Diabetes Dialogue: Understanding MARD, Comparing Popular CGM Sensors

In this episode of Diabetes Dialogue: Technology, Therapeutics, & Real-World Perspectives, hosts Diana Isaacs, PharmD, an endocrine clinical pharmacist, director of Education and Training in Diabetes Technology, and codirector of Endocrine Disorders in Pregnancy at the Cleveland Clinic, and Natalie Bellini, DNP, program director of Diabetes Technology at University Hospitals Diabetes and Metabolic Care Center, tackle a pair of recent publications related to leveraging diabetes technology.

The first study, which is part of a special edition of Diabetes Technology and Therapeutics, provides a review of mean absolute relative difference as well as how and why understanding its limitations is important to optimizing the benefit of continuous glucose monitoring technology. AN appropriate topic as they lead into the second study of interest for the episode, which is a January 2024 publication pitting the Dexcom G7 against the FreeStyle Libre 3.

The Myth of MARD

Mean absolute relative difference or MARD, as its commonly referred to, is a numeric metric used to describe the accuracy of a glucose sensor at a single point in time. Coined more than 20 years ago in the context of continuous glucose monitor evaluations, MARD has become a central part of evaluating the potential of a glucose sensor in the last 2 decades, but many lack a thorough understanding of its use, particularly the limitations of MARD.¹

MARD, its uses, and its limitations were the subject of a recent publication in Diabetes Technology & Therapeutics titled “The Myth of MARD (Mean Absolute Relative Difference): Limitations of MARD in the Clinical Assessment of Continuous Glucose Monitoring Data”. Penned by Robert Vigersky, MD, chief medical officer of Medtronic Diabetes, and John Shin, PhD, MBA, senior director of clinical research at Medtronic, the review details how MARD can vary based on several factors, including type of diabetes and age, site of sensor wear, and the percentage of collected values in each glycemic range during the study.¹

Comparing MARD with Dexcom G7 and FreeStyle Libre 3

In January 2024, a publication in the Journal of Diabetes Science and Technology attracted the attention of the medical community, as it compared the point accuracy of the Dexcom G7 and the Freestyle Libre 3 from Abbott. A single-arm, prospective, multicenter trial, the study enrolled 55 adult patients with type 1 or type 2 diabetes using one of the aforementioned CGM systems. Outcomes of interest for the trial included differences in MARD, number and percentage of matched glucose pairs within ±20mg/dL/ of reference values within glucose ranges: less than 54, 54 to 69, 70 to 180, 181 to 250, greater than 250 mg/dL, and combined.²

Results of this study suggested users of the FreeStyle Libre 3 sensor experienced significantly lower MARD relative to their counterparts using the Dexcom G7 sensor, with further analysis suggesting these patients also experienced a greater percentage of glucose values within t±20 mg/dL/±20 of reference (91.4% vs 78.6%). Study investigators pointed out MARD values for both sensors were similar during the first 12 hours, but MARD with FreeStyle Libre 3 was lower than the Dexcom G7 MARD during the next 12 hours (10.0% vs 15.1%, respectively, P < .0001) and throughout the study period.²

Relevant disclosures for Dr. Isaacs include Eli Lilly and Company, Novo Nordisk, Sanofi, Abbott Diabetes Care, Dexcom, Medtronic, and others. Relevant disclosures for Dr. Bellini include Abbott Diabetes Care, MannKind, Provention Bio, and others.

References:

1. _{^{Vigersky RA, Shin J. The Myth of MARD (Mean Absolute Relative Difference): Limitations of MARD in the Clinical Assessment of Continuous Glucose Monitoring Data. Diabetes Technol Ther. 2024;26(S3):38-44. doi:10.1089/dia.2023.0435}}
2. _{^{Hanson K, Kipnes M, Tran H. Comparison of Point Accuracy Between Two Widely Used Continuous Glucose Monitoring Systems. J Diabetes Sci Technol. Published online January 8, 2024. doi:10.1177/19322968231225676}}