Researchers from University of Utah Health have published new findings in Science Advances that examine the long-term effects of the ketogenic diet. The results raise fresh questions about whether the diet is safe or effective for improving metabolic health over extended periods. Although the ketogenic diet was first developed as a treatment for epilepsy, it has become widely popular in recent years for weight loss and for managing conditions such as obesity and type 2 diabetes. In this study, scientists used mice to explore how the diet affects metabolism over time, with results pointing to potentially harmful changes in how the body handles fats and carbohydrates.
The ketogenic diet is built around extremely high fat intake and very limited carbohydrates. It was originally introduced to help control seizures in people with epilepsy. By sharply reducing carbohydrates, the diet pushes the body into a state known as ketosis, in which fat is broken down into ketone bodies that serve as an alternative fuel for the brain. This metabolic shift helps stabilize brain activity and reduce seizures, similar to the effects of fasting. While the diet has since been promoted for weight loss and metabolic health, much of the research to date has focused on short-term effects rather than long-term outcomes.
“We’ve seen short-term studies and those just looking at weight, but not really any studies looking at what happens over the longer term or with other facets of metabolic health,” said Molly Gallop, PhD, now assistant professor of anatomy and physiology at Earlham College, who led the study as a postdoctoral fellow in nutrition and integrative physiology at U of U Health.
How the Long-Term Study Was Conducted
To address this gap, Gallop and her team designed a long-term experiment using adult male and female mice. The animals were assigned to one of four diets: a high-fat Western diet, a low-fat high-carbohydrate diet, a traditional ketogenic diet where nearly all calories came from fat, and a protein-matched low-fat diet. The mice were allowed to eat freely for nine months or longer.
Over the course of the study, researchers tracked changes in body weight and food consumption. They also measured blood lipid levels, fat buildup in the liver, and blood sugar and insulin levels. In addition, the team examined which genes were active in insulin-producing pancreatic cells and used advanced microscopy to better understand the cellular changes behind the metabolic effects they observed.
Keto Prevented Weight Gain but Changed Body Composition
Compared with mice eating a high-fat Western diet, those on the ketogenic diet gained significantly less weight. This effect was seen in both male and female mice. However, the weight changes were not evenly distributed. Any weight gain that did occur on the ketogenic diet was mainly due to increases in fat mass rather than lean tissue.
Fatty Liver Disease Emerged Despite Weight Control
Although the ketogenic diet helped prevent weight gain, it also led to serious metabolic problems, some of which appeared within just a few days.
“One thing that’s very clear is that if you have a really high-fat diet, the lipids have to go somewhere, and they usually end up in the blood and the liver,” said Amandine Chaix, PhD, assistant professor of nutrition and integrative physiology at U of U Health and senior author on the study.
Excess fat accumulation in the liver, known as fatty liver disease, is a key marker of metabolic disease and is often linked to obesity. “The ketogenic diet was definitely not protective in the sense of fatty liver disease,” Chaix added.
The researchers also found striking differences between sexes. Male mice developed severe fatty liver disease along with impaired liver function, which is considered an important indicator of metabolic illness. Female mice, on the other hand, showed no significant buildup of fat in the liver. The team plans to investigate why female mice appeared resistant to this effect in future studies.
Blood Sugar Control Was Disrupted on Keto
The ketogenic diet also produced unexpected effects on blood sugar regulation. After two to three months on the diet, mice showed low blood sugar and insulin levels. However, this apparent benefit came with a serious downside.
“The problem is that when you then give these mice a little bit of carbs, their carb response is completely skewed,” Chaix said. “Their blood glucose goes really high for really long, and that’s quite dangerous.”
Further analysis showed that the mice struggled to regulate blood sugar because their pancreatic cells were not releasing enough insulin. The researchers believe this problem may stem from prolonged exposure to high fat levels, which appeared to stress the pancreatic cells and interfere with their ability to move proteins properly. While the precise biological mechanism is still being studied, the team suspects this cellular stress plays a central role in the impaired glucose response.
Importantly, blood sugar regulation improved once the mice were taken off the ketogenic diet, suggesting that at least some of the metabolic damage may be reversible.
What the Findings Mean for People
Although results from mice do not always translate directly to humans, the study highlights potential long-term metabolic risks that have not been well explored before. These findings suggest that people considering the ketogenic diet should carefully weigh possible benefits against potential harms.
“I would urge anyone to talk to a health care provider if they’re thinking about going on a ketogenic diet,” Gallop cautioned.
The study appears in Science Advances under the title “A long-term ketogenic diet causes hyperlipidemia, liver dysfunction, and glucose intolerance from impaired insulin secretion in mice.”
The research was funded by the National Institutes of Health, including the National Institute on Aging (grant number R01AG065993), the National Institute of Diabetes and Digestive and Kidney Diseases (grant numbers P30DK020579, F32DK137475, T32DK110966, DK108833, and DK112826), the National Heart, Lung, and Blood Institute (grant number HL170575), and the National Cancer Institute (grant number R01CA222570). Additional support came from the Damon Runyon-Rachleff Innovation Award (DR 61-20) and the American Cancer Society (RSG-22-014-01-CCB). Content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.