Reliable and safe hypoglycemic episodes using the automated glucose clamp technique

Hypoglycemia represents a major challenge in the management of diabetes. Therefore, the accuracy of continuous glucose monitoring (CGM) systems in the hypoglycemic range is of particular interest to enable systems to reliably detect low blood glucose (BG) levels and prevent severe hypoglycemia.

In clinical studies, hypoglycemic events often occur infrequently and randomly making standardized evaluation protocols for CGM systems difficult. Therefore, it has been proposed that at least 7.5% of data points in performance studies should fall within the hypoglycemic range (defined as < 70 mg/dL) [1], to ensure sufficient representation of this clinically relevant zone.

To date, delayed and elevated insulin administration in combination with meals have been commonly used to induce hypoglycemia in clinical studies. While this method is relatively simple to implement, the resulting glucose profiles are highly dependent on meal composition, insulin dosage and timing, which limits reproducibility and standardization.

We therefore investigated if hypoglycemia could be safely and reliably induced with the automated glucose clamp technique.  

Methods

Data of two different clinical studies conducted using the ClampArt device were analyzed with respect to manageability and safety of intentionally induced hypoglycemic episodes.

Both studies included participants with Type 1 diabetes. In the first study, six subjects underwent two clamp procedures each. The second study involved eleven subjects, of whom ten participated in two and one in a single clamp experiment.

Both studies involved varying target levels of the participants’ BG concentrations. To achieve this, the ClampArt device was programmed to variable glucose target levels. The algorithm responsible for the automatic regulation of BG concentration calculated the necessary glucose infusion rates accordingly. During the clamp sessions, different intravenous insulin dosages were administered continuously to reduce BG levels as required and generate the option of regulating the glucose concentration (figure 1).

The design of both studies included episodes of hypoglycemic glucose concentrations. In the first study, there was one episode in each clamp with a glucose target of 60 mg/dL, lasting approximately 30 minutes. In the second study, two hypoglycemic targets were established, first at 60 mg/dL followed by 50 mg/dL. In case of intolerable symptoms or safety concerns, BG was immediately increased to levels outside the hypoglycemic range.

Figure 1: Example of clamp data with variable insulin infusion

In order to analyze the hypoglycemic episodes, we separated each episode and evaluated several characteristics. In figure 2 an exemplary clamp from study 2 with two hypoglycemic episodes with 60 mg/dL and 50 mg/dL is shown.

Figure 2: Example of clamp data with glucose infusion rates and marked hypoglycemic episodes

Results

In total we analyzed 30 episodes with a target level of 60 mg/dL and 20 episodes with a target of 50 mg/dL.

Table 1: Quality data of hypoglycemic events

During the hypoglycemic episodes, participants were continuously monitored by clinical staff. As the low BG concentrations were intentionally induced, associated hypoglycemic symptoms were not classified as adverse events. We analyzed the frequency and duration of unintentionally low BG levels as safety marker for the hypoglycemia management through the ClampArt system.

Table 2: Safety data of target level 60 mg/dl

In table 2 the results for hypoglycemic episodes with the target level of 60 mg/dL are shown. In 7 experiments, BG concentration fell below 54 mg/dL. None of the clamp procedures were terminated prematurely due to intolerable symptoms or other health-related complications.

Table 3 presents data on hypoglycemic episodes with the BG target level of 50 mg/dL. In none of these episodes were BG concentrations lower than 40 mg/dL measured.

Table 3: Safety data of target level 50 mg/dl

Prematurely terminated episodes were identified in cases where glucose infusion rates were manually increased despite no change in the predefined BG target level. In 4 instances, such premature increases in glucose infusion rates were observed. Nonetheless, minute-by-minute monitoring of BG concentrations did not reveal any critical deviations. Clinical staff responded proactively based on clinical signs, adjusting glucose infusion rates as a safety measure. Accordingly, continuous monitoring is crucial for subjects exhibiting values in these low BG concentration ranges.

Conclusion

The use of the ClampArt device represents a sophisticated method for inducing hypoglycemic episodes in a controlled and unbiased manner. For more details browse out Poster from DTM 2025.

Interested in evaluating hypoglycemic episodes in your clinical trial? Contact us now and get a free consultation.

[1] Eichenlaub M, Pleus S., Rothenbühler M., et al. Comparator Data Characteristics and Testing Procedures for the Clinical Performance Evaluation of Continuous Glucose Monitoring Systems, Diabetes Technology & Therapeutics 2024; 26(4): 263-275