NutriStack
What 92,018 supplement labels actually say.
We analyzed every currently marketed label in the NIH Dietary Supplement Label Database. Only 1 in 59 carries a seal from an independent testing program (USP Verified: 1 in 303), while 1 in 5 uses self-declared quality wording instead. 12.9% use a proprietary blend, which need not state how much of each ingredient it contains, and 40.4% of single-dose vitamin D labels state more than the adult upper limit per serving.
Finding 1
Independent testing seals are rare. Self-declared quality claims are not.
1,572 of 92,018 labels (1.71%) show a seal from a named independent testing program. 18,220 (19.8%) carry wording the brand declares about itself, such as GMP, “quality assured” or “lab tested”.
Independent program seals
| Program | Labels | Share |
|---|---|---|
| Informed Sport / Informed Choice | 670 | 0.73% |
| USP Verified | 304 | 0.33% |
| NSF Certified for Sport | 217 | 0.24% |
| NSF Contents Certified | 182 | 0.2% |
| IFOS (fish oil) | 67 | 0.07% |
| BSCG Certified Drug Free | 62 | 0.07% |
| Clean Label Project | 53 | 0.06% |
| Labdoor | 43 | 0.05% |
| ConsumerLab approved | 1 | 0.0% |
Self-declared wording
| Wording | Labels | Share |
|---|---|---|
| GMP / cGMP wording | 11,766 | 12.79% |
| 'Quality assured / guaranteed' wording | 5,334 | 5.8% |
| 'Lab tested / third-party tested' wording | 2,915 | 3.17% |
| FDA facility wording (registered/inspected) | 1,788 | 1.94% |
Current Good Manufacturing Practice compliance (21 CFR Part 111) is a legal requirement for all dietary supplement makers, and FDA does not approve dietary supplements or certify facilities; facility registration does not denote FDA approval. 17 labels refer to FDA approval. Our third-party testing guide explains what each seal does and does not check.
Finding 2
One label in eight uses a proprietary blend.
11,892 of 92,018 labels (12.9%) list at least one proprietary blend. A blend must state its total weight but not the amount of each ingredient inside it, so the buyer cannot tell how much of any one ingredient they get.
Share with a proprietary blend, by what the product name says it is
| Product type (from name) | Labels | With a blend | Share |
|---|---|---|---|
| Probiotic | 1,455 | 416 | 28.6% |
| Sleep | 545 | 138 | 25.3% |
| Weight loss / fat burner | 503 | 126 | 25.0% |
| Brain / focus / nootropic | 588 | 142 | 24.1% |
| Testosterone / male vitality | 103 | 21 | 20.4% |
| Pre-workout | 354 | 27 | 7.6% |
| Multivitamin (name) | 2,108 | 104 | 4.9% |
| Protein | 1,896 | 70 | 3.7% |
Under 21 CFR 101.36(c), a proprietary blend must declare its total weight and list its ingredients in descending order of predominance, but need not state the amount of each ingredient. The share has not fallen over time: by DSLD entry year it ranges from 8.3% to 15.2%. Counts rows whose name contains 'proprietary'. The local extract keeps the first 25 Supplement Facts rows per label, so 7,398 long labels may have unseen rows; the count is a lower bound. The extract does not carry per-ingredient amounts, so this study cannot measure how many blends omit individual amounts.
Finding 3
5,000 IU is now the most common vitamin D strength.
Of 1,048 labels whose product name states a vitamin D amount, 5,000 IU is the most common. Among the 846 with a one-unit serving, 342 (40.4%) state more per serving than the adult Tolerable Upper Intake Level of 100 mcg (4,000 IU).
Vitamin D strength stated in the product name
| Strength | Labels | Share |
|---|---|---|
| under 1,000 IU | 169 | 16.1% |
| 1,000 IU | 234 | 22.3% |
| 1,001 to 1,999 IU | 6 | 0.6% |
| 2,000 IU | 211 | 20.1% |
| 2,001 to 3,999 IU | 26 | 2.5% |
| 4,000 IU (equal to the UL) | 13 | 1.2% |
| 4,001 to 4,999 IU | 1 | 0.1% |
| 5,000 IU | 293 | 28.0% |
| 5,001 to 9,999 IU | 6 | 0.6% |
| 10,000 IU | 70 | 6.7% |
| 10,001 to 49,999 IU | 4 | 0.4% |
| 50,000 IU | 15 | 1.4% |
| over 50,000 IU | 0 | 0.0% |
Amounts come from the product name, not the Supplement Facts panel (the local extract has no per-ingredient amounts). mcg values are converted at 1 mcg = 40 IU. The UL comparison is limited to labels whose serving size is exactly one tablet, capsule, softgel, gummy, drop or similar unit, so the name amount equals the per-serving amount. The UL is a reference value for daily intake from all sources for the general adult population; it does not apply to people taking a product under medical supervision. UL source: NIH ODS.
Finding 4
Some single-nutrient products routinely exceed the upper limit.
Across seven nutrients, 219 of 971 single-nutrient labels (22.6%) state a per-serving amount above the adult upper limit. Niacin products almost all do; vitamin C products never do.
Labels stating more than the adult UL per serving
| Nutrient | Adult UL | Qualifying labels | Above UL | Share | Source |
|---|---|---|---|---|---|
| Niacin (nicotinic acid) | 35 mg | 104 | 101 | 97.1% | NIH ODS |
| Zinc | 40 mg | 111 | 68 | 61.3% | NIH ODS |
| Iodine | 1,100 mcg | 17 | 9 | 52.9% | NIH ODS |
| Iron | 45 mg | 90 | 24 | 26.7% | NIH ODS |
| Vitamin B6 | 100 mg | 109 | 16 | 14.7% | NIH ODS |
| Selenium | 400 mcg | 103 | 1 | 1.0% | NIH ODS |
| Vitamin C | 2,000 mg | 437 | 0 | 0.0% | NIH ODS |
Name-stated amounts only. A UL is not a toxicity threshold for any one person, and some products are intended for use under clinical supervision; this is a comparison of label text with a published reference value, not advice. A label qualifies when (1) the product name states the nutrient immediately followed by an amount (for example 'Zinc 50 mg'; 'Zinc Gluconate 50 mg' is excluded because the number may be compound weight), (2) the Supplement Facts rows include that nutrient, and (3) the serving size is exactly one discrete unit. Matches where the nutrient is joined to another by '&', 'and' or '+' are skipped (the amount may be a combined weight), as are figures over 3,000 mg in a single unit. Niacin excludes 'no-flush', inositol hexanicotinate and niacinamide products.
Finding 5
A third of multivitamins pair calcium with iron.
454 of 1,433 multivitamin-mineral labels (31.7%) list both calcium and iron in the same serving (36.2% of labels where the full panel was visible). The two compete for absorption, as the NIH iron and calcium fact sheets describe.
Note
Uses DSLD product type 'Multi-Vitamin and Mineral (MVM)'. 180 labels reach the 25-row cap without showing both minerals and are counted as not having both, so the count is a lower bound. Amounts are not assessed.
Finding 6
Four in ten labels ever recorded are already off the market.
92,821 of 214,780 label records (43.2%) are marked off market, a measure of how fast supplement products turn over.
Method
How the audit was done.
Every finding is reproducible from the published CSV, and each flagged product links to its public label page at the NIH so it can be checked.
Data and sample
- Source: NIH Office of Dietary Supplements, Dietary Supplement Label Database (DSLD), public domain. Labels entered 2011-10-26 to 2025-09-25.
- 214,780 label records; 121,959 marked on market; 92,018 after keeping one record per brand and product name.
- Seals come from the label’s own seal and symbol statements, matched against named programs.
- Nutrient amounts come from product names, matched conservatively, and compared only where the serving is exactly one tablet, capsule, softgel, gummy or drop.
Limitations
- DSLD coverage is not market share: it catalogs labels submitted or collected by NIH, weighted toward some brands and years, and says nothing about sales volume.
- Labels are not lab-tested contents. Every number describes what a label states.
- The local extract keeps up to 25 Supplement Facts rows and 12 seal statements per label and has no per-ingredient amounts, so amount findings rely on product names.
- Product-name matching is conservative and will miss labels that state amounts in other formats.
- DSLD entry date is when NIH entered the label, not when the product launched.
- Upper limits are reference values for the general adult population, not personal thresholds; some products are intended for use under medical supervision. This report describes label text and gives no advice about any product.
Reuse and citation
Our analysis is free to reuse under CC BY 4.0. Please cite: Williams B, Gassen D. Supplement Label Audit 2026. NutriStack. https://nutristackapp.com/reports/label-audit-2026
Read your own labels
Scan a label and see what it says.
NutriStack, our free iPhone app, reads supplement barcodes and labels, keeps your routine in one place, and shows the published research behind each ingredient.