Low-Dose Dasiglucagon Versus Oral Glucose for Prevention of Insulin-Induced Hypoglycemia in People With Type 1 Diabetes: A Phase 2, Randomized, Three-Arm Crossover Study

Near-normalization of blood glucose levels through intensive insulin therapy reduces the risk of diabetes late complications in individuals with type 1 diabetes, but the approach carries an inherent risk of inducing hypoglycemia (1,2). Data show that most people with type 1 diabetes experience nonsevere (<3.9 mmol/L [70 mg/dL]) hypoglycemia on a near-daily basis (3). For many individuals, recurrent episodes of nonsevere hypoglycemia can have negative physiological, emotional, and socioeconomic consequences and thus pose a significant challenge for optimal health management (4–6).

The American Diabetes Association currently recommends reversing nonsevere hypoglycemia by consuming 15–20 g of oral glucose (7). Although generally effective, research data suggest that for some people, reversal of hypoglycemia through food consumption may lead to undesirable effects. Nonsevere hypoglycemia has been shown to induce a significant craving of particularly high-carbohydrate content food, which if consumed in excess, may elicit rebound hyperglycemia and affect long-term glycemic control (8). In addition, research shows that excessive carbohydrate consumption caused by hypoglycemia may increase body weight (9). Meanwhile, epidemiological data show that more than half of the adult population with type 1 diabetes is overweight or obese and that the increasing prevalence has been rising faster than in the general population (10,11). As excess body weight is associated with increased risk of cardiovascular disease, lower quality of life, and other adverse health outcomes, efforts to prevent and reduce its occurrence are essential (12). Consequently, an alternative noncaloric method of managing manifested or impending nonsevere hypoglycemia has been warranted.

In recent years, studies have successfully demonstrated that subcutaneous (s.c.) low-dose glucagon can be used to effectively treat nonsevere hypoglycemia in people with type 1 diabetes (13,14). However, because the available glucagon preparations until recently required reconstitution before injection, applicability of low-dose glucagon has been limited to research settings. Dasiglucagon (Zealand Pharma, Søborg, Denmark) is a novel, soluble, ready-to-use glucagon analog that does not require reconstitution prior to use and, therefore, has the potential to transition the concept into clinical practice.

The objective of this study was to compare the efficacy of low-dose dasiglucagon with oral glucose—of the same magnitude as usually recommended—for prevention of insulin-induced hypoglycemia in people with type 1 diabetes using insulin pump or multiple daily injection (MDI) therapy.

The study was a randomized, phase 2, partially single-blind, three-arm, crossover trial enrolling 20 participants with type 1 diabetes, with 10 using insulin pumps and 10 using MDI therapy. Participants attended a screening visit, a ∼2-week insulin optimization run-in period, and three ∼6-h study visits, each separated by a washout period of ≥3 days.

The study was conducted at the Steno Diabetes Center Copenhagen Clinical Research Unit, Gentofte, Denmark. It was monitored by the Good Clinical Practice Unit at Bispebjerg and Frederiksberg Hospital, Copenhagen, Denmark, and approved by the Danish Medicines Agency (EudraCT: 2020-000551-12), the Regional Committee on Health Research Ethics (H-20013256), and the Danish Data Protection Agency (P-2020-219). The study was registered at ClinicalTrials.gov (NCT04449692) and conducted in accordance with the Declaration of Helsinki.

Participants were recruited from the outpatient diabetes clinic at Steno Diabetes Center Copenhagen. Key inclusion criteria were age 18–64 years, duration of type 1 diabetes ≥3 years, use of insulin pumps or MDI therapy for ≥6 months, glycated hemoglobin (HbA_1c) level ≤8% (64 mmol/mol), and current use of insulin aspart (Novo Nordisk, Bagsværd, Denmark). Key exclusion criteria were pregnancy or inadequate use of contraception, allergy or intolerance to glucagon or glucagon-like products, use of medications affecting glucose metabolism (other than insulin), and concomitant medical or psychological conditions making the individual unsuitable for study participation.

After providing oral and written informed consent, participants completed a screening visit for assessment of the eligibility criteria. Procedures included routine blood sampling, physical examination, review of medical history and medications, and registration of baseline characteristics (age, sex, treatment duration and modality, daily insulin dose, body weight, height, and blood pressure).

The order of the three study visits was determined by stratified permuted block randomization using random blocks of three and six and stratification by treatment modality. An allocation table generated by sealedenvelope.com was uploaded to the Research Electronic Data Capture (REDCap, hosted at the Capital Region of Denmark) system by a person not otherwise involved in the study. The randomization functionality of REDCap was used to randomize participants 1:1:1 to one of the three visit sequences after ensuring that all eligibility criteria were met. Participants were blinded to the dose of dasiglucagon but not blinded for the oral glucose intervention.

Upon enrollment, participants completed a ∼2-week run-in period to ensure that their basal insulin requirements and insulin sensitivity factors (ISF) were optimized before the first study visit. Participants were instructed to refrain from ethanol consumption and strenuous exercise for 24 h before each study visit, to consume ≥150 g of carbohydrates daily for 3 days before each study visit, and to avoid hypoglycemia (blood glucose <3.9 mmol/L [70 mg/dL]) 24 h prior to each study visit. To help avoid hypoglycemia, a Dexcom G6 continuous glucose monitor (Dexcom, San Diego, CA) was inserted into the abdominal subcutaneous tissue 24–48 h before each study visit, and the device was set to alert at glucose levels <4.5 mmol/L (81 mg/dL). Lastly, participants were instructed not to change the basal insulin rate (insulin pump users) or administer bolus insulin during the last 4 h before arrival. MDI users administered their long-acting insulin as usual.

Participants attended three in-clinic visits in random order. At each visit, participants arrived at the research facility at 8:00 a.m. after fasting overnight. On arrival, an antecubital intravenous catheter for blood sampling was inserted and covered by a heating pad. After collection of initial blood samples, a s.c. insulin bolus (insulin aspart, Novo Nordisk) was administered using the participants’ usual treatment modality (i.e., insulin pump or injection pen). The insulin dose was calculated based on the starting plasma glucose (PG_baseline) and each individual’s ISF and aimed at lowering PG to 3.0 mmol/L (54 mg/dL) ([PG_baseline − 3.0]/ISF) (15). When the PG reached 4.5 mmol/L (81 mg/dL), 80 µg s.c. dasiglucagon (D₈₀), 120 µg s.c. dasiglucagon (D₁₂₀), or 15 g oral glucose (CHO) from dextrose tablets (Dextro Energy GmbH & Co. KG, Krefeld, Germany) was administered (t = 0). The two doses of dasiglucagon were selected based on previous glucagon efficacy studies as well as previous characterizations of the pharmacokinetic and pharmacodynamic profile of dasiglucagon (16–19). Dasiglucagon was administered by the investigator as a s.c. injection into a lifted skinfold of the abdominal wall using a 1-mL syringe.

PG was measured on a YSI 2900 STAT (Yellow Springs, OH) every 5 minutes for the following hour and every 10 min for the subsequent 2 h before concluding the study visit (t = 180). Participants experiencing a PG <3.0 mmol/L (54 mg/dL) received 15 g oral glucose (dextrose tablets) rescue treatment. Adverse effects (nausea, headache, stomachache, palpitations, and injection site pain) were scored using a 0–100 visual analog scale (VAS) just prior to the intervention (t = 0) and 3 h later (t = 180) to evaluate whether any adverse events had occurred within 3 h after the intervention. Serum insulin and plasma dasiglucagon were measured at t = 0, 5, 15, 25, 40, 60, 90, 130, and 180 min following the intervention using a Cobas 600 analyzer (Roche Diagnostics GmbH, Mannheim, Germany) and Symbiosis Pro system (Spark Holland, Emmen, the Netherlands) in conjunction with a Xevo TQ-S mass spectrometer (Waters Corp., Milford, MA), respectively.

The primary end point was time below target PG range (<3.9 mmol/L [70 mg/dL]) from 0–180 min after the intervention. Key secondary end points included incidence of level 1 hypoglycemia (<3.9 mmol/L [70 mg/dL]), time to PG increase of 1.1 mmol/L (20 mg/dL), and incidence of level 2 hypoglycemia (<3.0 mmol/L [54 mg/dL]). Other PG end points were time in and above target glucose range, incidence of hyperglycemia, total area under the glucose curve (tAUC_glucose), time to peak PG, time from intervention to oral glucose rescue treatment (<3.0 mmol/L [54 mg/dL]), as well as average, peak, incremental peak, and nadir PG levels. Insulin end points included baseline and preintervention (t = 0) insulin concentration, baseline insulin bolus, and tAUC_insulin, whereas dasiglucagon end points included peak concentration, time to peak concentration, and tAUC_dasiglucagon. Safety outcomes were the baseline-adjusted VAS score of adverse effects (nausea, headache, stomachache, and injection site pain) and incidence rate of vomiting.

The sample size of 20 participants was calculated based on previous data (16,20) and used the following assumptions for the primary outcome (time <3.9 mmol/L [70 mg/dL]): SD (between-within subjects) of 5.4 min (equal to 3%), 80% power, and noninferiority margin of 2.7 min (equal to 1.5%). In cases where oral rescue treatment was provided, the nadir PG concentration was carried forward to the end of the observation period (t = 180) in the data analysis. Missing PG values were extrapolated from the closest measurements using linear interpolation. Normally distributed continuous outcomes were compared using a repeated-measures ANOVA analysis, whereas the Friedman test was used to compare nonnormally distributed continuous outcomes. In cases where the overall three-way comparisons were significant, the comparisons between treatments (CHO vs. D₈₀ and CHO vs. D₁₂₀) were statistically tested. Incidence rates were compared using the McNemar test. Comparison of the primary end point between interventions were also adjusted for the treatment modality (i.e., insulin pump or MDI). Statistical calculations were performed in SAS 9.4 (SAS Institute, Cary, NC) and presented with GraphPad Prism 6.01 (GraphPad Software, San Diego, CA). P values <0.05 were considered statistically significant. Unless otherwise stated, results are presented as mean ± SEM or median (interquartile range).

Between July 2020 and January 2021, 20 participants (8 women) were enrolled in and completed the study, with 10 using insulin pumps and 10 using MDI therapy. One participant completed only two study visits due to work-related reasons but counted toward the required 20 participants as defined per protocol. Participants had a median (interquartile range) age of 47 (23–64) years, duration of type 1 diabetes of 19 (3–58) years, BMI of 24.7 (21.6–36.4) kg/m², total daily insulin dose of 35 (20–65) IU, and HbA_1c level of 6.8% (4.9–7.7; 51 [41–61] mmol/mol). All participants used insulin aspart, whereas the MDI users’ long-acting insulin were insulin degludec (n = 7), insulin glargine (n = 2), and insulin detemir (n = 1). None of the baseline characteristics were significantly different between the two groups (Supplementary Table 1).

The 3-h postintervention PG profile and key efficacy outcomes are shown in Fig. 1, individual PG profiles are shown in Supplementary Fig. 1, and the full list of PG outcomes are summarized in Table 1. Ten participants (50%) with CHO, five (26%) with D₈₀, and four (20%) with D₁₂₀ experienced hypoglycemia (<3.9 mmol/L [70 mg/dL]) during the study (CHO vs. D₈₀, P = 0.096; CHO vs. D₁₂₀, P = 0.034). Participants spent more time in hypoglycemia (<3.9 mmol/L [70 mg/dL]) after CHO administration (14%) compared with D₈₀ (7%) and D₁₂₀ (6%), although the nonnormal distribution and larger-than-expected variation meant that the three-way nonparametric comparison was not statistically significant (P = 0.273). Treatment modality (i.e., insulin pump vs. MDI) did not affect the time in hypoglycemia (P = 0.735). Four participants experienced level 2 hypoglycemia and received oral glucose rescue treatment after CHO, three after D₈₀, and two after D₁₂₀ (CHO vs. D₈₀, P = 0.564; CHO vs. D₁₂₀, P = 0.317); for those, the median time to carbohydrate rescue was 120, 160, and 110 min, respectively (P = 0.320). Both doses of dasiglucagon raised PG levels significantly faster than CHO: the median (95% CI) time from intervention to first increase in PG of 1.1 mmol/L [20 mg/dL] was 30 (25–50), 15 (15–20), and 15 (15–20) min for CHO, D₈₀ and D₁₂₀ (CHO vs. D₈₀, P = 0.006; CHO vs. D₁₂₀, P = 0.003). The peak PG concentration did not significantly differ between the three interventions, but as a result of the faster glucose-elevating profile of dasiglucagon, both dasiglucagon doses produced a larger tAUC_glucose compared with CHO (CHO vs. D₈₀, P = 0.030; CHO vs. D₁₂₀, P = 0.0008).

Serum insulin and plasma dasiglucagon levels are shown in Fig. 2, and outcomes are listed in Table 1. The serum insulin concentration prior to each intervention, the tAUC_insulin, and the dose of the initial insulin bolus did not differ within participants between the three interventions. D₁₂₀ produced a higher peak dasiglucagon concentration (P < 0.0001) and a larger tAUC_dasiglucagon (P < 0.0001) than D₈₀, whereas the time to peak concentration did not differ between the two doses (P = 0.721).

Baseline (t = 0) VAS scores are available in Supplementary Table 2. There were no statistically significant differences between the interventions in the baseline-adjusted VAS scores for nausea, headache, palpitations, stomachache, and injection site pain (Table 2). After the CHO, D₈₀, and D₁₂₀ interventions, 0%, 16%, and 10% of the participants experienced a clinically significant (≥15-point) (21) increase in VAS-measured nausea from baseline, respectively. However, this result was also nonsignificant (CHO vs. D₈₀, P = 0.250; CHO vs. D₁₂₀, P = 0.50). One participant experienced vomiting ∼2 h after the D₈₀ intervention. Local tolerability after dasiglucagon administration was good, with no cases of postinjection erythema or edema. In general, the adverse effects profile of dasiglucagon was consistent with the known profile of glucagon, and no serious or unexpected adverse events were observed throughout the study.

This study demonstrated that low-dose (80 and 120 µg) s.c. dasiglucagon safely and effectively prevented insulin-induced hypoglycemia in people with type 1 diabetes using insulin pump or MDI therapy. Administration of dasiglucagon resulted in significantly fewer cases of hypoglycemia, numerically less time in the hypoglycemic range, and a substantially faster glucose response compared with CHO. Dasiglucagon was generally well tolerated, with the most frequent adverse effect being mild and transient nausea in line with the previously reported adverse effects profile of glucagon.

Dasiglucagon was recently approved by the U.S. Food and Drug Administration for rescue treatment of severe hypoglycemia in individuals with diabetes aged ≥6 years (22). In three randomized, double-blind, phase 3 studies, administration of 0.6 mg s.c. dasiglucagon was safe and effective for treatment of severe hypoglycemia in both adult and pediatric populations with type 1 diabetes (23–25). The pharmacokinetic and pharmacodynamic properties of low-dose dasiglucagon during euglycemic and hypoglycemic conditions were characterized in a previous study (19). In that study, dose-dependent increases in PG levels after 0.03–0.6 mg s.c. dasiglucagon administration were demonstrated in insulin-pump treated individuals with type 1 diabetes under intravenous insulin infusion conditions. More recently, a randomized, short-term, outpatient study showed that multiday use of dasiglucagon as part of the iLet bihormonal artificial pancreas configuration was safe and effective for treatment of individuals with type 1 diabetes (26). However, until now, no studies have explored the use of injection-based low-dose dasiglucagon for treatment of impending s.c. insulin-induced nonsevere hypoglycemia.

There are a number of reasons to use low doses of glucagon to prevent and treat nonsevere hypoglycemia. In cases of hypoglycemia, rapid restoration of euglycemia and relief of hypoglycemic symptoms are key (27). The current study has shown that low-dose dasiglucagon induced a significantly faster glucose-response compared with guideline-recommended oral glucose. In real-world settings, individuals with type 1 diabetes often overtreat hypoglycemia by consuming excess carbohydrates, thereby causing subsequent hyperglycemia that can affect long-term glycemic control (3,28,29). In this trial, dasiglucagon produced a peak PG concentration similar to that of oral glucose while maintaining the rapid glucose-elevating profile. The observed real-life carbohydrate overtreatment of recurrent hypoglycemic episodes may, however, have effects beyond glycemic control. The Diabetes Control and Complications Trial found that a higher incidence of nonsevere hypoglycemia was significantly associated with increased weight gain independently of glycemic control (9). At the same time, the prevalence of overweight and obesity among individuals with type 1 diabetes continues to rise (10). As treatment with oral carbohydrate leads to consumption of extra—and for some, excess—calories, the approach can be counterproductive from a weight management perspective. Low-dose glucagon promotes hepatic glucose output and, therefore, could potentially be a beneficial alternative to carbohydrates for individuals trying to maintain or decrease body weight. In a recent event-driven prospective survey among individuals with type 1 diabetes, most of the respondents reported that management of nonsevere hypoglycemia using low-dose glucagon would be a valuable alternative to carbohydrate consumption (30).

Use of low-dose glucagon for treatment of nonsevere hypoglycemia has been explored in previous studies using other glucagon preparations. Low doses of powder-formulation native glucagon (GlucaGen, Novo Nordisk) have proved effective in restoring euglycemia (17), but because reconstitution with sterile water is required prior to injection, the application has been limited to research settings. Syringe-based administration of low-dose Xeris glucagon (Gvoke, Xeris Pharmaceutical, Chicago, IL)—a glucagon formulation dissolved in the organic solvent DMSO—has been evaluated in both inpatient and outpatient settings, demonstrating successful use for treatment of nonsevere hypoglycemia, but also a high occurrence of transient injection-site discomfort (16,31). No head-to-head studies of dasiglucagon and Xeris glucagon have been performed. However, data have demonstrated that Xeris glucagon has a prolonged time from administration to reaching PG targets (3–4 min) compared with reconstituted glucagon for treatment of severe hypoglycemia (32), which is not the case for dasiglucagon (23). Thus, the fast glucose-elevating properties of dasiglucagon demonstrated in this study cannot necessarily be generalized to other glucagon formulations.

The current study provides a number of new insights. It is the first study to evaluate the efficacy of injection-based, low-dose dasiglucagon for treatment of impending nonsevere hypoglycemia. Likewise, it is the first study to compare the efficacy to guideline-recommended oral glucose by mimicking real-life s.c. insulin overbolus scenarios in a controlled inpatient setting, enabling a precise pharmacodynamic comparison of the interventions. Lastly, the current study is the first trial evaluating low-dose glucagon efficacy to include MDI-treated individuals with type 1 diabetes—a population without access to the hypoglycemia-reducing benefits of glucose-responsive insulin delivery systems and, thus, a population in particular need of an alternative method to manage impending and manifested episodes of nonsevere hypoglycemia.

A constraint in the generalizability of this current study is that the interventions were evaluated under controlled investigational settings: preceding insulin dosing, physical activity levels, carbohydrate intake, alcohol consumption, and hypoglycemic events prior to study start were all controlled to minimize the influence of confounding factors, which does not fully reflect real-world situations. However, these precautionary circumstances have been key in achieving the most accurate pharmacodynamic comparison of the different interventions. Overall, participant characteristics were very similar to the general population of individuals with type 1 diabetes, although most of the participants had an HbA_1c below average. The PG level of intervention (4.5 mmol/L [81 mg/dL]) used in the study is arbitrary and was chosen to compare the hypoglycemia prevention abilities of the interventions; in real-life, the intervention threshold would vary between individuals and situations and include other factors such as the preceding rate of change in glucose. Only a few participants received CHO rescue treatment in the acute postintervention phase (as evidenced by the median time to rescue treatment), and thus, an argument could be made for selecting a lower PG intervention threshold. This would have produced more cases of hypoglycemia, more (and earlier) episodes of rescue carbohydrate administration, and more time in the hypoglycemic range, which—due to its rapid glucose-elevating profile—would be strongly expected to favor the efficacy of dasiglucagon over oral glucose even further. The small sample size could be regarded as a limitation of the study; however, the crossover design reduces variability and increases the power as each participant served as their own control. Although a larger sample size would have generated more robust data, this study was able to detect clinically meaningful differences in hypoglycemic levels and glucose-elevating profiles. Going forward, more questions need to be addressed, including the effect on longer-term glycemic control, weight management, user experiences and perceptions, cost-effectiveness, and long-term safety. Therefore, an outpatient study in free-living conditions is necessary to address and reflect the full potential of using low-dose dasiglucagon to prevent and treat insulin-induced hypoglycemia. A phase 2, randomized, two-period crossover outpatient study exploring the use of dasiglucagon in real-world conditions is currently ongoing (33).

In conclusion, this study demonstrated that low-dose dasiglucagon—a novel, stable, ready-to-use glucagon analog—safely and effectively prevented s.c. insulin-induced hypoglycemia with a faster glucose-elevating profile than oral glucose. The fast-acting noncaloric properties of dasiglucagon offer potential as an alternative to oral glucose for treatment of impending or manifest nonsevere hypoglycemia in people with type 1 diabetes.

Acknowledgments. The authors thank the study participants and acknowledge the laboratory assistance from Birgitte Roed (Steno Diabetes Center Copenhagen, Gentofte, Denmark) and the analysis of insulin samples by Kirsten Piepgras Neergaard, Susanne Zederkopff, and Charlotte Gade Farcinsen Leth (Steno Diabetes Center Copenhagen, Gentofte, Denmark).

Funding and Duality of Interest. Steno Diabetes Center Copenhagen is the sponsor of this investigator-initiated trial. Zealand Pharma provided financial support to the conduct of the study, supplied the study medication, and analyzed the plasma dasiglucagon concentration of the samples. S.S. reports advisory board attendance for Medtronic Diabetes. K.N. serves as an adviser to Medtronic, Abbott, and Novo Nordisk, owns shares in Novo Nordisk, has received research grants to the institution from Novo Nordisk, Zealand Pharma, Medtronic, and Roche, and has received fees for speaking from Medtronic, Roche, Zealand Pharma, Novo Nordisk, and Dexcom. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. C.L. conducted the study. C.L. wrote and edited the manuscript. C.L., A.G.R., S.S., and K.N. all contributed to designing the study. C.L., A.G.R., S.S., and K.N. interpreted the data. A.G.R. participated in conducting the study and performed the statistical analysis. A.G.R., S.S., and K.N. reviewed, edited, and approved the manuscript. C.L. and K.N. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Publication. Parts of this study were presented as an oral presentation at the 81st Scientific Sessions of the American Diabetes Association, virtual meeting, 25–29 June 2021.