A bag of rice had spent eight months inside a recycled-cardboard box when scientists at the Zurich Cantonal Laboratory tested it in 2009. No one had poured oil onto the rice. Yet the laboratory found mineral-oil hydrocarbons in the food.5
To work out how they had arrived, the scientists had to follow the journey backwards. The box had once been newspapers. The newspapers had carried ink, and that ink had used high-boiling mineral oil to move black pigment through fast presses. When the old paper became new board, some of the oil came with it. The lighter parts could then leave the cardboard as vapour, cross the air inside the closed package and enter the rice while it waited on the shelf.45 The surprise was not that someone had put petroleum into a food recipe. It was that a useful ingredient from one industry had quietly acquired a second job in another.
That rice box gives us a way into mineral oil, but it is only one doorway. The same family of petroleum mixtures can carry printing ink, lubricate a machine, soften a cosmetic or work as a laxative. It can reach us through food, lungs and skin, and it can reach wildlife through water. The shared name makes these uses sound like one substance with one risk. They are neither. To know what might cause harm, we have to follow the oil: how it was refined, what was mixed with it, what job it was doing and where it went next.
The oil that rose from a salt well
Mineral oil’s modern story did not begin with a cereal box. In western Pennsylvania in the early 1850s, Samuel Kier was drilling salt wells when petroleum kept rising with the brine. At first it was a nuisance. Kier tried bottling some of it as medicine, but the wells produced more oil than that market could absorb, so he took the problem to the chemist James Curtis Booth.1
Booth saw a different use. If the dark petroleum could be distilled—heated so that parts with different boiling ranges separated—it might yield a cleaner fuel for lamps. Kier built a still that could handle one barrel at a time. When the lamp fuel sold, he built a larger one. The unwanted liquid from a salt well had become a set of useful materials because refiners had learned to separate it.1
That is still the essential idea. Modern refining goes much further. Petroleum is not one molecule. It is a changing mixture, and distillation divides it into broad fractions. Solvents, acid treatment and hydrogen-based refining can then remove or change more of what remains. Severe refining can produce the clear, nearly colourless materials called white oils. Untreated or mildly treated oils, especially older ones, could retain much higher levels of polycyclic aromatic hydrocarbons, or PAHs. This family of fused-ring chemicals includes known causes of cancer.2
So “mineral oil” is a family name, rather like “metal” or “plastic”. It tells you where the material came from, but not enough about the member in front of you. Food scientists use two broad groups. MOSH is the saturated part; MOAH is the aromatic part. A machinist may meet a base oil blended with additives and water. Metals and microbes can change it again during use. A pharmacist may use a highly refined oil intended to stay in the gut. Each version exists because petroleum mixtures can be tuned to lubricate, resist water, carry pigment, release food from machinery or soften a formulation.238
That usefulness is why the oil appears in so many places. It is also why the name alone cannot answer the question we actually care about. The next question is not simply, “Is mineral oil harmful?” It is, “Which oil, moving from which job, reached which body or ecosystem?”
How printing oil found its way into food
Return to the Zurich rice. Cold-set newspaper presses needed a fast ink. It had to carry carbon black and soak into porous paper without a separate drying step. High-boiling mineral oil did that job well. Then recycling joined two industries that had not designed their materials together. Printed newsprint became fibre for food boxes, bringing residues from yesterday’s ink into tomorrow’s packaging.45
Once the rice result appeared, the German Federal Institute for Risk Assessment and the Zurich laboratory faced another problem. The available method depended on specialist equipment. Most official food laboratories could not use it routinely, so the two laboratories adapted it. They separated the saturated and aromatic hydrocarbons by hand. Gas chromatography then let them separate and count parts of the mixture.23
In September 2011, they took the method and a development kit to a meeting of roughly 400 scientists, officials, businesses and civil-society representatives. The group also examined packaging barriers that might interrupt migration.23 An odd result in stored rice had now changed what laboratories could look for. It had not shown that every cardboard box contaminated its food, but it had made the route visible enough to test.
And the box is not the only possible route. Mineral-oil hydrocarbons may enter food through machinery lubricants, release agents, dust binders, ingredients, transport, glues or printing inks. A laboratory result can tell you what reached the food; it cannot, by itself, tell you which part of that history supplied it. In the Zurich case, the known packaging and migration work supported the recycled-board explanation. A current product would need its own investigation.35
What does finding these hydrocarbons mean for health? In 2023, the European Food Safety Authority revisited the evidence. For MOSH, it judged it likely to very likely—66–95 per cent certain—that present dietary exposure did not raise a health concern, while noting that possible long-term effects still need study.6
MOAH is where the food concern narrows. The possible DNA-damage and cancer concern centres on aromatic compounds with three or more rings. Routine “total MOAH” results do not tell laboratories how much of that particular group is present. That leaves a real concern, but not the sweeping conclusion that cereal consumers have been shown to develop cancer. The historical rice case established movement from board into food; it did not measure a typical British cereal exposure today or the dose of the three-or-more-ring compounds.56
The rules try to meet this problem from several directions, and none tells the whole story. Great Britain’s food-contact framework says materials must not transfer their parts to food at levels that could endanger health. The authorised list for plastic food-contact materials includes one defined paraffinic white mineral oil—but that permission covers a highly refined ingredient in plastic. It is not a certificate for recycled paperboard or for the finished cereal pack in your hand.78
A barrier inside a box can help, but “has an inner bag” is not a test result. Controlled experiments show that migration changes with the barrier material and with the particular mineral-oil component being measured. A pack maker therefore needs evidence for the relevant material, migrant and storage conditions, not a generic promise that a layer exists.9
The Forest Stewardship Council mark answers a useful but different question. It can trace forest material through a chain of custody. It does not test whether mineral-oil hydrocarbons move through a finished food pack.10 If you see the mark on a cereal box, read it as evidence about fibre sourcing, not as a migration certificate.
An official food alert is more direct. In 2025 the Food Standards Agency recalled a named confectionery product because it contained mineral-oil hydrocarbons, then updated the notice in January 2026.11 That alert told shoppers exactly which product to avoid. It did not turn every cardboard pack into the same case.
European Union countries have also moved towards controlling the food endpoint rather than waiting for every source to be found. They voted in May 2026 for maximum MOAH levels in food; as of 21 August, the European Commission catalogue still targeted final adoption for October. The measure was not yet Great Britain law.12 Its logic is still worth noticing: when an oil can enter through several doors, a limit at the food itself can stand guard while investigators trace the route.
For a shopper, the honest action is modest: follow named food alerts, and do not treat “recycled”, FSC or “inner bag” as proof of mineral-oil performance. The stronger choices sit upstream. Printers can change the ink source; packaging buyers can demand migration evidence for the finished construction; food businesses can test the endpoint. Recycling carried the old ink into a new job, but recycling itself is not the villain. The chemical choice and the missing barrier are the places to intervene.
When a laxative reaches the lungs
The food route is about what we swallow. A case reported in 2020 shows why swallowing and safety are not always the same thing.
A 66-year-old woman with multiple sclerosis had been treated for pneumonia twice before she arrived at hospital with breathing difficulty for a third time. Her blood oxygen was low. Scans showed cloudy patches in several parts of her lungs, but the pattern did not explain why the illness kept returning. Doctors washed fluid from her airways and examined a tissue sample. Both contained immune cells filled with fat.13
That finding changed the investigation. Fat-filled cells suggested that an oil had entered the airways, so the clinicians went back through the woman’s medicines and exposures. They found repeated use of mineral oil for long-term constipation and no other contributing exposure.13
As a laxative, highly refined mineral oil is meant to lubricate stool while passing through the digestive tract. The trouble begins if swallowing is impaired, or if the liquid comes back up and is breathed in. Once it reaches the small airways and air sacs, the body struggles to clear it. Inflammation can follow, causing exogenous lipoid pneumonia—lung disease caused by fat or oil that came from outside the body.
The doctors stopped the mineral oil. The woman improved, and scans after three and six months showed that the lung changes had largely cleared. A separate study of 17 children with known mineral-oil exposure found fat confirmed in their lungs and dense areas in the air spaces on every scan.1314 These cases do not tell us how often aspiration happens among all users. They do show a complete route: a useful oil was meant for the gut, entered the lung instead, produced damage and improved when the exposure stopped.
That is why a current US label for oral mineral oil tells people with difficulty swallowing not to use it and names other higher-risk groups.15 In Britain, National Health Service guidance describes non-oil options, including bulk-forming and osmotic laxatives, and directs people to a pharmacist when needed.16 Those alternatives work differently and are not suitable for everyone. The choice is not “natural versus petroleum”; it is the medicine and route that fit the person without creating an aspiration risk.
The factory where the mist disappeared
In 2013, workers at a US plant making parts for automotive water pumps reported breathing, skin and stomach problems. Their machines used a soluble-oil metalworking fluid. It lubricated and cooled the cutting, but the moving tools also threw some of the fluid into the air as mist. A manager asked the National Institute for Occupational Safety and Health—NIOSH—to investigate.17
NIOSH returned four times across 2013 and 2014. Investigators spoke with workers, examined rashes and measured the mist while different jobs were under way. They matched meter readings to video, so they could see which action made exposure rise. Forty-four per cent of machine-operator measurements exceeded the NIOSH recommended limit before the plant changed its controls. Blowing parts clean with compressed air created short, sharp peaks.17
The company improved local exhaust ventilation and added splash shields to the air guns. When NIOSH measured again, none of the sampled workers was above the recommended mist limit.17 That is what makes this more than a list of workplace hazards: the investigators found a route, the company interrupted it, and the next measurements showed that the interruption worked.
This investigation continues below.
Want the next one in your inbox?
The health evidence belongs to the whole fluid, not to the base oil alone. Metalworking fluids may contain additives and germ-killing chemicals. They can then collect metals or microbes in the sump. Repeated breathing of their mist has been linked with asthma, bronchitis and hypersensitivity pneumonitis. That last condition is an immune-driven swelling of the lungs. The plant investigation cannot tell us that mineral oil caused every reported symptom, and it says nothing about ordinary household vapour.18
British employers meet the same route through the Control of Substances Hazardous to Health rules and Health and Safety Executive guidance. Enclosures, working local extraction, careful fluid management, less compressed-air cleaning, skin protection and health checks can all break the pathway.19 A substitute fluid may help, but it brings its own formulation and performance questions. The proof is not the word on the drum. It is what is in the whole mixture and what air measurements show around the person using it.
The older cancer record belongs to a different workplace story. The International Agency for Research on Cancer examined workers who spent years with untreated or mildly treated oils on their skin, including people in machining and textile jobs. One record covered 344 cases of scrotal cancer in the West Midlands between 1936 and 1976. Sixty-two per cent of the men had worked in jobs likely to expose them to mineral oil, and the middle point in the recorded delay from exposure to diagnosis was 34 years.220
Across the wider evidence, the agency concluded that untreated or mildly treated mineral oils caused cancer in humans, with the strongest link to squamous-cell skin cancer. Severe refining changes that picture because it removes much of the aromatic material found in the older oils. Human cancer evidence for highly refined oils was inadequate.220
“Inadequate” is not a universal safety certificate. It means the old cancer finding cannot simply be pasted onto a modern white-oil cosmetic or medicine because the family name is the same. Nor can refining prevent aspiration or answer for every additive. A sound workplace specification therefore joins verified refining grade to closed handling, limited skin contact and measurements showing that the controls work.
When the next receptor is a river
Some lubricants are used in machines that can lose oil as part of normal work; others reach water or soil through a leak. Once that happens, the question changes again. The exposed body is no longer a person eating, breathing or touching the oil. It may be an alga, a small water animal or a fish meeting the finished lubricant.21
Tests follow those organisms in different ways. They can ask whether a product slows algal growth, stops small aquatic animals moving or breeding, or causes short- and long-term effects in fish. They also measure whether its ingredients break down and whether they build up in living things. A test result cannot prove that every spill has damaged a real ecosystem, but it can show which formulation presents the lower expected hazard before a release happens.21
The US Environmental Protection Agency confronted the language problem in a 2011 report for vessel operators. Products were already being sold with words such as “environmental”. The agency separated those broad claims from lubricants that had passed tests for breakdown, water toxicity and build-up in living things. It then compared vegetable oils, synthetic esters and polyalkylene glycols with the jobs they still had to perform.21
There was no perfect plant-based answer waiting at the end. Vegetable oils may oxidise or struggle at extreme temperatures. Synthetic esters can cost more or damage some seals. Polyalkylene glycols may be incompatible with other oils or coatings, and additives can change the toxicity of any base oil. The alternative is therefore a specification, not a slogan: the whole product must show lower environmental hazard and still work safely in the machine.
The European Union Ecolabel for lubricants uses that whole-product logic. It covers certain lubricants that can be lost fully, partly or by accident, and it requires environmental performance alongside technical fitness. The present criteria run through 2028.22 For a vessel, hydraulic system, chainsaw or other loss-prone use, a buyer can ask for that label or equivalent evidence covering breakdown, water toxicity, bioaccumulation and compatibility. Containment and maintenance still matter. A better-tested oil is not a licence to release it.
Follow the oil, not the family name
We began with rice in a cardboard box because that journey is easy to miss. Mineral oil left printing ink, survived a recycling loop, crossed the air inside a package and reached food without anyone pouring it there. But the box was never the whole story.
Kier’s unwanted petroleum became lamp fuel when Booth found a way to separate it. A refined laxative helped the gut until swallowing sent it towards a lung. A factory fluid became mist until engineers captured it. Older, less-refined oils stayed on workers’ skin long enough for cancers to appear decades later. A lubricant can remain between moving surfaces or escape into water. In every case, the useful job comes first; harm becomes possible when a particular mixture takes a route its specification and controls did not safely contain.
That gives us a better question than “Is mineral oil safe?” Ask what grade it is. Ask what else is in the mixture, where it is meant to stay and what evidence shows that it stays there. For food, that may be a named official alert or finished-pack migration testing. For medicine, it is an aspiration warning and advice matched to the person. At work, it is refining evidence, enclosure, extraction and exposure measurements. Near water, it is whole-product breakdown, water-toxicity, build-up and performance data.
No single badge can answer all of those questions. What protects us is evidence tied to the route—and a choice that interrupts it before the oil reaches the wrong place.
The Magic Wand