Ketogenic diet in brain cancer: what the ERGO2 and GLAD trials actually found
The ketogenic diet has attracted more interest in cancer research than almost any other dietary intervention. The theoretical basis is compelling: cancer cells rely heavily on glucose for energy via the Warburg effect, while most healthy cells can adapt to using ketone bodies derived from fat. A diet that drastically reduces glucose availability might, in theory, selectively disadvantage tumour cells.
In brain cancers specifically — where standard treatments have limited efficacy and where tumour cells show particularly high glucose dependence — this hypothesis has generated two of the most closely watched clinical trials in nutritional oncology. This post looks at what those trials actually found, and what that means for anyone trying to make sense of the ketogenic diet and cancer.
The ERGO2 trial
The ERGO2 trial was a randomised controlled trial conducted in Germany, published in the International Journal of Radiation Oncology Biology Physics in 2020. It enrolled patients with recurrent glioblastoma — one of the most aggressive brain cancers, with a median survival of around 15 months even with standard treatment — and randomised them to either a ketogenic diet alongside standard care or standard care alone.
The primary endpoint was overall survival. The trial enrolled 41 patients, which is small but appropriate given the rarity and severity of glioblastoma and the difficulty of conducting larger trials in this population.
The results were sobering: the ketogenic diet group did not show a statistically significant improvement in overall survival compared to the control group. Median overall survival was numerically similar between arms. There was no clear clinical benefit from the ketogenic dietary intervention in this population.
Importantly, the trial also documented the challenges of adherence. Glioblastoma patients face significant physical and cognitive burdens, and maintaining strict ketosis in this context proved difficult for many participants. The trial could not rule out the possibility that insufficient adherence limited any potential effect — but it also could not demonstrate benefit under real-world conditions.
The GLAD trial
The GLAD trial took a different approach. Published in Neurology in 2021, it studied patients with newly diagnosed glioma receiving standard treatment. The intervention was a “low glycaemic anti-inflammatory diet” — not strictly ketogenic, but carbohydrate-restricted and designed to reduce glycaemic load and systemic inflammation.
This was a Phase 2 trial, meaning it was designed to assess feasibility, safety, and preliminary efficacy signals rather than to provide definitive evidence of benefit.
The GLAD trial found that the low glycaemic, anti-inflammatory dietary pattern was feasible and well tolerated alongside standard glioma treatment — an important finding in itself, since concerns about patient burden and compliance have been a significant barrier to dietary intervention trials in oncology. The trial also found that dietary adherence was associated with favourable changes in metabolic markers including fasting insulin and inflammatory cytokines.
Critically, the GLAD trial was not powered to demonstrate a survival benefit, and it did not claim one. What it demonstrated was that a dietary modification programme is achievable in this patient population and produces the expected metabolic signals — establishing the groundwork for a larger Phase 3 trial.
What these trials tell us — and what they do not
Taken together, the ERGO2 and GLAD trials represent the most rigorous human evidence available on dietary carbohydrate restriction in brain cancer. Their conclusions are honest and limited.
The ketogenic diet did not improve survival in recurrent glioblastoma in the ERGO2 trial. This is an important negative finding that should temper enthusiasm in the advocacy literature, where the ketogenic diet is sometimes presented as a near-proven cancer treatment.
The low glycaemic anti-inflammatory approach in GLAD was feasible and produced expected metabolic effects, but was not designed or powered to demonstrate clinical benefit. Phase 3 evidence is still awaited.
Neither trial establishes the ketogenic diet as an evidence-based cancer treatment. They do establish that dietary carbohydrate restriction is clinically feasible in brain cancer patients — a necessary prerequisite for any future definitive trial.
The broader context
The ERGO2 and GLAD trials sit within a wider body of evidence that is largely preclinical. Laboratory studies have consistently shown that ketogenic dietary conditions disadvantage tumour cells across multiple cancer types. But the translation from laboratory to clinical benefit in humans has been far less straightforward — as is common in cancer research.
The reasons for this gap are likely multiple. Human tumours are metabolically heterogeneous: different cells within the same tumour may have different metabolic vulnerabilities. Some tumours upregulate pathways that allow them to use ketone bodies as an alternative fuel source, partially circumventing the intended mechanism. And the metabolic stress of cancer and its treatment makes strict dietary adherence extremely difficult to sustain.
This does not mean the hypothesis is wrong — it means the evidence required to confirm benefit in humans is not yet available.
What this means in practice
If you have a glioma diagnosis, the current evidence does not support adopting a ketogenic diet as a replacement for standard treatment, nor does it justify deferring treatment in favour of dietary intervention. The ERGO2 trial found no survival benefit; the GLAD trial showed feasibility but not efficacy. Any dietary modification should be discussed with your neuro-oncology team, who can assess your individual situation and nutritional status.
Reducing glycaemic load is broadly supported by the cancer metabolism evidence base. The metabolic rationale for limiting refined carbohydrates and sugar — regardless of full ketosis — is consistent with the broader literature on insulin, IGF-1, and cancer risk. You do not need to pursue strict ketosis to reduce the metabolic signals most associated with cancer promotion. A Mediterranean dietary pattern, which naturally limits refined carbohydrates and ultra-processed foods, captures much of the metabolic benefit with substantially more evidence behind it.
The absence of evidence of benefit is not evidence of absence. The ERGO2 trial was small and conducted in a very sick population facing a disease with limited treatment options. A larger, well-powered trial in a different glioma population — or in a cancer type with different metabolic characteristics — might yield different findings. The field is genuinely open.
Be sceptical of overclaiming. The ketogenic diet has attracted significant popular interest as a cancer treatment, and claims circulating online often go well beyond what the trial evidence supports. The ERGO2 and GLAD trials are the benchmark — not anecdotes, not cell culture data, not animal studies presented as human proof.
What the evidence does not show
The evidence does not show that the ketogenic diet prevents cancer in healthy individuals. No randomised trial has evaluated this question, and there are potential long-term considerations with strict ketogenic diets — including effects on cardiovascular lipid profiles, gut microbiome diversity, and bone density — that require weighing.
It does not show that ketogenic diets are ineffective in all cancer types. ERGO2 studied recurrent glioblastoma specifically; results may differ in other tumour types, treatment contexts, or patient populations.
And it does not resolve the optimal approach to dietary carbohydrate restriction in cancer: strict ketosis versus low glycaemic index versus Mediterranean-style carbohydrate moderation remain distinct interventions with distinct evidence bases.
The bottom line
The two most rigorous clinical trials of dietary carbohydrate restriction in brain cancer — ERGO2 and GLAD — tell an honest and mixed story. The ketogenic diet did not improve survival in recurrent glioblastoma. A low glycaemic, anti-inflammatory dietary approach was feasible and metabolically active but was not powered to demonstrate clinical benefit. The hypothesis underlying these interventions remains scientifically credible, and larger trials are needed. For now, the ketogenic diet in cancer sits in the category of promising but preliminary — a position the current evidence genuinely supports, and which requires neither dismissal nor exaggeration.
Scientific references
- Klement RJ, et al. Ketogenic diets in medical oncology: a systematic review with focus on clinical outcomes. Medical Oncology. 2020. PMID: 32619561. DOI: 10.1007/s12032-020-1337-2
- GLAD diet Phase 2 trial in newly diagnosed glioma. Neurology. 2021. PMID: 34233941
- Upadhyay M, et al. Metabolic alterations in cancer cells and the emerging role of oncometabolites as drivers and therapeutic targets. Pharmacology & Therapeutics. 2012. DOI: 10.1016/j.pharmthera.2012.11.003
This article is written for educational and informational purposes only. It is not intended as medical advice and should not be used to diagnose, treat, prevent, or cure any medical condition, including cancer. If you have a cancer diagnosis or are undergoing treatment, please discuss any lifestyle or dietary changes with your oncologist or medical team before making them. The studies cited reflect current research at the time of writing; the field is evolving rapidly and findings may be updated or superseded. Chloe Archard is a nutritional health coach, not a medical doctor.
