Plant Protein: Forms, Muscle Evidence and Safety

Confidence is limited for precise source-versus-source training differences: the direct trials are small or commercially supported, and some product terms and underlying-trial finances remain unresolved. The reviewed evidence does not establish universal equivalence, a guaranteed leucine threshold or a blanket safety ranking.

Introduction

“Plant protein” covers proteins obtained from different crops. A soy isolate, pea ingredient, rice protein and mixed blend can have different compositions. Whole dietary patterns and isolated supplements are also different interventions (ingredient analysis; FAO framework; dietary-pattern review).

This article separates human training outcomes from composition, short-term mechanisms and marketing. Review-level funding and constituent-trial funding are separate assessments: university authorship does not rule out manufacturer grants, product donations or sponsor participation (Roquette-supported trial; TiFN trial; donated-product study).

Table of contents

• Evidence summary
• What plant protein is
• Forms and types
• Protein quality and mechanisms
• Benefits by claim
• Hype and what works
• Risks and adverse effects
• Interactions and who needs advice
• Doses studied
• Animal and laboratory evidence excluded
• Money, countries and backers
• Source support and limits
• Regulatory context
• FAQ
• Sources

Evidence summary

The assessments below apply to specific claims and study settings. A nonsignificant difference is not proof of equivalence; financial eligibility is separate from study design.

What plant protein is

Grams of powder, protein and amino acids are different quantities. In Lynch’s trial, 29 g soy powder supplied 26 g protein, whereas 21 g whey powder supplied 19 g protein. Finished ingredients and flavourings change the label’s composition (Table 1; FDA label guidance).

Check the actual formulation and its role in the whole diet. Vegan or dairy-free wording does not remove other allergens or cross-contact risks (FDA; Health Canada). A powder sold as a dietary supplement is not represented as the sole item of a meal or diet under the US definition (FDA supplement definition).

Forms and types

These are compositional distinctions, not efficacy rankings. Gorissen analysed selected commercial ingredients and disclosed TiFN public–private support, in-kind samples from companies and dairy-company author fees. It is used for composition and financial context, not an independent clinical-outcome verdict (P9).

Protein quality and mechanisms

Protein supplies amino acids for tissue turnover. The FAO framework treats indispensable amino-acid supply, digestibility and the wider diet separately. Acute MPS measures a short-term synthesis response; it is not long-term muscle growth (FAO; tracer experiment).

Three common shortcuts are misleading:

1. Matching leucine alone does not match the protein dose, all amino acids or the wider diet. Lynch used different protein amounts to approximate leucine matching; Pinckaers found different blood amino-acid availability despite no detected MPS difference (P1; P4).
2. “Limiting amino acid” is relative to a reference requirement pattern; it does not necessarily mean absent or unusable protein (FAO scoring method).
3. PDCAAS and DIAAS estimate protein quality, not measured human muscle gain. The product, age-reference pattern and digestibility method matter. FAO prefers human ileal measurements but permits pig and then rat measurements when needed (FAO).

Compare amino-acid results using the same denominator: grams of amino acid per gram of protein differs from grams per gram of ingredient. Gorissen measured more leucine per gram of protein in corn than whey, so “every plant protein is lower” is inaccurate. These ingredient assays are not clinical muscle-growth tests (P9; FAO scoring framework).

Benefits by claim

Muscle strength and lean mass with training

Banaszek’s 15-person pilot used purchased supplements and reported no external funding. Both groups improved squat and deadlift strength, with no detected group difference; neither had a detected muscle-thickness or body-composition change. No unsupplemented control or equivalence margin was used. The author’s gym part-ownership began after data analysis, and differs from identified protein-seller funding (P2).

Lynch’s 12-week study detected no source difference in lean mass or strength during supervised training. Soy supplied 26 g protein versus 19 g whey, approximately matched for leucine. There was no unsupplemented control, 13 of 61 randomized participants did not complete, and habitual diet was assessed by self-reported logs. These findings neither isolate the added powder benefit nor prove equivalence across populations or doses. SCAN and ASU student grants were declared; ultimate grant funding and product purchase/donation terms remain unestablished (P1).

A plant powder can supply protein alongside training, but the reviewed trials do not establish identical muscle gains for every pea, soy, rice or blend product versus whey (soy trial; pea pilot; rice trial; review).

Why some larger trials do not determine this verdict

Babault randomized 161 men and analysed 137 after 12 weeks of training. The overall between-group muscle-thickness comparison was nonsignificant; the pea-versus-placebo signal came from the weaker-at-baseline subgroup, not all participants. Roquette funded the study, and three employee authors were credited with study design. The paper also states that its funders had no role; both statements should be visible, without inventing undisclosed control (P3).

Pinckaers gave 24 young men a single 30 g dose of pea or milk protein and measured five-hour myofibrillar MPS. No source difference was detected; this was not a formal equivalence test. The milk data also appeared in earlier publications from the same larger experiment, so they are not independent replications. Industry sponsors participated in design and product choice. The result is short-term context, not months of equal hypertrophy (P4).

Hevia-Larraín’s original abstract describes 19 habitual vegans and 19 omnivores, 12 weeks of training and total protein adjusted to about 1.6 g/kg/day. It reports no detected group differences. Dietary identity was not randomly assigned. The publisher’s accessible declarations identify FAPESP support, DuPont donations of both proteins and outside industry relationships for one author. Full original methods and statistical detail were unavailable; this is qualified context, not an independent-only equivalence result (P11).

What systematic reviews add

Reid-McCann’s 2025 review included 43 trials, searching through June 2023. Thirty trials contributed to a small pooled animal-protein advantage in muscle mass; 17 soy-versus-milk trials showed no detected mass difference. Overall strength and performance comparisons were nonsignificant, but lower-body strength favoured animal protein in the older-adult subgroup. Heterogeneity and risk-of-bias concerns limit inference. Northern Ireland government PhD funding for the review does not clear all constituent-trial finances (P6).

López-Moreno’s 2025 review pooled dietary-pattern strength comparisons, not isolated powders. It detected no overall source difference, but used small, short studies with some mismatched lifestyle interventions and bias concerns; three included reports had beef-industry support and one had Marlow Foods support. The review reports no funding (P12). Yimam’s 2026 review included 18 long-duration reports, 17 using soy and mostly involving peri- or postmenopausal women. Overall differences were nonsignificant; body-composition certainty was low and most strength/function certainty very low. It disclosed Danone APC support and an internship, so its outcomes are context under this article’s independent-only standard (P7).

Weight loss, fullness, glucose and healthy ageing

These comparisons do not establish a plant powder as a weight-loss or diabetes treatment. Yimam assessed heterogeneous body-composition and metabolic markers, with no overall source advantage and substantial uncertainty. Replacing another food differs from adding calories, and whole-diet findings do not establish one powder’s effect (long-term review; dietary-pattern review).

The small direct training trials mostly involved young or trained volunteers; they are not treatment protocols for frailty or sarcopenia. Reid-McCann found no trials measuring sarcopenia status as the outcome, and few clinically impaired populations. This differs from measuring lean mass alone (P1; P2; P6).

Soy, testosterone and thyroid claims

Reed’s 2021 analysis of 41 clinical reports detected no pooled significant change in the measured circulating male reproductive hormones. It mixed controlled and single-arm designs, foods/proteins and isoflavone extracts, and used some imputed data. The original reports no specific grant, while its acknowledgements disclose Messina’s soy-company consulting and Kurzer’s Soy Nutrition Institute advisory role. Included-trial finances were not fully cleared. This is qualified hormone context, not proof that every dose, extract or clinical circumstance is safe (original paper; PubMed record of the same review).

Soy foods, protein ingredients and concentrated isoflavone extracts are different exposures. NCCIH describes product- and population-specific benefits and safety limits; it does not guarantee any individual supplement. The thyroid-medicine question concerns absorption and monitoring, addressed below (NCCIH; levothyroxine label).

Hype and what works

Risks and adverse effects

Allergy

Soy is a US major food allergen. Pea can also cause reactions in some peanut-allergic people. Health Canada says severe reactions are expected to be uncommon, but recommends reading labels and asking an allergist; it does not predict an individual’s risk. Concentrated pea protein can create higher exposure than whole peas. Dairy-free and plant-based wording do not mean hypoallergenic (FDA; Health Canada).

Avoid known protein-allergy triggers and check the full ingredients and cross-contact information. A soy allergy does not automatically make pea a suitable substitute. Breathing difficulty, throat swelling or another severe systemic reaction needs emergency medical care (FDA; Health Canada).

Gastrointestinal symptoms and formulation

NCCIH lists bloating, nausea, diarrhoea and constipation among possible soy side effects. That does not give an adverse-event rate for every plant protein or show which ingredient caused symptoms in a mixed product. Check the full formula and seek advice for persistent symptoms (NCCIH; FDA product advice).

Contaminants

Clean Label Project’s 2025 whitepaper describes 160 products from 70 brands, retail sample selection and laboratory methods, and reports variation among sampled categories. Its web summary also mentions 165 powders, so the whitepaper’s study count is the clearer basis. These results do not establish that all plant proteins are hazardous or that an untested batch is safe. The FAQ discloses brand certification fees, grants and donations; the 2023 filing separately reports $440,031 certification revenue and $228,304 consultation revenue. That filing describes testing by an affiliated company without naming it in the programme statement. The inspected report and financial record do not identify the 2025 study’s donors or give individual raw batch results (study report; web summary; funding FAQ; 2023 filing).

California Proposition 65 safe-harbour levels are exposure-based warning thresholds. Exceeding one is not interchangeable with proving that a product violates every food-safety standard or that a user has been harmed. Interpretation needs the contaminant, amount per actual serving, frequency, background exposure and jurisdiction. OEHHA

As a practical check, ask which exact product and sample a certificate covers, its date, tested substances and numerical results. CLP says certification is product-specific, renewed annually and does not test every batch. A logo is not a clinical-efficacy test or assurance about untested material (testing FAQ).

Antinutrients

FAO describes how processing and antinutritional factors can change amino-acid availability. This does not establish that all isolates are antinutrient-free or that every laboratory effect causes clinical harm at a usual serving. Product composition and the relevant human exposure remain separate questions (FAO report).

Interactions and who needs advice

• Levothyroxine: the cited label names soybean flour and dietary fibre as possible absorption interferents. It separately gives four-hour spacing for interacting medicines; that is not a validated rule for every soy isolate. Follow your own medicine instructions and discuss major product or diet changes with the prescriber/pharmacist; do not adjust the dose yourself (label sections 2.1 and 7.9; NHS).
• MAO inhibitors: NCCIH reports interactions with certain tyramine-rich soy foods. This does not quantify the risk for every soy, pea or rice protein preparation. Check the exact food/product and medicine with a pharmacist (NCCIH).
• Warfarin and other medicines: review the complete product, including added vitamins, herbs and other ingredients, with a pharmacist. The words “plant protein” alone do not establish a clinically important interaction or its absence (FDA medicine advice).
• Kidney disease or a prescribed protein plan: protein and some mineral needs depend on the disease stage, treatment and nutrition status. Healthy-adult training studies are not a dosing guide; dialysis needs can differ (NIDDK CKD guidance).
• Pregnancy, breastfeeding, children, frailty or malnutrition: these adult trial results do not establish suitability or a dose for every clinical group. NCCIH flags possible risk or uncertainty for soy above customary food amounts in pregnancy/breastfeeding. Review the exact supplement with the appropriate clinician (NCCIH; FDA advice).

This is a material-interaction summary, not a claim to list every possible interaction.

Doses studied

These are study exposures, not personal recommendations. They should not be added on top of one another.

These protocols do not establish a universal meal ceiling, powder dose or leucine target. Protein from the whole diet also counts; approximate matching of one amino acid does not match the entire intervention (P1; P4; FAO).

Animal and laboratory evidence excluded

Animal muscle studies, cellular mTOR experiments and in-vitro digestion were not used to establish human muscle growth, hormone effects or clinical harms. Where protein-quality scores use animal digestibility measurements, that limitation belongs beside the score. FAO’s DIAAS framework can use non-human measurements; the score is not necessarily “invalid,” but it is not a human hypertrophy trial. P10

Money, countries and backers

Follow the money

Research funding at a glance

Funding & backers→Source & study→Claim & limits

24 disclosure entries. The counts below summarize independence tiers explicitly assigned in this article. They count disclosures, not studies, funding amounts or evidence quality.

No single explicit tier could be mapped. This graphic does not assign a new tier or infer independence.

Consult this article’s source and funding notes for named funders, countries, relationships and exceptions where available. Institutional backing, researcher interests and trial sponsorship are separate questions. Public funding alone does not establish independence; commercial ties alone do not prove a claim false. This overview is not a new financial audit.

Plant-ingredient manufacturers, powder brands, retailers and paid promoters benefit from stronger demand. Dairy sellers benefit from claims favouring whey. Researchers have publication and career incentives; certification organizations may benefit when consumers worry about contamination. These are incentive analyses, not allegations that results were falsified.

Documented examples:

• Roquette funded Babault’s trial and employed three authors. Roquette describes itself as privately family-owned, with global headquarters in France; its manufacturing footprint spans countries. Family share percentages were not independently reconstructed here. Its own corporate disclosure is used only for ownership, not efficacy P3, P19
• Tereos (France), Cargill and Kellogg (United States, locations as identified in the study) contributed to the TiFN pea project in the Netherlands. Sponsor involvement in study design is explicitly documented P4
• DuPont donated the supplements in the Brazilian habitual-vegan/omnivore trial. Historical sponsor names are retained; they are not claims about today’s ownership of every corresponding ingredient business P11
• The Northern Ireland Department for the Economy funded the Belfast review. That funding does not erase commercial support inside the trials it pooled P6

Arrows show documented relationships only. The diagram is not exhaustive corporate ownership tracing. Crop origin, ingredient processing location, trial location, brand headquarters and regulatory jurisdiction are separate facts; this article does not infer product quality from nationality.

Source support and limits

The table reports declared support, financial relationships, document scope and remaining limits. Direct commercial support excludes outcomes from this article’s independent-only verdict. Missing product terms and untraced constituent finances remain unknown. Study design and financial assessment are separate; a letter grade cannot certify a complete funding chain.

FDA receives appropriations and legally restricted industry user fees; these agency-level ties are disclosed without alleging that protein sellers paid for the cited guidance. P24

This is a focused assessment of the cited sources, not a registered systematic review of every plant-protein study. Full original methods were unavailable for the Hevia-Larraín comparison. Small studies, population differences, some review reporting inconsistencies, unresolved product terms and incomplete constituent-trial financial chains limit inference. Guidance and source location do not themselves establish independence.

Regulatory context

In the United States, FDA does not approve dietary supplements before marketing; firms have safety, labelling and manufacturing responsibilities. Some new dietary ingredients require a premarket safety notification, which is different from supplement approval. A powder may instead be a conventional food depending on its presentation. Regulatory status and allergen compliance do not establish muscle-building efficacy (FDA framework).

In India, FSSAI’s official portal points to the applicable health-supplement and nutraceutical framework. The exact ingredient, product category and current authorization must be checked for the actual product. An FSSAI licence or overseas research trial is not a clinical endorsement, and this article does not certify a particular formulation. P21

FAQ

Is plant protein as good as whey?

The reviewed studies support considering the actual product, protein amount, whole diet, person and outcome. Small trials with no detected source difference do not establish every plant powder as equivalent or inferior to whey (soy trial; pea pilot; review).

Is pea–rice automatically better than pea alone?

A blend changes composition; the useful question is whether its digestible amino-acid profile and protein amount meet the relevant dietary need. Its name alone does not establish an ideal ratio, specific score or smaller clinically effective serving (FAO mixture framework).

Does soy lower testosterone?

Reed’s pooled clinical analysis detected no significant change in the measured circulating male hormones. It has disclosed soy-industry author relationships and incompletely cleared constituent finances. That is qualified context, not universal safety proof for every extract or dose, and the reviewed evidence does not justify a routine testosterone-lowering claim (full original review).

Is “organic” safer?

It does not certify low heavy-metal content. Assess the actual batch and test scope. Equally, a category average from one survey cannot prove that every organic or plant-based powder is unsafe. P8, P17

Do I need a powder on a vegan diet?

A powder is one way to supply protein, not a requirement established by these comparisons. Adequacy depends on the entire diet’s protein, amino acids and other nutritional needs (FAO dietary framework).

Does no significant difference prove equality?

No. These small trials were not designed to establish universal equivalence. A convincing equivalence claim needs a suitable design, justified margin and sufficiently precise estimate; a P value above 0.05 is not enough (Lynch; Banaszek; Pinckaers).

Sources

Source index

• P1 — Lynch 2020; methods, funding and conflicts inspected
• P2 — Banaszek 2019; purchased supplements, no external funding
• P3 — Babault 2015; Roquette funding and employees
• P4 — Pinckaers 2024; Funding and Declarations
• P5 — Joy 2013; 48 g protocol, affiliations
• P6 — Reid-McCann 2025
• P7 — Yimam 2026; Danone disclosures
• P8 — Clean Label Project survey and certification promotion
• P9 — Gorissen 2018 commercial ingredient analysis
• P10 — FAO dietary protein quality report
• P11 — Hevia-Larraín 2021; DuPont donations
• P12 — López-Moreno 2025; trial funding
• P13 — Reed 2021 male hormone meta-analysis
• P14 — NCCIH soy fact sheet
• P15 — FDA food allergies
• P16 — Health Canada pea allergy advice
• P17 — OEHHA warning thresholds
• P18 — NHS levothyroxine guidance
• P19 — Roquette ownership and location self-disclosure
• P20 — FDA supplement regulation
• P21 — FSSAI official regulatory portal
• P22 — Messina 2022 personal industry-funding disclosure
• P23 — Levothyroxine label revised May 2026, section 7.9
• P24 — FDA funding and restricted user fees

• P25 — Clean Label Project 2025 study whitepaper
• Clean Label Project funding and testing FAQ
• P26 — Clean Label Project Foundation 2023 Form 990
• P27 — Reed 2021 full original journal paper, University of Essex repository

One daily research roundup

Get the topics, key findings and links from our new articles in one email. At most one digest a day, only when there is something new.

Email address

Leave empty

I want Pure City Research daily digests and have read the privacy policy.Get the daily digest

Confirm through your inbox. Unsubscribe anytime. Research only; Shop updates have a separate signup. Please do not send health information.