Who paid for the science on aspartame? We followed the money through 40 years of studies

Since 6 April 2018 every soft drink sold in Britain with more than five grams of sugar per 100ml has carried a levy. The industry's answer was to take the sugar out, and it did so faster than the Treasury expected: the forecast take fell from £520m to £240m before the tax was even collected. What replaced the sugar was mostly aspartame, acesulfame K and sucralose, the sweeteners that let a can taste roughly the same for a fraction of the calories. The question we were asked to answer is whether that swap is safe, who paid for the science that says it is, and whether the rise in cancer that people notice around them has anything to do with it. We set out expecting to find that the last of those does not hold. It does not. But the first two questions turned out to have better answers than the reassurance suggests, and the third has a UK-specific gap that nobody has filled.

What the cancer agency actually said

On 14 July 2023 the International Agency for Research on Cancer, the WHO's cancer arm, announced that aspartame was "possibly carcinogenic to humans", Group 2B on its four-step scale. The basis was "limited evidence" for one cancer, hepatocellular carcinoma of the liver, and "limited" evidence in animals and in mechanism. The full monograph, published in 2024, is precise about where the human evidence came from: "only three studies (including four cohorts) on the consumption of artificially sweetened beverages were available for assessing the association between aspartame and liver cancer", the European EPIC cohort, a pooled American analysis, and the American Cancer Society's cohort. In all three there was a positive association, "but bias or confounding could not be ruled out as an explanation for the findings". For every other cancer the evidence was rated inadequate.¹²³

Two things the press release did not say are in the monograph. The working group was split: "A minority of the Working Group supported a Group 2A classification", the higher "probably carcinogenic" category, "but these conclusions were not supported by the Working Group overall". And the 25 members included the senior author and a co-author of the French NutriNet-Santé study, the first author of the American Cancer Society study, the author of the Nurses' Health Study analysis, who chaired the human-evidence subgroup, and a scientist from the Ramazzini Institute in Bologna, whose rat and mouse studies were the animal evidence under review. IARC's rules say members serve "in their individual capacities as scientists" and all 25 were "assessed as free from conflicts of interest". The one invited specialist whose university had an arrangement with a company linked to the manufacturer Ajinomoto was, under the same rules, barred from drafting or voting.²³

The same day, the WHO's food additives committee, JECFA, did the second half of the job: not whether aspartame can cause cancer, but whether it does at the doses people eat. It kept the acceptable daily intake at 40 milligrams per kilogram of body weight, unchanged since 1981, and worked the example everyone quotes: a 70kg adult "would need to consume more than 9–14 cans per day" of diet drink to exceed it. It also did something more useful for a piece about children. It estimated actual exposure at "up to 10 mg/kg bw per day for children" on average and "up to 20 mg/kg bw per day for children" at the high end, half the limit, and concluded that "dietary exposure to aspartame does not pose a health concern". Its reasoning is that aspartame never reaches the bloodstream: it is broken down in the gut into phenylalanine, aspartic acid and methanol, all of which come in larger amounts from ordinary food. The largest single human study, of 474,000 Americans followed by the National Cancer Institute, found no link between aspartame drinks and blood or brain cancers in 2006.¹⁴¹¹

The French study that is usually blamed for the ruling did not carry it. Debras and colleagues followed 102,865 adults for a median of eight years and found a 15% higher overall cancer risk among higher aspartame consumers, and 22% for breast cancer. It was entirely publicly funded, by the French health ministry, INSERM and the European Research Council, with one author's only declared interest a non-profit food database. But its authors wrote that "the limited number of cases prevented us from studying associations for other cancer sites (e.g., … liver …)", and IARC says the same: the NutriNet-Santé study "did not investigate liver cancer separately". It is the best exposure data anyone has, and it is not the study the liver finding rests on.³⁵

Europe's regulator did its full re-evaluation in 2013 and found "no safety concerns at the current ADI of 40 mg/kg bw/day"; a 2019 academic audit of that review found the panel had judged all 73 studies suggesting harm unreliable while accepting 62 of 81 that did not, which EFSA disputed in print. On 10 September 2026, after IARC, after the monograph, EFSA published an update of its aspartame assessment and "confirmed the ADI of 40 mg/kg bw per day", finding "no new studies suitable" to change it. The American regulator was blunter: "The FDA disagrees with IARC's conclusion."⁶⁷⁸⁹¹⁹

Who paid for the reassurance

The safety case for aspartame rests on a literature, and much of that literature was commissioned. The review most often cited for the statement that aspartame is safe is a 98-page evaluation in Critical Reviews in Toxicology in 2007, which concluded that "critical review of all carcinogenicity studies conducted on aspartame found no credible evidence that aspartame is carcinogenic" and that "the weight of existing evidence is that aspartame is safe". The paper's funding line is behind a paywall, so we did not read it. We did not need to. The following year two of its authors wrote to Environmental Health Perspectives and declared: "The authors received payment from the Burdock Group during the preparation of an expert review of the safety of aspartame. The Burdock Group managed the independent review, which was financially supported by Ajinomoto Company Inc., a producer of aspartame." The corresponding author's affiliation was the Burdock Group, a food-industry consultancy. The lead scientist of the rat studies replied in the same issue that it was "a 'safety evaluation' sponsored entirely by Ajinomoto, the manufacturer of aspartame".²⁰²¹²²

The table above is built from each paper's own funding statement or, where that was unreadable, its authors' later declaration; the 2002 review was written from inside the NutraSweet Company, and the one publicly funded British review of sweeteners and weight, in 2021, reached a favourable conclusion too, which is a reminder that sponsorship predicts conclusions without dictating them. The names recur. The International Sweeteners Association, registered in Brussels and "funded through annual membership fee contributions", lists among its members Ajinomoto Europe, Cargill, Ingredion, Mars, PepsiCo, Coca-Cola Services and Tate and Lyle. The Calorie Control Council, in Washington since 1966, displays the Coca-Cola Company, PepsiCo, Keurig Dr Pepper, Tate and Lyle, Cargill and ADM. Between them they funded the 2018 intake review, the 2019 epidemiology review that found the human studies "do not support" a cancer link, the 2020 consensus workshop at which British academics from Bristol, Reading, Liverpool and Aberdeen sat alongside the British Nutrition Foundation, and a fund at the University of Toronto from which a favourable 2022 meta-analysis flowed. When the cancer agency ruled, the Calorie Control Council's president said the WHO's own committee had validated decades of research and that a person would need "14 12-oz cans of diet beverages or 74 packets of aspartame sweetener every day" to reach a level of concern.^232425273132

The third body is ILSI, the International Life Sciences Institute, founded by a Coca-Cola executive in 1978. A 2020 study of its internal documents, obtained under US freedom of information law by a campaign group and analysed by academics at Cambridge and Bocconi, reproduced its tax returns: $337,000 from Coca-Cola in 2013, $317,827 from PepsiCo to its North American arm in 2016, and six-figure sums from Mars, Mondelez, General Mills, Nestlé, Kellogg, Cargill and Unilever. The authors concluded that "ILSI's funding by corporate entities leads to industry influence over some of ILSI activities". ILSI Europe's task force paid for the 2016 Bristol review of sweeteners and weight. The campaign group's own donors are listed in the paper, which is more than most of the reviews in the table manage.²⁶³⁰

Does the money change the answer? The one systematic attempt to measure it looked at 31 reviews of artificially sweetened drinks and weight. The four sponsored by the sweetener industry reported favourable results in three cases; the 23 with no industry sponsor did so in one. That is a relative risk of 17, with a confidence interval so wide it should be read as "much more likely" rather than "seventeen times". Every review funded by the sugar industry reached the opposite conclusion. In 42% of the reviews the authors' financial conflicts were not disclosed, and authors with a conflict, disclosed or not, were more likely to conclude favourably. The lead author of that count works at the Ramazzini Institute, which is worth knowing; the pattern it found had already been found, in the other direction, for sugar: reviews with industry ties were five times more likely to find no link between sugary drinks and weight gain. And there is a documented precedent for the whole method. In 1965 the Sugar Research Foundation paid for a New England Journal of Medicine review that blamed fat rather than sugar for heart disease; it "set the review's objective, contributed articles for inclusion, and received drafts", and "the SRF's funding and role was not disclosed". That was established from the foundation's own archives in 2016.²⁸²⁹

The track record

The claim that the original approval was bought is the one people most want confirmed, and the public record both supports and limits it. Aspartame was discovered by G.D. Searle in 1965 and approved for dry foods in July 1974. In December 1975 the approval was suspended because a Food and Drug Administration task force had found, in the words of the 1987 audit by the US General Accounting Office, "serious deficiencies in Searle's operations and practices which undermine the basis for reliance on Searle's integrity in conducting high quality animal research". Technicians "did not keep accurate and consistent reports on the animals' condition"; protocols "were not followed"; on two of Searle's drugs, "certain types of tumors noted on raw data entries … were unaccountably changed". The task force said the problems "were so serious in some studies as to make it difficult, if not impossible, to draw conclusions regarding the full toxic potential of the products from the data", found Searle in violation of the food and drug law, and recommended a grand jury. Searle was not indicted. The affair produced the good laboratory practice rules that every study since has had to follow.¹⁶

An outside group of pathologists, paid by Searle under a contract stipulating independence, re-checked the key studies and found "few, if any, discrepancies that would significantly affect the studies' results". In 1980 a public board of inquiry of three scientists concluded that the rat data "do not rule out an oncogenic effect of aspartame" on the brain and that it "should not be approved for marketing until further animal testing was conducted". On 18 July 1981 a commissioner appointed three months earlier overturned the board and approved it, finding "a reasonable certainty that the ingestion of aspartame does not induce brain neoplasms (tumors) in the rat". Three of the five scientists on his own advisory panel had not been convinced. The GAO's conclusion in 1987 was that the FDA "adequately followed its food additive approval process", that the panel "felt no pressure to recommend approval or disapproval", and that "we found no evidence in FDA's files that anyone other than FDA or the Commissioner's panel was involved in the aspartame decision". It added the limit that should always be quoted with it: "We did not evaluate the scientific issues raised … nor did we determine aspartame's safety."¹⁶

The revolving door is real and it is on the record. In 1986 the GAO examined six former health department officials who had been involved with aspartame. The commissioner who approved it had, since November 1983, been "a Senior Scientific Consultant with Burson-Marsteller", which the report describes as "a public relations firm employed by Searle", recruited by another former FDA official who was working the Searle account. The GAO's finding was narrow: the officials "had no financial interest in Searle while employed by HHS", their later contacts with the FDA were "permissible under the postemployment statute", and "according to documentation provided by Burson-Marsteller, Dr. Hayes has not consulted with Searle". Another had gone to the soft drinks trade association, another to the grocery manufacturers, another to a law firm advising Searle. The FDA commissioner of the day told the Senate in 1987 that "there was not any actual conflict of interest". The senator who called the hearing said he did not have confidence that the product was "safe beyond doubt". The name most often attached to the story, Searle's chief executive Donald Rumsfeld, does not appear in either audit; the claim that he arranged the approval is a campaigners' account, not a finding. That is the honest shape of it: documented bad data, a documented walk from regulator to the regulated's PR firm, and no documented payment.¹⁷¹⁸

The other side's money

An investigation into funding has to look both ways. The studies that keep the cancer question alive come from one place: the Ramazzini Institute in Bologna, whose 2006 experiment fed aspartame to 1,800 rats until they died of old age and reported lymphomas and leukaemias rising with dose, "even at a daily dose of 20 mg/kg body weight, much less than the current acceptable daily intake". The institute is a cooperative with 37,138 members at the end of 2024, funded by their subscriptions, donations, legacies and an Italian tax allocation, and it says this guarantees "the independence of the Institute from economic power". Its scientists have declared no competing interests. What they have is a position. The 2021 paper declaring "new evidence for causation" was written by two members of the institute's scientific advisory board; the 2025 paper accusing IARC of "irregularities" and calling its aspartame evaluation "an unfortunate aberration that may have been guided by factors other than dispassionate scientific evaluation" was written by eight of them, including a former head of the IARC monographs programme, and opens by saying so. The rebuttal to the 2021 paper was written by a consultant whose firm "received financial support from the Calorie Control Council" and who declared "no financial conflicts of interest" in the same letter.^1012131533

On the science, the regulators have been consistent for twenty years and IARC has been more generous than they were. EFSA in 2006 found the rats had a "high background incidence of chronic inflammatory changes in the lungs" and dismissed the lymphomas as "an incidental finding"; the FDA in 2007 said the institute had refused to supply full data or slides; JECFA in 2023 called the studies "of uncertain relevance" because the animals were allowed to die naturally, when cancer rises anyway. The institute re-examined its own slides in 2020 and confirmed 72 of 78 diagnoses. The fairest published assessment comes from scientists at the US Environmental Protection Agency in 2013, who found that an international pathology review agreed with the institute "for most tumor types" but not "for lymphoma/leukemia of the respiratory tract", exactly the finding the aspartame case rests on. IARC's 2024 monograph split the difference: the studies did show more malignant tumours in two species, but with "serious questions about the adequacy of the design, conduct, interpretation, and reporting", so the animal evidence was "limited"; a minority thought it sufficient. Nobody on either side has had a paper retracted. The only corrections we found on the critical side were two notices to a 2017 review that had labelled the phenylalanine content of pure aspartame as if it were that of a diet drink.^3461419

What Britain did in 2018, and what it did not measure

The levy worked as a tax is supposed to. By 2020 the sales-weighted sugar content of drinks in its scope had fallen from 3.8g to 2.1g per 100ml, a 46% cut; the share of sales below the 5g threshold rose from 66% to 89%; sales of drinks in the higher band fell 57% while sales of drinks with no levy rose 66%. Irn-Bru halved its sugar in January 2018 "due to the introduction of low calorie sweeteners, including aspartame", and 27,000 people signed a petition to stop it. Public Health England's guidance to industry the year before had said sweeteners approved by EFSA "are a safe and acceptable alternative to using sugar and it is up to businesses if and how they wish to use them", while noting there "may be advantages in businesses not adding sweeteners" so that palates could adjust. Businesses added them. The one peer-reviewed measurement, from a large supermarket's shelves between 2017 and 2020, found a step change in the proportion of drinks containing sweeteners at the moment the levy came in. The trade body's 2026 report says 71.4% of soft drinks sold in Britain last year were low or no calorie, a record.^4043444748

Children drank the sugar-free versions instead. The National Diet and Nutrition Survey records the fall in sugar-sweetened drinks with precision: four-to-ten-year-olds went from 130 grams a day in 2008 to 52 by 2019 and 54 in the latest survey; teenagers from 275 to 124. It records no-added-sugar soft drinks too: in 2019 to 2023, 60.5% of four-to-ten-year-olds drank them, at an average of 332 grams a day among those who did, and a fifth of the age group drank sugar-sweetened ones. What the survey cannot record is what is in them. It measures food groups, not additives, and no sweetener appears anywhere in its tables. The government's own nutrition committee said so in April 2025, in a sentence that should be the starting point of any discussion of the levy: "It is unclear if NSS intakes have increased since introducing this policy, because the last survey of intakes (among children) was carried out in 2003. The UK National Diet and Nutrition Survey is not currently designed to assess intakes of NSS." It recommended that government "monitors the NSS content of food and drinks in the UK diet and their consumption" and "compels industry to make publicly available data on the amounts of individual NSS within foods". Twenty-three years after the last measurement and seven after the tax, that is a recommendation, not a plan.³⁵⁴¹⁴²

The same committee's advice on sweeteners is precautionary and it excludes the question this piece is about. It "recommends that intake of NSS be minimised", that younger children be given neither sugar-sweetened nor sweetener-sweetened drinks, and that for everyone "the long-term goal is to limit both sugar and NSS intake". Its reasons are the observational links to diabetes and heart disease that the WHO weighed in its May 2023 guideline, which it says should be "treated with caution". On cancer it says, in terms: "SACN has not reviewed the wider health issues considered by the WHO, such as some cancers". No UK body has. The Committee on Toxicity, which advises the Food Standards Agency, has published nothing on the 2023 classification. The FSA's consumer page on additives was published on 9 January 2018 and does not mention it; the agency's statement on the day, given through the Science Media Centre, was that "there are no plans to carry out a new risk assessment on aspartame". The NHS page on sweeteners was last reviewed on 20 February 2023, before both WHO decisions, is past its own review date, and answers the cancer question by deferring to a charity: "Cancer Research UK has said sweeteners do not cause cancer."^35364546

The chain that does not hold

Cancer is more common in Britain than it was, and liver cancer, the one the WHO agency named, is much more common: incidence rates have risen 184% since the early 1990s and 24% in the last decade, against 15% and 2% for all cancers combined. It would be easy to lay that beside the rise of the diet drink and call it a link. It is not one. Cancer Research UK attributes 23% of UK liver cancers to overweight and obesity, 20% to smoking, 10% to hepatitis B and C and 7% to alcohol, and its list of causes does not include sweeteners at all. Almost half of cases are in people over 75, whose diets were set long before 2018. The WHO's own guideline on sweeteners, published two months before the cancer ruling and based on a systematic review of the long-term studies, found that higher intake "was not associated with differences in overall cancer incidence or mortality". We searched three databases for any study attempting to connect sweetener consumption to British cancer trends and found none. The cohort studies that exist in the UK Biobank report no association for breast cancer and none for endometrial cancer.^34373839

So the sentence people want, that the sugar tax pushed children onto a sweetener and the sweetener is giving them cancer, cannot be written from the evidence. The children who drank the most diet drinks after 2018 are not old enough to appear in any cancer register for decades, the cancer in question is one the French study could not even examine, and the exposure the WHO's own committee estimates for high-consuming children is half the limit. What can be written is narrower and, in its way, harder to dismiss. Britain changed what its children drink on a national scale, on the strength of a safety literature that its manufacturers and trade bodies paid for a large part of, defended by regulators one of whom has a documented history of accepting the manufacturer's bad data, and then declined to measure the result. The independent evidence that exists, from a French cohort and an Italian laboratory, is disputed on its merits by people who are paid to dispute it and by people who are not. The right response to that is not fear. It is the survey that the government's own advisers asked for in 2025 and nobody has commissioned.

What this cannot tell you

It cannot tell you whether aspartame causes cancer. Nobody can; that is what "limited evidence" means, and it is why the WHO asked for "better studies with longer follow-up". It cannot tell you what UK children's sweetener intake is, because no survey has measured it since 2003; the levels quoted here are the WHO committee's international estimates. It could not read the funding line of the 2007 review itself, only the authors' later declaration of it, or the 2023 Reuters reporting on industry lobbying of the WHO, which was inaccessible, or the sweetener association's own declared lobbying spend. It cannot say what any company paid the officials it later employed, because the audits found no payment and we found nothing the audits did not. And it treats every attributable-cause figure and every hazard ratio as its authors published it, with the confidence intervals that make most of them wider than the headlines. What it can say is that the money is documented, the misconduct is documented, the cancer link is not, and the measurement that would settle the British question has not been done.