Abstract Body

Living with type 1 diabetes (T1D) is challenging as it requires intensive monitoring of glucose levels, nutritional intake and physical activity, and correct titration of insulin in order to obtain near-normal glucose levels. The Endocrine Society has proposed real-time continuous glucose monitoring (RT-CGM) as the gold standard for people with T1D. RT-CGM devices display interstitial glucose levels around the clock and are designed to set off alarms to warn people when glucose levels are trending too high or too low. The evolution in pump and CGM technology has led to the development of hybrid closed loop (HCL) systems where basal insulin delivery is automatically guided by sensor glucose values using an algorithm. CGM and HCL systems have demonstrated improvements in HbA1c, time spent in hypoglycaemia, hospitalisations for severe hypoglycaemia or ketoacidosis, and quality of life. However, there are still some shortcomings with currently available CGM devices.

Firstly, currently available sensors rely entirely on continuous glucose measurements and do not provide an alert for high ketone levels or impending diabetic ketoacidosis (DKA). Monitoring ketones is also advised for sick-day management, but in reality many at-risk patients do not have ketone test strips at home. Production of ketone bodies may occur as a result of insulin deficiency (e.g. in case of pump failure or inadequate bolus dosing), sickness, insufficient intake of carbohydrates (very low calorie diet), or sodium-glucose co-transporter-2 inhibitors (SGLT2-i) therapy. Continuous ketone monitoring (CKM) may facilitate earlier detection of ketones, thereby possibly reducing hospitalisations for DKA in high-risk patients.

The first-in-human results obtained in 12 volunteers of a CKM device were published in 2021 by Alva et al. The electrochemical sensor used wired enzyme to measure β-hydroxybutyrate (BHB), the major pathologic analyte. This sensor delivered a linear response over the 0-8 mM range with good accuracy and stability, both in vitro and in vivo, for 14 days. With a single retrospective calibration the mean absolute difference (MAD) for BHB concentrations <1.5 mM was 0.129 mM and 91.7% of the sensor results were within ±0.3 mM of the reference. For BHB ≥1.5mM the mean absolute relative difference (MARD) was 14.4%.

Teymouran et al. reported data of a new real-time CKM microneedle platform based on the electrochemical monitoring of BHB alongside with glucose. This sensor detects BHB based on the NAD-dependent dehydrogenase enzyme and a selective low-potential fouling-free anodic detection of NADH using an ionic liquid-based carbon paste transducer electrode. In vitro data showed that the sensor had a high sensitivity (with low detection limit, 50 µM), high selectivity in the presence of potential interferences, along with good stability. The BHB microneedle sensor has been coupled with an oxidase-based glucose microneedle sensor on the same array platform, leading to an attractive sensor array towards the simultaneous real-time continuous monitoring of both glucose and ketones. The ability to detect lactacte has also been demonstrated based on lactate oxidase catalyzed lactate oxidation to pyruvate.

A third sensor that has been developed by Indigo Diabetes nv is an implantable continuous multimetabolite sensor monitoring glucose, BHB, and lactate using near-infrared (NIR) spectroscopy technology with an expected lifetime of 2 years. In a first-in-man study, exploratory data on accuracy were promising (95.6 % of all data points for glucose ranging between 40-400 mg/dl were located in zone A, a MAD of 10.5 mg/dl for values between 40 and 70 mg/dl and a MARD of 10% for values between 70-180 mg/dl and of 4% for values >180 mg/dl were observed). Administration of paracetamol, acetylsalicylic acid, ibuprofen, sorbitol, caffeine, fructose, aspartame and vitamin C did not significantly influence the accuracy of glucose measurements. Also for BHB good accuracy was observed. Continuous measurement of ketones was compared to blood strip measurements over a physiological range of 0-4 mM; the sensor showed a MAD of 0.19 mM. The lactate concentrations measured over a range of 0-20 mM showed an MAD of 0.53 mM in relation to Biosen EKF reference measurements in blood.

A second area of concern relates to the design of current sensors, patient’s experiences and costs. The short sensor lifespan, the likelihood of accidental sensor dislocation, the occurrence of skin reactions, and privacy reasons (keep their diabetes hidden), limit the implementation of these sensors.

The Indigo Diabetes nv sensor is a miniaturized near-infrared spectrometer on a silicon photonics chip that measures optical transmittance in the interstitial fluid at up to 24 wavelengths between 1680 and 2400 nm. The sensor is covered by a biocompatible silicone envelope. It is implanted subcutaneously in the abdominal region and has an expected lifespan of 2 years or more.

In summary, there is a compelling need for a patient tailored device that is implantable, has a long lifespan and continuously monitors multiple biomarkers thereby helping to prevent episodes of hypoglycaemia or ketoacidosis under all circumstances (exercise, illness, SGLT2i therapy, very low carb diet), and possibly increasing quality of life.