Peppermint & the missing constituent?
...What "Pulegone-Free" Actually Means
There is a phrase moving through essential oil marketing that sounds like care and is mostly chemistry you weren’t told about. Pulegone-free. Low-pulegone. It appears on peppermint, spearmint, and various blends—all meant to reassure you. Yet it raises a question that the label is built to keep you from asking: ‘Free how?’
I want to walk through that question carefully, because pulegone is a genuinely useful teaching case. It is a constituent with a real toxicology file — not a marketing scare, not an internet rumor, but decades of peer-reviewed work and a federal carcinogenicity bioassay. And it is a constituent that the industry has learned to remove in a way that quietly changes the oil while leaving the label looking cleaner than the process was.
Both of those things are true at once. Holding them together is the whole point.
What pulegone is
Pulegone is a p-menthane monoterpene ketone — a ten-carbon molecule, structurally a near neighbor of menthone, carvone, and piperitone. It is the dominant constituent of pennyroyal oil (Mentha pulegium and Hedeoma pulegioides), where it can run as high as 85–97% of the oil. It also appears, at far lower levels, in peppermint, spearmint, and a scattering of other Mentha species. In the living plant, it is not an endpoint but an intermediate: peppermint’s own enzyme, (+)-pulegone reductase, converts it onward to menthone and ultimately to menthol as the leaf matures. Hold onto this point. It matters later.
What the toxicology actually says
This is where pulegone parts company from a constituent like piperitone, which has almost no dedicated mammalian safety literature. Pulegone has a deep one, and most of it traces back to a sobering source: pennyroyal poisonings, including fatalities, from the oil’s historical use as an abortifacient.
The mechanism is the part worth teaching, because it is a clean illustration of a principle that runs through all of constituent-level practice: the parent molecule is often not the problem — the metabolite is.
Pulegone itself is relatively inert. The damage is done by cytochrome P450 enzymes in the liver, which bioactivate it. One route converts pulegone to menthofuran, and then menthofuran to a reactive intermediate — a γ-ketoenal — that binds covalently to cellular proteins and injures hepatocytes (Thomassen and colleagues established this through the late 1980s and early 1990s; the original identification of pulegone as pennyroyal’s active hepatotoxin goes back to Gordon et al., 1982). For years, menthofuran was treated as the answer. It is not the whole answer. A second, parallel route depletes glutathione directly: in rats, pulegone extensively drained glutathione from both liver tissue and plasma, and its toxicity worsened sharply when glutathione synthesis was blocked — while menthofuran, given on its own, barely touched glutathione levels at all (Thomassen et al., 1990). So pulegone has at least two distinct paths to liver injury, which is precisely why menthofuran alone never accounted for all of it.
The in vivo picture is consistent. Moorthy et al. (1989, Toxicology) found that oral pulegone in rats produced dose- and time-dependent hepatotoxicity, with falling microsomal cytochrome P450 — and, tellingly, that menthone and carvone, close structural relatives, did not do the same thing. Small changes in a molecule produce large changes in what the liver does with it. That is the constituent principle stated in a single experiment.
The most recent and, for our purposes, most important study moved into human tissue. Working with precision-cut liver slices from five different patients, researchers found pulegone toxic in all five samples, and substantially more toxic than both menthofuran and acetaminophen. But the number that should stay with you is the variability: the response differed markedly from one person’s liver to the next. The authors concluded there is a high probability of significant differences in pulegone hepatotoxicity between people.
Read that again, because it is the entire thesis of biochemical individuality, demonstrated in human tissue with a single constituent. The same molecule, the same dose, five livers, five different outcomes. There is no population-level “safe” that survives contact with that finding intact — there is only this person, this enzyme profile, this exposure.
The NTP carcinogenicity bioassay: the actual numbers
In 2011 the U.S. National Toxicology Program published Technical Report 563 — a full two-year carcinogenicity study of pulegone in rats and mice. Because this report is the basis for nearly every regulatory action that followed, the specifics are worth stating plainly rather than gesturing at.
Pulegone (about 96% pure) was given by gavage in corn oil, five days a week, for up to two years, to groups of 50 male and 50 female F344/N rats and B6C3F1 mice per dose. Doses were 18.75, 37.5, and 75 mg/kg for male rats, and 37.5, 75, and 150 mg/kg for female rats and both sexes of mice.
The doses themselves tell a story before the tumors do. The high-dose groups — 75 mg/kg male rats, 150 mg/kg female rats — suffered so much mortality that NTP had to stop dosing them at week 60 and switch them to plain corn oil for the remainder of the study. Even so, only two of the high-dose male rats survived to the end, and none of the high-dose female rats did. This was not a subtle effect at the top dose.
The cancer findings:
Female rats — significantly increased urinary bladder tumors (papilloma, and papilloma or carcinoma combined) at the high dose.
Male and female mice — increased benign and malignant liver tumors.
Female mice — a small increase in rare bone tumors (osteoma or osteosarcoma).
Male rats — no increase in tumors.
Beyond cancer, pulegone produced a distinctive kidney lesion, hyaline glomerulopathy, across the dosed mice and the higher-dose rats; liver damage of several kinds; degeneration of the olfactory epithelium in the nose; and forestomach lesions. NTP’s formal conclusion used its strongest available language — clear evidence of carcinogenic activity — for female rats and for male and female mice.
Now, the honest framing, because a toxicology number without context is just a different kind of marketing. These were high, sustained gavage doses delivered directly to the stomach. Even the lowest two-year dose, 18.75 mg/kg/day, is enormously larger than any realistic human exposure from a properly distilled culinary or therapeutic oil. Some of the rodent site-specific findings — particularly the bladder and kidney lesions — come with genuine, ongoing scientific debate about how well the mechanism translates to humans. The NTP report itself states that extrapolating its results to human risk requires analysis beyond the report’s intent.
This is the line a constituent-literate practitioner has to be able to walk without falling off either side. The data are real and should not be waved away. And the data describe pennyroyal-scale and bioassay-scale exposures, not a drop of well-made peppermint oil in a diffuser. The case for caution scales with the constituent load: it is modest for peppermint, and it is not modest at all for pennyroyal. Pennyroyal is the oil that earns a hard stop.
What the regulators did — and what they didn’t
The regulatory response is narrower than the marketing implies, and the gap is instructive.
In 2018, the FDA removed synthetic pulegone from its approved food additives list. But the same action explicitly left naturally occurring pulegone — the pulegone already present in mint oils used as flavoring — untouched. And the FDA’s own stated reasoning was not that dietary pulegone was dangerous: the agency concluded that the intended use, in the small quantities involved, did not pose a public health risk. The delisting followed from the Delaney Clause, a provision requiring any additive shown to induce cancer in animals to be removed regardless of real-world exposure level. It was a legal trigger, not a risk verdict.
The European Union took the more transparent route: rather than a ban, it set numerical caps. Under Regulation (EC) 1334/2008, pulegone is limited to specific levels by product — for instance, 20 mg/kg in non-alcoholic mint beverages, 100 mg/kg in alcoholic ones, 350 mg/kg in chewing gum — and pure pulegone may not be added to food at all. For cosmetics, the Cosmetic Ingredient Review panel recommends that pulegone stay below 1% of a formulation.
So the regulatory picture is: a real concern, addressed with real limits, none of which amounts to “pulegone is poison.” Which brings us to the label.
“Pulegone-free”: free how?
Here is the question the marketing is built to skip. When an oil is sold as pulegone-free or low-pulegone, how did the pulegone get low? There are three answers, and they are not equivalent.
One: it was chemically converted. This is the most common industrial method, and the patent literature describes it without euphemism. Two U.S. patents (4,861,616 and 5,047,251) lay out a process for reducing pulegone in peppermint oil using sodium sulfite, which supplies hydrogen ions to hydrogenate the pulegone in place, at near-neutral pH. The patents are candid about what the process does: it does not lift pulegone out of an otherwise-intact oil — it transforms it. The resulting oil is described as having less than 0.5% pulegone and elevated menthone and menthol content compared to an untreated oil. That is the tell. The pulegone was hydrogenated into menthone, which carries downstream toward menthol, and the constituent profile shifts to compensate. The industrial reaction is, chemically, the same conversion peppermint’s own enzyme performs — just run non-enzymatically in a vat. That is a defensible reason to still call it peppermint oil. It is not a defensible reason to let a buyer assume the oil is what the plant made.
Two: it was rectified. Fractional distillation can lower pulegone, but pulegone boils close enough to menthone and other mid-range constituents that you cannot cleanly excise it without disturbing its neighbors. A rectified low-pulegone oil is also compositionally shifted — just by a different mechanism than chemical reduction.
Three: it was grown that way. This is the only low-pulegone oil that is not an altered oil. Because pulegone is a biosynthetic intermediate in the living plant, a late-season harvest of Mentha × piperita — and certain cultivars and growing conditions — yields an oil naturally low in pulegone, because the plant itself has already converted most of it to menthone and menthol. No vat, no rectification column. Just timing and agronomy.
Three oils. All three can legitimately print “low-pulegone” on a number. Only one of them is the oil that the plant actually produces.
What this means for practice
The lesson here is not “fear pulegone.” It is something more demanding and more useful: a label claim about a constituent is a claim about a number, and a number tells you nothing about provenance.
“Pulegone-free” does not tell you whether you are holding a late-harvest natural oil or a sulfite-reduced one. Those are different oils. They will have different menthone fractions, different menthofuran levels, different overall character — and a buyer who cares about working with what the plant made, rather than what a process left behind, cannot tell them apart from the front of the bottle.
So the practitioner’s move is the same one it always is. Ask the supplier directly: ‘Was this oil grown low in pulegone, or was it reduced or rectified to get there? And read the full GC/MS, where the signature of chemical reduction is legible if you know to look — an unusually high menthone or menthol fraction, paired with near-absent pulegone and depressed menthofuran, is the fingerprint of an oil that was processed rather than simply harvested at the right time.
Pulegone is worth knowing well, not because it is a villain, but because it teaches the whole method in miniature. A constituent with a metabolite-driven mechanism — the parent is quiet, the liver makes the danger. A constituent whose human toxicity varies from one person’s liver to the next — biochemical individuality, not as a slogan but as a measured result. A constituent whose regulation is more careful than the marketing built on top of it. And a constituent whose removal, sold as a clean reassurance, is often a chemical reaction the label declines to mention.
The oil-name on the bottle was never the unit of truth. The constituents are. And even the constituents, it turns out, don’t tell you the whole story unless you also ask how they got to be the numbers they are.
Selected references
Gordon WP, Forte AJ, McMurtry RJ, Gal J, Nelson SD. Hepatotoxicity and pulmonary toxicity of pennyroyal oil and its constituent terpenes in the mouse. Toxicol Appl Pharmacol. 1982.
Moorthy B, Madyastha P, Madyastha KM. Hepatotoxicity of pulegone in rats: its effects on microsomal enzymes, in vivo. Toxicology. 1989;55(3):327–337.
Thomassen D, Slattery JT, Nelson SD. Menthofuran-dependent and independent aspects of pulegone hepatotoxicity: roles of glutathione. J Pharmacol Exp Ther. 1990.
Khojasteh-Bakht SC, et al. Contribution of menthofuran to the hepatotoxicity of pulegone. (Matched AUC / time-course assessment.)
National Toxicology Program. Toxicology and Carcinogenesis Studies of Pulegone (CAS No. 89-82-7) in F344/N Rats and B6C3F1 Mice (Gavage Studies). NTP Technical Report 563. August 2011.
Human precision-cut liver slice study on inter-individual variability in R-pulegone and R-menthofuran toxicity (PMC6032148).
U.S. Patents 4,861,616 and 5,047,251 — Stable peppermint oil having reduced pulegone content and method of producing the same.
Regulation (EC) No. 1334/2008 on flavourings; FDA 2018 final rule removing synthetic pulegone from the food additive list.
Note: a few citation details (author lists, exact volume/page numbers for the Thomassen and Khojasteh-Bakht papers) should be verified against the primary sources before publication.
Tammy L. Davis, MCN, is a Master Clinical Neuroaromatherapist and founder of Aromagenomics. She trains licensed healthcare professionals in evidence-based neuroaromatherapy and consults as a private neuroaromatherapy concierge for complex clinical cases. She is the developer of the ANIS™ (Aromatic Neural Integration Solutions).