Does aspartame cause cancer?

Not at the amounts people actually consume. In 2023 two WHO bodies published on the same day: IARC classified aspartame as Group 2B, possibly carcinogenic, while JECFA reviewed the same evidence and reaffirmed the acceptable daily intake of 40mg per kilogram of body weight. These were widely reported as contradictory and are not, because they answer different questions. IARC identifies hazards — whether something can cause cancer under any conditions — and Group 2B is a low-confidence category that also contains aloe vera extract and pickled vegetables. JECFA assesses risk at realistic exposures. For a 70kg adult, exceeding the acceptable daily intake would require somewhere around nine to fourteen cans of diet soft drink every day. The evidence for harm at ordinary consumption is weak.

Did WHO say sweeteners are dangerous?

No, and this is probably the most consistently misreported nutrition guideline of recent years. WHO issued a conditional recommendation in 2023 against using non-sugar sweeteners as a means of controlling weight or reducing the risk of noncommunicable disease. The basis was a systematic review finding no evidence of long-term benefit for body fat in adults or children, together with observational associations of uncertain causality. That is a statement about lack of demonstrated benefit, not about demonstrated harm — and the guideline explicitly does not apply to people with diabetes, and is graded as conditional based on low-certainty evidence. The distinction between "has not been shown to help" and "has been shown to hurt" is the entire substance of the document, and it was lost almost immediately in coverage.

Is sugar or are sweeteners worse?

Sugar, on the current evidence, and it is not particularly close. The harms of free sugars are settled: dental caries with a clear dose-response, contribution to excess energy intake particularly through sugar-sweetened beverages, and the associated metabolic consequences. The harms of sweeteners at realistic intakes are not established despite decades of study and considerable effort to find them. For someone drinking several sugary drinks a day, switching to sweetened versions reduces sugar and calorie intake and is a reasonable step, and is specifically appropriate for people with diabetes. It is not a long-term weight control strategy, since the evidence does not support that. The sensible ranking is water first, sweetened drinks as a transitional improvement, sugar-sweetened drinks last.

Do sugar taxes actually work?

Yes, on purchasing, and the UK example shows something more interesting. Every jurisdiction that has introduced and properly evaluated a sugar-sweetened beverage tax has found reduced purchasing of the taxed products. The UK soft drinks industry levy went further by using a tiered structure with thresholds, which gave manufacturers a strong incentive to reformulate rather than pass on the cost — and they did, extensively, so the sugar content of the drinks supply fell without consumers having to change what they bought. That design is the transferable lesson: a tax that changes the product environment achieves more than one that relies on individual price sensitivity. The usual objection is that such taxes are regressive, which is partly answered by the fact that both consumption and the associated disease burden are higher in lower-income groups, so the benefits are also concentrated there.

What about erythritol and other newer sweeteners?

Erythritol is the one with a genuine open question attached. A 2023 study reported that higher circulating erythritol was associated with cardiovascular events across several cohorts, supported by laboratory work showing increased platelet aggregation. That deserves to be taken seriously and is under active investigation. It also needs an important qualification: the body produces erythritol itself, and production increases when metabolic dysfunction is present, so raised blood levels may be a marker of the underlying disease that causes cardiovascular events rather than a cause of them. Similar reverse-causation problems affect most observational sweetener research, since people take up sweeteners precisely because they already have weight or glycaemic problems. The honest summary is an unresolved signal rather than an established harm.

How does GMJ News cover sugar and sweeteners?

GMJ News covers the hazard-versus-risk distinction underlying the 2023 aspartame confusion, the settled evidence on free sugars and dental caries, free versus added versus total sugar definitions, sugar-sweetened beverage tax evidence and the reformulation effect of the tiered UK levy, the widely misreported WHO 2023 non-sugar sweetener guideline and what it actually states, the erythritol cardiovascular signal and its interpretation, reverse causation in sweetener epidemiology, and dental caries as the neglected settled outcome. Part of PHIG's contested science, nutrition policy and oral health programmes.

Home › Topics › Sugar and Artificial Sweeteners

Sugar and Artificial Sweeteners

GMJ News knowledge hub · last reviewed August 2026 · Georgian Medical Journal

Summary Statistics Glossary GMJ news FAQ Documents Organizations Research

The 2023 IARC classification of aspartame as Group 2B possibly carcinogenic, published simultaneously with a JECFA review reaffirming the acceptable daily intake of 40mg per kilogram, generated widespread confusion that was entirely predictable and entirely avoidable: the two bodies were answering different questions — hazard identification versus risk assessment at realistic exposures — and an adult would need to drink well over a dozen cans of diet soft drink daily to approach the ADI (WHO). Meanwhile the evidence on free sugars is settled and largely uncontested: they cause dental caries and contribute to excess energy intake, and sugar-sweetened beverage taxes reduce purchasing in every jurisdiction that has evaluated them. The genuinely open question is the one that receives least attention — WHO's 2023 conditional recommendation against using non-sugar sweeteners for weight control, based on low-certainty evidence of no long-term benefit, which is a statement about the absence of demonstrated benefit rather than about demonstrated harm.

Key messages

THE 2023 ASPARTAME CONFUSION: two agencies, two different questions

IARC classified aspartame as Group 2B, possibly carcinogenic, on the same day that JECFA reaffirmed the acceptable daily intake of 40mg per kilogram of body weight. This was reported as contradiction and was not. IARC performs hazard identification — can this agent cause cancer under any circumstances? JECFA performs risk assessment — does it cause harm at realistic human exposures? Group 2B contains aloe vera extract, pickled vegetables and working as a carpenter. An adult of 70kg would need approximately 9-14 cans of diet soft drink daily to exceed the ADI.

SETTLED: free sugars cause dental caries and contribute to excess energy intake

This is among the least contested findings in nutrition and receives the least attention because it is not controversial. WHO recommends reducing free sugars to below 10% of total energy intake, with a conditional further recommendation below 5% for additional benefit against caries. The relationship with dental caries is dose-dependent and demonstrable, and dental caries remains the most prevalent noncommunicable disease globally. Sugar-sweetened beverages contribute to weight gain in both cohort and trial evidence, largely because liquid calories produce weak compensatory reduction in subsequent intake.

SETTLED: sugar-sweetened beverage taxes reduce purchasing

Every jurisdiction that has implemented and evaluated an SSB tax — Mexico, Chile, the UK soft drinks industry levy, several US cities, South Africa — has found reduced purchasing of taxed products, with the UK levy notably producing substantial reformulation by manufacturers to avoid the higher tier, which reduced sugar in the supply without requiring consumer behaviour change at all. Effects on health outcomes take longer to demonstrate and evidence is accumulating. Regressivity is the standard objection and is partially answered by the fact that consumption and disease burden are both higher in lower-income groups, so benefits are also concentrated there.

THE UNDER-REPORTED FINDING: WHO advises against non-sugar sweeteners for weight control

WHO's 2023 guideline issued a conditional recommendation against using non-sugar sweeteners to achieve weight control or reduce noncommunicable disease risk, based on systematic review finding no long-term benefit for body fat in adults or children, and observational associations with type 2 diabetes and cardiovascular disease of uncertain causality. This is frequently misreported as a safety warning. It is a statement about absence of demonstrated benefit, with low certainty evidence, and it explicitly does not apply to people with diabetes. The distinction between "not shown to help" and "shown to harm" is the entire content of the guideline.

GENUINELY OPEN: sweeteners, the microbiome and cardiometabolic signals

Several mechanistic questions remain live. Some sweeteners alter gut microbiota composition in human studies, with unclear clinical consequence. Erythritol was associated with major adverse cardiovascular events in a 2023 study with supporting platelet aggregation experiments, though erythritol is also produced endogenously and the direction of causation is disputed. Observational associations between sweetener intake and cardiometabolic disease are heavily confounded by reverse causation, since people switch to sweeteners because of existing weight or glycaemic problems. These are real research questions, not settled harms.

WHAT FOLLOWS: sweeteners are better than sugar and worse than water

The practical position is unglamorous and defensible. For a person consuming substantial sugar-sweetened beverages, switching to sweetened alternatives reduces sugar and energy intake and is a reasonable transitional step, and is specifically appropriate in diabetes. It is not a strategy for long-term weight control, for which the evidence does not support it. Water, unsweetened tea and coffee remain the endpoint. The public health priority is reducing free sugar intake, on which the evidence is settled, rather than adjudicating sweetener safety, on which it is not — and the disproportionate attention paid to the latter has served the interests of the former remaining unaddressed.

Key statistics

Group 2B / ADI 40mg/kg

IARC hazard classification and JECFA risk assessment for aspartame, published simultaneously in 2023

WHO 2023

9-14 cans

of diet soft drink daily needed for a 70kg adult to exceed the aspartame ADI

JECFA/WHO

<10% energy

WHO recommendation for free sugars, with conditional further benefit below 5%

WHO

Reduced purchasing

in every evaluated SSB tax jurisdiction; UK levy also drove substantial reformulation

BMJ/Lancet PH

No long-term benefit

WHO 2023 conditional recommendation against non-sugar sweeteners for weight control

WHO 2023

Not a warning

the sweetener guideline concerns absence of benefit, not demonstrated harm — and excludes diabetes

WHO 2023

Sugar and sweeteners — where the disagreement actually lies

Source: Bars show strength of supporting evidence. The settled sugar evidence receives far less attention than the contested sweetener questions.

Glossary of key terms

Hazard identification versus risk assessment
ToxicologyThe distinction that resolves most chemical scares, and the reason the aspartame headlines were misleading. Hazard identification asks whether an agent is capable of causing harm under any conditions, including exposures far beyond realistic human intake — this is what IARC does. Risk assessment asks whether harm is likely at actual human exposure levels, incorporating dose, duration, route and population variability — this is what JECFA, EFSA and national regulators do. The two exercises are complementary and routinely produce apparently opposite headlines from the same evidence. Water is hazardous at sufficient dose; ultraviolet radiation is Group 1; neither observation is actionable without the exposure context. Teaching this distinction once does more for public health literacy than debunking any individual scare.

Free sugars, added sugars and total sugars
DefinitionsFree sugars are all monosaccharides and disaccharides added to foods by manufacturers, cooks or consumers, plus sugars naturally present in honey, syrups and fruit juices. Added sugars are a narrower category excluding honey and juice. Total sugars include those intrinsic to whole fruit, vegetables and milk, which are not targeted by any guideline because they are packaged with fibre, water and micronutrients and are consumed slowly. The distinction matters practically: whole fruit is not restricted by any sugar guideline while fruit juice is, because juicing removes the fibre matrix, increases the rate of ingestion and produces a glycaemic response closer to a soft drink. Product labelling frequently reports total sugars, which obscures the relevant quantity.

The UK soft drinks industry levy
Policy evaluationIntroduced in 2018 with a tiered structure — a higher rate above 8g sugar per 100ml and a lower rate between 5 and 8g — the levy was explicitly designed to incentivise reformulation rather than raise revenue or change consumer behaviour alone. It worked largely as intended: manufacturers reformulated extensively to fall below the thresholds, and the sugar content of the soft drink supply fell substantially without consumers needing to make different choices. Evaluations have reported reductions in sugar purchased from soft drinks and, in longer-term analyses, associations with reduced obesity in some age groups. It is now the most-cited example of a fiscal measure that shifts the product environment rather than placing the burden of change on individuals, and the tiered design is the transferable lesson.

Erythritol and the 2023 cardiovascular signal
Emerging scienceA 2023 Nature Medicine study reported that circulating erythritol was associated with major adverse cardiovascular events across several cohorts, and supported this with experiments showing enhanced platelet aggregation and thrombosis potential. The finding attracted substantial coverage and warrants attention. It also warrants qualification: erythritol is produced endogenously via the pentose phosphate pathway, and endogenous production increases in the presence of metabolic dysfunction, so elevated circulating levels may be a marker of the cardiometabolic disease that predicts events rather than a cause of them. The interventional arm involved a single high dose. This is a genuine open question under active investigation, appropriately described as an unresolved signal rather than an established harm.

Reverse causation in sweetener epidemiology
MethodologyThe dominant interpretive problem in this literature and the reason observational associations should be treated cautiously. People adopt non-sugar sweeteners predominantly because they already have overweight, obesity, type 2 diabetes or a family history of these conditions — precisely the characteristics that predict the outcomes being measured. Cohort studies therefore reliably find sweetener users have worse cardiometabolic outcomes, which is exactly what would be observed if sweeteners were entirely inert. Statistical adjustment for baseline BMI and glycaemia reduces but cannot eliminate this, since the decision to switch reflects a trajectory not captured at a single baseline. Randomised substitution trials, which avoid the problem, generally show modest favourable or neutral effects on weight.

Dental caries as the neglected outcome
Public health dentistryDental caries is the most prevalent noncommunicable disease worldwide, affecting an estimated 2 billion permanent teeth and 500 million primary teeth, and it is almost entirely attributable to free sugar exposure interacting with oral biofilm and modified by fluoride availability. It causes pain, infection, tooth loss, school and work absence, and in children is a leading cause of general anaesthesia admission in several high-income countries. It attracts a fraction of the attention given to sweetener safety, despite being settled, preventable and enormously burdensome. The frequency of sugar exposure matters as much as the total quantity, since each exposure produces a period of demineralisation — which is why sipping a sweet drink over an hour is worse than drinking it at once.