Study Deep Dive: Does Aspartame Increase Insulin? | Episode 20

In this episode, Layne Norton, PhD, dissects the landmark 2025 systematic review & meta-analysis in Advances in Nutrition: “The Effects of Aspartame on Glucose, Insulin, and Appetite-Regulating Hormone Responses in Humans”. (PMID: 4038180

This episode is perfect if you’ve heard claims like “diet soda spikes insulin,” “artificial sweeteners cause diabetes,” or “aspartame is toxic” — Layne uses this gold-standard review to separate science from fear-mongering.

If you’re into evidence-based nutrition, skeptical of viral health scares, or just want to know if your Diet Coke habit is secretly wrecking your metabolism — this episode has the receipts.

Subscribe for more deep-dive breakdowns of the latest research, and drop a comment: Do you use aspartame? Has this changed your view?

Study: The Effects of Aspartame on Glucose, Insulin, and Appetite-Regulating Hormone Responses in Humans: Systematic Review and Meta-Analyses

Introduction

Aspartame has been used as a low calorie sweetener in foods and beverages for ~50 years. It is a combination of 2 amino acids, aspartic acid and phenylalanine with a methyl ester group attached to it. It has utility as a sweetener because it is 200x sweeter than sucrose on a gram per gram basis.

While it has the same calorie density (4 kcal/gram) since it is 200x sweeter than sucrose, this allows beverages and foods to be sweetened with such a small amount of aspartame that they have almost no calories. For example, a can of diet coke contains 184mg of aspartame, which is 0.74 calories. Compare this with a can of regular coke which is 140 calories.

It was discovered by accident in 1965 when a research scientist named James Schlatter licked his fingers while turning a page in a book & noticed an intensely sweet taste. He had been working on synthesizing various di-peptides as ulcer drugs. 1

Despite it being accepted as safe by the FDA, many people have raised questions about its safety with many claims made that it causes cancer, cardiovascular disease(CVD), and type 2 diabetes (T2D). 2 3 4 The claims about CVD and T2D are usually ‘supported’ by the notion that artificial sweeteners (AS) like aspartame trigger an insulin response since your brain still senses a sweet taste and therefore your body releases insulin in response to this sweet taste. 5

While this is frightening for many consumers, the vast majority of these studies indicating potential harm of aspartame are either in animal models using supraphysiological doses or observational in nature and confounded greatly by the fact that people who consume greater amounts of AS also have overall poorer diet quality, consume more processed foods, less fruits/vegetables, tend to eat more calories, more frequently attempt weight loss, and are more likely to be obese. 6 Now some may look at this and say that aspartame & AS cause obesity due to increased consumption of ultraprocessed foods because aspartame increases insulin and this is driving increased hunger, consumption of processed foods, and therefore obesity. The issue is that reverse causality can also apply here:

Reverse causality: where what is perceived as the ‘effect’ is actually the cause. Instead of increased AS consumption causing obesity, obesity may be causing increased consumption of AS. That is, obese individuals are more likely to attempt to diet, and more likely to use AS. This is supported by data demonstrating that people who consume AS make more weight loss attempts than those who don’t. 6 7

According to the AI-driven research tool Consensus, a search of “is aspartame safe” yields mixed results. 40% of studies say yes, 13% are mixed, and 47% say no. 8 However, if we refine the search to only human randomized controlled trials, the gold standard for research studies (excluding animal data and observational data) we get a very different outcome: 83% of the studies concluded it was safe and 17% said not safe. 9 Of those that concluded it was not safe, however, as several of these studies were in unhealthy populations or people who were self-reported as ‘aspartame sensitive. 10

Mechanistically, these possible negative outcomes are also puzzling when considering how aspartame is metabolized, as it is converted into aspartic acid, phenylalanine, and methanol in the body. The breakdown is quite rapid, and it is complete as there are no documented cases of aspartame being found in the blood or any tissue in the body. 11 Aspartic acid and phenylalanine are both amino acids that our own body makes and I would add that both are present in very low doses (you get 20-30x more of each in a serving of chicken or steak). As such, the only plausible explanation for these possible negative effects of aspartame is from the production of methanol. Aspartame breaks down into ~50% phenylalanine, ~40% aspartic acid, and 10% methanol. For example, there is 184mg of aspartame in a 12 oz can of Diet Coke. This breaks down into 92mg phenylalanine, 73.6mg aspartic acid, and 18.4mg of methanol. Since aspartame is completely broken down into these compounds and does not stay intact in the body, we need only concern ourselves with the effects of the breakdown products.

This data calls into question whether or not the claims around insulin response, type 2 diabetes, and the risk factors associated with it are the result of aspartame ingestion or if this is a case of reverse causality/confounding variables.

Additionally, the human randomized controlled trials consistently demonstrate reductions in body fat levels when artificially sweetened beverages are consumed in place of sugar-sweetened beverages and the artificially sweetened beverages outperform water as a replacement for sugar-sweetened beverages. 12 13 14 15

The purpose of this study was to conduct a systematic review of the effect of aspartame on post-meal glycemic, insulin, and appetite-regulating hormone responses in human randomized controlled trials.

Methods

Inclusion criteria:

Population: Only studies in humans were included. The subjects could be any age, gender, or ethnicity, with or without obesity, with or without impaired glucose metabolism (pre-diabetes, Type 1 Diabetes (T1D), Type II Diabetes (T2D), and impaired glucose tolerance) but otherwise healthy. Studies of subjects with medical or clinical conditions other than those related to glucose metabolism were excluded.

Interventions: Included studies with any form of aspartame consumption (alone, in water, in combination with food, other low-calorie sweeteners (LCS), in tablet or capsule form. Any dosage was included and any pattern of consumption (ie, single or repeated exposure). Studies where it was not specifically stated that aspartame was used, but they simply referred to ‘LCS’ were excluded.

Comparisons: comparison arms must have used the same vehicle as aspartame without the inclusion of aspartame, without inclusion of aspartame using an alternate LCS, or without the inclusion of aspartame but with a caloric sweetener or other nutritive element such as sucrose, glucose, sugar alcohols, etc. In acute studies with high control, studies were only included if other aspects of the intervention that could influence the physiological responses were also controlled. These included nutritive components, beverage flavorings, and outcome assessment patterns. For longer-term studies in a more free-living setting, small differences between intervention and comparator arms were allowed, such as beverage flavorings. Long-term studies were excluded if if difference between aspartame and the comparator was known to affect digestive responses. For example, aspartame aspartame-sweetened beverage compared with a milk beverage or a multivitamin beverage.

Outcomes: Primary outcomes were glucose responses, insulin responses, and other appetite-regulating hormone responses. Secondary outcomes were energy intake, appetite, and adverse events.

Study design: Any controlled intervention study design. Parallel group studies, crossover studies were both acceptable. Studies that were excluded were in vitro studies, animal studies, and observational studies.

After searching for articles and excluding based on the above criteria, 101 articles were considered suitable for the review. Of the 101 articles, 73 reported results crossover experiments and 28 reported results on parallel-group experiments. Some of these studies had to be assessed differently as a few experiments reported on people with phenylketonuria (PKU), untreated T2D, postbariatric hypoglycemia, and a few studies only assessing digestive physiology during or after exercise.

Results

Meta-Analysis 1: Glucose responses to aspartame alone vs. comparator

In studies that compared aspartame + vehicle to vehicle alone or aspartame vs. other LCS, no effect on blood glucose responses were found

In studies that compared aspartame to sweet tasting carbohydrates like glucose, sucrose, fructose, there was a significant difference in glucose responses with the sweet tasting sugars causing a significantly greater response in blood glucose and the effect size was large

In studies that compared aspartame with non-sweet tasting carbohydrates similar results were found with non-sweet tasting carbohydrates causing a significantly greater blood glucose response and this effect size was also large

In studies comparing aspartame with other nutritive components like in a milkshake, a food, or a meal, the nutritive components produced a much greater glucose response than the aspartame treated groups. Once again the effect size was considered large.

Meta-Analysis 2: Glucose responses to aspartame with a nutritive component vs. nutritive component with no aspartame

In studies that compared a nutritive vehicle with aspartame vs. a nutritive vehicle alone without aspartame or a nutritive vehicle with other LCS, there were no differences in blood glucose responses

In studies comparing a nutritive vehicle with aspartame vs. a nutritive vehicle with sweet tasting sugars, there was a significantly lower blood glucose response with the aspartame condition vs. the nutritive vehicle with sweet tasting sugars with a large effect size.

In studies comparing a nutritive vehicle with aspartame vs. a nutritive vehicle with non-sweet tasting sugars, there was a significantly lower blood glucose response with the aspartame condition vs. the nutritive vehicle with non-sweet tasting carbohydrates and this effect size was also considered large.

In studies comparing a nutritive vehicle with aspartame vs. a nutritive vehicle with other nutritive components there was a significantly lower glucose response in the aspartame condition vs. the nutritive components condition and this effect size was considered large.

Meta-Analysis 3: Insulin responses to aspartame alone vs. comparator

In studies that compared aspartame + vehicle to vehicle alone, no effect on blood insulin responses were found

In studies that compared aspartame to sweet-tasting carbohydrates like glucose, sucrose, and fructose, there was a significant difference in insulin responses, with the sweet-tasting sugars causing a significantly greater response in blood insulin, and the effect size was large

In studies that compared aspartame with non-sweet tasting carbohydrates, similar results were found with non-sweet tasting carbohydrates, causing a significantly greater blood insulin response, and this effect size was also large

In studies comparing aspartame with other nutritive components, like in a milkshake, a food, or a meal, the nutritive components produced a much greater insulin response than the aspartame-treated groups. Once again, the effect size was considered large.

In studies that compared aspartame to other LCS, slightly higher blood insulin levels were found for the aspartame-treated groups, but this sub-analysis only contained 4 total studies. Aspartame-treated groups with stevia or monk fruit had slightly lower insulin levels compared to aspartame, and groups treated with allulose had similar insulin levels as those treated with aspartame.

Meta-Analysis 4: Insulin responses to aspartame with a nutritive component vs. nutritive component with no aspartame

In studies that compared a nutritive vehicle with aspartame vs. a nutritive vehicle alone without aspartame or a nutritive vehicle with other LCS, there were no differences in blood insulin responses

In studies comparing a nutritive vehicle with aspartame vs. a nutritive vehicle with sweet-tasting sugars, there was a significantly lower blood insulin response with the aspartame condition vs. the nutritive vehicle with sweet-tasting sugars, with a large effect size.

In studies comparing a nutritive vehicle with aspartame vs. a nutritive vehicle with other nutritive components, there was a significantly lower insulin response in the aspartame condition vs. the nutritive components condition, and this effect size was considered large.

Appetite, appetite-regulating hormones, and calorie intake

Where appetite and energy intake are concerned, the results mirrored those of the blood glucose and insulin responses. When aspartame was compared with placebo or other LCS there were no differences. But significant reductions were found when aspartame was compared with sugars or other nutritive compounds.

There were few effects found on appetite-regulating hormones

Other Results

Very few adverse events were reported across all studies

In one study, people who self-reported being aspartame sensitive demonstrated no effects of aspartame (in a cereal bar) on fasting glucose, insulin, insulin sensitivity, adverse events, though a decrease in glucose-dependent insulinotropic peptide (GIP) and an increase in glucagon-like-peptide-1 (GLP-1) were observed.

In studies using T2D subjects, similar results were observed for glucose and insulin responses compared to healthy adults.

In medium-term studies (2-30 days) intakes from 15-45 mg/kg BW/d (equivalent to ~7-20 diet cokes per day in someone who weighs 175 lbs) did not affect blood glucose, insulin, or adverse events compared to placebo. These results were consistent between healthy individuals and those with T2D.

In long-term studies (>30days), there was a range of measures and vehicles provided as comparators. The overall outcome was that there was no difference in HbA1c, blood glucose, insulin, insulin sensitivity, leptin, GLP-1, GIP, or glucagon. Studies assessing energy intake either found no difference or lower energy intake in groups consuming aspartame. In studies assessing appetite, there were either no differences between groups except for studies where people consumed aspartame-sweetened beverages vs. water or beverages sweetened with saccharin and reported lower hunger for those consuming aspartame-sweetened beverages. There were no differences in reported adverse events. One long-term experiment used a high dose of aspartame (2700mg/day, equivalent to ~15 Diet Cokes per day) in people with T2D and observed no changes in glucose metabolism or adverse events.

Discussion

Due to considerable heterogeneity in study design (acute, medium term, long term, different dosages, different modalities, etc), combining the studies was difficult. This is why meta-analyses were only performed for acute crossover studies

The summation of the research is that aspartame had no effect on blood glucose or insulin responses when compared to a non-caloric placebo and resulted in significantly lower blood glucose levels compared to any nutritive component it was tested against (sugar, carbs, etc). There were no differences in blood glucose levels compared to other LCS, but there were slightly higher insulin levels; however, this may be due to a small number of studies assessing aspartame vs. other LCS

The medium and long term studies tended to mirror the short term study results. Long term studies found no effect of aspartame consumption on blood glucose, insulin, or markers of insulin sensitivity like HbA1c.

Lack of effects were found in not only healthy individuals, but those with T2D, and self-reported ‘aspartame sensitivity.’

The results of this review are similar to other reviews on the topic. 16 17 18

These results also align with what we know about aspartame physiology. As discussed previously, Aspartame does not remain intact in the body, it rapidly breaks down into phenylalanine, aspartic acid, and methanol. As such, aspartic acid and phenylalanine are not compounds to be worried about (if you are concerned about them then you need to be way more concerned with any source of protein), so let’s focus on the methanol. How much methanol would we need to consume to see ill effects, and what are those?

The body itself produces approximately 300-600mg of methanol per day and we may consume up to 1000mg per day from various fruits and vegetables. 19 At high doses, methanol has toxic effects on the nervous system and can lead to blindness. This is due to the build up of the breakdown product formate (methanol is metabolized out of the body through conversion to formaldehyde then formate and finally to carbon dioxide). Ingestion of high doses of aspartame, up to ~4800mg (26 cans of diet coke equivalent in 8 hours) over an 8 hour period did not lead to increases in formate or methanol levels. 20 Further, the amount of methanol produced by fruits, vegetables, and alcohol is much greater than what is produced from aspartame. 21

Additionally, the deleterious effects of methanol exposure are nervous system toxicity, blindness, brain disorders, cognitive deficits, and Parkinsonism. Extremely high doses in the short term can cause metabolic acidosis, cardiopulmonary failure, coma, seizures, and death. At lower deleterious levels (still much higher than what could be obtained from diet) methanol can cause drowsiness, cns depression, confusion, headache, dizziness, ataxia, nausea, and vomiting. 22

What about formate aka formic acid exposure since that is the compound that methanol metabolizes into before elimination from the body? Most of the data in humans that examined formic acid ingestion demonstrated CNS depression, metabolic acidosis, nephropathy, hematuria, anuria, and possibly death. 23 These effects appear similar to methanol.

These doses are far beyond what is possible from aspartame ingestion. That said, acute methanol poisoning can cause hyperglycemia (elevated blood glucose). 24 This may be due to a stress response that elevates cortisol and glucagon, raising blood glucose levels as well as damage to the pancreas itself. 25 This, however, suggests that methanol may reduce insulin levels, rather than increase them.

Phenylalanine ingestion has been demonstrated to increase plasma insulin concentrations, but it’s important to point out the dosage and magnitude of the insulin response in this study. 26 The dosage was 1mmol phenylalanine/kg of lean body mass, which equated to approximately 9.75g phenylalanine (165mg phenylalanine/kg LBM). To ingest this amount of phenylalanine from aspartame, you’d need to consume over 100 cans of Diet Coke at one sitting. Further, the increase in insulin was relatively small. A 25g dose of glucose increased the insulin AUC ~5-6x more than the huge dose of phenylalanine. Further, this was offset by an increase in glucagon in the group given phenylalanine, resulting in no change in plasma glucose.

Aspartic acid ingestion also has been demonstrated to result in a small increase in insulin, but much less than phenylalanine, with the latter’s insulin response approximately 2.5x greater at the same 1mmol/kg LBM dose (7.85g AA or 133mg/kg LBM) than aspartic acid (both doses would require over 100 cans of diet coke at one sitting to achieve). So while these amino acids can increase insulin, the dosages used are orders of magnitude beyond what can be obtained through aspartame ingestion, and even then, only result in a small increase in insulin.

This makes sense, considering that if aspartame causes an insulin response with no ingestion of glucose, this would lead to a drop in blood glucose and hypoglycemia. This does not occur, indicating that either insulin is not increased in response to aspartame ingestion, or if it is increased, it is offset by an increase in glucagon.

This all fits with the results of this meta-analysis and systematic review

Conclusion

The findings of this review suggest few impacts of aspartame consumption on appetite-regulating hormones, with potential benefits for glucose metabolism, energy intake, and appetite when compared with the consumption of sugars, and no detrimental contraindications following consumption in the long term…

The majority of studies investigated blood glucose and insulin levels over the short term, and meta-analyses of these studies reveal no effects of aspartame when compared with vehicle or other LCS, and found lower blood glucose and insulin levels following aspartame compared with sugars or other nutritive components. Medium and long-term studies demonstrate few effects of aspartame consumption regardless of comparator.

References

Intro
What is Aspartame
Health Claims
How Aspartame is Metabolized
HRCTs
Outwork Nutrition
Study Overview
Inclusion Criteria
Results
David Protein
Results Cont.
Discussion
Conclusion