The headline number: protein by Kjeldahl
Almost every report you will see, from a brand's own certificate to the Citizens Protein Project papers, measures protein by the Kjeldahl method. The lab digests the sample, measures total nitrogen, and multiplies by a factor (6.25 for most foods, 6.38 for dairy) to estimate protein. Both Citizens Protein Project studies used it, following Bureau of Indian Standards and AOAC methods.
Kjeldahl has a known blind spot. It counts nitrogen, not protein. Anything with nitrogen in it, including free amino acids, creatine, taurine and even melamine, pushes the number up.
Amino spiking, and how a report catches it
"Amino spiking" is the practice of adding cheap nitrogen-rich ingredients such as glycine, taurine or creatine so the Kjeldahl result matches the label without paying for whey. The first Citizens Protein Project (36 products, published April 2024) flagged brands that measured above their label claim as possible spiking cases. The second study (34 whey products, published November 2025) went further and looked for taurine-based nitrogen spiking directly. It found signs of it in 89 per cent of the "medical" powders sold through pharmacies, against 12.5 per cent of the gym-market nutraceutical powders.
The way to check is the amino acid profile. A whey isolate has a predictable pattern: high leucine, high total essential amino acids, low glycine. The second study used benchmarks of at least 7 g leucine and at least 30 g essential amino acids per 100 g for a high-quality whey, and treated glycine plus alanine above 6 g as a sign of collagen or non-whey protein. If a report shows label-matching protein but a thin amino acid profile, the protein number is not telling you what you think.
Label claim versus measured
The plainest test is whether the tin delivers what it says. The first Citizens Protein Project found 25 of 36 products (69.4 per cent) delivered less protein than advertised, with deficits from under 10 per cent to over 50 per cent; two products from one manufacturer were 62 and 50.4 per cent short. The second study found the nutraceutical group averaged 75.6 g protein per 100 g, with 15 of 17 meeting a 60 g per 100 g "high protein" benchmark, while the medical group averaged 29.1 g per 100 g.
Testers express this differently. Unbox Health gives a Label Accuracy Score from 0 to 10. Trustified marks a product Pass or Fail at a stage it calls Macro Accuracy, and does not run the remaining tests after a fail there. Read the pass rule before the verdict.
Heavy metals and the FSSAI limits
Reports list lead, arsenic, cadmium and mercury, and sometimes copper. Protein powders have no line of their own in the Food Safety and Standards (Contaminants, Toxins and Residues) Regulations, 2011, so testers usually use the "foods not specified" limits: lead 2.5 mg/kg, arsenic 1.1 mg/kg, cadmium 1.5 mg/kg and mercury 1.0 mg/kg (being verified)
"Detected" is not "over the limit". The first Citizens Protein Project detected lead in 75 per cent of samples, cadmium in 27.8 per cent, arsenic in 13.9 per cent and copper in 94.4 per cent, mostly at trace levels, with no mercury or thallium. The second found no mercury and no pesticides, with the heavy-metal outliers mainly in the medical powders. A good report shows measured value, limit and unit side by side.
Aflatoxin
Aflatoxins are mould toxins. The FSSAI table sets 15 micrograms per kg total aflatoxins for cereals and separate figures for nuts, spices and other foods; testers usually compare a milk-based powder against 15 micrograms per kg, though the right line for whey is a matter of judgement (being verified) The first study found aflatoxin in 5 of 36 products, above 10 micrograms per kg in two. The second found it in two medical powders. Unbox Health reports "Total Aflatoxins B,G" and "Total Aflatoxins M1" separately, which is the right split, since M1 is the form that comes through milk.
Sugars, hormones and extras
The second Citizens Protein Project reported that 44 per cent of products contained at least 2 g sucrose or fructose per 100 g, often not on the label, and detected progesterone in two medical powders. The first study found undeclared herbal extracts, green tea, curcumin, Garcinia cambogia and ashwagandha, in some blends, and flagged them for liver risk. These lines rarely appear on a brand's own certificate.
What each tester does
Citizens Protein Project 1 and 2: academic studies led from Rajagiri Hospital, Kerala, published in the journal Medicine in April 2024 and November 2025. Products were bought blind, coded and sent to an accredited lab for Kjeldahl protein, amino acids, aflatoxin, pesticide residues, heavy metals by ICP-MS, steroids and full organic profiling. Results are per product, with raw numbers and no pass or fail stamp.
Trustified: a private Indian certification company. Its public list shows Pass, Fail or Expired for each product, the batch, the lab (Eurofins for most entries, TÜV NORD for at least one) and the test date. It says it tests label accuracy, heavy metals, chemicals and amino spiking in tiers it calls 6-level and 7-level testing. Brands can pay for certification, which is why "Expired" appears.
Unbox Health: buys from the open market with no brand involvement, tests at up to three NABL-accredited labs, and publishes a grade from D to A+ built from three scores out of 10: Label Accuracy, Non-Toxicity and Nutritional Profile. Each page shows batch number, purchase source, rating live date and expiry, and lists nine essential amino acids.
Why the same product can pass one test and fail another
One Science Nutrition Iso Gold is the example this site gets asked about most. Trustified lists it as FAIL, tested by TÜV NORD on 11 December 2024, batch 1229070 (manufactured October 2023), failing at the Macro Accuracy stage with no further tests run. The product is not on Unbox Health's whey list, and neither Citizens Protein Project paper names it. The brand sells the same product with its own certificates for other batches (being verified) All of these can be true at once, for five reasons.
Different batches. A test is of one batch. Batch 1229070 failing says nothing about a batch made in 2026.
Different samples. Most testers buy one pack. A single pack can be stored badly, counterfeit or simply unrepresentative.
Different methods. Kjeldahl protein and a summed amino acid profile can disagree by several grams per 100 g on the same tin.
Different pass rules. One tester may fail a product 8 per cent under label; another may score it 8 out of 10 and give it a B.
Different scope. A product can pass on protein and fail on lead, or pass on lead and fail on undeclared sugar. Ask which parameter failed.
A checklist for any report
Batch number and date. Lab name and accreditation. Method for protein. Amino acid profile, not just total protein. Measured value, limit and unit for each contaminant. The pass rule. Whether the sample was bought or supplied by the brand. A report missing three of these is marketing.
Sources
- Citizens protein project: a self-funded, transparent, and concerning report on analysis of popular protein supplements sold in the Indian market (Medicine, April 2024, full text via Europe PMC)
- The Citizens Protein Project 2 (Medicine, November 2025, full text via Europe PMC)
- Citizens Protein Project 2 results site
- Unsafe gains: 14% of Indian protein powders have toxins, 70% mislabelled (Business Standard, 17 Jul 2025)
- Trustified: protein powders results list
- Trustified: One Science Nutrition results page
- Unbox Health: whey protein powders ratings
- Unbox Health: Nakpro Impact Whey lab test review
- Amino spiking in India (Limitless Labs)
- Food Safety and Standards (Contaminants, Toxins and Residues) Regulations, 2011, compendium version IX, 3 Feb 2026 (PDF)