Last reviewed: July 2026

Artificial sweeteners compared: aspartame, sucralose, stevia and acesulfame-K

At real-world intakes, the approved sweeteners have a stronger safety record than their reputation suggests, and the acceptable daily intakes are much higher than most people picture. A 70 kg adult would need roughly 14 cans of aspartame-sweetened cola in a day to reach the EFSA acceptable daily intake of 40 mg per kg of body weight. When IARC classified aspartame as "possibly carcinogenic" in 2023, the joint expert committee reviewing the same evidence left the ADI unchanged, because possible-hazard classification describes whether something could cause harm at some dose, not whether it does at yours. The honest verdict: sweeteners are a large improvement on sugar and a smaller one on water.

The four you will actually meet

Dozens are approved in the EU; four dominate the shelf. Each is hundreds of times sweeter than sucrose, which is why the quantities in a product are measured in milligrams and why they contribute effectively no calories.

The acceptable daily intake is not a danger line. It is set by finding the highest dose showing no adverse effect in long-term animal studies and then dividing by 100 as a safety margin. Exceeding it once does not cause harm; it means you have eaten into a margin designed to be conservative for lifelong daily intake.

Translating the limits into cans

Abstract milligrams per kilogram are why these numbers feel meaningless. Converted for a 70 kg adult:

• Aspartame: ADI of 2,800 mg a day, roughly 14 cans of diet cola.
• Sucralose: ADI of 1,050 mg a day, on the order of 25 or more portions of a typical sucralose-sweetened drink.
• Acesulfame-K: ADI of 630 mg a day, usually reached alongside other sweeteners since it is commonly blended.
• Steviol glycosides: ADI of 280 mg steviol equivalents, in the region of 20 to 30 sachets or several litres of stevia-sweetened drink.

Actual population intake surveys across Europe consistently find average consumers well under these figures, including high consumers and children, with occasional exceptions for specific sweeteners in specific subgroups. Blends are a real consideration though: a single product often contains two or three sweeteners precisely to keep each one low and to smooth out the aftertastes, so reading the ingredient list, a skill covered in the sugar alcohols explainer, tells you more than the front of pack.

The 2023 aspartame headlines, in context

In July 2023 IARC placed aspartame in Group 2B, possibly carcinogenic to humans, and the news cycle did what news cycles do. Two things make that classification less alarming than it read. First, IARC assesses hazard, not risk: it asks whether a substance could cause cancer under some circumstance and at some dose, not whether typical exposure does. Group 2B also contains aloe vera extract and pickled vegetables. Second, and more directly, JECFA reviewed the same body of evidence at the same time and reaffirmed the 40 mg/kg acceptable daily intake, explicitly stating there was no sufficient reason to change it. Both statements were released on the same day. Only one of them got headlines.

The weight and appetite question

The theory that sweeteners drive compensatory eating or disrupt metabolic regulation is plausible and has been tested. Randomised controlled trials comparing sugar-sweetened to low-energy-sweetened beverages generally find modest weight reduction with the swap, on the order of one to two kilograms over months, which is what you would expect from removing calories. Observational studies frequently show the reverse association, but they are badly confounded by reverse causation: people already trying to lose weight are the ones who switch to diet drinks. When the same question is asked in a trial where the direction of causation is controlled, the effect flips. The gut-bacteria concern raised by some studies is real enough to keep researching and currently rests on small studies with inconsistent results.

Who genuinely should avoid which

• Phenylketonuria: aspartame must be avoided because it releases phenylalanine, which is why every label carries the warning.
• Digestive sensitivity: sugar alcohols such as maltitol and sorbitol, which are a separate category, cause bloating and laxative effects at moderate doses. The intense sweeteners above do not.
• People who dislike the taste: a perfectly good reason, and not one that needs a scientific justification.
• Anyone using them to justify the rest of the food: the halo effect is the more realistic harm. A zero-sugar drink alongside a large meal has not changed the meal.

Where they fit

The useful comparison is not sweetener versus nothing, it is sweetener versus the thing it replaces. A 330 ml regular cola carries about 35 g of sugar and 140 kcal; the zero-sugar versions of the same category sit at 2 to 3 kcal per 100 ml against the 11 g of sugar per 100 ml on the full-sugar label, and a glass of orange juice lands close to the regular cola on sugar despite its healthier reputation. Sweetened dairy shows the same pattern: a sweetener-based vanilla yogurt carries a fraction of the sugar of the standard flavoured tubs beside it. For someone drinking a litre of regular soft drink a day, switching is one of the largest single calorie reductions available without changing a meal, a point the diet soda article works through in more detail. For someone drinking water, switching to sweetened drinks gains nothing.

Bottom line

Approved sweeteners are among the most heavily tested ingredients in the food supply, and at realistic intakes the evidence does not support the fear attached to them. They are a good tool for cutting sugar and a poor reason to consider a product healthy. Water remains the free, unbeatable default, and everything else is a preference you are now equipped to make on the numbers rather than the headlines.