1 vs 0.7 ug; 1 cup = 244 g for both
• Calcium in sweetened refrigerated soy milk varied more between USDA samples than in whole cow's milk: 138 to 188 mg per 3.5 oz (100 g) for soy against 116 to 130 mg for whole milk. — USDA Foundation Foods samples: Soy milk, sweetened, plain, refrigerated (14 samples); Milk, whole, 3.25% milkfat, with added vitamin D (11 samples) (n = 25 samples) Evidence E source
  Calcium: soy milk 154.8 mg mean (138-188); whole milk 123 mg mean (116-130); protein: soy milk 2.56-3.38 g (sweetened) and 3.0-4.69 g (unsweetened), whole milk 2.93-3.51 g

What a real portion looks like

• The 2025-2030 US Dietary Guidelines advise including full-fat dairy with no added sugars and set a goal of 3 dairy servings a day on a 2,000-calorie pattern. — US population, federal dietary guidance Evidence E source
  3 dairy servings/day at 2,000 calories; full-fat dairy with no added sugars
• The guidelines' serving guide counts 1 cup of milk as one dairy serving and lists lactose-free products and fortified dairy alternatives as options in the dairy group. — US population, federal dietary guidance Evidence E source
  1 serving = 1 cup milk, 3/4 cup yogurt or 1 oz cheese; fortified dairy alternatives 'also options'
• The federal review behind the 2025-2030 guidelines marked the advisory committee's call for fat-free, low-fat and fortified soy options from age 2 as not implemented, and whole milk or fortified unsweetened soy beverage from 12 months as partly implemented. — Children from 12 months; people aged 2 and older Evidence E source
  Rec. 27 Partial; rec. 28 No (columns Yes / Partial / No)

What your body absorbs

• Calcium from dairy and from calcium-fortified foods such as soy beverages is absorbed at about the same rate, roughly 30%, according to the NIH fact sheet. — general population; fortified milk substitutes such as soy and almond beverages Evidence E source
  Calcium absorption about 30% from dairy products and fortified foods

What cooking and storage change

• Heating milk for about 10 minutes keeps all of its calcium but costs about a fifth of vitamin B12 and 15% of folate, by the USDA retention table. — USDA retention factors for Dairy and Egg Products, milk heated approximately 10 minutes Evidence E source
  Retained: calcium, riboflavin, potassium 100%; vitamin B12 80%; folate 85%; vitamin C 85%; thiamin 90%

What it does to your health

• Across 17 randomized trials in 504 adults, swapping about 17 fl oz (500 mL) a day of cow's milk for soymilk lowered non-HDL cholesterol by 10 mg/dL (0.26 mmol/L) and systolic blood pressure by 8 mm Hg. — 504 adults with a range of health statuses in 17 trials of at least 3 weeks, median 17 fl oz (500 mL) soymilk a day in place of 17 fl oz (500 mL) cow's milk (n = 504) Evidence A source
  Non-HDL-C -10 mg/dL (95% CI -17 to -3.9; metric: -0.26 mmol/L, -0.43 to -0.10); SBP -8.00 mm Hg (-14.89 to -1.11); DBP -4.74 mm Hg (-9.17 to -0.31)
• Pooling 35 randomized trials, soya foods lowered LDL cholesterol by 4.8 mg/dL overall, by 7.5 mg/dL in people with high cholesterol and by 11 mg/dL when whole foods such as soya milk were used. — 35 randomized trials (50 comparisons), 4 weeks to 1 year (n = 35 RCTs) Evidence A source
  LDL -4.83 mg/dL (95% CI -7.34 to -2.31); hypercholesterolemic -7.47 vs healthy -2.96; whole soya products -11.06 (-15.74 to -6.37)
• In 66 people with fatty liver disease on a calorie-cut diet, adding 8.1 fl oz (240 mL) of soy milk a day for 8 weeks lowered insulin resistance and systolic pressure more than diet alone. — 66 patients with non-alcoholic fatty liver disease, 500 calories deficit diet, 240 ml/day soy milk for 8 weeks (n = 66) Evidence B source
  HOMA-IR -0.45 vs -0.14 (P = 0.03); SBP -3.81 vs -1.48 mm Hg (P = 0.01); fasting glucose unchanged

Safety, limits and interactions

• In children, observational data link plant-based milk in place of cow milk with lower BMI, height and vitamin D, while fortified soy milk went with higher bone density in girls short on calcium. — children and adolescents aged 1-18 y in 6 studies (3 cross-sectional, 1 cohort, 2 clinical trials), total n = 15,815 (n = 15815) Evidence C source
  Lower BMI, height and serum vitamin D with plant-based milk vs cow milk; higher bone density with fortified soy milk in low-calcium adolescent girls

What people believe that the data does not support

• In the same 17 trials, added sugars in the soymilk made no meaningful difference to its effects against cow's milk on blood lipids, blood pressure or inflammation markers. — 504 adults in 17 randomized trials, sweetened and unsweetened soymilk compared with cow's milk (n = 504) Evidence A source
  No meaningful effect modification by added sugars across 19 outcomes

Who this works differently for

• The 2025-2030 US Dietary Guidelines say to stop infant formula at 12 months and give whole milk. — Infants moving off formula at 12 months Evidence E source
  Whole milk from 12 months; soy beverage not mentioned

What is still unknown

• A review of 8 trials with 263 people found soy milk alone lowered blood lipids in only some of them, so its cholesterol effect on its own remains unsettled. — 263 subjects in 8 randomized clinical trials of soy milk, to May 2018 (n = 263) Evidence A source
  3 trials no lipid change; LDL fell in 3, TG in 1, TC in 1; HDL rose in 1
• The 2025 US Dietary Guidelines committee asked for trials that study milk alternatives on their own, since most evidence lumps them with dairy milk. — USDA Nutrition Evidence Systematic Review for the 2025 Dietary Guidelines Advisory Committee Evidence E source
  Research gap: RCTs on milk alternatives examined independently

The bottom line

The brand variation is measurable. In 14 USDA samples of sweetened soy milk, calcium ran from 138 to 188 mg per 3.5 oz (100 g), while 11 samples of whole milk stayed between 116 and 130 mg. The NIH puts calcium absorption from dairy and from fortified foods such as soy drinks at about 30 percent, in a fact sheet that does not look at the form of calcium used in a carton. Warming milk for about 10 minutes keeps all its calcium and costs about a fifth of its vitamin B12 and 15 percent of its folate, by a 1975 USDA group table that has no factor for soy milk. Whether soy milk on its own lowers cholesterol is still open. A review of 8 small, short trials with 263 people found lipids fell in some and did not change in three, and there were too few to pool. A larger meta-analysis of 35 trials of soya foods found LDL fell by 4.8 mg/dL (95% CI 2.3 to 7.3) overall and by 7.5 mg/dL in people with high cholesterol. Those trials mixed soy milk with soybeans and soy nuts, and their results varied widely. In 66 people with fatty liver disease on a calorie-cut diet, 8.1 fl oz (240 mL) of soy milk a day for 8 weeks lowered insulin resistance (HOMA-IR down 0.45 against 0.14) and systolic pressure (down 3.8 against 1.5 mm Hg) more than diet with no drink. That trial was small and open-label, and it did not compare soy milk with cow's milk. Children are where to be careful. Across 6 studies of 15,815 young people aged 1 to 18, observational data link plant-based milk in place of cow's milk with lower BMI, height and vitamin D, though "plant-based milk" there lumps very different drinks together. One study found higher bone density in girls short on calcium who drank fortified soy milk. The 2025-2030 US Dietary Guidelines tell parents to move from formula to whole milk at 12 months and no longer mention soy drink there, where the previous edition allowed fortified unsweetened soy beverage. The new guidelines count "fortified dairy alternatives" as options in the dairy group without saying whether a cup of soy milk counts as a full serving. We have not collected evidence on allergy, pregnancy or medicines for either drink, and that gap is not a reassurance. The 2025 guidelines committee also asked for trials that study milk alternatives on their own, since most data lump them with dairy.

Sources

1. msym-r1-08 · USDA FNDDS survey foods, dataset
   Milk, whole (fdc_id 2705385, FNDDS 11111000): Energy 61 kcal, Protein 3.27 g, Calcium 123 mg, Total Sugars 4.81 g, Fatty acids total saturated 1.86 g, Vitamin D 1.1 ug per 100 g, 1 cup 244 g; Soy milk, unsweetened (fdc_id 2705405, FNDDS 11320100): Energy 38 kcal, Protein 3.55 g, Calcium 101 mg, Total Sugars 0.56 g, Fatty acids total saturated 0.314 g, Vitamin D 0.7 ug per 100 g, 1 cup 244 g
   USDA FoodData Central survey (FNDDS), release 2024-10-31, food codes 11111000 and 11320100 · checked 2026-09-29
2. msym-r1-09 · USDA Foundation Foods, dataset
   Soy milk, sweetened, plain, refrigerated: Calcium, Ca 154.8 mg, range 138.0-188.0, 14 samples; Milk, whole, 3.25% milkfat, with added vitamin D: Calcium, Ca 123.0 mg, range 116.0-130.0, 11 samples; Protein: Soy milk, sweetened 2.5625-3.375 g, Soy milk, unsweetened, shelf stable 3.0-4.6875 g, Milk, whole 2.93-3.51 g
   USDA FoodData Central Foundation Foods, release 2026-04-30 · checked 2026-09-29
3. msym-dga-01 · agency guideline
   When consuming dairy, include full-fat dairy with no added sugars. ... Dairy serving goals: 3 servings per day as part of a 2,000-calorie dietary pattern, adjusting as needed based on your individual caloric requirements.
   Dietary Guidelines for Americans, 2025-2030 (HHS, USDA), January 2026 · checked 2026-09-29
4. msym-dga-02 · agency guideline
   Dairy 1 cup milk; ¾ cup yogurt; 1 oz cheese ... Dairy: Whole, reduced-fat, low-fat, or nonfat dairy products, including fluid, dry, or evaporated milk; yogurt; and cheeses. Lactose-free and lactose-reduced products, as well as fortified dairy alternatives, are also options.
   Dietary Guidelines for Americans, 2025-2030: Daily Servings by Calorie Level (HHS, USDA) · checked 2026-09-29
5. msym-dga-03 · agency report
   Dietary Guideline Advisory Committee Recommendations Yes Partial No ... 27. For the Dairy and Fortified Soy Alternatives food group, plain cow milk (whole milk) or fortified unsweetened soy beverage can be offered beginning around 12 months of age. ☐ ☒ ☐ 28. Fat-free and low-fat dairy and fortified soy options are recommended for individuals ages 2 years and older. ☐ ☐ ☒
   The Scientific Foundation for the Dietary Guidelines for Americans, 2025-2030, Overview of Evidence Accepted and Rejected from the DGAC Report, p. v · checked 2026-09-29
6. msym-r1-07 · agency fact sheet
   In the United States, foods that are fortified with calcium include many fruit juices and drinks, milk substitutes (e.g., soy and almond beverages), tofu, and ready-to-eat cereals [1,8]. ... The absorption of calcium from dairy products and fortified foods is about 30% [1].
   NIH Office of Dietary Supplements, Calcium Fact Sheet for Health Professionals · checked 2026-09-29
7. msym-r1-10 · USDA Table of Nutrient Retention Factors, dataset
   USDA-RETN-R6:2151 MILK,HEATED APPROX 10MIN, group Dairy and Egg Products: vitamin_b12_ug 80% retained, vitamin_c_mg 85%, folate_total_ug 85%, thiamin_mg 90%, vitamin_b6_mg 90%, calcium_mg 100%, riboflavin_mg 100%, potassium_mg 100%
   USDA Table of Nutrient Retention Factors, Release 6, 2007, Retn_Code 2151 · checked 2026-09-29
8. msym-r1-01 · meta-analysis of RCTs
   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. 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 2026-09-29
9. msym-r1-03 · meta-analysis of RCTs
   Intake of soya products resulted in a significant reduction in serum LDL-cholesterol concentration, -4.83 (95% CI -7.34, -2.31) mg/dl, TAG, -4.92 (95% CI -7.79, -2.04) mg/dl, and total cholesterol (TC) concentrations, -5.33 (95% CI -8.35, -2.30) mg/dl. ... LDL reductions were more marked in hypercholesterolaemic patients, -7.47 (95% CI -11.79, -3.16) mg/dl, than in healthy subjects, -2.96 (95% CI -5.28, -0.65) mg/dl. LDL reduction was stronger when whole soya products (soya milk, soyabeans and nuts) were used as the test regimen, -11.06 (95% CI -15.74, -6.37) mg/dl, as opposed to when 'processed' soya extracts, -3.17 (95% CI -5.75, -0.58) mg/dl, were used.
   Br J Nutr, 2015 · checked 2026-09-29
10. msym-r1-05 · randomized controlled trial
    After 8-weeks of intervention, soy milk group had a greater significant reduction in serum insulin(-3.44 ± 5.02 vs. -1.09 ± 3.77 μIU/ml, P = 0.04), HOMA-IR (-0.45±0.64 vs -0.14 ± 0.47, P = 0.03), systolic (-3.81±4.15 vs -1.48±2.93 mmHg, P = 0.01) and diastolic (-2.39±2.80 vs. -0.94±2.76 mmHg, P = 0.04) BP, and also, a significant increase in QUICKI (0.02± 0.032 vs. 0.008±0.018, P = 0.04) compared to the control group.
    Complement Ther Med, 2019 · checked 2026-09-29
11. msym-r1-06 · systematic review
    Observational studies found that consumption of plant-based milk was associated with lower childhood body mass index (BMI), height, and serum vitamin D concentrations compared with cow milk. No association was found between soy milk consumption and BMI in adolescent girls. Low-quality clinical trials showed minimal effects on growth, and 1 study found that adolescent girls with low calcium intake who consumed fortified soy milk had higher bone density compared with those who did not consume soy milk.
    J Nutr, 2024 · checked 2026-09-29
12. msym-r1-02 · meta-analysis of RCTs
    There was no meaningful effect modification by added sugars across outcomes. ... We could not conduct the sub-study of the effect of lactose versus added sugars, as no eligible trials could be identified.
    BMC Med, 2024 · checked 2026-09-29
13. msym-dga-04 · agency guideline
    If feeding or supplementing your baby with infant formula, stop feeding your baby infant formula at 12 months of age and give them whole milk.
    Dietary Guidelines for Americans, 2025-2030 (HHS, USDA), January 2026 · checked 2026-09-29
14. msym-r1-04 · systematic review of RCTs
    Eight eligible trials comprising a total of 263 subjects were included in this review. Three studies reported no significant improvement in any of the serum lipid parameters following soy milk consumption. ... Evidence from clinical trials regarding the hypolipidemic effect of soy milk is limited and controversial.
    Complement Ther Med, 2019 · checked 2026-09-29
15. msym-r1-11 · agency systematic review
    When possible, rather than combining all types of dairy milk and milk alternatives into one category, categorizing intake by dairy milk or milk alternatives and then by fat and sweetened/flavored groupings will enhance accuracy and precision. Conduct longitudinal studies, particularly RCTs, examining sweetened/flavored milk independently. Conduct longitudinal studies, particularly RCTS, examining milk alternatives independently.
    USDA NESR 2024, Dairy Milk and Milk Alternatives and Growth, Body Composition, and Risk of Obesity · checked 2026-09-29