Milk or rice milk: the protein gap and the arsenic question
When 14 randomized trials in 543 people were pooled in a network meta-analysis, cow's milk and plant drinks, rice drink among them, showed few differences on cholesterol and other cardiometabolic markers. Oat drink lowered total cholesterol by 8.9 mg/dL (0.23 mmol/L) against rice drink (95% CI 0.05 to 0.40), and the authors call that evidence very uncertain. The clearer difference is in the glass. Per 100 g, whole milk carries 3.15 g of protein and unsweetened rice milk 0.28 g, while rice milk has nearly twice the carbohydrate, 9.17 g against 4.78 g. A 244 g cup, recalculated from those 100 g records, works out to about 7.7 g of protein from milk and 0.7 g from rice milk. That gap matters most for small children, and so does arsenic. UK government guidance says children under five should not have rice milk in place of breast milk, infant formula or cow's milk, because of inorganic arsenic.
Every statement here is tied to a source. The badge says what kind of evidence stands behind it: A is a meta-analysis or systematic review, E is the position of an expert body without its own analysis. How we verify →
We have no USDA entry for milk | rice milk itself in this cohort. The evidence below stands on published studies rather than on a composition table, and we would rather say that than show you a number for something else.
How much is actually in it
- Per 100 g, whole cow's milk carries 3.15 g of protein against 0.28 g in unsweetened rice milk, while rice milk carries nearly twice the carbohydrate, 9.17 g against 4.78 g. — Two USDA SR Legacy records: whole milk 3.25% without added vitamins A and D (FDC 172217) and unsweetened rice milk (FDC 171942), published 2019-04-01 (n = 2 records) Evidence E sourcePer 100 g, milk against rice milk: energy 61 against 47 calories, protein 3.15 against 0.28 g, fat 3.27 against 0.97 g, carbohydrate 4.78 against 9.17 g, sugars 5.05 against 5.28 g, calcium 113 against 118 mg, vitamin B12 0.45 against 0.63 ug, vitamin D 0.1 against 1.0 ug, potassium 132 against 27 mg
What a real portion looks like
- The US dietary survey counts a cup of either drink as 244 g, so a cup of whole milk gives about 7.7 g of protein and a cup of rice milk about 0.7 g. — USDA FNDDS 2021-2023 food codes 11111000 (Milk, whole, FDC 2705385) and 11360000 (Rice milk, FDC 2705411); protein from the SR Legacy records above (n = 2 foods) Evidence E source1 cup 244 g for both; protein per cup 3.15 x 2.44 = 7.7 g milk against 0.28 x 2.44 = 0.7 g rice milk
What your body absorbs
- In a randomized crossover study in 12 osteopenic postmenopausal women, calcium from fortified soymilk was absorbed at the same hourly rate as calcium from cow's milk, 0.65 against 0.66. — Osteopenic postmenopausal women aged about 57 years, single dose of radiolabelled fortified soymilk or cow's milk (n = 12) Evidence B sourceHourly fractional calcium absorption 0.65 fortified soymilk against 0.66 cow's milk, p > 0.05
What cooking and storage change
- By the USDA retention table for dairy, heating milk for about 10 minutes keeps all its calcium and riboflavin but loses about 20% of vitamin B12 and 15% of folate. — USDA Release 6 retention factors for the Dairy and Egg Products group, method MILK, HEATED APPROX 10MIN (code 2151) Evidence E sourceRetained: vitamin B12 80%, vitamin C 85%, folate 85%, thiamin 90%, vitamin B6 90%, calcium 100%, riboflavin 100%, potassium 100%
What it does to your health
- A network meta-analysis of 14 randomized trials found few cardiometabolic differences between plant drinks and cow's milk, with oat drink lowering total cholesterol by 8.9 mg/dL (0.23 mmol/L) against rice drink at very low certainty. — 543 participants in 14 RCTs of at least 3 weeks comparing soy, rice and oat drinks with each other or with cow's milk (n = 543) Evidence A sourceOat drink against rice drink total cholesterol -8.9 mg/dL (95% CI -15 to -1.9; metric: -0.23 mmol/L, -0.40 to -0.05), very uncertain; soy drink against cow's milk LDL -18 mg/dL (95% CI -33 to -3.9; metric: -0.47 mmol/L, -0.85 to -0.10), low certainty
Safety, limits and interactions
- The North American pediatric gastroenterology society states that an 8 oz serving of rice milk gives only about 8% of the protein in the same serving of cow's milk. — Infants and young children fed plant-based milks instead of cow's milk Evidence E sourceProtein per 8 oz: rice milk about 8% and almond milk about 2% of cow's milk; documented harms of misuse include kwashiorkor, failure to gain weight and severe nutrient deficiencies
- UK government guidance says children under five should not be given rice milk as a substitute for breast milk, infant formula or cow's milk, because of inorganic arsenic. — Young children aged under five years in England, Northern Ireland and Wales Evidence E sourceRice milk not to replace breast milk, infant formula or cow's milk under age 5; children over 5 and adults may drink it
- In a UK supermarket survey of 19 rice milks, every sample would fail the EU drinking water limit for arsenic, and 80% of those speciated exceeded 10 ug/L of inorganic arsenic. — 19 retail rice milk products from UK supermarkets, 15 with arsenic speciation (n = 19 products) Evidence C sourceTotal arsenic up to 3 times the EU 10 ug/L drinking water limit; inorganic arsenic above 10 ug/L in 80% of 15 speciated samples
Who this works differently for
- A systematic review of 6 studies in 15,815 children found plant-based milk drinkers had lower BMI, height and vitamin D levels than cow's milk drinkers. — Children aged 1 to 18 years in 3 cross-sectional, 1 cohort and 2 clinical studies (n = 15815) Evidence C sourceLower BMI, height and serum vitamin D with plant-based milk against cow milk in observational studies; minimal effects on growth in low-quality trials
What is still unknown
- Observational data link each daily cup of plant-based drink in five-year-olds with about 0.08 in (0.2 cm) less height and 0.22 lb (0.1 kg) less weight, and no randomized trial has compared these drinks with cow's milk in children. — Infants, children and adolescents in cross-sectional and cohort studies of plant-based milk alternatives Evidence C sourcePer cup of plant drink at age 5: height-for-age z -0.04 (95% CI -0.07 to -0.01), BMI-for-age z -0.06 (95% CI -0.09 to -0.03)
The bottom line
The USDA record gives rice milk 118 mg of calcium per 3.5 oz (100 g) against 113 mg in whole milk, a figure that comes from the brand's fortification, and an unfortified rice drink has almost none. The same holds for its vitamin B12 and vitamin D. We found no study of how well calcium from fortified rice milk is absorbed. The closest is a randomized crossover study in 12 osteopenic postmenopausal women, where calcium from fortified soymilk was absorbed at the same hourly rate as from cow's milk, 0.65 against 0.66, with one phosphate fortificant. Heating milk for about 10 minutes keeps all its calcium and riboflavin and loses about 20% of vitamin B12 and 15% of folate, by a 1975 USDA group factor for milk. No such factor exists for rice drink. Children are where the choice counts. The North American pediatric gastroenterology society states that 8 oz of rice milk gives about 8% of the protein in the same serving of cow's milk. It lists failure to gain weight, kwashiorkor, rickets and iron deficiency anemia among the documented harms of misusing plant drinks, described mostly in case reports. In a 2008 survey of 19 rice milks from UK supermarkets, every one would have failed the EU drinking water limit for arsenic, by up to three times. Of the 15 tested by arsenic type, 80% exceeded 10 micrograms per liter of inorganic arsenic, in a survey from one country that could only compare against water. The UK guidance says children over five and adults can drink rice milk as part of a balanced diet. A systematic review of 6 studies in 15,815 children found that plant drink users had lower BMI, height and vitamin D levels than cow's milk drinkers, mostly in cross-sectional data that did not separate rice drink. Another review puts each daily cup of plant drink at age five at about 0.08 in (0.2 cm) less height and 0.22 lb (0.1 kg) less weight (height-for-age z-score -0.04, 95% CI -0.07 to -0.01), an association open to residual confounding. No randomized trial has compared these drinks with cow's milk in children.
Each food on its own
- Milk, bones and the iodine nobody counts — everything on this food
- Rice milk: a sweet drink with almost no protein — everything on this food
Questions answered from the trials
Sources
- mkrm-daily-r3-01 · food composition dataset
Milk, whole, 3.25% milkfat, without added vitamin A and vitamin D (FDC 172217, SR Legacy) per 100 g: energy 61 kcal; protein 3.15 g; fat 3.27 g; carbohydrate 4.78 g; sugars 5.05 g; calcium 113.0 mg; vitamin B12 0.45 ug; vitamin D 0.1 ug; potassium 132.0 mg ... Beverages, rice milk, unsweetened (FDC 171942) per 100 g: energy 47 kcal; protein 0.28 g; fat 0.97 g; carbohydrate 9.17 g; sugars 5.28 g; calcium 118.0 mg; vitamin B12 0.63 ug; vitamin D 1.0 ug; potassium 27.0 mg
USDA FoodData Central, SR Legacy, records 172217 and 171942 · checked 2026-10-02 - mkrm-daily-r3-02 · food portion dataset
Milk, whole (FNDDS 11111000, FDC 2705385): 1 cup 244 g ... Rice milk (FNDDS 11360000, FDC 2705411, Plant-based milk): 1 cup 244 g
USDA FNDDS 2021-2023, food codes 11111000 and 11360000 · checked 2026-10-02 - mkrm-daily-r3-05 · randomized crossover trial
The mean hourly fractional calcium absorption from fortified soymilk was found to be comparable to that of cows' milk: alpha = 0.65+/-0.19 and alpha =0.66+/-0.22, p>0.05, respectively.
Asia Pac J Clin Nutr, 2010 · checked 2026-10-02 - mkrm-daily-r3-03 · retention dataset
MILK,HEATED APPROX 10MIN (USDA-RETN-R6:2151, Dairy and Egg Products): vitamin_b12_ug 80%; 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), code 2151 · checked 2026-10-02 - mkrm-daily-r3-04 · network meta-analysis
Overall, few differences were observed comparing plant-based drinks with another or with cow milk on cardiometabolic outcomes. ... Oat drink may slightly reduce total cholesterol compared with cow milk and rice drink [MD: -0.12 mmol/L (-0.24; -0.01); MD: -0.23 mmol/L (-0.40; -0.05)], but the evidence is very uncertain.
Adv Nutr, 2026 · checked 2026-10-02 - mkrm-daily-r3-06 · position statement
For almond or rice milk, an 8 oz serving provides only about 2% or 8%, respectively, of the protein equivalent found in a serving of CM. Adverse effects from the misuse of certain plant-based beverages have been well-documented and include failure to gain weight, decreased stature, kwashiorkor, electrolyte disorders, kidney stones, and severe nutrient deficiencies including iron deficiency anemia, rickets, and scurvy.
J Pediatr Gastroenterol Nutr, 2020 · checked 2026-10-02 - mkrm-daily-r3-07 · agency guidance
To help reduce their exposure, young children aged under five years should not be given rice milk as a substitute for ... Children over five years old and adults can drink rice milk as part of their balanced diet.
GOV.UK, Arsenic in rice (Food Standards Agency guidance, 2018) · checked 2026-10-02 - mkrm-daily-r3-08 · product survey
All rice milk samples analysed in a supermarket survey (n = 19) would fail the EU limit with up to 3 times this concentration recorded, while out of the subset that had arsenic species determined (n = 15), 80% had inorganic arsenic levels above 10 microg l(-1), with the remaining 3 samples approaching this value.
J Environ Monit, 2008 · checked 2026-10-02 - mkrm-daily-r3-10 · 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.
J Nutr, 2024 · checked 2026-10-02 - mkrm-daily-r3-09 · review
The height-for-age z-score was -0.04 [-0.07; -0.01] and the BMI-for-age z-score was -0.06 [-0.09; -0.03] for each cup of PBMA in five-year-olds, corresponding to reductions of 0.2 cm in height and 0.1 kg in weight. ... Our search did not retrieve any longitudinal studies on long-term outcomes or any randomized intervention trials comparing the consumption of cow's milk and PBMA.
Dtsch Arztebl Int, 2026 · checked 2026-10-02
Review. Reviewed on 2026-10-02 for evidence level, dose and population context, and claims a reader could misread. Bioma Learn has no human medical reviewer at this time, and we say so rather than invent one. Methodology. This is information, not medical advice.