The 17-Person Study That Ruined Splenda

One study. Seventeen people. And somehow it rewired what millions of people believe about artificial sweeteners.

The study was published in 2013 out of Washington University. Researchers took 17 obese individuals, gave them either sucralose or water, then had them drink 75 grams of pure glucose, which is roughly the amount in a standard oral glucose tolerance test. The sucralose group showed about a 20 percent higher insulin response compared to the water group, and that single result spread across the fitness internet like a fact that had always been true.

But to understand why that number means less than it sounds like, you need the full picture of how sweeteners interact with your body in the first place.

Your pancreas does not just respond to glucose arriving in your bloodstream. It can also respond to signals that glucose is coming, and one of those signals comes from something called sweet taste receptors, which are the same molecular sensors your tongue uses to detect sweetness. Your gut and pancreas have these receptors too, and when they detect something sweet, they can prime the system to release insulin ahead of the actual glucose load. The technical name for this anticipatory release is the cephalic phase insulin response, which is basically your body getting insulin ready at the door before the glucose even knocks.

That mechanism is real. That is not in dispute.

The question is whether non-nutritive sweeteners, the ones with no calories, are actually activating that response in a meaningful way under normal conditions. And that is where the 2013 study gets complicated.

The sucralose dose used in that study was roughly five times what you would consume from a single diet soda. Every participant was obese, which matters because metabolic responses in people with obesity, particularly around insulin sensitivity and pancreatic function, do not necessarily generalize to lean individuals. And the effect only appeared when sucralose was consumed alongside the glucose load. The study was not designed to test sucralose in isolation, and when researchers have tested sucralose alone without pairing it with carbohydrates, the effect on blood sugar and insulin disappears.

A 2020 study published in Cell Metabolism confirmed this distinction in a more controlled way. Short-term sucralose consumption paired with carbohydrates impaired the body's neural and metabolic sensitivity to sugar. Consumption without carbohydrates did not. The mechanism appears to require glucose to be present at the same time for the sweet receptor signal to actually change anything downstream.

So the receptor priming happens, but priming without fuel is like pressing a gas pedal that is not connected to an engine. The anticipatory signal fires, but if there is no glucose to follow it, the body resets and nothing accumulates.

Now the fitness industry did not get that nuance. It took the 20 percent figure and turned it into a rule that sucralose spikes insulin, full stop, which is a different claim entirely.

The broader research tells a different story. A systematic review published in 2025 covering 101 articles and 100 experiments found no effect of non-nutritive sweeteners on blood glucose or insulin as a general pattern. A network meta-analysis of 36 trials involving 472 participants found similar results for sweetener-containing beverages. These are not small samples built from a single population under a single condition. They are aggregated results across many different study designs and populations.

Stevia specifically has shown something worth noting. In a study comparing stevia, aspartame, and sucrose, the stevia group had lower postprandial insulin levels than both the aspartame group and the sugar group, despite eating a similar-sized lunch afterward. The mechanism behind that is not fully established, but some research suggests certain compounds in stevia may interact with insulin-regulating pathways in a way that is modestly favorable rather than harmful. That evidence is preliminary, but the direction is the opposite of what most people assume about sweeteners and insulin.

Aspartame sits comfortably in the no-effect category across the evidence base. More than 100 experiments and the signal remains flat.

Sucralose is the only one with any plausible mechanism for concern, and even that concern is narrow. The scenario where evidence exists for a meaningful insulin interaction is: high dose, consumed together with a significant carbohydrate load, in individuals who are already metabolically compromised. Outside of that specific set of conditions, the data does not support the claim.

There is one additional layer worth knowing about. A 2022 paper in Cell examined personalized microbiome responses to non-nutritive sweeteners and found that saccharin and sucralose in particular altered gut microbiome composition in ways that correlated with changes in glucose tolerance in some individuals. The key word is some. The responses were highly individual, driven by each person's baseline microbiome, and the effect was not uniform across participants. This area is still being mapped, but it means the sucralose story is not entirely closed. It just has different contours than the fitness internet described.

If you are trying to make a practical decision from all of this: aspartame and stevia have the cleanest profiles across the full body of research. Sucralose is likely fine for most people in normal amounts, and the only scenario where caution is reasonable is pairing it habitually with high-carbohydrate meals. If that describes your situation, swapping to stevia costs you nothing and removes the only context where any risk signal exists.

The broader lesson is not about sweeteners. It is about how a result gets transformed into a rule. Seventeen people in a specific metabolic condition, exposed to an atypical dose, produced one number under one set of conditions, and that number became a universal dietary law. The study was not wrong. The interpretation just traveled so far from the original data that it stopped resembling it.

That is what happens when a single result fills a vacuum instead of joining a conversation.

---

References

1. Pepino MY, Tiemann CD, Patterson BW, Wice BM, Klein S. Sucralose affects glycemic and hormonal responses to an oral glucose load. Diabetes Care. 2013;369:2530-2535. PMID: 23633524. Source
2. Dalenberg JR, Patel BP, Denis R, et al. Short-Term Consumption of Sucralose with, but Not without, Carbohydrate Impairs Neural and Metabolic Sensitivity to Sugar in Humans. Cell Metabolism. 2020;313:493-502. PMID: 32130881. Source
3. Anton SD, Martin CK, Han H, et al. Effects of stevia, aspartame, and sucrose on food intake, satiety, and postprandial glucose and insulin levels. Appetite. 2010;551:37-43. PMID: 20303371. Source
4. Network meta-analysis of 36 trials (472 participants). Non-nutritive sweetener beverages. Nutrients. 2023. PMC9965414.
5. Systematic review and meta-analyses: 101 articles, 100 experiments. Advances in Nutrition. 2025. PMC12205327.
6. Romo-Romo A, et al. Effects of the non-nutritive sweeteners on glucose metabolism and appetite regulating hormones. Clinical Nutrition. 2019. PMID: 31697573. Source
7. Suez J, et al. Personalized microbiome-driven effects of non-nutritive sweeteners on human glucose tolerance. Cell. 2022;18518:3307-3328. PMID: 35987213. Source

---

Join the free community:
Men: Iron Forge Brotherhood
Women: Powerhouse Fitness