Glucose ketone index (GKI): what it is and how it is measured
GKI is blood glucose divided by blood ketones. Where it came from, when two papers say to measure it, and why Lomba treats no value as a target.
By Lomba 4 min read
The glucose ketone index, or GKI, is one number made from two blood readings: glucose divided by ketones. It comes from a 2015 paper by three researchers at Boston College, who built it as a tool for studies of ketogenic diets in brain cancer.
We read it, along with a 2024 paper that changes when the readings are taken. Neither paper settles whether those diets help patients, and this post makes no claim about that.
What the two papers are
- The 2015 paper
- A methods paper, not a trial. It defines the index and applies it to glucose and ketone values from earlier studies: mice with brain tumors, and three human patients from two earlier reports.
- The 2024 paper
- A "Debate" article by 49 authors that proposes a research protocol for glioblastoma, a brain cancer. It reports no new data.
- What is measured
- Blood glucose and blood beta-hydroxybutyrate, the main ketone in blood, both in mmol/L.
- Who paid
- The 2015 work was supported in part by the US National Institutes of Health, the American Institute of Cancer Research and Boston College, and its authors declare no competing interests. The 2024 paper lists no funding and thanks several foundations for support. Four of its authors declare ties, such as patents on the use of exogenous ketones and consulting work.
Four takeaways
It is a ratio, so low ketone readings move it most.
The 2015 formula divides glucose in mg/dL by 18.016 to get mmol/L, then divides by ketones in mmol/L. With glucose at 5.0 mmol/L, ketones of 0.2 and 0.3 mmol/L give 25.0 and 16.7, while 2.0 and 2.1 give 2.5 and 2.4.
The two papers name different times to measure.
The 2015 paper says readings "should be measured" 2 to 3 hours "postprandial, twice a day if possible", which means after meals. The 2024 paper has its patients measure "at least twice daily: after the overnight fast (morning)", and 1 to 2 hours "prior to the last meal (evening)", or once daily, "maintaining consistency across measurements". It adds that "A morning fasted reading can be misleading", since ketones can show up after an overnight fast in healthy people too.
The 2024 paper wants blood readings, not urine.
The 2024 paper says "Urinary ketones are not accurate for GKI calculation". It prefers continuous glucose and ketone monitors for research and "during the learning phase".
The thresholds are predictions for brain cancer patients.
The 2015 paper says the "zone of metabolic management is likely entered with GKI values between 1 and 2 for humans", and that "Further studies will be needed to establish the validity of the predicted zone of management". For one of its three patients, ketones were measured in urine. The 2024 paper still calls its own range "proposed" and says "more research is needed to establish causal links between quantifiable metabolic changes and therapeutic outcomes".
Glucose 5.0 mmol/L divided by ketones 0.5 mmol/L is 10.0. Glucose 8.0 with ketones 0.8 gives 10.0 too, so the number alone does not say what the two readings were.
The 2015 paper asks for readings 2 to 3 hours “postprandial, twice a day if possible”. It does not say after which meals, so each meal has a window here.
The main takeaway
GKI was built for one purpose, in one group of patients: following ketogenic diets in brain cancer research.
The 2015 paper's thresholds are predictions from mouse studies and three human cases, and the 2024 paper's range comes from expert opinion. Both say the validation is still to come, and neither tests the thresholds in anyone else.
So the number is best read as a description: how your two readings stand against each other, set beside your own earlier readings taken the same way.
What Lomba does with that
Lomba shows GKI as a description. You type in a glucose and a ketone reading from your meter, glucose in either unit, and Lomba works out the index. For mg/dL it divides by 18, where the paper divides by 18.016, a difference of under 0.1%.
You can mark each reading as fasted, before a meal or after one. The chart shows every reading over time, and its line follows the first fasted reading of each morning. Beside the latest reading Lomba shows your usual: the middle half of your counted mornings in the last 90 days.
It has limits, and we would rather say them:
- No target. Lomba shows the index as "a descriptive ratio, not a target". Its chart draws five bands behind the readings. The bands are a common convention for reading GKI, not a published or validated scale, and not thresholds for any condition.
- Mornings only. The line and your usual use fasted morning readings. The 2024 paper also asks for an evening reading, which Lomba keeps as a dot, not a line.
- Not medical advice. What Lomba shows is a pattern in your own days, not proof of cause, and not medical advice. What a GKI value means for your health is a conversation with your doctor.
Sources
- Meidenbauer JJ and others. The glucose ketone index calculator: a simple tool to monitor therapeutic efficacy for metabolic management of brain cancer. Nutrition & Metabolism, 11 March 2015.
- Duraj T and others. Clinical research framework proposal for ketogenic metabolic therapy in glioblastoma. BMC Medicine, 5 December 2024.
The two papers share two authors. Neither is a trial: the first is a methods paper, the second a proposal for research.