Aspartame at a normal dose is not the hazard. The sugar in a “natural” drink often is.

Food-safety agencies still treat aspartame as acceptable for people without phenylketonuria, including the amount in diet soda. A bottled yerba mate can wear an organic leaf on the label and still be a cane-sugar drink. A date is real food, and it is also a small block of sugar.

The July 2023 headline did not say a can of diet soda causes cancer. The cancer-research arm of the World Health Organization put aspartame in Group 2B, “possibly carcinogenic,” on limited evidence. That is a weak bin. In 2011 the same agency put radiofrequency fields from cell phones there, and a microwave oven uses energy in that same frequency range. The same month, the food-additive committee looked at dietary doses and left the limit where it was. In September 2026 the European Food Safety Authority said again that current dietary exposure is not linked, causally, to cancer or to the other harms that list usually names.

The limit is a number, and normal drinking sits under it

Aspartame is a methyl ester of two amino acids, phenylalanine and aspartic acid, about 200 times sweeter than sugar. After you swallow it, it is fully split in the gut. Nothing called aspartame circulates in the blood. The three pieces (phenylalanine, aspartic acid, and methanol) are the same substances released from ordinary food. That is the Joint FAO/WHO Expert Committee on Food Additives in 2023, not a slogan.

The acceptable daily intake used by that committee and by EFSA is 40 mg per kilogram of body weight per day, for a lifetime. It is the highest chronic animal dose with no observed adverse effect, 4,000 mg/kg, divided by 100. The U.S. FDA uses 50 mg/kg. JECFA’s exposure estimates for what people actually eat stay under the line: up to about 5 mg/kg a day on average in adults, and about 12 mg/kg in high adult consumers.

Can counts are the first whole can that meets the 40 mg/kg line. The World Health Organization’s worked example for a 70 kg adult is more than 9 cans at 300 mg, or more than 14 cans at 200 mg, assuming no other source. Brand formulas differ and are not always printed, so the page uses that official range.

Diet soda, in milligrams

In the United States, Coca-Cola says Diet Coke is sweetened with aspartame, and Coke Zero Sugar with aspartame plus acesulfame potassium. The line “phenylketonurics: contains phenylalanine” is how you know. The WHO example is 200 or 300 mg a can. A 1989 dosing study used the same scale: 600 mg stood for 36 ounces of diet drink, which is 200 mg in 12 ounces.

Two cans at 250 mg is 500 mg. For a 70 kg adult that is 7 mg/kg, about one-sixth of the European limit. Six cans, 1,500 mg, is still under 2,800 mg.

Packets, gum, and yogurt count toward the same daily total.

People who drank the beverages

Lim and colleagues followed 285,079 men and 188,905 women, ages 50 to 71, in the NIH-AARP study. Intake came from soda, fruit drinks, iced tea, and packets in coffee and tea. Over about five years: 1,888 hematopoietic cancers and 315 malignant gliomas. At 600 mg a day or more, the relative risk of hematopoietic cancer versus none was 0.98 (95% CI 0.76–1.27). For glioma, at 400 mg a day or more, it was 0.73 (95% CI 0.46–1.15). Diet was self-reported and follow-up was short. It is strong evidence against a near-term wave of those cancers. It is one cohort.

Blood levels, not vibes

• Stegink and colleagues, 1981. At 34 mg/kg taken at once, near a full day’s limit, blood methanol stayed below detection. At 200 mg/kg, on the order of fifty to seventy cans for a 70 kg adult, the toxic metabolite formate did not rise, and eye exams did not change.
• Stegink, Filer, and Baker, 1989. Six adults drank eight servings of 600 mg, an hour apart. Phenylalanine never cleared the top of a normal after-meal range. Methanol and formate stayed inside normal limits. Six people, one day: a heavy-user simulation, not a lifetime trial.

Methanol is about 10 percent of aspartame by weight, so 250 mg in a can is about 25 mg of methanol. At a full day’s limit taken at once, the 1981 study could not detect a rise in blood methanol. EFSA’s 2013 panel also concluded that aspartame does not harm the brain or nervous system and does not affect behavior or cognition in children or adults, and that phenylalanine from aspartame at the intake limit is not a fetal risk except in phenylketonuria. The same can, set next to an egg, a cup of milk, and tomato juice, is further down.

Diet soda does hydrate you. Water is the comparison, and it ties.

The claim is that caffeine makes the can a diuretic, so the drink does not count as fluid. The test of that claim is urine after the same volume of diet cola and still water.

Maughan and colleagues, 2016, gave 72 men a liter of still water or a liter of another common drink, then collected urine for 4 hours. The diet cola was Diet Coke: 100 percent water, about 4 kcal per liter, and 127 mg of caffeine per liter. Urine output after diet cola was not different from urine output after still water. Still water produced 1,337 grams of urine over those 4 hours. Cola, coffee, tea, and a sports drink landed in the same “not different from water” group. Milk and an oral rehydration solution produced less urine, so they held onto more of the liter. Diet cola did not beat water. It matched it.

A second, smaller trial asked the day-long version. Grandjean and colleagues, 2000, had 18 men drink their fluid allowance as water, or as water plus caffeinated diet cola and the other test drinks. Body weight, blood, and urine markers of hydration did not differ. Average weight loss across every condition was 0.30 percent. Eighteen men is a small sample. The result still points the same way as the liter test: the caffeine in the cola did not cancel the water.

A 12-ounce can is about a third of the liter in the 2016 study, so the caffeine in one can is on the order of 45 mg, against 127 mg in the liter that already matched water. The can counts as fluid. Water remains the drink with nothing else in it. The caffeine is still a stimulant. Teeth are the next section, and a diet soda is fine there too.

Diet soda is fine for teeth

A normal can is fine. The alarming photos come from leaving a tooth in soda for 25 hours, or for 14 days. Saliva clears a drink in minutes. Those baths are how a lab ranks drinks against each other. They are not what happens when someone finishes a can.

von Fraunhofer and Rogers, 2004, ran the 14-day version and kept replacing the liquid. Diet Coca-Cola, pH 3.22, removed about 3.07 mg of enamel per square centimeter. Regular Coca-Cola, pH 2.48, removed about 2.78. The diet cola and the regular cola were the same. Tap water and root beer removed none. Diet Mountain Dew removed about 14.8 mg/cm², and a canned iced tea about 9.0. Even in a two-week soak, the diet cola was on the low end.

Ehlen and colleagues, 2008, used 25 hours and say outright that a mouth is not that bath. Mean enamel lesions were about 61 micrometers for Diet Coke, 57 for apple juice, 92 for Coke, 100 for Red Bull, and 131 for lemon-lime Gatorade. Diet Coke was among the gentler drinks in that set.

Regular cola also carries 39 grams of sugar, and sugar is what feeds cavities. A diet soda skips that. Citrus sodas, canned tea, sports drinks, and energy drinks are the ones these tests mark as harder on enamel. A diet soda with a meal is an ordinary drink, and it is fine for teeth.

von Fraunhofer 2004 · Ehlen 2008, 25-hour lesions · Moynihan 2014, sugar and cavities

An egg, a glass of milk, and tomato juice already contain more

Aspartame does not introduce a new chemical. Digestion splits it into aspartic acid, phenylalanine, and methanol, about 40 percent, 50 percent, and 10 percent by weight. The 250 mg can used above is therefore about 100 mg of aspartic acid, 125 mg of phenylalanine, and 25 mg of methanol. Those three are already in ordinary food, and a plain breakfast beats the can on each of them.

The egg is the aspartic acid. USDA’s record for a raw whole egg lists 1.33 grams of aspartic acid per 100 grams. One large egg is 50 grams, so about 665 mg. That is roughly seven times the can. The same egg also has 0.68 grams of phenylalanine per 100 grams, about 340 mg, which is already more phenylalanine than the can’s 125 mg.

Milk is the phenylalanine comparison on its own. One cup of whole milk, 244 grams, lists about 398 mg of phenylalanine and 659 mg of aspartic acid. About three times the phenylalanine in the can, and about seven times the aspartic acid.

The tomato is the methanol, and the measurement is juice, because that is what was analyzed. Hou and colleagues, 2008, found 10.92 mg of methanol per 100 mL in fresh tomato juice, the highest of the vegetables they squeezed. Twelve ounces, the same volume as the soda, is about 39 mg, against about 25 mg from the aspartame. About one and a half times, not a different universe. Juice left for three hours measured higher, about 68 mg in that volume from the refrigerator and about 86 mg at room temperature, because enzymes keep freeing methanol from pectin. The table uses the fresh number.

One egg, one cup of milk, and those 12 ounces of tomato juice come to about 1,324 mg of aspartic acid, 738 mg of phenylalanine, and 39 mg of methanol. The can is 100, 125, and 25. EFSA’s 2013 panel said the contribution of aspartame to intake of these three is low beside the rest of the diet. The amino-acid gap is the obvious one. The methanol gap is smaller, and it still favors the juice.

The egg and the milk are the defense of the sweetener. They are ordinary protein foods, and a normal meal already contains more of the same two amino acids than the can. The juice does the same job for methanol. The diet soda is a small addition of chemicals that food already provides in larger amounts.

USDA egg · USDA whole milk · Hou 2008, methanol in juice · EFSA 2013, contribution from aspartame is low

Why a reasonable person could think it is harmful

The belief did not come from nowhere. It came from a real classification, a real set of animal papers, a scary molecule name, headaches people actually feel, and surveys that mix up cause and effect. Each of those is weaker than the sentence people repeat.

A cancer agency used the word “possibly”

On 14 July 2023, IARC classified aspartame as Group 2B. The basis was limited evidence for liver cancer in humans, limited evidence in animals, and limited mechanistic evidence. IARC’s own summary says chance, bias, or confounding could not be ruled out of the three beverage studies it used, and that a Group label is not a risk at a given dose. The three studies were Stepien 2016, Jones 2022, and McCullough 2022. Artificially sweetened drinks stood in for aspartame.

JECFA, meeting the same month, reaffirmed 0–40 mg/kg and said there was no convincing evidence of adverse effects after ingestion. The FDA wrote that “possibly carcinogenic” does not mean aspartame is actually linked to cancer, and that the studies had significant shortcomings. The short paper is Riboli and colleagues, Lancet Oncology, 2023.

The group label itself settles very little. Group 2B means limited evidence, and the list is full of ordinary exposures. In 2011 IARC put radiofrequency electromagnetic fields in that same group, from limited evidence on wireless phones. That is the cell-phone classification. The frequencies they reviewed run from 30 kHz to 300 GHz. A microwave oven runs at 2.45 GHz, inside that range. The studies they weighed were about phones, so the oven inherits the headline only because it uses the same kind of energy. A phone and a can of diet soda share the word “possible.” A normal phone call and a normal can are still dose questions, and the dose committee did not move the aspartame limit.

Rats in Italy developed tumors

Most oral cancer studies of aspartame were negative. The exception is the Ramazzini series: Soffritti 2006, Soffritti 2007, and Soffritti 2010. The authors reported lymphomas, leukemias, and, in male mice, liver and lung tumors. JECFA did not use them to set the limit because the animals were kept until natural death, so late-life tumors are hard to read, and because of problems in design and reporting. IARC counted them only as limited animal evidence and recorded its own doubts about litter effects and diagnoses. Those papers are why the fear exists. They are not why the intake limit moved. It did not move.

A French cohort found a 15 percent higher cancer rate

Debras and colleagues, 2022, in NutriNet-Santé, reported a hazard ratio of 1.15 (95% CI 1.03–1.28) for higher aspartame consumers versus non-consumers, and 1.22 for breast cancer. The authors list residual confounding and reverse causality as limitations. Schernhammer and colleagues, 2012, found more non-Hodgkin lymphoma and multiple myeloma in men who drank diet soda daily, and not in women. JECFA’s reading: some cohorts reported signals, no cancer type was consistent, and exposure was often a blurry stand-in for aspartame.

Methanol is a poison, in a different quantity

The word is doing the work. Methanol poisoning is real, and it is formate that injures people. The human dosing studies above are the relevant scale: a can, a day’s limit, even an absurd single-day stack. Blood formate did not accumulate. A forwarded 1990s message that blamed aspartame for multiple sclerosis and lupus is not a study; those diseases are not what the clinical file shows.

Headaches are the claim with a split trial record

People who get a headache after a diet soda are not inventing the headache. The question is whether aspartame caused it. Schiffman and colleagues, 1987, in the New England Journal of Medicine, challenged 40 people who believed aspartame triggered their headaches. Headache followed aspartame in 35 percent and placebo in 45 percent. Not a difference. Van den Eeden and colleagues, 1994, in Neurology, found the opposite lean: among people who already blamed aspartame, headache days were 33 percent on aspartame and 24 percent on placebo. The gap was carried by the volunteers who had been “very sure” beforehand. So a subset may be susceptible. That is a reason for that person to stop. It is not evidence that the ingredient harms the general population at a normal dose.

Diet soda shows up next to illness in surveys

People who already have weight gain or diabetes switch to diet drinks. A snapshot then “finds” diet soda beside the disease. That is reverse causation, and JECFA flagged it for the non-cancer outcomes. Separately, in May 2023 the WHO advised against using non-sugar sweeteners as a weight-control strategy. That is a statement about body weight, and it did not revoke the aspartame limit. Replacing sugar removes the sugar from that drink. It does not repair a diet. Acid in cola can still erode enamel. Caffeine is a different ingredient.

EFSA’s September 2026 opinion, written with all of this already published, kept 40 mg/kg and said the evidence does not support a causal link at current dietary exposure with cancer, cardiovascular disease, liver or kidney toxicity, neurotoxicity, developmental effects, or disturbed glucose regulation.