Shilajit and Urolithin A: What a New 10-Sample Study Found

Who paid for it

Read this first, because it shapes how to read everything else. The paper's first author, Nodari Rizun, is the founder and CEO of Pürblack Inc., a US company that sells shilajit resin, and he holds shilajit patents. Pürblack paid the article fee and supplied the ten samples. The authors declare all of this and say the samples were coded and analysed blind, with the elemental and urolithin A measurements run in outside laboratories. The other authors are at institutions in Serbia and the US and declare no competing interest. doi:10.1038/s41538-026-01197-5

ShilajitDB earns no money from Pürblack or any other brand (see our disclosure). Pürblack's resins grade A+ here because of their independent finished-product lab reports, which we describe in our Pürblack review. This paper changes no grade. Industry funding does not make the measurements wrong, but it is the reason to look hardest at the framing and the choices of comparison, which is what the rest of this page does.

What the researchers did

They crushed shilajit-bearing rock from five mountain regions (the abstract names the Himalayas, Andes, Altai, Caucasus and Carpathians; the sample list also names Mongolian and Sierra Madre sources) and turned each batch into a resin with their own process: hot water, then ethanol, then high-pressure homogenisation, then evaporation. Three of the ten are blends of rock from two regions. They then measured:

• how each resin melts, by differential scanning calorimetry;
• 17 elements by ICP-MS, including arsenic, cadmium, lead and mercury;
• urolithin A and B by LC-MS/MS, using two different extraction solvents;
• how well each resin stopped the growth of five bacteria and five fungi in a dish.

These are lab-prepared resins, not retail jars. The paper does not say which, if any, match a product Pürblack sells, so you can't read a result across to a particular jar. It also describes the samples inconsistently: the methods list ten samples from five regions, while the funding statement calls them samples "from a single origin".

The heavy metals

Mercury was not detected in any sample. The three metals that matter most for a supplement came out as follows (all in parts per million, which is the same as mg/kg). The lead-per-dose column is ours: it uses the paper's own assumption of 0.5 g of resin a day.

Cadmium was 0.01 to 0.10 ppm in every sample. Urolithin A is the methanol/ethyl acetate extract, the higher of the two the authors ran. Amber is above 0.5 µg of lead a day at 0.5 g. Source: Tables 2 and 3 of the paper.

The authors conclude that all three metals sit "well within" FDA and WHO tolerable intakes. That is true against the limits that govern supplements: the highest lead figure, 3.01 ppm, gives 1.5 µg a day at 0.5 g, under the 5 µg a day we list for USP <2232> in our heavy-metals guide. Two things the headline leaves out:

• Half the samples are above California's lead figure. The paper itself cites Proposition 65's 0.5 µg a day for lead and calls it much stricter than the FDA guideline. At 0.5 g a day, five of the ten resins (0, 1, 3, 7 and 9) supply more than that. We use the same figure when we read a lab report; see is shilajit safe.
• The lead benchmark the paper leads with has been withdrawn. It compares lead to a WHO weekly intake of 25 µg per kg of body weight, about 250 µg a day. The WHO and FAO's expert committee withdrew that figure in 2010 because it could no longer be considered health protective (see the Hong Kong Centre for Food Safety's summary). Measured against 250 µg, almost anything passes.

The more useful finding is the spread. Lead differed nearly nine-fold between the lowest and highest resin (0.35 and 3.01 ppm), and arsenic nearly tenfold. Rock from one region is not a guarantee, and a result for one resin tells you little about the next. That is the case for a numeric report on the finished product, batch by batch, which is what our grade rewards (see how to read a COA). For scale, the seven Pürblack finished products we track have independent Cambium Analytica reports giving lead of 0.26 to 1.04 mg/kg, inside the range of this paper's samples.

Urolithin A: the headline, and the dose

Urolithins are compounds that bacteria in the human gut make from ellagic acid, which comes from foods such as pomegranate, walnuts and berries. Urolithin A is the one researchers have studied most, as a purified supplement for muscle and mitochondrial health. The paper reports that shilajit contains it, and calls it the resin's "most promising bioactive component". The measurements:

• It depends on the solvent. Pure methanol pulled out 0.07 to 1.63 ppm. A half-and-half mix of methanol and ethyl acetate pulled out 3.5 to 58.5 ppm, 20 to 65 times more. A single "urolithin A content" number is only meaningful with the method attached.
• Urolithin B was never found. And treating the extracts with acid lowered the urolithin A reading in every sample.
• The highest values were in the blends. The three highest resins (4, 5 and 6) were blends of rock from two regions, and the authors link high values to oak-dominated regions as a plausible source of ellagic-acid precursors. With ten samples, three of them blends the authors made, and no statistical test of that link in the paper, treat it as a hypothesis.

Now the dose. In a 2022 double-blind trial, Liu and colleagues gave adults aged 65 to 90 1,000 mg of purified urolithin A a day for four months. PubMed 35050355. The richest resin in this paper had 58.5 ppm, which is 58.5 µg per gram, so a typical 0.5 g dose carries about 29 µg of urolithin A. That is roughly 34,000 times less than the trial dose. Even the 300 to 4,200 ppm range from another analysis that the authors cite would give 0.15 to 2.1 mg, which at best is still around 500 times less than 1,000 mg.

That comparison is not proof that shilajit's urolithin A does nothing; the trial tested a purified compound, not shilajit, and no one has tested the reverse. But it does mean a shilajit label cannot borrow the human results for urolithin A, and the paper measured no biological effect of urolithin A in shilajit at all. The "most promising" line is the authors' framing of where to look next, which they say needs standardisation and bioavailability studies first.

The antimicrobial result

The resins stopped some bacteria (Bacillus subtilis, Staphylococcus aureus) and some fungi only at 3 to 25 milligrams per millilitre, tens to hundreds of times the amount the control antibiotics needed, and did nothing to E. coli, Proteus mirabilis or Klebsiella pneumoniae. The authors themselves call this weak and say it does not support using shilajit as an antimicrobial. A dish is not a gut, so it says nothing about taking shilajit either way.

What this study does not show

• That shilajit does anything for energy, strength, testosterone or ageing. No person, animal or cell was dosed.
• That any product you can buy contains these amounts. The samples were made in a lab from rock, not taken from retail jars.
• That origin determines quality. Ten samples can show variation; they can't establish rules about regions.
• That urolithin A is why shilajit might work. The paper measures it; it does not test it.
• That the results are final. The paper is an early-access manuscript and may change before the final version.

What to take from it as a buyer

• Treat a brand's "contains urolithin A" claim like a fulvic acid percentage: ask which method and what dose it delivers, and don't read it as a benefit.
• Look for a numeric lab report for the finished product, because metals varied this much between sources. Our best third-party tested list is ranked on exactly that.
• Weigh heavy metals per daily dose, not per gram. The conversion is in is shilajit safe.
• Read industry-funded research the way you would read a brand's own claims: useful data, framed by the people who sell the product.

For what human studies have actually tested, see the clinical trials, explained, and for how resin is made and bought, the shilajit resin guide.

Common questions

Source: Rizun et al., npj Science of Food, 2026. This page is general information from published research, not medical advice.