New Research with Mice Reveals the Dangers of Sweeteners — and the Bigger Lesson for Preclinical Labs

People land on this page because they’re searching for artificial sweeteners and mouse studies. Fair. The headlines are loud.

But if you run mouse or rat research, the more useful takeaway isn’t “sweeteners are evil.” It’s this:

Even minor changes in diet and water can change what you measure. Standardize them, or they’ll show up later as unexplained variability.

And yes, there have been plenty of “accidental” discoveries from unintended variables. If Alexander Flemming had better sterile technique, we might not have penicillin, the microwave oven may never have been created if Percy Spencer remembered he had a chocolate bar in his pocket, and the artificial sweetener saccharin itself was the result of Constantine Fahlberg not washing his hands after leaving the lab. You could get lucky and discover something novel… or you could be stuck looking at inconsistent data, wondering what’s going wrong.

With that in mind, below is an overview of what recent mouse-focused research has reported, followed by the practical part: what to control, what to document, and how to keep anesthesia and recovery physiology from becoming its own confounder.

What research with mice shows

The following are study findings in animal models, using specific doses, durations, and disease contexts. They show how “background inputs” can shift physiology in ways that can collide with your endpoints.

Aspartame

Doing work in anxiety or epigenetics? A study from Florida State University found that mice given free access to water sweetened with aspartame in an amount equivalent to having 6-8 cans of diet soda per day. showed more anxious behaviors. Not only did those mice show elevated signs of anxiety (measured through maze tests), the results were apparently transferred to their offspring and their offspring’s offspring.

The researchers also looked at the nervous systems of the mice and found significant changes in the amygdala, which plays a part in regulating anxiety. When the mice were given diazepam (brand name Valium), they stopped displaying anxiety behavior across all affected generations.

Sucralose

Doing work in immunology? Researchers at the Francis Crick Institute in London examined the effects of sucralose on the immune system. The team administered different dosages of sucralose to mice and compared their immune reactions to mice that received a similar dosage of a different artificial sweetener, sodium saccharin.

The researchers found that mice that received high doses of sucralose had lower T-cell levels compared to those that received saccharin, suppressing their immune systems. While a suppressed immune system is undesirable for most people, the team is investigating how high dosages of sucralose might be used to treat people suffering from autoimmune diseases.

Additionally, a 2023 Nature paper reported that high doses of sucralose in mice limited T cell proliferation and differentiation, producing measurable immunomodulatory effects.

Reporting around the paper highlights that the observed immune effects occurred at high intake, and that dose context matters when interpreting relevance.

Erythritol

Doing work in stroke models? A University of Colorado Boulder summary describes research suggesting erythritol can impair brain blood vessel cell function and reduce clot breakdown potential, framing this as conditions that could increase stroke risk.

A 2025 paper in Journal of Applied Physiology reports mechanistic findings in this direction.

The necessary context: dose and model matter

We’re not calling for you to put the can of Diet Coke down, or comparing bread to yoga mats, or trying to sell you our all natural organic detox supplements. These papers are not all designed to represent “typical human consumption,” and many use specific models to answer specific mechanistic questions. 

However, all these studies provide the same useful reminder—even common additives can alter physiology in measurable ways, and preclinical research depends on controlling variables you don’t want driving the signal. Variability can come from a variety of sources, some of which you might not think about when planning your studies.

The implications of new research for preclinical labs

1) Additives can be confounders without being “toxic”

Even if your standard chow doesn’t have any of these sweeteners, dietary additives can still change:

• palatability and water intake
• metabolic signaling
• immune tone
• vascular reactivity

If intake changes, stress changes. When stress changes, physiology changes. If physiology changes, your variability increases.

2) Confounders don’t announce themselves

Most confounders show up later as:

• wider variance
• inconsistent recovery
• “we had to increase n”
• results that don’t reproduce across cohorts

This is a 3Rs problem, because avoidable variability can lead to avoidable animal use.

3) “Diet and water” aren’t the only hidden variables

In rodent studies, the most common confound that labs underestimate is anesthesia and recovery physiology. Even when the study isn’t “about anesthesia,” anesthetic delivery, oxygen strategy, temperature drift, ventilation decisions, and monitoring practices can shift physiology enough to change outcomes.

What to control and document so your endpoints stay your endpoints

You don’t need a perfect protocol, but one that’s consistent and defensible.

Diet and water controls

Start with the basics and write them down:

• chow formulation, vendor, and lot
• water source and treatment
• any additives, sweeteners, or vehicles
• timing rules for feeding, fasting, and handling around procedures
• acclimation practices if intake is measured

If your model is specifically looking at metabolism, immune response, vascular tone, cognition, or microbiome-related outcomes, these details belong in the methods section, not in someone’s memory.

Anesthesia, warming, and monitoring controls

This is where a lot of “mysterious variability” is born.

A practical baseline to control:

• anesthesia approach and carrier gas strategy
• warming method during the procedure and recovery
• temperature monitoring plan (core vs surface is a real distinction)
• basic physiologic monitoring when stability matters
• ventilation decisions for longer or deeper anesthesia windows

Watch the SomnoFlo workflow video

If your study outcomes depend on stable physiology, anesthesia is not a background detail. It’s part of the experimental environment.

Watch the SomnoFlo video on general low-flow vaporizer benefits.

Build stability without adding complexity

If you’re tightening protocols to reduce variability and support refinement, these Kent tools are designed to make stable workflows easier to run consistently:

• SomnoFlo® — Rodent anesthesia system for consistent inhalant delivery and predictable recovery conditions that support refinement and reproducibility.
• SomnoFlo® O2Care — Oxygen blending anesthesia system that supports a more physiologically relevant carrier-gas strategy, improving refinement when oxygen choice matters.
• RightTemp® / RightTemp® Jr. — Rodent warming and temperature control platform that reduces hypothermia-related drift, supporting refinement and reducing avoidable variability.
• PhysioSuite® — Small animal physiological monitoring system that helps teams track stability during anesthesia and recovery, supporting refinement and better data quality.
• CODA® Non-Invasive Blood Pressure — Mouse/rat non-invasive blood pressure monitor that supports cohort-scale BP measurement without surgical implantation, aligning with refinement and reduction when appropriate.

What does the new research with mice reveal? 

For preclinical labs, the new research on the dangers of sweeteners is a practical lesson that’s already clear: inputs you don’t control can become biology you didn’t intend to study. This new research with mice reveals the dangers of sweeteners like aspartame, sucralose, and erythritol—is a reminder to treat diet and water as real variables that may impact your work.

If you want fewer surprises and more reliable outcomes, start by documenting diet and water—and then make anesthesia, warming, and monitoring a stable part of the workflow instead of a variable that changes from case to case.