SANTIAGO — For millions of people who reach for Diet Coke or a stevia-sweetened tea each morning, the assumption is uncomplicated: zero calories means zero consequences. Sucralose and stevia are the two most widely used non-caloric sweeteners in the world, approved by food safety regulators across four continents, and marketed on the premise that they pass through the body without being stored or processed the way sugar is. A study published this year in Frontiers in Nutrition suggests that premise misses a dimension regulators have not yet tested: what happens to the children and grandchildren of people who consume them.
Researchers at Universidad de Chile fed mice either sucralose or stevia at concentrations designed to match typical human intake, then bred those mice for two consecutive generations. Both the first and second generations received only plain water; they consumed no sweetener themselves. By the end of the study, the offspring showed measurable biological changes, in their gut microbiome composition and in the expression of genes governing metabolism and inflammation, that matched patterns seen in the parents who had consumed the sweeteners directly.
The study enrolled 47 mice split into three groups and was led by Dr. Francisca Concha Celume. Her team described their findings in the study, which drew wide public attention in late August after science media covered the results. The research did not produce its most striking results in a single generation: effects unfolded across both offspring generations in different ways, and the two sweeteners diverged sharply in how lasting their biological footprint proved to be.
Sucralose produced the more extensive and durable changes. In the parent mice that consumed it, fecal microbiome composition shifted toward higher bacterial diversity but with lower concentrations of short-chain fatty acids, the metabolic byproducts that gut bacteria generate when fermenting dietary fiber. Short-chain fatty acids serve as primary energy sources for cells lining the colon and play a central role in regulating immune function and glucose metabolism. Their depletion was not contained to the parent generation. First-generation male offspring of sucralose-consuming mice showed impaired glucose tolerance, a precursor pattern to type 2 diabetes. Second-generation male offspring showed elevated fasting blood sugar. The effects reached the grandchildren of animals that had consumed the sweetener.
Stevia’s trajectory was different. Its effects on the microbiome and gene expression were smaller and did not carry forward as consistently to the second generation. But second-generation female offspring of stevia-consuming mice showed elevated fasting blood sugar, a result that did not appear in the first generation of stevia offspring. The pattern reveals a divergence between the two sweeteners in which generation bears the greatest metabolic burden, and complicates the assumption that stevia, as a plant-derived compound, is necessarily the safer choice.
The researchers identified epigenetic modification as a probable mechanism for the transmission, changes to the marks on DNA that control how genes are expressed without altering the underlying sequence itself. Epigenetic changes can be inherited, and animal research has increasingly documented cases in which diet-induced epigenetic marks pass from parent to offspring. Whether that process explains the sweetener findings specifically has not been confirmed; the study cannot distinguish between epigenetic inheritance and other modes of intergenerational biological transfer.
Regulators have not assessed multigenerational effects in approving these compounds. Current safety frameworks evaluate a sweetener’s effects on the individual who consumes it at typical doses across a typical lifetime. The Chilean team’s controlled breeding design applied a different standard, one the existing approval process does not include. The sample of 47 mice is too small to produce definitive conclusions, and mouse physiology differs from human physiology in ways that frequently prevent direct extrapolation. The researchers called explicitly for follow-up studies using human data and longer observation windows.
That gap in the regulatory frame is not new. In 2023, the World Health Organization issued guidance recommending against using non-sugar sweeteners for weight management, citing evidence in its guidance that long-term use does not produce sustained reductions in body fat and may be associated with higher risks of type 2 diabetes and cardiovascular disease in adults. That guidance predated any study designed to assess multigenerational effects and does not address them.
The findings connect to a broader body of evidence on how specific dietary choices shape long-term metabolic outcomes. Research on metabolic health has consistently found that the specific molecules an individual consumes produce distinct biological effects even when caloric totals are matched, suggesting that the conventional nutrition framework understates the significance of what those calories are made of. The Frontiers in Nutrition study raises the possibility that this significance may extend across generations.
Concerns about long-term supplement safety have grown more acute as large studies have produced unexpected signals in compounds previously assumed to be benign, a pattern suggesting that current safety evaluation methods may systematically miss long-horizon effects.
What the Chilean study cannot yet answer is whether the intergenerational effects depend on the dose consumed, whether they reverse if parents stop using sweeteners before conception, or whether they compound when both parents are exposed. The sample was too small to test those variables. Whether the gut microbiome changes are the cause of the metabolic outcomes or a parallel signal of the same underlying disruption also remains open. A finding this novel in its design generates more questions than it settles.
For the billions of people who use sucralose or stevia daily, the study offers no directive, only a displacement: safety has been tested within a lifetime. It has not been tested across two.
