The Assumption That Defined an Industry
For decades, the pitch was simple. Non-nutritive sweeteners deliver sweetness without calories, and because they aren't metabolized like sugar, they pass through the body without consequence. The FDA approved them as safe. The American Diabetes Association included them in dietary guidance. An estimated 40% of American adults consume them regularly, typically in diet sodas, yogurts, and tabletop packets, motivated by the belief that replacing sugar with a calorie-free alternative is strictly better for metabolic health.
That belief rested on the assumption that these molecules are biologically inert. A 2022 randomized controlled trial published in Cell tested that assumption directly and found it wrong for two of the four most popular sweeteners.
What 120 Adults Showed in Two Weeks
Jotham Suez and colleagues at the Weizmann Institute of Science enrolled 120 healthy adults who normally avoided artificial sweeteners, then randomly assigned them to one of six groups for two weeks: daily sachets of saccharin, sucralose, aspartame, or stevia, all at doses below the FDA's acceptable daily intake, or controls receiving either a glucose-only sachet or no supplement. Every participant wore a continuous glucose monitor and completed nine oral glucose tolerance tests over the 29-day protocol, including a baseline week and a follow-up week after supplementation stopped.
The results split cleanly. Saccharin elevated glycemic responses starting in the first week (incremental area under the curve mean difference 783.5, p = 0.007) and persisting through the second (mean difference 811.2, p = 0.009). Sucralose followed the same pattern, with an even larger elevation emerging by the second week of supplementation (mean difference 976.3, p = 0.009), a trajectory indicating that the glycemic disruption established itself within the first seven days of exposure rather than building gradually over the full two-week period. Both effects reversed within a week of stopping.
Aspartame and stevia showed no significant glycemic effect at any time point during the supplementation period, and their glucose tolerance curves remained statistically indistinguishable from those of the two control groups throughout the entire 29-day protocol. Neither did the glucose control group or the no-supplement group, ruling out the possibility that the glucose carrier in the sachets or the testing protocol itself was driving the change.
How Gut Bacteria Mediate the Effect
The mechanism ran through the microbiome. All four sweetener groups showed distinct compositional shifts in stool bacteria, oral bacteria, and plasma metabolites. But Suez's team went further than compositional analysis: they transplanted fecal microbiomes from the top and bottom glycemic responders in each sweetener group into germ-free mice raised without any gut bacteria of their own. The mice that received microbiomes from human "top responders" developed impaired glucose tolerance. Mice that received microbiomes from "bottom responders" did not. Glycemic responses in the recipient mice largely mirrored those in their specific human donors, establishing a direct causal chain from sweetener consumption to microbial alteration to metabolic impairment that could be transferred between organisms in a controlled laboratory setting.
This design is what separates the finding from a correlation: the fecal transplant experiments demonstrate that the altered microbiome itself, not some direct pharmacological action of the sweetener molecule on host tissue, is sufficient to reproduce the glucose impairment in a new organism.
The Strongest Counterargument
In 2021, Joan Serrano and colleagues published a rigorous counterpoint in Microbiome: a double-blind, placebo-controlled trial in which 46 healthy adults received 400 mg/day of pure saccharin, roughly the maximum acceptable daily intake set by JECFA, for two weeks. A parallel arm ran 10 weeks in mice at 250 mg/kg/day. No microbiome changes were detected. No glucose intolerance emerged. The null result held in both species at high doses.
The discrepancy may come down to the vehicle: Suez's sachets contained saccharin mixed with a glucose carrier, mimicking commercial sweetener packets like Sweet'N Low, while Serrano used pure saccharin in capsules. The glucose carrier may interact with the sweetener in ways the pure compound does not. If so, the real-world relevance of Suez's finding is strengthened (since people consume sweeteners in packets, not purified powder), but the mechanistic attribution becomes murkier. This remains unresolved.
What We Didn't Prove
The trial enrolled 120 participants randomized across six groups, yielding roughly 20 per arm, a sample adequate for detecting the reported group-level glycemic effects but too small to characterize individual variation in microbiome response or to predict which people are most susceptible to sweetener-driven disruption of blood sugar regulation. That provides enough power to detect group-level effects but limits the ability to characterize individual variation or to identify which baseline microbiome features predict who will respond. The wide confidence intervals on the iAUC differences reflect this constraint, and the study was not designed to generate the kind of individualized microbiome signatures that would let a clinician predict in advance which patients will experience glucose impairment from a given sweetener.
The study lasted only two weeks, so whether these glycemic alterations accumulate, stabilize, or resolve over months or years of habitual use is unknown. The one-week follow-up after supplementation ended showed that glycemic responses returned to baseline, indicating short-term reversibility, but whether chronic daily exposure over months or years might gradually produce more durable microbiome shifts or self-sustaining metabolic changes is a question that a fourteen-day trial cannot begin to resolve.
Participants were Israeli adults who normally avoided sweeteners, and because gut microbiome composition varies substantially across populations, diets, and geographies, the same sweetener could produce different effects in someone with a radically different baseline microbiome. Generalizability to populations in Asia, Africa, or Latin America has not been tested, and given the documented differences in baseline microbiome profiles across dietary cultures, extending these conclusions globally without replication in diverse cohorts would be premature.
Finally, the question of dose context deserves scrutiny. Suez's 2014 Nature study used saccharin at doses exceeding the FDA's acceptable daily intake, drawing criticism. The 2022 trial used sub-ADI doses to address this, but the saccharin dose (180 mg/day) was still at the low end of real-world variation, and whether the effect scales linearly with dose is not established.
The Bottom Line
A randomized trial of 120 healthy adults found that saccharin and sucralose, consumed in commercial-style sachets at doses below the FDA safety limit for just two weeks, significantly impaired blood sugar control. The impairment was caused by changes to the gut microbiome, confirmed when fecal transplants from affected humans reproduced the effect in germ-free mice. The finding challenges the foundational assumption behind the non-nutritive sweetener market, an industry worth billions of dollars annually and endorsed by major health organizations, that these molecules pass through the human body without leaving a metabolic trace.
What You Can Do
If you use artificial sweeteners and are concerned about metabolic health, consider your specific sweetener. This trial found no glycemic effect from aspartame or stevia, while saccharin (the main ingredient in Sweet'N Low) and sucralose (the active compound in Splenda) both impaired glucose tolerance. Switching to stevia-based products or reducing sweetener use altogether are both evidence-supported options, and neither requires waiting for further replication since the intervention carries no known risk. If you wear a continuous glucose monitor, you can run the experiment on yourself: consume your usual sweetener consistently for a week, remove it for a week, and compare your post-meal glucose curves. The core lesson is that "zero calories" does not mean "zero metabolic effect." Your gut bacteria read the ingredient list even when the nutrition label says nothing is there.