The Need to Change Regulatory Evaluation of Hypoglycemia in Trials of Diabetes Treatments

Introduction

Diabetes is the most challenging of all hormonal deficiencies to manage because the deficit hormone even when administered with best of care practices and technologies carries substantial risk of hypoglycemia. Treating hyperglycemia while at the same time minimizing the risk of hypoglycemia is an important part of managing diabetes. Hypoglycemia itself and the fear of hypoglycemia prevent the early use of insulin therapy in type 2 diabetes (T2D) and is the major impediment to achieving good glycemic control and avoiding diabetic complication in all people who require insulin therapy. On average, people with longer standing type 1 diabetes (T1D) experience each year about two events of severe hypoglycemia (SH). Up to one in ten patients with T1D will die of hypoglycemia. Hypoglycemia risk in general is highly predictive of mortality in people with T1D. The prevalence estimates for hypoglycemia in T2D range from 1% to 3%.¹ Hypoglycemia risk therefore must be a key clinical outcome for any trial that evaluates insulin products, agents with the risk of hypoglycemia, and products that are specifically intended to reduce hypoglycemia risk. Reduction of hypoglycemia risk is also the primary benefit of technologies aimed at managing diabetes, such as insulin pumps and glucose monitoring systems.

The Costs of Hypoglycemia

Over 100,000 emergency department visits occur annually for SH.² These episodes result in both clinical morbidity and an economic burden on society.³ In the United States, patients with T2D who have experienced severe hypoglycemia, compared to those who have not experienced SH, have greater impairment of activity and work productivity, greater utilization of healthcare resources, and higher healthcare costs.⁴ Furthermore, mean direct costs of severe hypoglycemic events requiring medical assistance have been reported to be $1,161 per episode and mean indirect costs have been estimated to be as high as $579 per episode (2012 US dollars).⁵ A real-world evidence claims review over five years (2010-2015) of 66,179 patients using basal insulin and 81,876 patients using basal-bolus insulin revealed that 1.1% (basal) and 3.2% (basal-bolus) experienced at least one hospitalization because of SH.⁶

The Need for a New Regulatory Paradigm

The development of regulated products that reduce the risk of hypoglycemia should be encouraged and facilitated. A change in clinical trial and regulatory approaches for demonstrating hypoglycemia benefits of these products is long overdue.

The Diabetes Control and Complications Trial (DCCT) established the definition of SH, and FDA has longed used that definition for the sole endpoint that can support a claim for hypoglycemia risk reduction in a diabetes drug product label. Severe hypoglycemia is an infrequent event, and therefore, requires a very large trial to adequately power. Such trials are not generally feasible until after initial approval. Degludec (Tresiba™, Novo) carries the only mention of reduced hypoglycemia risk in an FDA-approved diabetes drug product label. This claim is based on the result of DEVOTE, which was a treat-to-target, event-driven trial of 7,637 T2D patients randomized to either TRESIBA or insulin glargine (Lantus™, Sanofi). FDA allowed this hypoglycemia benefit claim based on a difference in SH events. Only trials of this very large size would be sufficiently powered to compare SH events.

FDA’s position on the definition of an endpoint to show hypoglycemia benefit in clinical trials has not changed since the DCCT, despite several published recommendations starting with the American Diabetes Association (ADA) Workgroup position paper in 2005. In 2017, the International Hypoglycemia Study Group (IHSG) proposed a revised classification of low glucose in diabetes to include three levels. They were level 1: a glucose alert value of 70 mg/dL (3.9 mmol/L) or less; level 2: a glucose level of 54 mg/dL (3.0 mmol/L) or less; and level 3: SH, defined by severe cognitive impairment requiring external assistance for recovery. They considered serious, clinically important hypoglycemia to be a level 2 or level 3 glucose reading.⁷ In a recent report, the IHSG level 2 definition of serious hypoglycemia was recently validated in a series of clinical trials, demonstrating its ability to discriminate between the hypoglycemia risk of basal insulins potentially more accurate than the prior widely accepted definition of hypoglycemia developed by the ADA in 2005. Escalation from using the ADA 2005 definition of SH to the ISHG level 2 definition of hypoglycemia to the ISHG, level 3 definition of SH for the SWITCH 2 trial resulted in a progressively lower rate ratio for the two insulins tested, but a progressively wider confidence interval because the number of subjects reaching level 3 hypoglycemia was very small. The optimal definition of hypoglycemia for distinguishing outcomes between the two insulins was therefore concluded to be located at level 2 or less rate ratio.⁸ Similar results were noted for DEVOTE data, even though the hypoglycemic endpoints in this trial were different from those of the SWITCH trial.⁸

While the DCCT gold standard definition of SH was necessary in the era before glucose measurement technology was available, this should no longer be the sole definition of meaningful hypoglycemia events for regulatory purposes. Continuous glucose monitoring (CGM) technology has matured to the point that it provides a reliable, meaningful, and more robust means of evaluating hypoglycemia risk than the DCCT definition of the 1990s. Continuous glucose monitoring enables hypoglycemia to be practically evaluated in smaller phase 2 trials, which can help to encourage and optimize further development of products that treat or help to manage diabetes.

Therapies Can Reduce Hypoglycemia Risk

The importance of improving the evaluation of hypoglycemia is not limited to demonstrating small differences in hypoglycemia risk between competing insulin products. The development and commercialization of T1D disease modifying therapies will depend on demonstrating clinical benefits of intervening against diabetic autoimmunity. FDA has stated its acceptance of stimulated C-peptide as the primary efficacy endpoint for treatments that are aimed at preserving or restoring endogenous insulin secretion.⁹ Improving hypoglycemia risk is both the key secondary efficacy endpoint for those trials, and the main clinical value proposition for T1D modifying therapies.¹⁰ The challenges of T1D intervention trials preclude the large patient numbers that would be required to power a trial to show a benefit on SH. Continuous glucose monitoring can allow hypoglycemia benefits to be demonstrated in T1D trials of feasible size. These CGM data will help patients, prescribers, payers, and regulators to understand the value of these treatments.

Recommendations

Based on reasoning summarized by IHSG and others,^8,11 we recommend that (1) a numerical cut point of <54 mg/dL (<3.0 mmol/L) be accepted by regulatory agencies as a clinically relevant level with which to define meaningful hypoglycemia for trials of diabetes therapies; and (2) more frequent and standardized utilization of CGM in clinical trials for purposes of regulatory evaluation and clinical decision-making.¹¹