The glucose ketone index: a compelling mechanistic hypothesis on the path to clinical validation

Key points

• The glucose ketone index (GKI) is an elegant, hypothesis-generating measure of systemic metabolic state that extends beyond glucose alone by relating circulating glucose to β-hydroxybutyrate.
• The GKI has been proposed as a quantitative biomarker for monitoring nutrition-based interventions in chronic diseases and cancer, but it requires validation in terms of clinical application, along with adjusted clinical thresholds.
• The priority is prospective, disease- and stage-stratified validation within a broader personalized metabolic phenotype; meanwhile, public health guidance should emphasize individualized metabolic balance rather than the indiscriminate pursuit of ketosis.

Introduction

In their recent lead article in Frontiers in Science, Lee et al. make a valuable contribution to metabolic oncology research by advancing the glucose ketone index (GKI)—the ratio of circulating glucose to β-hydroxybutyrate—as a practical measure of systemic metabolic state (1). Their hypothesis is grounded in altered glucose metabolism, including aerobic glycolysis, and mitochondrial dysfunction (2). The Warburg effect is among the best-established phenomena in tumor biology, while mitochondrial dysfunction is now recognized as a genuinely convergent feature across malignancy, metabolic disease, cardiovascular disease, and neurodegeneration (2). Because the GKI can be measured inexpensively and repeatedly at the bedside, it could offer an accessible way to track systemic metabolic changes in patients, adding to the relatively limited range of oncology biomarkers that can be readily monitored. New approaches to precision medicine could use the GKI as a suitable parameter for validation in n-of-1 trials.

A pragmatic starting point, not a finished tool

Lee et al. propose a GKI target of approximately 1.0 (1). This is a principled estimate worth validating: it was derived by applying the glucose:ketone ratio to a carefully assembled set of published brain tumor cases and selecting the point at which glucose and ketone concentrations, measured in mmol/L, are numerically equal. This provides a biologically intuitive anchor for testing metabolic pressure on glucose-dependent tumor cells; however, it should not yet be treated as an established optimum. The distinction between mechanistic plausibility and clinical validation is particularly important because metabolism is one of the most dynamic physiological systems in human biology, especially in diseases marked by profound metabolic alterations, such as cancer. Therefore, trials must refine and, where necessary, recalibrate the proposed threshold against clinically meaningful outcomes across disease contexts, stages, and populations.

Educating without overreach

What matters most at this stage is how the concept is communicated beyond the research community. Currently, the GKI is a research and hypothesis-generating tool, not a validated clinical target, and certainly not a lifestyle prescription for extending healthspan or targeting aging. This distinction matters because the underlying biology is more intricate than a single favorable number suggests. Recent studies have shown that some cancer cells can reroute β-hydroxybutyrate through a non-canonical cytosolic pathway to accelerate biosynthesis (3). At the same time, ketolysis has been shown to drive effector function in CD8⁺ T cells via histone acetylation, while β-hydroxybutyrate enhances the metabolic fitness of chimeric antigen receptor (CAR) T cells, improving their proliferation and antitumor activity (4, 5). This suggests that a genuinely exciting therapeutic angle may operate through the immune compartment rather than through tumor suppression alone.

In clinical practice, if ketone bodies can restrain some tumor cells while fueling others and simultaneously sharpen antitumor immunity, the clinically meaningful question is not simply whether the GKI is low enough but also what level is appropriate for a particular patient and disease stage given the potentially different effects across tissue compartments.

Interpreting the GKI in context

Nutritional ketosis may be induced through diverse approaches, including individualized carbohydrate reduction, time-restricted eating, intermittent fasting, structured exercise, reduction in visceral adiposity, or a combination of these, without necessarily adhering to a traditional ketogenic diet. Importantly, the growing public interest in ketogenic diets and prolonged fasting has highlighted the need for improved health literacy and careful medical supervision. In vulnerable populations, particularly patients with advanced cancer, cachexia, or sarcopenia, unsupervised fasting or excessive caloric restriction may further accelerate the loss of lean body mass and worsen clinical outcomes, although some reports describe the ketogenic diet as muscle-sparing and anti-inflammatory (6). Similarly, the indiscriminate adoption of high-fat ketogenic diets by patients with obesity, metabolic syndrome, dyslipidemia, or established cardiovascular disease may inadvertently increase saturated fat intake and adversely affect cardiometabolic risk if nutritional quality and lipid composition are not carefully individualized.

Multiple dietary patterns can produce nutritional ketosis while differing substantially in caloric intake, protein composition, fatty acid profile, micronutrient density, fiber content, and inflammatory effects. Consequently, identical ketone concentrations may emerge from profoundly different nutritional exposures with potentially distinct biological consequences. Conversely, similar GKI values may arise under entirely different physiological conditions. A physically active individual with preserved skeletal muscle mass, excellent insulin sensitivity, and intermittent fasting may have a metabolic profile that is fundamentally different from that of an individual with advanced malignancy, cachexia, or malnutrition, despite having comparable glucose and ketone concentrations.

Accordingly, neither ketosis itself nor the GKI should be interpreted independently of the broader metabolic phenotype. The principal contribution of the GKI may therefore lie not in defining universal metabolic targets, but in supporting precision metabolic medicine.

From hypothesis to a validated, personalized tool

The next encouraging step is disease-specific validation. The promise of the GKI will be realized not by defending a single, universal threshold, but by testing it prospectively, disease by disease and stage by stage, against outcomes that matter to patients. Precision medicine increasingly treats glucose and ketone metabolism as nodes within a broader, individualized phenotyping framework rather than as isolated parameters (7, 8). High-performance, AI-supported medicine also makes it increasingly realistic to track such multidimensional signatures longitudinally rather than as single snapshots (9). Because metabolism is inherently dynamic—responding to aging, treatment, physical activity, circadian biology, microbiome composition, and environmental exposures—the optimal metabolic intervention may itself require continuous recalibration.

A productive research agenda would first define the intended clinical use of the GKI, such as treatment monitoring, prognosis, or metabolic risk stratification, and then prospectively establish disease-specific ranges. These would almost certainly differ among glioblastoma, other solid tumors, and non-oncologic metabolic conditions. The GKI should be interpreted alongside broader phenotyping, including at least body composition, metabolomics, and cognitive and nutritional status, and tested against survival, progression, and quality-of-life outcomes rather than assumed a priori. Studies should also standardize how and when the GKI is measured and determine whether it adds clinically useful information beyond glucose and ketone concentrations alone.

N-of-1 and adaptive trial designs could be valuable for testing short-term metabolic responses and individual GKI trajectories. These could provide an initial basis for determining which findings should then be evaluated across large populations. Prospective and randomized studies will also be necessary to establish relationships with longer-term clinical outcomes. Subsequently, adaptive metabolic interventions could be individualized according to continuously evolving biological data rather than predefined dietary algorithms.

Framed this way, the path forward extends Lee et al.’s proposal of the GKI as a potential tool for treating cancer and other age-related diseases. Rather than serving as a stand-alone biomarker of disease state, the GKI may ultimately become one component of a broader panel of personalized clinical tools.

Conclusion

The GKI is a genuinely promising, hypothesis-generating advance over measuring glucose alone, and Lee et al. deserve credit for proposing a concrete, testable starting target rather than leaving the concept purely theoretical. The task ahead is not to discard that number but rather to earn it through prospective, disease-stratified validation trials that convert a compelling mechanistic hypothesis into clinically meaningful, context-specific thresholds. In the meantime, clinicians and the public should understand that the GKI belongs in the research setting and is not yet ready for general health guidance.

Generative AI statement

The author declared that generative AI was used in the creation of this manuscript. The author used ChatGPT 4.0 and Claude 5.0 to arrange the reference list and proofread the English language.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited and reviewed by: Adrienne C. Scheck, University of Arizona, United States