Mould-free and mycotoxin-tested coffee is a real product category, sold at a premium on the claim that ordinary coffee is contaminated with fungal toxins that cause fatigue, brain fog and worse. The underlying fact is true: coffee can contain mycotoxins, they are measured routinely, and several countries regulate how much is allowed. What the marketing leaves out is everything that happens between the green bean and the cup, which is where most of the toxin goes.
What mycotoxins are and which one matters
Mycotoxins are toxic compounds produced by certain moulds. In coffee the one that matters is ochratoxin A, produced mainly by Aspergillus and Penicillium species that can grow on beans when drying or storage goes wrong, particularly in humid conditions. It is of concern because it is nephrotoxic, meaning it damages the kidneys, and it is classified as a possible human carcinogen.
Aflatoxins, the mycotoxins best known from peanuts and maize, are the other candidate, but they are much less of an issue in coffee. The reviews below consistently find ochratoxin A to be the relevant compound.
How common is it?
The most comprehensive answer comes from Fakhri and colleagues[1], who meta-analysed studies published between 2010 and 2022 on ochratoxin A in coffee products, then modelled consumer health risk using Monte Carlo simulation. The overall prevalence of detectable ochratoxin A in coffee products was 58.01%.
That number sounds alarming and needs immediate context. Prevalence here means detectable by sensitive laboratory methods, not present at a harmful level. Modern analytical chemistry detects contaminants at parts per billion, so a high detection rate across a food supply is normal rather than newsworthy. The question that matters is concentration against a health-based threshold.
Li and colleagues[2] reviewed ochratoxin A across nine food categories, including cereal, wine, coffee, beer, cocoa and dried fruit, and found occurrence and contamination levels varying greatly with geography, climate and storage conditions. Coffee is one contributor to total dietary exposure, not the only one, which is a point the risk assessments return to.
What roasting does
This is the part the marketing omits, and it is the most important finding in this article. van der Stegen and colleagues[3] took a commercial lot of green coffee naturally contaminated with ochratoxin A and roasted it under a range of commercial conditions, from 2.5 to 10 minutes and from light medium to dark. The differences between roasting times and colours did not reach statistical significance, but the reduction itself was highly significant: 69% across the combined results.
They also reviewed the wider literature, noting that of nine studies available on ochratoxin A reduction during roasting, seven reported reductions between 69% and 96%, and that those seven included all four studies using naturally contaminated beans with mycotoxin-appropriate sampling and realistic roasting conditions. Three mechanisms appear to contribute: physical removal with the chaff, isomerisation, and thermal degradation.
Blanc and colleagues[4] followed ochratoxin A along an industrial soluble coffee manufacturing line using naturally contaminated Thai robusta. Green coffee cleaning removed a small proportion, and roasting did most of the work, leaving roast and ground coffee with only 16% of the original ochratoxin A. A further 20% reduction occurred during soluble coffee manufacture, so the finished powder contained 13% of what the green beans held.
The older review by Levi[5] reached a similar conclusion decades earlier, reporting that experimental roasting destroyed 70% to 80% of toxin added to green coffee, and that findings in commercial green coffee were so infrequent and at such low levels that further study was discontinued.
Is the amount left actually risky?
Risk assessors answer this with the margin of exposure, which compares estimated intake against the dose associated with effects in animal studies. A margin above 10,000 is conventionally treated as low concern.
In the global assessment by Fakhri and colleagues[1], mean margins of exposure for adult consumers were above 10,000 in most countries. Two were below: Panama at 9,526 and the Philippines at 8,873. The highest pooled concentrations were found in the Philippines at 39.55 micrograms per kilogram, Turkey at 39.32 and Panama at 21.33. The authors' conclusion was that monitoring and control plans should be carried out in different countries, which is a regulatory recommendation rather than advice to consumers to stop drinking coffee.
A national assessment illustrates how these numbers usually come out. Sevim and colleagues[6] analysed 41 Turkish coffees and modelled exposure across three age groups. Some samples exceeded Turkey's permissible maximum of 5 micrograms per kilogram, and the mean was 8.41. Even so, margins of exposure were above 10,000 for all age categories, and the authors concluded the levels did not pose a risk to those groups, while emphasising the need to minimise and control concentrations and to consider other dietary sources.
That combination, some samples over the legal limit and margins of exposure still in the reassuring range, is typical of this field and is worth understanding. Regulatory limits are set conservatively, well below the level where harm is expected, so exceeding one is a compliance problem rather than a poisoning.
Reading this from the Philippines
This is the rare article where the Philippine data is not a footnote. In Fakhri and colleagues[1], the Philippines had the highest pooled ochratoxin A concentration of any country in the analysis, and one of only two adult margins of exposure below the 10,000 threshold.
The likely explanation is climate and post-harvest handling rather than anything about Philippine coffee itself. Ochratoxin-producing fungi thrive in warm, humid conditions, and the toxin forms when beans dry too slowly or are stored damp. That makes it a drying and storage problem, which is exactly what Fakhri and colleagues[1] addressed by recommending monitoring and control programmes.
For a coffee drinker here, the practical reading is measured. The exposure estimate is a population average built from the coffee tested, and it still sits close to a conservative threshold rather than at a level associated with harm. Roasting removes most of what is present, and buying from roasters who source carefully and control storage is a more meaningful response than paying a premium for a mycotoxin-free label. The local industry context is covered in our article on the Philippine coffee industry.
Does the brewing method change what reaches you?
Roasting is the main filter, but it is not the last step. Ochratoxin A is water soluble to a degree, so how much transfers from grounds into the cup depends on the brewing method, contact time and the coffee-to-water ratio. Studies of brewed infusions have found that only part of the toxin present in the ground coffee ends up in the drink, with the remainder staying in the spent grounds.
That adds another reduction on top of roasting, and it means measurements reported per kilogram of ground coffee overstate what a drinker actually consumes. It also means the comparisons people want, whether espresso is safer than filter or whether instant is worse, cannot be answered confidently from the evidence here. What can be said is that the industrial process for soluble coffee reduced ochratoxin A further rather than concentrating it, according to Blanc and colleagues[4], which is the opposite of the usual assumption about instant.
Robusta, arabica and where the toxin comes from
Contamination is not evenly spread across the coffee world. It tracks growing and post-harvest conditions: humid climates, wet harvests, slow drying, and storage in conditions where beans reabsorb moisture. Li and colleagues[2] identified geography, climate and storage among the main determinants across food commodities generally, and coffee follows the same rule.
Robusta has historically shown higher contamination rates than arabica in survey data, which reflects where and how each is typically grown and processed rather than anything intrinsic to the species. The difference between an origin with careful drying infrastructure and one without is larger than the difference between species, which is why Fakhri and colleagues[1] framed their recommendations as country-level monitoring rather than advice about which bean to buy.
What the mould-free claims get wrong
Several things, on the evidence above. Testing green coffee and reporting the result says little about the cup, since roasting removes most of the toxin. Detectable is not the same as harmful, and at parts-per-billion sensitivity almost everything is detectable. And the symptoms attributed to coffee mycotoxins in marketing, fatigue, brain fog, headaches, are not what ochratoxin A toxicology describes, which is chronic kidney damage at sustained high exposures.
None of that means mycotoxin testing is worthless. A roaster testing incoming green coffee is doing quality control, and in producing countries with humid harvests it is genuinely useful. The gap is between that legitimate practice and the consumer-facing claim that ordinary coffee is making people ill.
Storage at home
Coffee mould is a post-harvest problem, and the same principle applies in a kitchen: moisture and warmth allow fungal growth. Roasted coffee is dry and inhospitable, which is why mould on roasted beans is rare, but brewing equipment, damp grounds left in a portafilter or a reusable filter, and coffee stored somewhere humid are all more plausible sources of mould exposure than the beans themselves.
Practical storage advice, which is about staleness more than safety, is covered in our guide to storing coffee beans and in our article on staling.
How regulators handle it
Coffee is one of the food categories with legally binding maximum levels for ochratoxin A in several jurisdictions, set separately for roasted and soluble coffee, and it is routinely sampled at import. The Turkish limit of 5 micrograms per kilogram referenced by Sevim and colleagues[6] is typical of the range in which these limits sit.
The existence of a limit is why the surveys exist in the first place, and it is also why the numbers look worse in isolation than they are. A regulatory maximum is set with a deliberate safety margin below the level where harm would be expected, so samples above it trigger enforcement rather than illness. When a study reports that some samples exceeded a limit while margins of exposure remained above 10,000, both statements are true and the second is the one about health.
What a drinker can actually control
Very little of this is in a consumer's hands, and it is worth being honest about that rather than offering reassuring rituals. Ochratoxin A forms before the coffee reaches you, during drying and storage at origin and in transit, and the largest reduction happens during roasting, which is also not your decision. Nothing done in a kitchen removes it.
What remains is buying from roasters who care where their green coffee comes from and how it was stored, which is the same thing that produces better-tasting coffee, and keeping your own storage dry, which is about staleness more than safety. If a producer publishes test results, that is a reasonable signal of quality control. It is not evidence that untested coffee is dangerous, which is the inference the marketing wants.
The limits of this evidence
Mycotoxin distribution in a lot of coffee is notoriously uneven, which is why sampling procedure matters so much and why older studies with poor sampling produced inconsistent results. Detection methods differ between studies, with immunoassay and chromatographic methods giving different sensitivities. The risk assessments model exposure from consumption surveys, which are estimates. And margins of exposure are based on animal toxicology extrapolated to humans, which carries its own assumptions. Nobody has followed coffee drinkers to see whether ochratoxin exposure from coffee affects their kidneys.
The practical answer
Ochratoxin A is present at detectable levels in a majority of coffee products worldwide, and roasting destroys most of it, leaving roughly 13% to 31% of what was in the green bean. Risk assessments generally find exposure well below levels of concern, with a small number of countries, including the Philippines, sitting closer to the threshold, which is an argument for better drying, storage and monitoring rather than for avoiding coffee. Mould-free coffee marketing overstates both the prevalence of a problem in the cup and the symptoms attributed to it. See also our article on coffee and kidney health and our article on safe caffeine limits. Nothing here is medical advice.