Continuous measurement of ketone levels with a wearable device is feasible, new research suggests.
A small sensing device worn on the upper arm, similar to a continuous glucose monitor (CGM), effectively measured interstitial fluid ketone levels for 14 days in 12 healthy individuals who followed low-carbohydrate diets.
This first-ever human study of a continuous ketone monitor (CKM) was published online April 9 in the Journal of Diabetes Science and Technology by Shridhara Alva, PhD, of Abbott Diabetes Care, Alameda, California and colleagues.
The sensor, which uses wired enzyme electrochemistry technology like that used in CGMs, was based on β-hydroxybutyrate dehydrogenase chemistry. It requires a one-time calibration for 14 days of wear. Overall, the sensor performed well against capillary ketone values, the authors report.
Continuous ketone sensing of interstitial fluid offers advantages over measuring urine or blood ketone levels using strip-based technology, as those only provide episodic information that confirms ketosis, or diabetic ketoacidosis (DKA), after they’re already ongoing, the researchers say.
“Early identification of production of ketones may warn of impending ketoacidosis that could reduce the complications of DKA and perhaps even prevent it. Real-time continuous ketone monitoring could also help clinicians manage ketoacidosis,” they point out.
And, they say, a CKM could also be useful for people who are on low carbohydrate diets, in which the replacement of carbohydrate with fat prompts the body to use ketones instead of glucose for energy, leading to the metabolic state of ketosis. In such situations, a CKM device “may serve as a tool to monitor the effectiveness of their diet and indicate the effect of diet or exercise on the ketone levels,” the authors write.
In an accompanying commentary, Jennifer Y. Zhang, BA, of the Diabetes Technology Society, Burlingame, California and colleagues say that CKM fits in with the growing trend of wearable sensors that collect and store health data, particularly among people with diabetes.
“CKMs have the potential to enrich the idea of comprehensive self-monitoring. As this idea continues to expand with an ever-increasing array of new devices, people with diabetes and those looking to optimize their metabolic health in general will have a highly precise and comprehensive set of unobtrusive tools to reach their goals,” the editorialists write.
The editorialists point out some additional potential benefits of CKMs for people with diabetes, including:
Minimizing the risk for DKA among those with diabetes who have had frequent DKA hospitalizations
For sick day management in those with type 1 diabetes
For those using sodium glucose co-transporter (SGLT-2) inhibitor therapy, which has been linked to euglycemic DKA
However, they also note several challenges, including the need to gather data from people with dynamically changing blood glucose and ketone levels, and particularly where impending DKA might be expected, such as prolonged fasting or fasting combined with exercise.
The editorialists also suggest that future studies might include venous plasma measurements as a reference, rather than fingersticks. Moreover, they say, the technology will need to be tested among individuals hospitalized with DKA and during the process of resolution.
And they note, just as with CGM currently, cost and reimbursement will need to be worked out.
CKM Stacks Up Well Against Fingerstick Ketone Measurements
Of the 12 study subjects, one had type 1 diabetes and the other 11 did not have diabetes. All followed a low carbohydrate (ketogenic) diet throughout the study. None were taking SGLT2 inhibitors.
The sensor system was similar to that of the FreeStyle Libre CGM, and the applicator was identical. But instead of measuring glucose, this sensor measures β-hydroxybutyrate dehydrogenase in the subcutaneous tissue, transfers the ketone signal to the receiver, and the data is then uploaded to a computer.
Each study subject wore multiple sensors on the backs of both upper arms for up to 14 days and performed eight daily fingerstick measurements using ketone test strips. Out of 36 ketone sensors and 12 background sensors tested, 31 ketone sensors and 11 background sensors had evaluable data for a total of 3132 paired data points over the 14 days. Reference measurements ranged from 0-5.1 mM, with a median of 0.6 mM.
For reference ketone concentrations less than 1.5 mM, the overall mean absolute difference was 0.129 mM, with 83.4% of sensor results within ± 0.225 mM and 91.7% of the sensor results within ± 0.3 mM of the reference.
For reference ketone levels of 1.5 mM or greater, the overall mean absolute relative difference was 14.4%, with 76% of the sensor results within 20% and 89.7% of the results within 30% of the reference.
Overall, 82.4% of the values were within 0.225 mM/20% and 91.4% within 0.3 mM/30% of the reference.
The authors of the study conclude: “A prospective calibrated system is required for real time monitoring of ketones. However, this first human study suggests that a continuous ketone sensor similar to continuous glucose sensors is achievable.”
Alva and coauthors Hyun Cho and Junli Ou are employees of Abbott Diabetes Care, which funded the study. Co-author Kristin Castorino conducts research sponsored by Medtronic, Dexcom, Abbott, Insulet, and Novo Nordisk, and has been a speaker for Dexcom and Abbott. Zhang has disclosed no relevant financial relationships.
Miriam E. Tucker is a freelance journalist based in the Washington, DC area. She is a regular contributor to Medscape, with other work appearing in the Washington Post, NPR’s Shots blog, and Diabetes Forecast magazine. She can be found on Twitter @MiriamETucker.