CGM-A1c Genetics

CGM-A1C GENETICS

GWAS have reported over 200 hemoglobin A1c (HbA1c)-associated genetic variants, the majority of which affect HbA1c levels independently of glucose. Individuals who carry these nonglycemic HbA1c variants may have undetected hyperglycemia if HbA1c is used to diagnose diabetes or estimate glycemic control.

Study Design

Aims & Outcomes

Aims

Estimate the effects of nonglycemic HbA1c variants on HbA1c-glucose discordance and glycated proteins, by assessing the difference between measured and predicted HbA1c using 10–14 days of continuous glucose monitoring, plus albumin and fructosamine measurements.
Use recall-by-genotype for enriched recruitment of subjects with African or European ancestry.

Outcomes

HbA1c correlated well with CGM glucose across genotype groups, but the relationship differed by G6PD genotype, polygenic score, and ancestry.
HbA1c-glycemia discordance differed by polygenic score in European ancestry, and by G6PD carrier status in African ancestry.
HbA1c and glycated albumin were higher in African ancestry vs. European ancestry despite similar CGM glucose — suggesting glycation propensity may differ by ancestry.

Published Research

From a 200-variant discovery to a bedside regression equation.

This work traces one thread: a large genetic discovery, this team’s own recall-by-genotype study built to test it directly, and a more recent replication in an independent clinical trial cohort.

PLoS Medicine · 2017

Impact of Common Genetic Determinants of Hemoglobin A1c on Type 2 Diabetes Risk and Diagnosis in Ancestrally Diverse Populations: A Transethnic Genome-Wide Meta-analysis

Wheeler E, Leong A, Liu CT, Hivert MF, Strawbridge RJ, Podmore C, et al. PLoS Medicine. 2017;14(9):e1002383. doi:10.1371/journal.pmed.1002383

This large multi-ancestry meta-analysis is the foundational discovery behind this entire study: it identified over 200 genetic variants associated with HbA1c, and showed that most of them shift HbA1c levels through pathways unrelated to blood glucose itself — for example, by affecting red blood cell lifespan rather than glycemic control. That distinction matters clinically: a person carrying one of these “nonglycemic” variants can have a falsely reassuring HbA1c despite real, elevated blood sugar. This paper is what motivated a direct, prospective test of the idea using continuous glucose monitoring — which is exactly what this study does.

Illustrative summary of the discovery’s central point — proportions are schematic, not the paper’s exact reported split. See the paper for full results by ancestry and variant.

Diabetes (ADA Abstract) · 2023

Aggregate Effect of Hemoglobin A1c–Related Genetic Variants on A1c–Glycemia Discordance by Ancestry

Thangthaeng N, Facibene MN, Kartik S, Mandla R, Schroeder PH, Norgil NR, Mercader JM, Leong A. Diabetes. 2023;72(Suppl_1):194-LB. doi:10.2337/db23-194-LB

This is this team’s own study, presented as a conference abstract. The team recalled 177 patients from the Mass General Brigham Biobank, deliberately enriching recruitment for carriers of the African-ancestry G6PD variant, the alpha-thalassemia 3.7kb deletion, and people at the extremes of a European polygenic score built from glucose-independent HbA1c variants. Each participant wore a continuous glucose monitor for 14 days, letting the team directly compare each person’s lab HbA1c to what their real glucose exposure would predict.

Co-authored by current team members Nopporn Thangthaeng and Ravi Mandla; former team members Mariella Facibene, Saadhvi Kartik, Philip Schroeder, MS, and Nara Norgil; and collaborator Josep Mercader, PhD.

Illustrative summary of the discordance pattern by recruitment group — direction and rough scale only, not the poster’s actual measured values. See the abstract for full statistics.

Diabetes (ADA Abstract) · 2025

The Impact of the African-Ancestry–Specific G6PD Variant on the Relationship between HbA1c and Glucose

Leong A, Tripputi M, Szczerbinski L, Rosin S, Kretowski AJ, Mercader JM, Younes N, Li JH, Florez JC, for the GRADE Research Group. Diabetes. 2025;74(Suppl_1):1332-P. doi:10.2337/db25-1332-P

Using the independent GRADE Study cohort — 1,287 people with type 2 diabetes, including 303 non-Hispanic Black participants who underwent 10 days of continuous glucose monitoring plus HbA1c and glycated albumin testing — this more recent analysis replicates the core finding in a much larger, independently recruited sample: the relationship between average glucose and HbA1c differs by G6PD genotype and race. The team concludes that genotype-specific regression equations, or use of glycated albumin and CGM directly, could reduce undetected hyperglycemia in diverse populations.

Co-authored with collaborator Josep Mercader, PhD, and the GRADE Research Group.

Illustrative summary of the genotype-specific regression finding — the two lines are schematic, not the abstract’s fitted equations. See the abstract for actual coefficients.