Aspartame carries a cancer hazard classification from the International Agency for Research on Cancer (IARC), part of the World Health Organization. It also carries a safety verdict, from JECFA, the WHO and FAO's own food-safety risk committee, that hasn't moved since before that classification existed. Both are true. The distance between them is where most of the anxiety about diet soda actually comes from.
The short version: at the levels people actually consume it, the agencies that assess real-world safety — the FDA, JECFA, and the European Food Safety Authority (EFSA) — still classify aspartame as safe. What changed in 2023 was a separate, narrower finding: whether aspartame could, in principle, contribute to cancer under some circumstance, based on limited evidence. That's a legitimate scientific question, and it isn't the same question as "is my can of diet soda a health risk." This article covers both, along with two genuinely new 2026 developments worth knowing about: a regulatory finding on heating sucralose, and how stevia's evidence picture differs from the other two.
If you're here because of a headline along the lines of "WHO says aspartame causes cancer," that headline oversimplified the finding. Nothing below asks you to panic, and nothing below asks you to dismiss a real regulatory finding either.
The short answer: hazard vs. risk (this is the whole story)
Almost everything confusing about sweetener safety news comes down to two different questions getting mixed together: could this substance cause harm under some circumstance (hazard), and does it actually cause harm at the levels people are exposed to in real life (risk)? IARC only answers the first question. JECFA, the FDA, and EFSA answer the second.
That distinction isn't unique to aspartame. IARC's highest-confidence category, Group 1 ("carcinogenic to humans"), includes processed meat and alcoholic beverages, evaluated as definite causes of cancer, alongside tobacco (WHO). Aspartame sits three tiers below that, in Group 2B, the category IARC uses for limited or inconsistent evidence. A useful rule for reading any hazard headline: it tells you whether a substance could plausibly cause harm somewhere, to someone, under some level of exposure. It does not tell you how likely that harm is for you, at the amount you actually consume.
What IARC's 2023 "possible carcinogen" (Group 2B) classification actually means
On 14 July 2023, IARC classified aspartame as possibly carcinogenic to humans (Group 2B), citing "limited evidence" for cancer in humans (WHO, 2023). Group 2B is used when there's limited but not convincing evidence of cancer in humans, or convincing evidence in animals but not both — the third of four IARC categories, ranked below Group 1 ("carcinogenic") and Group 2A ("probably carcinogenic").
The evidence behind the classification was specific and limited in scope. IARC's working group of 25 independent experts reviewed three studies, covering four large cohorts including the U.S. NIH-AARP and Cancer Prevention Study-II cohorts, that used artificially sweetened beverage consumption as a proxy for aspartame exposure. Each found a positive association with liver cancer (specifically hepatocellular carcinoma), but the same review noted that chance, bias, or confounding could not be ruled out as the real explanation (IARC/WHO summary, 2023). Animal evidence was similarly rated limited, after the working group raised specific concerns about the design of the positive studies available. Mechanistic evidence, whether aspartame does something in cells consistent with how carcinogens behave, was also rated limited, based mainly on evidence that aspartame can trigger oxidative stress in lab systems.
"Limited" is a specific technical grade here, not IARC being coy. It means the evidence pointed somewhere, wasn't strong enough to be convincing, and needs more research, which is exactly why IARC and JECFA both called for better-designed cohort studies and randomized trials going forward.
What JECFA's risk assessment says — and why the acceptable daily intake didn't change
JECFA, the Joint FAO/WHO Expert Committee on Food Additives, ran its own, separate evaluation over the same period and reached a different kind of conclusion, because it was answering a different question. After reviewing animal and human data, including the same studies IARC looked at, JECFA concluded that the evidence of an association between aspartame and cancer in humans "is not convincing," and found no mechanism by which oral aspartame exposure could plausibly cause cancer (WHO, 2023). It reaffirmed the acceptable daily intake (ADI) of 0–40 mg per kilogram of body weight, the same figure it had set in earlier reviews.
An ADI is the amount of a substance a person can consume every day, for life, without an appreciable health risk, built with a wide safety margin. WHO's own illustration is useful here: at 40 mg/kg, a 70 kg (154 lb) adult would need to drink more than 9 to 14 cans of diet soda a day, every day, for life, to exceed it, assuming zero aspartame from any other source (WHO, 2023). Actual average intake is far below that for most people. That's the number that didn't move in 2023, and it's the reason regulators kept saying aspartame remains safe at normal levels of use even while IARC's hazard finding was making headlines.
Aspartame specifically: the current safety verdict, in plain language
Pulling the two threads together: aspartame is approved for use in food in the U.S., the EU, and most other countries, and the agencies responsible for judging real-world safety haven't changed that position. The FDA has reviewed more than 100 studies on aspartame and maintains that it's safe for the general population under approved conditions of use, at its own acceptable daily intake of 50 mg/kg body weight per day — slightly higher than JECFA and EFSA's more commonly cited 40 mg/kg figure, though both describe the same basic safety margin, and both sit well above typical consumption (FDA). The one group that genuinely needs to avoid aspartame is people with phenylketonuria (PKU), a rare inherited condition that prevents the body from properly breaking down phenylalanine, one of aspartame's three metabolic byproducts; products containing aspartame are required to carry a PKU warning for exactly this reason.
The FDA has also stated directly that it disagrees with IARC's classification, citing what it calls significant shortcomings in the studies IARC relied on (FDA). That disagreement between a hazard-identification body and a risk-assessment body isn't unusual or alarming on its own; it reflects the different jobs the two are doing, as covered above.
Beyond IARC and JECFA's 2023 reviews, a handful of other studies are worth knowing about, because together they show a genuinely mixed picture rather than a one-sided one:
- The French NutriNet-Santé cohort, which tracked artificial sweetener intake from every dietary source (not beverages alone) in 102,865 adults over a median 7.8 years, found that people who consumed more aspartame had a higher rate of cancer overall (hazard ratio 1.15), and specifically breast cancer and obesity-related cancers, compared with non-consumers (Debras et al., 2022, PLOS Medicine). This is a large, well-conducted study, and its own authors are explicit that a cohort design like this cannot establish causation: people who consume more artificial sweeteners also tend to differ from non-consumers in other ways researchers can't fully account for, and reverse causation is a possibility the authors specifically tested for and could not fully rule out.
- A separate U.S. cohort of nearly 99,000 postmenopausal women, the Women's Health Initiative, found no significant association between artificially sweetened beverages and liver cancer or chronic liver disease mortality, in contrast to sugar-sweetened beverages, which did show a significant association in the same study (Zhao et al., 2023, JAMA).
- An umbrella review pooling multiple meta-analyses found the evidence linking artificially sweetened beverages to cancer outcomes specifically (colorectal, pancreatic, and other cancers, plus cancer mortality) was weak, while the evidence linking them to metabolic outcomes, obesity, type 2 diabetes, and cardiovascular disease, was rated highly suggestive (Diaz et al., 2023, Advances in Nutrition).
Put together, that's not a picture of settled reassurance, and it's not a picture of proven harm either. It's an honest "mixed, with the strongest specific finding pointing toward metabolic effects rather than cancer." None of it has moved the ADI, and none of the major regulators have withdrawn approval.
Sucralose and the 2026 heating question
Sucralose (Splenda) is chemically distinct from aspartame, made by replacing three hydroxyl groups on a sucrose molecule with chlorine atoms, and until 2026 its regulatory story was comparatively quiet: no cancer-classification controversy, a stable ADI, and general use in baked goods on the strength of being marketed as "heat stable." That last point is exactly what a new, genuinely dated 2026 development complicates.
BfR's official notification on sucralose above 120°C
On 26 February 2026, Germany's Federal Institute for Risk Assessment (BfR) issued an official communication restating a finding it first raised in 2019: heating sucralose above 120°C (248°F) for a sustained period causes it to gradually break down and lose chlorine atoms, which can form chlorinated organic compounds, including polychlorinated dibenzo-p-dioxins (PCDDs), dibenzofurans (PCDFs), and chloropropanols, some with harmful or carcinogenic potential (BfR, 2026). Temperatures between 120°C and 150°C are common in both industrial food processing and ordinary home cooking, such as baking, frying, and roasting, of foods that contain sucralose.
BfR's recommendation is specific and practical, not a warning to avoid sucralose altogether: don't heat sucralose-containing food above 120°C, or add sucralose only after cooking rather than before (BfR, 2026).
What EFSA's February 2026 re-evaluation adds
Five days before BfR's communication, on 17 February 2026, the European Food Safety Authority published its own scheduled re-evaluation of sucralose, and its findings are more layered than a single headline can capture (EFSA Journal, 2026).
On the core safety question, EFSA's Panel on Food Additives and Flavourings found no need to change sucralose's acceptable daily intake of 15 mg/kg body weight per day, and concluded there's no safety concern at its currently authorized uses; current exposure estimates, even at the high end, stay under that limit.
On heating specifically, the picture is more uncertain than settled. EFSA reviewed older, deliberately extreme laboratory studies (sucralose heated to 120–400°C in unrealistic amounts and conditions, which did form chlorinated compounds) alongside newer studies using realistic baking conditions, for example biscuits baked 12 minutes at 180°C and wafers baked 75 seconds at 160°C, the kind of process an industrial bakery would actually use. In those realistic tests, sucralose did not cause a statistically significant increase in the chlorinated compounds tested for. But a 2024 study found sucralose becomes measurably unstable starting around 85–90°C in dry conditions, and identified new chlorinated degradation products in baked goods at 220°C (Hellwig, 2024, Journal of Agricultural and Food Chemistry), a finding EFSA explicitly weighed and could not fully resolve with the data available.
Because of that residual uncertainty, EFSA declined the specific request in front of it: extending sucralose's authorized use into more categories of "fine bakery wares." Its stated reasoning was direct: home cooking varies far more in temperature, time, and technique than standardized industrial baking, so the same risk of chlorination that looked negligible under controlled industrial conditions "cannot be excluded" in home kitchens (EFSA Journal, 2026). That's the specific, bounded finding worth taking seriously: not that sucralose is newly proven dangerous, but that regulators still can't rule out a heating-related risk in home cooking, and are asking for more data before expanding where it's used in baking.
Practical takeaway — baking/heating vs. cold use
For the way most people actually use sucralose, in coffee, cold drinks, yogurt, or sprinkled on food after it's cooked, none of this changes anything; that use isn't in question. The caution applies specifically to prolonged heating above 120°C: baking, frying, or roasting sucralose-containing food, or using sucralose as a sugar substitute in a recipe that goes into a hot oven. If you bake with sucralose regularly, the straightforward option both BfR and EFSA point toward is adding it after cooking rather than before, or choosing a different sweetener for high-heat recipes until the remaining uncertainty is resolved.
Stevia: a different profile, not lumped in with the cancer conversation
Stevia doesn't belong in the same conversation as aspartame and sucralose's cancer-adjacent headlines, because it has a genuinely different evidence history. The sweetening compounds extracted and purified from the stevia plant, steviol glycosides, the E960 group of food additives, haven't been linked to cancer in any regulatory review, and none of the 2023 or 2026 developments covered above apply to them.
Steviol glycosides carry their own ADI, set by JECFA and maintained by EFSA at 4 mg/kg body weight per day, expressed as steviol equivalents (FDA; EFSA Journal, 2024). EFSA's most recent review of that limit, in 2024, actually went the other direction from the sucralose story: a manufacturer asked EFSA to raise the ADI, and EFSA said no, the data didn't justify a higher limit. That review flagged one genuine nuance worth knowing: at current permitted uses, toddlers' estimated intake at the high end already sits close to the ADI, a margin issue for very young children specifically, not a finding that stevia is unsafe for the general population.
There's an important labeling distinction worth understanding, too. What's FDA-approved and widely sold as "stevia" in the U.S. is purified, high-purity steviol glycosides (often labeled Reb A, Rebaudioside A, or similar), not whole-leaf stevia or crude stevia extract, which the FDA has explicitly said is not generally recognized as safe and cannot legally be imported for use as a sweetener (FDA). If a product's label just says "stevia," it's almost certainly referring to the purified, regulated form.
Stevia also doesn't share sucralose's heating question, and unlike aspartame, it doesn't lose its sweetness when heated. None of that makes stevia a "winner"; it just means the specific 2023 and 2026 concerns covered in this article are genuinely aspartame- and sucralose-specific, not generic to all non-sugar sweeteners.
So, which sweetener should you use? A calm framework, not a verdict
There isn't a single "safest" sweetener this article can honestly hand you, because the three substances above have different evidence pictures, different open questions, and different practical considerations. A framework is more useful than a verdict:
- If you have phenylketonuria, aspartame isn't an option regardless of anything else in this piece. That's a firm medical restriction, not a risk-tolerance question.
- If you bake or cook with sucralose at high heat regularly, the current EFSA and BfR position is to add it after cooking, or use it only in cold or low-heat applications, until the open heating question is resolved.
- If you're specifically concerned about the cancer-adjacent findings above, stevia (specifically, purified steviol glycosides) currently has the cleanest evidence picture of the three, though it isn't unlimited: its own ADI exists for a reason, and very young children shouldn't be treated as having unlimited headroom under it.
- If your goal in using any of these is weight control or reducing diabetes or heart disease risk, it's worth knowing that WHO's own 2023 guideline recommends against relying on any non-sugar sweetener for that purpose specifically. Its review of the evidence found no long-term benefit for weight control, and flagged a possible association between long-term use and increased risk of type 2 diabetes and cardiovascular disease (WHO, 2023). That recommendation applies to aspartame, sucralose, and stevia alike.
Sweeteners are common in ultra-processed foods generally, and if you're trying to eat fewer of those for reasons beyond sweeteners specifically, additive load, low fiber, easy overconsumption, that broader shift will do more for your health than swapping which sweetener is in your coffee. None of the evidence above is a reason to panic about a diet soda you already drank. It's also not a reason to stop paying attention to genuinely new findings, like the 2026 sucralose update, as they arrive.
Frequently Asked Questions (FAQ)
Does aspartame cause cancer?
The evidence doesn't show that it does at normal intake. IARC's 2023 classification (Group 2B, "possible" carcinogen) was based on limited evidence for liver cancer in a handful of human studies. JECFA, the WHO and FAO's risk-assessment body, reviewed the same evidence, called it "not convincing," and didn't change the acceptable daily intake. Cohort studies since then are mixed: one French cohort found a modest association with cancer overall; a separate U.S. cohort found none for liver cancer specifically. No study has shown aspartame causes cancer in humans at normal consumption levels.
What is the difference between a "possible carcinogen" classification and an actual health risk?
A hazard classification (IARC's job) asks whether a substance could cause cancer under any circumstance, to anyone, at any exposure. A risk assessment (JECFA's, the FDA's, and EFSA's job) asks how likely that harm actually is at real-world exposure. Processed meat and alcohol both sit in IARC's highest-confidence category, Group 1, alongside tobacco, but nobody treats an occasional hot dog like a cigarette, because actual exposure matters. Aspartame's Group 2B rating means the evidence is limited and inconclusive, not that typical consumption is dangerous.
Is it safe to heat sucralose, like in baking?
No acute danger has been shown from occasional baking, but the question isn't fully closed either. Heating sucralose above 120°C for a sustained period can form chlorinated compounds, some with harmful potential, per Germany's BfR. EFSA's February 2026 re-evaluation found realistic industrial baking tests showed no significant increase, but flagged that home cooking's less controlled temperatures make the risk harder to rule out, and declined to expand sucralose's approved baking uses on that basis. Practical guidance from both agencies: avoid heating sucralose-containing food above 120°C, or add it after cooking.
Is stevia a safer choice than aspartame or sucralose?
Of the three, purified steviol glycosides (sold as "stevia" in the U.S.) currently have the cleanest evidence record: no cancer link in any regulatory review, and no role in either the 2023 IARC finding or the 2026 sucralose question. That doesn't mean unlimited: it has its own acceptable daily intake, and a 2024 EFSA review found young children's typical intake already sits close to that limit, a reason for moderation in that group, not a general safety concern.
What does the FDA say about aspartame's acceptable daily intake?
The FDA's own acceptable daily intake for aspartame is 50 mg per kilogram of body weight per day, a bit higher than the 40 mg/kg figure most commonly cited internationally by JECFA and the European Food Safety Authority, though both describe the same basic margin. At JECFA's 40 mg/kg figure, WHO calculates a 70 kg adult would need more than 9 to 14 cans of diet soda a day to exceed it. The FDA's chart puts its 50 mg/kg limit at around 75 sweetener packets a day for a 60 kg person. Either way, typical intake sits far below both.
Is there any other research linking artificial sweeteners to cancer risk?
Yes, and it's genuinely mixed. The French NutriNet-Santé cohort (102,865 adults) found people who consumed more aspartame and acesulfame-K had a modestly higher rate of cancer overall, breast cancer, and obesity-related cancers, but as an observational cohort it shows association, not causation. A separate U.S. cohort of postmenopausal women found no link between artificially sweetened beverages and liver cancer. A 2023 umbrella review found the case for cancer outcomes specifically was weak, while metabolic outcomes like obesity and type 2 diabetes had stronger evidence. Nothing here is definitive, and none of it has changed a regulator's position.
Which sweetener is best if I want to avoid all of these concerns?
No sweetener here can be honestly called risk-free; each has its own open question: aspartame's hazard classification, sucralose's heating question, stevia's smaller pediatric margin. If avoiding all of it matters, the most evidence-backed move isn't switching sweeteners; it's WHO's broader advice to gradually reduce how much sweet taste you rely on overall, rather than treating any sweetener as a permanent substitute.
Has anything changed about sweetener safety guidance in 2026?
For sucralose, yes, in a bounded way: EFSA's February 2026 re-evaluation kept the ADI unchanged but declined to approve sucralose in more baked-good categories, citing unresolved heating uncertainty in home cooking, reinforcing BfR's 120°C caution, restated in its own February 2026 communication. For aspartame and stevia, core guidance hasn't moved since 2023 and 2024: both remain within existing limits, with no new cancer finding attached in 2026.
Sources
- World Health Organization. Aspartame hazard and risk assessment results released. 14 July 2023. https://www.who.int/news/item/14-07-2023-aspartame-hazard-and-risk-assessment-results-released
- International Agency for Research on Cancer / WHO. Summary of findings of the evaluation of aspartame at the IARC Monographs Programme's 134th Meeting and JECFA's 96th Meeting. July 2023. https://www.iarc.who.int/wp-content/uploads/2023/07/Summary_of_findings_Aspartame.pdf
- Riboli E, Beland FA, Lachenmeier DW, et al. Carcinogenicity of aspartame, methyleugenol, and isoeugenol. The Lancet Oncology. 2023;24(8):848-850. PMID: 37454664. https://pubmed.ncbi.nlm.nih.gov/37454664/
- U.S. Food and Drug Administration. Aspartame and Other Sweeteners in Food. https://www.fda.gov/food/food-additives-petitions/aspartame-and-other-sweeteners-food
- National Cancer Institute. Artificial Sweeteners and Cancer. Reviewed August 29, 2023. https://www.cancer.gov/about-cancer/causes-prevention/risk/diet/artificial-sweeteners-fact-sheet
- Debras C, Chazelas E, Srour B, et al. Artificial sweeteners and cancer risk: Results from the NutriNet-Santé population-based cohort study. PLOS Medicine. 2022;19(3):e1003950. PMID: 35324894. https://pubmed.ncbi.nlm.nih.gov/35324894/
- Zhao L, Zhang X, Coday M, et al. Sugar- and Artificially-Sweetened Beverages and Risk of Liver Cancer and Chronic Liver Disease Mortality. JAMA. 2023;330(6):537-546. PMID: 37552302. https://pubmed.ncbi.nlm.nih.gov/37552302/
- Diaz C, Rezende LFM, Sabag A, et al. Artificially Sweetened Beverages and Health Outcomes: An Umbrella Review. Advances in Nutrition. 2023;14(4):710-717. PMID: 37187453. https://pubmed.ncbi.nlm.nih.gov/37187453/
- German Federal Institute for Risk Assessment (BfR). Sucralose – Heating above 120°C can produce substances that are harmful to health. Communication No. 015/2026. 26 February 2026. https://www.bfr.bund.de/en/notification/sucralose-heating-above-120c-can-produce-substances-that-are-harmful-to-health/
- EFSA Panel on Food Additives and Flavourings (FAF). Re-evaluation of sucralose (E 955) as a food additive and evaluation of a new application on extension of use of sucralose (E 955) in fine bakery wares. EFSA Journal. 2026;24(2):e9854. doi: 10.2903/j.efsa.2026.9854. Full text: https://pmc.ncbi.nlm.nih.gov/articles/PMC12911467/
- Hellwig M. Formation of Chlorinated Carbohydrate Degradation Products and Amino Acids during Heating of Sucralose in Model Systems and Food. Journal of Agricultural and Food Chemistry. 2024;72(47):26441-26450. PMID: 39556422. https://pubmed.ncbi.nlm.nih.gov/39556422/
- EFSA Panel on Food Additives and Flavourings (FAF). Scientific opinion on the extension of the authorisation of use of the food additive steviol glycosides (E 960a–d) and the modification of the acceptable daily intake (ADI) for steviol. EFSA Journal. 2024;22(11):e9045. PMID: 39502654. https://pubmed.ncbi.nlm.nih.gov/39502654/
- World Health Organization. WHO advises not to use non-sugar sweeteners for weight control in newly released guideline. 15 May 2023. https://www.who.int/news/item/15-05-2023-who-advises-not-to-use-non-sugar-sweeteners-for-weight-control-in-newly-released-guideline
- World Health Organization. Cancer: Carcinogenicity of the consumption of red meat and processed meat. https://www.who.int/news-room/questions-and-answers/item/cancer-carcinogenicity-of-the-consumption-of-red-meat-and-processed-meat
Olivia Smith, RDN
A registered dietitian who separates a hazard classification from an actual health risk before calling anything dangerous.
