Kero keeps pushing into territory far beyond weight loss, but the smartest reading of the newest studies is a reality check, not a victory lap. The cleanest human signal so far comes from a UK trial in treatment-resistant depression, while the diabetes data are smaller and more biomarker-driven than most headlines make them sound. If you are asking whether to treat these findings as actionable today, the answer depends on whether you are talking about epilepsy, where keto is already established, or depression and type 2 diabetes, where the science is still early.
What the depression trial actually showed
The strongest piece of evidence in this batch is the randomized clinical trial in JAMA Psychiatry. It enrolled 88 adults ages 18 to 65 in the United Kingdom, all with treatment-resistant depression and PHQ-9 scores of 15 or higher, and split them evenly between keto and a phytochemical control diet. The keto side was tightly managed, with fewer than 30 grams of carbohydrates a day and weekly dietetic support, and the team followed participants from February 22 to June 15, 2024.
That design matters because it is the first UK-based randomized controlled trial of keto for treatment-resistant depression, not just a before-and-after weight-loss experiment. The researchers also used a well-matched control diet and encouraged people in both arms to keep their weight stable, which helped separate diet composition from simple calorie loss. Treatment-resistant depression itself is not a vague label either. It usually means symptoms that persist after at least two antidepressant trials at adequate doses, and it is tied to poorer quality of life, higher healthcare costs, and increased suicide risk.
The primary endpoint was the PHQ-9 change from baseline to week 6, and that is where the signal appeared. The keto group improved by a mean of -10.5 points, compared with -8.3 in the control group, for a between-group difference of -2.18 points with P = .05. At 12 weeks, the gap was smaller and no longer statistically significant. The trial also tracked remission, anxiety, anhedonia, cognitive impairment, quality of life, and functioning, which gives the study breadth, but not the kind of large, durable effect that would justify calling keto a replacement for standard psychiatric care.
Min Gao of Oxford University framed the result as encouraging but modest, and that is the right scale for it. This is a real clinical signal, but it is not proof that keto is a stand-alone antidepressant.
Why the diabetes signal is interesting, and still incomplete
The diabetes story is a different kind of promise. Keto shifts the body into ketosis, where it burns stored fat and produces ketones for fuel, and the diet is typically high in fat, moderate in protein, and very low in carbohydrates. In practice, some people feel steadier appetite and energy on that pattern, which is part of why it keeps coming back in metabolic research.
The April 21, 2026 study highlighted by the Endocrine Society looked at 51 adults with type 2 diabetes, ages 55 to 62, and 71% of them were female. It compared a ketogenic diet with a low-fat weight-maintaining diet and found a greater reduction in the proinsulin-to-C-peptide ratio in the keto group, a marker of beta-cell stress. The authors said beta-cell function improved after three months of keto and that the change happened without substantial weight loss, which is important because it suggests something beyond simple scale change.
That is promising, but it is still a biomarker study, not a long-term remission trial. The press release says the work was published in the Journal of the Endocrine Society and involved researchers from the University of Alabama at Birmingham and the University of Glasgow, with Marian Yurchishin as the lead researcher. The clinical question people actually care about is whether keto can meaningfully change the course of type 2 diabetes, not just nudge a pancreas-stress marker in the right direction. This study moves that question forward, but it does not settle it.
The roundup that brought all of this together also pointed to two other pieces of evidence: a preclinical mouse study suggesting a keto-like diet may help manage blood sugar, and a third human study suggesting that some people with type 2 diabetes could reach remission. That mix matters because it shows where the field is leaning, but it also shows the evidence hierarchy clearly. A mouse study is not a treatment recommendation, and a remission claim still needs context before it becomes practical guidance.
Where keto already belongs in medicine
Keto is not new to medicine. It has been used for decades to treat epilepsy, and that is still the clearest example of a ketogenic diet with established clinical utility. What is new is the attempt to move it into psychiatry and metabolic disease, where the targets are more complex and the stakes for sustainability are much higher.
That is why the depression and diabetes studies deserve attention without hype. The depression trial was carefully run, but the effect was modest and short-term. The diabetes study suggested better beta-cell function, but it was small and did not hinge on a hard endpoint like remission. Taken together, they show a diet that may be doing more than the internet sometimes gives it credit for, but not nearly enough to call it a cure-all.
Who should act on this now
If you are using keto for epilepsy, you are on ground with a real medical track record. If you are looking at it for treatment-resistant depression or type 2 diabetes, the evidence is strong enough to make the diet worth discussing, but not strong enough to treat it as a self-directed substitute for standard care. Even in the depression trial, people got weekly support and a tightly controlled carb target under 30 grams a day, and sustainability was still an open question.
That is the practical line here: keto has moved from fringe enthusiasm to serious investigation, but for depression and diabetes it is still a tool under study, not a finished answer.
Every story on Keto Diet Magazine is assembled by an automated editorial system that works from verified research, official records, and credible reporting, then clears automated accuracy and moderation checks before it goes live. The standards that system follows are set and overseen by the people who run the publication. Read our full editorial policy.
Know something we missed? Have a correction or additional information?
Submit a Tip
